Backmerge tag 'v5.13-rc7' into drm-next

Backmerge Linux 5.13-rc7 to make some pulls from later bases apply, and to bake in the conflicts so far.

Dave Airlie committed 5 years ago

Commit f45fbbb2 parents d472b36 13311e71141 files · +13286 −5752Browse files

Unified · Split

1141 files

.clang-format +11 −2
.mailmap +3 −0
Documentation/admin-guide/sysctl/kernel.rst +14 −3
Documentation/devicetree/bindings/connector/usb-connector.yaml +15 −0
Documentation/devicetree/bindings/hwmon/ti,ads7828.yaml +1 −1
Documentation/devicetree/bindings/i2c/i2c-mpc.yaml +7 −0
Documentation/devicetree/bindings/media/renesas,drif.yaml +1 −3
Documentation/devicetree/bindings/net/renesas,ether.yaml +1 −1
Documentation/devicetree/bindings/sound/amlogic,gx-sound-card.yaml +2 −2
Documentation/devicetree/bindings/spi/spi-mux.yaml +1 −1
Documentation/driver-api/usb/usb.rst +2 −0
Documentation/riscv/vm-layout.rst +2 −2
Documentation/userspace-api/seccomp_filter.rst +8 −8
Documentation/virt/kvm/mmu.rst +2 −2
Documentation/virt/kvm/vcpu-requests.rst +5 −3
Documentation/vm/slub.rst +5 −5
MAINTAINERS +40 −12
Makefile +10 −2
arch/arc/include/uapi/asm/sigcontext.h +1 −0
arch/arc/kernel/signal.c +43 −0
arch/arc/kernel/vmlinux.lds.S +1 −1
arch/arm/boot/dts/imx6dl-yapp4-common.dtsi +5 −1
arch/arm/boot/dts/imx6q-dhcom-som.dtsi +12 −0
arch/arm/boot/dts/imx6qdl-emcon-avari.dtsi +1 −1
arch/arm/boot/dts/imx7d-meerkat96.dts +1 −1
arch/arm/boot/dts/imx7d-pico.dtsi +1 −1
arch/arm/include/asm/cpuidle.h +3 −2
arch/arm/mach-imx/pm-imx27.c +1 −0
arch/arm/mach-omap1/board-ams-delta.c +0 −14
arch/arm/mach-omap1/board-h2.c +3 −1
arch/arm/mach-omap1/pm.c +7 −3
arch/arm/mach-omap2/board-n8x0.c +1 −1
arch/arm64/Kbuild +1 −2
arch/arm64/Kconfig.platforms +1 −0
arch/arm64/boot/dts/freescale/fsl-ls1028a-kontron-sl28-var1.dts +2 −1
arch/arm64/boot/dts/freescale/fsl-ls1028a-kontron-sl28-var4.dts +2 −3
arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi +2 −2
arch/arm64/boot/dts/freescale/imx8mq-zii-ultra-rmb3.dts +5 −5
arch/arm64/boot/dts/freescale/imx8mq-zii-ultra.dtsi +7 −16
arch/arm64/boot/dts/ti/k3-am64-main.dtsi +6 −5
arch/arm64/boot/dts/ti/k3-am64-mcu.dtsi +2 −1
arch/arm64/boot/dts/ti/k3-am65-main.dtsi +6 −4
arch/arm64/boot/dts/ti/k3-am65-mcu.dtsi +2 −2
arch/arm64/boot/dts/ti/k3-am65-wakeup.dtsi +7 −6
arch/arm64/boot/dts/ti/k3-am654-base-board.dts +0 −31
arch/arm64/boot/dts/ti/k3-j7200-main.dtsi +6 −2
arch/arm64/boot/dts/ti/k3-j7200-mcu-wakeup.dtsi +4 −3
arch/arm64/boot/dts/ti/k3-j721e-main.dtsi +6 −4
arch/arm64/boot/dts/ti/k3-j721e-mcu-wakeup.dtsi +6 −5
arch/arm64/include/asm/barrier.h +1 −1
arch/arm64/include/asm/kvm_asm.h +3 −0
arch/arm64/include/asm/kvm_emulate.h +5 −0
arch/arm64/kvm/arm.c +17 −3
arch/arm64/kvm/hyp/exception.c +17 −1
arch/arm64/kvm/hyp/include/hyp/adjust_pc.h +0 −18
arch/arm64/kvm/hyp/nvhe/hyp-main.c +8 −0
arch/arm64/kvm/hyp/nvhe/mem_protect.c +2 −2
arch/arm64/kvm/hyp/nvhe/setup.c +1 −1
arch/arm64/kvm/hyp/nvhe/switch.c +1 −2
arch/arm64/kvm/hyp/vhe/switch.c +1 −2
arch/arm64/kvm/mmu.c +6 −6
arch/arm64/kvm/reset.c +24 −4
arch/arm64/kvm/sys_regs.c +21 −21
arch/arm64/mm/mmu.c +2 −1
arch/mips/alchemy/board-xxs1500.c +1 −0
arch/mips/include/asm/mips-boards/launch.h +5 −0
arch/mips/lib/mips-atomic.c +6 −6
arch/mips/mm/cache.c +14 −16
arch/mips/ralink/of.c +2 −0
arch/powerpc/boot/dts/fsl/p1010si-post.dtsi +8 −0
arch/powerpc/boot/dts/fsl/p2041si-post.dtsi +16 −0
arch/powerpc/include/asm/jump_label.h +1 −1
arch/powerpc/include/asm/kvm_host.h +1 −0
arch/powerpc/include/asm/pte-walk.h +29 −0
arch/powerpc/kernel/eeh.c +1 −22
arch/powerpc/kernel/io-workarounds.c +3 −13
arch/powerpc/kernel/iommu.c +5 −6
arch/powerpc/kernel/kprobes.c +2 −2
arch/powerpc/kernel/signal_64.c +4 −5
arch/powerpc/kvm/book3s_hv.c +1 −2
arch/powerpc/kvm/book3s_hv_rm_mmu.c +2 −13
arch/powerpc/kvm/book3s_hv_rmhandlers.S +7 −0
arch/powerpc/mm/mem.c +1 −0
arch/powerpc/perf/core-book3s.c +1 −1
arch/riscv/Kconfig +10 −9
arch/riscv/Kconfig.socs +1 −0
arch/riscv/Makefile +10 −1
arch/riscv/boot/dts/microchip/Makefile +1 −0
arch/riscv/boot/dts/sifive/Makefile +1 −0
arch/riscv/boot/dts/sifive/fu740-c000.dtsi +1 −1
arch/riscv/errata/sifive/Makefile +1 −1
arch/riscv/include/asm/alternative-macros.h +2 −2
arch/riscv/include/asm/kexec.h +2 −2
arch/riscv/include/asm/pgtable.h +2 −3
arch/riscv/kernel/machine_kexec.c +6 −5
arch/riscv/kernel/probes/kprobes.c +2 −0
arch/riscv/kernel/setup.c +2 −2
arch/riscv/kernel/stacktrace.c +7 −7
arch/riscv/kernel/traps.c +9 −4
arch/riscv/kernel/vmlinux-xip.lds.S +14 −1
arch/riscv/mm/init.c +6 −2
arch/riscv/mm/kasan_init.c +5 −5
arch/s390/kernel/entry.S +2 −2
arch/x86/Makefile +3 −2
arch/x86/events/intel/uncore_snbep.c +6 −3
arch/x86/include/asm/apic.h +1 −0
arch/x86/include/asm/disabled-features.h +2 −5
arch/x86/include/asm/fpu/api.h +1 −5
arch/x86/include/asm/fpu/internal.h +10 −10
arch/x86/include/asm/kvm-x86-ops.h +1 −0
arch/x86/include/asm/kvm_host.h +1 −0
arch/x86/include/asm/thermal.h +3 −1
arch/x86/kernel/alternative.c +46 −18
arch/x86/kernel/apic/apic.c +1 −0
arch/x86/kernel/apic/vector.c +20 −0
arch/x86/kernel/cpu/perfctr-watchdog.c +2 −2
arch/x86/kernel/cpu/sgx/virt.c +1 −0
arch/x86/kernel/fpu/signal.c +35 −19
arch/x86/kernel/fpu/xstate.c +0 −57
arch/x86/kernel/setup.c +30 −14
arch/x86/kvm/cpuid.c +1 −0
arch/x86/kvm/emulate.c +3 −2
arch/x86/kvm/hyperv.c +8 −0
arch/x86/kvm/kvm_emulate.h +1 −2
arch/x86/kvm/lapic.c +26 −10
arch/x86/kvm/mmu/mmu.c +25 −1
arch/x86/kvm/mmu/paging_tmpl.h +9 −5
arch/x86/kvm/mmu/tdp_mmu.c +3 −3
arch/x86/kvm/svm/avic.c +5 −7
arch/x86/kvm/svm/sev.c +17 −9
arch/x86/kvm/svm/svm.c +1 −3
arch/x86/kvm/svm/svm.h +1 −1
arch/x86/kvm/trace.h +3 −3
arch/x86/kvm/vmx/capabilities.h +1 −2
arch/x86/kvm/vmx/posted_intr.c +14 −0
arch/x86/kvm/vmx/posted_intr.h +1 −0
arch/x86/kvm/vmx/vmx.c +4 −3
arch/x86/kvm/x86.c +37 −15
arch/x86/mm/fault.c +2 −2
arch/x86/mm/ioremap.c +3 −1
arch/x86/mm/mem_encrypt_identity.c +6 −5
arch/x86/mm/numa.c +7 −1
arch/x86/pci/fixup.c +44 −0
arch/x86/platform/efi/quirks.c +12 −0
arch/x86/realmode/init.c +8 −6
crypto/async_tx/async_xor.c +2 −1
drivers/acpi/acpi_apd.c +1 −0
drivers/acpi/acpica/utdelete.c +8 −0
drivers/acpi/bus.c +8 −19
drivers/acpi/internal.h +2 −2
drivers/acpi/power.c +46 −13
drivers/acpi/scan.c +1 −1
drivers/acpi/sleep.c +2 −4
drivers/base/core.c +36 −32
drivers/base/memory.c +3 −3
drivers/block/loop.c +7 −18
drivers/block/loop.h +1 −0
drivers/bluetooth/btusb.c +23 −2
drivers/bus/mhi/pci_generic.c +38 −4
drivers/bus/ti-sysc.c +54 −6
drivers/cpufreq/Kconfig.arm +0 −10
drivers/cpufreq/cppc_cpufreq.c +12 −233
drivers/dma/Kconfig +2 −0
drivers/dma/fsl-dpaa2-qdma/dpaa2-qdma.c +3 −0
drivers/dma/idxd/cdev.c +1 −0
drivers/dma/idxd/init.c +62 −5
drivers/dma/ipu/ipu_irq.c +1 −1
drivers/dma/mediatek/mtk-uart-apdma.c +14 −13
drivers/dma/pl330.c +4 −2
drivers/dma/qcom/Kconfig +1 −0
drivers/dma/sf-pdma/Kconfig +1 −0
drivers/dma/sh/rcar-dmac.c +1 −1
drivers/dma/ste_dma40.c +3 −0
drivers/dma/stm32-mdma.c +2 −2
drivers/dma/xilinx/xilinx_dpdma.c +28 −3
drivers/dma/xilinx/zynqmp_dma.c +1 −1
drivers/firmware/efi/cper.c +1 −3
drivers/firmware/efi/fdtparams.c +3 −0
drivers/firmware/efi/libstub/file.c +1 −1
drivers/firmware/efi/memattr.c +0 −5
drivers/gpio/gpio-wcd934x.c +1 −1
drivers/gpu/drm/drm_auth.c +2 −1
drivers/gpu/drm/drm_ioctl.c +5 −4
drivers/gpu/drm/mcde/mcde_dsi.c +1 −1
drivers/gpu/drm/meson/meson_drv.c +5 −4
drivers/gpu/drm/msm/adreno/a6xx_gpu.c +114 −41
drivers/gpu/drm/msm/adreno/a6xx_gpu.h +1 −1
drivers/gpu/drm/msm/dsi/phy/dsi_phy_10nm.c +1 −0
drivers/gpu/drm/msm/dsi/phy/dsi_phy_7nm.c +1 −0
drivers/gpu/drm/msm/msm_gem.c +7 −0
drivers/gpu/drm/sun4i/sun8i_dw_hdmi.c +27 −4
drivers/gpu/drm/sun4i/sun8i_dw_hdmi.h +3 −2
drivers/gpu/drm/sun4i/sun8i_hdmi_phy.c +36 −5
drivers/gpu/drm/tegra/drm.h +1 −1
drivers/gpu/drm/tegra/hub.c +1 −1
drivers/gpu/drm/tegra/sor.c +42 −28
drivers/gpu/drm/ttm/ttm_bo.c +4 −1
drivers/gpu/drm/ttm/ttm_device.c +1 −7
drivers/gpu/drm/vc4/vc4_kms.c +1 −1
drivers/gpu/host1x/bus.c +24 −6
drivers/hid/Kconfig +17 −2
drivers/hid/Makefile +1 −0
drivers/hid/amd-sfh-hid/amd_sfh_client.c +10 −9
drivers/hid/amd-sfh-hid/amd_sfh_hid.c +0 −3
drivers/hid/hid-a4tech.c +2 −0
drivers/hid/hid-asus.c +20 −12
drivers/hid/hid-core.c +3 −1
drivers/hid/hid-debug.c +3 −0
drivers/hid/hid-ft260.c +16 −13
drivers/hid/hid-gt683r.c +1 −0
drivers/hid/hid-ids.h +7 −2
drivers/hid/hid-input.c +3 −0
drivers/hid/hid-logitech-hidpp.c +1 −0
drivers/hid/hid-magicmouse.c +5 −2
drivers/hid/hid-multitouch.c +37 −9
drivers/hid/hid-quirks.c +4 −0
drivers/hid/hid-semitek.c +40 −0
drivers/hid/hid-sensor-custom.c +5 −3
drivers/hid/hid-sensor-hub.c +9 −4
drivers/hid/hid-thrustmaster.c +1 −1
drivers/hid/i2c-hid/i2c-hid-core.c +10 −3
drivers/hid/intel-ish-hid/ipc/hw-ish.h +2 −0
drivers/hid/intel-ish-hid/ipc/pci-ish.c +2 −0
drivers/hid/surface-hid/surface_hid_core.c +3 −3
drivers/hid/usbhid/hid-core.c +1 −1
drivers/hid/usbhid/hid-pidff.c +1 −0
drivers/hwmon/corsair-psu.c +14 −0
drivers/hwmon/dell-smm-hwmon.c +2 −2
drivers/hwmon/pmbus/fsp-3y.c +25 −7
drivers/hwmon/pmbus/isl68137.c +2 −2
drivers/hwmon/pmbus/q54sj108a2.c +1 −1
drivers/hwmon/scpi-hwmon.c +9 −0
drivers/hwmon/tps23861.c +15 −2
drivers/i2c/busses/Kconfig +1 −1
drivers/i2c/busses/i2c-ali1563.c +1 −1
drivers/i2c/busses/i2c-altera.c +4 −5
drivers/i2c/busses/i2c-cadence.c +1 −1
drivers/i2c/busses/i2c-designware-master.c +1 −1
drivers/i2c/busses/i2c-eg20t.c +2 −1
drivers/i2c/busses/i2c-i801.c +2 −4
drivers/i2c/busses/i2c-icy.c +0 −1
drivers/i2c/busses/i2c-mpc.c +79 −2
drivers/i2c/busses/i2c-mt65xx.c +5 −0
drivers/i2c/busses/i2c-nomadik.c +1 −1
drivers/i2c/busses/i2c-ocores.c +4 −4
drivers/i2c/busses/i2c-pnx.c +4 −4
drivers/i2c/busses/i2c-qcom-geni.c +21 −2
drivers/i2c/busses/i2c-s3c2410.c +3 −0
drivers/i2c/busses/i2c-sh_mobile.c +1 −1
drivers/i2c/busses/i2c-st.c +2 −2
drivers/i2c/busses/i2c-stm32f4.c +1 −1
drivers/i2c/busses/i2c-tegra-bpmp.c +2 −2
drivers/i2c/muxes/i2c-arb-gpio-challenge.c +2 −2
drivers/iio/adc/ad7124.c +20 −16
drivers/iio/adc/ad7192.c +10 −9
drivers/iio/adc/ad7768-1.c +6 −2
drivers/iio/adc/ad7793.c +1 −0
drivers/iio/adc/ad7923.c +3 −1
drivers/iio/dac/ad5770r.c +11 −5
drivers/iio/gyro/fxas21002c_core.c +2 −0
drivers/infiniband/core/uverbs_cmd.c +5 −0
drivers/infiniband/hw/mlx4/main.c +1 −7
drivers/infiniband/hw/mlx5/cq.c +4 −5
drivers/infiniband/hw/mlx5/doorbell.c +6 −1
drivers/infiniband/hw/mlx5/fs.c +8 −3
drivers/infiniband/hw/mlx5/mr.c +4 −4
drivers/infiniband/ulp/ipoib/ipoib_netlink.c +1 −0
drivers/interconnect/qcom/bcm-voter.c +3 −1
drivers/iommu/amd/iommu.c +3 −1
drivers/iommu/intel/dmar.c +3 −1
drivers/iommu/intel/iommu.c +7 −2
drivers/iommu/intel/pasid.c +2 −1
drivers/iommu/virtio-iommu.c +1 −0
drivers/irqchip/irq-gic-v3.c +35 −1
drivers/isdn/hardware/mISDN/netjet.c +0 −1
drivers/md/bcache/bcache.h +0 −1
drivers/md/bcache/request.c +7 −13
drivers/md/bcache/stats.c +0 −14
drivers/md/bcache/stats.h +0 −1
drivers/md/bcache/sysfs.c +0 −4
drivers/md/dm-snap.c +2 −1
drivers/md/dm-verity-verify-sig.c +1 −1
drivers/md/raid5.c +0 −2
drivers/misc/cardreader/rtl8411.c +1 −0
drivers/misc/cardreader/rts5209.c +1 −0
drivers/misc/cardreader/rts5227.c +2 −0
drivers/misc/cardreader/rts5228.c +1 −0
drivers/misc/cardreader/rts5229.c +1 −0
drivers/misc/cardreader/rts5249.c +3 −0
drivers/misc/cardreader/rts5260.c +1 −0
drivers/misc/cardreader/rts5261.c +1 −0
drivers/misc/cardreader/rtsx_pcr.c +31 −13
drivers/misc/kgdbts.c +2 −1
drivers/misc/mei/interrupt.c +3 −0
drivers/mmc/host/renesas_sdhi_core.c +7 −2
drivers/mtd/nand/raw/cs553x_nand.c +11 −1
drivers/mtd/nand/raw/fsmc_nand.c +11 −1
drivers/mtd/nand/raw/lpc32xx_slc.c +14 −1
drivers/mtd/nand/raw/ndfc.c +11 −1
drivers/mtd/nand/raw/sharpsl.c +11 −1
drivers/mtd/nand/raw/tmio_nand.c +5 −3
drivers/mtd/nand/raw/txx9ndfmc.c +3 −2
drivers/mtd/parsers/ofpart_core.c +14 −12
drivers/net/appletalk/cops.c +2 −2
drivers/net/bonding/bond_main.c +1 −1
drivers/net/caif/caif_serial.c +1 −0
drivers/net/can/usb/mcba_usb.c +15 −2
drivers/net/dsa/bcm_sf2.c +2 −3
drivers/net/dsa/microchip/ksz9477.c +1 −0
drivers/net/dsa/mt7530.c +0 −8
drivers/net/dsa/ocelot/felix_vsc9959.c +10 −5
drivers/net/dsa/sja1105/sja1105_dynamic_config.c +19 −4
drivers/net/dsa/sja1105/sja1105_main.c +51 −23
drivers/net/ethernet/amazon/ena/ena_netdev.c +28 −26
drivers/net/ethernet/atheros/alx/main.c +1 −0
drivers/net/ethernet/broadcom/bnx2.c +1 −1
drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c +3 −1
drivers/net/ethernet/broadcom/bnxt/bnxt.c +117 −29
drivers/net/ethernet/broadcom/bnxt/bnxt.h +10 −0
drivers/net/ethernet/cadence/macb_main.c +3 −0
drivers/net/ethernet/chelsio/cxgb4/cxgb4.h +0 −2
drivers/net/ethernet/chelsio/cxgb4/cxgb4_ethtool.c +35 −13
drivers/net/ethernet/chelsio/cxgb4/cxgb4_filter.c +2 −2
drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c +3 −5
drivers/net/ethernet/chelsio/cxgb4/cxgb4_tc_flower.c +5 −9
drivers/net/ethernet/chelsio/cxgb4/cxgb4_tc_mqprio.c +6 −3
drivers/net/ethernet/chelsio/cxgb4/sge.c +6 −0
drivers/net/ethernet/chelsio/cxgb4/t4_hw.c +29 −17
drivers/net/ethernet/chelsio/inline_crypto/ch_ktls/chcr_ktls.c +79 −1
drivers/net/ethernet/chelsio/inline_crypto/ch_ktls/chcr_ktls.h +2 −0
drivers/net/ethernet/chelsio/inline_crypto/chtls/chtls_io.c +4 −2
drivers/net/ethernet/ec_bhf.c +3 −1
drivers/net/ethernet/emulex/benet/be_main.c +1 −0
drivers/net/ethernet/freescale/fec_main.c +19 −5
drivers/net/ethernet/freescale/fec_ptp.c +5 −3
drivers/net/ethernet/google/gve/gve_main.c +12 −9
drivers/net/ethernet/google/gve/gve_tx.c +7 −3
drivers/net/ethernet/hisilicon/hns/hns_ethtool.c +3 −3
drivers/net/ethernet/hisilicon/hns3/hns3_enet.c +51 −59
drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c +23 −41
drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_mbx.c +3 −1
drivers/net/ethernet/intel/i40e/i40e_txrx.c +6 −1
drivers/net/ethernet/intel/i40e/i40e_xsk.c +6 −2
drivers/net/ethernet/intel/ice/ice.h +5 −3
drivers/net/ethernet/intel/ice/ice_ethtool.c +6 −45
drivers/net/ethernet/intel/ice/ice_hw_autogen.h +1 −0
drivers/net/ethernet/intel/ice/ice_lib.c +22 −8
drivers/net/ethernet/intel/ice/ice_main.c +15 −0
drivers/net/ethernet/intel/ice/ice_txrx.c +13 −4
drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c +13 −6
drivers/net/ethernet/intel/ice/ice_xsk.c +9 −2
drivers/net/ethernet/intel/igb/igb.h +1 −1
drivers/net/ethernet/intel/igb/igb_main.c +32 −23
drivers/net/ethernet/intel/igb/igb_ptp.c +10 −13
drivers/net/ethernet/intel/igc/igc_main.c +5 −6
drivers/net/ethernet/intel/ixgbe/ixgbe_main.c +8 −8
drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c +7 −9
drivers/net/ethernet/intel/ixgbe/ixgbe_xsk.c +8 −6
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c +3 −0
drivers/net/ethernet/korina.c +6 −6
drivers/net/ethernet/lantiq_xrx200.c +11 −6
drivers/net/ethernet/marvell/mvpp2/mvpp2.h +22 −0
drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c +45 −9
drivers/net/ethernet/marvell/octeontx2/nic/otx2_ethtool.c +4 −0
drivers/net/ethernet/mediatek/mtk_eth_soc.c +52 −25
drivers/net/ethernet/mediatek/mtk_eth_soc.h +22 −2
drivers/net/ethernet/mellanox/mlx4/en_ethtool.c +1 −3
drivers/net/ethernet/mellanox/mlx4/fw.c +3 −0
drivers/net/ethernet/mellanox/mlx4/fw.h +1 −0
drivers/net/ethernet/mellanox/mlx4/main.c +6 −0
drivers/net/ethernet/mellanox/mlx4/port.c +103 −4
drivers/net/ethernet/mellanox/mlx5/core/dev.c +19 −0
drivers/net/ethernet/mellanox/mlx5/core/en/devlink.c +2 −0
drivers/net/ethernet/mellanox/mlx5/core/en/ptp.c +0 −1
drivers/net/ethernet/mellanox/mlx5/core/en/ptp.h +22 −0
drivers/net/ethernet/mellanox/mlx5/core/en/rep/bond.c +2 −0
drivers/net/ethernet/mellanox/mlx5/core/en/rep/neigh.c +6 −9
drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.c +2 −6
drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c +31 −4
drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c +0 −3
drivers/net/ethernet/mellanox/mlx5/core/en_arfs.c +1 −1
drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c +10 −2
drivers/net/ethernet/mellanox/mlx5/core/en_fs.c +1 −0
drivers/net/ethernet/mellanox/mlx5/core/en_main.c +81 −33
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +32 −7
drivers/net/ethernet/mellanox/mlx5/core/en_tc.h +3 −0
drivers/net/ethernet/mellanox/mlx5/core/en_tx.c +3 −22
drivers/net/ethernet/mellanox/mlx5/core/eq.c +4 −2
drivers/net/ethernet/mellanox/mlx5/core/eswitch.c +7 −0
drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +2 −1
drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_termtbl.c +26 −35
drivers/net/ethernet/mellanox/mlx5/core/fw_reset.c +3 −0
drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c +6 −0
drivers/net/ethernet/mellanox/mlx5/core/lib/fs_chains.c +1 −1
drivers/net/ethernet/mellanox/mlx5/core/lib/fs_chains.h +5 −0
drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.c +3 −0
drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.h +1 −4
drivers/net/ethernet/mellanox/mlx5/core/main.c +7 −7
drivers/net/ethernet/mellanox/mlx5/core/mr.c +1 −1
drivers/net/ethernet/mellanox/mlx5/core/pci_irq.c +17 −5
drivers/net/ethernet/mellanox/mlx5/core/rdma.c +3 −0
drivers/net/ethernet/mellanox/mlx5/core/sf/dev/dev.c +1 −0
drivers/net/ethernet/mellanox/mlx5/core/sf/devlink.c +11 −7
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_fw.c +2 −1
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste_v1.c +16 −10
drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5dr.h +5 −4
drivers/net/ethernet/mellanox/mlx5/core/transobj.c +24 −6
drivers/net/ethernet/mellanox/mlx5/core/vport.c +0 −2
drivers/net/ethernet/mellanox/mlxsw/core_thermal.c +4 −2
drivers/net/ethernet/mellanox/mlxsw/reg.h +1 −1
drivers/net/ethernet/mellanox/mlxsw/spectrum_qdisc.c +4 −1
drivers/net/ethernet/microchip/encx24j600.c +1 −1
drivers/net/ethernet/microchip/encx24j600_hw.h +1 −1
drivers/net/ethernet/mscc/ocelot.c +5 −0
drivers/net/ethernet/myricom/myri10ge/myri10ge.c +1 −0
drivers/net/ethernet/pensando/Kconfig +1 −0
drivers/net/ethernet/qlogic/netxen/netxen_nic_main.c +2 −0
drivers/net/ethernet/qlogic/qed/qed_dcbx.c +3 −1
drivers/net/ethernet/qlogic/qla3xxx.c +1 −1
drivers/net/ethernet/qlogic/qlcnic/qlcnic_main.c +1 −0
drivers/net/ethernet/qualcomm/rmnet/rmnet_vnd.c +9 −9
drivers/net/ethernet/realtek/r8169_main.c +1 −1
drivers/net/ethernet/renesas/sh_eth.c +1 −1
drivers/net/ethernet/sfc/nic.c +1 −0
drivers/net/ethernet/stmicro/stmmac/dwmac1000.h +4 −4
drivers/net/ethernet/stmicro/stmmac/stmmac_main.c +25 −13
drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c +2 −0
drivers/net/ethernet/ti/netcp_core.c +2 −2
drivers/net/ethernet/xilinx/ll_temac_main.c +24 −3
drivers/net/hamradio/mkiss.c +1 −0
drivers/net/ieee802154/mrf24j40.c +2 −2
drivers/net/ipa/ipa.h +2 −0
drivers/net/ipa/ipa_mem.c +2 −1
drivers/net/mdio/mdio-octeon.c +0 −2
drivers/net/mdio/mdio-thunder.c +0 −1
drivers/net/mhi/net.c +1 −1
drivers/net/phy/dp83867.c +1 −5
drivers/net/phy/mdio_bus.c +2 −1
drivers/net/usb/cdc_eem.c +2 −2
drivers/net/usb/cdc_ncm.c +1 −1
drivers/net/usb/hso.c +22 −23
drivers/net/usb/lan78xx.c +1 −0
drivers/net/usb/qmi_wwan.c +1 −1
drivers/net/usb/r8152.c +38 −6
drivers/net/usb/smsc75xx.c +9 −3
drivers/net/virtio_net.c +8 −12
drivers/net/vrf.c +2 −4
drivers/net/wireguard/Makefile +1 −2
drivers/net/wireguard/allowedips.c +99 −90
drivers/net/wireguard/allowedips.h +7 −7
drivers/net/wireguard/main.c +16 −1
drivers/net/wireguard/peer.c +20 −7
drivers/net/wireguard/peer.h +3 −0
drivers/net/wireguard/selftest/allowedips.c +79 −86
drivers/net/wireguard/socket.c +1 −1
drivers/net/wireless/ath/ath10k/htt.h +1 −0
drivers/net/wireless/ath/ath10k/htt_rx.c +191 −10
drivers/net/wireless/ath/ath10k/rx_desc.h +13 −1
drivers/net/wireless/ath/ath11k/dp_rx.c +34 −0
drivers/net/wireless/ath/ath11k/dp_rx.h +1 −0
drivers/net/wireless/ath/ath11k/mac.c +6 −0
drivers/net/wireless/mac80211_hwsim.c +5 −0
drivers/net/wireless/mediatek/mt76/mac80211.c +26 −0
drivers/net/wireless/mediatek/mt76/mt7615/init.c +0 −1
drivers/net/wireless/mediatek/mt76/mt7615/mac.c +3 −2
drivers/net/wireless/mediatek/mt76/mt7615/sdio_mcu.c +12 −7
drivers/net/wireless/mediatek/mt76/mt7615/usb_mcu.c +0 −3
drivers/net/wireless/mediatek/mt76/mt76_connac_mcu.c +4 −0
drivers/net/wireless/mediatek/mt76/mt76x0/pci.c +77 −4
drivers/net/wireless/mediatek/mt76/mt7921/init.c +2 −2
drivers/net/wireless/mediatek/mt76/mt7921/mac.c +3 −2
drivers/net/wireless/mediatek/mt76/mt7921/main.c +1 −2
drivers/net/wireless/mediatek/mt76/mt7921/mcu.c +10 −7
drivers/net/xen-netback/interface.c +6 −0
drivers/nfc/nfcmrvl/fw_dnld.h +1 −1
drivers/nfc/nfcmrvl/i2c.c +1 −1
drivers/nfc/nfcmrvl/nfcmrvl.h +1 −1
drivers/nfc/nfcmrvl/spi.c +1 −1
drivers/nfc/nfcmrvl/uart.c +1 −1
drivers/nfc/nfcmrvl/usb.c +1 −1
drivers/nvme/host/Kconfig +2 −1
drivers/nvme/host/core.c +3 −1
drivers/nvme/host/fabrics.c +5 −0
drivers/nvme/host/fc.c +17 −8
drivers/nvme/host/rdma.c +3 −2
drivers/nvme/target/core.c +27 −21
drivers/nvme/target/loop.c +8 −3
drivers/nvme/target/nvmet.h +1 −1
drivers/nvme/target/tcp.c +1 −1
drivers/pci/controller/dwc/Makefile +2 −1
drivers/pci/controller/dwc/pcie-tegra194-acpi.c +108 −0
drivers/pci/controller/dwc/pcie-tegra194.c +18 −120
drivers/pci/controller/pci-aardvark.c +40 −9
drivers/pci/of.c +9 −0
drivers/pci/probe.c +2 −1
drivers/pci/quirks.c +92 −1
drivers/phy/broadcom/phy-brcm-usb-init.h +2 −2
drivers/phy/cadence/phy-cadence-sierra.c +1 −0
drivers/phy/mediatek/phy-mtk-tphy.c +2 −0
drivers/phy/microchip/sparx5_serdes.c +4 −0
drivers/phy/ralink/phy-mt7621-pci.c +1 −1
drivers/phy/ti/phy-j721e-wiz.c +1 −0
drivers/pinctrl/aspeed/pinctrl-aspeed-g5.c +2 −2
drivers/pinctrl/aspeed/pinctrl-aspeed-g6.c +2 −2
drivers/pinctrl/aspeed/pinctrl-aspeed.c +2 −1
drivers/pinctrl/aspeed/pinmux-aspeed.c +2 −1
drivers/pinctrl/qcom/Kconfig +1 −1
drivers/pinctrl/qcom/pinctrl-sdx55.c +9 −9
drivers/pinctrl/ralink/pinctrl-rt2880.c +1 −1
drivers/platform/mellanox/mlxreg-hotplug.c +2 −2
drivers/platform/surface/aggregator/controller.c +1 −1
drivers/platform/surface/surface_aggregator_registry.c +5 −2
drivers/platform/surface/surface_dtx.c +1 −0
drivers/platform/x86/thinkpad_acpi.c +1 −0
drivers/ptp/ptp_clock.c +3 −3
drivers/ptp/ptp_ocp.c +3 −1
drivers/regulator/Kconfig +1 −1
drivers/regulator/atc260x-regulator.c +10 −9
drivers/regulator/bd718x7-regulator.c +1 −1
drivers/regulator/core.c +6 −0
drivers/regulator/cros-ec-regulator.c +2 −1
drivers/regulator/da9121-regulator.c +7 −3
drivers/regulator/fan53555.c +1 −2
drivers/regulator/fan53880.c +3 −0
drivers/regulator/fixed.c +6 −1
drivers/regulator/helpers.c +1 −1
drivers/regulator/hi6421v600-regulator.c +2 −2
drivers/regulator/hi655x-regulator.c +1 −1
drivers/regulator/max77620-regulator.c +11 −6
drivers/regulator/mt6315-regulator.c +1 −1
drivers/regulator/rt4801-regulator.c +2 −2
drivers/regulator/rtmv20-regulator.c +42 −2
drivers/regulator/scmi-regulator.c +1 −1
drivers/s390/block/dasd_diag.c +7 −1
drivers/s390/block/dasd_fba.c +7 −1
drivers/s390/block/dasd_int.h +0 −1
drivers/s390/cio/vfio_ccw_cp.c +4 −0
drivers/s390/cio/vfio_ccw_drv.c +10 −2
drivers/s390/cio/vfio_ccw_fsm.c +1 −0
drivers/s390/cio/vfio_ccw_ops.c +0 −2
drivers/s390/crypto/ap_queue.c +9 −2
drivers/scsi/aic7xxx/aicasm/aicasm_gram.y +0 −1
drivers/scsi/aic7xxx/aicasm/aicasm_symbol.h +1 −1
drivers/scsi/aic7xxx/scsi_message.h +11 −0
drivers/scsi/bnx2fc/bnx2fc_io.c +1 −0
drivers/scsi/hisi_sas/hisi_sas_v3_hw.c +4 −4
drivers/scsi/hosts.c +26 −21
drivers/scsi/libsas/sas_port.c +2 −2
drivers/scsi/lpfc/lpfc_sli.c +1 −3
drivers/scsi/qedf/qedf_main.c +9 −11
drivers/scsi/qla2xxx/qla_target.c +2 −0
drivers/scsi/scsi_devinfo.c +1 −0
drivers/scsi/ufs/ufs-mediatek.c +14 −1
drivers/scsi/vmw_pvscsi.c +7 −1
drivers/soc/amlogic/meson-clk-measure.c +1 −3
drivers/soundwire/qcom.c +10 −2
drivers/spi/Kconfig +1 −1
drivers/spi/spi-bcm2835.c +8 −2
drivers/spi/spi-bitbang.c +14 −4
drivers/spi/spi-fsl-dspi.c +3 −1
drivers/spi/spi-fsl-spi.c +4 −0
drivers/spi/spi-omap-uwire.c +8 −1
drivers/spi/spi-omap2-mcspi.c +20 −13
drivers/spi/spi-pxa2xx.c +8 −1
drivers/spi/spi-sc18is602.c +8 −1
drivers/spi/spi-sprd.c +1 −0
drivers/spi/spi-stm32-qspi.c +4 −1
drivers/spi/spi-zynq-qspi.c +8 −8
drivers/spi/spi.c +38 −13
drivers/staging/emxx_udc/emxx_udc.c +2 −2
drivers/staging/iio/cdc/ad7746.c +0 −1
drivers/staging/ralink-gdma/ralink-gdma.c +0 −2
drivers/staging/rtl8723bs/os_dep/ioctl_cfg80211.c +3 −3
drivers/target/target_core_iblock.c +2 −2
drivers/target/target_core_transport.c +2 −4
drivers/target/target_core_user.c +7 −5
drivers/tee/optee/call.c +4 −2
drivers/tee/optee/optee_msg.h +4 −2
drivers/thermal/intel/int340x_thermal/int340x_thermal_zone.c +4 −0
drivers/thermal/intel/therm_throt.c +11 −4
drivers/thermal/intel/x86_pkg_temp_thermal.c +1 −1
drivers/thermal/qcom/qcom-spmi-adc-tm5.c +1 −1
drivers/thermal/ti-soc-thermal/ti-bandgap.c +1 −1
drivers/thunderbolt/dma_port.c +6 −5
drivers/thunderbolt/usb4.c +5 −4
drivers/tty/serial/8250/8250.h +17 −15
drivers/tty/serial/8250/8250_aspeed_vuart.c +1 −0
drivers/tty/serial/8250/8250_dw.c +1 −0
drivers/tty/serial/8250/8250_exar.c +5 −1
drivers/tty/serial/8250/8250_pci.c +35 −12
drivers/tty/serial/8250/8250_port.c +12 −0
drivers/tty/serial/rp2.c +17 −35
drivers/tty/serial/serial-tegra.c +1 −1
drivers/tty/serial/serial_core.c +5 −3
drivers/tty/serial/sh-sci.c +2 −2
drivers/usb/cdns3/cdns3-gadget.c +6 −6
drivers/usb/cdns3/cdnsp-gadget.c +7 −7
drivers/usb/cdns3/cdnsp-ring.c +4 −3
drivers/usb/chipidea/udc.c +1 −0
drivers/usb/chipidea/usbmisc_imx.c +14 −2
drivers/usb/core/devio.c +8 −3
drivers/usb/core/hub.c +7 −0
drivers/usb/dwc3/core.c +1 −7
drivers/usb/dwc3/debug.h +3 −0
drivers/usb/dwc3/debugfs.c +2 −19
drivers/usb/dwc3/dwc3-meson-g12a.c +10 −3
drivers/usb/dwc3/ep0.c +3 −0
drivers/usb/dwc3/gadget.c +19 −12
drivers/usb/gadget/config.c +8 −0
drivers/usb/gadget/function/f_ecm.c +1 −1
drivers/usb/gadget/function/f_eem.c +3 −3
drivers/usb/gadget/function/f_fs.c +3 −0
drivers/usb/gadget/function/f_hid.c +2 −1
drivers/usb/gadget/function/f_loopback.c +1 −1
drivers/usb/gadget/function/f_ncm.c +5 −5
drivers/usb/gadget/function/f_printer.c +2 −1
drivers/usb/gadget/function/f_rndis.c +1 −1
drivers/usb/gadget/function/f_serial.c +1 −1
drivers/usb/gadget/function/f_sourcesink.c +2 −1
drivers/usb/gadget/function/f_subset.c +1 −1
drivers/usb/gadget/function/f_tcm.c +2 −1
drivers/usb/gadget/udc/renesas_usb3.c +3 −2
drivers/usb/host/xhci-pci.c +6 −1
drivers/usb/host/xhci-ring.c +7 −7
drivers/usb/host/xhci.h +1 −0
drivers/usb/misc/brcmstb-usb-pinmap.c +2 −0
drivers/usb/misc/trancevibrator.c +2 −2
drivers/usb/misc/uss720.c +1 −0
drivers/usb/musb/musb_core.c +1 −2
drivers/usb/serial/cp210x.c +78 −6
drivers/usb/serial/ftdi_sio.c +4 −0
drivers/usb/serial/ftdi_sio_ids.h +8 −0
drivers/usb/serial/omninet.c +5 −3
drivers/usb/serial/option.c +4 −0
drivers/usb/serial/pl2303.c +1 −0
drivers/usb/serial/pl2303.h +1 −0
drivers/usb/serial/quatech2.c +3 −3
drivers/usb/serial/ti_usb_3410_5052.c +3 −0
drivers/usb/typec/mux.c +5 −4
drivers/usb/typec/mux/intel_pmc_mux.c +11 −4
drivers/usb/typec/tcpm/tcpm.c +107 −43
drivers/usb/typec/tcpm/wcove.c +1 −1
drivers/usb/typec/ucsi/ucsi.c +2 −1
drivers/vdpa/mlx5/net/mlx5_vnet.c +18 −1
drivers/vfio/pci/Kconfig +1 −0
drivers/vfio/pci/vfio_pci_config.c +1 −1
drivers/vfio/platform/vfio_platform_common.c +1 −1
drivers/vfio/vfio_iommu_type1.c +1 −1
drivers/video/fbdev/core/fb_defio.c +35 −0
drivers/video/fbdev/core/fbmem.c +4 −0
drivers/video/fbdev/hgafb.c +1 −1
fs/afs/cmservice.c +5 −0
fs/afs/dir.c +3 −1
fs/afs/fsclient.c +4 −0
fs/afs/main.c +2 −2
fs/afs/vlclient.c +1 −0
fs/afs/write.c +9 −6
fs/btrfs/block-group.c +4 −4
fs/btrfs/compression.c +12 −5
fs/btrfs/disk-io.c +18 −8
fs/btrfs/extent-tree.c +1 −1
fs/btrfs/file-item.c +81 −27
fs/btrfs/file.c +2 −2
fs/btrfs/inode.c +18 −1
fs/btrfs/reflink.c +22 −16
fs/btrfs/tree-log.c +29 −8
fs/btrfs/zoned.c +18 −5
fs/cifs/cifs_ioctl.h +19 −6
fs/cifs/cifspdu.h +2 −1
fs/cifs/ioctl.c +105 −38
fs/cifs/smb2pdu.c +7 −0
fs/cifs/trace.h +17 −12
fs/coredump.c +1 −1
fs/debugfs/file.c +1 −1
fs/debugfs/inode.c +6 −3
fs/ext4/extents.c +23 −20
fs/ext4/fast_commit.c +90 −80
fs/ext4/fast_commit.h +0 −19
fs/ext4/ialloc.c +4 −2
fs/ext4/mballoc.c +1 −1
fs/ext4/namei.c +4 −2
fs/ext4/super.c +9 −2
fs/ext4/sysfs.c +4 −0
fs/gfs2/file.c +4 −1
fs/gfs2/glock.c +24 −4
fs/gfs2/glops.c +1 −1
fs/gfs2/log.c +3 −3
fs/gfs2/log.h +1 −0
fs/gfs2/lops.c +6 −1
fs/gfs2/lops.h +1 −0
fs/gfs2/util.c +1 −0
fs/hugetlbfs/inode.c +1 −0
fs/io-wq.c +15 −14
fs/io-wq.h +1 −1
fs/io_uring.c +43 −15
fs/netfs/Kconfig +1 −1
fs/netfs/read_helper.c +1 −1
fs/nfs/client.c +1 −1
fs/nfs/filelayout/filelayout.c +1 −1
fs/nfs/namespace.c +1 −1
fs/nfs/nfs4_fs.h +1 −0
fs/nfs/nfs4client.c +1 −1
fs/nfs/nfs4file.c +1 −1
fs/nfs/nfs4proc.c +32 −9
fs/nfs/nfstrace.h +0 −4
fs/nfs/pagelist.c +6 −14
fs/nfs/pnfs.c +8 −9
fs/nfs/super.c +1 −1
fs/notify/fanotify/fanotify_user.c +26 −8
fs/notify/fdinfo.c +1 −1
fs/ocfs2/file.c +50 −5
fs/proc/base.c +13 −0
fs/xfs/libxfs/xfs_ag_resv.c +15 −3
fs/xfs/libxfs/xfs_bmap.c +0 −12
fs/xfs/libxfs/xfs_inode_buf.c +42 −4
fs/xfs/libxfs/xfs_trans_inode.c +17 −0
fs/xfs/xfs_inode.c +29 −0
fs/xfs/xfs_ioctl.c +34 −67
fs/xfs/xfs_message.h +2 −0
include/asm-generic/vmlinux.lds.h +1 −0
include/dt-bindings/usb/pd.h +76 −13
include/linux/arch_topology.h +0 −1
include/linux/avf/virtchnl.h +1 −0
include/linux/cgroup-defs.h +3 −3
include/linux/cgroup.h +1 −1
include/linux/compiler_attributes.h +1 −0
include/linux/device.h +2 −4
include/linux/entry-kvm.h +2 −1
include/linux/fanotify.h +4 −0
include/linux/fb.h +3 −0
include/linux/hid.h +1 −2
include/linux/host1x.h +24 −6
include/linux/huge_mm.h +7 −1
include/linux/hugetlb.h +8 −0
include/linux/init.h +2 −1
include/linux/kvm_host.h +16 −2
include/linux/mfd/rohm-bd70528.h +1 −3
include/linux/mfd/rohm-bd71828.h +5 −5
include/linux/mlx4/device.h +1 −0
include/linux/mlx5/driver.h +26 −22
include/linux/mlx5/mlx5_ifc.h +2 −0
include/linux/mlx5/mpfs.h +18 −0
include/linux/mlx5/transobj.h +1 −0
include/linux/mm.h +3 −0
include/linux/mm_types.h +20 −7
include/linux/pci.h +2 −0
include/linux/pgtable.h +8 −0
include/linux/phy.h +5 −0
include/linux/platform_data/ti-sysc.h +1 −0
include/linux/ptp_clock_kernel.h +1 −1
include/linux/rmap.h +1 −0
include/linux/rtsx_pci.h +2 −0
include/linux/sched.h +8 −0
include/linux/socket.h +0 −2
include/linux/spi/spi.h +2 −2
include/linux/sunrpc/xprt.h +2 −0
include/linux/swapops.h +11 −4
include/linux/tick.h +7 −0
include/linux/usb/pd.h +1 −1
include/linux/usb/pd_ext_sdb.h +0 −4
include/net/caif/caif_dev.h +1 −1
include/net/caif/cfcnfg.h +1 −1
include/net/caif/cfserl.h +1 −0
include/net/cfg80211.h +2 −2
include/net/mac80211.h +7 −2
include/net/net_namespace.h +13 −1
include/net/netfilter/nf_flow_table.h +0 −1
include/net/netfilter/nf_tables.h +0 −6
include/net/nfc/nci_core.h +1 −0
include/net/pkt_cls.h +11 −0
include/net/pkt_sched.h +1 −6
include/net/sch_generic.h +34 −1
include/net/sock.h +16 −5
include/net/tls.h +9 −1
include/sound/soc-dai.h +1 −1
include/uapi/asm-generic/unistd.h +1 −2
include/uapi/linux/in.h +3 −0
include/uapi/linux/input-event-codes.h +1 −0
include/uapi/linux/io_uring.h +10 −9
include/uapi/linux/kvm.h +3 −2
include/uapi/linux/virtio_ids.h +1 −1
init/Kconfig +1 −40
init/main.c +1 −1
kernel/bpf/Kconfig +89 −0
kernel/bpf/bpf_lsm.c +2 −0
kernel/bpf/btf.c +12 −0
kernel/bpf/helpers.c +24 −18
kernel/bpf/ringbuf.c +11 −13
kernel/bpf/syscall.c +2 −1
kernel/bpf/verifier.c +118 −44
kernel/cgroup/cgroup-v1.c +5 −1
kernel/cgroup/cgroup.c +9 −12
kernel/cgroup/cpuset.c +1 −1
kernel/cgroup/rdma.c +1 −1
kernel/cgroup/rstat.c +1 −1
kernel/crash_core.c +1 −0
kernel/entry/common.c +3 −2
kernel/events/core.c +2 −0
kernel/irq_work.c +0 −3
kernel/printk/printk_safe.c +1 −1
kernel/sched/core.c +0 −1
kernel/sched/debug.c +2 −1
kernel/sched/fair.c +42 −30
kernel/sched/pelt.h +1 −10
kernel/seccomp.c +16 −14
kernel/sysctl.c +24 −5
kernel/time/tick-sched.c +1 −0
kernel/trace/bpf_trace.c +12 −20
kernel/trace/ftrace.c +7 −1
kernel/trace/trace.c +1 −12
kernel/trace/trace_clock.c +3 −3
kernel/workqueue.c +10 −2
lib/crc64.c +1 −1
lib/percpu-refcount.c +3 −3
mm/debug_vm_pgtable.c +2 −2
mm/huge_memory.c +29 −27
mm/hugetlb.c +125 −26
mm/internal.h +39 −14
mm/kasan/init.c +2 −2
mm/kfence/core.c +3 −3
mm/memory-failure.c +33 −3
mm/memory.c +45 −0
mm/migrate.c +1 −0
mm/page_alloc.c +2 −0
mm/page_vma_mapped.c +17 −10
mm/pgtable-generic.c +2 −3
mm/rmap.c +27 −12
mm/slab_common.c +1 −2
mm/slub.c +16 −21
mm/sparse.c +11 −2
mm/swapfile.c +1 −1
mm/truncate.c +19 −24
net/Kconfig +0 −27
net/appletalk/aarp.c +1 −1
net/batman-adv/bat_iv_ogm.c +3 −1
net/bluetooth/hci_core.c +6 −1
net/bluetooth/hci_sock.c +2 −2
net/bluetooth/smp.c +3 −3
net/bridge/br_private.h +2 −2
net/bridge/br_vlan_tunnel.c +24 −14
net/caif/caif_dev.c +9 −4
net/caif/caif_usb.c +13 −1
net/caif/cfcnfg.c +11 −5
net/caif/cfserl.c +5 −0
net/can/bcm.c +49 −13
net/can/isotp.c +81 −29
net/can/j1939/transport.c +40 −14
net/can/raw.c +48 −14
net/compat.c +1 −1
net/core/dev.c +24 −5
net/core/devlink.c +2 −2
net/core/fib_rules.c +1 −1
net/core/filter.c +1 −0
net/core/neighbour.c +1 −0
net/core/net_namespace.c +13 −7
net/core/rtnetlink.c +4 −0
net/core/skbuff.c +3 −1
net/core/sock.c +16 −8
net/dsa/master.c +3 −2
net/dsa/slave.c +7 −5
net/dsa/tag_8021q.c +1 −1
net/ethtool/eeprom.c +1 −1
net/ethtool/ioctl.c +5 −5
net/ethtool/stats.c +1 −1
net/ethtool/strset.c +2 −0
net/hsr/hsr_device.c +2 −0
net/hsr/hsr_forward.c +21 −9
net/hsr/hsr_forward.h +4 −4
net/hsr/hsr_main.h +2 −2
net/hsr/hsr_slave.c +5 −6
net/ieee802154/nl-mac.c +6 −4
net/ieee802154/nl-phy.c +3 −1
net/ieee802154/nl802154.c +5 −4
net/ipv4/af_inet.c +2 −2
net/ipv4/bpf_tcp_ca.c +2 −0
net/ipv4/cipso_ipv4.c +1 −0
net/ipv4/devinet.c +1 −1
net/ipv4/icmp.c +7 −0
net/ipv4/igmp.c +1 −0
net/ipv4/ipconfig.c +8 −5
net/ipv4/ping.c +7 −5
net/ipv4/route.c +14 −1
net/ipv4/udp.c +10 −0
net/ipv6/addrconf.c +1 −1
net/ipv6/mcast.c +0 −3
net/ipv6/netfilter/nft_fib_ipv6.c +18 −4
net/ipv6/reassembly.c +3 −1
net/ipv6/route.c +6 −2
net/ipv6/sit.c +3 −0
net/ipv6/udp.c +3 −0
net/mac80211/debugfs.c +9 −2
net/mac80211/ieee80211_i.h +14 −24
net/mac80211/iface.c +15 −15
net/mac80211/key.c +7 −0
net/mac80211/key.h +2 −0
net/mac80211/main.c +6 −1
net/mac80211/mlme.c +8 −0
net/mac80211/rc80211_minstrel_ht.c +1 −1
net/mac80211/rx.c +119 −38
net/mac80211/scan.c +16 −5
net/mac80211/sta_info.c +5 −1
net/mac80211/sta_info.h +32 −1
net/mac80211/tx.c +36 −16
net/mac80211/util.c +11 −13
net/mac80211/wpa.c +9 −4
net/mptcp/options.c +3 −2
net/mptcp/pm_netlink.c +4 −4
net/mptcp/protocol.c +59 −29
net/mptcp/protocol.h +2 −2
net/mptcp/sockopt.c +2 −2
net/mptcp/subflow.c +90 −92
net/netfilter/ipvs/ip_vs_ctl.c +1 −1
net/netfilter/nf_conntrack_proto.c +1 −1
net/netfilter/nf_flow_table_core.c +1 −2
net/netfilter/nf_flow_table_offload.c +4 −3
net/netfilter/nf_synproxy_core.c +5 −0
net/netfilter/nf_tables_api.c +100 −69
net/netfilter/nfnetlink_cthelper.c +6 −2
net/netfilter/nft_ct.c +1 −1
net/netfilter/nft_set_pipapo.c +2 −2
net/netfilter/nft_set_pipapo.h +2 −0
net/netfilter/nft_set_pipapo_avx2.c +3 −0
net/netlink/af_netlink.c +4 −2
net/nfc/llcp_sock.c +2 −0
net/nfc/nci/core.c +1 −0
net/nfc/nci/hci.c +5 −0
net/nfc/rawsock.c +1 −1
net/openvswitch/meter.c +8 −0
net/packet/af_packet.c +31 −20
net/qrtr/qrtr.c +1 −1
net/rds/connection.c +17 −6
net/rds/recv.c +1 −1
net/rds/tcp.c +2 −2
net/rds/tcp.h +2 −1
net/rds/tcp_listen.c +6 −0
net/sched/act_ct.c +17 −14
net/sched/cls_api.c +1 −1
net/sched/sch_cake.c +11 −7
net/sched/sch_dsmark.c +2 −1
net/sched/sch_fq_pie.c +13 −6
net/sched/sch_generic.c +48 −2
net/sched/sch_htb.c +4 −4
net/sctp/socket.c +1 −0
net/sctp/sysctl.c +1 −1
net/smc/smc_ism.c +10 −1
net/socket.c +0 −13
net/sunrpc/clnt.c +0 −7
net/sunrpc/xprt.c +32 −8
net/sunrpc/xprtrdma/rpc_rdma.c +18 −9
net/sunrpc/xprtrdma/transport.c +6 −6
net/sunrpc/xprtrdma/verbs.c +15 −3
net/sunrpc/xprtrdma/xprt_rdma.h +1 −0
net/sunrpc/xprtsock.c +2 −0
net/tipc/core.c +4 −2
net/tipc/core.h +3 −7
net/tipc/discover.c +2 −2
net/tipc/link.c +5 −0
net/tipc/link.h +1 −0
net/tipc/msg.c +2 −7
net/tipc/net.c +3 −12
net/tipc/node.c +6 −6
net/tipc/socket.c +4 −1
net/tipc/udp_media.c +2 −0
net/tls/tls_device.c +49 −11
net/tls/tls_device_fallback.c +7 −0
net/tls/tls_main.c +1 −0
net/tls/tls_sw.c +6 −5
net/unix/af_unix.c +4 −3
net/wireless/Makefile +1 −1
net/wireless/core.c +5 −8
net/wireless/pmsr.c +14 −2
net/wireless/sysfs.c +4 −0
net/wireless/util.c +8 −2
net/x25/af_x25.c +1 −1
samples/bpf/xdpsock_user.c +1 −1
samples/vfio-mdev/mdpy-fb.c +9 −4
scripts/Makefile.modfinal +1 −1
scripts/link-vmlinux.sh +1 −1
scripts/recordmcount.h +10 −5
sound/core/control_led.c +26 −7
sound/core/seq/seq_timer.c +9 −1
sound/core/timer.c +2 −1
sound/firewire/amdtp-stream.c +1 −1
sound/hda/intel-dsp-config.c +4 −0
sound/pci/hda/hda_codec.c +5 −0
sound/pci/hda/hda_generic.c +1 −0
sound/pci/hda/hda_intel.c +3 −0
sound/pci/hda/patch_cirrus.c +3 −4
sound/pci/hda/patch_realtek.c +62 −1
sound/soc/amd/raven/acp3x-pcm-dma.c +0 −10
sound/soc/amd/raven/acp3x.h +1 −0
sound/soc/amd/raven/pci-acp3x.c +15 −0
sound/soc/codecs/ak5558.c +1 −1
sound/soc/codecs/cs35l32.c +3 −0
sound/soc/codecs/cs35l33.c +1 −0
sound/soc/codecs/cs35l34.c +3 −0
sound/soc/codecs/cs42l42.c +3 −0
sound/soc/codecs/cs42l56.c +6 −1
sound/soc/codecs/cs42l73.c +3 −0
sound/soc/codecs/cs53l30.c +3 −0
sound/soc/codecs/da7219.c +1 −4
sound/soc/codecs/lpass-rx-macro.c +1 −0
sound/soc/codecs/lpass-tx-macro.c +1 −0
sound/soc/codecs/max98088.c +11 −2
sound/soc/codecs/rt5659.c +21 −5
sound/soc/codecs/rt5682-sdw.c +2 −1
sound/soc/codecs/rt711-sdca.c +2 −2
sound/soc/codecs/sti-sas.c +1 −0
sound/soc/codecs/tas2562.h +7 −7
sound/soc/fsl/Kconfig +1 −0
sound/soc/fsl/fsl-asoc-card.c +1 −0
sound/soc/generic/audio-graph-card.c +29 −28
sound/soc/generic/simple-card.c +83 −85
sound/soc/intel/boards/bytcr_rt5640.c +25 −0
sound/soc/qcom/lpass-cpu.c +80 −11
sound/soc/qcom/lpass.h +4 −0
sound/soc/soc-core.c +2 −0
sound/soc/soc-topology.c +3 −3
sound/soc/sof/intel/hda-dai.c +5 −0
sound/soc/sof/pm.c +1 −0
sound/soc/stm/stm32_sai_sub.c +1 −4
sound/usb/format.c +1 −1
sound/usb/mixer_quirks.c +1 −1
sound/usb/mixer_scarlett_gen2.c +51 −30
sound/usb/mixer_scarlett_gen2.h +1 −1
tools/arch/mips/include/uapi/asm/perf_regs.h +40 −0
tools/arch/x86/include/asm/disabled-features.h +2 −5
tools/arch/x86/include/uapi/asm/kvm.h +2 −0
tools/bootconfig/include/linux/bootconfig.h +4 −0
tools/bootconfig/main.c +1 −0
tools/bpf/bpftool/Documentation/bpftool-cgroup.rst +3 −1
tools/bpf/bpftool/Documentation/bpftool-prog.rst +1 −1
tools/bpf/bpftool/bash-completion/bpftool +3 −3
tools/bpf/bpftool/cgroup.c +2 −1
tools/bpf/bpftool/prog.c +1 −1
tools/include/uapi/asm-generic/unistd.h +1 −2
tools/include/uapi/linux/fs.h +1 −1
tools/include/uapi/linux/in.h +3 −0
tools/include/uapi/linux/kvm.h +3 −2
tools/include/uapi/linux/perf_event.h +1 −1
tools/lib/bpf/libbpf.c +3 −0
tools/lib/bpf/libbpf_internal.h +5 −0
tools/lib/bpf/xsk.c +1 −1
tools/objtool/arch/x86/decode.c +4 −0
tools/objtool/elf.c +24 −1
tools/perf/Documentation/perf-intel-pt.txt +3 −3
tools/perf/Documentation/perf-script.txt +4 −3
tools/perf/Makefile.config +0 −1
tools/perf/arch/mips/entry/syscalls/syscall_n64.tbl +1 −1
tools/perf/arch/powerpc/entry/syscalls/syscall.tbl +1 −1
tools/perf/arch/s390/entry/syscalls/syscall.tbl +1 −1
tools/perf/arch/x86/entry/syscalls/syscall_64.tbl +1 −1
tools/perf/builtin-buildid-list.c +3 −0
tools/perf/builtin-record.c +6 −0
tools/perf/builtin-stat.c +10 −3
tools/perf/check-headers.sh +1 −0
tools/perf/perf.c +2 −2
tools/perf/pmu-events/arch/powerpc/power10/cache.json +20 −10
tools/perf/pmu-events/arch/powerpc/power10/floating_point.json +1 −1
tools/perf/pmu-events/arch/powerpc/power10/frontend.json +77 −47
tools/perf/pmu-events/arch/powerpc/power10/locks.json +2 −2
tools/perf/pmu-events/arch/powerpc/power10/marked.json +28 −33
tools/perf/pmu-events/arch/powerpc/power10/memory.json +37 −42
tools/perf/pmu-events/arch/powerpc/power10/others.json +54 −79
tools/perf/pmu-events/arch/powerpc/power10/pipeline.json +65 −70
tools/perf/pmu-events/arch/powerpc/power10/pmc.json +4 −4
tools/perf/pmu-events/arch/powerpc/power10/translation.json +11 −11
tools/perf/pmu-events/jevents.c +1 −1
tools/perf/scripts/python/exported-sql-viewer.py +9 −3
tools/perf/tests/attr/base-record +1 −1
tools/perf/tests/pfm.c +2 −2
tools/perf/tests/shell/stat_bpf_counters.sh +2 −2
tools/perf/trace/beauty/include/linux/socket.h +0 −2
tools/perf/util/bpf_counter.c +6 −4
tools/perf/util/dwarf-aux.c +6 −2
tools/perf/util/env.c +1 −0
tools/perf/util/event.h +1 −1
tools/perf/util/evlist.c +0 −3
tools/perf/util/evsel.c +1 −0
tools/perf/util/evsel.h +2 −2
tools/perf/util/intel-pt-decoder/intel-pt-decoder.c +5 −1
tools/perf/util/intel-pt.c +4 −2
tools/perf/util/machine.c +2 −1
tools/perf/util/metricgroup.c +8 −6
tools/perf/util/parse-events.c +11 −2
tools/perf/util/parse-events.l +1 −0
tools/perf/util/perf_api_probe.c +10 −0
tools/perf/util/perf_api_probe.h +1 −0
tools/perf/util/pfm.c +10 −1
tools/perf/util/probe-finder.c +3 −0
tools/perf/util/session.c +1 −0
tools/perf/util/stat-display.c +3 −5
tools/perf/util/symbol-elf.c +1 −0
tools/testing/selftests/bpf/network_helpers.c +1 −1
tools/testing/selftests/bpf/network_helpers.h +1 −0
tools/testing/selftests/bpf/prog_tests/ringbuf.c +48 −1
tools/testing/selftests/bpf/prog_tests/tc_redirect.c +785 −0
tools/testing/selftests/bpf/progs/test_tc_neigh.c +10 −23
tools/testing/selftests/bpf/progs/test_tc_neigh_fib.c +6 −3
tools/testing/selftests/bpf/progs/test_tc_peer.c +37 −19
tools/testing/selftests/bpf/test_tc_redirect.sh +0 −216
tools/testing/selftests/bpf/test_verifier.c +1 −1
tools/testing/selftests/bpf/verifier/and.c +2 −0
tools/testing/selftests/bpf/verifier/bounds.c +14 −0
tools/testing/selftests/bpf/verifier/dead_code.c +2 −0
tools/testing/selftests/bpf/verifier/jmp32.c +22 −0
tools/testing/selftests/bpf/verifier/jset.c +6 −4
tools/testing/selftests/bpf/verifier/stack_ptr.c +0 −2
tools/testing/selftests/bpf/verifier/unpriv.c +2 −0
tools/testing/selftests/bpf/verifier/value_ptr_arith.c +4 −11
tools/testing/selftests/kvm/.gitignore +1 −0
tools/testing/selftests/kvm/Makefile +2 −1
tools/testing/selftests/kvm/demand_paging_test.c +109 −65
tools/testing/selftests/kvm/hardware_disable_test.c +31 −1
tools/testing/selftests/kvm/include/kvm_util.h +9 −5
tools/testing/selftests/kvm/include/test_util.h +12 −0
tools/testing/selftests/kvm/kvm_page_table_test.c +1 −1
tools/testing/selftests/kvm/lib/kvm_util.c +263 −67
tools/testing/selftests/kvm/lib/kvm_util_internal.h +15 −2
tools/testing/selftests/kvm/lib/perf_test_util.c +4 −2
tools/testing/selftests/kvm/lib/rbtree.c +1 −0
tools/testing/selftests/kvm/lib/test_util.c +35 −16
tools/testing/selftests/kvm/lib/x86_64/processor.c +4 −12
tools/testing/selftests/kvm/memslot_modification_stress_test.c +11 −7
tools/testing/selftests/kvm/memslot_perf_test.c +1037 −0
tools/testing/selftests/kvm/x86_64/get_cpuid_test.c +5 −0
tools/testing/selftests/kvm/x86_64/get_msr_index_features.c +2 −6
tools/testing/selftests/nci/.gitignore +1 −0
tools/testing/selftests/net/fib_tests.sh +25 −0
tools/testing/selftests/net/icmp.sh +74 −0
tools/testing/selftests/net/mptcp/mptcp_connect.sh +17 −7
tools/testing/selftests/net/udpgro_fwd.sh +1 −1
tools/testing/selftests/net/veth.sh +3 −2
tools/testing/selftests/netfilter/Makefile +1 −1
tools/testing/selftests/netfilter/nft_fib.sh +221 −0
tools/testing/selftests/proc/.gitignore +1 −0
tools/testing/selftests/tc-testing/tc-tests/qdiscs/fq_pie.json +4 −4
tools/testing/selftests/wireguard/netns.sh +1 −0
tools/testing/selftests/wireguard/qemu/kernel.config +0 −1
virt/kvm/kvm_main.c +4 −2
virt/lib/irqbypass.c +6 −10

@@ -109,8 +109,8 @@ ForEachMacros: · 108 unmodified lines
109	109	- 'css_for_each_child'
110	110	- 'css_for_each_descendant_post'
111	111	- 'css_for_each_descendant_pre'
112		− - 'cxl_for_each_cmd'
113	112	- 'device_for_each_child_node'
	113	+ - 'displayid_iter_for_each'
114	114	- 'dma_fence_chain_for_each'
115	115	- 'do_for_each_ftrace_op'
116	116	- 'drm_atomic_crtc_for_each_plane'
@@ -136,6 +136,7 @@ ForEachMacros: · 19 unmodified lines
136	136	- 'drm_mm_for_each_node_in_range'
137	137	- 'drm_mm_for_each_node_safe'
138	138	- 'flow_action_for_each'
	139	+ - 'for_each_acpi_dev_match'
139	140	- 'for_each_active_dev_scope'
140	141	- 'for_each_active_drhd_unit'
141	142	- 'for_each_active_iommu'
@@ -171,7 +172,6 @@ ForEachMacros: · 29 unmodified lines
171	172	- 'for_each_dapm_widgets'
172	173	- 'for_each_dev_addr'
173	174	- 'for_each_dev_scope'
174		− - 'for_each_displayid_db'
175	175	- 'for_each_dma_cap_mask'
176	176	- 'for_each_dpcm_be'
177	177	- 'for_each_dpcm_be_rollback'
@@ -179,6 +179,7 @@ ForEachMacros: · 1 unmodified line
179	179	- 'for_each_dpcm_fe'
180	180	- 'for_each_drhd_unit'
181	181	- 'for_each_dss_dev'
	182	+ - 'for_each_dtpm_table'
182	183	- 'for_each_efi_memory_desc'
183	184	- 'for_each_efi_memory_desc_in_map'
184	185	- 'for_each_element'
@@ -215,6 +216,7 @@ ForEachMacros: · 30 unmodified lines
215	216	- 'for_each_migratetype_order'
216	217	- 'for_each_msi_entry'
217	218	- 'for_each_msi_entry_safe'
	219	+ - 'for_each_msi_vector'
218	220	- 'for_each_net'
219	221	- 'for_each_net_continue_reverse'
220	222	- 'for_each_netdev'
@@ -270,6 +272,12 @@ ForEachMacros: · 49 unmodified lines
270	272	- 'for_each_prime_number_from'
271	273	- 'for_each_process'
272	274	- 'for_each_process_thread'
	275	+ - 'for_each_prop_codec_conf'
	276	+ - 'for_each_prop_dai_codec'
	277	+ - 'for_each_prop_dai_cpu'
	278	+ - 'for_each_prop_dlc_codecs'
	279	+ - 'for_each_prop_dlc_cpus'
	280	+ - 'for_each_prop_dlc_platforms'
273	281	- 'for_each_property_of_node'
274	282	- 'for_each_registered_fb'
275	283	- 'for_each_requested_gpio'
@@ -430,6 +438,7 @@ ForEachMacros: · 154 unmodified lines
430	438	- 'queue_for_each_hw_ctx'
431	439	- 'radix_tree_for_each_slot'
432	440	- 'radix_tree_for_each_tagged'
	441	+ - 'rb_for_each'
433	442	- 'rbtree_postorder_for_each_entry_safe'
434	443	- 'rdma_for_each_block'
435	444	- 'rdma_for_each_port'

.mailmap+3−0

@@ -243,6 +243,9 @@ Maxime Ripard <mripard@kernel.org> <maxime.ripard@free-electrons.com> · 242 unmodified lines
243	243	Mayuresh Janorkar <mayur@ti.com>
244	244	Michael Buesch <m@bues.ch>
245	245	Michel Dänzer <michel@tungstengraphics.com>
	246	+Michel Lespinasse <michel@lespinasse.org>
	247	+Michel Lespinasse <michel@lespinasse.org> <walken@google.com>
	248	+Michel Lespinasse <michel@lespinasse.org> <walken@zoy.org>
246	249	Miguel Ojeda <ojeda@kernel.org> <miguel.ojeda.sandonis@gmail.com>
247	250	Mike Rapoport <rppt@kernel.org> <mike@compulab.co.il>
248	251	Mike Rapoport <rppt@kernel.org> <mike.rapoport@gmail.com>

Documentation/admin-guide/sysctl/kernel.rst+14−3

@@ -1458,11 +1458,22 @@ unprivileged_bpf_disabled · 1457 unmodified lines
1458	1458	=========================
1459	1459
1460	1460	Writing 1 to this entry will disable unprivileged calls to ``bpf()``;
1461		−once disabled, calling ``bpf()`` without ``CAP_SYS_ADMIN`` will return
1462		−``-EPERM``.
	1461	+once disabled, calling ``bpf()`` without ``CAP_SYS_ADMIN`` or ``CAP_BPF``
	1462	+will return ``-EPERM``. Once set to 1, this can't be cleared from the
	1463	+running kernel anymore.
1463	1464
1464		−Once set, this can't be cleared.
	1465	+Writing 2 to this entry will also disable unprivileged calls to ``bpf()``,
	1466	+however, an admin can still change this setting later on, if needed, by
	1467	+writing 0 or 1 to this entry.
1465	1468
	1469	+If ``BPF_UNPRIV_DEFAULT_OFF`` is enabled in the kernel config, then this
	1470	+entry will default to 2 instead of 0.
	1471	+
	1472	+= =============================================================
	1473	+0 Unprivileged calls to ``bpf()`` are enabled
	1474	+1 Unprivileged calls to ``bpf()`` are disabled without recovery
	1475	+2 Unprivileged calls to ``bpf()`` are disabled
	1476	+= =============================================================
1466	1477
1467	1478	watchdog
1468	1479	========

Documentation/devicetree/bindings/connector/usb-connector.yaml+15−0

@@ -149,6 +149,17 @@ properties: · 148 unmodified lines
149	149	maxItems: 6
150	150	$ref: /schemas/types.yaml#/definitions/uint32-array
151	151
	152	+ sink-vdos-v1:
	153	+ description: An array of u32 with each entry, a Vendor Defined Message Object (VDO),
	154	+ providing additional information corresponding to the product, the detailed bit
	155	+ definitions and the order of each VDO can be found in
	156	+ "USB Power Delivery Specification Revision 2.0, Version 1.3" chapter 6.4.4.3.1 Discover
	157	+ Identity. User can specify the VDO array via VDO_IDH/_CERT/_PRODUCT/_CABLE/_AMA defined in
	158	+ dt-bindings/usb/pd.h.
	159	+ minItems: 3
	160	+ maxItems: 6
	161	+ $ref: /schemas/types.yaml#/definitions/uint32-array
	162	+
152	163	op-sink-microwatt:
153	164	description: Sink required operating power in microwatt, if source can't
154	165	offer the power, Capability Mismatch is set. Required for power sink and
@@ -207,6 +218,10 @@ properties: · 52 unmodified lines
207	218	SNK_READY for non-pd link.
208	219	type: boolean
209	220
	221	+dependencies:
	222	+ sink-vdos-v1: [ 'sink-vdos' ]
	223	+ sink-vdos: [ 'sink-vdos-v1' ]
	224	+
210	225	required:
211	226	- compatible
212	227

Documentation/devicetree/bindings/hwmon/ti,ads7828.yaml+1−1

@@ -49,7 +49,7 @@ examples: · 48 unmodified lines
49	49	#size-cells = <0>;
50	50
51	51	adc@48 {
52		− comatible = "ti,ads7828";
	52	+ compatible = "ti,ads7828";
53	53	reg = <0x48>;
54	54	vref-supply = <&vref>;
55	55	ti,differential-input;

Documentation/devicetree/bindings/i2c/i2c-mpc.yaml+7−0

@@ -46,6 +46,13 @@ properties: · 45 unmodified lines
46	46	description: \|
47	47	I2C bus timeout in microseconds
48	48
	49	+ fsl,i2c-erratum-a004447:
	50	+ $ref: /schemas/types.yaml#/definitions/flag
	51	+ description: \|
	52	+ Indicates the presence of QorIQ erratum A-004447, which
	53	+ says that the standard i2c recovery scheme mechanism does
	54	+ not work and an alternate implementation is needed.
	55	+
49	56	required:
50	57	- compatible
51	58	- reg

Documentation/devicetree/bindings/media/renesas,drif.yaml+1−3

@@ -67,9 +67,7 @@ properties: · 66 unmodified lines
67	67	maxItems: 1
68	68
69	69	clock-names:
70		− maxItems: 1
71		− items:
72		− - const: fck
	70	+ const: fck
73	71
74	72	resets:
75	73	maxItems: 1

Documentation/devicetree/bindings/net/renesas,ether.yaml+1−1

@@ -10,7 +10,7 @@ allOf: · 9 unmodified lines
10	10	- $ref: ethernet-controller.yaml#
11	11
12	12	maintainers:
13		− - Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
	13	+ - Sergei Shtylyov <sergei.shtylyov@gmail.com>
14	14
15	15	properties:
16	16	compatible:

Documentation/devicetree/bindings/sound/amlogic,gx-sound-card.yaml+2−2

@@ -57,7 +57,7 @@ patternProperties: · 56 unmodified lines
57	57	rate
58	58
59	59	sound-dai:
60		− $ref: /schemas/types.yaml#/definitions/phandle
	60	+ $ref: /schemas/types.yaml#/definitions/phandle-array
61	61	description: phandle of the CPU DAI
62	62
63	63	patternProperties:
@@ -71,7 +71,7 @@ patternProperties: · 7 unmodified lines
71	71
72	72	properties:
73	73	sound-dai:
74		− $ref: /schemas/types.yaml#/definitions/phandle
	74	+ $ref: /schemas/types.yaml#/definitions/phandle-array
75	75	description: phandle of the codec DAI
76	76
77	77	required:

Documentation/devicetree/bindings/spi/spi-mux.yaml+1−1

@@ -72,7 +72,7 @@ examples: · 71 unmodified lines
72	72
73	73	mux-controls = <&mux>;
74	74
75		− spi-flash@0 {
	75	+ flash@0 {
76	76	compatible = "jedec,spi-nor";
77	77	reg = <0>;
78	78	spi-max-frequency = <40000000>;

Documentation/driver-api/usb/usb.rst+2−0

@@ -123,6 +123,8 @@ are in ``drivers/usb/common/common.c``. · 122 unmodified lines
123	123	In addition, some functions useful for creating debugging output are
124	124	defined in ``drivers/usb/common/debug.c``.
125	125
	126	+.. _usb_header:
	127	+
126	128	Host-Side Data Types and Macros
127	129	===============================
128	130

Documentation/riscv/vm-layout.rst+2−2

@@ -58,6 +58,6 @@ RISC-V Linux Kernel SV39 · 57 unmodified lines
58	58	\|
59	59	____________________________________________________________\|____________________________________________________________
60	60	\| \| \| \|
61		− ffffffff00000000 \| -4 GB \| ffffffff7fffffff \| 2 GB \| modules
62		− ffffffff80000000 \| -2 GB \| ffffffffffffffff \| 2 GB \| kernel, BPF
	61	+ ffffffff00000000 \| -4 GB \| ffffffff7fffffff \| 2 GB \| modules, BPF
	62	+ ffffffff80000000 \| -2 GB \| ffffffffffffffff \| 2 GB \| kernel
63	63	__________________\|____________\|__________________\|_________\|____________________________________________________________

Documentation/userspace-api/seccomp_filter.rst+8−8

@@ -250,14 +250,14 @@ Users can read via ``ioctl(SECCOMP_IOCTL_NOTIF_RECV)`` (or ``poll()``) on a · 249 unmodified lines
250	250	seccomp notification fd to receive a ``struct seccomp_notif``, which contains
251	251	five members: the input length of the structure, a unique-per-filter ``id``,
252	252	the ``pid`` of the task which triggered this request (which may be 0 if the
253		−task is in a pid ns not visible from the listener's pid namespace), a ``flags``
254		−member which for now only has ``SECCOMP_NOTIF_FLAG_SIGNALED``, representing
255		−whether or not the notification is a result of a non-fatal signal, and the
256		−``data`` passed to seccomp. Userspace can then make a decision based on this
257		−information about what to do, and ``ioctl(SECCOMP_IOCTL_NOTIF_SEND)`` a
258		−response, indicating what should be returned to userspace. The ``id`` member of
259		−``struct seccomp_notif_resp`` should be the same ``id`` as in ``struct
260		−seccomp_notif``.
	253	+task is in a pid ns not visible from the listener's pid namespace). The
	254	+notification also contains the ``data`` passed to seccomp, and a filters flag.
	255	+The structure should be zeroed out prior to calling the ioctl.
	256	+
	257	+Userspace can then make a decision based on this information about what to do,
	258	+and ``ioctl(SECCOMP_IOCTL_NOTIF_SEND)`` a response, indicating what should be
	259	+returned to userspace. The ``id`` member of ``struct seccomp_notif_resp`` should
	260	+be the same ``id`` as in ``struct seccomp_notif``.
261	261
262	262	It is worth noting that ``struct seccomp_data`` contains the values of register
263	263	arguments to the syscall, but does not contain pointers to memory. The task's

Documentation/virt/kvm/mmu.rst+2−2

@@ -171,8 +171,8 @@ Shadow pages contain the following information: · 170 unmodified lines
171	171	shadow pages) so role.quadrant takes values in the range 0..3. Each
172	172	quadrant maps 1GB virtual address space.
173	173	role.access:
174		− Inherited guest access permissions in the form uwx. Note execute
175		− permission is positive, not negative.
	174	+ Inherited guest access permissions from the parent ptes in the form uwx.
	175	+ Note execute permission is positive, not negative.
176	176	role.invalid:
177	177	The page is invalid and should not be used. It is a root page that is
178	178	currently pinned (by a cpu hardware register pointing to it); once it is

Documentation/virt/kvm/vcpu-requests.rst+5−3

@@ -118,10 +118,12 @@ KVM_REQ_MMU_RELOAD · 117 unmodified lines
118	118	necessary to inform each VCPU to completely refresh the tables. This
119	119	request is used for that.
120	120
121		−KVM_REQ_PENDING_TIMER
	121	+KVM_REQ_UNBLOCK
122	122
123		− This request may be made from a timer handler run on the host on behalf
124		− of a VCPU. It informs the VCPU thread to inject a timer interrupt.
	123	+ This request informs the vCPU to exit kvm_vcpu_block. It is used for
	124	+ example from timer handlers that run on the host on behalf of a vCPU,
	125	+ or in order to update the interrupt routing and ensure that assigned
	126	+ devices will wake up the vCPU.
125	127
126	128	KVM_REQ_UNHALT
127	129

Documentation/vm/slub.rst+5−5

@@ -181,7 +181,7 @@ SLUB Debug output · 180 unmodified lines
181	181	Here is a sample of slub debug output::
182	182
183	183	====================================================================
184		− BUG kmalloc-8: Redzone overwritten
	184	+ BUG kmalloc-8: Right Redzone overwritten
185	185	--------------------------------------------------------------------
186	186
187	187	INFO: 0xc90f6d28-0xc90f6d2b. First byte 0x00 instead of 0xcc
@@ -189,10 +189,10 @@ Here is a sample of slub debug output:: · 1 unmodified line
189	189	INFO: Object 0xc90f6d20 @offset=3360 fp=0xc90f6d58
190	190	INFO: Allocated in get_modalias+0x61/0xf5 age=53 cpu=1 pid=554
191	191
192		− Bytes b4 0xc90f6d10: 00 00 00 00 00 00 00 00 5a 5a 5a 5a 5a 5a 5a 5a ........ZZZZZZZZ
193		− Object 0xc90f6d20: 31 30 31 39 2e 30 30 35 1019.005
194		− Redzone 0xc90f6d28: 00 cc cc cc .
195		− Padding 0xc90f6d50: 5a 5a 5a 5a 5a 5a 5a 5a ZZZZZZZZ
	192	+ Bytes b4 (0xc90f6d10): 00 00 00 00 00 00 00 00 5a 5a 5a 5a 5a 5a 5a 5a ........ZZZZZZZZ
	193	+ Object (0xc90f6d20): 31 30 31 39 2e 30 30 35 1019.005
	194	+ Redzone (0xc90f6d28): 00 cc cc cc .
	195	+ Padding (0xc90f6d50): 5a 5a 5a 5a 5a 5a 5a 5a ZZZZZZZZ
196	196
197	197	[<c010523d>] dump_trace+0x63/0x1eb
198	198	[<c01053df>] show_trace_log_lvl+0x1a/0x2f

MAINTAINERS+40−12

@@ -3877,6 +3877,7 @@ L: linux-btrfs@vger.kernel.org · 3876 unmodified lines
3877	3877	S: Maintained
3878	3878	W: http://btrfs.wiki.kernel.org/
3879	3879	Q: http://patchwork.kernel.org/project/linux-btrfs/list/
	3880	+C: irc://irc.libera.chat/btrfs
3880	3881	T: git git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux.git
3881	3882	F: Documentation/filesystems/btrfs.rst
3882	3883	F: fs/btrfs/
@@ -4138,6 +4139,14 @@ S: Odd Fixes · 255 unmodified lines
4138	4139	F: Documentation/devicetree/bindings/arm/cavium-thunder2.txt
4139	4140	F: arch/arm64/boot/dts/cavium/thunder2-99xx*
4140	4141
	4142	+CBS/ETF/TAPRIO QDISCS
	4143	+M: Vinicius Costa Gomes <vinicius.gomes@intel.com>
	4144	+S: Maintained
	4145	+L: netdev@vger.kernel.org
	4146	+F: net/sched/sch_cbs.c
	4147	+F: net/sched/sch_etf.c
	4148	+F: net/sched/sch_taprio.c
	4149	+
4141	4150	CC2520 IEEE-802.15.4 RADIO DRIVER
4142	4151	M: Varka Bhadram <varkabhadram@gmail.com>
4143	4152	L: linux-wpan@vger.kernel.org
@@ -5569,7 +5578,6 @@ F: drivers/soc/fsl/dpio · 1425 unmodified lines
5569	5578
5570	5579	DPAA2 ETHERNET DRIVER
5571	5580	M: Ioana Ciornei <ioana.ciornei@nxp.com>
5572		−M: Ioana Radulescu <ruxandra.radulescu@nxp.com>
5573	5581	L: netdev@vger.kernel.org
5574	5582	S: Maintained
5575	5583	F: Documentation/networking/device_drivers/ethernet/freescale/dpaa2/ethernet-driver.rst
@@ -6953,6 +6961,7 @@ F: net/core/failover.c · 1377 unmodified lines
6953	6961	FANOTIFY
6954	6962	M: Jan Kara <jack@suse.cz>
6955	6963	R: Amir Goldstein <amir73il@gmail.com>
	6964	+R: Matthew Bobrowski <repnop@google.com>
6956	6965	L: linux-fsdevel@vger.kernel.org
6957	6966	S: Maintained
6958	6967	F: fs/notify/fanotify/
@@ -12704,9 +12713,9 @@ F: drivers/rtc/rtc-ntxec.c · 5745 unmodified lines
12704	12713	F: include/linux/mfd/ntxec.h
12705	12714
12706	12715	NETRONOME ETHERNET DRIVERS
12707		−M: Simon Horman <simon.horman@netronome.com>
	12716	+M: Simon Horman <simon.horman@corigine.com>
12708	12717	R: Jakub Kicinski <kuba@kernel.org>
12709		−L: oss-drivers@netronome.com
	12718	+L: oss-drivers@corigine.com
12710	12719	S: Maintained
12711	12720	F: drivers/net/ethernet/netronome/
12712	12721
@@ -12733,7 +12742,6 @@ M: "David S. Miller" <davem@davemloft.net> · 20 unmodified lines
12733	12742	M: Jakub Kicinski <kuba@kernel.org>
12734	12743	L: netdev@vger.kernel.org
12735	12744	S: Maintained
12736		−W: http://www.linuxfoundation.org/en/Net
12737	12745	Q: https://patchwork.kernel.org/project/netdevbpf/list/
12738	12746	T: git git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net.git
12739	12747	T: git git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git
@@ -12778,7 +12786,6 @@ M: "David S. Miller" <davem@davemloft.net> · 38 unmodified lines
12778	12786	M: Jakub Kicinski <kuba@kernel.org>
12779	12787	L: netdev@vger.kernel.org
12780	12788	S: Maintained
12781		−W: http://www.linuxfoundation.org/en/Net
12782	12789	Q: https://patchwork.kernel.org/project/netdevbpf/list/
12783	12790	B: mailto:netdev@vger.kernel.org
12784	12791	T: git git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net.git
@@ -12920,8 +12927,10 @@ F: include/uapi/linux/nexthop.h · 135 unmodified lines
12920	12927	F: net/ipv4/nexthop.c
12921	12928
12922	12929	NFC SUBSYSTEM
	12930	+M: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
	12931	+L: linux-nfc@lists.01.org (subscribers-only)
12923	12932	L: netdev@vger.kernel.org
12924		−S: Orphan
	12933	+S: Maintained
12925	12934	F: Documentation/devicetree/bindings/net/nfc/
12926	12935	F: drivers/nfc/
12927	12936	F: include/linux/platform_data/nfcmrvl.h
@@ -12932,7 +12941,7 @@ F: net/nfc/ · 4 unmodified lines
12932	12941	NFC VIRTUAL NCI DEVICE DRIVER
12933	12942	M: Bongsu Jeon <bongsu.jeon@samsung.com>
12934	12943	L: netdev@vger.kernel.org
12935		−L: linux-nfc@lists.01.org (moderated for non-subscribers)
	12944	+L: linux-nfc@lists.01.org (subscribers-only)
12936	12945	S: Supported
12937	12946	F: drivers/nfc/virtual_ncidev.c
12938	12947	F: tools/testing/selftests/nci/
@@ -13229,9 +13238,8 @@ F: Documentation/devicetree/bindings/sound/tfa9879.txt · 290 unmodified lines
13229	13238	F: sound/soc/codecs/tfa9879*
13230	13239
13231	13240	NXP-NCI NFC DRIVER
13232		−M: Clément Perrochaud <clement.perrochaud@effinnov.com>
13233	13241	R: Charles Gorand <charles.gorand@effinnov.com>
13234		−L: linux-nfc@lists.01.org (moderated for non-subscribers)
	13242	+L: linux-nfc@lists.01.org (subscribers-only)
13235	13243	S: Supported
13236	13244	F: drivers/nfc/nxp-nci
13237	13245
@@ -14134,6 +14142,7 @@ F: drivers/pci/controller/pci-v3-semi.c · 896 unmodified lines
14134	14142	PCI ENDPOINT SUBSYSTEM
14135	14143	M: Kishon Vijay Abraham I <kishon@ti.com>
14136	14144	M: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
	14145	+R: Krzysztof Wilczyński <kw@linux.com>
14137	14146	L: linux-pci@vger.kernel.org
14138	14147	S: Supported
14139	14148	F: Documentation/PCI/endpoint/*
@@ -14182,6 +14191,7 @@ F: drivers/pci/controller/pci-xgene-msi.c · 42 unmodified lines
14182	14191	PCI NATIVE HOST BRIDGE AND ENDPOINT DRIVERS
14183	14192	M: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
14184	14193	R: Rob Herring <robh@kernel.org>
	14194	+R: Krzysztof Wilczyński <kw@linux.com>
14185	14195	L: linux-pci@vger.kernel.org
14186	14196	S: Supported
14187	14197	Q: http://patchwork.ozlabs.org/project/linux-pci/list/
@@ -14341,10 +14351,12 @@ PER-CPU MEMORY ALLOCATOR · 153 unmodified lines
14341	14351	M: Dennis Zhou <dennis@kernel.org>
14342	14352	M: Tejun Heo <tj@kernel.org>
14343	14353	M: Christoph Lameter <cl@linux.com>
	14354	+L: linux-mm@kvack.org
14344	14355	S: Maintained
14345	14356	T: git git://git.kernel.org/pub/scm/linux/kernel/git/dennis/percpu.git
14346	14357	F: arch/*/include/asm/percpu.h
14347	14358	F: include/linux/percpu*.h
	14359	+F: lib/percpu*.c
14348	14360	F: mm/percpu*.c
14349	14361
14350	14362	PER-TASK DELAY ACCOUNTING
@@ -15969,6 +15981,7 @@ S390 IUCV NETWORK LAYER · 1618 unmodified lines
15969	15981	M: Julian Wiedmann <jwi@linux.ibm.com>
15970	15982	M: Karsten Graul <kgraul@linux.ibm.com>
15971	15983	L: linux-s390@vger.kernel.org
	15984	+L: netdev@vger.kernel.org
15972	15985	S: Supported
15973	15986	W: http://www.ibm.com/developerworks/linux/linux390/
15974	15987	F: drivers/s390/net/iucv
@@ -15979,6 +15992,7 @@ S390 NETWORK DRIVERS · 4 unmodified lines
15979	15992	M: Julian Wiedmann <jwi@linux.ibm.com>
15980	15993	M: Karsten Graul <kgraul@linux.ibm.com>
15981	15994	L: linux-s390@vger.kernel.org
	15995	+L: netdev@vger.kernel.org
15982	15996	S: Supported
15983	15997	W: http://www.ibm.com/developerworks/linux/linux390/
15984	15998	F: drivers/s390/net/
@@ -16157,7 +16171,7 @@ F: include/media/drv-intf/s3c_camif.h · 172 unmodified lines
16157	16171	SAMSUNG S3FWRN5 NFC DRIVER
16158	16172	M: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
16159	16173	M: Krzysztof Opasiak <k.opasiak@samsung.com>
16160		−L: linux-nfc@lists.01.org (moderated for non-subscribers)
	16174	+L: linux-nfc@lists.01.org (subscribers-only)
16161	16175	S: Maintained
16162	16176	F: Documentation/devicetree/bindings/net/nfc/samsung,s3fwrn5.yaml
16163	16177	F: drivers/nfc/s3fwrn5
@@ -16570,6 +16584,7 @@ F: drivers/misc/sgi-xp/ · 406 unmodified lines
16570	16584
16571	16585	SHARED MEMORY COMMUNICATIONS (SMC) SOCKETS
16572	16586	M: Karsten Graul <kgraul@linux.ibm.com>
	16587	+M: Guvenc Gulce <guvenc@linux.ibm.com>
16573	16588	L: linux-s390@vger.kernel.org
16574	16589	S: Supported
16575	16590	W: http://www.ibm.com/developerworks/linux/linux390/
@@ -17328,6 +17343,12 @@ L: linux-i2c@vger.kernel.org · 752 unmodified lines
17328	17343	S: Maintained
17329	17344	F: drivers/i2c/busses/i2c-stm32*
17330	17345
	17346	+ST STM32 SPI DRIVER
	17347	+M: Alain Volmat <alain.volmat@foss.st.com>
	17348	+L: linux-spi@vger.kernel.org
	17349	+S: Maintained
	17350	+F: drivers/spi/spi-stm32.c
	17351	+
17331	17352	ST STPDDC60 DRIVER
17332	17353	M: Daniel Nilsson <daniel.nilsson@flex.com>
17333	17354	L: linux-hwmon@vger.kernel.org
@@ -17686,7 +17707,6 @@ R: Mika Westerberg <mika.westerberg@linux.intel.com> · 352 unmodified lines
17686	17707	L: linux-i2c@vger.kernel.org
17687	17708	S: Maintained
17688	17709	F: drivers/i2c/busses/i2c-designware-*
17689		−F: include/linux/platform_data/i2c-designware.h
17690	17710
17691	17711	SYNOPSYS DESIGNWARE MMC/SD/SDIO DRIVER
17692	17712	M: Jaehoon Chung <jh80.chung@samsung.com>
@@ -18342,7 +18362,7 @@ F: sound/soc/codecs/tas571x* · 649 unmodified lines
18342	18362	TI TRF7970A NFC DRIVER
18343	18363	M: Mark Greer <mgreer@animalcreek.com>
18344	18364	L: linux-wireless@vger.kernel.org
18345		−L: linux-nfc@lists.01.org (moderated for non-subscribers)
	18365	+L: linux-nfc@lists.01.org (subscribers-only)
18346	18366	S: Supported
18347	18367	F: Documentation/devicetree/bindings/net/nfc/trf7970a.txt
18348	18368	F: drivers/nfc/trf7970a.c
@@ -18878,6 +18898,13 @@ S: Maintained · 529 unmodified lines
18878	18898	F: drivers/usb/host/isp116x*
18879	18899	F: include/linux/usb/isp116x.h
18880	18900
	18901	+USB ISP1760 DRIVER
	18902	+M: Rui Miguel Silva <rui.silva@linaro.org>
	18903	+L: linux-usb@vger.kernel.org
	18904	+S: Maintained
	18905	+F: drivers/usb/isp1760/*
	18906	+F: Documentation/devicetree/bindings/usb/nxp,isp1760.yaml
	18907	+
18881	18908	USB LAN78XX ETHERNET DRIVER
18882	18909	M: Woojung Huh <woojung.huh@microchip.com>
18883	18910	M: UNGLinuxDriver@microchip.com
@@ -20022,6 +20049,7 @@ F: arch/x86/xen/swiotlb · 1138 unmodified lines
20022	20049	F: drivers/xen/swiotlb
20023	20050
20024	20051	XFS FILESYSTEM
	20052	+C: irc://irc.oftc.net/xfs
20025	20053	M: Darrick J. Wong <djwong@kernel.org>
20026	20054	M: linux-xfs@vger.kernel.org
20027	20055	L: linux-xfs@vger.kernel.org

Makefile+10−2

@@ -2,8 +2,8 @@ · 1 unmodified line
2	2	VERSION = 5
3	3	PATCHLEVEL = 13
4	4	SUBLEVEL = 0
5		−EXTRAVERSION = -rc3
6		−NAME = Frozen Wasteland
	5	+EXTRAVERSION = -rc7
	6	+NAME = Opossums on Parade
7	7
8	8	# DOCUMENTATION
9	9	# To see a list of typical targets execute "make help"
@@ -928,6 +928,14 @@ CC_FLAGS_LTO += -fvisibility=hidden · 918 unmodified lines
928	928
929	929	# Limit inlining across translation units to reduce binary size
930	930	KBUILD_LDFLAGS += -mllvm -import-instr-limit=5
	931	+
	932	+# Check for frame size exceeding threshold during prolog/epilog insertion
	933	+# when using lld < 13.0.0.
	934	+ifneq ($(CONFIG_FRAME_WARN),0)
	935	+ifeq ($(shell test $(CONFIG_LLD_VERSION) -lt 130000; echo $$?),0)
	936	+KBUILD_LDFLAGS += -plugin-opt=-warn-stack-size=$(CONFIG_FRAME_WARN)
	937	+endif
	938	+endif
931	939	endif
932	940
933	941	ifdef CONFIG_LTO

arch/arc/include/uapi/asm/sigcontext.h+1−0

@@ -18,6 +18,7 @@ · 17 unmodified lines
18	18	*/
19	19	struct sigcontext {
20	20	struct user_regs_struct regs;
	21	+ struct user_regs_arcv2 v2abi;
21	22	};
22	23
23	24	#endif /* _ASM_ARC_SIGCONTEXT_H */

arch/arc/kernel/signal.c+43−0

@@ -61,6 +61,41 @@ struct rt_sigframe { · 60 unmodified lines
61	61	unsigned int sigret_magic;
62	62	};
63	63
	64	+static int save_arcv2_regs(struct sigcontext mctx, struct pt_regs regs)
	65	+{
	66	+ int err = 0;
	67	+#ifndef CONFIG_ISA_ARCOMPACT
	68	+ struct user_regs_arcv2 v2abi;
	69	+
	70	+ v2abi.r30 = regs->r30;
	71	+#ifdef CONFIG_ARC_HAS_ACCL_REGS
	72	+ v2abi.r58 = regs->r58;
	73	+ v2abi.r59 = regs->r59;
	74	+#else
	75	+ v2abi.r58 = v2abi.r59 = 0;
	76	+#endif
	77	+ err = __copy_to_user(&mctx->v2abi, &v2abi, sizeof(v2abi));
	78	+#endif
	79	+ return err;
	80	+}
	81	+
	82	+static int restore_arcv2_regs(struct sigcontext mctx, struct pt_regs regs)
	83	+{
	84	+ int err = 0;
	85	+#ifndef CONFIG_ISA_ARCOMPACT
	86	+ struct user_regs_arcv2 v2abi;
	87	+
	88	+ err = __copy_from_user(&v2abi, &mctx->v2abi, sizeof(v2abi));
	89	+
	90	+ regs->r30 = v2abi.r30;
	91	+#ifdef CONFIG_ARC_HAS_ACCL_REGS
	92	+ regs->r58 = v2abi.r58;
	93	+ regs->r59 = v2abi.r59;
	94	+#endif
	95	+#endif
	96	+ return err;
	97	+}
	98	+
64	99	static int
65	100	stash_usr_regs(struct rt_sigframe __user sf, struct pt_regs regs,
66	101	sigset_t *set)
@@ -94,6 +129,10 @@ stash_usr_regs(struct rt_sigframe __user sf, struct pt_regs regs, · 27 unmodified lines
94	129
95	130	err = __copy_to_user(&(sf->uc.uc_mcontext.regs.scratch), &uregs.scratch,
96	131	sizeof(sf->uc.uc_mcontext.regs.scratch));
	132	+
	133	+ if (is_isa_arcv2())
	134	+ err \|= save_arcv2_regs(&(sf->uc.uc_mcontext), regs);
	135	+
97	136	err \|= __copy_to_user(&sf->uc.uc_sigmask, set, sizeof(sigset_t));
98	137
99	138	return err ? -EFAULT : 0;
@@ -109,6 +148,10 @@ static int restore_usr_regs(struct pt_regs regs, struct rt_sigframe __user sf) · 9 unmodified lines
109	148	err \|= __copy_from_user(&uregs.scratch,
110	149	&(sf->uc.uc_mcontext.regs.scratch),
111	150	sizeof(sf->uc.uc_mcontext.regs.scratch));
	151	+
	152	+ if (is_isa_arcv2())
	153	+ err \|= restore_arcv2_regs(&(sf->uc.uc_mcontext), regs);
	154	+
112	155	if (err)
113	156	return -EFAULT;
114	157

arch/arc/kernel/vmlinux.lds.S+1−1

@@ -57,7 +57,6 @@ SECTIONS · 56 unmodified lines
57	57	.init.ramfs : { INIT_RAM_FS }
58	58
59	59	. = ALIGN(PAGE_SIZE);
60		− _stext = .;
61	60
62	61	HEAD_TEXT_SECTION
63	62	INIT_TEXT_SECTION(L1_CACHE_BYTES)
@@ -83,6 +82,7 @@ SECTIONS · 19 unmodified lines
83	82
84	83	.text : {
85	84	_text = .;
	85	+ _stext = .;
86	86	TEXT_TEXT
87	87	SCHED_TEXT
88	88	CPUIDLE_TEXT

arch/arm/boot/dts/imx6dl-yapp4-common.dtsi+5−1

@@ -105,9 +105,13 @@ · 104 unmodified lines
105	105	phy-reset-gpios = <&gpio1 25 GPIO_ACTIVE_LOW>;
106	106	phy-reset-duration = <20>;
107	107	phy-supply = <&sw2_reg>;
108		− phy-handle = <&ethphy0>;
109	108	status = "okay";
110	109
	110	+ fixed-link {
	111	+ speed = <1000>;
	112	+ full-duplex;
	113	+ };
	114	+
111	115	mdio {
112	116	#address-cells = <1>;
113	117	#size-cells = <0>;

arch/arm/boot/dts/imx6q-dhcom-som.dtsi+12−0

@@ -406,6 +406,18 @@ · 405 unmodified lines
406	406	vin-supply = <&sw1_reg>;
407	407	};
408	408
	409	+&reg_pu {
	410	+ vin-supply = <&sw1_reg>;
	411	+};
	412	+
	413	+&reg_vdd1p1 {
	414	+ vin-supply = <&sw2_reg>;
	415	+};
	416	+
	417	+&reg_vdd2p5 {
	418	+ vin-supply = <&sw2_reg>;
	419	+};
	420	+
409	421	&uart1 {
410	422	pinctrl-names = "default";
411	423	pinctrl-0 = <&pinctrl_uart1>;

arch/arm/boot/dts/imx6qdl-emcon-avari.dtsi+1−1

@@ -126,7 +126,7 @@ · 125 unmodified lines
126	126	compatible = "nxp,pca8574";
127	127	reg = <0x3a>;
128	128	gpio-controller;
129		− #gpio-cells = <1>;
	129	+ #gpio-cells = <2>;
130	130	};
131	131	};
132	132

arch/arm/boot/dts/imx7d-meerkat96.dts+1−1

@@ -193,7 +193,7 @@ · 192 unmodified lines
193	193	pinctrl-names = "default";
194	194	pinctrl-0 = <&pinctrl_usdhc1>;
195	195	keep-power-in-suspend;
196		− tuning-step = <2>;
	196	+ fsl,tuning-step = <2>;
197	197	vmmc-supply = <&reg_3p3v>;
198	198	no-1-8-v;
199	199	broken-cd;

arch/arm/boot/dts/imx7d-pico.dtsi+1−1

@@ -351,7 +351,7 @@ · 350 unmodified lines
351	351	pinctrl-2 = <&pinctrl_usdhc1_200mhz>;
352	352	cd-gpios = <&gpio5 0 GPIO_ACTIVE_LOW>;
353	353	bus-width = <4>;
354		− tuning-step = <2>;
	354	+ fsl,tuning-step = <2>;
355	355	vmmc-supply = <&reg_3p3v>;
356	356	wakeup-source;
357	357	no-1-8-v;

arch/arm/include/asm/cpuidle.h+3−2

@@ -7,9 +7,11 @@ · 6 unmodified lines
7	7	#ifdef CONFIG_CPU_IDLE
8	8	extern int arm_cpuidle_simple_enter(struct cpuidle_device *dev,
9	9	struct cpuidle_driver *drv, int index);
	10	+#define __cpuidle_method_section __used __section("__cpuidle_method_of_table")
10	11	#else
11	12	static inline int arm_cpuidle_simple_enter(struct cpuidle_device *dev,
12	13	struct cpuidle_driver *drv, int index) { return -ENODEV; }
	14	+#define __cpuidle_method_section __maybe_unused /* drop silently */
13	15	#endif
14	16
15	17	/* Common ARM WFI state */
@@ -42,8 +44,7 @@ struct of_cpuidle_method { · 26 unmodified lines
42	44
43	45	#define CPUIDLE_METHOD_OF_DECLARE(name, _method, _ops) \
44	46	static const struct of_cpuidle_method __cpuidle_method_of_table_##name \
45		− __used __section("__cpuidle_method_of_table") \
46		− = { .method = _method, .ops = _ops }
	47	+ __cpuidle_method_section = { .method = _method, .ops = _ops }
47	48
48	49	extern int arm_cpuidle_suspend(int index);
49	50

arch/arm/mach-imx/pm-imx27.c+1−0

@@ -12,6 +12,7 @@ · 11 unmodified lines
12	12	#include <linux/suspend.h>
13	13	#include <linux/io.h>
14	14
	15	+#include "common.h"
15	16	#include "hardware.h"
16	17
17	18	static int mx27_suspend_enter(suspend_state_t state)

arch/arm/mach-omap1/board-ams-delta.c+0−14

@@ -458,20 +458,6 @@ static struct gpiod_lookup_table leds_gpio_table = { · 457 unmodified lines
458	458
459	459	#ifdef CONFIG_LEDS_TRIGGERS
460	460	DEFINE_LED_TRIGGER(ams_delta_camera_led_trigger);
461		−
462		−static int ams_delta_camera_power(struct device *dev, int power)
463		−{
464		− /*
465		− * turn on camera LED
466		− */
467		− if (power)
468		− led_trigger_event(ams_delta_camera_led_trigger, LED_FULL);
469		− else
470		− led_trigger_event(ams_delta_camera_led_trigger, LED_OFF);
471		− return 0;
472		−}
473		−#else
474		−#define ams_delta_camera_power NULL
475	461	#endif
476	462
477	463	static struct platform_device ams_delta_audio_device = {

arch/arm/mach-omap1/board-h2.c+3−1

@@ -320,7 +320,7 @@ static int tps_setup(struct i2c_client client, void context) · 319 unmodified lines
320	320	{
321	321	if (!IS_BUILTIN(CONFIG_TPS65010))
322	322	return -ENOSYS;
323		−
	323	+
324	324	tps65010_config_vregs1(TPS_LDO2_ENABLE \| TPS_VLDO2_3_0V \|
325	325	TPS_LDO1_ENABLE \| TPS_VLDO1_3_0V);
326	326
@@ -394,6 +394,8 @@ static void __init h2_init(void) · 67 unmodified lines
394	394	BUG_ON(gpio_request(H2_NAND_RB_GPIO_PIN, "NAND ready") < 0);
395	395	gpio_direction_input(H2_NAND_RB_GPIO_PIN);
396	396
	397	+ gpiod_add_lookup_table(&isp1301_gpiod_table);
	398	+
397	399	omap_cfg_reg(L3_1610_FLASH_CS2B_OE);
398	400	omap_cfg_reg(M8_1610_FLASH_CS2B_WE);
399	401

arch/arm/mach-omap1/pm.c+7−3

@@ -655,9 +655,13 @@ static int __init omap_pm_init(void) · 654 unmodified lines
655	655	irq = INT_7XX_WAKE_UP_REQ;
656	656	else if (cpu_is_omap16xx())
657	657	irq = INT_1610_WAKE_UP_REQ;
658		− if (request_irq(irq, omap_wakeup_interrupt, 0, "peripheral wakeup",
659		− NULL))
660		− pr_err("Failed to request irq %d (peripheral wakeup)\n", irq);
	658	+ else
	659	+ irq = -1;
	660	+
	661	+ if (irq >= 0) {
	662	+ if (request_irq(irq, omap_wakeup_interrupt, 0, "peripheral wakeup", NULL))
	663	+ pr_err("Failed to request irq %d (peripheral wakeup)\n", irq);
	664	+ }
661	665
662	666	/* Program new power ramp-up time
663	667	* (0 for most boards since we don't lower voltage when in deep sleep)

arch/arm/mach-omap2/board-n8x0.c+1−1

@@ -322,6 +322,7 @@ static int n8x0_mmc_get_cover_state(struct device *dev, int slot) · 321 unmodified lines
322	322
323	323	static void n8x0_mmc_callback(void *data, u8 card_mask)
324	324	{
	325	+#ifdef CONFIG_MMC_OMAP
325	326	int bit, *openp, index;
326	327
327	328	if (board_is_n800()) {
@@ -339,7 +340,6 @@ static void n8x0_mmc_callback(void *data, u8 card_mask) · 11 unmodified lines
339	340	else
340	341	*openp = 0;
341	342
342		−#ifdef CONFIG_MMC_OMAP
343	343	omap_mmc_notify_cover_event(mmc_device, index, *openp);
344	344	#else
345	345	pr_warn("MMC: notify cover event not available\n");

arch/arm64/Kbuild+1−2

@@ -1,6 +1,5 @@
1	1	# SPDX-License-Identifier: GPL-2.0-only
2		−obj-y += kernel/ mm/
3		−obj-$(CONFIG_NET) += net/
	2	+obj-y += kernel/ mm/ net/
4	3	obj-$(CONFIG_KVM) += kvm/
5	4	obj-$(CONFIG_XEN) += xen/
6	5	obj-$(CONFIG_CRYPTO) += crypto/

arch/arm64/Kconfig.platforms+1−0

@@ -165,6 +165,7 @@ config ARCH_MEDIATEK · 164 unmodified lines
165	165
166	166	config ARCH_MESON
167	167	bool "Amlogic Platforms"
	168	+ select COMMON_CLK
168	169	select MESON_IRQ_GPIO
169	170	help
170	171	This enables support for the arm64 based Amlogic SoCs

arch/arm64/boot/dts/freescale/fsl-ls1028a-kontron-sl28-var1.dts+2−1

@@ -46,7 +46,8 @@ · 45 unmodified lines
46	46	eee-broken-100tx;
47	47	qca,clk-out-frequency = <125000000>;
48	48	qca,clk-out-strength = <AR803X_STRENGTH_FULL>;
49		− vddio-supply = <&vddh>;
	49	+ qca,keep-pll-enabled;
	50	+ vddio-supply = <&vddio>;
50	51
51	52	vddio: vddio-regulator {
52	53	regulator-name = "VDDIO";

arch/arm64/boot/dts/freescale/fsl-ls1028a-kontron-sl28-var4.dts+2−3

@@ -31,11 +31,10 @@ · 30 unmodified lines
31	31	reg = <0x4>;
32	32	eee-broken-1000t;
33	33	eee-broken-100tx;
34		−
35	34	qca,clk-out-frequency = <125000000>;
36	35	qca,clk-out-strength = <AR803X_STRENGTH_FULL>;
37		−
38		− vddio-supply = <&vddh>;
	36	+ qca,keep-pll-enabled;
	37	+ vddio-supply = <&vddio>;
39	38
40	39	vddio: vddio-regulator {
41	40	regulator-name = "VDDIO";

arch/arm64/boot/dts/freescale/fsl-ls1028a.dtsi+2−2

@@ -197,8 +197,8 @@ · 196 unmodified lines
197	197	ddr: memory-controller@1080000 {
198	198	compatible = "fsl,qoriq-memory-controller";
199	199	reg = <0x0 0x1080000 0x0 0x1000>;
200		− interrupts = <GIC_SPI 144 IRQ_TYPE_LEVEL_HIGH>;
201		− big-endian;
	200	+ interrupts = <GIC_SPI 17 IRQ_TYPE_LEVEL_HIGH>;
	201	+ little-endian;
202	202	};
203	203
204	204	dcfg: syscon@1e00000 {

arch/arm64/boot/dts/freescale/imx8mq-zii-ultra-rmb3.dts+5−5

@@ -88,11 +88,11 @@ · 87 unmodified lines
88	88	pinctrl-0 = <&pinctrl_codec2>;
89	89	reg = <0x18>;
90	90	#sound-dai-cells = <0>;
91		− HPVDD-supply = <&reg_3p3v>;
92		− SPRVDD-supply = <&reg_3p3v>;
93		− SPLVDD-supply = <&reg_3p3v>;
94		− AVDD-supply = <&reg_3p3v>;
95		− IOVDD-supply = <&reg_3p3v>;
	91	+ HPVDD-supply = <&reg_gen_3p3>;
	92	+ SPRVDD-supply = <&reg_gen_3p3>;
	93	+ SPLVDD-supply = <&reg_gen_3p3>;
	94	+ AVDD-supply = <&reg_gen_3p3>;
	95	+ IOVDD-supply = <&reg_gen_3p3>;
96	96	DVDD-supply = <&vgen4_reg>;
97	97	reset-gpios = <&gpio3 4 GPIO_ACTIVE_HIGH>;
98	98	};

arch/arm64/boot/dts/freescale/imx8mq-zii-ultra.dtsi+7−16

@@ -45,8 +45,8 @@ · 44 unmodified lines
45	45	reg_12p0_main: regulator-12p0-main {
46	46	compatible = "regulator-fixed";
47	47	regulator-name = "12V_MAIN";
48		− regulator-min-microvolt = <5000000>;
49		− regulator-max-microvolt = <5000000>;
	48	+ regulator-min-microvolt = <12000000>;
	49	+ regulator-max-microvolt = <12000000>;
50	50	regulator-always-on;
51	51	};
52	52
@@ -77,15 +77,6 @@ · 24 unmodified lines
77	77	regulator-always-on;
78	78	};
79	79
80		− reg_3p3v: regulator-3p3v {
81		− compatible = "regulator-fixed";
82		− vin-supply = <&reg_3p3_main>;
83		− regulator-name = "GEN_3V3";
84		− regulator-min-microvolt = <3300000>;
85		− regulator-max-microvolt = <3300000>;
86		− regulator-always-on;
87		− };
88		−
89	80	reg_usdhc2_vmmc: regulator-vsd-3v3 {
90	81	pinctrl-names = "default";
91	82	pinctrl-0 = <&pinctrl_reg_usdhc2>;
@@ -415,11 +406,11 @@ · 323 unmodified lines
415	406	pinctrl-0 = <&pinctrl_codec1>;
416	407	reg = <0x18>;
417	408	#sound-dai-cells = <0>;
418		− HPVDD-supply = <&reg_3p3v>;
419		− SPRVDD-supply = <&reg_3p3v>;
420		− SPLVDD-supply = <&reg_3p3v>;
421		− AVDD-supply = <&reg_3p3v>;
422		− IOVDD-supply = <&reg_3p3v>;
	409	+ HPVDD-supply = <&reg_gen_3p3>;
	410	+ SPRVDD-supply = <&reg_gen_3p3>;
	411	+ SPLVDD-supply = <&reg_gen_3p3>;
	412	+ AVDD-supply = <&reg_gen_3p3>;
	413	+ IOVDD-supply = <&reg_gen_3p3>;
423	414	DVDD-supply = <&vgen4_reg>;
424	415	reset-gpios = <&gpio3 3 GPIO_ACTIVE_LOW>;
425	416	};

arch/arm64/boot/dts/ti/k3-am64-main.dtsi+6−5

@@ -42,12 +42,12 @@ · 41 unmodified lines
42	42	};
43	43	};
44	44
45		− dmss: dmss {
	45	+ dmss: bus@48000000 {
46	46	compatible = "simple-mfd";
47	47	#address-cells = <2>;
48	48	#size-cells = <2>;
49	49	dma-ranges;
50		− ranges;
	50	+ ranges = <0x00 0x48000000 0x00 0x48000000 0x00 0x06400000>;
51	51
52	52	ti,sci-dev-id = <25>;
53	53
@@ -134,7 +134,7 @@ · 80 unmodified lines
134	134	};
135	135	};
136	136
137		− dmsc: dmsc@44043000 {
	137	+ dmsc: system-controller@44043000 {
138	138	compatible = "ti,k2g-sci";
139	139	ti,host-id = <12>;
140	140	mbox-names = "rx", "tx";
@@ -148,7 +148,7 @@ · 7 unmodified lines
148	148	#power-domain-cells = <2>;
149	149	};
150	150
151		− k3_clks: clocks {
	151	+ k3_clks: clock-controller {
152	152	compatible = "ti,k2g-sci-clk";
153	153	#clock-cells = <2>;
154	154	};
@@ -373,8 +373,9 @@ · 218 unmodified lines
373	373	clocks = <&k3_clks 145 0>;
374	374	};
375	375
376		− main_gpio_intr: interrupt-controller0 {
	376	+ main_gpio_intr: interrupt-controller@a00000 {
377	377	compatible = "ti,sci-intr";
	378	+ reg = <0x00 0x00a00000 0x00 0x800>;
378	379	ti,intr-trigger-type = <1>;
379	380	interrupt-controller;
380	381	interrupt-parent = <&gic500>;

arch/arm64/boot/dts/ti/k3-am64-mcu.dtsi+2−1

@@ -74,8 +74,9 @@ · 73 unmodified lines
74	74	clocks = <&k3_clks 148 0>;
75	75	};
76	76
77		− mcu_gpio_intr: interrupt-controller1 {
	77	+ mcu_gpio_intr: interrupt-controller@4210000 {
78	78	compatible = "ti,sci-intr";
	79	+ reg = <0x00 0x04210000 0x00 0x200>;
79	80	ti,intr-trigger-type = <1>;
80	81	interrupt-controller;
81	82	interrupt-parent = <&gic500>;

arch/arm64/boot/dts/ti/k3-am65-main.dtsi+6−4

@@ -433,8 +433,9 @@ · 432 unmodified lines
433	433	#phy-cells = <0>;
434	434	};
435	435
436		− intr_main_gpio: interrupt-controller0 {
	436	+ intr_main_gpio: interrupt-controller@a00000 {
437	437	compatible = "ti,sci-intr";
	438	+ reg = <0x0 0x00a00000 0x0 0x400>;
438	439	ti,intr-trigger-type = <1>;
439	440	interrupt-controller;
440	441	interrupt-parent = <&gic500>;
@@ -444,18 +445,19 @@ · 3 unmodified lines
444	445	ti,interrupt-ranges = <0 392 32>;
445	446	};
446	447
447		− main-navss {
	448	+ main_navss: bus@30800000 {
448	449	compatible = "simple-mfd";
449	450	#address-cells = <2>;
450	451	#size-cells = <2>;
451		− ranges;
	452	+ ranges = <0x0 0x30800000 0x0 0x30800000 0x0 0xbc00000>;
452	453	dma-coherent;
453	454	dma-ranges;
454	455
455	456	ti,sci-dev-id = <118>;
456	457
457		− intr_main_navss: interrupt-controller1 {
	458	+ intr_main_navss: interrupt-controller@310e0000 {
458	459	compatible = "ti,sci-intr";
	460	+ reg = <0x0 0x310e0000 0x0 0x2000>;
459	461	ti,intr-trigger-type = <4>;
460	462	interrupt-controller;
461	463	interrupt-parent = <&gic500>;

arch/arm64/boot/dts/ti/k3-am65-mcu.dtsi+2−2

@@ -116,11 +116,11 @@ · 115 unmodified lines
116	116	};
117	117	};
118	118
119		− mcu-navss {
	119	+ mcu_navss: bus@28380000 {
120	120	compatible = "simple-mfd";
121	121	#address-cells = <2>;
122	122	#size-cells = <2>;
123		− ranges;
	123	+ ranges = <0x00 0x28380000 0x00 0x28380000 0x00 0x03880000>;
124	124	dma-coherent;
125	125	dma-ranges;
126	126

arch/arm64/boot/dts/ti/k3-am65-wakeup.dtsi+7−6

@@ -6,24 +6,24 @@ · 5 unmodified lines
6	6	*/
7	7
8	8	&cbass_wakeup {
9		− dmsc: dmsc {
	9	+ dmsc: system-controller@44083000 {
10	10	compatible = "ti,am654-sci";
11	11	ti,host-id = <12>;
12		− #address-cells = <1>;
13		− #size-cells = <1>;
14		− ranges;
15	12
16	13	mbox-names = "rx", "tx";
17	14
18	15	mboxes= <&secure_proxy_main 11>,
19	16	<&secure_proxy_main 13>;
20	17
	18	+ reg-names = "debug_messages";
	19	+ reg = <0x44083000 0x1000>;
	20	+
21	21	k3_pds: power-controller {
22	22	compatible = "ti,sci-pm-domain";
23	23	#power-domain-cells = <2>;
24	24	};
25	25
26		− k3_clks: clocks {
	26	+ k3_clks: clock-controller {
27	27	compatible = "ti,k2g-sci-clk";
28	28	#clock-cells = <2>;
29	29	};
@@ -69,8 +69,9 @@ · 39 unmodified lines
69	69	power-domains = <&k3_pds 115 TI_SCI_PD_EXCLUSIVE>;
70	70	};
71	71
72		− intr_wkup_gpio: interrupt-controller2 {
	72	+ intr_wkup_gpio: interrupt-controller@42200000 {
73	73	compatible = "ti,sci-intr";
	74	+ reg = <0x42200000 0x200>;
74	75	ti,intr-trigger-type = <1>;
75	76	interrupt-controller;
76	77	interrupt-parent = <&gic500>;

arch/arm64/boot/dts/ti/k3-am654-base-board.dts+0−31

@@ -85,12 +85,6 @@ · 84 unmodified lines
85	85	gpios = <&wkup_gpio0 27 GPIO_ACTIVE_LOW>;
86	86	};
87	87	};
88		−
89		− clk_ov5640_fixed: clock {
90		− compatible = "fixed-clock";
91		− #clock-cells = <0>;
92		− clock-frequency = <24000000>;
93		− };
94	88	};
95	89
96	90	&wkup_pmx0 {
@@ -287,23 +281,6 @@ · 190 unmodified lines
287	281	pinctrl-names = "default";
288	282	pinctrl-0 = <&main_i2c1_pins_default>;
289	283	clock-frequency = <400000>;
290		−
291		− ov5640: camera@3c {
292		− compatible = "ovti,ov5640";
293		− reg = <0x3c>;
294		−
295		− clocks = <&clk_ov5640_fixed>;
296		− clock-names = "xclk";
297		−
298		− port {
299		− csi2_cam0: endpoint {
300		− remote-endpoint = <&csi2_phy0>;
301		− clock-lanes = <0>;
302		− data-lanes = <1 2>;
303		− };
304		− };
305		− };
306		−
307	284	};
308	285
309	286	&main_i2c2 {
@@ -496,14 +473,6 @@ · 186 unmodified lines
496	473	};
497	474	};
498	475
499		−&csi2_0 {
500		− csi2_phy0: endpoint {
501		− remote-endpoint = <&csi2_cam0>;
502		− clock-lanes = <0>;
503		− data-lanes = <1 2>;
504		− };
505		−};
506		−
507	476	&mcu_cpsw {
508	477	pinctrl-names = "default";
509	478	pinctrl-0 = <&mcu_cpsw_pins_default &mcu_mdio_pins_default>;

arch/arm64/boot/dts/ti/k3-j7200-main.dtsi+6−2

@@ -68,8 +68,9 @@ · 67 unmodified lines
68	68	};
69	69	};
70	70
71		− main_gpio_intr: interrupt-controller0 {
	71	+ main_gpio_intr: interrupt-controller@a00000 {
72	72	compatible = "ti,sci-intr";
	73	+ reg = <0x00 0x00a00000 0x00 0x800>;
73	74	ti,intr-trigger-type = <1>;
74	75	interrupt-controller;
75	76	interrupt-parent = <&gic500>;
@@ -85,9 +86,12 @@ · 9 unmodified lines
85	86	#size-cells = <2>;
86	87	ranges = <0x00 0x30000000 0x00 0x30000000 0x00 0x0c400000>;
87	88	ti,sci-dev-id = <199>;
	89	+ dma-coherent;
	90	+ dma-ranges;
88	91
89		− main_navss_intr: interrupt-controller1 {
	92	+ main_navss_intr: interrupt-controller@310e0000 {
90	93	compatible = "ti,sci-intr";
	94	+ reg = <0x00 0x310e0000 0x00 0x4000>;
91	95	ti,intr-trigger-type = <4>;
92	96	interrupt-controller;
93	97	interrupt-parent = <&gic500>;

arch/arm64/boot/dts/ti/k3-j7200-mcu-wakeup.dtsi+4−3

@@ -6,7 +6,7 @@ · 5 unmodified lines
6	6	*/
7	7
8	8	&cbass_mcu_wakeup {
9		− dmsc: dmsc@44083000 {
	9	+ dmsc: system-controller@44083000 {
10	10	compatible = "ti,k2g-sci";
11	11	ti,host-id = <12>;
12	12
@@ -23,7 +23,7 @@ · 10 unmodified lines
23	23	#power-domain-cells = <2>;
24	24	};
25	25
26		− k3_clks: clocks {
	26	+ k3_clks: clock-controller {
27	27	compatible = "ti,k2g-sci-clk";
28	28	#clock-cells = <2>;
29	29	};
@@ -96,8 +96,9 @@ · 66 unmodified lines
96	96	clock-names = "fclk";
97	97	};
98	98
99		− wkup_gpio_intr: interrupt-controller2 {
	99	+ wkup_gpio_intr: interrupt-controller@42200000 {
100	100	compatible = "ti,sci-intr";
	101	+ reg = <0x00 0x42200000 0x00 0x400>;
101	102	ti,intr-trigger-type = <1>;
102	103	interrupt-controller;
103	104	interrupt-parent = <&gic500>;

arch/arm64/boot/dts/ti/k3-j721e-main.dtsi+6−4

@@ -76,8 +76,9 @@ · 75 unmodified lines
76	76	};
77	77	};
78	78
79		− main_gpio_intr: interrupt-controller0 {
	79	+ main_gpio_intr: interrupt-controller@a00000 {
80	80	compatible = "ti,sci-intr";
	81	+ reg = <0x00 0x00a00000 0x00 0x800>;
81	82	ti,intr-trigger-type = <1>;
82	83	interrupt-controller;
83	84	interrupt-parent = <&gic500>;
@@ -87,18 +88,19 @@ · 3 unmodified lines
87	88	ti,interrupt-ranges = <8 392 56>;
88	89	};
89	90
90		− main-navss {
	91	+ main_navss: bus@30000000 {
91	92	compatible = "simple-mfd";
92	93	#address-cells = <2>;
93	94	#size-cells = <2>;
94		− ranges;
	95	+ ranges = <0x00 0x30000000 0x00 0x30000000 0x00 0x0c400000>;
95	96	dma-coherent;
96	97	dma-ranges;
97	98
98	99	ti,sci-dev-id = <199>;
99	100
100		− main_navss_intr: interrupt-controller1 {
	101	+ main_navss_intr: interrupt-controller@310e0000 {
101	102	compatible = "ti,sci-intr";
	103	+ reg = <0x0 0x310e0000 0x0 0x4000>;
102	104	ti,intr-trigger-type = <4>;
103	105	interrupt-controller;
104	106	interrupt-parent = <&gic500>;

arch/arm64/boot/dts/ti/k3-j721e-mcu-wakeup.dtsi+6−5

@@ -6,7 +6,7 @@ · 5 unmodified lines
6	6	*/
7	7
8	8	&cbass_mcu_wakeup {
9		− dmsc: dmsc@44083000 {
	9	+ dmsc: system-controller@44083000 {
10	10	compatible = "ti,k2g-sci";
11	11	ti,host-id = <12>;
12	12
@@ -23,7 +23,7 @@ · 10 unmodified lines
23	23	#power-domain-cells = <2>;
24	24	};
25	25
26		− k3_clks: clocks {
	26	+ k3_clks: clock-controller {
27	27	compatible = "ti,k2g-sci-clk";
28	28	#clock-cells = <2>;
29	29	};
@@ -96,8 +96,9 @@ · 66 unmodified lines
96	96	clock-names = "fclk";
97	97	};
98	98
99		− wkup_gpio_intr: interrupt-controller2 {
	99	+ wkup_gpio_intr: interrupt-controller@42200000 {
100	100	compatible = "ti,sci-intr";
	101	+ reg = <0x00 0x42200000 0x00 0x400>;
101	102	ti,intr-trigger-type = <1>;
102	103	interrupt-controller;
103	104	interrupt-parent = <&gic500>;
@@ -249,11 +250,11 @@ · 145 unmodified lines
249	250	};
250	251	};
251	252
252		− mcu-navss {
	253	+ mcu_navss: bus@28380000 {
253	254	compatible = "simple-mfd";
254	255	#address-cells = <2>;
255	256	#size-cells = <2>;
256		− ranges;
	257	+ ranges = <0x00 0x28380000 0x00 0x28380000 0x00 0x03880000>;
257	258	dma-coherent;
258	259	dma-ranges;
259	260

arch/arm64/include/asm/barrier.h+1−1

@@ -74,7 +74,7 @@ static inline unsigned long array_index_mask_nospec(unsigned long idx, · 73 unmodified lines
74	74	* This insanity brought to you by speculative system register reads,
75	75	* out-of-order memory accesses, sequence locks and Thomas Gleixner.
76	76	*
77		− * http://lists.infradead.org/pipermail/linux-arm-kernel/2019-February/631195.html
	77	+ * https://lore.kernel.org/r/alpine.DEB.2.21.1902081950260.1662@nanos.tec.linutronix.de/
78	78	*/
79	79	#define arch_counter_enforce_ordering(val) do { \
80	80	u64 tmp, _val = (val); \

arch/arm64/include/asm/kvm_asm.h+3−0

@@ -63,6 +63,7 @@ · 62 unmodified lines
63	63	#define __KVM_HOST_SMCCC_FUNC___pkvm_cpu_set_vector 18
64	64	#define __KVM_HOST_SMCCC_FUNC___pkvm_prot_finalize 19
65	65	#define __KVM_HOST_SMCCC_FUNC___pkvm_mark_hyp 20
	66	+#define __KVM_HOST_SMCCC_FUNC___kvm_adjust_pc 21
66	67
67	68	#ifndef __ASSEMBLY__
68	69
@@ -201,6 +202,8 @@ extern void __kvm_timer_set_cntvoff(u64 cntvoff); · 132 unmodified lines
201	202
202	203	extern int __kvm_vcpu_run(struct kvm_vcpu *vcpu);
203	204
	205	+extern void __kvm_adjust_pc(struct kvm_vcpu *vcpu);
	206	+
204	207	extern u64 __vgic_v3_get_gic_config(void);
205	208	extern u64 __vgic_v3_read_vmcr(void);
206	209	extern void __vgic_v3_write_vmcr(u32 vmcr);

arch/arm64/include/asm/kvm_emulate.h+5−0

@@ -463,4 +463,9 @@ static __always_inline void kvm_incr_pc(struct kvm_vcpu *vcpu) · 462 unmodified lines
463	463	vcpu->arch.flags \|= KVM_ARM64_INCREMENT_PC;
464	464	}
465	465
	466	+static inline bool vcpu_has_feature(struct kvm_vcpu *vcpu, int feature)
	467	+{
	468	+ return test_bit(feature, vcpu->arch.features);
	469	+}
	470	+
466	471	#endif /* __ARM64_KVM_EMULATE_H__ */

arch/arm64/kvm/arm.c+17−3

@@ -720,11 +720,13 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu) · 719 unmodified lines
720	720	return ret;
721	721	}
722	722
723		− if (run->immediate_exit)
724		− return -EINTR;
725		−
726	723	vcpu_load(vcpu);
727	724
	725	+ if (run->immediate_exit) {
	726	+ ret = -EINTR;
	727	+ goto out;
	728	+ }
	729	+
728	730	kvm_sigset_activate(vcpu);
729	731
730	732	ret = 1;
@@ -897,6 +899,18 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu) · 166 unmodified lines
897	899
898	900	kvm_sigset_deactivate(vcpu);
899	901
	902	+out:
	903	+ /*
	904	+ * In the unlikely event that we are returning to userspace
	905	+ * with pending exceptions or PC adjustment, commit these
	906	+ * adjustments in order to give userspace a consistent view of
	907	+ * the vcpu state. Note that this relies on __kvm_adjust_pc()
	908	+ * being preempt-safe on VHE.
	909	+ */
	910	+ if (unlikely(vcpu->arch.flags & (KVM_ARM64_PENDING_EXCEPTION \|
	911	+ KVM_ARM64_INCREMENT_PC)))
	912	+ kvm_call_hyp(__kvm_adjust_pc, vcpu);
	913	+
900	914	vcpu_put(vcpu);
901	915	return ret;
902	916	}

arch/arm64/kvm/hyp/exception.c+17−1

@@ -296,7 +296,7 @@ static void enter_exception32(struct kvm_vcpu *vcpu, u32 mode, u32 vect_offset) · 295 unmodified lines
296	296	*vcpu_pc(vcpu) = vect_offset;
297	297	}
298	298
299		−void kvm_inject_exception(struct kvm_vcpu *vcpu)
	299	+static void kvm_inject_exception(struct kvm_vcpu *vcpu)
300	300	{
301	301	if (vcpu_el1_is_32bit(vcpu)) {
302	302	switch (vcpu->arch.flags & KVM_ARM64_EXCEPT_MASK) {
@@ -329,3 +329,19 @@ void kvm_inject_exception(struct kvm_vcpu *vcpu) · 26 unmodified lines
329	329	}
330	330	}
331	331	}
	332	+
	333	+/*
	334	+ * Adjust the guest PC (and potentially exception state) depending on
	335	+ * flags provided by the emulation code.
	336	+ */
	337	+void __kvm_adjust_pc(struct kvm_vcpu *vcpu)
	338	+{
	339	+ if (vcpu->arch.flags & KVM_ARM64_PENDING_EXCEPTION) {
	340	+ kvm_inject_exception(vcpu);
	341	+ vcpu->arch.flags &= ~(KVM_ARM64_PENDING_EXCEPTION \|
	342	+ KVM_ARM64_EXCEPT_MASK);
	343	+ } else if (vcpu->arch.flags & KVM_ARM64_INCREMENT_PC) {
	344	+ kvm_skip_instr(vcpu);
	345	+ vcpu->arch.flags &= ~KVM_ARM64_INCREMENT_PC;
	346	+ }
	347	+}

arch/arm64/kvm/hyp/include/hyp/adjust_pc.h+0−18

@@ -13,8 +13,6 @@ · 12 unmodified lines
13	13	#include <asm/kvm_emulate.h>
14	14	#include <asm/kvm_host.h>
15	15
16		−void kvm_inject_exception(struct kvm_vcpu *vcpu);
17		−
18	16	static inline void kvm_skip_instr(struct kvm_vcpu *vcpu)
19	17	{
20	18	if (vcpu_mode_is_32bit(vcpu)) {
@@ -43,22 +41,6 @@ static inline void __kvm_skip_instr(struct kvm_vcpu *vcpu) · 22 unmodified lines
43	41	write_sysreg_el2(*vcpu_pc(vcpu), SYS_ELR);
44	42	}
45	43
46		−/*
47		− * Adjust the guest PC on entry, depending on flags provided by EL1
48		− * for the purpose of emulation (MMIO, sysreg) or exception injection.
49		− */
50		−static inline void __adjust_pc(struct kvm_vcpu *vcpu)
51		−{
52		− if (vcpu->arch.flags & KVM_ARM64_PENDING_EXCEPTION) {
53		− kvm_inject_exception(vcpu);
54		− vcpu->arch.flags &= ~(KVM_ARM64_PENDING_EXCEPTION \|
55		− KVM_ARM64_EXCEPT_MASK);
56		− } else if (vcpu->arch.flags & KVM_ARM64_INCREMENT_PC) {
57		− kvm_skip_instr(vcpu);
58		− vcpu->arch.flags &= ~KVM_ARM64_INCREMENT_PC;
59		− }
60		−}
61		−
62	44	/*
63	45	* Skip an instruction while host sysregs are live.
64	46	* Assumes host is always 64-bit.

arch/arm64/kvm/hyp/nvhe/hyp-main.c+8−0

@@ -28,6 +28,13 @@ static void handle___kvm_vcpu_run(struct kvm_cpu_context *host_ctxt) · 27 unmodified lines
28	28	cpu_reg(host_ctxt, 1) = __kvm_vcpu_run(kern_hyp_va(vcpu));
29	29	}
30	30
	31	+static void handle___kvm_adjust_pc(struct kvm_cpu_context *host_ctxt)
	32	+{
	33	+ DECLARE_REG(struct kvm_vcpu *, vcpu, host_ctxt, 1);
	34	+
	35	+ __kvm_adjust_pc(kern_hyp_va(vcpu));
	36	+}
	37	+
31	38	static void handle___kvm_flush_vm_context(struct kvm_cpu_context *host_ctxt)
32	39	{
33	40	__kvm_flush_vm_context();
@@ -170,6 +177,7 @@ typedef void (hcall_t)(struct kvm_cpu_context ); · 136 unmodified lines
170	177
171	178	static const hcall_t host_hcall[] = {
172	179	HANDLE_FUNC(__kvm_vcpu_run),
	180	+ HANDLE_FUNC(__kvm_adjust_pc),
173	181	HANDLE_FUNC(__kvm_flush_vm_context),
174	182	HANDLE_FUNC(__kvm_tlb_flush_vmid_ipa),
175	183	HANDLE_FUNC(__kvm_tlb_flush_vmid),

arch/arm64/kvm/hyp/nvhe/mem_protect.c+2−2

@@ -23,8 +23,8 @@ · 22 unmodified lines
23	23	extern unsigned long hyp_nr_cpus;
24	24	struct host_kvm host_kvm;
25	25
26		−struct hyp_pool host_s2_mem;
27		−struct hyp_pool host_s2_dev;
	26	+static struct hyp_pool host_s2_mem;
	27	+static struct hyp_pool host_s2_dev;
28	28
29	29	/*
30	30	* Copies of the host's CPU features registers holding sanitized values.

arch/arm64/kvm/hyp/nvhe/setup.c+1−1

@@ -17,7 +17,6 @@ · 16 unmodified lines
17	17	#include <nvhe/trap_handler.h>
18	18
19	19	struct hyp_pool hpool;
20		−struct kvm_pgtable_mm_ops pkvm_pgtable_mm_ops;
21	20	unsigned long hyp_nr_cpus;
22	21
23	22	#define hyp_percpu_size ((unsigned long)__per_cpu_end - \
@@ -27,6 +26,7 @@ static void *vmemmap_base; · 3 unmodified lines
27	26	static void *hyp_pgt_base;
28	27	static void *host_s2_mem_pgt_base;
29	28	static void *host_s2_dev_pgt_base;
	29	+static struct kvm_pgtable_mm_ops pkvm_pgtable_mm_ops;
30	30
31	31	static int divide_memory_pool(void *virt, unsigned long size)
32	32	{

arch/arm64/kvm/hyp/nvhe/switch.c+1−2

@@ -4,7 +4,6 @@ · 3 unmodified lines
4	4	* Author: Marc Zyngier <marc.zyngier@arm.com>
5	5	*/
6	6
7		−#include <hyp/adjust_pc.h>
8	7	#include <hyp/switch.h>
9	8	#include <hyp/sysreg-sr.h>
10	9
@@ -201,7 +200,7 @@ int __kvm_vcpu_run(struct kvm_vcpu *vcpu) · 190 unmodified lines
201	200	*/
202	201	__debug_save_host_buffers_nvhe(vcpu);
203	202
204		− __adjust_pc(vcpu);
	203	+ __kvm_adjust_pc(vcpu);
205	204
206	205	/*
207	206	* We must restore the 32-bit state before the sysregs, thanks

arch/arm64/kvm/hyp/vhe/switch.c+1−2

@@ -4,7 +4,6 @@ · 3 unmodified lines
4	4	* Author: Marc Zyngier <marc.zyngier@arm.com>
5	5	*/
6	6
7		−#include <hyp/adjust_pc.h>
8	7	#include <hyp/switch.h>
9	8
10	9	#include <linux/arm-smccc.h>
@@ -132,7 +131,7 @@ static int __kvm_vcpu_run_vhe(struct kvm_vcpu *vcpu) · 121 unmodified lines
132	131	__load_guest_stage2(vcpu->arch.hw_mmu);
133	132	__activate_traps(vcpu);
134	133
135		− __adjust_pc(vcpu);
	134	+ __kvm_adjust_pc(vcpu);
136	135
137	136	sysreg_restore_guest_state_vhe(guest_ctxt);
138	137	__debug_switch_to_guest(vcpu);

arch/arm64/kvm/mmu.c+6−6

@@ -1156,13 +1156,13 @@ out_unlock: · 1155 unmodified lines
1156	1156	bool kvm_unmap_gfn_range(struct kvm kvm, struct kvm_gfn_range range)
1157	1157	{
1158	1158	if (!kvm->arch.mmu.pgt)
1159		− return 0;
	1159	+ return false;
1160	1160
1161	1161	__unmap_stage2_range(&kvm->arch.mmu, range->start << PAGE_SHIFT,
1162	1162	(range->end - range->start) << PAGE_SHIFT,
1163	1163	range->may_block);
1164	1164
1165		− return 0;
	1165	+ return false;
1166	1166	}
1167	1167
1168	1168	bool kvm_set_spte_gfn(struct kvm kvm, struct kvm_gfn_range range)
@@ -1170,7 +1170,7 @@ bool kvm_set_spte_gfn(struct kvm kvm, struct kvm_gfn_range range) · 1 unmodified line
1170	1170	kvm_pfn_t pfn = pte_pfn(range->pte);
1171	1171
1172	1172	if (!kvm->arch.mmu.pgt)
1173		− return 0;
	1173	+ return false;
1174	1174
1175	1175	WARN_ON(range->end - range->start != 1);
1176	1176
@@ -1190,7 +1190,7 @@ bool kvm_set_spte_gfn(struct kvm kvm, struct kvm_gfn_range range) · 13 unmodified lines
1190	1190	PAGE_SIZE, __pfn_to_phys(pfn),
1191	1191	KVM_PGTABLE_PROT_R, NULL);
1192	1192
1193		− return 0;
	1193	+ return false;
1194	1194	}
1195	1195
1196	1196	bool kvm_age_gfn(struct kvm kvm, struct kvm_gfn_range range)
@@ -1200,7 +1200,7 @@ bool kvm_age_gfn(struct kvm kvm, struct kvm_gfn_range range) · 3 unmodified lines
1200	1200	pte_t pte;
1201	1201
1202	1202	if (!kvm->arch.mmu.pgt)
1203		− return 0;
	1203	+ return false;
1204	1204
1205	1205	WARN_ON(size != PAGE_SIZE && size != PMD_SIZE && size != PUD_SIZE);
1206	1206
@@ -1213,7 +1213,7 @@ bool kvm_age_gfn(struct kvm kvm, struct kvm_gfn_range range) · 6 unmodified lines
1213	1213	bool kvm_test_age_gfn(struct kvm kvm, struct kvm_gfn_range range)
1214	1214	{
1215	1215	if (!kvm->arch.mmu.pgt)
1216		− return 0;
	1216	+ return false;
1217	1217
1218	1218	return kvm_pgtable_stage2_is_young(kvm->arch.mmu.pgt,
1219	1219	range->start << PAGE_SHIFT);

arch/arm64/kvm/reset.c+24−4

@@ -166,6 +166,25 @@ static int kvm_vcpu_enable_ptrauth(struct kvm_vcpu *vcpu) · 165 unmodified lines
166	166	return 0;
167	167	}
168	168
	169	+static bool vcpu_allowed_register_width(struct kvm_vcpu *vcpu)
	170	+{
	171	+ struct kvm_vcpu *tmp;
	172	+ bool is32bit;
	173	+ int i;
	174	+
	175	+ is32bit = vcpu_has_feature(vcpu, KVM_ARM_VCPU_EL1_32BIT);
	176	+ if (!cpus_have_const_cap(ARM64_HAS_32BIT_EL1) && is32bit)
	177	+ return false;
	178	+
	179	+ /* Check that the vcpus are either all 32bit or all 64bit */
	180	+ kvm_for_each_vcpu(i, tmp, vcpu->kvm) {
	181	+ if (vcpu_has_feature(tmp, KVM_ARM_VCPU_EL1_32BIT) != is32bit)
	182	+ return false;
	183	+ }
	184	+
	185	+ return true;
	186	+}
	187	+
169	188	/**
170	189	* kvm_reset_vcpu - sets core registers and sys_regs to reset value
171	190	* @vcpu: The VCPU pointer
@@ -217,13 +236,14 @@ int kvm_reset_vcpu(struct kvm_vcpu *vcpu) · 45 unmodified lines
217	236	}
218	237	}
219	238
	239	+ if (!vcpu_allowed_register_width(vcpu)) {
	240	+ ret = -EINVAL;
	241	+ goto out;
	242	+ }
	243	+
220	244	switch (vcpu->arch.target) {
221	245	default:
222	246	if (test_bit(KVM_ARM_VCPU_EL1_32BIT, vcpu->arch.features)) {
223		− if (!cpus_have_const_cap(ARM64_HAS_32BIT_EL1)) {
224		− ret = -EINVAL;
225		− goto out;
226		− }
227	247	pstate = VCPU_RESET_PSTATE_SVC;
228	248	} else {
229	249	pstate = VCPU_RESET_PSTATE_EL1;

arch/arm64/kvm/sys_regs.c+21−21

@@ -399,14 +399,14 @@ static bool trap_bvr(struct kvm_vcpu *vcpu, · 398 unmodified lines
399	399	struct sys_reg_params *p,
400	400	const struct sys_reg_desc *rd)
401	401	{
402		− u64 *dbg_reg = &vcpu->arch.vcpu_debug_state.dbg_bvr[rd->reg];
	402	+ u64 *dbg_reg = &vcpu->arch.vcpu_debug_state.dbg_bvr[rd->CRm];
403	403
404	404	if (p->is_write)
405	405	reg_to_dbg(vcpu, p, rd, dbg_reg);
406	406	else
407	407	dbg_to_reg(vcpu, p, rd, dbg_reg);
408	408
409		− trace_trap_reg(__func__, rd->reg, p->is_write, *dbg_reg);
	409	+ trace_trap_reg(__func__, rd->CRm, p->is_write, *dbg_reg);
410	410
411	411	return true;
412	412	}
@@ -414,7 +414,7 @@ static bool trap_bvr(struct kvm_vcpu *vcpu, · 1 unmodified line
414	414	static int set_bvr(struct kvm_vcpu vcpu, const struct sys_reg_desc rd,
415	415	const struct kvm_one_reg reg, void __user uaddr)
416	416	{
417		− __u64 *r = &vcpu->arch.vcpu_debug_state.dbg_bvr[rd->reg];
	417	+ __u64 *r = &vcpu->arch.vcpu_debug_state.dbg_bvr[rd->CRm];
418	418
419	419	if (copy_from_user(r, uaddr, KVM_REG_SIZE(reg->id)) != 0)
420	420	return -EFAULT;
@@ -424,7 +424,7 @@ static int set_bvr(struct kvm_vcpu vcpu, const struct sys_reg_desc rd, · 3 unmodified lines
424	424	static int get_bvr(struct kvm_vcpu vcpu, const struct sys_reg_desc rd,
425	425	const struct kvm_one_reg reg, void __user uaddr)
426	426	{
427		− __u64 *r = &vcpu->arch.vcpu_debug_state.dbg_bvr[rd->reg];
	427	+ __u64 *r = &vcpu->arch.vcpu_debug_state.dbg_bvr[rd->CRm];
428	428
429	429	if (copy_to_user(uaddr, r, KVM_REG_SIZE(reg->id)) != 0)
430	430	return -EFAULT;
@@ -434,21 +434,21 @@ static int get_bvr(struct kvm_vcpu vcpu, const struct sys_reg_desc rd, · 3 unmodified lines
434	434	static void reset_bvr(struct kvm_vcpu *vcpu,
435	435	const struct sys_reg_desc *rd)
436	436	{
437		− vcpu->arch.vcpu_debug_state.dbg_bvr[rd->reg] = rd->val;
	437	+ vcpu->arch.vcpu_debug_state.dbg_bvr[rd->CRm] = rd->val;
438	438	}
439	439
440	440	static bool trap_bcr(struct kvm_vcpu *vcpu,
441	441	struct sys_reg_params *p,
442	442	const struct sys_reg_desc *rd)
443	443	{
444		− u64 *dbg_reg = &vcpu->arch.vcpu_debug_state.dbg_bcr[rd->reg];
	444	+ u64 *dbg_reg = &vcpu->arch.vcpu_debug_state.dbg_bcr[rd->CRm];
445	445
446	446	if (p->is_write)
447	447	reg_to_dbg(vcpu, p, rd, dbg_reg);
448	448	else
449	449	dbg_to_reg(vcpu, p, rd, dbg_reg);
450	450
451		− trace_trap_reg(__func__, rd->reg, p->is_write, *dbg_reg);
	451	+ trace_trap_reg(__func__, rd->CRm, p->is_write, *dbg_reg);
452	452
453	453	return true;
454	454	}
@@ -456,7 +456,7 @@ static bool trap_bcr(struct kvm_vcpu *vcpu, · 1 unmodified line
456	456	static int set_bcr(struct kvm_vcpu vcpu, const struct sys_reg_desc rd,
457	457	const struct kvm_one_reg reg, void __user uaddr)
458	458	{
459		− __u64 *r = &vcpu->arch.vcpu_debug_state.dbg_bcr[rd->reg];
	459	+ __u64 *r = &vcpu->arch.vcpu_debug_state.dbg_bcr[rd->CRm];
460	460
461	461	if (copy_from_user(r, uaddr, KVM_REG_SIZE(reg->id)) != 0)
462	462	return -EFAULT;
@@ -467,7 +467,7 @@ static int set_bcr(struct kvm_vcpu vcpu, const struct sys_reg_desc rd, · 4 unmodified lines
467	467	static int get_bcr(struct kvm_vcpu vcpu, const struct sys_reg_desc rd,
468	468	const struct kvm_one_reg reg, void __user uaddr)
469	469	{
470		− __u64 *r = &vcpu->arch.vcpu_debug_state.dbg_bcr[rd->reg];
	470	+ __u64 *r = &vcpu->arch.vcpu_debug_state.dbg_bcr[rd->CRm];
471	471
472	472	if (copy_to_user(uaddr, r, KVM_REG_SIZE(reg->id)) != 0)
473	473	return -EFAULT;
@@ -477,22 +477,22 @@ static int get_bcr(struct kvm_vcpu vcpu, const struct sys_reg_desc rd, · 3 unmodified lines
477	477	static void reset_bcr(struct kvm_vcpu *vcpu,
478	478	const struct sys_reg_desc *rd)
479	479	{
480		− vcpu->arch.vcpu_debug_state.dbg_bcr[rd->reg] = rd->val;
	480	+ vcpu->arch.vcpu_debug_state.dbg_bcr[rd->CRm] = rd->val;
481	481	}
482	482
483	483	static bool trap_wvr(struct kvm_vcpu *vcpu,
484	484	struct sys_reg_params *p,
485	485	const struct sys_reg_desc *rd)
486	486	{
487		− u64 *dbg_reg = &vcpu->arch.vcpu_debug_state.dbg_wvr[rd->reg];
	487	+ u64 *dbg_reg = &vcpu->arch.vcpu_debug_state.dbg_wvr[rd->CRm];
488	488
489	489	if (p->is_write)
490	490	reg_to_dbg(vcpu, p, rd, dbg_reg);
491	491	else
492	492	dbg_to_reg(vcpu, p, rd, dbg_reg);
493	493
494		− trace_trap_reg(__func__, rd->reg, p->is_write,
495		− vcpu->arch.vcpu_debug_state.dbg_wvr[rd->reg]);
	494	+ trace_trap_reg(__func__, rd->CRm, p->is_write,
	495	+ vcpu->arch.vcpu_debug_state.dbg_wvr[rd->CRm]);
496	496
497	497	return true;
498	498	}
@@ -500,7 +500,7 @@ static bool trap_wvr(struct kvm_vcpu *vcpu, · 1 unmodified line
500	500	static int set_wvr(struct kvm_vcpu vcpu, const struct sys_reg_desc rd,
501	501	const struct kvm_one_reg reg, void __user uaddr)
502	502	{
503		− __u64 *r = &vcpu->arch.vcpu_debug_state.dbg_wvr[rd->reg];
	503	+ __u64 *r = &vcpu->arch.vcpu_debug_state.dbg_wvr[rd->CRm];
504	504
505	505	if (copy_from_user(r, uaddr, KVM_REG_SIZE(reg->id)) != 0)
506	506	return -EFAULT;
@@ -510,7 +510,7 @@ static int set_wvr(struct kvm_vcpu vcpu, const struct sys_reg_desc rd, · 3 unmodified lines
510	510	static int get_wvr(struct kvm_vcpu vcpu, const struct sys_reg_desc rd,
511	511	const struct kvm_one_reg reg, void __user uaddr)
512	512	{
513		− __u64 *r = &vcpu->arch.vcpu_debug_state.dbg_wvr[rd->reg];
	513	+ __u64 *r = &vcpu->arch.vcpu_debug_state.dbg_wvr[rd->CRm];
514	514
515	515	if (copy_to_user(uaddr, r, KVM_REG_SIZE(reg->id)) != 0)
516	516	return -EFAULT;
@@ -520,21 +520,21 @@ static int get_wvr(struct kvm_vcpu vcpu, const struct sys_reg_desc rd, · 3 unmodified lines
520	520	static void reset_wvr(struct kvm_vcpu *vcpu,
521	521	const struct sys_reg_desc *rd)
522	522	{
523		− vcpu->arch.vcpu_debug_state.dbg_wvr[rd->reg] = rd->val;
	523	+ vcpu->arch.vcpu_debug_state.dbg_wvr[rd->CRm] = rd->val;
524	524	}
525	525
526	526	static bool trap_wcr(struct kvm_vcpu *vcpu,
527	527	struct sys_reg_params *p,
528	528	const struct sys_reg_desc *rd)
529	529	{
530		− u64 *dbg_reg = &vcpu->arch.vcpu_debug_state.dbg_wcr[rd->reg];
	530	+ u64 *dbg_reg = &vcpu->arch.vcpu_debug_state.dbg_wcr[rd->CRm];
531	531
532	532	if (p->is_write)
533	533	reg_to_dbg(vcpu, p, rd, dbg_reg);
534	534	else
535	535	dbg_to_reg(vcpu, p, rd, dbg_reg);
536	536
537		− trace_trap_reg(__func__, rd->reg, p->is_write, *dbg_reg);
	537	+ trace_trap_reg(__func__, rd->CRm, p->is_write, *dbg_reg);
538	538
539	539	return true;
540	540	}
@@ -542,7 +542,7 @@ static bool trap_wcr(struct kvm_vcpu *vcpu, · 1 unmodified line
542	542	static int set_wcr(struct kvm_vcpu vcpu, const struct sys_reg_desc rd,
543	543	const struct kvm_one_reg reg, void __user uaddr)
544	544	{
545		− __u64 *r = &vcpu->arch.vcpu_debug_state.dbg_wcr[rd->reg];
	545	+ __u64 *r = &vcpu->arch.vcpu_debug_state.dbg_wcr[rd->CRm];
546	546
547	547	if (copy_from_user(r, uaddr, KVM_REG_SIZE(reg->id)) != 0)
548	548	return -EFAULT;
@@ -552,7 +552,7 @@ static int set_wcr(struct kvm_vcpu vcpu, const struct sys_reg_desc rd, · 3 unmodified lines
552	552	static int get_wcr(struct kvm_vcpu vcpu, const struct sys_reg_desc rd,
553	553	const struct kvm_one_reg reg, void __user uaddr)
554	554	{
555		− __u64 *r = &vcpu->arch.vcpu_debug_state.dbg_wcr[rd->reg];
	555	+ __u64 *r = &vcpu->arch.vcpu_debug_state.dbg_wcr[rd->CRm];
556	556
557	557	if (copy_to_user(uaddr, r, KVM_REG_SIZE(reg->id)) != 0)
558	558	return -EFAULT;
@@ -562,7 +562,7 @@ static int get_wcr(struct kvm_vcpu vcpu, const struct sys_reg_desc rd, · 3 unmodified lines
562	562	static void reset_wcr(struct kvm_vcpu *vcpu,
563	563	const struct sys_reg_desc *rd)
564	564	{
565		− vcpu->arch.vcpu_debug_state.dbg_wcr[rd->reg] = rd->val;
	565	+ vcpu->arch.vcpu_debug_state.dbg_wcr[rd->CRm] = rd->val;
566	566	}
567	567
568	568	static void reset_amair_el1(struct kvm_vcpu vcpu, const struct sys_reg_desc r)

arch/arm64/mm/mmu.c+2−1

@@ -515,7 +515,8 @@ static void __init map_mem(pgd_t *pgdp) · 514 unmodified lines
515	515	*/
516	516	BUILD_BUG_ON(pgd_index(direct_map_end - 1) == pgd_index(direct_map_end));
517	517
518		− if (rodata_full \|\| crash_mem_map \|\| debug_pagealloc_enabled())
	518	+ if (rodata_full \|\| crash_mem_map \|\| debug_pagealloc_enabled() \|\|
	519	+ IS_ENABLED(CONFIG_KFENCE))
519	520	flags \|= NO_BLOCK_MAPPINGS \| NO_CONT_MAPPINGS;
520	521
521	522	/*

arch/mips/alchemy/board-xxs1500.c+1−0

@@ -18,6 +18,7 @@ · 17 unmodified lines
18	18	#include <asm/reboot.h>
19	19	#include <asm/setup.h>
20	20	#include <asm/mach-au1x00/au1000.h>
	21	+#include <asm/mach-au1x00/gpio-au1000.h>
21	22	#include <prom.h>
22	23
23	24	const char *get_system_type(void)

arch/mips/include/asm/mips-boards/launch.h+5−0

@@ -3,6 +3,9 @@ · 2 unmodified lines
3	3	*
4	4	*/
5	5
	6	+#ifndef _ASM_MIPS_BOARDS_LAUNCH_H
	7	+#define _ASM_MIPS_BOARDS_LAUNCH_H
	8	+
6	9	#ifndef _ASSEMBLER_
7	10
8	11	struct cpulaunch {
@@ -34,3 +37,5 @@ struct cpulaunch { · 25 unmodified lines
34	37
35	38	/* Polling period in count cycles for secondary CPU's */
36	39	#define LAUNCHPERIOD 10000
	40	+
	41	+#endif /* _ASM_MIPS_BOARDS_LAUNCH_H */

arch/mips/lib/mips-atomic.c+6−6

@@ -37,7 +37,7 @@ · 36 unmodified lines
37	37	*/
38	38	notrace void arch_local_irq_disable(void)
39	39	{
40		− preempt_disable();
	40	+ preempt_disable_notrace();
41	41
42	42	__asm__ __volatile__(
43	43	" .set push \n"
@@ -53,7 +53,7 @@ notrace void arch_local_irq_disable(void) · 9 unmodified lines
53	53	: /* no inputs */
54	54	: "memory");
55	55
56		− preempt_enable();
	56	+ preempt_enable_notrace();
57	57	}
58	58	EXPORT_SYMBOL(arch_local_irq_disable);
59	59
@@ -61,7 +61,7 @@ notrace unsigned long arch_local_irq_save(void) · 1 unmodified line
61	61	{
62	62	unsigned long flags;
63	63
64		− preempt_disable();
	64	+ preempt_disable_notrace();
65	65
66	66	__asm__ __volatile__(
67	67	" .set push \n"
@@ -78,7 +78,7 @@ notrace unsigned long arch_local_irq_save(void) · 10 unmodified lines
78	78	: /* no inputs */
79	79	: "memory");
80	80
81		− preempt_enable();
	81	+ preempt_enable_notrace();
82	82
83	83	return flags;
84	84	}
@@ -88,7 +88,7 @@ notrace void arch_local_irq_restore(unsigned long flags) · 3 unmodified lines
88	88	{
89	89	unsigned long __tmp1;
90	90
91		− preempt_disable();
	91	+ preempt_disable_notrace();
92	92
93	93	__asm__ __volatile__(
94	94	" .set push \n"
@@ -106,7 +106,7 @@ notrace void arch_local_irq_restore(unsigned long flags) · 11 unmodified lines
106	106	: "0" (flags)
107	107	: "memory");
108	108
109		− preempt_enable();
	109	+ preempt_enable_notrace();
110	110	}
111	111	EXPORT_SYMBOL(arch_local_irq_restore);
112	112

arch/mips/mm/cache.c+14−16

@@ -158,31 +158,29 @@ unsigned long _page_cachable_default; · 157 unmodified lines
158	158	EXPORT_SYMBOL(_page_cachable_default);
159	159
160	160	#define PM(p) __pgprot(_page_cachable_default \| (p))
161		−#define PVA(p) PM(_PAGE_VALID \| _PAGE_ACCESSED \| (p))
162	161
163	162	static inline void setup_protection_map(void)
164	163	{
165	164	protection_map[0] = PM(_PAGE_PRESENT \| _PAGE_NO_EXEC \| _PAGE_NO_READ);
166		− protection_map[1] = PVA(_PAGE_PRESENT \| _PAGE_NO_EXEC);
167		− protection_map[2] = PVA(_PAGE_PRESENT \| _PAGE_NO_EXEC \| _PAGE_NO_READ);
168		− protection_map[3] = PVA(_PAGE_PRESENT \| _PAGE_NO_EXEC);
169		− protection_map[4] = PVA(_PAGE_PRESENT);
170		− protection_map[5] = PVA(_PAGE_PRESENT);
171		− protection_map[6] = PVA(_PAGE_PRESENT);
172		− protection_map[7] = PVA(_PAGE_PRESENT);
	165	+ protection_map[1] = PM(_PAGE_PRESENT \| _PAGE_NO_EXEC);
	166	+ protection_map[2] = PM(_PAGE_PRESENT \| _PAGE_NO_EXEC \| _PAGE_NO_READ);
	167	+ protection_map[3] = PM(_PAGE_PRESENT \| _PAGE_NO_EXEC);
	168	+ protection_map[4] = PM(_PAGE_PRESENT);
	169	+ protection_map[5] = PM(_PAGE_PRESENT);
	170	+ protection_map[6] = PM(_PAGE_PRESENT);
	171	+ protection_map[7] = PM(_PAGE_PRESENT);
173	172
174	173	protection_map[8] = PM(_PAGE_PRESENT \| _PAGE_NO_EXEC \| _PAGE_NO_READ);
175		− protection_map[9] = PVA(_PAGE_PRESENT \| _PAGE_NO_EXEC);
176		− protection_map[10] = PVA(_PAGE_PRESENT \| _PAGE_NO_EXEC \| _PAGE_WRITE \|
	174	+ protection_map[9] = PM(_PAGE_PRESENT \| _PAGE_NO_EXEC);
	175	+ protection_map[10] = PM(_PAGE_PRESENT \| _PAGE_NO_EXEC \| _PAGE_WRITE \|
177	176	_PAGE_NO_READ);
178		− protection_map[11] = PVA(_PAGE_PRESENT \| _PAGE_NO_EXEC \| _PAGE_WRITE);
179		− protection_map[12] = PVA(_PAGE_PRESENT);
180		− protection_map[13] = PVA(_PAGE_PRESENT);
181		− protection_map[14] = PVA(_PAGE_PRESENT);
182		− protection_map[15] = PVA(_PAGE_PRESENT);
	177	+ protection_map[11] = PM(_PAGE_PRESENT \| _PAGE_NO_EXEC \| _PAGE_WRITE);
	178	+ protection_map[12] = PM(_PAGE_PRESENT);
	179	+ protection_map[13] = PM(_PAGE_PRESENT);
	180	+ protection_map[14] = PM(_PAGE_PRESENT \| _PAGE_WRITE);
	181	+ protection_map[15] = PM(_PAGE_PRESENT \| _PAGE_WRITE);
183	182	}
184	183
185		−#undef _PVA
186	184	#undef PM
187	185
188	186	void cpu_cache_init(void)

arch/mips/ralink/of.c+2−0

@@ -8,6 +8,7 @@ · 7 unmodified lines
8	8
9	9	#include <linux/io.h>
10	10	#include <linux/clk.h>
	11	+#include <linux/export.h>
11	12	#include <linux/init.h>
12	13	#include <linux/sizes.h>
13	14	#include <linux/of_fdt.h>
@@ -25,6 +26,7 @@ · 11 unmodified lines
25	26
26	27	__iomem void *rt_sysc_membase;
27	28	__iomem void *rt_memc_membase;
	29	+EXPORT_SYMBOL_GPL(rt_sysc_membase);
28	30
29	31	__iomem void plat_of_remap_node(const char node)
30	32	{

arch/powerpc/boot/dts/fsl/p1010si-post.dtsi+8−0

@@ -122,7 +122,15 @@ · 121 unmodified lines
122	122	};
123	123
124	124	/include/ "pq3-i2c-0.dtsi"
	125	+ i2c@3000 {
	126	+ fsl,i2c-erratum-a004447;
	127	+ };
	128	+
125	129	/include/ "pq3-i2c-1.dtsi"
	130	+ i2c@3100 {
	131	+ fsl,i2c-erratum-a004447;
	132	+ };
	133	+
126	134	/include/ "pq3-duart-0.dtsi"
127	135	/include/ "pq3-espi-0.dtsi"
128	136	spi0: spi@7000 {

arch/powerpc/boot/dts/fsl/p2041si-post.dtsi+16−0

@@ -371,7 +371,23 @@ · 370 unmodified lines
371	371	};
372	372
373	373	/include/ "qoriq-i2c-0.dtsi"
	374	+ i2c@118000 {
	375	+ fsl,i2c-erratum-a004447;
	376	+ };
	377	+
	378	+ i2c@118100 {
	379	+ fsl,i2c-erratum-a004447;
	380	+ };
	381	+
374	382	/include/ "qoriq-i2c-1.dtsi"
	383	+ i2c@119000 {
	384	+ fsl,i2c-erratum-a004447;
	385	+ };
	386	+
	387	+ i2c@119100 {
	388	+ fsl,i2c-erratum-a004447;
	389	+ };
	390	+
375	391	/include/ "qoriq-duart-0.dtsi"
376	392	/include/ "qoriq-duart-1.dtsi"
377	393	/include/ "qoriq-gpio-0.dtsi"

arch/powerpc/include/asm/jump_label.h+1−1

@@ -50,7 +50,7 @@ l_yes: · 49 unmodified lines
50	50	1098: nop; \
51	51	.pushsection __jump_table, "aw"; \
52	52	.long 1098b - ., LABEL - .; \
53		− FTR_ENTRY_LONG KEY; \
	53	+ FTR_ENTRY_LONG KEY - .; \
54	54	.popsection
55	55	#endif
56	56

arch/powerpc/include/asm/kvm_host.h+1−0

@@ -51,6 +51,7 @@ · 50 unmodified lines
51	51	/* PPC-specific vcpu->requests bit members */
52	52	#define KVM_REQ_WATCHDOG KVM_ARCH_REQ(0)
53	53	#define KVM_REQ_EPR_EXIT KVM_ARCH_REQ(1)
	54	+#define KVM_REQ_PENDING_TIMER KVM_ARCH_REQ(2)
54	55
55	56	#include <linux/mmu_notifier.h>
56	57

arch/powerpc/include/asm/pte-walk.h+29−0

@@ -31,6 +31,35 @@ static inline pte_t find_init_mm_pte(unsigned long ea, unsigned hshift) · 30 unmodified lines
31	31	pgd_t *pgdir = init_mm.pgd;
32	32	return __find_linux_pte(pgdir, ea, NULL, hshift);
33	33	}
	34	+
	35	+/*
	36	+ * Convert a kernel vmap virtual address (vmalloc or ioremap space) to a
	37	+ * physical address, without taking locks. This can be used in real-mode.
	38	+ */
	39	+static inline phys_addr_t ppc_find_vmap_phys(unsigned long addr)
	40	+{
	41	+ pte_t *ptep;
	42	+ phys_addr_t pa;
	43	+ int hugepage_shift;
	44	+
	45	+ /*
	46	+ * init_mm does not free page tables, and does not do THP. It may
	47	+ * have huge pages from huge vmalloc / ioremap etc.
	48	+ */
	49	+ ptep = find_init_mm_pte(addr, &hugepage_shift);
	50	+ if (WARN_ON(!ptep))
	51	+ return 0;
	52	+
	53	+ pa = PFN_PHYS(pte_pfn(*ptep));
	54	+
	55	+ if (!hugepage_shift)
	56	+ hugepage_shift = PAGE_SHIFT;
	57	+
	58	+ pa \|= addr & ((1ul << hugepage_shift) - 1);
	59	+
	60	+ return pa;
	61	+}
	62	+
34	63	/*
35	64	* This is what we should always use. Any other lockless page table lookup needs
36	65	* careful audit against THP split.

arch/powerpc/kernel/eeh.c+1−22

@@ -346,28 +346,7 @@ void eeh_slot_error_detail(struct eeh_pe *pe, int severity) · 345 unmodified lines
346	346	*/
347	347	static inline unsigned long eeh_token_to_phys(unsigned long token)
348	348	{
349		− pte_t *ptep;
350		− unsigned long pa;
351		− int hugepage_shift;
352		−
353		− /*
354		− * We won't find hugepages here(this is iomem). Hence we are not
355		− * worried about _PAGE_SPLITTING/collapse. Also we will not hit
356		− * page table free, because of init_mm.
357		− */
358		− ptep = find_init_mm_pte(token, &hugepage_shift);
359		− if (!ptep)
360		− return token;
361		−
362		− pa = pte_pfn(*ptep);
363		−
364		− /* On radix we can do hugepage mappings for io, so handle that */
365		− if (!hugepage_shift)
366		− hugepage_shift = PAGE_SHIFT;
367		−
368		− pa <<= PAGE_SHIFT;
369		− pa \|= token & ((1ul << hugepage_shift) - 1);
370		− return pa;
	349	+ return ppc_find_vmap_phys(token);
371	350	}
372	351
373	352	/*

arch/powerpc/kernel/io-workarounds.c+3−13

@@ -55,7 +55,6 @@ static struct iowa_bus *iowa_pci_find(unsigned long vaddr, unsigned long paddr) · 54 unmodified lines
55	55	#ifdef CONFIG_PPC_INDIRECT_MMIO
56	56	struct iowa_bus *iowa_mem_find_bus(const PCI_IO_ADDR addr)
57	57	{
58		− unsigned hugepage_shift;
59	58	struct iowa_bus *bus;
60	59	int token;
61	60
@@ -65,22 +64,13 @@ struct iowa_bus *iowa_mem_find_bus(const PCI_IO_ADDR addr) · 3 unmodified lines
65	64	bus = &iowa_busses[token - 1];
66	65	else {
67	66	unsigned long vaddr, paddr;
68		− pte_t *ptep;
69	67
70	68	vaddr = (unsigned long)PCI_FIX_ADDR(addr);
71	69	if (vaddr < PHB_IO_BASE \|\| vaddr >= PHB_IO_END)
72	70	return NULL;
73		− /*
74		− * We won't find huge pages here (iomem). Also can't hit
75		− * a page table free due to init_mm
76		− */
77		− ptep = find_init_mm_pte(vaddr, &hugepage_shift);
78		− if (ptep == NULL)
79		− paddr = 0;
80		− else {
81		− WARN_ON(hugepage_shift);
82		− paddr = pte_pfn(*ptep) << PAGE_SHIFT;
83		− }
	71	+
	72	+ paddr = ppc_find_vmap_phys(vaddr);
	73	+
84	74	bus = iowa_pci_find(vaddr, paddr);
85	75
86	76	if (bus == NULL)

arch/powerpc/kernel/iommu.c+5−6

@@ -898,7 +898,6 @@ void iommu_alloc_coherent(struct device dev, struct iommu_table *tbl, · 897 unmodified lines
898	898	unsigned int order;
899	899	unsigned int nio_pages, io_order;
900	900	struct page *page;
901		− size_t size_io = size;
902	901
903	902	size = PAGE_ALIGN(size);
904	903	order = get_order(size);
@@ -925,9 +924,8 @@ void iommu_alloc_coherent(struct device dev, struct iommu_table *tbl, · 20 unmodified lines
925	924	memset(ret, 0, size);
926	925
927	926	/* Set up tces to cover the allocated range */
928		− size_io = IOMMU_PAGE_ALIGN(size_io, tbl);
929		− nio_pages = size_io >> tbl->it_page_shift;
930		− io_order = get_iommu_order(size_io, tbl);
	927	+ nio_pages = size >> tbl->it_page_shift;
	928	+ io_order = get_iommu_order(size, tbl);
931	929	mapping = iommu_alloc(dev, tbl, ret, nio_pages, DMA_BIDIRECTIONAL,
932	930	mask >> tbl->it_page_shift, io_order, 0);
933	931	if (mapping == DMA_MAPPING_ERROR) {
@@ -942,9 +940,10 @@ void iommu_free_coherent(struct iommu_table *tbl, size_t size, · 8 unmodified lines
942	940	void *vaddr, dma_addr_t dma_handle)
943	941	{
944	942	if (tbl) {
945		− size_t size_io = IOMMU_PAGE_ALIGN(size, tbl);
946		− unsigned int nio_pages = size_io >> tbl->it_page_shift;
	943	+ unsigned int nio_pages;
947	944
	945	+ size = PAGE_ALIGN(size);
	946	+ nio_pages = size >> tbl->it_page_shift;
948	947	iommu_free(tbl, dma_handle, nio_pages);
949	948	size = PAGE_ALIGN(size);
950	949	free_pages((unsigned long)vaddr, get_order(size));

arch/powerpc/kernel/kprobes.c+2−2

@@ -108,7 +108,6 @@ int arch_prepare_kprobe(struct kprobe *p) · 107 unmodified lines
108	108	int ret = 0;
109	109	struct kprobe *prev;
110	110	struct ppc_inst insn = ppc_inst_read((struct ppc_inst *)p->addr);
111		− struct ppc_inst prefix = ppc_inst_read((struct ppc_inst *)(p->addr - 1));
112	111
113	112	if ((unsigned long)p->addr & 0x03) {
114	113	printk("Attempt to register kprobe at an unaligned address\n");
@@ -116,7 +115,8 @@ int arch_prepare_kprobe(struct kprobe *p) · 1 unmodified line
116	115	} else if (IS_MTMSRD(insn) \|\| IS_RFID(insn) \|\| IS_RFI(insn)) {
117	116	printk("Cannot register a kprobe on rfi/rfid or mtmsr[d]\n");
118	117	ret = -EINVAL;
119		− } else if (ppc_inst_prefixed(prefix)) {
	118	+ } else if ((unsigned long)p->addr & ~PAGE_MASK &&
	119	+ ppc_inst_prefixed(ppc_inst_read((struct ppc_inst *)(p->addr - 1)))) {
120	120	printk("Cannot register a kprobe on the second word of prefixed instruction\n");
121	121	ret = -EINVAL;
122	122	}

arch/powerpc/kernel/signal_64.c+4−5

@@ -902,6 +902,10 @@ int handle_rt_signal64(struct ksignal ksig, sigset_t set, · 901 unmodified lines
902	902	unsafe_copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set), badframe_block);
903	903	user_write_access_end();
904	904
	905	+ /* Save the siginfo outside of the unsafe block. */
	906	+ if (copy_siginfo_to_user(&frame->info, &ksig->info))
	907	+ goto badframe;
	908	+
905	909	/* Make sure signal handler doesn't get spurious FP exceptions */
906	910	tsk->thread.fp_state.fpscr = 0;
907	911
@@ -915,11 +919,6 @@ int handle_rt_signal64(struct ksignal ksig, sigset_t set, · 7 unmodified lines
915	919	regs->nip = (unsigned long) &frame->tramp[0];
916	920	}
917	921
918		−
919		− /* Save the siginfo outside of the unsafe block. */
920		− if (copy_siginfo_to_user(&frame->info, &ksig->info))
921		− goto badframe;
922		−
923	922	/* Allocate a dummy caller frame for the signal handler. */
924	923	newsp = ((unsigned long)frame) - __SIGNAL_FRAMESIZE;
925	924	err \|= put_user(regs->gpr[1], (unsigned long __user *)newsp);

arch/powerpc/kvm/book3s_hv.c+1−2

@@ -3936,7 +3936,7 @@ static void kvmppc_vcore_blocked(struct kvmppc_vcore *vc) · 3935 unmodified lines
3936	3936	break;
3937	3937	}
3938	3938	cur = ktime_get();
3939		− } while (single_task_running() && ktime_before(cur, stop));
	3939	+ } while (kvm_vcpu_can_poll(cur, stop));
3940	3940
3941	3941	spin_lock(&vc->lock);
3942	3942	vc->vcore_state = VCORE_INACTIVE;
@@ -4455,7 +4455,6 @@ static int kvmppc_vcpu_run_hv(struct kvm_vcpu *vcpu) · 512 unmodified lines
4455	4455	mtspr(SPRN_EBBRR, ebb_regs[1]);
4456	4456	mtspr(SPRN_BESCR, ebb_regs[2]);
4457	4457	mtspr(SPRN_TAR, user_tar);
4458		− mtspr(SPRN_FSCR, current->thread.fscr);
4459	4458	}
4460	4459	mtspr(SPRN_VRSAVE, user_vrsave);
4461	4460

arch/powerpc/kvm/book3s_hv_rm_mmu.c+2−13

@@ -23,20 +23,9 @@ · 22 unmodified lines
23	23	#include <asm/pte-walk.h>
24	24
25	25	/* Translate address of a vmalloc'd thing to a linear map address */
26		−static void real_vmalloc_addr(void x)
	26	+static void real_vmalloc_addr(void addr)
27	27	{
28		− unsigned long addr = (unsigned long) x;
29		− pte_t *p;
30		− /*
31		− * assume we don't have huge pages in vmalloc space...
32		− * So don't worry about THP collapse/split. Called
33		− * Only in realmode with MSR_EE = 0, hence won't need irq_save/restore.
34		− */
35		− p = find_init_mm_pte(addr, NULL);
36		− if (!p \|\| !pte_present(*p))
37		− return NULL;
38		− addr = (pte_pfn(*p) << PAGE_SHIFT) \| (addr & ~PAGE_MASK);
39		− return __va(addr);
	28	+ return __va(ppc_find_vmap_phys((unsigned long)addr));
40	29	}
41	30
42	31	/* Return 1 if we need to do a global tlbie, 0 if we can use tlbiel */

arch/powerpc/kvm/book3s_hv_rmhandlers.S+7−0

@@ -59,6 +59,7 @@ END_FTR_SECTION_IFCLR(CPU_FTR_ARCH_300) · 58 unmodified lines
59	59	#define STACK_SLOT_UAMOR (SFS-88)
60	60	#define STACK_SLOT_DAWR1 (SFS-96)
61	61	#define STACK_SLOT_DAWRX1 (SFS-104)
	62	+#define STACK_SLOT_FSCR (SFS-112)
62	63	/* the following is used by the P9 short path */
63	64	#define STACK_SLOT_NVGPRS (SFS-152) /* 18 gprs */
64	65
@@ -686,6 +687,8 @@ BEGIN_FTR_SECTION · 621 unmodified lines
686	687	std r6, STACK_SLOT_DAWR0(r1)
687	688	std r7, STACK_SLOT_DAWRX0(r1)
688	689	std r8, STACK_SLOT_IAMR(r1)
	690	+ mfspr r5, SPRN_FSCR
	691	+ std r5, STACK_SLOT_FSCR(r1)
689	692	END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S)
690	693	BEGIN_FTR_SECTION
691	694	mfspr r6, SPRN_DAWR1
@@ -1663,6 +1666,10 @@ FTR_SECTION_ELSE · 971 unmodified lines
1663	1666	ld r7, STACK_SLOT_HFSCR(r1)
1664	1667	mtspr SPRN_HFSCR, r7
1665	1668	ALT_FTR_SECTION_END_IFCLR(CPU_FTR_ARCH_300)
	1669	+BEGIN_FTR_SECTION
	1670	+ ld r5, STACK_SLOT_FSCR(r1)
	1671	+ mtspr SPRN_FSCR, r5
	1672	+END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S)
1666	1673	/*
1667	1674	* Restore various registers to 0, where non-zero values
1668	1675	* set by the guest could disrupt the host.

arch/powerpc/mm/mem.c+1−0

@@ -20,6 +20,7 @@ · 19 unmodified lines
20	20	#include <asm/machdep.h>
21	21	#include <asm/rtas.h>
22	22	#include <asm/kasan.h>
	23	+#include <asm/sparsemem.h>
23	24	#include <asm/svm.h>
24	25
25	26	#include <mm/mmu_decl.h>

arch/powerpc/perf/core-book3s.c+1−1

@@ -2254,7 +2254,7 @@ unsigned long perf_instruction_pointer(struct pt_regs *regs) · 2253 unmodified lines
2254	2254	bool use_siar = regs_use_siar(regs);
2255	2255	unsigned long siar = mfspr(SPRN_SIAR);
2256	2256
2257		− if (ppmu->flags & PPMU_P10_DD1) {
	2257	+ if (ppmu && (ppmu->flags & PPMU_P10_DD1)) {
2258	2258	if (siar)
2259	2259	return siar;
2260	2260	else

arch/riscv/Kconfig+10−9

@@ -34,6 +34,7 @@ config RISCV · 33 unmodified lines
34	34	select ARCH_OPTIONAL_KERNEL_RWX if ARCH_HAS_STRICT_KERNEL_RWX
35	35	select ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
36	36	select ARCH_SUPPORTS_HUGETLBFS if MMU
	37	+ select ARCH_USE_MEMTEST
37	38	select ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT if MMU
38	39	select ARCH_WANT_FRAME_POINTERS
39	40	select ARCH_WANT_HUGE_PMD_SHARE if 64BIT
@@ -60,11 +61,11 @@ config RISCV · 20 unmodified lines
60	61	select GENERIC_TIME_VSYSCALL if MMU && 64BIT
61	62	select HANDLE_DOMAIN_IRQ
62	63	select HAVE_ARCH_AUDITSYSCALL
63		− select HAVE_ARCH_JUMP_LABEL
64		− select HAVE_ARCH_JUMP_LABEL_RELATIVE
	64	+ select HAVE_ARCH_JUMP_LABEL if !XIP_KERNEL
	65	+ select HAVE_ARCH_JUMP_LABEL_RELATIVE if !XIP_KERNEL
65	66	select HAVE_ARCH_KASAN if MMU && 64BIT
66	67	select HAVE_ARCH_KASAN_VMALLOC if MMU && 64BIT
67		− select HAVE_ARCH_KGDB
	68	+ select HAVE_ARCH_KGDB if !XIP_KERNEL
68	69	select HAVE_ARCH_KGDB_QXFER_PKT
69	70	select HAVE_ARCH_MMAP_RND_BITS if MMU
70	71	select HAVE_ARCH_SECCOMP_FILTER
@@ -79,9 +80,9 @@ config RISCV · 8 unmodified lines
79	80	select HAVE_GCC_PLUGINS
80	81	select HAVE_GENERIC_VDSO if MMU && 64BIT
81	82	select HAVE_IRQ_TIME_ACCOUNTING
82		− select HAVE_KPROBES
83		− select HAVE_KPROBES_ON_FTRACE
84		− select HAVE_KRETPROBES
	83	+ select HAVE_KPROBES if !XIP_KERNEL
	84	+ select HAVE_KPROBES_ON_FTRACE if !XIP_KERNEL
	85	+ select HAVE_KRETPROBES if !XIP_KERNEL
85	86	select HAVE_PCI
86	87	select HAVE_PERF_EVENTS
87	88	select HAVE_PERF_REGS
@@ -230,11 +231,11 @@ config ARCH_RV64I · 142 unmodified lines
230	231	bool "RV64I"
231	232	select 64BIT
232	233	select ARCH_SUPPORTS_INT128 if CC_HAS_INT128 && GCC_VERSION >= 50000
233		− select HAVE_DYNAMIC_FTRACE if MMU && $(cc-option,-fpatchable-function-entry=8)
	234	+ select HAVE_DYNAMIC_FTRACE if !XIP_KERNEL && MMU && $(cc-option,-fpatchable-function-entry=8)
234	235	select HAVE_DYNAMIC_FTRACE_WITH_REGS if HAVE_DYNAMIC_FTRACE
235		− select HAVE_FTRACE_MCOUNT_RECORD
	236	+ select HAVE_FTRACE_MCOUNT_RECORD if !XIP_KERNEL
236	237	select HAVE_FUNCTION_GRAPH_TRACER
237		− select HAVE_FUNCTION_TRACER
	238	+ select HAVE_FUNCTION_TRACER if !XIP_KERNEL
238	239	select SWIOTLB if MMU
239	240
240	241	endchoice

arch/riscv/Kconfig.socs+1−0

@@ -14,6 +14,7 @@ config SOC_SIFIVE · 13 unmodified lines
14	14	select CLK_SIFIVE
15	15	select CLK_SIFIVE_PRCI
16	16	select SIFIVE_PLIC
	17	+ select RISCV_ERRATA_ALTERNATIVE
17	18	select ERRATA_SIFIVE
18	19	help
19	20	This enables support for SiFive SoC platform hardware.

arch/riscv/Makefile+10−1

@@ -16,7 +16,7 @@ ifeq ($(CONFIG_DYNAMIC_FTRACE),y) · 15 unmodified lines
16	16	CC_FLAGS_FTRACE := -fpatchable-function-entry=8
17	17	endif
18	18
19		−ifeq ($(CONFIG_64BIT)$(CONFIG_CMODEL_MEDLOW),yy)
	19	+ifeq ($(CONFIG_CMODEL_MEDLOW),y)
20	20	KBUILD_CFLAGS_MODULE += -mcmodel=medany
21	21	endif
22	22
@@ -38,6 +38,15 @@ else · 15 unmodified lines
38	38	KBUILD_LDFLAGS += -melf32lriscv
39	39	endif
40	40
	41	+ifeq ($(CONFIG_LD_IS_LLD),y)
	42	+ KBUILD_CFLAGS += -mno-relax
	43	+ KBUILD_AFLAGS += -mno-relax
	44	+ifneq ($(LLVM_IAS),1)
	45	+ KBUILD_CFLAGS += -Wa,-mno-relax
	46	+ KBUILD_AFLAGS += -Wa,-mno-relax
	47	+endif
	48	+endif
	49	+
41	50	# ISA string setting
42	51	riscv-march-$(CONFIG_ARCH_RV32I) := rv32ima
43	52	riscv-march-$(CONFIG_ARCH_RV64I) := rv64ima

arch/riscv/boot/dts/microchip/Makefile+1−0

@@ -1,2 +1,3 @@
1	1	# SPDX-License-Identifier: GPL-2.0
2	2	dtb-$(CONFIG_SOC_MICROCHIP_POLARFIRE) += microchip-mpfs-icicle-kit.dtb
	3	+obj-$(CONFIG_BUILTIN_DTB) += $(addsuffix .o, $(dtb-y))

arch/riscv/boot/dts/sifive/Makefile+1−0

@@ -1,3 +1,4 @@
1	1	# SPDX-License-Identifier: GPL-2.0
2	2	dtb-$(CONFIG_SOC_SIFIVE) += hifive-unleashed-a00.dtb \
3	3	hifive-unmatched-a00.dtb
	4	+obj-$(CONFIG_BUILTIN_DTB) += $(addsuffix .o, $(dtb-y))

arch/riscv/boot/dts/sifive/fu740-c000.dtsi+1−1

@@ -273,7 +273,7 @@ · 272 unmodified lines
273	273	cache-size = <2097152>;
274	274	cache-unified;
275	275	interrupt-parent = <&plic0>;
276		− interrupts = <19 20 21 22>;
	276	+ interrupts = <19 21 22 20>;
277	277	reg = <0x0 0x2010000 0x0 0x1000>;
278	278	};
279	279	gpio: gpio@10060000 {

arch/riscv/errata/sifive/Makefile+1−1

@@ -1,2 +1,2 @@
1		−obj-y += errata_cip_453.o
	1	+obj-$(CONFIG_ERRATA_SIFIVE_CIP_453) += errata_cip_453.o
2	2	obj-y += errata.o

arch/riscv/include/asm/alternative-macros.h+2−2

@@ -51,7 +51,7 @@ · 50 unmodified lines
51	51	REG_ASM " " newlen "\n" \
52	52	".word " errata_id "\n"
53	53
54		−#define ALT_NEW_CONSTENT(vendor_id, errata_id, enable, new_c) \
	54	+#define ALT_NEW_CONTENT(vendor_id, errata_id, enable, new_c) \
55	55	".if " __stringify(enable) " == 1\n" \
56	56	".pushsection .alternative, \"a\"\n" \
57	57	ALT_ENTRY("886b", "888f", __stringify(vendor_id), __stringify(errata_id), "889f - 888f") \
@@ -69,7 +69,7 @@ · 11 unmodified lines
69	69	"886 :\n" \
70	70	old_c "\n" \
71	71	"887 :\n" \
72		− ALT_NEW_CONSTENT(vendor_id, errata_id, enable, new_c)
	72	+ ALT_NEW_CONTENT(vendor_id, errata_id, enable, new_c)
73	73
74	74	#define _ALTERNATIVE_CFG(old_c, new_c, vendor_id, errata_id, CONFIG_k) \
75	75	__ALTERNATIVE_CFG(old_c, new_c, vendor_id, errata_id, IS_ENABLED(CONFIG_k))

arch/riscv/include/asm/kexec.h+2−2

@@ -42,8 +42,8 @@ struct kimage_arch { · 41 unmodified lines
42	42	unsigned long fdt_addr;
43	43	};
44	44
45		−const extern unsigned char riscv_kexec_relocate[];
46		−const extern unsigned int riscv_kexec_relocate_size;
	45	+extern const unsigned char riscv_kexec_relocate[];
	46	+extern const unsigned int riscv_kexec_relocate_size;
47	47
48	48	typedef void (*riscv_kexec_method)(unsigned long first_ind_entry,
49	49	unsigned long jump_addr,

arch/riscv/include/asm/pgtable.h+2−3

@@ -30,9 +30,8 @@ · 29 unmodified lines
30	30
31	31	#define BPF_JIT_REGION_SIZE (SZ_128M)
32	32	#ifdef CONFIG_64BIT
33		−/* KASLR should leave at least 128MB for BPF after the kernel */
34		−#define BPF_JIT_REGION_START PFN_ALIGN((unsigned long)&_end)
35		−#define BPF_JIT_REGION_END (BPF_JIT_REGION_START + BPF_JIT_REGION_SIZE)
	33	+#define BPF_JIT_REGION_START (BPF_JIT_REGION_END - BPF_JIT_REGION_SIZE)
	34	+#define BPF_JIT_REGION_END (MODULES_END)
36	35	#else
37	36	#define BPF_JIT_REGION_START (PAGE_OFFSET - BPF_JIT_REGION_SIZE)
38	37	#define BPF_JIT_REGION_END (VMALLOC_END)

arch/riscv/kernel/machine_kexec.c+6−5

@@ -14,8 +14,9 @@ · 13 unmodified lines
14	14	#include <asm/set_memory.h> /* For set_memory_x() */
15	15	#include <linux/compiler.h> /* For unreachable() */
16	16	#include <linux/cpu.h> /* For cpu_down() */
	17	+#include <linux/reboot.h>
17	18
18		−/**
	19	+/*
19	20	* kexec_image_info - Print received image details
20	21	*/
21	22	static void
@@ -39,7 +40,7 @@ kexec_image_info(const struct kimage *image) · 17 unmodified lines
39	40	}
40	41	}
41	42
42		−/**
	43	+/*
43	44	* machine_kexec_prepare - Initialize kexec
44	45	*
45	46	* This function is called from do_kexec_load, when the user has
@@ -100,7 +101,7 @@ machine_kexec_prepare(struct kimage *image) · 54 unmodified lines
100	101	}
101	102
102	103
103		−/**
	104	+/*
104	105	* machine_kexec_cleanup - Cleanup any leftovers from
105	106	* machine_kexec_prepare
106	107	*
@@ -135,7 +136,7 @@ void machine_shutdown(void) · 28 unmodified lines
135	136	#endif
136	137	}
137	138
138		−/**
	139	+/*
139	140	* machine_crash_shutdown - Prepare to kexec after a kernel crash
140	141	*
141	142	* This function is called by crash_kexec just before machine_kexec
@@ -151,7 +152,7 @@ machine_crash_shutdown(struct pt_regs *regs) · 9 unmodified lines
151	152	pr_info("Starting crashdump kernel...\n");
152	153	}
153	154
154		−/**
	155	+/*
155	156	* machine_kexec - Jump to the loaded kimage
156	157	*
157	158	* This function is called by kernel_kexec which is called by the

arch/riscv/kernel/probes/kprobes.c+2−0

@@ -84,6 +84,7 @@ int __kprobes arch_prepare_kprobe(struct kprobe *p) · 83 unmodified lines
84	84	return 0;
85	85	}
86	86
	87	+#ifdef CONFIG_MMU
87	88	void *alloc_insn_page(void)
88	89	{
89	90	return __vmalloc_node_range(PAGE_SIZE, 1, VMALLOC_START, VMALLOC_END,
@@ -91,6 +92,7 @@ void *alloc_insn_page(void) · 1 unmodified line
91	92	VM_FLUSH_RESET_PERMS, NUMA_NO_NODE,
92	93	__builtin_return_address(0));
93	94	}
	95	+#endif
94	96
95	97	/* install breakpoint in text */
96	98	void __kprobes arch_arm_kprobe(struct kprobe *p)

arch/riscv/kernel/setup.c+2−2

@@ -231,13 +231,13 @@ static void __init init_resources(void) · 230 unmodified lines
231	231
232	232	/* Clean-up any unused pre-allocated resources */
233	233	mem_res_sz = (num_resources - res_idx + 1) * sizeof(*mem_res);
234		− memblock_free((phys_addr_t) mem_res, mem_res_sz);
	234	+ memblock_free(__pa(mem_res), mem_res_sz);
235	235	return;
236	236
237	237	error:
238	238	/* Better an empty resource tree than an inconsistent one */
239	239	release_child_resources(&iomem_resource);
240		− memblock_free((phys_addr_t) mem_res, mem_res_sz);
	240	+ memblock_free(__pa(mem_res), mem_res_sz);
241	241	}
242	242
243	243

arch/riscv/kernel/stacktrace.c+7−7

@@ -27,10 +27,10 @@ void notrace walk_stackframe(struct task_struct task, struct pt_regs regs, · 26 unmodified lines
27	27	fp = frame_pointer(regs);
28	28	sp = user_stack_pointer(regs);
29	29	pc = instruction_pointer(regs);
30		− } else if (task == NULL \|\| task == current) {
31		− fp = (unsigned long)__builtin_frame_address(0);
32		− sp = sp_in_global;
33		− pc = (unsigned long)walk_stackframe;
	30	+ } else if (task == current) {
	31	+ fp = (unsigned long)__builtin_frame_address(1);
	32	+ sp = (unsigned long)__builtin_frame_address(0);
	33	+ pc = (unsigned long)__builtin_return_address(0);
34	34	} else {
35	35	/* task blocked in __switch_to */
36	36	fp = task->thread.s[0];
@@ -106,15 +106,15 @@ static bool print_trace_address(void *arg, unsigned long pc) · 69 unmodified lines
106	106	return true;
107	107	}
108	108
109		−void dump_backtrace(struct pt_regs regs, struct task_struct task,
	109	+noinline void dump_backtrace(struct pt_regs regs, struct task_struct task,
110	110	const char *loglvl)
111	111	{
112		− pr_cont("%sCall Trace:\n", loglvl);
113	112	walk_stackframe(task, regs, print_trace_address, (void *)loglvl);
114	113	}
115	114
116	115	void show_stack(struct task_struct task, unsigned long sp, const char *loglvl)
117	116	{
	117	+ pr_cont("%sCall Trace:\n", loglvl);
118	118	dump_backtrace(NULL, task, loglvl);
119	119	}
120	120
@@ -139,7 +139,7 @@ unsigned long get_wchan(struct task_struct *task) · 18 unmodified lines
139	139
140	140	#ifdef CONFIG_STACKTRACE
141	141
142		−void arch_stack_walk(stack_trace_consume_fn consume_entry, void *cookie,
	142	+noinline void arch_stack_walk(stack_trace_consume_fn consume_entry, void *cookie,
143	143	struct task_struct task, struct pt_regs regs)
144	144	{
145	145	walk_stackframe(task, regs, consume_entry, cookie);

arch/riscv/kernel/traps.c+9−4

@@ -86,8 +86,13 @@ static void do_trap_error(struct pt_regs *regs, int signo, int code, · 85 unmodified lines
86	86	}
87	87	}
88	88
	89	+#if defined (CONFIG_XIP_KERNEL) && defined (CONFIG_RISCV_ERRATA_ALTERNATIVE)
	90	+#define __trap_section __section(".xip.traps")
	91	+#else
	92	+#define __trap_section
	93	+#endif
89	94	#define DO_ERROR_INFO(name, signo, code, str) \
90		−asmlinkage __visible void name(struct pt_regs *regs) \
	95	+asmlinkage __visible __trap_section void name(struct pt_regs *regs) \
91	96	{ \
92	97	do_trap_error(regs, signo, code, regs->epc, "Oops - " str); \
93	98	}
@@ -111,7 +116,7 @@ DO_ERROR_INFO(do_trap_store_misaligned, · 17 unmodified lines
111	116	int handle_misaligned_load(struct pt_regs *regs);
112	117	int handle_misaligned_store(struct pt_regs *regs);
113	118
114		−asmlinkage void do_trap_load_misaligned(struct pt_regs *regs)
	119	+asmlinkage void __trap_section do_trap_load_misaligned(struct pt_regs *regs)
115	120	{
116	121	if (!handle_misaligned_load(regs))
117	122	return;
@@ -119,7 +124,7 @@ asmlinkage void do_trap_load_misaligned(struct pt_regs *regs) · 1 unmodified line
119	124	"Oops - load address misaligned");
120	125	}
121	126
122		−asmlinkage void do_trap_store_misaligned(struct pt_regs *regs)
	127	+asmlinkage void __trap_section do_trap_store_misaligned(struct pt_regs *regs)
123	128	{
124	129	if (!handle_misaligned_store(regs))
125	130	return;
@@ -146,7 +151,7 @@ static inline unsigned long get_break_insn_length(unsigned long pc) · 20 unmodified lines
146	151	return GET_INSN_LENGTH(insn);
147	152	}
148	153
149		−asmlinkage __visible void do_trap_break(struct pt_regs *regs)
	154	+asmlinkage __visible __trap_section void do_trap_break(struct pt_regs *regs)
150	155	{
151	156	#ifdef CONFIG_KPROBES
152	157	if (kprobe_single_step_handler(regs))

arch/riscv/kernel/vmlinux-xip.lds.S+14−1

@@ -99,9 +99,22 @@ SECTIONS · 98 unmodified lines
99	99	}
100	100	PERCPU_SECTION(L1_CACHE_BYTES)
101	101
102		− . = ALIGN(PAGE_SIZE);
	102	+ . = ALIGN(8);
	103	+ .alternative : {
	104	+ __alt_start = .;
	105	+ *(.alternative)
	106	+ __alt_end = .;
	107	+ }
103	108	__init_end = .;
104	109
	110	+ . = ALIGN(16);
	111	+ .xip.traps : {
	112	+ __xip_traps_start = .;
	113	+ *(.xip.traps)
	114	+ __xip_traps_end = .;
	115	+ }
	116	+
	117	+ . = ALIGN(PAGE_SIZE);
105	118	.sdata : {
106	119	__global_pointer$ = . + 0x800;
107	120	(.sdata)

arch/riscv/mm/init.c+6−2

@@ -746,14 +746,18 @@ void __init protect_kernel_text_data(void) · 745 unmodified lines
746	746	unsigned long init_data_start = (unsigned long)__init_data_begin;
747	747	unsigned long rodata_start = (unsigned long)__start_rodata;
748	748	unsigned long data_start = (unsigned long)_data;
749		− unsigned long max_low = (unsigned long)(__va(PFN_PHYS(max_low_pfn)));
	749	+#if defined(CONFIG_64BIT) && defined(CONFIG_MMU)
	750	+ unsigned long end_va = kernel_virt_addr + load_sz;
	751	+#else
	752	+ unsigned long end_va = (unsigned long)(__va(PFN_PHYS(max_low_pfn)));
	753	+#endif
750	754
751	755	set_memory_ro(text_start, (init_text_start - text_start) >> PAGE_SHIFT);
752	756	set_memory_ro(init_text_start, (init_data_start - init_text_start) >> PAGE_SHIFT);
753	757	set_memory_nx(init_data_start, (rodata_start - init_data_start) >> PAGE_SHIFT);
754	758	/* rodata section is marked readonly in mark_rodata_ro */
755	759	set_memory_nx(rodata_start, (data_start - rodata_start) >> PAGE_SHIFT);
756		− set_memory_nx(data_start, (max_low - data_start) >> PAGE_SHIFT);
	760	+ set_memory_nx(data_start, (end_va - data_start) >> PAGE_SHIFT);
757	761	}
758	762
759	763	void mark_rodata_ro(void)

arch/riscv/mm/kasan_init.c+5−5

@@ -169,7 +169,7 @@ static void __init kasan_shallow_populate(void start, void end) · 168 unmodified lines
169	169
170	170	void __init kasan_init(void)
171	171	{
172		− phys_addr_t _start, _end;
	172	+ phys_addr_t p_start, p_end;
173	173	u64 i;
174	174
175	175	/*
@@ -189,9 +189,9 @@ void __init kasan_init(void) · 13 unmodified lines
189	189	(void )kasan_mem_to_shadow((void )VMALLOC_END));
190	190
191	191	/* Populate the linear mapping */
192		− for_each_mem_range(i, &_start, &_end) {
193		− void start = (void )__va(_start);
194		− void end = (void )__va(_end);
	192	+ for_each_mem_range(i, &p_start, &p_end) {
	193	+ void start = (void )__va(p_start);
	194	+ void end = (void )__va(p_end);
195	195
196	196	if (start >= end)
197	197	break;
@@ -201,7 +201,7 @@ void __init kasan_init(void) · 3 unmodified lines
201	201
202	202	/* Populate kernel, BPF, modules mapping */
203	203	kasan_populate(kasan_mem_to_shadow((const void *)MODULES_VADDR),
204		− kasan_mem_to_shadow((const void *)BPF_JIT_REGION_END));
	204	+ kasan_mem_to_shadow((const void *)MODULES_VADDR + SZ_2G));
205	205
206	206	for (i = 0; i < PTRS_PER_PTE; i++)
207	207	set_pte(&kasan_early_shadow_pte[i],

arch/s390/kernel/entry.S+2−2

@@ -651,9 +651,9 @@ ENDPROC(stack_overflow) · 650 unmodified lines
651	651	.Lcleanup_sie_mcck:
652	652	larl %r13,.Lsie_entry
653	653	slgr %r9,%r13
654		− larl %r13,.Lsie_skip
	654	+ lghi %r13,.Lsie_skip - .Lsie_entry
655	655	clgr %r9,%r13
656		− jh .Lcleanup_sie_int
	656	+ jhe .Lcleanup_sie_int
657	657	oi __LC_CPU_FLAGS+7, _CIF_MCCK_GUEST
658	658	.Lcleanup_sie_int:
659	659	BPENTER __SF_SIE_FLAGS(%r15),(_TIF_ISOLATE_BP\|_TIF_ISOLATE_BP_GUEST)

arch/x86/Makefile+3−2

@@ -200,8 +200,9 @@ endif · 199 unmodified lines
200	200	KBUILD_LDFLAGS += -m elf_$(UTS_MACHINE)
201	201
202	202	ifdef CONFIG_LTO_CLANG
203		−KBUILD_LDFLAGS += -plugin-opt=-code-model=kernel \
204		− -plugin-opt=-stack-alignment=$(if $(CONFIG_X86_32),4,8)
	203	+ifeq ($(shell test $(CONFIG_LLD_VERSION) -lt 130000; echo $$?),0)
	204	+KBUILD_LDFLAGS += -plugin-opt=-stack-alignment=$(if $(CONFIG_X86_32),4,8)
	205	+endif
205	206	endif
206	207
207	208	ifdef CONFIG_X86_NEED_RELOCS

arch/x86/events/intel/uncore_snbep.c+6−3

@@ -1406,6 +1406,8 @@ static int snbep_pci2phy_map_init(int devid, int nodeid_loc, int idmap_loc, bool · 1405 unmodified lines
1406	1406	die_id = i;
1407	1407	else
1408	1408	die_id = topology_phys_to_logical_pkg(i);
	1409	+ if (die_id < 0)
	1410	+ die_id = -ENODEV;
1409	1411	map->pbus_to_dieid[bus] = die_id;
1410	1412	break;
1411	1413	}
@@ -1452,14 +1454,14 @@ static int snbep_pci2phy_map_init(int devid, int nodeid_loc, int idmap_loc, bool · 40 unmodified lines
1452	1454	i = -1;
1453	1455	if (reverse) {
1454	1456	for (bus = 255; bus >= 0; bus--) {
1455		− if (map->pbus_to_dieid[bus] >= 0)
	1457	+ if (map->pbus_to_dieid[bus] != -1)
1456	1458	i = map->pbus_to_dieid[bus];
1457	1459	else
1458	1460	map->pbus_to_dieid[bus] = i;
1459	1461	}
1460	1462	} else {
1461	1463	for (bus = 0; bus <= 255; bus++) {
1462		− if (map->pbus_to_dieid[bus] >= 0)
	1464	+ if (map->pbus_to_dieid[bus] != -1)
1463	1465	i = map->pbus_to_dieid[bus];
1464	1466	else
1465	1467	map->pbus_to_dieid[bus] = i;
@@ -5097,9 +5099,10 @@ static struct intel_uncore_type icx_uncore_m2m = { · 3631 unmodified lines
5097	5099	.perf_ctr = SNR_M2M_PCI_PMON_CTR0,
5098	5100	.event_ctl = SNR_M2M_PCI_PMON_CTL0,
5099	5101	.event_mask = SNBEP_PMON_RAW_EVENT_MASK,
	5102	+ .event_mask_ext = SNR_M2M_PCI_PMON_UMASK_EXT,
5100	5103	.box_ctl = SNR_M2M_PCI_PMON_BOX_CTL,
5101	5104	.ops = &snr_m2m_uncore_pci_ops,
5102		− .format_group = &skx_uncore_format_group,
	5105	+ .format_group = &snr_m2m_uncore_format_group,
5103	5106	};
5104	5107
5105	5108	static struct attribute *icx_upi_uncore_formats_attr[] = {

arch/x86/include/asm/apic.h+1−0

@@ -174,6 +174,7 @@ static inline int apic_is_clustered_box(void) · 173 unmodified lines
174	174	extern int setup_APIC_eilvt(u8 lvt_off, u8 vector, u8 msg_type, u8 mask);
175	175	extern void lapic_assign_system_vectors(void);
176	176	extern void lapic_assign_legacy_vector(unsigned int isairq, bool replace);
	177	+extern void lapic_update_legacy_vectors(void);
177	178	extern void lapic_online(void);
178	179	extern void lapic_offline(void);
179	180	extern bool apic_needs_pit(void);

arch/x86/include/asm/disabled-features.h+2−5

@@ -56,11 +56,8 @@ · 55 unmodified lines
56	56	# define DISABLE_PTI (1 << (X86_FEATURE_PTI & 31))
57	57	#endif
58	58
59		−#ifdef CONFIG_IOMMU_SUPPORT
60		−# define DISABLE_ENQCMD 0
61		−#else
62		−# define DISABLE_ENQCMD (1 << (X86_FEATURE_ENQCMD & 31))
63		−#endif
	59	+/* Force disable because it's broken beyond repair */
	60	+#define DISABLE_ENQCMD (1 << (X86_FEATURE_ENQCMD & 31))
64	61
65	62	#ifdef CONFIG_X86_SGX
66	63	# define DISABLE_SGX 0

arch/x86/include/asm/fpu/api.h+1−5

@@ -106,10 +106,6 @@ extern int cpu_has_xfeatures(u64 xfeatures_mask, const char **feature_name); · 105 unmodified lines
106	106	*/
107	107	#define PASID_DISABLED 0
108	108
109		−#ifdef CONFIG_IOMMU_SUPPORT
110		−/* Update current's PASID MSR/state by mm's PASID. */
111		−void update_pasid(void);
112		−#else
113	109	static inline void update_pasid(void) { }
114		−#endif
	110	+
115	111	#endif /* _ASM_X86_FPU_API_H */

arch/x86/include/asm/fpu/internal.h+10−10

@@ -578,19 +578,19 @@ static inline void switch_fpu_finish(struct fpu *new_fpu) · 577 unmodified lines
578	578	* PKRU state is switched eagerly because it needs to be valid before we
579	579	* return to userland e.g. for a copy_to_user() operation.
580	580	*/
581		− if (current->mm) {
	581	+ if (!(current->flags & PF_KTHREAD)) {
	582	+ /*
	583	+ * If the PKRU bit in xsave.header.xfeatures is not set,
	584	+ * then the PKRU component was in init state, which means
	585	+ * XRSTOR will set PKRU to 0. If the bit is not set then
	586	+ * get_xsave_addr() will return NULL because the PKRU value
	587	+ * in memory is not valid. This means pkru_val has to be
	588	+ * set to 0 and not to init_pkru_value.
	589	+ */
582	590	pk = get_xsave_addr(&new_fpu->state.xsave, XFEATURE_PKRU);
583		− if (pk)
584		− pkru_val = pk->pkru;
	591	+ pkru_val = pk ? pk->pkru : 0;
585	592	}
586	593	__write_pkru(pkru_val);
587		−
588		− /*
589		− * Expensive PASID MSR write will be avoided in update_pasid() because
590		− * TIF_NEED_FPU_LOAD was set. And the PASID state won't be updated
591		− * unless it's different from mm->pasid to reduce overhead.
592		− */
593		− update_pasid();
594	594	}
595	595
596	596	#endif /* _ASM_X86_FPU_INTERNAL_H */

arch/x86/include/asm/kvm-x86-ops.h+1−0

@@ -99,6 +99,7 @@ KVM_X86_OP_NULL(post_block) · 98 unmodified lines
99	99	KVM_X86_OP_NULL(vcpu_blocking)
100	100	KVM_X86_OP_NULL(vcpu_unblocking)
101	101	KVM_X86_OP_NULL(update_pi_irte)
	102	+KVM_X86_OP_NULL(start_assignment)
102	103	KVM_X86_OP_NULL(apicv_post_state_restore)
103	104	KVM_X86_OP_NULL(dy_apicv_has_pending_interrupt)
104	105	KVM_X86_OP_NULL(set_hv_timer)

arch/x86/include/asm/kvm_host.h+1−0

@@ -1352,6 +1352,7 @@ struct kvm_x86_ops { · 1351 unmodified lines
1352	1352
1353	1353	int (update_pi_irte)(struct kvm kvm, unsigned int host_irq,
1354	1354	uint32_t guest_irq, bool set);
	1355	+ void (start_assignment)(struct kvm kvm);
1355	1356	void (apicv_post_state_restore)(struct kvm_vcpu vcpu);
1356	1357	bool (dy_apicv_has_pending_interrupt)(struct kvm_vcpu vcpu);
1357	1358

arch/x86/include/asm/thermal.h+3−1

@@ -3,11 +3,13 @@ · 2 unmodified lines
3	3	#define _ASM_X86_THERMAL_H
4	4
5	5	#ifdef CONFIG_X86_THERMAL_VECTOR
	6	+void therm_lvt_init(void);
6	7	void intel_init_thermal(struct cpuinfo_x86 *c);
7	8	bool x86_thermal_enabled(void);
8	9	void intel_thermal_interrupt(void);
9	10	#else
10		−static inline void intel_init_thermal(struct cpuinfo_x86 *c) { }
	11	+static inline void therm_lvt_init(void) { }
	12	+static inline void intel_init_thermal(struct cpuinfo_x86 *c) { }
11	13	#endif
12	14
13	15	#endif /* _ASM_X86_THERMAL_H */

arch/x86/kernel/alternative.c+46−18

@@ -182,42 +182,70 @@ done: · 181 unmodified lines
182	182	n_dspl, (unsigned long)orig_insn + n_dspl + repl_len);
183	183	}
184	184
	185	+/*
	186	+ * optimize_nops_range() - Optimize a sequence of single byte NOPs (0x90)
	187	+ *
	188	+ * @instr: instruction byte stream
	189	+ * @instrlen: length of the above
	190	+ * @off: offset within @instr where the first NOP has been detected
	191	+ *
	192	+ * Return: number of NOPs found (and replaced).
	193	+ */
	194	+static __always_inline int optimize_nops_range(u8 *instr, u8 instrlen, int off)
	195	+{
	196	+ unsigned long flags;
	197	+ int i = off, nnops;
	198	+
	199	+ while (i < instrlen) {
	200	+ if (instr[i] != 0x90)
	201	+ break;
	202	+
	203	+ i++;
	204	+ }
	205	+
	206	+ nnops = i - off;
	207	+
	208	+ if (nnops <= 1)
	209	+ return nnops;
	210	+
	211	+ local_irq_save(flags);
	212	+ add_nops(instr + off, nnops);
	213	+ local_irq_restore(flags);
	214	+
	215	+ DUMP_BYTES(instr, instrlen, "%px: [%d:%d) optimized NOPs: ", instr, off, i);
	216	+
	217	+ return nnops;
	218	+}
	219	+
185	220	/*
186	221	* "noinline" to cause control flow change and thus invalidate I$ and
187	222	* cause refetch after modification.
188	223	*/
189	224	static void __init_or_module noinline optimize_nops(struct alt_instr a, u8 instr)
190	225	{
191		− unsigned long flags;
192	226	struct insn insn;
193		− int nop, i = 0;
	227	+ int i = 0;
194	228
195	229	/*
196		− * Jump over the non-NOP insns, the remaining bytes must be single-byte
197		− * NOPs, optimize them.
	230	+ * Jump over the non-NOP insns and optimize single-byte NOPs into bigger
	231	+ * ones.
198	232	*/
199	233	for (;;) {
200	234	if (insn_decode_kernel(&insn, &instr[i]))
201	235	return;
202	236
	237	+ /*
	238	+ * See if this and any potentially following NOPs can be
	239	+ * optimized.
	240	+ */
203	241	if (insn.length == 1 && insn.opcode.bytes[0] == 0x90)
204		− break;
205		−
206		− if ((i += insn.length) >= a->instrlen)
207		− return;
208		− }
	242	+ i += optimize_nops_range(instr, a->instrlen, i);
	243	+ else
	244	+ i += insn.length;
209	245
210		− for (nop = i; i < a->instrlen; i++) {
211		− if (WARN_ONCE(instr[i] != 0x90, "Not a NOP at 0x%px\n", &instr[i]))
	246	+ if (i >= a->instrlen)
212	247	return;
213	248	}
214		−
215		− local_irq_save(flags);
216		− add_nops(instr + nop, i - nop);
217		− local_irq_restore(flags);
218		−
219		− DUMP_BYTES(instr, a->instrlen, "%px: [%d:%d) optimized NOPs: ",
220		− instr, nop, a->instrlen);
221	249	}
222	250
223	251	/*

arch/x86/kernel/apic/apic.c+1−0

@@ -2604,6 +2604,7 @@ static void __init apic_bsp_setup(bool upmode) · 2603 unmodified lines
2604	2604	end_local_APIC_setup();
2605	2605	irq_remap_enable_fault_handling();
2606	2606	setup_IO_APIC();
	2607	+ lapic_update_legacy_vectors();
2607	2608	}
2608	2609
2609	2610	#ifdef CONFIG_UP_LATE_INIT

arch/x86/kernel/apic/vector.c+20−0

@@ -738,6 +738,26 @@ void lapic_assign_legacy_vector(unsigned int irq, bool replace) · 737 unmodified lines
738	738	irq_matrix_assign_system(vector_matrix, ISA_IRQ_VECTOR(irq), replace);
739	739	}
740	740
	741	+void __init lapic_update_legacy_vectors(void)
	742	+{
	743	+ unsigned int i;
	744	+
	745	+ if (IS_ENABLED(CONFIG_X86_IO_APIC) && nr_ioapics > 0)
	746	+ return;
	747	+
	748	+ /*
	749	+ * If the IO/APIC is disabled via config, kernel command line or
	750	+ * lack of enumeration then all legacy interrupts are routed
	751	+ * through the PIC. Make sure that they are marked as legacy
	752	+ * vectors. PIC_CASCADE_IRQ has already been marked in
	753	+ * lapic_assign_system_vectors().
	754	+ */
	755	+ for (i = 0; i < nr_legacy_irqs(); i++) {
	756	+ if (i != PIC_CASCADE_IR)
	757	+ lapic_assign_legacy_vector(i, true);
	758	+ }
	759	+}
	760	+
741	761	void __init lapic_assign_system_vectors(void)
742	762	{
743	763	unsigned int i, vector = 0;

arch/x86/kernel/cpu/perfctr-watchdog.c+2−2

@@ -63,7 +63,7 @@ static inline unsigned int nmi_perfctr_msr_to_bit(unsigned int msr) · 62 unmodified lines
63	63	case 15:
64	64	return msr - MSR_P4_BPU_PERFCTR0;
65	65	}
66		− fallthrough;
	66	+ break;
67	67	case X86_VENDOR_ZHAOXIN:
68	68	case X86_VENDOR_CENTAUR:
69	69	return msr - MSR_ARCH_PERFMON_PERFCTR0;
@@ -96,7 +96,7 @@ static inline unsigned int nmi_evntsel_msr_to_bit(unsigned int msr) · 26 unmodified lines
96	96	case 15:
97	97	return msr - MSR_P4_BSU_ESCR0;
98	98	}
99		− fallthrough;
	99	+ break;
100	100	case X86_VENDOR_ZHAOXIN:
101	101	case X86_VENDOR_CENTAUR:
102	102	return msr - MSR_ARCH_PERFMON_EVENTSEL0;

arch/x86/kernel/cpu/sgx/virt.c+1−0

@@ -212,6 +212,7 @@ static int sgx_vepc_release(struct inode inode, struct file file) · 211 unmodified lines
212	212	list_splice_tail(&secs_pages, &zombie_secs_pages);
213	213	mutex_unlock(&zombie_secs_pages_lock);
214	214
	215	+ xa_destroy(&vepc->page_array);
215	216	kfree(vepc);
216	217
217	218	return 0;

arch/x86/kernel/fpu/signal.c+35−19

@@ -307,13 +307,17 @@ static int __fpu__restore_sig(void __user buf, void __user buf_fx, int size) · 306 unmodified lines
307	307	return 0;
308	308	}
309	309
310		− if (!access_ok(buf, size))
311		− return -EACCES;
	310	+ if (!access_ok(buf, size)) {
	311	+ ret = -EACCES;
	312	+ goto out;
	313	+ }
312	314
313		− if (!static_cpu_has(X86_FEATURE_FPU))
314		− return fpregs_soft_set(current, NULL,
315		− 0, sizeof(struct user_i387_ia32_struct),
316		− NULL, buf) != 0;
	315	+ if (!static_cpu_has(X86_FEATURE_FPU)) {
	316	+ ret = fpregs_soft_set(current, NULL, 0,
	317	+ sizeof(struct user_i387_ia32_struct),
	318	+ NULL, buf);
	319	+ goto out;
	320	+ }
317	321
318	322	if (use_xsave()) {
319	323	struct _fpx_sw_bytes fx_sw_user;
@@ -369,6 +373,25 @@ static int __fpu__restore_sig(void __user buf, void __user buf_fx, int size) · 49 unmodified lines
369	373	fpregs_unlock();
370	374	return 0;
371	375	}
	376	+
	377	+ /*
	378	+ * The above did an FPU restore operation, restricted to
	379	+ * the user portion of the registers, and failed, but the
	380	+ * microcode might have modified the FPU registers
	381	+ * nevertheless.
	382	+ *
	383	+ * If the FPU registers do not belong to current, then
	384	+ * invalidate the FPU register state otherwise the task might
	385	+ * preempt current and return to user space with corrupted
	386	+ * FPU registers.
	387	+ *
	388	+ * In case current owns the FPU registers then no further
	389	+ * action is required. The fixup below will handle it
	390	+ * correctly.
	391	+ */
	392	+ if (test_thread_flag(TIF_NEED_FPU_LOAD))
	393	+ __cpu_invalidate_fpregs_state();
	394	+
372	395	fpregs_unlock();
373	396	} else {
374	397	/*
@@ -377,7 +400,7 @@ static int __fpu__restore_sig(void __user buf, void __user buf_fx, int size) · 2 unmodified lines
377	400	*/
378	401	ret = __copy_from_user(&env, buf, sizeof(env));
379	402	if (ret)
380		− goto err_out;
	403	+ goto out;
381	404	envp = &env;
382	405	}
383	406
@@ -405,16 +428,9 @@ static int __fpu__restore_sig(void __user buf, void __user buf_fx, int size) · 21 unmodified lines
405	428	if (use_xsave() && !fx_only) {
406	429	u64 init_bv = xfeatures_mask_user() & ~user_xfeatures;
407	430
408		− if (using_compacted_format()) {
409		− ret = copy_user_to_xstate(&fpu->state.xsave, buf_fx);
410		− } else {
411		− ret = __copy_from_user(&fpu->state.xsave, buf_fx, state_size);
412		−
413		− if (!ret && state_size > offsetof(struct xregs_state, header))
414		− ret = validate_user_xstate_header(&fpu->state.xsave.header);
415		− }
	431	+ ret = copy_user_to_xstate(&fpu->state.xsave, buf_fx);
416	432	if (ret)
417		− goto err_out;
	433	+ goto out;
418	434
419	435	sanitize_restored_user_xstate(&fpu->state, envp, user_xfeatures,
420	436	fx_only);
@@ -434,7 +450,7 @@ static int __fpu__restore_sig(void __user buf, void __user buf_fx, int size) · 13 unmodified lines
434	450	ret = __copy_from_user(&fpu->state.fxsave, buf_fx, state_size);
435	451	if (ret) {
436	452	ret = -EFAULT;
437		− goto err_out;
	453	+ goto out;
438	454	}
439	455
440	456	sanitize_restored_user_xstate(&fpu->state, envp, user_xfeatures,
@@ -452,7 +468,7 @@ static int __fpu__restore_sig(void __user buf, void __user buf_fx, int size) · 11 unmodified lines
452	468	} else {
453	469	ret = __copy_from_user(&fpu->state.fsave, buf_fx, state_size);
454	470	if (ret)
455		− goto err_out;
	471	+ goto out;
456	472
457	473	fpregs_lock();
458	474	ret = copy_kernel_to_fregs_err(&fpu->state.fsave);
@@ -463,7 +479,7 @@ static int __fpu__restore_sig(void __user buf, void __user buf_fx, int size) · 4 unmodified lines
463	479	fpregs_deactivate(fpu);
464	480	fpregs_unlock();
465	481
466		−err_out:
	482	+out:
467	483	if (ret)
468	484	fpu__clear_user_states(fpu);
469	485	return ret;

arch/x86/kernel/fpu/xstate.c+0−57

@@ -1402,60 +1402,3 @@ int proc_pid_arch_status(struct seq_file m, struct pid_namespace ns, · 1401 unmodified lines
1402	1402	return 0;
1403	1403	}
1404	1404	#endif /* CONFIG_PROC_PID_ARCH_STATUS */
1405		−
1406		−#ifdef CONFIG_IOMMU_SUPPORT
1407		−void update_pasid(void)
1408		−{
1409		− u64 pasid_state;
1410		− u32 pasid;
1411		−
1412		− if (!cpu_feature_enabled(X86_FEATURE_ENQCMD))
1413		− return;
1414		−
1415		− if (!current->mm)
1416		− return;
1417		−
1418		− pasid = READ_ONCE(current->mm->pasid);
1419		− /* Set the valid bit in the PASID MSR/state only for valid pasid. */
1420		− pasid_state = pasid == PASID_DISABLED ?
1421		− pasid : pasid \| MSR_IA32_PASID_VALID;
1422		−
1423		− /*
1424		− * No need to hold fregs_lock() since the task's fpstate won't
1425		− * be changed by others (e.g. ptrace) while the task is being
1426		− * switched to or is in IPI.
1427		− */
1428		− if (!test_thread_flag(TIF_NEED_FPU_LOAD)) {
1429		− /* The MSR is active and can be directly updated. */
1430		− wrmsrl(MSR_IA32_PASID, pasid_state);
1431		− } else {
1432		− struct fpu *fpu = &current->thread.fpu;
1433		− struct ia32_pasid_state *ppasid_state;
1434		− struct xregs_state *xsave;
1435		−
1436		− /*
1437		− * The CPU's xstate registers are not currently active. Just
1438		− * update the PASID state in the memory buffer here. The
1439		− * PASID MSR will be loaded when returning to user mode.
1440		− */
1441		− xsave = &fpu->state.xsave;
1442		− xsave->header.xfeatures \|= XFEATURE_MASK_PASID;
1443		− ppasid_state = get_xsave_addr(xsave, XFEATURE_PASID);
1444		− /*
1445		− * Since XFEATURE_MASK_PASID is set in xfeatures, ppasid_state
1446		− * won't be NULL and no need to check its value.
1447		− *
1448		− * Only update the task's PASID state when it's different
1449		− * from the mm's pasid.
1450		− */
1451		− if (ppasid_state->pasid != pasid_state) {
1452		− /*
1453		− * Invalid fpregs so that state restoring will pick up
1454		− * the PASID state.
1455		− */
1456		− __fpu_invalidate_fpregs_state(fpu);
1457		− ppasid_state->pasid = pasid_state;
1458		− }
1459		− }
1460		−}
1461		−#endif /* CONFIG_IOMMU_SUPPORT */

arch/x86/kernel/setup.c+30−14

@@ -44,6 +44,7 @@ · 43 unmodified lines
44	44	#include <asm/pci-direct.h>
45	45	#include <asm/prom.h>
46	46	#include <asm/proto.h>
	47	+#include <asm/thermal.h>
47	48	#include <asm/unwind.h>
48	49	#include <asm/vsyscall.h>
49	50	#include <linux/vmalloc.h>
@@ -637,11 +638,11 @@ static void __init trim_snb_memory(void) · 587 unmodified lines
637	638	* them from accessing certain memory ranges, namely anything below
638	639	* 1M and in the pages listed in bad_pages[] above.
639	640	*
640		− * To avoid these pages being ever accessed by SNB gfx devices
641		− * reserve all memory below the 1 MB mark and bad_pages that have
642		− * not already been reserved at boot time.
	641	+ * To avoid these pages being ever accessed by SNB gfx devices reserve
	642	+ * bad_pages that have not already been reserved at boot time.
	643	+ * All memory below the 1 MB mark is anyway reserved later during
	644	+ * setup_arch(), so there is no need to reserve it here.
643	645	*/
644		− memblock_reserve(0, 1<<20);
645	646
646	647	for (i = 0; i < ARRAY_SIZE(bad_pages); i++) {
647	648	if (memblock_reserve(bad_pages[i], PAGE_SIZE))
@@ -733,14 +734,14 @@ static void __init early_reserve_memory(void) · 85 unmodified lines
733	734	* The first 4Kb of memory is a BIOS owned area, but generally it is
734	735	* not listed as such in the E820 table.
735	736	*
736		− * Reserve the first memory page and typically some additional
737		− * memory (64KiB by default) since some BIOSes are known to corrupt
738		− * low memory. See the Kconfig help text for X86_RESERVE_LOW.
	737	+ * Reserve the first 64K of memory since some BIOSes are known to
	738	+ * corrupt low memory. After the real mode trampoline is allocated the
	739	+ * rest of the memory below 640k is reserved.
739	740	*
740	741	* In addition, make sure page 0 is always reserved because on
741	742	* systems with L1TF its contents can be leaked to user processes.
742	743	*/
743		− memblock_reserve(0, ALIGN(reserve_low, PAGE_SIZE));
	744	+ memblock_reserve(0, SZ_64K);
744	745
745	746	early_reserve_initrd();
746	747
@@ -751,6 +752,7 @@ static void __init early_reserve_memory(void) · 4 unmodified lines
751	752
752	753	reserve_ibft_region();
753	754	reserve_bios_regions();
	755	+ trim_snb_memory();
754	756	}
755	757
756	758	/*
@@ -1081,14 +1083,20 @@ void __init setup_arch(char **cmdline_p) · 324 unmodified lines
1081	1083	(max_pfn_mapped<<PAGE_SHIFT) - 1);
1082	1084	#endif
1083	1085
1084		− reserve_real_mode();
1085		−
1086	1086	/*
1087		− * Reserving memory causing GPU hangs on Sandy Bridge integrated
1088		− * graphics devices should be done after we allocated memory under
1089		− * 1M for the real mode trampoline.
	1087	+ * Find free memory for the real mode trampoline and place it
	1088	+ * there.
	1089	+ * If there is not enough free memory under 1M, on EFI-enabled
	1090	+ * systems there will be additional attempt to reclaim the memory
	1091	+ * for the real mode trampoline at efi_free_boot_services().
	1092	+ *
	1093	+ * Unconditionally reserve the entire first 1M of RAM because
	1094	+ * BIOSes are know to corrupt low memory and several
	1095	+ * hundred kilobytes are not worth complex detection what memory gets
	1096	+ * clobbered. Moreover, on machines with SandyBridge graphics or in
	1097	+ * setups that use crashkernel the entire 1M is reserved anyway.
1090	1098	*/
1091		− trim_snb_memory();
	1099	+ reserve_real_mode();
1092	1100
1093	1101	init_mem_mapping();
1094	1102
@@ -1226,6 +1234,14 @@ void __init setup_arch(char **cmdline_p) · 131 unmodified lines
1226	1234
1227	1235	x86_init.timers.wallclock_init();
1228	1236
	1237	+ /*
	1238	+ * This needs to run before setup_local_APIC() which soft-disables the
	1239	+ * local APIC temporarily and that masks the thermal LVT interrupt,
	1240	+ * leading to softlockups on machines which have configured SMI
	1241	+ * interrupt delivery.
	1242	+ */
	1243	+ therm_lvt_init();
	1244	+
1229	1245	mcheck_init();
1230	1246
1231	1247	register_refined_jiffies(CLOCK_TICK_RATE);

arch/x86/kvm/cpuid.c+1−0

@@ -655,6 +655,7 @@ static int __do_cpuid_func_emulated(struct kvm_cpuid_array *array, u32 func) · 654 unmodified lines
655	655	if (kvm_cpu_cap_has(X86_FEATURE_RDTSCP))
656	656	entry->ecx = F(RDPID);
657	657	++array->nent;
	658	+ break;
658	659	default:
659	660	break;
660	661	}

arch/x86/kvm/emulate.c+3−2

@@ -5111,7 +5111,7 @@ done: · 5110 unmodified lines
5111	5111	return rc;
5112	5112	}
5113	5113
5114		−int x86_decode_insn(struct x86_emulate_ctxt ctxt, void insn, int insn_len)
	5114	+int x86_decode_insn(struct x86_emulate_ctxt ctxt, void insn, int insn_len, int emulation_type)
5115	5115	{
5116	5116	int rc = X86EMUL_CONTINUE;
5117	5117	int mode = ctxt->mode;
@@ -5322,7 +5322,8 @@ done_prefixes: · 204 unmodified lines
5322	5322
5323	5323	ctxt->execute = opcode.u.execute;
5324	5324
5325		− if (unlikely(ctxt->ud) && likely(!(ctxt->d & EmulateOnUD)))
	5325	+ if (unlikely(emulation_type & EMULTYPE_TRAP_UD) &&
	5326	+ likely(!(ctxt->d & EmulateOnUD)))
5326	5327	return EMULATION_FAILED;
5327	5328
5328	5329	if (unlikely(ctxt->d &

arch/x86/kvm/hyperv.c+8−0

@@ -1172,6 +1172,7 @@ void kvm_hv_invalidate_tsc_page(struct kvm *kvm) · 1171 unmodified lines
1172	1172	{
1173	1173	struct kvm_hv *hv = to_kvm_hv(kvm);
1174	1174	u64 gfn;
	1175	+ int idx;
1175	1176
1176	1177	if (hv->hv_tsc_page_status == HV_TSC_PAGE_BROKEN \|\|
1177	1178	hv->hv_tsc_page_status == HV_TSC_PAGE_UNSET \|\|
@@ -1190,9 +1191,16 @@ void kvm_hv_invalidate_tsc_page(struct kvm *kvm) · 12 unmodified lines
1190	1191	gfn = hv->hv_tsc_page >> HV_X64_MSR_TSC_REFERENCE_ADDRESS_SHIFT;
1191	1192
1192	1193	hv->tsc_ref.tsc_sequence = 0;
	1194	+
	1195	+ /*
	1196	+ * Take the srcu lock as memslots will be accessed to check the gfn
	1197	+ * cache generation against the memslots generation.
	1198	+ */
	1199	+ idx = srcu_read_lock(&kvm->srcu);
1193	1200	if (kvm_write_guest(kvm, gfn_to_gpa(gfn),
1194	1201	&hv->tsc_ref, sizeof(hv->tsc_ref.tsc_sequence)))
1195	1202	hv->hv_tsc_page_status = HV_TSC_PAGE_BROKEN;
	1203	+ srcu_read_unlock(&kvm->srcu, idx);
1196	1204
1197	1205	out_unlock:
1198	1206	mutex_unlock(&hv->hv_lock);

arch/x86/kvm/kvm_emulate.h+1−2

@@ -314,7 +314,6 @@ struct x86_emulate_ctxt { · 313 unmodified lines
314	314	int interruptibility;
315	315
316	316	bool perm_ok; /* do not check permissions if true */
317		− bool ud; /* inject an #UD if host doesn't support insn */
318	317	bool tf; /* TF value before instruction (after for syscall/sysret) */
319	318
320	319	bool have_exception;
@@ -491,7 +490,7 @@ enum x86_intercept { · 170 unmodified lines
491	490	#define X86EMUL_MODE_HOST X86EMUL_MODE_PROT64
492	491	#endif
493	492
494		−int x86_decode_insn(struct x86_emulate_ctxt ctxt, void insn, int insn_len);
	493	+int x86_decode_insn(struct x86_emulate_ctxt ctxt, void insn, int insn_len, int emulation_type);
495	494	bool x86_page_table_writing_insn(struct x86_emulate_ctxt *ctxt);
496	495	#define EMULATION_FAILED -1
497	496	#define EMULATION_OK 0

arch/x86/kvm/lapic.c+26−10

@@ -1410,6 +1410,9 @@ int kvm_lapic_reg_read(struct kvm_lapic *apic, u32 offset, int len, · 1409 unmodified lines
1410	1410	if (!apic_x2apic_mode(apic))
1411	1411	valid_reg_mask \|= APIC_REG_MASK(APIC_ARBPRI);
1412	1412
	1413	+ if (alignment + len > 4)
	1414	+ return 1;
	1415	+
1413	1416	if (offset > 0x3f0 \|\| !(valid_reg_mask & APIC_REG_MASK(offset)))
1414	1417	return 1;
1415	1418
@@ -1494,6 +1497,15 @@ static void limit_periodic_timer_frequency(struct kvm_lapic *apic) · 78 unmodified lines
1494	1497
1495	1498	static void cancel_hv_timer(struct kvm_lapic *apic);
1496	1499
	1500	+static void cancel_apic_timer(struct kvm_lapic *apic)
	1501	+{
	1502	+ hrtimer_cancel(&apic->lapic_timer.timer);
	1503	+ preempt_disable();
	1504	+ if (apic->lapic_timer.hv_timer_in_use)
	1505	+ cancel_hv_timer(apic);
	1506	+ preempt_enable();
	1507	+}
	1508	+
1497	1509	static void apic_update_lvtt(struct kvm_lapic *apic)
1498	1510	{
1499	1511	u32 timer_mode = kvm_lapic_get_reg(apic, APIC_LVTT) &
@@ -1502,11 +1514,7 @@ static void apic_update_lvtt(struct kvm_lapic *apic) · 2 unmodified lines
1502	1514	if (apic->lapic_timer.timer_mode != timer_mode) {
1503	1515	if (apic_lvtt_tscdeadline(apic) != (timer_mode ==
1504	1516	APIC_LVT_TIMER_TSCDEADLINE)) {
1505		− hrtimer_cancel(&apic->lapic_timer.timer);
1506		− preempt_disable();
1507		− if (apic->lapic_timer.hv_timer_in_use)
1508		− cancel_hv_timer(apic);
1509		− preempt_enable();
	1517	+ cancel_apic_timer(apic);
1510	1518	kvm_lapic_set_reg(apic, APIC_TMICT, 0);
1511	1519	apic->lapic_timer.period = 0;
1512	1520	apic->lapic_timer.tscdeadline = 0;
@@ -1598,11 +1606,19 @@ static void __kvm_wait_lapic_expire(struct kvm_vcpu *vcpu) · 85 unmodified lines
1598	1606	guest_tsc = kvm_read_l1_tsc(vcpu, rdtsc());
1599	1607	apic->lapic_timer.advance_expire_delta = guest_tsc - tsc_deadline;
1600	1608
	1609	+ if (lapic_timer_advance_dynamic) {
	1610	+ adjust_lapic_timer_advance(vcpu, apic->lapic_timer.advance_expire_delta);
	1611	+ /*
	1612	+ * If the timer fired early, reread the TSC to account for the
	1613	+ * overhead of the above adjustment to avoid waiting longer
	1614	+ * than is necessary.
	1615	+ */
	1616	+ if (guest_tsc < tsc_deadline)
	1617	+ guest_tsc = kvm_read_l1_tsc(vcpu, rdtsc());
	1618	+ }
	1619	+
1601	1620	if (guest_tsc < tsc_deadline)
1602	1621	__wait_lapic_expire(vcpu, tsc_deadline - guest_tsc);
1603		−
1604		− if (lapic_timer_advance_dynamic)
1605		− adjust_lapic_timer_advance(vcpu, apic->lapic_timer.advance_expire_delta);
1606	1622	}
1607	1623
1608	1624	void kvm_wait_lapic_expire(struct kvm_vcpu *vcpu)
@@ -1661,7 +1677,7 @@ static void apic_timer_expired(struct kvm_lapic *apic, bool from_timer_fn) · 52 unmodified lines
1661	1677	}
1662	1678
1663	1679	atomic_inc(&apic->lapic_timer.pending);
1664		− kvm_make_request(KVM_REQ_PENDING_TIMER, vcpu);
	1680	+ kvm_make_request(KVM_REQ_UNBLOCK, vcpu);
1665	1681	if (from_timer_fn)
1666	1682	kvm_vcpu_kick(vcpu);
1667	1683	}
@@ -2084,7 +2100,7 @@ int kvm_lapic_reg_write(struct kvm_lapic *apic, u32 reg, u32 val) · 416 unmodified lines
2084	2100	if (apic_lvtt_tscdeadline(apic))
2085	2101	break;
2086	2102
2087		− hrtimer_cancel(&apic->lapic_timer.timer);
	2103	+ cancel_apic_timer(apic);
2088	2104	kvm_lapic_set_reg(apic, APIC_TMICT, val);
2089	2105	start_apic_timer(apic);
2090	2106	break;

arch/x86/kvm/mmu/mmu.c+25−1

@@ -4739,9 +4739,33 @@ static void init_kvm_softmmu(struct kvm_vcpu *vcpu) · 4738 unmodified lines
4739	4739	context->inject_page_fault = kvm_inject_page_fault;
4740	4740	}
4741	4741
	4742	+static union kvm_mmu_role kvm_calc_nested_mmu_role(struct kvm_vcpu *vcpu)
	4743	+{
	4744	+ union kvm_mmu_role role = kvm_calc_shadow_root_page_role_common(vcpu, false);
	4745	+
	4746	+ /*
	4747	+ * Nested MMUs are used only for walking L2's gva->gpa, they never have
	4748	+ * shadow pages of their own and so "direct" has no meaning. Set it
	4749	+ * to "true" to try to detect bogus usage of the nested MMU.
	4750	+ */
	4751	+ role.base.direct = true;
	4752	+
	4753	+ if (!is_paging(vcpu))
	4754	+ role.base.level = 0;
	4755	+ else if (is_long_mode(vcpu))
	4756	+ role.base.level = is_la57_mode(vcpu) ? PT64_ROOT_5LEVEL :
	4757	+ PT64_ROOT_4LEVEL;
	4758	+ else if (is_pae(vcpu))
	4759	+ role.base.level = PT32E_ROOT_LEVEL;
	4760	+ else
	4761	+ role.base.level = PT32_ROOT_LEVEL;
	4762	+
	4763	+ return role;
	4764	+}
	4765	+
4742	4766	static void init_kvm_nested_mmu(struct kvm_vcpu *vcpu)
4743	4767	{
4744		− union kvm_mmu_role new_role = kvm_calc_mmu_role_common(vcpu, false);
	4768	+ union kvm_mmu_role new_role = kvm_calc_nested_mmu_role(vcpu);
4745	4769	struct kvm_mmu *g_context = &vcpu->arch.nested_mmu;
4746	4770
4747	4771	if (new_role.as_u64 == g_context->mmu_role.as_u64)

arch/x86/kvm/mmu/paging_tmpl.h+9−5

@@ -90,8 +90,8 @@ struct guest_walker { · 89 unmodified lines
90	90	gpa_t pte_gpa[PT_MAX_FULL_LEVELS];
91	91	pt_element_t __user *ptep_user[PT_MAX_FULL_LEVELS];
92	92	bool pte_writable[PT_MAX_FULL_LEVELS];
93		− unsigned pt_access;
94		− unsigned pte_access;
	93	+ unsigned int pt_access[PT_MAX_FULL_LEVELS];
	94	+ unsigned int pte_access;
95	95	gfn_t gfn;
96	96	struct x86_exception fault;
97	97	};
@@ -418,13 +418,15 @@ retry_walk: · 320 unmodified lines
418	418	}
419	419
420	420	walker->ptes[walker->level - 1] = pte;
	421	+
	422	+ /* Convert to ACC__MASK flags for struct guest_walker. /
	423	+ walker->pt_access[walker->level - 1] = FNAME(gpte_access)(pt_access ^ walk_nx_mask);
421	424	} while (!is_last_gpte(mmu, walker->level, pte));
422	425
423	426	pte_pkey = FNAME(gpte_pkeys)(vcpu, pte);
424	427	accessed_dirty = have_ad ? pte_access & PT_GUEST_ACCESSED_MASK : 0;
425	428
426	429	/* Convert to ACC__MASK flags for struct guest_walker. /
427		− walker->pt_access = FNAME(gpte_access)(pt_access ^ walk_nx_mask);
428	430	walker->pte_access = FNAME(gpte_access)(pte_access ^ walk_nx_mask);
429	431	errcode = permission_fault(vcpu, mmu, walker->pte_access, pte_pkey, access);
430	432	if (unlikely(errcode))
@@ -463,7 +465,8 @@ retry_walk: · 32 unmodified lines
463	465	}
464	466
465	467	pgprintk("%s: pte %llx pte_access %x pt_access %x\n",
466		− __func__, (u64)pte, walker->pte_access, walker->pt_access);
	468	+ __func__, (u64)pte, walker->pte_access,
	469	+ walker->pt_access[walker->level - 1]);
467	470	return 1;
468	471
469	472	error:
@@ -643,7 +646,7 @@ static int FNAME(fetch)(struct kvm_vcpu *vcpu, gpa_t addr, · 173 unmodified lines
643	646	bool huge_page_disallowed = exec && nx_huge_page_workaround_enabled;
644	647	struct kvm_mmu_page *sp = NULL;
645	648	struct kvm_shadow_walk_iterator it;
646		− unsigned direct_access, access = gw->pt_access;
	649	+ unsigned int direct_access, access;
647	650	int top_level, level, req_level, ret;
648	651	gfn_t base_gfn = gw->gfn;
649	652
@@ -675,6 +678,7 @@ static int FNAME(fetch)(struct kvm_vcpu *vcpu, gpa_t addr, · 25 unmodified lines
675	678	sp = NULL;
676	679	if (!is_shadow_present_pte(*it.sptep)) {
677	680	table_gfn = gw->table_gfn[it.level - 2];
	681	+ access = gw->pt_access[it.level - 2];
678	682	sp = kvm_mmu_get_page(vcpu, table_gfn, addr, it.level-1,
679	683	false, access);
680	684	}

arch/x86/kvm/mmu/tdp_mmu.c+3−3

@@ -1192,9 +1192,9 @@ bool kvm_tdp_mmu_set_spte_gfn(struct kvm kvm, struct kvm_gfn_range range) · 1191 unmodified lines
1192	1192	}
1193	1193
1194	1194	/*
1195		− * Remove write access from all the SPTEs mapping GFNs [start, end). If
1196		− * skip_4k is set, SPTEs that map 4k pages, will not be write-protected.
1197		− * Returns true if an SPTE has been changed and the TLBs need to be flushed.
	1195	+ * Remove write access from all SPTEs at or above min_level that map GFNs
	1196	+ * [start, end). Returns true if an SPTE has been changed and the TLBs need to
	1197	+ * be flushed.
1198	1198	*/
1199	1199	static bool wrprot_gfn_range(struct kvm kvm, struct kvm_mmu_page root,
1200	1200	gfn_t start, gfn_t end, int min_level)

arch/x86/kvm/svm/avic.c+5−7

@@ -28,10 +28,8 @@ · 27 unmodified lines
28	28	#include "svm.h"
29	29
30	30	/* enable / disable AVIC */
31		−int avic;
32		−#ifdef CONFIG_X86_LOCAL_APIC
33		−module_param(avic, int, S_IRUGO);
34		−#endif
	31	+bool avic;
	32	+module_param(avic, bool, S_IRUGO);
35	33
36	34	#define SVM_AVIC_DOORBELL 0xc001011b
37	35
@@ -223,7 +221,7 @@ static u64 avic_get_physical_id_entry(struct kvm_vcpu vcpu, · 185 unmodified lines
223	221	return &avic_physical_id_table[index];
224	222	}
225	223
226		−/**
	224	+/*
227	225	* Note:
228	226	* AVIC hardware walks the nested page table to check permissions,
229	227	* but does not use the SPA address specified in the leaf page
@@ -766,7 +764,7 @@ out: · 536 unmodified lines
766	764	return ret;
767	765	}
768	766
769		−/**
	767	+/*
770	768	* Note:
771	769	* The HW cannot support posting multicast/broadcast
772	770	* interrupts to a vCPU. So, we still use legacy interrupt
@@ -1007,7 +1005,7 @@ void avic_vcpu_put(struct kvm_vcpu *vcpu) · 234 unmodified lines
1007	1005	WRITE_ONCE(*(svm->avic_physical_id_cache), entry);
1008	1006	}
1009	1007
1010		−/**
	1008	+/*
1011	1009	* This function is called during VCPU halt/unhalt.
1012	1010	*/
1013	1011	static void avic_set_running(struct kvm_vcpu *vcpu, bool is_run)

arch/x86/kvm/svm/sev.c+17−9

@@ -199,9 +199,19 @@ static void sev_asid_free(struct kvm_sev_info *sev) · 198 unmodified lines
199	199	sev->misc_cg = NULL;
200	200	}
201	201
202		−static void sev_unbind_asid(struct kvm *kvm, unsigned int handle)
	202	+static void sev_decommission(unsigned int handle)
203	203	{
204	204	struct sev_data_decommission decommission;
	205	+
	206	+ if (!handle)
	207	+ return;
	208	+
	209	+ decommission.handle = handle;
	210	+ sev_guest_decommission(&decommission, NULL);
	211	+}
	212	+
	213	+static void sev_unbind_asid(struct kvm *kvm, unsigned int handle)
	214	+{
205	215	struct sev_data_deactivate deactivate;
206	216
207	217	if (!handle)
@@ -214,9 +224,7 @@ static void sev_unbind_asid(struct kvm *kvm, unsigned int handle) · 6 unmodified lines
214	224	sev_guest_deactivate(&deactivate, NULL);
215	225	up_read(&sev_deactivate_lock);
216	226
217		− /* decommission handle */
218		− decommission.handle = handle;
219		− sev_guest_decommission(&decommission, NULL);
	227	+ sev_decommission(handle);
220	228	}
221	229
222	230	static int sev_guest_init(struct kvm kvm, struct kvm_sev_cmd argp)
@@ -341,8 +349,10 @@ static int sev_launch_start(struct kvm kvm, struct kvm_sev_cmd argp) · 118 unmodified lines
341	349
342	350	/* Bind ASID to this guest */
343	351	ret = sev_bind_asid(kvm, start.handle, error);
344		− if (ret)
	352	+ if (ret) {
	353	+ sev_decommission(start.handle);
345	354	goto e_free_session;
	355	+ }
346	356
347	357	/* return handle to userspace */
348	358	params.handle = start.handle;
@@ -1103,10 +1113,9 @@ __sev_send_start_query_session_length(struct kvm kvm, struct kvm_sev_cmd argp, · 754 unmodified lines
1103	1113	struct sev_data_send_start data;
1104	1114	int ret;
1105	1115
	1116	+ memset(&data, 0, sizeof(data));
1106	1117	data.handle = sev->handle;
1107	1118	ret = sev_issue_cmd(kvm, SEV_CMD_SEND_START, &data, &argp->error);
1108		− if (ret < 0)
1109		− return ret;
1110	1119
1111	1120	params->session_len = data.session_len;
1112	1121	if (copy_to_user((void __user *)(uintptr_t)argp->data, params,
@@ -1215,10 +1224,9 @@ __sev_send_update_data_query_lengths(struct kvm kvm, struct kvm_sev_cmd argp, · 102 unmodified lines
1215	1224	struct sev_data_send_update_data data;
1216	1225	int ret;
1217	1226
	1227	+ memset(&data, 0, sizeof(data));
1218	1228	data.handle = sev->handle;
1219	1229	ret = sev_issue_cmd(kvm, SEV_CMD_SEND_UPDATE_DATA, &data, &argp->error);
1220		− if (ret < 0)
1221		− return ret;
1222	1230
1223	1231	params->hdr_len = data.hdr_len;
1224	1232	params->trans_len = data.trans_len;

arch/x86/kvm/svm/svm.c+1−3

@@ -1010,9 +1010,7 @@ static __init int svm_hardware_setup(void) · 1009 unmodified lines
1010	1010	}
1011	1011
1012	1012	if (avic) {
1013		− if (!npt_enabled \|\|
1014		− !boot_cpu_has(X86_FEATURE_AVIC) \|\|
1015		− !IS_ENABLED(CONFIG_X86_LOCAL_APIC)) {
	1013	+ if (!npt_enabled \|\| !boot_cpu_has(X86_FEATURE_AVIC)) {
1016	1014	avic = false;
1017	1015	} else {
1018	1016	pr_info("AVIC enabled\n");

arch/x86/kvm/svm/svm.h+1−1

@@ -480,7 +480,7 @@ extern struct kvm_x86_nested_ops svm_nested_ops; · 479 unmodified lines
480	480
481	481	#define VMCB_AVIC_APIC_BAR_MASK 0xFFFFFFFFFF000ULL
482	482
483		−extern int avic;
	483	+extern bool avic;
484	484
485	485	static inline void avic_update_vapic_bar(struct vcpu_svm *svm, u64 data)
486	486	{

arch/x86/kvm/trace.h+3−3

@@ -1550,16 +1550,16 @@ TRACE_EVENT(kvm_nested_vmenter_failed, · 1549 unmodified lines
1550	1550	TP_ARGS(msg, err),
1551	1551
1552	1552	TP_STRUCT__entry(
1553		− __field(const char *, msg)
	1553	+ __string(msg, msg)
1554	1554	__field(u32, err)
1555	1555	),
1556	1556
1557	1557	TP_fast_assign(
1558		− __entry->msg = msg;
	1558	+ __assign_str(msg, msg);
1559	1559	__entry->err = err;
1560	1560	),
1561	1561
1562		− TP_printk("%s%s", __entry->msg, !__entry->err ? "" :
	1562	+ TP_printk("%s%s", __get_str(msg), !__entry->err ? "" :
1563	1563	__print_symbolic(__entry->err, VMX_VMENTER_INSTRUCTION_ERRORS))
1564	1564	);
1565	1565

arch/x86/kvm/vmx/capabilities.h+1−2

@@ -90,8 +90,7 @@ static inline bool cpu_has_vmx_preemption_timer(void) · 89 unmodified lines
90	90
91	91	static inline bool cpu_has_vmx_posted_intr(void)
92	92	{
93		− return IS_ENABLED(CONFIG_X86_LOCAL_APIC) &&
94		− vmcs_config.pin_based_exec_ctrl & PIN_BASED_POSTED_INTR;
	93	+ return vmcs_config.pin_based_exec_ctrl & PIN_BASED_POSTED_INTR;
95	94	}
96	95
97	96	static inline bool cpu_has_load_ia32_efer(void)

arch/x86/kvm/vmx/posted_intr.c+14−0

@@ -237,6 +237,20 @@ bool pi_has_pending_interrupt(struct kvm_vcpu *vcpu) · 236 unmodified lines
237	237	}
238	238
239	239
	240	+/*
	241	+ * Bail out of the block loop if the VM has an assigned
	242	+ * device, but the blocking vCPU didn't reconfigure the
	243	+ * PI.NV to the wakeup vector, i.e. the assigned device
	244	+ * came along after the initial check in pi_pre_block().
	245	+ */
	246	+void vmx_pi_start_assignment(struct kvm *kvm)
	247	+{
	248	+ if (!irq_remapping_cap(IRQ_POSTING_CAP))
	249	+ return;
	250	+
	251	+ kvm_make_all_cpus_request(kvm, KVM_REQ_UNBLOCK);
	252	+}
	253	+
240	254	/*
241	255	* pi_update_irte - set IRTE for Posted-Interrupts
242	256	*

arch/x86/kvm/vmx/posted_intr.h+1−0

@@ -95,5 +95,6 @@ void __init pi_init_cpu(int cpu); · 94 unmodified lines
95	95	bool pi_has_pending_interrupt(struct kvm_vcpu *vcpu);
96	96	int pi_update_irte(struct kvm *kvm, unsigned int host_irq, uint32_t guest_irq,
97	97	bool set);
	98	+void vmx_pi_start_assignment(struct kvm *kvm);
98	99
99	100	#endif /* __KVM_X86_VMX_POSTED_INTR_H */

arch/x86/kvm/vmx/vmx.c+4−3

@@ -4843,7 +4843,7 @@ static int handle_exception_nmi(struct kvm_vcpu *vcpu) · 4842 unmodified lines
4843	4843	struct vcpu_vmx *vmx = to_vmx(vcpu);
4844	4844	struct kvm_run *kvm_run = vcpu->run;
4845	4845	u32 intr_info, ex_no, error_code;
4846		− unsigned long cr2, rip, dr6;
	4846	+ unsigned long cr2, dr6;
4847	4847	u32 vect_info;
4848	4848
4849	4849	vect_info = vmx->idt_vectoring_info;
@@ -4933,8 +4933,7 @@ static int handle_exception_nmi(struct kvm_vcpu *vcpu) · 83 unmodified lines
4933	4933	vmx->vcpu.arch.event_exit_inst_len =
4934	4934	vmcs_read32(VM_EXIT_INSTRUCTION_LEN);
4935	4935	kvm_run->exit_reason = KVM_EXIT_DEBUG;
4936		− rip = kvm_rip_read(vcpu);
4937		− kvm_run->debug.arch.pc = vmcs_readl(GUEST_CS_BASE) + rip;
	4936	+ kvm_run->debug.arch.pc = kvm_get_linear_rip(vcpu);
4938	4937	kvm_run->debug.arch.exception = ex_no;
4939	4938	break;
4940	4939	case AC_VECTOR:
@@ -6248,6 +6247,7 @@ void vmx_set_virtual_apic_mode(struct kvm_vcpu *vcpu) · 1307 unmodified lines
6248	6247	switch (kvm_get_apic_mode(vcpu)) {
6249	6248	case LAPIC_MODE_INVALID:
6250	6249	WARN_ONCE(true, "Invalid local APIC state");
	6250	+ break;
6251	6251	case LAPIC_MODE_DISABLED:
6252	6252	break;
6253	6253	case LAPIC_MODE_XAPIC:
@@ -7721,6 +7721,7 @@ static struct kvm_x86_ops vmx_x86_ops __initdata = { · 1467 unmodified lines
7721	7721	.nested_ops = &vmx_nested_ops,
7722	7722
7723	7723	.update_pi_irte = pi_update_irte,
	7724	+ .start_assignment = vmx_pi_start_assignment,
7724	7725
7725	7726	#ifdef CONFIG_X86_64
7726	7727	.set_hv_timer = vmx_set_hv_timer,

arch/x86/kvm/x86.c+37−15

@@ -3072,6 +3072,19 @@ static void kvm_vcpu_flush_tlb_all(struct kvm_vcpu *vcpu) · 3071 unmodified lines
3072	3072	static void kvm_vcpu_flush_tlb_guest(struct kvm_vcpu *vcpu)
3073	3073	{
3074	3074	++vcpu->stat.tlb_flush;
	3075	+
	3076	+ if (!tdp_enabled) {
	3077	+ /*
	3078	+ * A TLB flush on behalf of the guest is equivalent to
	3079	+ * INVPCID(all), toggling CR4.PGE, etc., which requires
	3080	+ * a forced sync of the shadow page tables. Unload the
	3081	+ * entire MMU here and the subsequent load will sync the
	3082	+ * shadow page tables, and also flush the TLB.
	3083	+ */
	3084	+ kvm_mmu_unload(vcpu);
	3085	+ return;
	3086	+ }
	3087	+
3075	3088	static_call(kvm_x86_tlb_flush_guest)(vcpu);
3076	3089	}
3077	3090
@@ -3101,10 +3114,14 @@ static void record_steal_time(struct kvm_vcpu *vcpu) · 23 unmodified lines
3101	3114	* expensive IPIs.
3102	3115	*/
3103	3116	if (guest_pv_has(vcpu, KVM_FEATURE_PV_TLB_FLUSH)) {
	3117	+ u8 st_preempted = xchg(&st->preempted, 0);
	3118	+
3104	3119	trace_kvm_pv_tlb_flush(vcpu->vcpu_id,
3105		− st->preempted & KVM_VCPU_FLUSH_TLB);
3106		− if (xchg(&st->preempted, 0) & KVM_VCPU_FLUSH_TLB)
	3120	+ st_preempted & KVM_VCPU_FLUSH_TLB);
	3121	+ if (st_preempted & KVM_VCPU_FLUSH_TLB)
3107	3122	kvm_vcpu_flush_tlb_guest(vcpu);
	3123	+ } else {
	3124	+ st->preempted = 0;
3108	3125	}
3109	3126
3110	3127	vcpu->arch.st.preempted = 0;
@@ -7089,7 +7106,10 @@ static unsigned emulator_get_hflags(struct x86_emulate_ctxt *ctxt) · 3978 unmodified lines
7089	7106
7090	7107	static void emulator_set_hflags(struct x86_emulate_ctxt *ctxt, unsigned emul_flags)
7091	7108	{
7092		− emul_to_vcpu(ctxt)->arch.hflags = emul_flags;
	7109	+ struct kvm_vcpu *vcpu = emul_to_vcpu(ctxt);
	7110	+
	7111	+ vcpu->arch.hflags = emul_flags;
	7112	+ kvm_mmu_reset_context(vcpu);
7093	7113	}
7094	7114
7095	7115	static int emulator_pre_leave_smm(struct x86_emulate_ctxt *ctxt,
@@ -7226,6 +7246,11 @@ static void init_emulate_ctxt(struct kvm_vcpu *vcpu) · 130 unmodified lines
7226	7246	BUILD_BUG_ON(HF_SMM_MASK != X86EMUL_SMM_MASK);
7227	7247	BUILD_BUG_ON(HF_SMM_INSIDE_NMI_MASK != X86EMUL_SMM_INSIDE_NMI_MASK);
7228	7248
	7249	+ ctxt->interruptibility = 0;
	7250	+ ctxt->have_exception = false;
	7251	+ ctxt->exception.vector = -1;
	7252	+ ctxt->perm_ok = false;
	7253	+
7229	7254	init_decode_cache(ctxt);
7230	7255	vcpu->arch.emulate_regs_need_sync_from_vcpu = false;
7231	7256	}
@@ -7561,14 +7586,7 @@ int x86_decode_emulated_instruction(struct kvm_vcpu *vcpu, int emulation_type, · 329 unmodified lines
7561	7586	kvm_vcpu_check_breakpoint(vcpu, &r))
7562	7587	return r;
7563	7588
7564		− ctxt->interruptibility = 0;
7565		− ctxt->have_exception = false;
7566		− ctxt->exception.vector = -1;
7567		− ctxt->perm_ok = false;
7568		−
7569		− ctxt->ud = emulation_type & EMULTYPE_TRAP_UD;
7570		−
7571		− r = x86_decode_insn(ctxt, insn, insn_len);
	7589	+ r = x86_decode_insn(ctxt, insn, insn_len, emulation_type);
7572	7590
7573	7591	trace_kvm_emulate_insn_start(vcpu);
7574	7592	++vcpu->stat.insn_emulation;
@@ -8243,6 +8261,7 @@ void kvm_arch_exit(void) · 668 unmodified lines
8243	8261	kvm_x86_ops.hardware_enable = NULL;
8244	8262	kvm_mmu_module_exit();
8245	8263	free_percpu(user_return_msrs);
	8264	+ kmem_cache_destroy(x86_emulator_cache);
8246	8265	kmem_cache_destroy(x86_fpu_cache);
8247	8266	#ifdef CONFIG_KVM_XEN
8248	8267	static_key_deferred_flush(&kvm_xen_enabled);
@@ -8360,6 +8379,9 @@ static void kvm_sched_yield(struct kvm_vcpu *vcpu, unsigned long dest_id) · 111 unmodified lines
8360	8379
8361	8380	vcpu->stat.directed_yield_attempted++;
8362	8381
	8382	+ if (single_task_running())
	8383	+ goto no_yield;
	8384	+
8363	8385	rcu_read_lock();
8364	8386	map = rcu_dereference(vcpu->kvm->arch.apic_map);
8365	8387
@@ -9496,7 +9518,7 @@ static int vcpu_run(struct kvm_vcpu *vcpu) · 1130 unmodified lines
9496	9518	if (r <= 0)
9497	9519	break;
9498	9520
9499		− kvm_clear_request(KVM_REQ_PENDING_TIMER, vcpu);
	9521	+ kvm_clear_request(KVM_REQ_UNBLOCK, vcpu);
9500	9522	if (kvm_cpu_has_pending_timer(vcpu))
9501	9523	kvm_inject_pending_timer_irqs(vcpu);
9502	9524
@@ -10115,8 +10137,7 @@ int kvm_arch_vcpu_ioctl_set_guest_debug(struct kvm_vcpu *vcpu, · 612 unmodified lines
10115	10137	kvm_update_dr7(vcpu);
10116	10138
10117	10139	if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP)
10118		− vcpu->arch.singlestep_rip = kvm_rip_read(vcpu) +
10119		− get_segment_base(vcpu, VCPU_SREG_CS);
	10140	+ vcpu->arch.singlestep_rip = kvm_get_linear_rip(vcpu);
10120	10141
10121	10142	/*
10122	10143	* Trigger an rflags update that will inject or remove the trace
@@ -11499,7 +11520,8 @@ bool kvm_arch_can_dequeue_async_page_present(struct kvm_vcpu *vcpu) · 1376 unmodified lines
11499	11520
11500	11521	void kvm_arch_start_assignment(struct kvm *kvm)
11501	11522	{
11502		− atomic_inc(&kvm->arch.assigned_device_count);
	11523	+ if (atomic_inc_return(&kvm->arch.assigned_device_count) == 1)
	11524	+ static_call_cond(kvm_x86_start_assignment)(kvm);
11503	11525	}
11504	11526	EXPORT_SYMBOL_GPL(kvm_arch_start_assignment);
11505	11527

arch/x86/mm/fault.c+2−2

@@ -836,8 +836,8 @@ __bad_area_nosemaphore(struct pt_regs *regs, unsigned long error_code, · 835 unmodified lines
836	836
837	837	if (si_code == SEGV_PKUERR)
838	838	force_sig_pkuerr((void __user *)address, pkey);
839		−
840		− force_sig_fault(SIGSEGV, si_code, (void __user *)address);
	839	+ else
	840	+ force_sig_fault(SIGSEGV, si_code, (void __user *)address);
841	841
842	842	local_irq_disable();
843	843	}

arch/x86/mm/ioremap.c+3−1

@@ -118,7 +118,9 @@ static void __ioremap_check_other(resource_size_t addr, struct ioremap_desc *des · 117 unmodified lines
118	118	if (!IS_ENABLED(CONFIG_EFI))
119	119	return;
120	120
121		− if (efi_mem_type(addr) == EFI_RUNTIME_SERVICES_DATA)
	121	+ if (efi_mem_type(addr) == EFI_RUNTIME_SERVICES_DATA \|\|
	122	+ (efi_mem_type(addr) == EFI_BOOT_SERVICES_DATA &&
	123	+ efi_mem_attributes(addr) & EFI_MEMORY_RUNTIME))
122	124	desc->flags \|= IORES_MAP_ENCRYPTED;
123	125	}
124	126

arch/x86/mm/mem_encrypt_identity.c+6−5

@@ -504,10 +504,6 @@ void __init sme_enable(struct boot_params *bp) · 503 unmodified lines
504	504	#define AMD_SME_BIT BIT(0)
505	505	#define AMD_SEV_BIT BIT(1)
506	506
507		− /* Check the SEV MSR whether SEV or SME is enabled */
508		− sev_status = __rdmsr(MSR_AMD64_SEV);
509		− feature_mask = (sev_status & MSR_AMD64_SEV_ENABLED) ? AMD_SEV_BIT : AMD_SME_BIT;
510		−
511	507	/*
512	508	* Check for the SME/SEV feature:
513	509	* CPUID Fn8000_001F[EAX]
@@ -519,11 +515,16 @@ void __init sme_enable(struct boot_params *bp) · 5 unmodified lines
519	515	eax = 0x8000001f;
520	516	ecx = 0;
521	517	native_cpuid(&eax, &ebx, &ecx, &edx);
522		− if (!(eax & feature_mask))
	518	+ /* Check whether SEV or SME is supported */
	519	+ if (!(eax & (AMD_SEV_BIT \| AMD_SME_BIT)))
523	520	return;
524	521
525	522	me_mask = 1UL << (ebx & 0x3f);
526	523
	524	+ /* Check the SEV MSR whether SEV or SME is enabled */
	525	+ sev_status = __rdmsr(MSR_AMD64_SEV);
	526	+ feature_mask = (sev_status & MSR_AMD64_SEV_ENABLED) ? AMD_SEV_BIT : AMD_SME_BIT;
	527	+
527	528	/* Check if memory encryption is enabled */
528	529	if (feature_mask == AMD_SME_BIT) {
529	530	/*

arch/x86/mm/numa.c+7−1

@@ -254,7 +254,13 @@ int __init numa_cleanup_meminfo(struct numa_meminfo *mi) · 253 unmodified lines
254	254
255	255	/* make sure all non-reserved blocks are inside the limits */
256	256	bi->start = max(bi->start, low);
257		− bi->end = min(bi->end, high);
	257	+
	258	+ /* preserve info for non-RAM areas above 'max_pfn': */
	259	+ if (bi->end > high) {
	260	+ numa_add_memblk_to(bi->nid, high, bi->end,
	261	+ &numa_reserved_meminfo);
	262	+ bi->end = high;
	263	+ }
258	264
259	265	/* and there's no empty block */
260	266	if (bi->start >= bi->end)

arch/x86/pci/fixup.c+44−0

@@ -779,4 +779,48 @@ DECLARE_PCI_FIXUP_RESUME(PCI_VENDOR_ID_AMD, 0x1571, pci_amd_enable_64bit_bar); · 778 unmodified lines
779	779	DECLARE_PCI_FIXUP_RESUME(PCI_VENDOR_ID_AMD, 0x15b1, pci_amd_enable_64bit_bar);
780	780	DECLARE_PCI_FIXUP_RESUME(PCI_VENDOR_ID_AMD, 0x1601, pci_amd_enable_64bit_bar);
781	781
	782	+#define RS690_LOWER_TOP_OF_DRAM2 0x30
	783	+#define RS690_LOWER_TOP_OF_DRAM2_VALID 0x1
	784	+#define RS690_UPPER_TOP_OF_DRAM2 0x31
	785	+#define RS690_HTIU_NB_INDEX 0xA8
	786	+#define RS690_HTIU_NB_INDEX_WR_ENABLE 0x100
	787	+#define RS690_HTIU_NB_DATA 0xAC
	788	+
	789	+/*
	790	+ * Some BIOS implementations support RAM above 4GB, but do not configure the
	791	+ * PCI host to respond to bus master accesses for these addresses. These
	792	+ * implementations set the TOP_OF_DRAM_SLOT1 register correctly, so PCI DMA
	793	+ * works as expected for addresses below 4GB.
	794	+ *
	795	+ * Reference: "AMD RS690 ASIC Family Register Reference Guide" (pg. 2-57)
	796	+ * https://www.amd.com/system/files/TechDocs/43372_rs690_rrg_3.00o.pdf
	797	+ */
	798	+static void rs690_fix_64bit_dma(struct pci_dev *pdev)
	799	+{
	800	+ u32 val = 0;
	801	+ phys_addr_t top_of_dram = __pa(high_memory - 1) + 1;
	802	+
	803	+ if (top_of_dram <= (1ULL << 32))
	804	+ return;
	805	+
	806	+ pci_write_config_dword(pdev, RS690_HTIU_NB_INDEX,
	807	+ RS690_LOWER_TOP_OF_DRAM2);
	808	+ pci_read_config_dword(pdev, RS690_HTIU_NB_DATA, &val);
	809	+
	810	+ if (val)
	811	+ return;
	812	+
	813	+ pci_info(pdev, "Adjusting top of DRAM to %pa for 64-bit DMA support\n", &top_of_dram);
	814	+
	815	+ pci_write_config_dword(pdev, RS690_HTIU_NB_INDEX,
	816	+ RS690_UPPER_TOP_OF_DRAM2 \| RS690_HTIU_NB_INDEX_WR_ENABLE);
	817	+ pci_write_config_dword(pdev, RS690_HTIU_NB_DATA, top_of_dram >> 32);
	818	+
	819	+ pci_write_config_dword(pdev, RS690_HTIU_NB_INDEX,
	820	+ RS690_LOWER_TOP_OF_DRAM2 \| RS690_HTIU_NB_INDEX_WR_ENABLE);
	821	+ pci_write_config_dword(pdev, RS690_HTIU_NB_DATA,
	822	+ top_of_dram \| RS690_LOWER_TOP_OF_DRAM2_VALID);
	823	+}
	824	+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_ATI, 0x7910, rs690_fix_64bit_dma);
	825	+
782	826	#endif

arch/x86/platform/efi/quirks.c+12−0

@@ -450,6 +450,18 @@ void __init efi_free_boot_services(void) · 449 unmodified lines
450	450	size -= rm_size;
451	451	}
452	452
	453	+ /*
	454	+ * Don't free memory under 1M for two reasons:
	455	+ * - BIOS might clobber it
	456	+ * - Crash kernel needs it to be reserved
	457	+ */
	458	+ if (start + size < SZ_1M)
	459	+ continue;
	460	+ if (start < SZ_1M) {
	461	+ size -= (SZ_1M - start);
	462	+ start = SZ_1M;
	463	+ }
	464	+
453	465	memblock_free_late(start, size);
454	466	}
455	467

arch/x86/realmode/init.c+8−6

@@ -29,14 +29,16 @@ void __init reserve_real_mode(void) · 28 unmodified lines
29	29
30	30	/* Has to be under 1M so we can execute real-mode AP code. */
31	31	mem = memblock_find_in_range(0, 1<<20, size, PAGE_SIZE);
32		− if (!mem) {
	32	+ if (!mem)
33	33	pr_info("No sub-1M memory is available for the trampoline\n");
34		− return;
35		− }
	34	+ else
	35	+ set_real_mode_mem(mem);
36	36
37		− memblock_reserve(mem, size);
38		− set_real_mode_mem(mem);
39		− crash_reserve_low_1M();
	37	+ /*
	38	+ * Unconditionally reserve the entire fisrt 1M, see comment in
	39	+ * setup_arch().
	40	+ */
	41	+ memblock_reserve(0, SZ_1M);
40	42	}
41	43
42	44	static void sme_sev_setup_real_mode(struct trampoline_header *th)

crypto/async_tx/async_xor.c+2−1

@@ -233,7 +233,8 @@ async_xor_offs(struct page *dest, unsigned int offset, · 232 unmodified lines
233	233	if (submit->flags & ASYNC_TX_XOR_DROP_DST) {
234	234	src_cnt--;
235	235	src_list++;
236		− src_offs++;
	236	+ if (src_offs)
	237	+ src_offs++;
237	238	}
238	239
239	240	/* wait for any prerequisite operations */

drivers/acpi/acpi_apd.c+1−0

@@ -226,6 +226,7 @@ static const struct acpi_device_id acpi_apd_device_ids[] = { · 225 unmodified lines
226	226	{ "AMDI0010", APD_ADDR(wt_i2c_desc) },
227	227	{ "AMD0020", APD_ADDR(cz_uart_desc) },
228	228	{ "AMDI0020", APD_ADDR(cz_uart_desc) },
	229	+ { "AMDI0022", APD_ADDR(cz_uart_desc) },
229	230	{ "AMD0030", },
230	231	{ "AMD0040", APD_ADDR(fch_misc_desc)},
231	232	{ "HYGO0010", APD_ADDR(wt_i2c_desc) },

drivers/acpi/acpica/utdelete.c+8−0

@@ -285,6 +285,14 @@ static void acpi_ut_delete_internal_obj(union acpi_operand_object *object) · 284 unmodified lines
285	285	}
286	286	break;
287	287
	288	+ case ACPI_TYPE_LOCAL_ADDRESS_HANDLER:
	289	+
	290	+ ACPI_DEBUG_PRINT((ACPI_DB_ALLOCATIONS,
	291	+ "***** Address handler %p\n", object));
	292	+
	293	+ acpi_os_delete_mutex(object->address_space.context_mutex);
	294	+ break;
	295	+
288	296	default:
289	297
290	298	break;

drivers/acpi/bus.c+8−19

@@ -330,32 +330,21 @@ static void acpi_bus_osc_negotiate_platform_control(void) · 329 unmodified lines
330	330	if (ACPI_FAILURE(acpi_run_osc(handle, &context)))
331	331	return;
332	332
333		− capbuf_ret = context.ret.pointer;
334		− if (context.ret.length <= OSC_SUPPORT_DWORD) {
335		− kfree(context.ret.pointer);
336		− return;
337		− }
	333	+ kfree(context.ret.pointer);
338	334
339		− /*
340		− * Now run _OSC again with query flag clear and with the caps
341		− * supported by both the OS and the platform.
342		− */
	335	+ /* Now run _OSC again with query flag clear */
343	336	capbuf[OSC_QUERY_DWORD] = 0;
344		− capbuf[OSC_SUPPORT_DWORD] = capbuf_ret[OSC_SUPPORT_DWORD];
345		− kfree(context.ret.pointer);
346	337
347	338	if (ACPI_FAILURE(acpi_run_osc(handle, &context)))
348	339	return;
349	340
350	341	capbuf_ret = context.ret.pointer;
351		− if (context.ret.length > OSC_SUPPORT_DWORD) {
352		− osc_sb_apei_support_acked =
353		− capbuf_ret[OSC_SUPPORT_DWORD] & OSC_SB_APEI_SUPPORT;
354		− osc_pc_lpi_support_confirmed =
355		− capbuf_ret[OSC_SUPPORT_DWORD] & OSC_SB_PCLPI_SUPPORT;
356		− osc_sb_native_usb4_support_confirmed =
357		− capbuf_ret[OSC_SUPPORT_DWORD] & OSC_SB_NATIVE_USB4_SUPPORT;
358		− }
	342	+ osc_sb_apei_support_acked =
	343	+ capbuf_ret[OSC_SUPPORT_DWORD] & OSC_SB_APEI_SUPPORT;
	344	+ osc_pc_lpi_support_confirmed =
	345	+ capbuf_ret[OSC_SUPPORT_DWORD] & OSC_SB_PCLPI_SUPPORT;
	346	+ osc_sb_native_usb4_support_confirmed =
	347	+ capbuf_ret[OSC_SUPPORT_DWORD] & OSC_SB_NATIVE_USB4_SUPPORT;
359	348
360	349	kfree(context.ret.pointer);
361	350	}

drivers/acpi/internal.h+2−2

@@ -134,7 +134,7 @@ int acpi_power_init(void); · 133 unmodified lines
134	134	void acpi_power_resources_list_free(struct list_head *list);
135	135	int acpi_extract_power_resources(union acpi_object *package, unsigned int start,
136	136	struct list_head *list);
137		−int acpi_add_power_resource(acpi_handle handle);
	137	+struct acpi_device *acpi_add_power_resource(acpi_handle handle);
138	138	void acpi_power_add_remove_device(struct acpi_device *adev, bool add);
139	139	int acpi_power_wakeup_list_init(struct list_head list, int system_level);
140	140	int acpi_device_sleep_wake(struct acpi_device *dev,
@@ -142,7 +142,7 @@ int acpi_device_sleep_wake(struct acpi_device *dev, · 1 unmodified line
142	142	int acpi_power_get_inferred_state(struct acpi_device device, int state);
143	143	int acpi_power_on_resources(struct acpi_device *device, int state);
144	144	int acpi_power_transition(struct acpi_device *device, int state);
145		−void acpi_turn_off_unused_power_resources(void);
	145	+void acpi_turn_off_unused_power_resources(bool init);
146	146
147	147	/* --------------------------------------------------------------------------
148	148	Device Power Management

drivers/acpi/power.c+46−13

@@ -52,6 +52,7 @@ struct acpi_power_resource { · 51 unmodified lines
52	52	u32 system_level;
53	53	u32 order;
54	54	unsigned int ref_count;
	55	+ unsigned int users;
55	56	bool wakeup_enabled;
56	57	struct mutex resource_lock;
57	58	struct list_head dependents;
@@ -147,6 +148,7 @@ int acpi_extract_power_resources(union acpi_object *package, unsigned int start, · 89 unmodified lines
147	148
148	149	for (i = start; i < package->package.count; i++) {
149	150	union acpi_object *element = &package->package.elements[i];
	151	+ struct acpi_device *rdev;
150	152	acpi_handle rhandle;
151	153
152	154	if (element->type != ACPI_TYPE_LOCAL_REFERENCE) {
@@ -163,13 +165,16 @@ int acpi_extract_power_resources(union acpi_object *package, unsigned int start, · 10 unmodified lines
163	165	if (acpi_power_resource_is_dup(package, start, i))
164	166	continue;
165	167
166		− err = acpi_add_power_resource(rhandle);
167		− if (err)
	168	+ rdev = acpi_add_power_resource(rhandle);
	169	+ if (!rdev) {
	170	+ err = -ENODEV;
168	171	break;
169		−
	172	+ }
170	173	err = acpi_power_resources_list_add(rhandle, list);
171	174	if (err)
172	175	break;
	176	+
	177	+ to_power_resource(rdev)->users++;
173	178	}
174	179	if (err)
175	180	acpi_power_resources_list_free(list);
@@ -907,7 +912,7 @@ static void acpi_power_add_resource_to_list(struct acpi_power_resource *resource · 731 unmodified lines
907	912	mutex_unlock(&power_resource_list_lock);
908	913	}
909	914
910		−int acpi_add_power_resource(acpi_handle handle)
	915	+struct acpi_device *acpi_add_power_resource(acpi_handle handle)
911	916	{
912	917	struct acpi_power_resource *resource;
913	918	struct acpi_device *device = NULL;
@@ -918,11 +923,11 @@ int acpi_add_power_resource(acpi_handle handle) · 4 unmodified lines
918	923
919	924	acpi_bus_get_device(handle, &device);
920	925	if (device)
921		− return 0;
	926	+ return device;
922	927
923	928	resource = kzalloc(sizeof(*resource), GFP_KERNEL);
924	929	if (!resource)
925		− return -ENOMEM;
	930	+ return NULL;
926	931
927	932	device = &resource->device;
928	933	acpi_init_device_object(device, handle, ACPI_BUS_TYPE_POWER);
@@ -959,11 +964,11 @@ int acpi_add_power_resource(acpi_handle handle) · 30 unmodified lines
959	964
960	965	acpi_power_add_resource_to_list(resource);
961	966	acpi_device_add_finalize(device);
962		− return 0;
	967	+ return device;
963	968
964	969	err:
965	970	acpi_release_power_resource(&device->dev);
966		− return result;
	971	+ return NULL;
967	972	}
968	973
969	974	#ifdef CONFIG_ACPI_SLEEP
@@ -997,7 +1002,38 @@ void acpi_resume_power_resources(void) · 27 unmodified lines
997	1002	}
998	1003	#endif
999	1004
1000		−void acpi_turn_off_unused_power_resources(void)
	1005	+static void acpi_power_turn_off_if_unused(struct acpi_power_resource *resource,
	1006	+ bool init)
	1007	+{
	1008	+ if (resource->ref_count > 0)
	1009	+ return;
	1010	+
	1011	+ if (init) {
	1012	+ if (resource->users > 0)
	1013	+ return;
	1014	+ } else {
	1015	+ int result, state;
	1016	+
	1017	+ result = acpi_power_get_state(resource->device.handle, &state);
	1018	+ if (result \|\| state == ACPI_POWER_RESOURCE_STATE_OFF)
	1019	+ return;
	1020	+ }
	1021	+
	1022	+ dev_info(&resource->device.dev, "Turning OFF\n");
	1023	+ __acpi_power_off(resource);
	1024	+}
	1025	+
	1026	+/**
	1027	+ * acpi_turn_off_unused_power_resources - Turn off power resources not in use.
	1028	+ * @init: Control switch.
	1029	+ *
	1030	+ * If @ainit is set, unconditionally turn off all of the ACPI power resources
	1031	+ * without any users.
	1032	+ *
	1033	+ * Otherwise, turn off all ACPI power resources without active references (that
	1034	+ * is, the ones that should be "off" at the moment) that are "on".
	1035	+ */
	1036	+void acpi_turn_off_unused_power_resources(bool init)
1001	1037	{
1002	1038	struct acpi_power_resource *resource;
1003	1039
@@ -1006,10 +1042,7 @@ void acpi_turn_off_unused_power_resources(void) · 2 unmodified lines
1006	1042	list_for_each_entry_reverse(resource, &acpi_power_resource_list, list_node) {
1007	1043	mutex_lock(&resource->resource_lock);
1008	1044
1009		− if (!resource->ref_count) {
1010		− dev_info(&resource->device.dev, "Turning OFF\n");
1011		− __acpi_power_off(resource);
1012		− }
	1045	+ acpi_power_turn_off_if_unused(resource, init);
1013	1046
1014	1047	mutex_unlock(&resource->resource_lock);
1015	1048	}

drivers/acpi/scan.c+1−1

@@ -2360,7 +2360,7 @@ int __init acpi_scan_init(void) · 2359 unmodified lines
2360	2360	}
2361	2361	}
2362	2362
2363		− acpi_turn_off_unused_power_resources();
	2363	+ acpi_turn_off_unused_power_resources(true);
2364	2364
2365	2365	acpi_scan_initialized = true;
2366	2366

drivers/acpi/sleep.c+2−4

@@ -504,7 +504,7 @@ static void acpi_pm_start(u32 acpi_state) · 503 unmodified lines
504	504	*/
505	505	static void acpi_pm_end(void)
506	506	{
507		− acpi_turn_off_unused_power_resources();
	507	+ acpi_turn_off_unused_power_resources(false);
508	508	acpi_scan_lock_release();
509	509	/*
510	510	* This is necessary in case acpi_pm_finish() is not called during a
@@ -1009,10 +1009,8 @@ static void acpi_sleep_hibernate_setup(void) · 498 unmodified lines
1009	1009	return;
1010	1010
1011	1011	acpi_get_table(ACPI_SIG_FACS, 1, (struct acpi_table_header **)&facs);
1012		− if (facs) {
	1012	+ if (facs)
1013	1013	s4_hardware_signature = facs->hardware_signature;
1014		− acpi_put_table((struct acpi_table_header *)facs);
1015		− }
1016	1014	}
1017	1015	#else /* !CONFIG_HIBERNATION */
1018	1016	static inline void acpi_sleep_hibernate_setup(void) {}

drivers/base/core.c+36−32

@@ -194,6 +194,17 @@ int device_links_read_lock_held(void) · 193 unmodified lines
194	194	{
195	195	return srcu_read_lock_held(&device_links_srcu);
196	196	}
	197	+
	198	+static void device_link_synchronize_removal(void)
	199	+{
	200	+ synchronize_srcu(&device_links_srcu);
	201	+}
	202	+
	203	+static void device_link_remove_from_lists(struct device_link *link)
	204	+{
	205	+ list_del_rcu(&link->s_node);
	206	+ list_del_rcu(&link->c_node);
	207	+}
197	208	#else /* !CONFIG_SRCU */
198	209	static DECLARE_RWSEM(device_links_lock);
199	210
@@ -224,6 +235,16 @@ int device_links_read_lock_held(void) · 24 unmodified lines
224	235	return lockdep_is_held(&device_links_lock);
225	236	}
226	237	#endif
	238	+
	239	+static inline void device_link_synchronize_removal(void)
	240	+{
	241	+}
	242	+
	243	+static void device_link_remove_from_lists(struct device_link *link)
	244	+{
	245	+ list_del(&link->s_node);
	246	+ list_del(&link->c_node);
	247	+}
227	248	#endif /* !CONFIG_SRCU */
228	249
229	250	static bool device_is_ancestor(struct device dev, struct device target)
@@ -445,8 +466,13 @@ static struct attribute *devlink_attrs[] = { · 215 unmodified lines
445	466	};
446	467	ATTRIBUTE_GROUPS(devlink);
447	468
448		−static void device_link_free(struct device_link *link)
	469	+static void device_link_release_fn(struct work_struct *work)
449	470	{
	471	+ struct device_link *link = container_of(work, struct device_link, rm_work);
	472	+
	473	+ /* Ensure that all references to the link object have been dropped. */
	474	+ device_link_synchronize_removal();
	475	+
450	476	while (refcount_dec_not_one(&link->rpm_active))
451	477	pm_runtime_put(link->supplier);
452	478
@@ -455,24 +481,19 @@ static void device_link_free(struct device_link *link) · 2 unmodified lines
455	481	kfree(link);
456	482	}
457	483
458		−#ifdef CONFIG_SRCU
459		−static void __device_link_free_srcu(struct rcu_head *rhead)
460		−{
461		− device_link_free(container_of(rhead, struct device_link, rcu_head));
462		−}
463		−
464	484	static void devlink_dev_release(struct device *dev)
465	485	{
466	486	struct device_link *link = to_devlink(dev);
467	487
468		− call_srcu(&device_links_srcu, &link->rcu_head, __device_link_free_srcu);
469		−}
470		−#else
471		−static void devlink_dev_release(struct device *dev)
472		−{
473		− device_link_free(to_devlink(dev));
	488	+ INIT_WORK(&link->rm_work, device_link_release_fn);
	489	+ /*
	490	+ * It may take a while to complete this work because of the SRCU
	491	+ * synchronization in device_link_release_fn() and if the consumer or
	492	+ * supplier devices get deleted when it runs, so put it into the "long"
	493	+ * workqueue.
	494	+ */
	495	+ queue_work(system_long_wq, &link->rm_work);
474	496	}
475		−#endif
476	497
477	498	static struct class devlink_class = {
478	499	.name = "devlink",
@@ -846,7 +867,6 @@ out: · 367 unmodified lines
846	867	}
847	868	EXPORT_SYMBOL_GPL(device_link_add);
848	869
849		−#ifdef CONFIG_SRCU
850	870	static void __device_link_del(struct kref *kref)
851	871	{
852	872	struct device_link *link = container_of(kref, struct device_link, kref);
@@ -856,25 +876,9 @@ static void __device_link_del(struct kref *kref) · 3 unmodified lines
856	876
857	877	pm_runtime_drop_link(link);
858	878
859		− list_del_rcu(&link->s_node);
860		− list_del_rcu(&link->c_node);
861		− device_unregister(&link->link_dev);
862		−}
863		−#else /* !CONFIG_SRCU */
864		−static void __device_link_del(struct kref *kref)
865		−{
866		− struct device_link *link = container_of(kref, struct device_link, kref);
867		−
868		− dev_info(link->consumer, "Dropping the link to %s\n",
869		− dev_name(link->supplier));
870		−
871		− pm_runtime_drop_link(link);
872		−
873		− list_del(&link->s_node);
874		− list_del(&link->c_node);
	879	+ device_link_remove_from_lists(link);
875	880	device_unregister(&link->link_dev);
876	881	}
877		−#endif /* !CONFIG_SRCU */
878	882
879	883	static void device_link_put_kref(struct device_link *link)
880	884	{

drivers/base/memory.c+3−3

@@ -218,14 +218,14 @@ static int memory_block_offline(struct memory_block *mem) · 217 unmodified lines
218	218	struct zone *zone;
219	219	int ret;
220	220
221		− zone = page_zone(pfn_to_page(start_pfn));
222		−
223	221	/*
224	222	* Unaccount before offlining, such that unpopulated zone and kthreads
225	223	* can properly be torn down in offline_pages().
226	224	*/
227		− if (nr_vmemmap_pages)
	225	+ if (nr_vmemmap_pages) {
	226	+ zone = page_zone(pfn_to_page(start_pfn));
228	227	adjust_present_page_count(zone, -nr_vmemmap_pages);
	228	+ }
229	229
230	230	ret = offline_pages(start_pfn + nr_vmemmap_pages,
231	231	nr_pages - nr_vmemmap_pages);

drivers/block/loop.c+7−18

@@ -1879,29 +1879,18 @@ static int lo_compat_ioctl(struct block_device *bdev, fmode_t mode, · 1878 unmodified lines
1879	1879
1880	1880	static int lo_open(struct block_device *bdev, fmode_t mode)
1881	1881	{
1882		− struct loop_device *lo;
	1882	+ struct loop_device *lo = bdev->bd_disk->private_data;
1883	1883	int err;
1884	1884
1885		− /*
1886		− * take loop_ctl_mutex to protect lo pointer from race with
1887		− * loop_control_ioctl(LOOP_CTL_REMOVE), however, to reduce contention
1888		− * release it prior to updating lo->lo_refcnt.
1889		− */
1890		− err = mutex_lock_killable(&loop_ctl_mutex);
1891		− if (err)
1892		− return err;
1893		− lo = bdev->bd_disk->private_data;
1894		− if (!lo) {
1895		− mutex_unlock(&loop_ctl_mutex);
1896		− return -ENXIO;
1897		− }
1898	1885	err = mutex_lock_killable(&lo->lo_mutex);
1899		− mutex_unlock(&loop_ctl_mutex);
1900	1886	if (err)
1901	1887	return err;
1902		− atomic_inc(&lo->lo_refcnt);
	1888	+ if (lo->lo_state == Lo_deleting)
	1889	+ err = -ENXIO;
	1890	+ else
	1891	+ atomic_inc(&lo->lo_refcnt);
1903	1892	mutex_unlock(&lo->lo_mutex);
1904		− return 0;
	1893	+ return err;
1905	1894	}
1906	1895
1907	1896	static void lo_release(struct gendisk *disk, fmode_t mode)
@@ -2285,7 +2274,7 @@ static long loop_control_ioctl(struct file *file, unsigned int cmd, · 377 unmodified lines
2285	2274	mutex_unlock(&lo->lo_mutex);
2286	2275	break;
2287	2276	}
2288		− lo->lo_disk->private_data = NULL;
	2277	+ lo->lo_state = Lo_deleting;
2289	2278	mutex_unlock(&lo->lo_mutex);
2290	2279	idr_remove(&loop_index_idr, lo->lo_number);
2291	2280	loop_remove(lo);

drivers/block/loop.h+1−0

@@ -22,6 +22,7 @@ enum { · 21 unmodified lines
22	22	Lo_unbound,
23	23	Lo_bound,
24	24	Lo_rundown,
	25	+ Lo_deleting,
25	26	};
26	27
27	28	struct loop_func_table;

drivers/bluetooth/btusb.c+23−2

@@ -388,6 +388,8 @@ static const struct usb_device_id blacklist_table[] = { · 387 unmodified lines
388	388	/* Realtek 8822CE Bluetooth devices */
389	389	{ USB_DEVICE(0x0bda, 0xb00c), .driver_info = BTUSB_REALTEK \|
390	390	BTUSB_WIDEBAND_SPEECH },
	391	+ { USB_DEVICE(0x0bda, 0xc822), .driver_info = BTUSB_REALTEK \|
	392	+ BTUSB_WIDEBAND_SPEECH },
391	393
392	394	/* Realtek 8852AE Bluetooth devices */
393	395	{ USB_DEVICE(0x0bda, 0xc852), .driver_info = BTUSB_REALTEK \|
@@ -2527,10 +2529,17 @@ static int btusb_intel_download_firmware_newgen(struct hci_dev *hdev, · 2133 unmodified lines
2527	2529	}
2528	2530
2529	2531	btusb_setup_intel_newgen_get_fw_name(ver, fwname, sizeof(fwname), "sfi");
2530		− err = request_firmware(&fw, fwname, &hdev->dev);
	2532	+ err = firmware_request_nowarn(&fw, fwname, &hdev->dev);
2531	2533	if (err < 0) {
	2534	+ if (!test_bit(BTUSB_BOOTLOADER, &data->flags)) {
	2535	+ /* Firmware has already been loaded */
	2536	+ set_bit(BTUSB_FIRMWARE_LOADED, &data->flags);
	2537	+ return 0;
	2538	+ }
	2539	+
2532	2540	bt_dev_err(hdev, "Failed to load Intel firmware file %s (%d)",
2533	2541	fwname, err);
	2542	+
2534	2543	return err;
2535	2544	}
2536	2545
@@ -2680,12 +2689,24 @@ download: · 143 unmodified lines
2680	2689	err = btusb_setup_intel_new_get_fw_name(ver, params, fwname,
2681	2690	sizeof(fwname), "sfi");
2682	2691	if (err < 0) {
	2692	+ if (!test_bit(BTUSB_BOOTLOADER, &data->flags)) {
	2693	+ /* Firmware has already been loaded */
	2694	+ set_bit(BTUSB_FIRMWARE_LOADED, &data->flags);
	2695	+ return 0;
	2696	+ }
	2697	+
2683	2698	bt_dev_err(hdev, "Unsupported Intel firmware naming");
2684	2699	return -EINVAL;
2685	2700	}
2686	2701
2687		− err = request_firmware(&fw, fwname, &hdev->dev);
	2702	+ err = firmware_request_nowarn(&fw, fwname, &hdev->dev);
2688	2703	if (err < 0) {
	2704	+ if (!test_bit(BTUSB_BOOTLOADER, &data->flags)) {
	2705	+ /* Firmware has already been loaded */
	2706	+ set_bit(BTUSB_FIRMWARE_LOADED, &data->flags);
	2707	+ return 0;
	2708	+ }
	2709	+
2689	2710	bt_dev_err(hdev, "Failed to load Intel firmware file %s (%d)",
2690	2711	fwname, err);
2691	2712	return err;

drivers/bus/mhi/pci_generic.c+38−4

@@ -311,8 +311,8 @@ static const struct mhi_channel_config mhi_foxconn_sdx55_channels[] = { · 310 unmodified lines
311	311	MHI_CHANNEL_CONFIG_DL(5, "DIAG", 32, 1),
312	312	MHI_CHANNEL_CONFIG_UL(12, "MBIM", 32, 0),
313	313	MHI_CHANNEL_CONFIG_DL(13, "MBIM", 32, 0),
314		− MHI_CHANNEL_CONFIG_UL(32, "AT", 32, 0),
315		− MHI_CHANNEL_CONFIG_DL(33, "AT", 32, 0),
	314	+ MHI_CHANNEL_CONFIG_UL(32, "DUN", 32, 0),
	315	+ MHI_CHANNEL_CONFIG_DL(33, "DUN", 32, 0),
316	316	MHI_CHANNEL_CONFIG_HW_UL(100, "IP_HW0_MBIM", 128, 2),
317	317	MHI_CHANNEL_CONFIG_HW_DL(101, "IP_HW0_MBIM", 128, 3),
318	318	};
@@ -708,7 +708,7 @@ static void mhi_pci_remove(struct pci_dev *pdev) · 389 unmodified lines
708	708	struct mhi_pci_device *mhi_pdev = pci_get_drvdata(pdev);
709	709	struct mhi_controller *mhi_cntrl = &mhi_pdev->mhi_cntrl;
710	710
711		− del_timer(&mhi_pdev->health_check_timer);
	711	+ del_timer_sync(&mhi_pdev->health_check_timer);
712	712	cancel_work_sync(&mhi_pdev->recovery_work);
713	713
714	714	if (test_and_clear_bit(MHI_PCI_DEV_STARTED, &mhi_pdev->status)) {
@@ -935,9 +935,43 @@ static int __maybe_unused mhi_pci_resume(struct device *dev) · 220 unmodified lines
935	935	return ret;
936	936	}
937	937
	938	+static int __maybe_unused mhi_pci_freeze(struct device *dev)
	939	+{
	940	+ struct mhi_pci_device *mhi_pdev = dev_get_drvdata(dev);
	941	+ struct mhi_controller *mhi_cntrl = &mhi_pdev->mhi_cntrl;
	942	+
	943	+ /* We want to stop all operations, hibernation does not guarantee that
	944	+ * device will be in the same state as before freezing, especially if
	945	+ * the intermediate restore kernel reinitializes MHI device with new
	946	+ * context.
	947	+ */
	948	+ if (test_and_clear_bit(MHI_PCI_DEV_STARTED, &mhi_pdev->status)) {
	949	+ mhi_power_down(mhi_cntrl, false);
	950	+ mhi_unprepare_after_power_down(mhi_cntrl);
	951	+ }
	952	+
	953	+ return 0;
	954	+}
	955	+
	956	+static int __maybe_unused mhi_pci_restore(struct device *dev)
	957	+{
	958	+ struct mhi_pci_device *mhi_pdev = dev_get_drvdata(dev);
	959	+
	960	+ /* Reinitialize the device */
	961	+ queue_work(system_long_wq, &mhi_pdev->recovery_work);
	962	+
	963	+ return 0;
	964	+}
	965	+
938	966	static const struct dev_pm_ops mhi_pci_pm_ops = {
939	967	SET_RUNTIME_PM_OPS(mhi_pci_runtime_suspend, mhi_pci_runtime_resume, NULL)
940		− SET_SYSTEM_SLEEP_PM_OPS(mhi_pci_suspend, mhi_pci_resume)
	968	+#ifdef CONFIG_PM_SLEEP
	969	+ .suspend = mhi_pci_suspend,
	970	+ .resume = mhi_pci_resume,
	971	+ .freeze = mhi_pci_freeze,
	972	+ .thaw = mhi_pci_restore,
	973	+ .restore = mhi_pci_restore,
	974	+#endif
941	975	};
942	976
943	977	static struct pci_driver mhi_pci_driver = {

drivers/bus/ti-sysc.c+54−6

@@ -1334,6 +1334,34 @@ err_allow_idle: · 1333 unmodified lines
1334	1334	return error;
1335	1335	}
1336	1336
	1337	+static int sysc_reinit_module(struct sysc *ddata, bool leave_enabled)
	1338	+{
	1339	+ struct device *dev = ddata->dev;
	1340	+ int error;
	1341	+
	1342	+ /* Disable target module if it is enabled */
	1343	+ if (ddata->enabled) {
	1344	+ error = sysc_runtime_suspend(dev);
	1345	+ if (error)
	1346	+ dev_warn(dev, "reinit suspend failed: %i\n", error);
	1347	+ }
	1348	+
	1349	+ /* Enable target module */
	1350	+ error = sysc_runtime_resume(dev);
	1351	+ if (error)
	1352	+ dev_warn(dev, "reinit resume failed: %i\n", error);
	1353	+
	1354	+ if (leave_enabled)
	1355	+ return error;
	1356	+
	1357	+ /* Disable target module if no leave_enabled was set */
	1358	+ error = sysc_runtime_suspend(dev);
	1359	+ if (error)
	1360	+ dev_warn(dev, "reinit suspend failed: %i\n", error);
	1361	+
	1362	+ return error;
	1363	+}
	1364	+
1337	1365	static int __maybe_unused sysc_noirq_suspend(struct device *dev)
1338	1366	{
1339	1367	struct sysc *ddata;
@@ -1344,12 +1372,18 @@ static int __maybe_unused sysc_noirq_suspend(struct device *dev) · 4 unmodified lines
1344	1372	(SYSC_QUIRK_LEGACY_IDLE \| SYSC_QUIRK_NO_IDLE))
1345	1373	return 0;
1346	1374
1347		− return pm_runtime_force_suspend(dev);
	1375	+ if (!ddata->enabled)
	1376	+ return 0;
	1377	+
	1378	+ ddata->needs_resume = 1;
	1379	+
	1380	+ return sysc_runtime_suspend(dev);
1348	1381	}
1349	1382
1350	1383	static int __maybe_unused sysc_noirq_resume(struct device *dev)
1351	1384	{
1352	1385	struct sysc *ddata;
	1386	+ int error = 0;
1353	1387
1354	1388	ddata = dev_get_drvdata(dev);
1355	1389
@@ -1357,7 +1391,19 @@ static int __maybe_unused sysc_noirq_resume(struct device *dev) · 1 unmodified line
1357	1391	(SYSC_QUIRK_LEGACY_IDLE \| SYSC_QUIRK_NO_IDLE))
1358	1392	return 0;
1359	1393
1360		− return pm_runtime_force_resume(dev);
	1394	+ if (ddata->cfg.quirks & SYSC_QUIRK_REINIT_ON_RESUME) {
	1395	+ error = sysc_reinit_module(ddata, ddata->needs_resume);
	1396	+ if (error)
	1397	+ dev_warn(dev, "noirq_resume failed: %i\n", error);
	1398	+ } else if (ddata->needs_resume) {
	1399	+ error = sysc_runtime_resume(dev);
	1400	+ if (error)
	1401	+ dev_warn(dev, "noirq_resume failed: %i\n", error);
	1402	+ }
	1403	+
	1404	+ ddata->needs_resume = 0;
	1405	+
	1406	+ return error;
1361	1407	}
1362	1408
1363	1409	static const struct dev_pm_ops sysc_pm_ops = {
@@ -1408,9 +1454,9 @@ static const struct sysc_revision_quirk sysc_revision_quirks[] = { · 44 unmodified lines
1408	1454	SYSC_QUIRK_SWSUP_SIDLE \| SYSC_QUIRK_LEGACY_IDLE),
1409	1455	/* Uarts on omap4 and later */
1410	1456	SYSC_QUIRK("uart", 0, 0x50, 0x54, 0x58, 0x50411e03, 0xffff00ff,
1411		− SYSC_QUIRK_SWSUP_SIDLE_ACT \| SYSC_QUIRK_LEGACY_IDLE),
	1457	+ SYSC_QUIRK_SWSUP_SIDLE \| SYSC_QUIRK_LEGACY_IDLE),
1412	1458	SYSC_QUIRK("uart", 0, 0x50, 0x54, 0x58, 0x47422e03, 0xffffffff,
1413		− SYSC_QUIRK_SWSUP_SIDLE_ACT \| SYSC_QUIRK_LEGACY_IDLE),
	1459	+ SYSC_QUIRK_SWSUP_SIDLE \| SYSC_QUIRK_LEGACY_IDLE),
1414	1460
1415	1461	/* Quirks that need to be set based on the module address */
1416	1462	SYSC_QUIRK("mcpdm", 0x40132000, 0, 0x10, -ENODEV, 0x50000800, 0xffffffff,
@@ -1459,6 +1505,8 @@ static const struct sysc_revision_quirk sysc_revision_quirks[] = { · 42 unmodified lines
1459	1505	SYSC_QUIRK_SWSUP_SIDLE \| SYSC_QUIRK_SWSUP_MSTANDBY),
1460	1506	SYSC_QUIRK("tptc", 0, 0, -ENODEV, -ENODEV, 0x40007c00, 0xffffffff,
1461	1507	SYSC_QUIRK_SWSUP_SIDLE \| SYSC_QUIRK_SWSUP_MSTANDBY),
	1508	+ SYSC_QUIRK("sata", 0, 0xfc, 0x1100, -ENODEV, 0x5e412000, 0xffffffff,
	1509	+ SYSC_QUIRK_SWSUP_SIDLE \| SYSC_QUIRK_SWSUP_MSTANDBY),
1462	1510	SYSC_QUIRK("usb_host_hs", 0, 0, 0x10, 0x14, 0x50700100, 0xffffffff,
1463	1511	SYSC_QUIRK_SWSUP_SIDLE \| SYSC_QUIRK_SWSUP_MSTANDBY),
1464	1512	SYSC_QUIRK("usb_host_hs", 0, 0, 0x10, -ENODEV, 0x50700101, 0xffffffff,
@@ -1466,7 +1514,8 @@ static const struct sysc_revision_quirk sysc_revision_quirks[] = { · 1 unmodified line
1466	1514	SYSC_QUIRK("usb_otg_hs", 0, 0x400, 0x404, 0x408, 0x00000050,
1467	1515	0xffffffff, SYSC_QUIRK_SWSUP_SIDLE \| SYSC_QUIRK_SWSUP_MSTANDBY),
1468	1516	SYSC_QUIRK("usb_otg_hs", 0, 0, 0x10, -ENODEV, 0x4ea2080d, 0xffffffff,
1469		− SYSC_QUIRK_SWSUP_SIDLE \| SYSC_QUIRK_SWSUP_MSTANDBY),
	1517	+ SYSC_QUIRK_SWSUP_SIDLE \| SYSC_QUIRK_SWSUP_MSTANDBY \|
	1518	+ SYSC_QUIRK_REINIT_ON_RESUME),
1470	1519	SYSC_QUIRK("wdt", 0, 0, 0x10, 0x14, 0x502a0500, 0xfffff0f0,
1471	1520	SYSC_MODULE_QUIRK_WDT),
1472	1521	/* PRUSS on am3, am4 and am5 */
@@ -1524,7 +1573,6 @@ static const struct sysc_revision_quirk sysc_revision_quirks[] = { · 51 unmodified lines
1524	1573	SYSC_QUIRK("prcm", 0, 0, -ENODEV, -ENODEV, 0x40000400, 0xffffffff, 0),
1525	1574	SYSC_QUIRK("rfbi", 0x4832a800, 0, 0x10, 0x14, 0x00000010, 0xffffffff, 0),
1526	1575	SYSC_QUIRK("rfbi", 0x58002000, 0, 0x10, 0x14, 0x00000010, 0xffffffff, 0),
1527		− SYSC_QUIRK("sata", 0, 0xfc, 0x1100, -ENODEV, 0x5e412000, 0xffffffff, 0),
1528	1576	SYSC_QUIRK("scm", 0, 0, 0x10, -ENODEV, 0x40000900, 0xffffffff, 0),
1529	1577	SYSC_QUIRK("scm", 0, 0, -ENODEV, -ENODEV, 0x4e8b0100, 0xffffffff, 0),
1530	1578	SYSC_QUIRK("scm", 0, 0, -ENODEV, -ENODEV, 0x4f000100, 0xffffffff, 0),

drivers/cpufreq/Kconfig.arm+0−10

@@ -19,16 +19,6 @@ config ACPI_CPPC_CPUFREQ · 18 unmodified lines
19	19
20	20	If in doubt, say N.
21	21
22		−config ACPI_CPPC_CPUFREQ_FIE
23		− bool "Frequency Invariance support for CPPC cpufreq driver"
24		− depends on ACPI_CPPC_CPUFREQ && GENERIC_ARCH_TOPOLOGY
25		− default y
26		− help
27		− This extends frequency invariance support in the CPPC cpufreq driver,
28		− by using CPPC delivered and reference performance counters.
29		−
30		− If in doubt, say N.
31		−
32	22	config ARM_ALLWINNER_SUN50I_CPUFREQ_NVMEM
33	23	tristate "Allwinner nvmem based SUN50I CPUFreq driver"
34	24	depends on ARCH_SUNXI

drivers/cpufreq/cppc_cpufreq.c+12−233

@@ -10,18 +10,14 @@ · 9 unmodified lines
10	10
11	11	#define pr_fmt(fmt) "CPPC Cpufreq:" fmt
12	12
13		−#include <linux/arch_topology.h>
14	13	#include <linux/kernel.h>
15	14	#include <linux/module.h>
16	15	#include <linux/delay.h>
17	16	#include <linux/cpu.h>
18	17	#include <linux/cpufreq.h>
19	18	#include <linux/dmi.h>
20		−#include <linux/irq_work.h>
21		−#include <linux/kthread.h>
22	19	#include <linux/time.h>
23	20	#include <linux/vmalloc.h>
24		−#include <uapi/linux/sched/types.h>
25	21
26	22	#include <asm/unaligned.h>
27	23
@@ -61,204 +57,6 @@ static struct cppc_workaround_oem_info wa_info[] = { · 33 unmodified lines
61	57	}
62	58	};
63	59
64		−#ifdef CONFIG_ACPI_CPPC_CPUFREQ_FIE
65		−
66		−/* Frequency invariance support */
67		−struct cppc_freq_invariance {
68		− int cpu;
69		− struct irq_work irq_work;
70		− struct kthread_work work;
71		− struct cppc_perf_fb_ctrs prev_perf_fb_ctrs;
72		− struct cppc_cpudata *cpu_data;
73		−};
74		−
75		−static DEFINE_PER_CPU(struct cppc_freq_invariance, cppc_freq_inv);
76		−static struct kthread_worker *kworker_fie;
77		−static bool fie_disabled;
78		−
79		−static struct cpufreq_driver cppc_cpufreq_driver;
80		−static unsigned int hisi_cppc_cpufreq_get_rate(unsigned int cpu);
81		−static int cppc_perf_from_fbctrs(struct cppc_cpudata *cpu_data,
82		− struct cppc_perf_fb_ctrs fb_ctrs_t0,
83		− struct cppc_perf_fb_ctrs fb_ctrs_t1);
84		−
85		−/**
86		− * cppc_scale_freq_workfn - CPPC arch_freq_scale updater for frequency invariance
87		− * @work: The work item.
88		− *
89		− * The CPPC driver register itself with the topology core to provide its own
90		− * implementation (cppc_scale_freq_tick()) of topology_scale_freq_tick() which
91		− * gets called by the scheduler on every tick.
92		− *
93		− * Note that the arch specific counters have higher priority than CPPC counters,
94		− * if available, though the CPPC driver doesn't need to have any special
95		− * handling for that.
96		− *
97		− * On an invocation of cppc_scale_freq_tick(), we schedule an irq work (since we
98		− * reach here from hard-irq context), which then schedules a normal work item
99		− * and cppc_scale_freq_workfn() updates the per_cpu arch_freq_scale variable
100		− * based on the counter updates since the last tick.
101		− */
102		−static void cppc_scale_freq_workfn(struct kthread_work *work)
103		−{
104		− struct cppc_freq_invariance *cppc_fi;
105		− struct cppc_perf_fb_ctrs fb_ctrs = {0};
106		− struct cppc_cpudata *cpu_data;
107		− unsigned long local_freq_scale;
108		− u64 perf;
109		−
110		− cppc_fi = container_of(work, struct cppc_freq_invariance, work);
111		− cpu_data = cppc_fi->cpu_data;
112		−
113		− if (cppc_get_perf_ctrs(cppc_fi->cpu, &fb_ctrs)) {
114		− pr_warn("%s: failed to read perf counters\n", __func__);
115		− return;
116		− }
117		−
118		− cppc_fi->prev_perf_fb_ctrs = fb_ctrs;
119		− perf = cppc_perf_from_fbctrs(cpu_data, cppc_fi->prev_perf_fb_ctrs,
120		− fb_ctrs);
121		−
122		− perf <<= SCHED_CAPACITY_SHIFT;
123		− local_freq_scale = div64_u64(perf, cpu_data->perf_caps.highest_perf);
124		− if (WARN_ON(local_freq_scale > 1024))
125		− local_freq_scale = 1024;
126		−
127		− per_cpu(arch_freq_scale, cppc_fi->cpu) = local_freq_scale;
128		−}
129		−
130		−static void cppc_irq_work(struct irq_work *irq_work)
131		−{
132		− struct cppc_freq_invariance *cppc_fi;
133		−
134		− cppc_fi = container_of(irq_work, struct cppc_freq_invariance, irq_work);
135		− kthread_queue_work(kworker_fie, &cppc_fi->work);
136		−}
137		−
138		−static void cppc_scale_freq_tick(void)
139		−{
140		− struct cppc_freq_invariance *cppc_fi = &per_cpu(cppc_freq_inv, smp_processor_id());
141		−
142		− /*
143		− * cppc_get_perf_ctrs() can potentially sleep, call that from the right
144		− * context.
145		− */
146		− irq_work_queue(&cppc_fi->irq_work);
147		−}
148		−
149		−static struct scale_freq_data cppc_sftd = {
150		− .source = SCALE_FREQ_SOURCE_CPPC,
151		− .set_freq_scale = cppc_scale_freq_tick,
152		−};
153		−
154		−static void cppc_freq_invariance_policy_init(struct cpufreq_policy *policy,
155		− struct cppc_cpudata *cpu_data)
156		−{
157		− struct cppc_perf_fb_ctrs fb_ctrs = {0};
158		− struct cppc_freq_invariance *cppc_fi;
159		− int i, ret;
160		−
161		− if (cppc_cpufreq_driver.get == hisi_cppc_cpufreq_get_rate)
162		− return;
163		−
164		− if (fie_disabled)
165		− return;
166		−
167		− for_each_cpu(i, policy->cpus) {
168		− cppc_fi = &per_cpu(cppc_freq_inv, i);
169		− cppc_fi->cpu = i;
170		− cppc_fi->cpu_data = cpu_data;
171		− kthread_init_work(&cppc_fi->work, cppc_scale_freq_workfn);
172		− init_irq_work(&cppc_fi->irq_work, cppc_irq_work);
173		−
174		− ret = cppc_get_perf_ctrs(i, &fb_ctrs);
175		− if (ret) {
176		− pr_warn("%s: failed to read perf counters: %d\n",
177		− __func__, ret);
178		− fie_disabled = true;
179		− } else {
180		− cppc_fi->prev_perf_fb_ctrs = fb_ctrs;
181		− }
182		− }
183		−}
184		−
185		−static void __init cppc_freq_invariance_init(void)
186		−{
187		− struct sched_attr attr = {
188		− .size = sizeof(struct sched_attr),
189		− .sched_policy = SCHED_DEADLINE,
190		− .sched_nice = 0,
191		− .sched_priority = 0,
192		− /*
193		− * Fake (unused) bandwidth; workaround to "fix"
194		− * priority inheritance.
195		− */
196		− .sched_runtime = 1000000,
197		− .sched_deadline = 10000000,
198		− .sched_period = 10000000,
199		− };
200		− int ret;
201		−
202		− if (cppc_cpufreq_driver.get == hisi_cppc_cpufreq_get_rate)
203		− return;
204		−
205		− if (fie_disabled)
206		− return;
207		−
208		− kworker_fie = kthread_create_worker(0, "cppc_fie");
209		− if (IS_ERR(kworker_fie))
210		− return;
211		−
212		− ret = sched_setattr_nocheck(kworker_fie->task, &attr);
213		− if (ret) {
214		− pr_warn("%s: failed to set SCHED_DEADLINE: %d\n", __func__,
215		− ret);
216		− kthread_destroy_worker(kworker_fie);
217		− return;
218		− }
219		−
220		− /* Register for freq-invariance */
221		− topology_set_scale_freq_source(&cppc_sftd, cpu_present_mask);
222		−}
223		−
224		−static void cppc_freq_invariance_exit(void)
225		−{
226		− struct cppc_freq_invariance *cppc_fi;
227		− int i;
228		−
229		− if (cppc_cpufreq_driver.get == hisi_cppc_cpufreq_get_rate)
230		− return;
231		−
232		− if (fie_disabled)
233		− return;
234		−
235		− topology_clear_scale_freq_source(SCALE_FREQ_SOURCE_CPPC, cpu_present_mask);
236		−
237		− for_each_possible_cpu(i) {
238		− cppc_fi = &per_cpu(cppc_freq_inv, i);
239		− irq_work_sync(&cppc_fi->irq_work);
240		− }
241		−
242		− kthread_destroy_worker(kworker_fie);
243		− kworker_fie = NULL;
244		−}
245		−
246		−#else
247		−static inline void
248		−cppc_freq_invariance_policy_init(struct cpufreq_policy *policy,
249		− struct cppc_cpudata *cpu_data)
250		−{
251		−}
252		−
253		−static inline void cppc_freq_invariance_init(void)
254		−{
255		−}
256		−
257		−static inline void cppc_freq_invariance_exit(void)
258		−{
259		−}
260		−#endif /* CONFIG_ACPI_CPPC_CPUFREQ_FIE */
261		−
262	60	/* Callback function used to retrieve the max frequency from DMI */
263	61	static void cppc_find_dmi_mhz(const struct dmi_header dm, void private)
264	62	{
@@ -547,12 +345,9 @@ static int cppc_cpufreq_cpu_init(struct cpufreq_policy *policy) · 282 unmodified lines
547	345	cpu_data->perf_ctrls.desired_perf = caps->highest_perf;
548	346
549	347	ret = cppc_set_perf(cpu, &cpu_data->perf_ctrls);
550		− if (ret) {
	348	+ if (ret)
551	349	pr_debug("Err setting perf value:%d on CPU:%d. ret:%d\n",
552	350	caps->highest_perf, cpu, ret);
553		− } else {
554		− cppc_freq_invariance_policy_init(policy, cpu_data);
555		− }
556	351
557	352	return ret;
558	353	}
@@ -565,12 +360,12 @@ static inline u64 get_delta(u64 t1, u64 t0) · 6 unmodified lines
565	360	return (u32)t1 - (u32)t0;
566	361	}
567	362
568		−static int cppc_perf_from_fbctrs(struct cppc_cpudata *cpu_data,
569		− struct cppc_perf_fb_ctrs fb_ctrs_t0,
570		− struct cppc_perf_fb_ctrs fb_ctrs_t1)
	363	+static int cppc_get_rate_from_fbctrs(struct cppc_cpudata *cpu_data,
	364	+ struct cppc_perf_fb_ctrs fb_ctrs_t0,
	365	+ struct cppc_perf_fb_ctrs fb_ctrs_t1)
571	366	{
572	367	u64 delta_reference, delta_delivered;
573		− u64 reference_perf;
	368	+ u64 reference_perf, delivered_perf;
574	369
575	370	reference_perf = fb_ctrs_t0.reference_perf;
576	371
@@ -579,21 +374,12 @@ static int cppc_perf_from_fbctrs(struct cppc_cpudata *cpu_data, · 2 unmodified lines
579	374	delta_delivered = get_delta(fb_ctrs_t1.delivered,
580	375	fb_ctrs_t0.delivered);
581	376
582		− /* Check to avoid divide-by zero and invalid delivered_perf */
583		− if (!delta_reference \|\| !delta_delivered)
584		− return cpu_data->perf_ctrls.desired_perf;
585		−
586		− return (reference_perf * delta_delivered) / delta_reference;
587		−}
588		−
589		−static int cppc_get_rate_from_fbctrs(struct cppc_cpudata *cpu_data,
590		− struct cppc_perf_fb_ctrs fb_ctrs_t0,
591		− struct cppc_perf_fb_ctrs fb_ctrs_t1)
592		−{
593		− u64 delivered_perf;
594		−
595		− delivered_perf = cppc_perf_from_fbctrs(cpu_data, fb_ctrs_t0,
596		− fb_ctrs_t1);
	377	+ /* Check to avoid divide-by zero */
	378	+ if (delta_reference \|\| delta_delivered)
	379	+ delivered_perf = (reference_perf * delta_delivered) /
	380	+ delta_reference;
	381	+ else
	382	+ delivered_perf = cpu_data->perf_ctrls.desired_perf;
597	383
598	384	return cppc_cpufreq_perf_to_khz(cpu_data, delivered_perf);
599	385	}
@@ -718,8 +504,6 @@ static void cppc_check_hisi_workaround(void) · 118 unmodified lines
718	504
719	505	static int __init cppc_cpufreq_init(void)
720	506	{
721		− int ret;
722		−
723	507	if ((acpi_disabled) \|\| !acpi_cpc_valid())
724	508	return -ENODEV;
725	509
@@ -727,11 +511,7 @@ static int __init cppc_cpufreq_init(void) · 1 unmodified line
727	511
728	512	cppc_check_hisi_workaround();
729	513
730		− ret = cpufreq_register_driver(&cppc_cpufreq_driver);
731		− if (!ret)
732		− cppc_freq_invariance_init();
733		−
734		− return ret;
	514	+ return cpufreq_register_driver(&cppc_cpufreq_driver);
735	515	}
736	516
737	517	static inline void free_cpu_data(void)
@@ -748,7 +528,6 @@ static inline void free_cpu_data(void) · 10 unmodified lines
748	528
749	529	static void __exit cppc_cpufreq_exit(void)
750	530	{
751		− cppc_freq_invariance_exit();
752	531	cpufreq_unregister_driver(&cppc_cpufreq_driver);
753	532
754	533	free_cpu_data();

drivers/dma/Kconfig+2−0

@@ -59,6 +59,7 @@ config DMA_OF · 58 unmodified lines
59	59	#devices
60	60	config ALTERA_MSGDMA
61	61	tristate "Altera / Intel mSGDMA Engine"
	62	+ depends on HAS_IOMEM
62	63	select DMA_ENGINE
63	64	help
64	65	Enable support for Altera / Intel mSGDMA controller.
@@ -701,6 +702,7 @@ config XILINX_ZYNQMP_DMA · 636 unmodified lines
701	702
702	703	config XILINX_ZYNQMP_DPDMA
703	704	tristate "Xilinx DPDMA Engine"
	705	+ depends on HAS_IOMEM && OF
704	706	select DMA_ENGINE
705	707	select DMA_VIRTUAL_CHANNELS
706	708	help

drivers/dma/fsl-dpaa2-qdma/dpaa2-qdma.c+3−0

@@ -332,6 +332,7 @@ static int __cold dpaa2_qdma_setup(struct fsl_mc_device *ls_dev) · 331 unmodified lines
332	332	}
333	333
334	334	if (priv->dpdmai_attr.version.major > DPDMAI_VER_MAJOR) {
	335	+ err = -EINVAL;
335	336	dev_err(dev, "DPDMAI major version mismatch\n"
336	337	"Found %u.%u, supported version is %u.%u\n",
337	338	priv->dpdmai_attr.version.major,
@@ -341,6 +342,7 @@ static int __cold dpaa2_qdma_setup(struct fsl_mc_device *ls_dev) · 3 unmodified lines
341	342	}
342	343
343	344	if (priv->dpdmai_attr.version.minor > DPDMAI_VER_MINOR) {
	345	+ err = -EINVAL;
344	346	dev_err(dev, "DPDMAI minor version mismatch\n"
345	347	"Found %u.%u, supported version is %u.%u\n",
346	348	priv->dpdmai_attr.version.major,
@@ -475,6 +477,7 @@ static int __cold dpaa2_qdma_dpio_setup(struct dpaa2_qdma_priv *priv) · 128 unmodified lines
475	477	ppriv->store =
476	478	dpaa2_io_store_create(DPAA2_QDMA_STORE_SIZE, dev);
477	479	if (!ppriv->store) {
	480	+ err = -ENOMEM;
478	481	dev_err(dev, "dpaa2_io_store_create() failed\n");
479	482	goto err_store;
480	483	}

drivers/dma/idxd/cdev.c+1−0

@@ -110,6 +110,7 @@ static int idxd_cdev_open(struct inode inode, struct file filp) · 109 unmodified lines
110	110	pasid = iommu_sva_get_pasid(sva);
111	111	if (pasid == IOMMU_PASID_INVALID) {
112	112	iommu_sva_unbind_device(sva);
	113	+ rc = -EINVAL;
113	114	goto failed;
114	115	}
115	116

drivers/dma/idxd/init.c+62−5

@@ -168,6 +168,32 @@ static int idxd_setup_interrupts(struct idxd_device *idxd) · 167 unmodified lines
168	168	return rc;
169	169	}
170	170
	171	+static void idxd_cleanup_interrupts(struct idxd_device *idxd)
	172	+{
	173	+ struct pci_dev *pdev = idxd->pdev;
	174	+ struct idxd_irq_entry *irq_entry;
	175	+ int i, msixcnt;
	176	+
	177	+ msixcnt = pci_msix_vec_count(pdev);
	178	+ if (msixcnt <= 0)
	179	+ return;
	180	+
	181	+ irq_entry = &idxd->irq_entries[0];
	182	+ free_irq(irq_entry->vector, irq_entry);
	183	+
	184	+ for (i = 1; i < msixcnt; i++) {
	185	+
	186	+ irq_entry = &idxd->irq_entries[i];
	187	+ if (idxd->hw.cmd_cap & BIT(IDXD_CMD_RELEASE_INT_HANDLE))
	188	+ idxd_device_release_int_handle(idxd, idxd->int_handles[i],
	189	+ IDXD_IRQ_MSIX);
	190	+ free_irq(irq_entry->vector, irq_entry);
	191	+ }
	192	+
	193	+ idxd_mask_error_interrupts(idxd);
	194	+ pci_free_irq_vectors(pdev);
	195	+}
	196	+
171	197	static int idxd_setup_wqs(struct idxd_device *idxd)
172	198	{
173	199	struct device *dev = &idxd->pdev->dev;
@@ -242,6 +268,7 @@ static int idxd_setup_engines(struct idxd_device *idxd) · 68 unmodified lines
242	268	engine->idxd = idxd;
243	269	device_initialize(&engine->conf_dev);
244	270	engine->conf_dev.parent = &idxd->conf_dev;
	271	+ engine->conf_dev.bus = &dsa_bus_type;
245	272	engine->conf_dev.type = &idxd_engine_device_type;
246	273	rc = dev_set_name(&engine->conf_dev, "engine%d.%d", idxd->id, engine->id);
247	274	if (rc < 0) {
@@ -303,6 +330,19 @@ static int idxd_setup_groups(struct idxd_device *idxd) · 55 unmodified lines
303	330	return rc;
304	331	}
305	332
	333	+static void idxd_cleanup_internals(struct idxd_device *idxd)
	334	+{
	335	+ int i;
	336	+
	337	+ for (i = 0; i < idxd->max_groups; i++)
	338	+ put_device(&idxd->groups[i]->conf_dev);
	339	+ for (i = 0; i < idxd->max_engines; i++)
	340	+ put_device(&idxd->engines[i]->conf_dev);
	341	+ for (i = 0; i < idxd->max_wqs; i++)
	342	+ put_device(&idxd->wqs[i]->conf_dev);
	343	+ destroy_workqueue(idxd->wq);
	344	+}
	345	+
306	346	static int idxd_setup_internals(struct idxd_device *idxd)
307	347	{
308	348	struct device *dev = &idxd->pdev->dev;
@@ -531,12 +571,12 @@ static int idxd_probe(struct idxd_device *idxd) · 222 unmodified lines
531	571	dev_dbg(dev, "Loading RO device config\n");
532	572	rc = idxd_device_load_config(idxd);
533	573	if (rc < 0)
534		− goto err;
	574	+ goto err_config;
535	575	}
536	576
537	577	rc = idxd_setup_interrupts(idxd);
538	578	if (rc)
539		− goto err;
	579	+ goto err_config;
540	580
541	581	dev_dbg(dev, "IDXD interrupt setup complete.\n");
542	582
@@ -549,6 +589,8 @@ static int idxd_probe(struct idxd_device *idxd) · 6 unmodified lines
549	589	dev_dbg(dev, "IDXD device %d probed successfully\n", idxd->id);
550	590	return 0;
551	591
	592	+ err_config:
	593	+ idxd_cleanup_internals(idxd);
552	594	err:
553	595	if (device_pasid_enabled(idxd))
554	596	idxd_disable_system_pasid(idxd);
@@ -556,6 +598,18 @@ static int idxd_probe(struct idxd_device *idxd) · 1 unmodified line
556	598	return rc;
557	599	}
558	600
	601	+static void idxd_cleanup(struct idxd_device *idxd)
	602	+{
	603	+ struct device *dev = &idxd->pdev->dev;
	604	+
	605	+ perfmon_pmu_remove(idxd);
	606	+ idxd_cleanup_interrupts(idxd);
	607	+ idxd_cleanup_internals(idxd);
	608	+ if (device_pasid_enabled(idxd))
	609	+ idxd_disable_system_pasid(idxd);
	610	+ iommu_dev_disable_feature(dev, IOMMU_DEV_FEAT_SVA);
	611	+}
	612	+
559	613	static int idxd_pci_probe(struct pci_dev pdev, const struct pci_device_id id)
560	614	{
561	615	struct device *dev = &pdev->dev;
@@ -608,7 +662,7 @@ static int idxd_pci_probe(struct pci_dev pdev, const struct pci_device_id id) · 46 unmodified lines
608	662	rc = idxd_register_devices(idxd);
609	663	if (rc) {
610	664	dev_err(dev, "IDXD sysfs setup failed\n");
611		− goto err;
	665	+ goto err_dev_register;
612	666	}
613	667
614	668	idxd->state = IDXD_DEV_CONF_READY;
@@ -618,6 +672,8 @@ static int idxd_pci_probe(struct pci_dev pdev, const struct pci_device_id id) · 3 unmodified lines
618	672
619	673	return 0;
620	674
	675	+ err_dev_register:
	676	+ idxd_cleanup(idxd);
621	677	err:
622	678	pci_iounmap(pdev, idxd->reg_base);
623	679	err_iomap:
@@ -745,12 +801,12 @@ static int __init idxd_init_module(void) · 121 unmodified lines
745	801	* If the CPU does not support MOVDIR64B or ENQCMDS, there's no point in
746	802	* enumerating the device. We can not utilize it.
747	803	*/
748		− if (!boot_cpu_has(X86_FEATURE_MOVDIR64B)) {
	804	+ if (!cpu_feature_enabled(X86_FEATURE_MOVDIR64B)) {
749	805	pr_warn("idxd driver failed to load without MOVDIR64B.\n");
750	806	return -ENODEV;
751	807	}
752	808
753		− if (!boot_cpu_has(X86_FEATURE_ENQCMD))
	809	+ if (!cpu_feature_enabled(X86_FEATURE_ENQCMD))
754	810	pr_warn("Platform does not have ENQCMD(S) support.\n");
755	811	else
756	812	support_enqcmd = true;
@@ -787,6 +843,7 @@ module_init(idxd_init_module); · 30 unmodified lines
787	843
788	844	static void __exit idxd_exit_module(void)
789	845	{
	846	+ idxd_unregister_driver();
790	847	pci_unregister_driver(&idxd_pci_driver);
791	848	idxd_cdev_remove();
792	849	idxd_unregister_bus_type();

drivers/dma/ipu/ipu_irq.c+1−1

@@ -230,7 +230,7 @@ out: · 229 unmodified lines
230	230	}
231	231
232	232	/**
233		− * ipu_irq_map() - map an IPU interrupt source to an IRQ number
	233	+ * ipu_irq_unmap() - unmap an IPU interrupt source
234	234	* @source: interrupt source bit position (see ipu_irq_map())
235	235	* @return: 0 or negative error code
236	236	*/

drivers/dma/mediatek/mtk-uart-apdma.c+14−13

@@ -131,10 +131,7 @@ static unsigned int mtk_uart_apdma_read(struct mtk_chan *c, unsigned int reg) · 130 unmodified lines
131	131
132	132	static void mtk_uart_apdma_desc_free(struct virt_dma_desc *vd)
133	133	{
134		− struct dma_chan *chan = vd->tx.chan;
135		− struct mtk_chan *c = to_mtk_uart_apdma_chan(chan);
136		−
137		− kfree(c->desc);
	134	+ kfree(container_of(vd, struct mtk_uart_apdma_desc, vd));
138	135	}
139	136
140	137	static void mtk_uart_apdma_start_tx(struct mtk_chan *c)
@@ -207,14 +204,9 @@ static void mtk_uart_apdma_start_rx(struct mtk_chan *c) · 66 unmodified lines
207	204
208	205	static void mtk_uart_apdma_tx_handler(struct mtk_chan *c)
209	206	{
210		− struct mtk_uart_apdma_desc *d = c->desc;
211		−
212	207	mtk_uart_apdma_write(c, VFF_INT_FLAG, VFF_TX_INT_CLR_B);
213	208	mtk_uart_apdma_write(c, VFF_INT_EN, VFF_INT_EN_CLR_B);
214	209	mtk_uart_apdma_write(c, VFF_EN, VFF_EN_CLR_B);
215		−
216		− list_del(&d->vd.node);
217		− vchan_cookie_complete(&d->vd);
218	210	}
219	211
220	212	static void mtk_uart_apdma_rx_handler(struct mtk_chan *c)
@@ -245,9 +237,17 @@ static void mtk_uart_apdma_rx_handler(struct mtk_chan *c) · 24 unmodified lines
245	237
246	238	c->rx_status = d->avail_len - cnt;
247	239	mtk_uart_apdma_write(c, VFF_RPT, wg);
	240	+}
248	241
249		− list_del(&d->vd.node);
250		− vchan_cookie_complete(&d->vd);
	242	+static void mtk_uart_apdma_chan_complete_handler(struct mtk_chan *c)
	243	+{
	244	+ struct mtk_uart_apdma_desc *d = c->desc;
	245	+
	246	+ if (d) {
	247	+ list_del(&d->vd.node);
	248	+ vchan_cookie_complete(&d->vd);
	249	+ c->desc = NULL;
	250	+ }
251	251	}
252	252
253	253	static irqreturn_t mtk_uart_apdma_irq_handler(int irq, void *dev_id)
@@ -261,6 +261,7 @@ static irqreturn_t mtk_uart_apdma_irq_handler(int irq, void *dev_id) · 7 unmodified lines
261	261	mtk_uart_apdma_rx_handler(c);
262	262	else if (c->dir == DMA_MEM_TO_DEV)
263	263	mtk_uart_apdma_tx_handler(c);
	264	+ mtk_uart_apdma_chan_complete_handler(c);
264	265	spin_unlock_irqrestore(&c->vc.lock, flags);
265	266
266	267	return IRQ_HANDLED;
@@ -348,7 +349,7 @@ static struct dma_async_tx_descriptor *mtk_uart_apdma_prep_slave_sg · 81 unmodified lines
348	349	return NULL;
349	350
350	351	/* Now allocate and setup the descriptor */
351		− d = kzalloc(sizeof(*d), GFP_ATOMIC);
	352	+ d = kzalloc(sizeof(*d), GFP_NOWAIT);
352	353	if (!d)
353	354	return NULL;
354	355
@@ -366,7 +367,7 @@ static void mtk_uart_apdma_issue_pending(struct dma_chan *chan) · 11 unmodified lines
366	367	unsigned long flags;
367	368
368	369	spin_lock_irqsave(&c->vc.lock, flags);
369		− if (vchan_issue_pending(&c->vc)) {
	370	+ if (vchan_issue_pending(&c->vc) && !c->desc) {
370	371	vd = vchan_next_desc(&c->vc);
371	372	c->desc = to_mtk_uart_apdma_desc(&vd->tx);
372	373

drivers/dma/pl330.c+4−2

@@ -2694,13 +2694,15 @@ static struct dma_async_tx_descriptor *pl330_prep_dma_cyclic( · 2693 unmodified lines
2694	2694	for (i = 0; i < len / period_len; i++) {
2695	2695	desc = pl330_get_desc(pch);
2696	2696	if (!desc) {
	2697	+ unsigned long iflags;
	2698	+
2697	2699	dev_err(pch->dmac->ddma.dev, "%s:%d Unable to fetch desc\n",
2698	2700	__func__, __LINE__);
2699	2701
2700	2702	if (!first)
2701	2703	return NULL;
2702	2704
2703		− spin_lock_irqsave(&pl330->pool_lock, flags);
	2705	+ spin_lock_irqsave(&pl330->pool_lock, iflags);
2704	2706
2705	2707	while (!list_empty(&first->node)) {
2706	2708	desc = list_entry(first->node.next,
@@ -2710,7 +2712,7 @@ static struct dma_async_tx_descriptor *pl330_prep_dma_cyclic( · 3 unmodified lines
2710	2712
2711	2713	list_move_tail(&first->node, &pl330->desc_pool);
2712	2714
2713		− spin_unlock_irqrestore(&pl330->pool_lock, flags);
	2715	+ spin_unlock_irqrestore(&pl330->pool_lock, iflags);
2714	2716
2715	2717	return NULL;
2716	2718	}

drivers/dma/qcom/Kconfig+1−0

@@ -33,6 +33,7 @@ config QCOM_GPI_DMA · 32 unmodified lines
33	33
34	34	config QCOM_HIDMA_MGMT
35	35	tristate "Qualcomm Technologies HIDMA Management support"
	36	+ depends on HAS_IOMEM
36	37	select DMA_ENGINE
37	38	help
38	39	Enable support for the Qualcomm Technologies HIDMA Management.

drivers/dma/sf-pdma/Kconfig+1−0

@@ -1,5 +1,6 @@
1	1	config SF_PDMA
2	2	tristate "Sifive PDMA controller driver"
	3	+ depends on HAS_IOMEM
3	4	select DMA_ENGINE
4	5	select DMA_VIRTUAL_CHANNELS
5	6	help

drivers/dma/sh/rcar-dmac.c+1−1

@@ -1913,7 +1913,7 @@ static int rcar_dmac_probe(struct platform_device *pdev) · 1912 unmodified lines
1913	1913
1914	1914	/* Enable runtime PM and initialize the device. */
1915	1915	pm_runtime_enable(&pdev->dev);
1916		− ret = pm_runtime_get_sync(&pdev->dev);
	1916	+ ret = pm_runtime_resume_and_get(&pdev->dev);
1917	1917	if (ret < 0) {
1918	1918	dev_err(&pdev->dev, "runtime PM get sync failed (%d)\n", ret);
1919	1919	return ret;

drivers/dma/ste_dma40.c+3−0

@@ -3675,6 +3675,9 @@ static int __init d40_probe(struct platform_device *pdev) · 3674 unmodified lines
3675	3675
3676	3676	kfree(base->lcla_pool.base_unaligned);
3677	3677
	3678	+ if (base->lcpa_base)
	3679	+ iounmap(base->lcpa_base);
	3680	+
3678	3681	if (base->phy_lcpa)
3679	3682	release_mem_region(base->phy_lcpa,
3680	3683	base->lcpa_size);

drivers/dma/stm32-mdma.c+2−2

@@ -1452,7 +1452,7 @@ static int stm32_mdma_alloc_chan_resources(struct dma_chan *c) · 1451 unmodified lines
1452	1452	return -ENOMEM;
1453	1453	}
1454	1454
1455		− ret = pm_runtime_get_sync(dmadev->ddev.dev);
	1455	+ ret = pm_runtime_resume_and_get(dmadev->ddev.dev);
1456	1456	if (ret < 0)
1457	1457	return ret;
1458	1458
@@ -1718,7 +1718,7 @@ static int stm32_mdma_pm_suspend(struct device *dev) · 259 unmodified lines
1718	1718	u32 ccr, id;
1719	1719	int ret;
1720	1720
1721		− ret = pm_runtime_get_sync(dev);
	1721	+ ret = pm_runtime_resume_and_get(dev);
1722	1722	if (ret < 0)
1723	1723	return ret;
1724	1724

drivers/dma/xilinx/xilinx_dpdma.c+28−3

@@ -113,6 +113,7 @@ · 112 unmodified lines
113	113	#define XILINX_DPDMA_CH_VDO 0x020
114	114	#define XILINX_DPDMA_CH_PYLD_SZ 0x024
115	115	#define XILINX_DPDMA_CH_DESC_ID 0x028
	116	+#define XILINX_DPDMA_CH_DESC_ID_MASK GENMASK(15, 0)
116	117
117	118	/* DPDMA descriptor fields */
118	119	#define XILINX_DPDMA_DESC_CONTROL_PREEMBLE 0xa5
@@ -866,7 +867,8 @@ static void xilinx_dpdma_chan_queue_transfer(struct xilinx_dpdma_chan *chan) · 747 unmodified lines
866	867	* will be used, but it should be enough.
867	868	*/
868	869	list_for_each_entry(sw_desc, &desc->descriptors, node)
869		− sw_desc->hw.desc_id = desc->vdesc.tx.cookie;
	870	+ sw_desc->hw.desc_id = desc->vdesc.tx.cookie
	871	+ & XILINX_DPDMA_CH_DESC_ID_MASK;
870	872
871	873	sw_desc = list_first_entry(&desc->descriptors,
872	874	struct xilinx_dpdma_sw_desc, node);
@@ -1086,7 +1088,8 @@ static void xilinx_dpdma_chan_vsync_irq(struct xilinx_dpdma_chan *chan) · 213 unmodified lines
1086	1088	if (!chan->running \|\| !pending)
1087	1089	goto out;
1088	1090
1089		− desc_id = dpdma_read(chan->reg, XILINX_DPDMA_CH_DESC_ID);
	1091	+ desc_id = dpdma_read(chan->reg, XILINX_DPDMA_CH_DESC_ID)
	1092	+ & XILINX_DPDMA_CH_DESC_ID_MASK;
1090	1093
1091	1094	/* If the retrigger raced with vsync, retry at the next frame. */
1092	1095	sw_desc = list_first_entry(&pending->descriptors,
@@ -1459,7 +1462,7 @@ static void xilinx_dpdma_enable_irq(struct xilinx_dpdma_device *xdev) · 366 unmodified lines
1459	1462	*/
1460	1463	static void xilinx_dpdma_disable_irq(struct xilinx_dpdma_device *xdev)
1461	1464	{
1462		− dpdma_write(xdev->reg, XILINX_DPDMA_IDS, XILINX_DPDMA_INTR_ERR_ALL);
	1465	+ dpdma_write(xdev->reg, XILINX_DPDMA_IDS, XILINX_DPDMA_INTR_ALL);
1463	1466	dpdma_write(xdev->reg, XILINX_DPDMA_EIDS, XILINX_DPDMA_EINTR_ALL);
1464	1467	}
1465	1468
@@ -1596,6 +1599,26 @@ static struct dma_chan of_dma_xilinx_xlate(struct of_phandle_args dma_spec, · 130 unmodified lines
1596	1599	return dma_get_slave_channel(&xdev->chan[chan_id]->vchan.chan);
1597	1600	}
1598	1601
	1602	+static void dpdma_hw_init(struct xilinx_dpdma_device *xdev)
	1603	+{
	1604	+ unsigned int i;
	1605	+ void __iomem *reg;
	1606	+
	1607	+ /* Disable all interrupts */
	1608	+ xilinx_dpdma_disable_irq(xdev);
	1609	+
	1610	+ /* Stop all channels */
	1611	+ for (i = 0; i < ARRAY_SIZE(xdev->chan); i++) {
	1612	+ reg = xdev->reg + XILINX_DPDMA_CH_BASE
	1613	+ + XILINX_DPDMA_CH_OFFSET * i;
	1614	+ dpdma_clr(reg, XILINX_DPDMA_CH_CNTL, XILINX_DPDMA_CH_CNTL_ENABLE);
	1615	+ }
	1616	+
	1617	+ /* Clear the interrupt status registers */
	1618	+ dpdma_write(xdev->reg, XILINX_DPDMA_ISR, XILINX_DPDMA_INTR_ALL);
	1619	+ dpdma_write(xdev->reg, XILINX_DPDMA_EISR, XILINX_DPDMA_EINTR_ALL);
	1620	+}
	1621	+
1599	1622	static int xilinx_dpdma_probe(struct platform_device *pdev)
1600	1623	{
1601	1624	struct xilinx_dpdma_device *xdev;
@@ -1622,6 +1645,8 @@ static int xilinx_dpdma_probe(struct platform_device *pdev) · 20 unmodified lines
1622	1645	if (IS_ERR(xdev->reg))
1623	1646	return PTR_ERR(xdev->reg);
1624	1647
	1648	+ dpdma_hw_init(xdev);
	1649	+
1625	1650	xdev->irq = platform_get_irq(pdev, 0);
1626	1651	if (xdev->irq < 0) {
1627	1652	dev_err(xdev->dev, "failed to get platform irq\n");

drivers/dma/xilinx/zynqmp_dma.c+1−1

@@ -468,7 +468,7 @@ static int zynqmp_dma_alloc_chan_resources(struct dma_chan *dchan) · 467 unmodified lines
468	468	struct zynqmp_dma_desc_sw *desc;
469	469	int i, ret;
470	470
471		− ret = pm_runtime_get_sync(chan->dev);
	471	+ ret = pm_runtime_resume_and_get(chan->dev);
472	472	if (ret < 0)
473	473	return ret;
474	474

drivers/firmware/efi/cper.c+1−3

@@ -276,8 +276,7 @@ static int cper_dimm_err_location(struct cper_mem_err_compact mem, char msg) · 275 unmodified lines
276	276	if (!msg \|\| !(mem->validation_bits & CPER_MEM_VALID_MODULE_HANDLE))
277	277	return 0;
278	278
279		− n = 0;
280		− len = CPER_REC_LEN - 1;
	279	+ len = CPER_REC_LEN;
281	280	dmi_memdev_name(mem->mem_dev_handle, &bank, &device);
282	281	if (bank && device)
283	282	n = snprintf(msg, len, "DIMM location: %s %s ", bank, device);
@@ -286,7 +285,6 @@ static int cper_dimm_err_location(struct cper_mem_err_compact mem, char msg) · 2 unmodified lines
286	285	"DIMM location: not present. DMI handle: 0x%.4x ",
287	286	mem->mem_dev_handle);
288	287
289		− msg[n] = '\0';
290	288	return n;
291	289	}
292	290

drivers/firmware/efi/fdtparams.c+3−0

@@ -98,6 +98,9 @@ u64 __init efi_get_fdt_params(struct efi_memory_map_data *mm) · 97 unmodified lines
98	98	BUILD_BUG_ON(ARRAY_SIZE(target) != ARRAY_SIZE(name));
99	99	BUILD_BUG_ON(ARRAY_SIZE(target) != ARRAY_SIZE(dt_params[0].params));
100	100
	101	+ if (!fdt)
	102	+ return 0;
	103	+
101	104	for (i = 0; i < ARRAY_SIZE(dt_params); i++) {
102	105	node = fdt_path_offset(fdt, dt_params[i].path);
103	106	if (node < 0)

drivers/firmware/efi/libstub/file.c+1−1

@@ -103,7 +103,7 @@ static int find_file_option(const efi_char16_t *cmdline, int cmdline_len, · 102 unmodified lines
103	103	return 0;
104	104
105	105	/* Skip any leading slashes */
106		− while (cmdline[i] == L'/' \|\| cmdline[i] == L'\\')
	106	+ while (i < cmdline_len && (cmdline[i] == L'/' \|\| cmdline[i] == L'\\'))
107	107	i++;
108	108
109	109	while (--result_len > 0 && i < cmdline_len) {

drivers/firmware/efi/memattr.c+0−5

@@ -67,11 +67,6 @@ static bool entry_is_valid(const efi_memory_desc_t in, efi_memory_desc_t out) · 66 unmodified lines
67	67	return false;
68	68	}
69	69
70		− if (!(in->attribute & (EFI_MEMORY_RO \| EFI_MEMORY_XP))) {
71		− pr_warn("Entry attributes invalid: RO and XP bits both cleared\n");
72		− return false;
73		− }
74		−
75	70	if (PAGE_SIZE > EFI_PAGE_SIZE &&
76	71	(!PAGE_ALIGNED(in->phys_addr) \|\|
77	72	!PAGE_ALIGNED(in->num_pages << EFI_PAGE_SHIFT))) {

drivers/gpio/gpio-wcd934x.c+1−1

@@ -7,7 +7,7 @@ · 6 unmodified lines
7	7	#include <linux/slab.h>
8	8	#include <linux/of_device.h>
9	9
10		−#define WCD_PIN_MASK(p) BIT(p - 1)
	10	+#define WCD_PIN_MASK(p) BIT(p)
11	11	#define WCD_REG_DIR_CTL_OFFSET 0x42
12	12	#define WCD_REG_VAL_CTL_OFFSET 0x43
13	13	#define WCD934X_NPINS 5

drivers/gpu/drm/drm_auth.c+2−1

@@ -315,9 +315,10 @@ int drm_master_open(struct drm_file *file_priv) · 314 unmodified lines
315	315	void drm_master_release(struct drm_file *file_priv)
316	316	{
317	317	struct drm_device *dev = file_priv->minor->dev;
318		− struct drm_master *master = file_priv->master;
	318	+ struct drm_master *master;
319	319
320	320	mutex_lock(&dev->master_mutex);
	321	+ master = file_priv->master;
321	322	if (file_priv->magic)
322	323	idr_remove(&file_priv->master->magic_map, file_priv->magic);
323	324

drivers/gpu/drm/drm_ioctl.c+5−4

@@ -117,17 +117,18 @@ int drm_getunique(struct drm_device dev, void data, · 116 unmodified lines
117	117	struct drm_file *file_priv)
118	118	{
119	119	struct drm_unique *u = data;
120		− struct drm_master *master = file_priv->master;
	120	+ struct drm_master *master;
121	121
122		− mutex_lock(&master->dev->master_mutex);
	122	+ mutex_lock(&dev->master_mutex);
	123	+ master = file_priv->master;
123	124	if (u->unique_len >= master->unique_len) {
124	125	if (copy_to_user(u->unique, master->unique, master->unique_len)) {
125		− mutex_unlock(&master->dev->master_mutex);
	126	+ mutex_unlock(&dev->master_mutex);
126	127	return -EFAULT;
127	128	}
128	129	}
129	130	u->unique_len = master->unique_len;
130		− mutex_unlock(&master->dev->master_mutex);
	131	+ mutex_unlock(&dev->master_mutex);
131	132
132	133	return 0;
133	134	}

drivers/gpu/drm/mcde/mcde_dsi.c+1−1

@@ -577,7 +577,7 @@ static void mcde_dsi_setup_video_mode(struct mcde_dsi *d, · 576 unmodified lines
577	577	* porches and sync.
578	578	*/
579	579	/* (ps/s) / (pixels/s) = ps/pixels */
580		− pclk = DIV_ROUND_UP_ULL(1000000000000, mode->clock);
	580	+ pclk = DIV_ROUND_UP_ULL(1000000000000, (mode->clock * 1000));
581	581	dev_dbg(d->dev, "picoseconds between two pixels: %llu\n",
582	582	pclk);
583	583

drivers/gpu/drm/meson/meson_drv.c+5−4

@@ -474,11 +474,12 @@ static int meson_probe_remote(struct platform_device *pdev, · 473 unmodified lines
474	474	static void meson_drv_shutdown(struct platform_device *pdev)
475	475	{
476	476	struct meson_drm *priv = dev_get_drvdata(&pdev->dev);
477		− struct drm_device *drm = priv->drm;
478	477
479		− DRM_DEBUG_DRIVER("\n");
480		− drm_kms_helper_poll_fini(drm);
481		− drm_atomic_helper_shutdown(drm);
	478	+ if (!priv)
	479	+ return;
	480	+
	481	+ drm_kms_helper_poll_fini(priv->drm);
	482	+ drm_atomic_helper_shutdown(priv->drm);
482	483	}
483	484
484	485	static int meson_drv_probe(struct platform_device *pdev)

drivers/gpu/drm/msm/adreno/a6xx_gpu.c+114−41

@@ -157,7 +157,7 @@ static void a6xx_submit(struct msm_gpu gpu, struct msm_gem_submit submit) · 156 unmodified lines
157	157	* GPU registers so we need to add 0x1a800 to the register value on A630
158	158	* to get the right value from PM4.
159	159	*/
160		− get_stats_counter(ring, REG_A6XX_GMU_ALWAYS_ON_COUNTER_L + 0x1a800,
	160	+ get_stats_counter(ring, REG_A6XX_CP_ALWAYS_ON_COUNTER_LO,
161	161	rbmemptr_stats(ring, index, alwayson_start));
162	162
163	163	/* Invalidate CCU depth and color */
@@ -187,7 +187,7 @@ static void a6xx_submit(struct msm_gpu gpu, struct msm_gem_submit submit) · 23 unmodified lines
187	187
188	188	get_stats_counter(ring, REG_A6XX_RBBM_PERFCTR_CP_0_LO,
189	189	rbmemptr_stats(ring, index, cpcycles_end));
190		− get_stats_counter(ring, REG_A6XX_GMU_ALWAYS_ON_COUNTER_L + 0x1a800,
	190	+ get_stats_counter(ring, REG_A6XX_CP_ALWAYS_ON_COUNTER_LO,
191	191	rbmemptr_stats(ring, index, alwayson_end));
192	192
193	193	/* Write the fence to the scratch register */
@@ -206,8 +206,8 @@ static void a6xx_submit(struct msm_gpu gpu, struct msm_gem_submit submit) · 12 unmodified lines
206	206	OUT_RING(ring, submit->seqno);
207	207
208	208	trace_msm_gpu_submit_flush(submit,
209		− gmu_read64(&a6xx_gpu->gmu, REG_A6XX_GMU_ALWAYS_ON_COUNTER_L,
210		− REG_A6XX_GMU_ALWAYS_ON_COUNTER_H));
	209	+ gpu_read64(gpu, REG_A6XX_CP_ALWAYS_ON_COUNTER_LO,
	210	+ REG_A6XX_CP_ALWAYS_ON_COUNTER_HI));
211	211
212	212	a6xx_flush(gpu, ring);
213	213	}
@@ -462,6 +462,113 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state) · 248 unmodified lines
462	462	gpu_write(gpu, REG_A6XX_RBBM_CLOCK_CNTL, state ? clock_cntl_on : 0);
463	463	}
464	464
	465	+/* For a615, a616, a618, A619, a630, a640 and a680 */
	466	+static const u32 a6xx_protect[] = {
	467	+ A6XX_PROTECT_RDONLY(0x00000, 0x04ff),
	468	+ A6XX_PROTECT_RDONLY(0x00501, 0x0005),
	469	+ A6XX_PROTECT_RDONLY(0x0050b, 0x02f4),
	470	+ A6XX_PROTECT_NORDWR(0x0050e, 0x0000),
	471	+ A6XX_PROTECT_NORDWR(0x00510, 0x0000),
	472	+ A6XX_PROTECT_NORDWR(0x00534, 0x0000),
	473	+ A6XX_PROTECT_NORDWR(0x00800, 0x0082),
	474	+ A6XX_PROTECT_NORDWR(0x008a0, 0x0008),
	475	+ A6XX_PROTECT_NORDWR(0x008ab, 0x0024),
	476	+ A6XX_PROTECT_RDONLY(0x008de, 0x00ae),
	477	+ A6XX_PROTECT_NORDWR(0x00900, 0x004d),
	478	+ A6XX_PROTECT_NORDWR(0x0098d, 0x0272),
	479	+ A6XX_PROTECT_NORDWR(0x00e00, 0x0001),
	480	+ A6XX_PROTECT_NORDWR(0x00e03, 0x000c),
	481	+ A6XX_PROTECT_NORDWR(0x03c00, 0x00c3),
	482	+ A6XX_PROTECT_RDONLY(0x03cc4, 0x1fff),
	483	+ A6XX_PROTECT_NORDWR(0x08630, 0x01cf),
	484	+ A6XX_PROTECT_NORDWR(0x08e00, 0x0000),
	485	+ A6XX_PROTECT_NORDWR(0x08e08, 0x0000),
	486	+ A6XX_PROTECT_NORDWR(0x08e50, 0x001f),
	487	+ A6XX_PROTECT_NORDWR(0x09624, 0x01db),
	488	+ A6XX_PROTECT_NORDWR(0x09e70, 0x0001),
	489	+ A6XX_PROTECT_NORDWR(0x09e78, 0x0187),
	490	+ A6XX_PROTECT_NORDWR(0x0a630, 0x01cf),
	491	+ A6XX_PROTECT_NORDWR(0x0ae02, 0x0000),
	492	+ A6XX_PROTECT_NORDWR(0x0ae50, 0x032f),
	493	+ A6XX_PROTECT_NORDWR(0x0b604, 0x0000),
	494	+ A6XX_PROTECT_NORDWR(0x0be02, 0x0001),
	495	+ A6XX_PROTECT_NORDWR(0x0be20, 0x17df),
	496	+ A6XX_PROTECT_NORDWR(0x0f000, 0x0bff),
	497	+ A6XX_PROTECT_RDONLY(0x0fc00, 0x1fff),
	498	+ A6XX_PROTECT_NORDWR(0x11c00, 0x0000), /* note: infinite range */
	499	+};
	500	+
	501	+/* These are for a620 and a650 */
	502	+static const u32 a650_protect[] = {
	503	+ A6XX_PROTECT_RDONLY(0x00000, 0x04ff),
	504	+ A6XX_PROTECT_RDONLY(0x00501, 0x0005),
	505	+ A6XX_PROTECT_RDONLY(0x0050b, 0x02f4),
	506	+ A6XX_PROTECT_NORDWR(0x0050e, 0x0000),
	507	+ A6XX_PROTECT_NORDWR(0x00510, 0x0000),
	508	+ A6XX_PROTECT_NORDWR(0x00534, 0x0000),
	509	+ A6XX_PROTECT_NORDWR(0x00800, 0x0082),
	510	+ A6XX_PROTECT_NORDWR(0x008a0, 0x0008),
	511	+ A6XX_PROTECT_NORDWR(0x008ab, 0x0024),
	512	+ A6XX_PROTECT_RDONLY(0x008de, 0x00ae),
	513	+ A6XX_PROTECT_NORDWR(0x00900, 0x004d),
	514	+ A6XX_PROTECT_NORDWR(0x0098d, 0x0272),
	515	+ A6XX_PROTECT_NORDWR(0x00e00, 0x0001),
	516	+ A6XX_PROTECT_NORDWR(0x00e03, 0x000c),
	517	+ A6XX_PROTECT_NORDWR(0x03c00, 0x00c3),
	518	+ A6XX_PROTECT_RDONLY(0x03cc4, 0x1fff),
	519	+ A6XX_PROTECT_NORDWR(0x08630, 0x01cf),
	520	+ A6XX_PROTECT_NORDWR(0x08e00, 0x0000),
	521	+ A6XX_PROTECT_NORDWR(0x08e08, 0x0000),
	522	+ A6XX_PROTECT_NORDWR(0x08e50, 0x001f),
	523	+ A6XX_PROTECT_NORDWR(0x08e80, 0x027f),
	524	+ A6XX_PROTECT_NORDWR(0x09624, 0x01db),
	525	+ A6XX_PROTECT_NORDWR(0x09e60, 0x0011),
	526	+ A6XX_PROTECT_NORDWR(0x09e78, 0x0187),
	527	+ A6XX_PROTECT_NORDWR(0x0a630, 0x01cf),
	528	+ A6XX_PROTECT_NORDWR(0x0ae02, 0x0000),
	529	+ A6XX_PROTECT_NORDWR(0x0ae50, 0x032f),
	530	+ A6XX_PROTECT_NORDWR(0x0b604, 0x0000),
	531	+ A6XX_PROTECT_NORDWR(0x0b608, 0x0007),
	532	+ A6XX_PROTECT_NORDWR(0x0be02, 0x0001),
	533	+ A6XX_PROTECT_NORDWR(0x0be20, 0x17df),
	534	+ A6XX_PROTECT_NORDWR(0x0f000, 0x0bff),
	535	+ A6XX_PROTECT_RDONLY(0x0fc00, 0x1fff),
	536	+ A6XX_PROTECT_NORDWR(0x18400, 0x1fff),
	537	+ A6XX_PROTECT_NORDWR(0x1a800, 0x1fff),
	538	+ A6XX_PROTECT_NORDWR(0x1f400, 0x0443),
	539	+ A6XX_PROTECT_RDONLY(0x1f844, 0x007b),
	540	+ A6XX_PROTECT_NORDWR(0x1f887, 0x001b),
	541	+ A6XX_PROTECT_NORDWR(0x1f8c0, 0x0000), /* note: infinite range */
	542	+};
	543	+
	544	+static void a6xx_set_cp_protect(struct msm_gpu *gpu)
	545	+{
	546	+ struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
	547	+ const u32 *regs = a6xx_protect;
	548	+ unsigned i, count = ARRAY_SIZE(a6xx_protect), count_max = 32;
	549	+
	550	+ BUILD_BUG_ON(ARRAY_SIZE(a6xx_protect) > 32);
	551	+ BUILD_BUG_ON(ARRAY_SIZE(a650_protect) > 48);
	552	+
	553	+ if (adreno_is_a650(adreno_gpu)) {
	554	+ regs = a650_protect;
	555	+ count = ARRAY_SIZE(a650_protect);
	556	+ count_max = 48;
	557	+ }
	558	+
	559	+ /*
	560	+ * Enable access protection to privileged registers, fault on an access
	561	+ * protect violation and select the last span to protect from the start
	562	+ * address all the way to the end of the register address space
	563	+ */
	564	+ gpu_write(gpu, REG_A6XX_CP_PROTECT_CNTL, BIT(0) \| BIT(1) \| BIT(3));
	565	+
	566	+ for (i = 0; i < count - 1; i++)
	567	+ gpu_write(gpu, REG_A6XX_CP_PROTECT(i), regs[i]);
	568	+ /* last CP_PROTECT to have "infinite" length on the last entry */
	569	+ gpu_write(gpu, REG_A6XX_CP_PROTECT(count_max - 1), regs[i]);
	570	+}
	571	+
465	572	static void a6xx_set_ubwc_config(struct msm_gpu *gpu)
466	573	{
467	574	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
@@ -489,7 +596,7 @@ static void a6xx_set_ubwc_config(struct msm_gpu *gpu) · 21 unmodified lines
489	596	rgb565_predicator << 11 \| amsbc << 4 \| lower_bit << 1);
490	597	gpu_write(gpu, REG_A6XX_TPL1_NC_MODE_CNTL, lower_bit << 1);
491	598	gpu_write(gpu, REG_A6XX_SP_NC_MODE_CNTL,
492		− uavflagprd_inv >> 4 \| lower_bit << 1);
	599	+ uavflagprd_inv << 4 \| lower_bit << 1);
493	600	gpu_write(gpu, REG_A6XX_UCHE_MODE_CNTL, lower_bit << 21);
494	601	}
495	602
@@ -776,41 +883,7 @@ static int a6xx_hw_init(struct msm_gpu *gpu) · 280 unmodified lines
776	883	}
777	884
778	885	/* Protect registers from the CP */
779		− gpu_write(gpu, REG_A6XX_CP_PROTECT_CNTL, 0x00000003);
780		−
781		− gpu_write(gpu, REG_A6XX_CP_PROTECT(0),
782		− A6XX_PROTECT_RDONLY(0x600, 0x51));
783		− gpu_write(gpu, REG_A6XX_CP_PROTECT(1), A6XX_PROTECT_RW(0xae50, 0x2));
784		− gpu_write(gpu, REG_A6XX_CP_PROTECT(2), A6XX_PROTECT_RW(0x9624, 0x13));
785		− gpu_write(gpu, REG_A6XX_CP_PROTECT(3), A6XX_PROTECT_RW(0x8630, 0x8));
786		− gpu_write(gpu, REG_A6XX_CP_PROTECT(4), A6XX_PROTECT_RW(0x9e70, 0x1));
787		− gpu_write(gpu, REG_A6XX_CP_PROTECT(5), A6XX_PROTECT_RW(0x9e78, 0x187));
788		− gpu_write(gpu, REG_A6XX_CP_PROTECT(6), A6XX_PROTECT_RW(0xf000, 0x810));
789		− gpu_write(gpu, REG_A6XX_CP_PROTECT(7),
790		− A6XX_PROTECT_RDONLY(0xfc00, 0x3));
791		− gpu_write(gpu, REG_A6XX_CP_PROTECT(8), A6XX_PROTECT_RW(0x50e, 0x0));
792		− gpu_write(gpu, REG_A6XX_CP_PROTECT(9), A6XX_PROTECT_RDONLY(0x50f, 0x0));
793		− gpu_write(gpu, REG_A6XX_CP_PROTECT(10), A6XX_PROTECT_RW(0x510, 0x0));
794		− gpu_write(gpu, REG_A6XX_CP_PROTECT(11),
795		− A6XX_PROTECT_RDONLY(0x0, 0x4f9));
796		− gpu_write(gpu, REG_A6XX_CP_PROTECT(12),
797		− A6XX_PROTECT_RDONLY(0x501, 0xa));
798		− gpu_write(gpu, REG_A6XX_CP_PROTECT(13),
799		− A6XX_PROTECT_RDONLY(0x511, 0x44));
800		− gpu_write(gpu, REG_A6XX_CP_PROTECT(14), A6XX_PROTECT_RW(0xe00, 0xe));
801		− gpu_write(gpu, REG_A6XX_CP_PROTECT(15), A6XX_PROTECT_RW(0x8e00, 0x0));
802		− gpu_write(gpu, REG_A6XX_CP_PROTECT(16), A6XX_PROTECT_RW(0x8e50, 0xf));
803		− gpu_write(gpu, REG_A6XX_CP_PROTECT(17), A6XX_PROTECT_RW(0xbe02, 0x0));
804		− gpu_write(gpu, REG_A6XX_CP_PROTECT(18),
805		− A6XX_PROTECT_RW(0xbe20, 0x11f3));
806		− gpu_write(gpu, REG_A6XX_CP_PROTECT(19), A6XX_PROTECT_RW(0x800, 0x82));
807		− gpu_write(gpu, REG_A6XX_CP_PROTECT(20), A6XX_PROTECT_RW(0x8a0, 0x8));
808		− gpu_write(gpu, REG_A6XX_CP_PROTECT(21), A6XX_PROTECT_RW(0x8ab, 0x19));
809		− gpu_write(gpu, REG_A6XX_CP_PROTECT(22), A6XX_PROTECT_RW(0x900, 0x4d));
810		− gpu_write(gpu, REG_A6XX_CP_PROTECT(23), A6XX_PROTECT_RW(0x98d, 0x76));
811		− gpu_write(gpu, REG_A6XX_CP_PROTECT(24),
812		− A6XX_PROTECT_RDONLY(0x980, 0x4));
813		− gpu_write(gpu, REG_A6XX_CP_PROTECT(25), A6XX_PROTECT_RW(0xa630, 0x0));
	886	+ a6xx_set_cp_protect(gpu);
814	887
815	888	/* Enable expanded apriv for targets that support it */
816	889	if (gpu->hw_apriv) {
@@ -1211,7 +1284,7 @@ static int a6xx_pm_suspend(struct msm_gpu *gpu) · 394 unmodified lines
1211	1284	if (ret)
1212	1285	return ret;
1213	1286
1214		− if (adreno_gpu->base.hw_apriv \|\| a6xx_gpu->has_whereami)
	1287	+ if (a6xx_gpu->shadow_bo)
1215	1288	for (i = 0; i < gpu->nr_rings; i++)
1216	1289	a6xx_gpu->shadow[i] = 0;
1217	1290

drivers/gpu/drm/msm/adreno/a6xx_gpu.h+1−1

@@ -44,7 +44,7 @@ struct a6xx_gpu { · 43 unmodified lines
44	44	* REG_CP_PROTECT_REG(n) - this will block both reads and writes for _len
45	45	* registers starting at _reg.
46	46	*/
47		−#define A6XX_PROTECT_RW(_reg, _len) \
	47	+#define A6XX_PROTECT_NORDWR(_reg, _len) \
48	48	((1 << 31) \| \
49	49	(((_len) & 0x3FFF) << 18) \| ((_reg) & 0x3FFFF))
50	50

drivers/gpu/drm/msm/dsi/phy/dsi_phy_10nm.c+1−0

@@ -432,6 +432,7 @@ static unsigned long dsi_pll_10nm_vco_recalc_rate(struct clk_hw *hw, · 431 unmodified lines
432	432	pll_freq += div_u64(tmp64, multiplier);
433	433
434	434	vco_rate = pll_freq;
	435	+ pll_10nm->vco_current_rate = vco_rate;
435	436
436	437	DBG("DSI PLL%d returning vco rate = %lu, dec = %x, frac = %x",
437	438	pll_10nm->phy->id, (unsigned long)vco_rate, dec, frac);

drivers/gpu/drm/msm/dsi/phy/dsi_phy_7nm.c+1−0

@@ -460,6 +460,7 @@ static unsigned long dsi_pll_7nm_vco_recalc_rate(struct clk_hw *hw, · 459 unmodified lines
460	460	pll_freq += div_u64(tmp64, multiplier);
461	461
462	462	vco_rate = pll_freq;
	463	+ pll_7nm->vco_current_rate = vco_rate;
463	464
464	465	DBG("DSI PLL%d returning vco rate = %lu, dec = %x, frac = %x",
465	466	pll_7nm->phy->id, (unsigned long)vco_rate, dec, frac);

drivers/gpu/drm/msm/msm_gem.c+7−0

@@ -1240,6 +1240,13 @@ static struct drm_gem_object _msm_gem_new(struct drm_device dev, · 1239 unmodified lines
1240	1240
1241	1241	to_msm_bo(obj)->vram_node = &vma->node;
1242	1242
	1243	+ /* Call chain get_pages() -> update_inactive() tries to
	1244	+ * access msm_obj->mm_list, but it is not initialized yet.
	1245	+ * To avoid NULL pointer dereference error, initialize
	1246	+ * mm_list to be empty.
	1247	+ */
	1248	+ INIT_LIST_HEAD(&msm_obj->mm_list);
	1249	+
1243	1250	msm_gem_lock(obj);
1244	1251	pages = get_pages(obj);
1245	1252	msm_gem_unlock(obj);

drivers/gpu/drm/sun4i/sun8i_dw_hdmi.c+27−4

@@ -209,7 +209,7 @@ static int sun8i_dw_hdmi_bind(struct device dev, struct device master, · 208 unmodified lines
209	209	goto err_disable_clk_tmds;
210	210	}
211	211
212		− ret = sun8i_hdmi_phy_probe(hdmi, phy_node);
	212	+ ret = sun8i_hdmi_phy_get(hdmi, phy_node);
213	213	of_node_put(phy_node);
214	214	if (ret) {
215	215	dev_err(dev, "Couldn't get the HDMI PHY\n");
@@ -242,7 +242,6 @@ static int sun8i_dw_hdmi_bind(struct device dev, struct device master, · 26 unmodified lines
242	242
243	243	cleanup_encoder:
244	244	drm_encoder_cleanup(encoder);
245		− sun8i_hdmi_phy_remove(hdmi);
246	245	err_disable_clk_tmds:
247	246	clk_disable_unprepare(hdmi->clk_tmds);
248	247	err_assert_ctrl_reset:
@@ -263,7 +262,6 @@ static void sun8i_dw_hdmi_unbind(struct device dev, struct device master, · 14 unmodified lines
263	262	struct sun8i_dw_hdmi *hdmi = dev_get_drvdata(dev);
264	263
265	264	dw_hdmi_unbind(hdmi->hdmi);
266		− sun8i_hdmi_phy_remove(hdmi);
267	265	clk_disable_unprepare(hdmi->clk_tmds);
268	266	reset_control_assert(hdmi->rst_ctrl);
269	267	gpiod_set_value(hdmi->ddc_en, 0);
@@ -320,7 +318,32 @@ static struct platform_driver sun8i_dw_hdmi_pltfm_driver = { · 50 unmodified lines
320	318	.of_match_table = sun8i_dw_hdmi_dt_ids,
321	319	},
322	320	};
323		−module_platform_driver(sun8i_dw_hdmi_pltfm_driver);
	321	+
	322	+static int __init sun8i_dw_hdmi_init(void)
	323	+{
	324	+ int ret;
	325	+
	326	+ ret = platform_driver_register(&sun8i_dw_hdmi_pltfm_driver);
	327	+ if (ret)
	328	+ return ret;
	329	+
	330	+ ret = platform_driver_register(&sun8i_hdmi_phy_driver);
	331	+ if (ret) {
	332	+ platform_driver_unregister(&sun8i_dw_hdmi_pltfm_driver);
	333	+ return ret;
	334	+ }
	335	+
	336	+ return ret;
	337	+}
	338	+
	339	+static void __exit sun8i_dw_hdmi_exit(void)
	340	+{
	341	+ platform_driver_unregister(&sun8i_dw_hdmi_pltfm_driver);
	342	+ platform_driver_unregister(&sun8i_hdmi_phy_driver);
	343	+}
	344	+
	345	+module_init(sun8i_dw_hdmi_init);
	346	+module_exit(sun8i_dw_hdmi_exit);
324	347
325	348	MODULE_AUTHOR("Jernej Skrabec <jernej.skrabec@siol.net>");
326	349	MODULE_DESCRIPTION("Allwinner DW HDMI bridge");

drivers/gpu/drm/sun4i/sun8i_dw_hdmi.h+3−2

@@ -195,14 +195,15 @@ struct sun8i_dw_hdmi { · 194 unmodified lines
195	195	struct gpio_desc *ddc_en;
196	196	};
197	197
	198	+extern struct platform_driver sun8i_hdmi_phy_driver;
	199	+
198	200	static inline struct sun8i_dw_hdmi *
199	201	encoder_to_sun8i_dw_hdmi(struct drm_encoder *encoder)
200	202	{
201	203	return container_of(encoder, struct sun8i_dw_hdmi, encoder);
202	204	}
203	205
204		−int sun8i_hdmi_phy_probe(struct sun8i_dw_hdmi hdmi, struct device_node node);
205		−void sun8i_hdmi_phy_remove(struct sun8i_dw_hdmi *hdmi);
	206	+int sun8i_hdmi_phy_get(struct sun8i_dw_hdmi hdmi, struct device_node node);
206	207
207	208	void sun8i_hdmi_phy_init(struct sun8i_hdmi_phy *phy);
208	209	void sun8i_hdmi_phy_set_ops(struct sun8i_hdmi_phy *phy,

drivers/gpu/drm/sun4i/sun8i_hdmi_phy.c+36−5

@@ -5,6 +5,7 @@ · 4 unmodified lines
5	5
6	6	#include <linux/delay.h>
7	7	#include <linux/of_address.h>
	8	+#include <linux/of_platform.h>
8	9
9	10	#include "sun8i_dw_hdmi.h"
10	11
@@ -597,10 +598,30 @@ static const struct of_device_id sun8i_hdmi_phy_of_table[] = { · 586 unmodified lines
597	598	{ /* sentinel */ }
598	599	};
599	600
600		−int sun8i_hdmi_phy_probe(struct sun8i_dw_hdmi hdmi, struct device_node node)
	601	+int sun8i_hdmi_phy_get(struct sun8i_dw_hdmi hdmi, struct device_node node)
	602	+{
	603	+ struct platform_device *pdev = of_find_device_by_node(node);
	604	+ struct sun8i_hdmi_phy *phy;
	605	+
	606	+ if (!pdev)
	607	+ return -EPROBE_DEFER;
	608	+
	609	+ phy = platform_get_drvdata(pdev);
	610	+ if (!phy)
	611	+ return -EPROBE_DEFER;
	612	+
	613	+ hdmi->phy = phy;
	614	+
	615	+ put_device(&pdev->dev);
	616	+
	617	+ return 0;
	618	+}
	619	+
	620	+static int sun8i_hdmi_phy_probe(struct platform_device *pdev)
601	621	{
602	622	const struct of_device_id *match;
603		− struct device *dev = hdmi->dev;
	623	+ struct device *dev = &pdev->dev;
	624	+ struct device_node *node = dev->of_node;
604	625	struct sun8i_hdmi_phy *phy;
605	626	struct resource res;
606	627	void __iomem *regs;
@@ -704,7 +725,7 @@ int sun8i_hdmi_phy_probe(struct sun8i_dw_hdmi hdmi, struct device_node node) · 97 unmodified lines
704	725	clk_prepare_enable(phy->clk_phy);
705	726	}
706	727
707		− hdmi->phy = phy;
	728	+ platform_set_drvdata(pdev, phy);
708	729
709	730	return 0;
710	731
@@ -728,9 +749,9 @@ err_put_clk_bus: · 17 unmodified lines
728	749	return ret;
729	750	}
730	751
731		−void sun8i_hdmi_phy_remove(struct sun8i_dw_hdmi *hdmi)
	752	+static int sun8i_hdmi_phy_remove(struct platform_device *pdev)
732	753	{
733		− struct sun8i_hdmi_phy *phy = hdmi->phy;
	754	+ struct sun8i_hdmi_phy *phy = platform_get_drvdata(pdev);
734	755
735	756	clk_disable_unprepare(phy->clk_mod);
736	757	clk_disable_unprepare(phy->clk_bus);
@@ -744,4 +765,14 @@ void sun8i_hdmi_phy_remove(struct sun8i_dw_hdmi *hdmi) · 7 unmodified lines
744	765	clk_put(phy->clk_pll1);
745	766	clk_put(phy->clk_mod);
746	767	clk_put(phy->clk_bus);
	768	+ return 0;
747	769	}
	770	+
	771	+struct platform_driver sun8i_hdmi_phy_driver = {
	772	+ .probe = sun8i_hdmi_phy_probe,
	773	+ .remove = sun8i_hdmi_phy_remove,
	774	+ .driver = {
	775	+ .name = "sun8i-hdmi-phy",
	776	+ .of_match_table = sun8i_hdmi_phy_of_table,
	777	+ },
	778	+};

drivers/gpu/drm/tegra/drm.h+1−1

@@ -25,7 +25,7 @@ · 24 unmodified lines
25	25	#include "trace.h"
26	26
27	27	/* XXX move to include/uapi/drm/drm_fourcc.h? */
28		−#define DRM_FORMAT_MOD_NVIDIA_SECTOR_LAYOUT BIT(22)
	28	+#define DRM_FORMAT_MOD_NVIDIA_SECTOR_LAYOUT BIT_ULL(22)
29	29
30	30	struct reset_control;
31	31

drivers/gpu/drm/tegra/hub.c+1−1

@@ -510,7 +510,7 @@ static void tegra_shared_plane_atomic_update(struct drm_plane *plane, · 509 unmodified lines
510	510	* dGPU sector layout.
511	511	*/
512	512	if (tegra_plane_state->tiling.sector_layout == TEGRA_BO_SECTOR_LAYOUT_GPU)
513		− base \|= BIT(39);
	513	+ base \|= BIT_ULL(39);
514	514	#endif
515	515
516	516	tegra_plane_writel(p, tegra_plane_state->format, DC_WIN_COLOR_DEPTH);

drivers/gpu/drm/tegra/sor.c+42−28

@@ -3125,21 +3125,21 @@ static int tegra_sor_init(struct host1x_client *client) · 3124 unmodified lines
3125	3125	if (err < 0) {
3126	3126	dev_err(sor->dev, "failed to acquire SOR reset: %d\n",
3127	3127	err);
3128		− return err;
	3128	+ goto rpm_put;
3129	3129	}
3130	3130
3131	3131	err = reset_control_assert(sor->rst);
3132	3132	if (err < 0) {
3133	3133	dev_err(sor->dev, "failed to assert SOR reset: %d\n",
3134	3134	err);
3135		− return err;
	3135	+ goto rpm_put;
3136	3136	}
3137	3137	}
3138	3138
3139	3139	err = clk_prepare_enable(sor->clk);
3140	3140	if (err < 0) {
3141	3141	dev_err(sor->dev, "failed to enable clock: %d\n", err);
3142		− return err;
	3142	+ goto rpm_put;
3143	3143	}
3144	3144
3145	3145	usleep_range(1000, 3000);
@@ -3150,7 +3150,7 @@ static int tegra_sor_init(struct host1x_client *client) · 4 unmodified lines
3150	3150	dev_err(sor->dev, "failed to deassert SOR reset: %d\n",
3151	3151	err);
3152	3152	clk_disable_unprepare(sor->clk);
3153		− return err;
	3153	+ goto rpm_put;
3154	3154	}
3155	3155
3156	3156	reset_control_release(sor->rst);
@@ -3171,6 +3171,12 @@ static int tegra_sor_init(struct host1x_client *client) · 14 unmodified lines
3171	3171	}
3172	3172
3173	3173	return 0;
	3174	+
	3175	+rpm_put:
	3176	+ if (sor->rst)
	3177	+ pm_runtime_put(sor->dev);
	3178	+
	3179	+ return err;
3174	3180	}
3175	3181
3176	3182	static int tegra_sor_exit(struct host1x_client *client)
@@ -3739,12 +3745,8 @@ static int tegra_sor_probe(struct platform_device *pdev) · 562 unmodified lines
3739	3745	if (!sor->aux)
3740	3746	return -EPROBE_DEFER;
3741	3747
3742		− if (get_device(&sor->aux->ddc.dev)) {
3743		− if (try_module_get(sor->aux->ddc.owner))
3744		− sor->output.ddc = &sor->aux->ddc;
3745		− else
3746		− put_device(&sor->aux->ddc.dev);
3747		− }
	3748	+ if (get_device(sor->aux->dev))
	3749	+ sor->output.ddc = &sor->aux->ddc;
3748	3750	}
3749	3751
3750	3752	if (!sor->aux) {
@@ -3772,12 +3774,13 @@ static int tegra_sor_probe(struct platform_device *pdev) · 21 unmodified lines
3772	3774
3773	3775	err = tegra_sor_parse_dt(sor);
3774	3776	if (err < 0)
3775		− return err;
	3777	+ goto put_aux;
3776	3778
3777	3779	err = tegra_output_probe(&sor->output);
3778		− if (err < 0)
3779		− return dev_err_probe(&pdev->dev, err,
3780		− "failed to probe output\n");
	3780	+ if (err < 0) {
	3781	+ dev_err_probe(&pdev->dev, err, "failed to probe output\n");
	3782	+ goto put_aux;
	3783	+ }
3781	3784
3782	3785	if (sor->ops && sor->ops->probe) {
3783	3786	err = sor->ops->probe(sor);
@@ -3916,17 +3919,10 @@ static int tegra_sor_probe(struct platform_device *pdev) · 132 unmodified lines
3916	3919	platform_set_drvdata(pdev, sor);
3917	3920	pm_runtime_enable(&pdev->dev);
3918	3921
3919		− INIT_LIST_HEAD(&sor->client.list);
	3922	+ host1x_client_init(&sor->client);
3920	3923	sor->client.ops = &sor_client_ops;
3921	3924	sor->client.dev = &pdev->dev;
3922	3925
3923		− err = host1x_client_register(&sor->client);
3924		− if (err < 0) {
3925		− dev_err(&pdev->dev, "failed to register host1x client: %d\n",
3926		− err);
3927		− goto rpm_disable;
3928		− }
3929		−
3930	3926	/*
3931	3927	* On Tegra210 and earlier, provide our own implementation for the
3932	3928	* pad output clock.
@@ -3938,13 +3934,13 @@ static int tegra_sor_probe(struct platform_device *pdev) · 5 unmodified lines
3938	3934	sor->index);
3939	3935	if (!name) {
3940	3936	err = -ENOMEM;
3941		− goto unregister;
	3937	+ goto uninit;
3942	3938	}
3943	3939
3944	3940	err = host1x_client_resume(&sor->client);
3945	3941	if (err < 0) {
3946	3942	dev_err(sor->dev, "failed to resume: %d\n", err);
3947		− goto unregister;
	3943	+ goto uninit;
3948	3944	}
3949	3945
3950	3946	sor->clk_pad = tegra_clk_sor_pad_register(sor, name);
@@ -3955,17 +3951,30 @@ static int tegra_sor_probe(struct platform_device *pdev) · 4 unmodified lines
3955	3951	err = PTR_ERR(sor->clk_pad);
3956	3952	dev_err(sor->dev, "failed to register SOR pad clock: %d\n",
3957	3953	err);
3958		− goto unregister;
	3954	+ goto uninit;
	3955	+ }
	3956	+
	3957	+ err = __host1x_client_register(&sor->client);
	3958	+ if (err < 0) {
	3959	+ dev_err(&pdev->dev, "failed to register host1x client: %d\n",
	3960	+ err);
	3961	+ goto uninit;
3959	3962	}
3960	3963
3961	3964	return 0;
3962	3965
3963		−unregister:
3964		− host1x_client_unregister(&sor->client);
3965		−rpm_disable:
	3966	+uninit:
	3967	+ host1x_client_exit(&sor->client);
3966	3968	pm_runtime_disable(&pdev->dev);
3967	3969	remove:
	3970	+ if (sor->aux)
	3971	+ sor->output.ddc = NULL;
	3972	+
3968	3973	tegra_output_remove(&sor->output);
	3974	+put_aux:
	3975	+ if (sor->aux)
	3976	+ put_device(sor->aux->dev);
	3977	+
3969	3978	return err;
3970	3979	}
3971	3980
@@ -3983,6 +3992,11 @@ static int tegra_sor_remove(struct platform_device *pdev) · 11 unmodified lines
3983	3992
3984	3993	pm_runtime_disable(&pdev->dev);
3985	3994
	3995	+ if (sor->aux) {
	3996	+ put_device(sor->aux->dev);
	3997	+ sor->output.ddc = NULL;
	3998	+ }
	3999	+
3986	4000	tegra_output_remove(&sor->output);
3987	4001
3988	4002	return 0;

drivers/gpu/drm/ttm/ttm_bo.c+4−1

@@ -1139,7 +1139,10 @@ int ttm_bo_swapout(struct ttm_buffer_object bo, struct ttm_operation_ctx ctx, · 1138 unmodified lines
1139	1139	if (!ttm_bo_evict_swapout_allowable(bo, ctx, &place, &locked, NULL))
1140	1140	return -EBUSY;
1141	1141
1142		− if (!ttm_bo_get_unless_zero(bo)) {
	1142	+ if (!bo->ttm \|\| !ttm_tt_is_populated(bo->ttm) \|\|
	1143	+ bo->ttm->page_flags & TTM_PAGE_FLAG_SG \|\|
	1144	+ bo->ttm->page_flags & TTM_PAGE_FLAG_SWAPPED \|\|
	1145	+ !ttm_bo_get_unless_zero(bo)) {
1143	1146	if (locked)
1144	1147	dma_resv_unlock(bo->base.resv);
1145	1148	return -EBUSY;

drivers/gpu/drm/ttm/ttm_device.c+1−7

@@ -143,14 +143,8 @@ int ttm_device_swapout(struct ttm_device bdev, struct ttm_operation_ctx ctx, · 142 unmodified lines
143	143
144	144	for (j = 0; j < TTM_MAX_BO_PRIORITY; ++j) {
145	145	list_for_each_entry(bo, &man->lru[j], lru) {
146		− uint32_t num_pages;
	146	+ uint32_t num_pages = PFN_UP(bo->base.size);
147	147
148		− if (!bo->ttm \|\|
149		− bo->ttm->page_flags & TTM_PAGE_FLAG_SG \|\|
150		− bo->ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)
151		− continue;
152		−
153		− num_pages = bo->ttm->num_pages;
154	148	ret = ttm_bo_swapout(bo, ctx, gfp_flags);
155	149	/* ttm_bo_swapout has dropped the lru_lock */
156	150	if (!ret)

drivers/gpu/drm/vc4/vc4_kms.c+1−1

@@ -372,7 +372,7 @@ static void vc4_atomic_commit_tail(struct drm_atomic_state *state) · 371 unmodified lines
372	372	if (!old_hvs_state->fifo_state[channel].in_use)
373	373	continue;
374	374
375		− ret = drm_crtc_commit_wait(old_hvs_state->fifo_state[i].pending_commit);
	375	+ ret = drm_crtc_commit_wait(old_hvs_state->fifo_state[channel].pending_commit);
376	376	if (ret)
377	377	drm_err(dev, "Timed out waiting for commit\n");
378	378	}

drivers/gpu/host1x/bus.c+24−6

@@ -735,6 +735,29 @@ void host1x_driver_unregister(struct host1x_driver *driver) · 734 unmodified lines
735	735	}
736	736	EXPORT_SYMBOL(host1x_driver_unregister);
737	737
	738	+/**
	739	+ * __host1x_client_init() - initialize a host1x client
	740	+ * @client: host1x client
	741	+ * @key: lock class key for the client-specific mutex
	742	+ */
	743	+void __host1x_client_init(struct host1x_client client, struct lock_class_key key)
	744	+{
	745	+ INIT_LIST_HEAD(&client->list);
	746	+ __mutex_init(&client->lock, "host1x client lock", key);
	747	+ client->usecount = 0;
	748	+}
	749	+EXPORT_SYMBOL(__host1x_client_init);
	750	+
	751	+/**
	752	+ * host1x_client_exit() - uninitialize a host1x client
	753	+ * @client: host1x client
	754	+ */
	755	+void host1x_client_exit(struct host1x_client *client)
	756	+{
	757	+ mutex_destroy(&client->lock);
	758	+}
	759	+EXPORT_SYMBOL(host1x_client_exit);
	760	+
738	761	/**
739	762	* __host1x_client_register() - register a host1x client
740	763	* @client: host1x client
@@ -747,16 +770,11 @@ EXPORT_SYMBOL(host1x_driver_unregister); · 6 unmodified lines
747	770	* device and call host1x_device_init(), which will in turn call each client's
748	771	* &host1x_client_ops.init implementation.
749	772	*/
750		−int __host1x_client_register(struct host1x_client *client,
751		− struct lock_class_key *key)
	773	+int __host1x_client_register(struct host1x_client *client)
752	774	{
753	775	struct host1x *host1x;
754	776	int err;
755	777
756		− INIT_LIST_HEAD(&client->list);
757		− __mutex_init(&client->lock, "host1x client lock", key);
758		− client->usecount = 0;
759		−
760	778	mutex_lock(&devices_lock);
761	779
762	780	list_for_each_entry(host1x, &devices, list) {

drivers/hid/Kconfig+17−2

@@ -93,11 +93,11 @@ menu "Special HID drivers" · 92 unmodified lines
93	93	depends on HID
94	94
95	95	config HID_A4TECH
96		− tristate "A4 tech mice"
	96	+ tristate "A4TECH mice"
97	97	depends on HID
98	98	default !EXPERT
99	99	help
100		− Support for A4 tech X5 and WOP-35 / Trust 450L mice.
	100	+ Support for some A4TECH mice with two scroll wheels.
101	101
102	102	config HID_ACCUTOUCH
103	103	tristate "Accutouch touch device"
@@ -922,6 +922,21 @@ config HID_SAMSUNG · 818 unmodified lines
922	922	help
923	923	Support for Samsung InfraRed remote control or keyboards.
924	924
	925	+config HID_SEMITEK
	926	+ tristate "Semitek USB keyboards"
	927	+ depends on HID
	928	+ help
	929	+ Support for Semitek USB keyboards that are not fully compliant
	930	+ with the HID standard.
	931	+
	932	+ There are many variants, including:
	933	+ - GK61, GK64, GK68, GK84, GK96, etc.
	934	+ - SK61, SK64, SK68, SK84, SK96, etc.
	935	+ - Dierya DK61/DK66
	936	+ - Tronsmart TK09R
	937	+ - Woo-dy
	938	+ - X-Bows Nature/Knight
	939	+
925	940	config HID_SONY
926	941	tristate "Sony PS2/3/4 accessories"
927	942	depends on USB_HID

drivers/hid/Makefile+1−0

@@ -106,6 +106,7 @@ obj-$(CONFIG_HID_ROCCAT) += hid-roccat.o hid-roccat-common.o \ · 105 unmodified lines
106	106	obj-$(CONFIG_HID_RMI) += hid-rmi.o
107	107	obj-$(CONFIG_HID_SAITEK) += hid-saitek.o
108	108	obj-$(CONFIG_HID_SAMSUNG) += hid-samsung.o
	109	+obj-$(CONFIG_HID_SEMITEK) += hid-semitek.o
109	110	obj-$(CONFIG_HID_SMARTJOYPLUS) += hid-sjoy.o
110	111	obj-$(CONFIG_HID_SONY) += hid-sony.o
111	112	obj-$(CONFIG_HID_SPEEDLINK) += hid-speedlink.o

drivers/hid/amd-sfh-hid/amd_sfh_client.c+10−9

@@ -88,6 +88,7 @@ static void amd_sfh_work(struct work_struct *work) · 87 unmodified lines
88	88	sensor_index = req_node->sensor_idx;
89	89	report_id = req_node->report_id;
90	90	node_type = req_node->report_type;
	91	+ kfree(req_node);
91	92
92	93	if (node_type == HID_FEATURE_REPORT) {
93	94	report_size = get_feature_report(sensor_index, report_id,
@@ -142,7 +143,7 @@ int amd_sfh_hid_client_init(struct amd_mp2_dev *privdata) · 48 unmodified lines
142	143	int rc, i;
143	144
144	145	dev = &privdata->pdev->dev;
145		− cl_data = kzalloc(sizeof(*cl_data), GFP_KERNEL);
	146	+ cl_data = devm_kzalloc(dev, sizeof(*cl_data), GFP_KERNEL);
146	147	if (!cl_data)
147	148	return -ENOMEM;
148	149
@@ -175,12 +176,12 @@ int amd_sfh_hid_client_init(struct amd_mp2_dev *privdata) · 26 unmodified lines
175	176	rc = -EINVAL;
176	177	goto cleanup;
177	178	}
178		− cl_data->feature_report[i] = kzalloc(feature_report_size, GFP_KERNEL);
	179	+ cl_data->feature_report[i] = devm_kzalloc(dev, feature_report_size, GFP_KERNEL);
179	180	if (!cl_data->feature_report[i]) {
180	181	rc = -ENOMEM;
181	182	goto cleanup;
182	183	}
183		− cl_data->input_report[i] = kzalloc(input_report_size, GFP_KERNEL);
	184	+ cl_data->input_report[i] = devm_kzalloc(dev, input_report_size, GFP_KERNEL);
184	185	if (!cl_data->input_report[i]) {
185	186	rc = -ENOMEM;
186	187	goto cleanup;
@@ -189,7 +190,8 @@ int amd_sfh_hid_client_init(struct amd_mp2_dev *privdata) · 2 unmodified lines
189	190	info.sensor_idx = cl_idx;
190	191	info.dma_address = cl_data->sensor_dma_addr[i];
191	192
192		− cl_data->report_descr[i] = kzalloc(cl_data->report_descr_sz[i], GFP_KERNEL);
	193	+ cl_data->report_descr[i] =
	194	+ devm_kzalloc(dev, cl_data->report_descr_sz[i], GFP_KERNEL);
193	195	if (!cl_data->report_descr[i]) {
194	196	rc = -ENOMEM;
195	197	goto cleanup;
@@ -214,11 +216,11 @@ cleanup: · 18 unmodified lines
214	216	cl_data->sensor_virt_addr[i],
215	217	cl_data->sensor_dma_addr[i]);
216	218	}
217		− kfree(cl_data->feature_report[i]);
218		− kfree(cl_data->input_report[i]);
219		− kfree(cl_data->report_descr[i]);
	219	+ devm_kfree(dev, cl_data->feature_report[i]);
	220	+ devm_kfree(dev, cl_data->input_report[i]);
	221	+ devm_kfree(dev, cl_data->report_descr[i]);
220	222	}
221		− kfree(cl_data);
	223	+ devm_kfree(dev, cl_data);
222	224	return rc;
223	225	}
224	226
@@ -241,6 +243,5 @@ int amd_sfh_hid_client_deinit(struct amd_mp2_dev *privdata) · 16 unmodified lines
241	243	cl_data->sensor_dma_addr[i]);
242	244	}
243	245	}
244		− kfree(cl_data);
245	246	return 0;
246	247	}

drivers/hid/amd-sfh-hid/amd_sfh_hid.c+0−3

@@ -162,9 +162,6 @@ void amdtp_hid_remove(struct amdtp_cl_data *cli_data) · 161 unmodified lines
162	162	int i;
163	163
164	164	for (i = 0; i < cli_data->num_hid_devices; ++i) {
165		− kfree(cli_data->feature_report[i]);
166		− kfree(cli_data->input_report[i]);
167		− kfree(cli_data->report_descr[i]);
168	165	if (cli_data->hid_sensor_hubs[i]) {
169	166	kfree(cli_data->hid_sensor_hubs[i]->driver_data);
170	167	hid_destroy_device(cli_data->hid_sensor_hubs[i]);

drivers/hid/hid-a4tech.c+2−0

@@ -147,6 +147,8 @@ static const struct hid_device_id a4_devices[] = { · 146 unmodified lines
147	147	.driver_data = A4_2WHEEL_MOUSE_HACK_B8 },
148	148	{ HID_USB_DEVICE(USB_VENDOR_ID_A4TECH, USB_DEVICE_ID_A4TECH_RP_649),
149	149	.driver_data = A4_2WHEEL_MOUSE_HACK_B8 },
	150	+ { HID_USB_DEVICE(USB_VENDOR_ID_A4TECH, USB_DEVICE_ID_A4TECH_NB_95),
	151	+ .driver_data = A4_2WHEEL_MOUSE_HACK_B8 },
150	152	{ }
151	153	};
152	154	MODULE_DEVICE_TABLE(hid, a4_devices);

drivers/hid/hid-asus.c+20−12

@@ -79,10 +79,9 @@ MODULE_DESCRIPTION("Asus HID Keyboard and TouchPad"); · 78 unmodified lines
79	79	#define QUIRK_T100_KEYBOARD BIT(6)
80	80	#define QUIRK_T100CHI BIT(7)
81	81	#define QUIRK_G752_KEYBOARD BIT(8)
82		−#define QUIRK_T101HA_DOCK BIT(9)
83		−#define QUIRK_T90CHI BIT(10)
84		−#define QUIRK_MEDION_E1239T BIT(11)
85		−#define QUIRK_ROG_NKEY_KEYBOARD BIT(12)
	82	+#define QUIRK_T90CHI BIT(9)
	83	+#define QUIRK_MEDION_E1239T BIT(10)
	84	+#define QUIRK_ROG_NKEY_KEYBOARD BIT(11)
86	85
87	86	#define I2C_KEYBOARD_QUIRKS (QUIRK_FIX_NOTEBOOK_REPORT \| \
88	87	QUIRK_NO_INIT_REPORTS \| \
@@ -335,7 +334,7 @@ static int asus_raw_event(struct hid_device *hdev, · 246 unmodified lines
335	334	if (drvdata->quirks & QUIRK_MEDION_E1239T)
336	335	return asus_e1239t_event(drvdata, data, size);
337	336
338		− if (drvdata->quirks & QUIRK_ROG_NKEY_KEYBOARD) {
	337	+ if (drvdata->quirks & QUIRK_USE_KBD_BACKLIGHT) {
339	338	/*
340	339	* Skip these report ID, the device emits a continuous stream associated
341	340	* with the AURA mode it is in which looks like an 'echo'.
@@ -355,6 +354,16 @@ static int asus_raw_event(struct hid_device *hdev, · 13 unmodified lines
355	354	return -1;
356	355	}
357	356	}
	357	+ if (drvdata->quirks & QUIRK_ROG_NKEY_KEYBOARD) {
	358	+ /*
	359	+ * G713 and G733 send these codes on some keypresses, depending on
	360	+ * the key pressed it can trigger a shutdown event if not caught.
	361	+ */
	362	+ if(data[0] == 0x02 && data[1] == 0x30) {
	363	+ return -1;
	364	+ }
	365	+ }
	366	+
358	367	}
359	368
360	369	return 0;
@@ -1072,11 +1081,6 @@ static int asus_probe(struct hid_device hdev, const struct hid_device_id id) · 711 unmodified lines
1072	1081	return ret;
1073	1082	}
1074	1083
1075		− /* use hid-multitouch for T101HA touchpad */
1076		− if (id->driver_data & QUIRK_T101HA_DOCK &&
1077		− hdev->collection->usage == HID_GD_MOUSE)
1078		− return -ENODEV;
1079		−
1080	1084	ret = hid_hw_start(hdev, HID_CONNECT_DEFAULT);
1081	1085	if (ret) {
1082	1086	hid_err(hdev, "Asus hw start failed: %d\n", ret);
@@ -1230,8 +1234,6 @@ static const struct hid_device_id asus_devices[] = { · 147 unmodified lines
1230	1234	{ HID_USB_DEVICE(USB_VENDOR_ID_ASUSTEK,
1231	1235	USB_DEVICE_ID_ASUSTEK_T100TAF_KEYBOARD),
1232	1236	QUIRK_T100_KEYBOARD \| QUIRK_NO_CONSUMER_USAGES },
1233		− { HID_USB_DEVICE(USB_VENDOR_ID_ASUSTEK,
1234		− USB_DEVICE_ID_ASUSTEK_T101HA_KEYBOARD), QUIRK_T101HA_DOCK },
1235	1237	{ HID_USB_DEVICE(USB_VENDOR_ID_CHICONY, USB_DEVICE_ID_ASUS_AK1D) },
1236	1238	{ HID_USB_DEVICE(USB_VENDOR_ID_TURBOX, USB_DEVICE_ID_ASUS_MD_5110) },
1237	1239	{ HID_USB_DEVICE(USB_VENDOR_ID_JESS, USB_DEVICE_ID_ASUS_MD_5112) },
@@ -1239,6 +1241,12 @@ static const struct hid_device_id asus_devices[] = { · 1 unmodified line
1239	1241	USB_DEVICE_ID_ASUSTEK_T100CHI_KEYBOARD), QUIRK_T100CHI },
1240	1242	{ HID_USB_DEVICE(USB_VENDOR_ID_ITE, USB_DEVICE_ID_ITE_MEDION_E1239T),
1241	1243	QUIRK_MEDION_E1239T },
	1244	+ /*
	1245	+ * Note bind to the HID_GROUP_GENERIC group, so that we only bind to the keyboard
	1246	+ * part, while letting hid-multitouch.c handle the touchpad.
	1247	+ */
	1248	+ { HID_DEVICE(BUS_USB, HID_GROUP_GENERIC,
	1249	+ USB_VENDOR_ID_ASUSTEK, USB_DEVICE_ID_ASUSTEK_T101HA_KEYBOARD) },
1242	1250	{ }
1243	1251	};
1244	1252	MODULE_DEVICE_TABLE(hid, asus_devices);

drivers/hid/hid-core.c+3−1

@@ -2005,6 +2005,9 @@ int hid_connect(struct hid_device *hdev, unsigned int connect_mask) · 2004 unmodified lines
2005	2005	case BUS_I2C:
2006	2006	bus = "I2C";
2007	2007	break;
	2008	+ case BUS_VIRTUAL:
	2009	+ bus = "VIRTUAL";
	2010	+ break;
2008	2011	default:
2009	2012	bus = "<UNKNOWN>";
2010	2013	}
@@ -2588,7 +2591,6 @@ int hid_check_keys_pressed(struct hid_device *hid) · 577 unmodified lines
2588	2591
2589	2592	return 0;
2590	2593	}
2591		−
2592	2594	EXPORT_SYMBOL_GPL(hid_check_keys_pressed);
2593	2595
2594	2596	static int __init hid_init(void)

drivers/hid/hid-debug.c+3−0

@@ -930,6 +930,9 @@ static const char *keys[KEY_MAX + 1] = { · 929 unmodified lines
930	930	[KEY_APPSELECT] = "AppSelect",
931	931	[KEY_SCREENSAVER] = "ScreenSaver",
932	932	[KEY_VOICECOMMAND] = "VoiceCommand",
	933	+ [KEY_ASSISTANT] = "Assistant",
	934	+ [KEY_KBD_LAYOUT_NEXT] = "KbdLayoutNext",
	935	+ [KEY_EMOJI_PICKER] = "EmojiPicker",
933	936	[KEY_BRIGHTNESS_MIN] = "BrightnessMin",
934	937	[KEY_BRIGHTNESS_MAX] = "BrightnessMax",
935	938	[KEY_BRIGHTNESS_AUTO] = "BrightnessAuto",

drivers/hid/hid-ft260.c+16−13

@@ -201,7 +201,7 @@ struct ft260_i2c_write_request_report { · 200 unmodified lines
201	201	u8 address; /* 7-bit I2C address */
202	202	u8 flag; /* I2C transaction condition */
203	203	u8 length; /* data payload length */
204		− u8 data[60]; /* data payload */
	204	+ u8 data[FT260_WR_DATA_MAX]; /* data payload */
205	205	} __packed;
206	206
207	207	struct ft260_i2c_read_request_report {
@@ -249,7 +249,10 @@ static int ft260_hid_feature_report_get(struct hid_device *hdev, · 41 unmodified lines
249	249
250	250	ret = hid_hw_raw_request(hdev, report_id, buf, len, HID_FEATURE_REPORT,
251	251	HID_REQ_GET_REPORT);
252		− memcpy(data, buf, len);
	252	+ if (likely(ret == len))
	253	+ memcpy(data, buf, len);
	254	+ else if (ret >= 0)
	255	+ ret = -EIO;
253	256	kfree(buf);
254	257	return ret;
255	258	}
@@ -298,7 +301,7 @@ static int ft260_xfer_status(struct ft260_device *dev) · 42 unmodified lines
298	301
299	302	ret = ft260_hid_feature_report_get(hdev, FT260_I2C_STATUS,
300	303	(u8 *)&report, sizeof(report));
301		− if (ret < 0) {
	304	+ if (unlikely(ret < 0)) {
302	305	hid_err(hdev, "failed to retrieve status: %d\n", ret);
303	306	return ret;
304	307	}
@@ -429,6 +432,9 @@ static int ft260_smbus_write(struct ft260_device *dev, u8 addr, u8 cmd, · 124 unmodified lines
429	432	struct ft260_i2c_write_request_report *rep =
430	433	(struct ft260_i2c_write_request_report *)dev->write_buf;
431	434
	435	+ if (data_len >= sizeof(rep->data))
	436	+ return -EINVAL;
	437	+
432	438	rep->address = addr;
433	439	rep->data[0] = cmd;
434	440	rep->length = data_len + 1;
@@ -721,10 +727,9 @@ static int ft260_get_system_config(struct hid_device *hdev, · 286 unmodified lines
721	727
722	728	ret = ft260_hid_feature_report_get(hdev, FT260_SYSTEM_SETTINGS,
723	729	(u8 *)cfg, len);
724		− if (ret != len) {
	730	+ if (ret < 0) {
725	731	hid_err(hdev, "failed to retrieve system status\n");
726		− if (ret >= 0)
727		− return -EIO;
	732	+ return ret;
728	733	}
729	734	return 0;
730	735	}
@@ -777,8 +782,8 @@ static int ft260_byte_show(struct hid_device hdev, int id, u8 cfg, int len, · 46 unmodified lines
777	782	int ret;
778	783
779	784	ret = ft260_hid_feature_report_get(hdev, id, cfg, len);
780		− if (ret != len && ret >= 0)
781		− return -EIO;
	785	+ if (ret < 0)
	786	+ return ret;
782	787
783	788	return scnprintf(buf, PAGE_SIZE, "%hi\n", *field);
784	789	}
@@ -789,8 +794,8 @@ static int ft260_word_show(struct hid_device hdev, int id, u8 cfg, int len, · 4 unmodified lines
789	794	int ret;
790	795
791	796	ret = ft260_hid_feature_report_get(hdev, id, cfg, len);
792		− if (ret != len && ret >= 0)
793		− return -EIO;
	797	+ if (ret < 0)
	798	+ return ret;
794	799
795	800	return scnprintf(buf, PAGE_SIZE, "%hi\n", le16_to_cpu(*field));
796	801	}
@@ -941,10 +946,8 @@ static int ft260_probe(struct hid_device hdev, const struct hid_device_id id) · 144 unmodified lines
941	946
942	947	ret = ft260_hid_feature_report_get(hdev, FT260_CHIP_VERSION,
943	948	(u8 *)&version, sizeof(version));
944		− if (ret != sizeof(version)) {
	949	+ if (ret < 0) {
945	950	hid_err(hdev, "failed to retrieve chip version\n");
946		− if (ret >= 0)
947		− ret = -EIO;
948	951	goto err_hid_close;
949	952	}
950	953

drivers/hid/hid-gt683r.c+1−0

@@ -54,6 +54,7 @@ static const struct hid_device_id gt683r_led_id[] = { · 53 unmodified lines
54	54	{ HID_USB_DEVICE(USB_VENDOR_ID_MSI, USB_DEVICE_ID_MSI_GT683R_LED_PANEL) },
55	55	{ }
56	56	};
	57	+MODULE_DEVICE_TABLE(hid, gt683r_led_id);
57	58
58	59	static void gt683r_brightness_set(struct led_classdev *led_cdev,
59	60	enum led_brightness brightness)

drivers/hid/hid-ids.h+7−2

@@ -26,6 +26,7 @@ · 25 unmodified lines
26	26	#define USB_DEVICE_ID_A4TECH_WCP32PU 0x0006
27	27	#define USB_DEVICE_ID_A4TECH_X5_005D 0x000a
28	28	#define USB_DEVICE_ID_A4TECH_RP_649 0x001a
	29	+#define USB_DEVICE_ID_A4TECH_NB_95 0x022b
29	30
30	31	#define USB_VENDOR_ID_AASHIMA 0x06d6
31	32	#define USB_DEVICE_ID_AASHIMA_GAMEPAD 0x0025
@@ -299,8 +300,6 @@ · 267 unmodified lines
299	300
300	301	#define USB_VENDOR_ID_CORSAIR 0x1b1c
301	302	#define USB_DEVICE_ID_CORSAIR_K90 0x1b02
302		−
303		−#define USB_VENDOR_ID_CORSAIR 0x1b1c
304	303	#define USB_DEVICE_ID_CORSAIR_K70R 0x1b09
305	304	#define USB_DEVICE_ID_CORSAIR_K95RGB 0x1b11
306	305	#define USB_DEVICE_ID_CORSAIR_M65RGB 0x1b12
@@ -751,6 +750,7 @@ · 444 unmodified lines
751	750	#define USB_DEVICE_ID_LENOVO_X1_COVER 0x6085
752	751	#define USB_DEVICE_ID_LENOVO_X1_TAB 0x60a3
753	752	#define USB_DEVICE_ID_LENOVO_X1_TAB3 0x60b5
	753	+#define USB_DEVICE_ID_LENOVO_OPTICAL_USB_MOUSE_600E 0x600e
754	754	#define USB_DEVICE_ID_LENOVO_PIXART_USB_MOUSE_608D 0x608d
755	755	#define USB_DEVICE_ID_LENOVO_PIXART_USB_MOUSE_6019 0x6019
756	756	#define USB_DEVICE_ID_LENOVO_PIXART_USB_MOUSE_602E 0x602e
@@ -1051,6 +1051,7 @@ · 294 unmodified lines
1051	1051	#define USB_DEVICE_ID_SAITEK_X52 0x075c
1052	1052	#define USB_DEVICE_ID_SAITEK_X52_2 0x0255
1053	1053	#define USB_DEVICE_ID_SAITEK_X52_PRO 0x0762
	1054	+#define USB_DEVICE_ID_SAITEK_X65 0x0b6a
1054	1055
1055	1056	#define USB_VENDOR_ID_SAMSUNG 0x0419
1056	1057	#define USB_DEVICE_ID_SAMSUNG_IR_REMOTE 0x0001
@@ -1060,6 +1061,9 @@ · 3 unmodified lines
1060	1061	#define USB_DEVICE_ID_SEMICO_USB_KEYKOARD 0x0023
1061	1062	#define USB_DEVICE_ID_SEMICO_USB_KEYKOARD2 0x0027
1062	1063
	1064	+#define USB_VENDOR_ID_SEMITEK 0x1ea7
	1065	+#define USB_DEVICE_ID_SEMITEK_KEYBOARD 0x0907
	1066	+
1063	1067	#define USB_VENDOR_ID_SENNHEISER 0x1395
1064	1068	#define USB_DEVICE_ID_SENNHEISER_BTD500USB 0x002c
1065	1069
@@ -1161,6 +1165,7 @@ · 95 unmodified lines
1161	1165	#define USB_DEVICE_ID_SYNAPTICS_DELL_K12A 0x2819
1162	1166	#define USB_DEVICE_ID_SYNAPTICS_ACER_SWITCH5_012 0x2968
1163	1167	#define USB_DEVICE_ID_SYNAPTICS_TP_V103 0x5710
	1168	+#define USB_DEVICE_ID_SYNAPTICS_DELL_K15A 0x6e21
1164	1169	#define USB_DEVICE_ID_SYNAPTICS_ACER_ONE_S1002 0x73f4
1165	1170	#define USB_DEVICE_ID_SYNAPTICS_ACER_ONE_S1003 0x73f5
1166	1171	#define USB_DEVICE_ID_SYNAPTICS_ACER_SWITCH5 0x81a7

drivers/hid/hid-input.c+3−0

@@ -964,6 +964,9 @@ static void hidinput_configure_usage(struct hid_input *hidinput, struct hid_fiel · 963 unmodified lines
964	964
965	965	case 0x0cd: map_key_clear(KEY_PLAYPAUSE); break;
966	966	case 0x0cf: map_key_clear(KEY_VOICECOMMAND); break;
	967	+
	968	+ case 0x0d9: map_key_clear(KEY_EMOJI_PICKER); break;
	969	+
967	970	case 0x0e0: map_abs_clear(ABS_VOLUME); break;
968	971	case 0x0e2: map_key_clear(KEY_MUTE); break;
969	972	case 0x0e5: map_key_clear(KEY_BASSBOOST); break;

drivers/hid/hid-logitech-hidpp.c+1−0

@@ -1263,6 +1263,7 @@ static int hidpp20_battery_map_status_voltage(u8 data[3], int *voltage, · 1262 unmodified lines
1263	1263	int status;
1264	1264
1265	1265	long flags = (long) data[2];
	1266	+ *level = POWER_SUPPLY_CAPACITY_LEVEL_UNKNOWN;
1266	1267
1267	1268	if (flags & 0x80)
1268	1269	switch (flags & 0x07) {

drivers/hid/hid-magicmouse.c+5−2

@@ -693,7 +693,7 @@ static int magicmouse_probe(struct hid_device *hdev, · 692 unmodified lines
693	693	if (id->vendor == USB_VENDOR_ID_APPLE &&
694	694	id->product == USB_DEVICE_ID_APPLE_MAGICTRACKPAD2 &&
695	695	hdev->type != HID_TYPE_USBMOUSE)
696		− return 0;
	696	+ return -ENODEV;
697	697
698	698	msc = devm_kzalloc(&hdev->dev, sizeof(*msc), GFP_KERNEL);
699	699	if (msc == NULL) {
@@ -779,7 +779,10 @@ err_stop_hw: · 79 unmodified lines
779	779	static void magicmouse_remove(struct hid_device *hdev)
780	780	{
781	781	struct magicmouse_sc *msc = hid_get_drvdata(hdev);
782		− cancel_delayed_work_sync(&msc->work);
	782	+
	783	+ if (msc)
	784	+ cancel_delayed_work_sync(&msc->work);
	785	+
783	786	hid_hw_stop(hdev);
784	787	}
785	788

drivers/hid/hid-multitouch.c+37−9

@@ -70,6 +70,7 @@ MODULE_LICENSE("GPL"); · 69 unmodified lines
70	70	#define MT_QUIRK_WIN8_PTP_BUTTONS BIT(18)
71	71	#define MT_QUIRK_SEPARATE_APP_REPORT BIT(19)
72	72	#define MT_QUIRK_FORCE_MULTI_INPUT BIT(20)
	73	+#define MT_QUIRK_DISABLE_WAKEUP BIT(21)
73	74
74	75	#define MT_INPUTMODE_TOUCHSCREEN 0x02
75	76	#define MT_INPUTMODE_TOUCHPAD 0x03
@@ -191,6 +192,7 @@ static void mt_post_parse(struct mt_device td, struct mt_application app); · 115 unmodified lines
191	192	#define MT_CLS_EXPORT_ALL_INPUTS 0x0013
192	193	/* reserved 0x0014 */
193	194	#define MT_CLS_WIN_8_FORCE_MULTI_INPUT 0x0015
	195	+#define MT_CLS_WIN_8_DISABLE_WAKEUP 0x0016
194	196
195	197	/* vendor specific classes */
196	198	#define MT_CLS_3M 0x0101
@@ -283,6 +285,15 @@ static const struct mt_class mt_classes[] = { · 86 unmodified lines
283	285	MT_QUIRK_WIN8_PTP_BUTTONS \|
284	286	MT_QUIRK_FORCE_MULTI_INPUT,
285	287	.export_all_inputs = true },
	288	+ { .name = MT_CLS_WIN_8_DISABLE_WAKEUP,
	289	+ .quirks = MT_QUIRK_ALWAYS_VALID \|
	290	+ MT_QUIRK_IGNORE_DUPLICATES \|
	291	+ MT_QUIRK_HOVERING \|
	292	+ MT_QUIRK_CONTACT_CNT_ACCURATE \|
	293	+ MT_QUIRK_STICKY_FINGERS \|
	294	+ MT_QUIRK_WIN8_PTP_BUTTONS \|
	295	+ MT_QUIRK_DISABLE_WAKEUP,
	296	+ .export_all_inputs = true },
286	297
287	298	/*
288	299	* vendor specific classes
@@ -604,9 +615,13 @@ static struct mt_report_data mt_allocate_report_data(struct mt_device td, · 315 unmodified lines
604	615	if (!(HID_MAIN_ITEM_VARIABLE & field->flags))
605	616	continue;
606	617
607		− for (n = 0; n < field->report_count; n++) {
608		− if (field->usage[n].hid == HID_DG_CONTACTID)
609		− rdata->is_mt_collection = true;
	618	+ if (field->logical == HID_DG_FINGER \|\| td->hdev->group != HID_GROUP_MULTITOUCH_WIN_8) {
	619	+ for (n = 0; n < field->report_count; n++) {
	620	+ if (field->usage[n].hid == HID_DG_CONTACTID) {
	621	+ rdata->is_mt_collection = true;
	622	+ break;
	623	+ }
	624	+ }
610	625	}
611	626	}
612	627
@@ -759,7 +774,8 @@ static int mt_touch_input_mapping(struct hid_device hdev, struct hid_input hi, · 146 unmodified lines
759	774	return 1;
760	775	case HID_DG_CONFIDENCE:
761	776	if ((cls->name == MT_CLS_WIN_8 \|\|
762		− cls->name == MT_CLS_WIN_8_FORCE_MULTI_INPUT) &&
	777	+ cls->name == MT_CLS_WIN_8_FORCE_MULTI_INPUT \|\|
	778	+ cls->name == MT_CLS_WIN_8_DISABLE_WAKEUP) &&
763	779	(field->application == HID_DG_TOUCHPAD \|\|
764	780	field->application == HID_DG_TOUCHSCREEN))
765	781	app->quirks \|= MT_QUIRK_CONFIDENCE;
@@ -1576,13 +1592,13 @@ static int mt_input_configured(struct hid_device hdev, struct hid_input hi) · 810 unmodified lines
1576	1592	/* we do not set suffix = "Touchscreen" */
1577	1593	hi->input->name = hdev->name;
1578	1594	break;
1579		− case HID_DG_STYLUS:
1580		− /* force BTN_STYLUS to allow tablet matching in udev */
1581		− __set_bit(BTN_STYLUS, hi->input->keybit);
1582		− break;
1583	1595	case HID_VD_ASUS_CUSTOM_MEDIA_KEYS:
1584	1596	suffix = "Custom Media Keys";
1585	1597	break;
	1598	+ case HID_DG_STYLUS:
	1599	+ /* force BTN_STYLUS to allow tablet matching in udev */
	1600	+ __set_bit(BTN_STYLUS, hi->input->keybit);
	1601	+ fallthrough;
1586	1602	case HID_DG_PEN:
1587	1603	suffix = "Stylus";
1588	1604	break;
@@ -1749,8 +1765,14 @@ static int mt_probe(struct hid_device hdev, const struct hid_device_id id) · 160 unmodified lines
1749	1765	#ifdef CONFIG_PM
1750	1766	static int mt_suspend(struct hid_device *hdev, pm_message_t state)
1751	1767	{
	1768	+ struct mt_device *td = hid_get_drvdata(hdev);
	1769	+
1752	1770	/* High latency is desirable for power savings during S3/S0ix */
1753		− mt_set_modes(hdev, HID_LATENCY_HIGH, true, true);
	1771	+ if (td->mtclass.quirks & MT_QUIRK_DISABLE_WAKEUP)
	1772	+ mt_set_modes(hdev, HID_LATENCY_HIGH, false, false);
	1773	+ else
	1774	+ mt_set_modes(hdev, HID_LATENCY_HIGH, true, true);
	1775	+
1754	1776	return 0;
1755	1777	}
1756	1778
@@ -1809,6 +1831,12 @@ static const struct hid_device_id mt_devices[] = { · 52 unmodified lines
1809	1831	MT_USB_DEVICE(USB_VENDOR_ID_ANTON,
1810	1832	USB_DEVICE_ID_ANTON_TOUCH_PAD) },
1811	1833
	1834	+ /* Asus T101HA */
	1835	+ { .driver_data = MT_CLS_WIN_8_DISABLE_WAKEUP,
	1836	+ HID_DEVICE(BUS_USB, HID_GROUP_MULTITOUCH_WIN_8,
	1837	+ USB_VENDOR_ID_ASUSTEK,
	1838	+ USB_DEVICE_ID_ASUSTEK_T101HA_KEYBOARD) },
	1839	+
1812	1840	/* Asus T304UA */
1813	1841	{ .driver_data = MT_CLS_ASUS,
1814	1842	HID_DEVICE(BUS_USB, HID_GROUP_MULTITOUCH_WIN_8,

drivers/hid/hid-quirks.c+4−0

@@ -110,6 +110,7 @@ static const struct hid_device_id hid_quirks[] = { · 109 unmodified lines
110	110	{ HID_USB_DEVICE(USB_VENDOR_ID_KYE, USB_DEVICE_ID_KYE_PENSKETCH_M912), HID_QUIRK_MULTI_INPUT },
111	111	{ HID_USB_DEVICE(USB_VENDOR_ID_KYE, USB_DEVICE_ID_KYE_EASYPEN_M406XE), HID_QUIRK_MULTI_INPUT },
112	112	{ HID_USB_DEVICE(USB_VENDOR_ID_KYE, USB_DEVICE_ID_PIXART_USB_OPTICAL_MOUSE_ID2), HID_QUIRK_ALWAYS_POLL },
	113	+ { HID_USB_DEVICE(USB_VENDOR_ID_LENOVO, USB_DEVICE_ID_LENOVO_OPTICAL_USB_MOUSE_600E), HID_QUIRK_ALWAYS_POLL },
113	114	{ HID_USB_DEVICE(USB_VENDOR_ID_LENOVO, USB_DEVICE_ID_LENOVO_PIXART_USB_MOUSE_608D), HID_QUIRK_ALWAYS_POLL },
114	115	{ HID_USB_DEVICE(USB_VENDOR_ID_LENOVO, USB_DEVICE_ID_LENOVO_PIXART_USB_MOUSE_6019), HID_QUIRK_ALWAYS_POLL },
115	116	{ HID_USB_DEVICE(USB_VENDOR_ID_LENOVO, USB_DEVICE_ID_LENOVO_PIXART_USB_MOUSE_602E), HID_QUIRK_ALWAYS_POLL },
@@ -158,6 +159,7 @@ static const struct hid_device_id hid_quirks[] = { · 42 unmodified lines
158	159	{ HID_USB_DEVICE(USB_VENDOR_ID_SAITEK, USB_DEVICE_ID_SAITEK_X52), HID_QUIRK_INCREMENT_USAGE_ON_DUPLICATE },
159	160	{ HID_USB_DEVICE(USB_VENDOR_ID_SAITEK, USB_DEVICE_ID_SAITEK_X52_2), HID_QUIRK_INCREMENT_USAGE_ON_DUPLICATE },
160	161	{ HID_USB_DEVICE(USB_VENDOR_ID_SAITEK, USB_DEVICE_ID_SAITEK_X52_PRO), HID_QUIRK_INCREMENT_USAGE_ON_DUPLICATE },
	162	+ { HID_USB_DEVICE(USB_VENDOR_ID_SAITEK, USB_DEVICE_ID_SAITEK_X65), HID_QUIRK_INCREMENT_USAGE_ON_DUPLICATE },
161	163	{ HID_USB_DEVICE(USB_VENDOR_ID_SEMICO, USB_DEVICE_ID_SEMICO_USB_KEYKOARD2), HID_QUIRK_NO_INIT_REPORTS },
162	164	{ HID_USB_DEVICE(USB_VENDOR_ID_SEMICO, USB_DEVICE_ID_SEMICO_USB_KEYKOARD), HID_QUIRK_NO_INIT_REPORTS },
163	165	{ HID_USB_DEVICE(USB_VENDOR_ID_SENNHEISER, USB_DEVICE_ID_SENNHEISER_BTD500USB), HID_QUIRK_NOGET },
@@ -176,6 +178,7 @@ static const struct hid_device_id hid_quirks[] = { · 12 unmodified lines
176	178	{ HID_USB_DEVICE(USB_VENDOR_ID_SYNAPTICS, USB_DEVICE_ID_SYNAPTICS_QUAD_HD), HID_QUIRK_NO_INIT_REPORTS },
177	179	{ HID_USB_DEVICE(USB_VENDOR_ID_SYNAPTICS, USB_DEVICE_ID_SYNAPTICS_TP_V103), HID_QUIRK_NO_INIT_REPORTS },
178	180	{ HID_USB_DEVICE(USB_VENDOR_ID_SYNAPTICS, USB_DEVICE_ID_SYNAPTICS_DELL_K12A), HID_QUIRK_NO_INIT_REPORTS },
	181	+ { HID_USB_DEVICE(USB_VENDOR_ID_SYNAPTICS, USB_DEVICE_ID_SYNAPTICS_DELL_K15A), HID_QUIRK_NO_INIT_REPORTS },
179	182	{ HID_USB_DEVICE(USB_VENDOR_ID_TOPMAX, USB_DEVICE_ID_TOPMAX_COBRAPAD), HID_QUIRK_BADPAD },
180	183	{ HID_USB_DEVICE(USB_VENDOR_ID_TOUCHPACK, USB_DEVICE_ID_TOUCHPACK_RTS), HID_QUIRK_MULTI_INPUT },
181	184	{ HID_USB_DEVICE(USB_VENDOR_ID_TPV, USB_DEVICE_ID_TPV_OPTICAL_TOUCHSCREEN_8882), HID_QUIRK_NOGET },
@@ -211,6 +214,7 @@ static const struct hid_device_id hid_have_special_driver[] = { · 29 unmodified lines
211	214	{ HID_USB_DEVICE(USB_VENDOR_ID_A4TECH, USB_DEVICE_ID_A4TECH_WCP32PU) },
212	215	{ HID_USB_DEVICE(USB_VENDOR_ID_A4TECH, USB_DEVICE_ID_A4TECH_X5_005D) },
213	216	{ HID_USB_DEVICE(USB_VENDOR_ID_A4TECH, USB_DEVICE_ID_A4TECH_RP_649) },
	217	+ { HID_USB_DEVICE(USB_VENDOR_ID_A4TECH, USB_DEVICE_ID_A4TECH_NB_95) },
214	218	#endif
215	219	#if IS_ENABLED(CONFIG_HID_ACCUTOUCH)
216	220	{ HID_USB_DEVICE(USB_VENDOR_ID_ELO, USB_DEVICE_ID_ELO_ACCUTOUCH_2216) },

drivers/hid/hid-semitek.c+40−0

@@ -0,0 +1,40 @@
	1	+// SPDX-License-Identifier: GPL-2.0-or-later
	2	+/*
	3	+ * HID driver for Semitek keyboards
	4	+ *
	5	+ * Copyright (c) 2021 Benjamin Moody
	6	+ */
	7	+
	8	+#include <linux/device.h>
	9	+#include <linux/hid.h>
	10	+#include <linux/module.h>
	11	+
	12	+#include "hid-ids.h"
	13	+
	14	+static __u8 semitek_report_fixup(struct hid_device hdev, __u8 *rdesc,
	15	+ unsigned int *rsize)
	16	+{
	17	+ /* In the report descriptor for interface 2, fix the incorrect
	18	+ description of report ID 0x04 (the report contains a
	19	+ bitmask, not an array of keycodes.) */
	20	+ if (*rsize == 0xcb && rdesc[0x83] == 0x81 && rdesc[0x84] == 0x00) {
	21	+ hid_info(hdev, "fixing up Semitek report descriptor\n");
	22	+ rdesc[0x84] = 0x02;
	23	+ }
	24	+ return rdesc;
	25	+}
	26	+
	27	+static const struct hid_device_id semitek_devices[] = {
	28	+ { HID_USB_DEVICE(USB_VENDOR_ID_SEMITEK, USB_DEVICE_ID_SEMITEK_KEYBOARD) },
	29	+ { }
	30	+};
	31	+MODULE_DEVICE_TABLE(hid, semitek_devices);
	32	+
	33	+static struct hid_driver semitek_driver = {
	34	+ .name = "semitek",
	35	+ .id_table = semitek_devices,
	36	+ .report_fixup = semitek_report_fixup,
	37	+};
	38	+module_hid_driver(semitek_driver);
	39	+
	40	+MODULE_LICENSE("GPL");

drivers/hid/hid-sensor-custom.c+5−3

@@ -387,7 +387,7 @@ static ssize_t store_value(struct device dev, struct device_attribute attr, · 386 unmodified lines
387	387	struct hid_sensor_custom *sensor_inst = dev_get_drvdata(dev);
388	388	int index, field_index, usage;
389	389	char name[HID_CUSTOM_NAME_LENGTH];
390		− int value;
	390	+ int value, ret;
391	391
392	392	if (sscanf(attr->attr.name, "feature-%x-%x-%s", &index, &usage,
393	393	name) == 3) {
@@ -403,8 +403,10 @@ static ssize_t store_value(struct device dev, struct device_attribute attr, · 9 unmodified lines
403	403
404	404	report_id = sensor_inst->fields[field_index].attribute.
405	405	report_id;
406		− sensor_hub_set_feature(sensor_inst->hsdev, report_id,
407		− index, sizeof(value), &value);
	406	+ ret = sensor_hub_set_feature(sensor_inst->hsdev, report_id,
	407	+ index, sizeof(value), &value);
	408	+ if (ret)
	409	+ return ret;
408	410	} else
409	411	return -EINVAL;
410	412

drivers/hid/hid-sensor-hub.c+9−4

@@ -209,16 +209,21 @@ int sensor_hub_set_feature(struct hid_sensor_hub_device *hsdev, u32 report_id, · 208 unmodified lines
209	209	buffer_size = buffer_size / sizeof(__s32);
210	210	if (buffer_size) {
211	211	for (i = 0; i < buffer_size; ++i) {
212		− hid_set_field(report->field[field_index], i,
213		− (__force __s32)cpu_to_le32(*buf32));
	212	+ ret = hid_set_field(report->field[field_index], i,
	213	+ (__force __s32)cpu_to_le32(*buf32));
	214	+ if (ret)
	215	+ goto done_proc;
	216	+
214	217	++buf32;
215	218	}
216	219	}
217	220	if (remaining_bytes) {
218	221	value = 0;
219	222	memcpy(&value, (u8 *)buf32, remaining_bytes);
220		− hid_set_field(report->field[field_index], i,
221		− (__force __s32)cpu_to_le32(value));
	223	+ ret = hid_set_field(report->field[field_index], i,
	224	+ (__force __s32)cpu_to_le32(value));
	225	+ if (ret)
	226	+ goto done_proc;
222	227	}
223	228	hid_hw_request(hsdev->hdev, report, HID_REQ_SET_REPORT);
224	229	hid_hw_wait(hsdev->hdev);

drivers/hid/hid-thrustmaster.c+1−1

@@ -312,7 +312,7 @@ static int thrustmaster_probe(struct hid_device *hdev, const struct hid_device_i · 311 unmodified lines
312	312	}
313	313
314	314	tm_wheel->change_request = kzalloc(sizeof(struct usb_ctrlrequest), GFP_KERNEL);
315		− if (!tm_wheel->model_request) {
	315	+ if (!tm_wheel->change_request) {
316	316	ret = -ENOMEM;
317	317	goto error5;
318	318	}

drivers/hid/i2c-hid/i2c-hid-core.c+10−3

@@ -45,6 +45,7 @@ · 44 unmodified lines
45	45	#define I2C_HID_QUIRK_BOGUS_IRQ BIT(4)
46	46	#define I2C_HID_QUIRK_RESET_ON_RESUME BIT(5)
47	47	#define I2C_HID_QUIRK_BAD_INPUT_SIZE BIT(6)
	48	+#define I2C_HID_QUIRK_NO_WAKEUP_AFTER_RESET BIT(7)
48	49
49	50
50	51	/* flags */
@@ -178,6 +179,11 @@ static const struct i2c_hid_quirks { · 127 unmodified lines
178	179	I2C_HID_QUIRK_RESET_ON_RESUME },
179	180	{ USB_VENDOR_ID_ITE, I2C_DEVICE_ID_ITE_LENOVO_LEGION_Y720,
180	181	I2C_HID_QUIRK_BAD_INPUT_SIZE },
	182	+ /*
	183	+ * Sending the wakeup after reset actually break ELAN touchscreen controller
	184	+ */
	185	+ { USB_VENDOR_ID_ELAN, HID_ANY_ID,
	186	+ I2C_HID_QUIRK_NO_WAKEUP_AFTER_RESET },
181	187	{ 0, 0 }
182	188	};
183	189
@@ -461,7 +467,8 @@ static int i2c_hid_hwreset(struct i2c_client *client) · 277 unmodified lines
461	467	}
462	468
463	469	/* At least some SIS devices need this after reset */
464		− ret = i2c_hid_set_power(client, I2C_HID_PWR_ON);
	470	+ if (!(ihid->quirks & I2C_HID_QUIRK_NO_WAKEUP_AFTER_RESET))
	471	+ ret = i2c_hid_set_power(client, I2C_HID_PWR_ON);
465	472
466	473	out_unlock:
467	474	mutex_unlock(&ihid->reset_lock);
@@ -990,8 +997,8 @@ int i2c_hid_core_probe(struct i2c_client client, struct i2chid_ops ops, · 522 unmodified lines
990	997	hid->vendor = le16_to_cpu(ihid->hdesc.wVendorID);
991	998	hid->product = le16_to_cpu(ihid->hdesc.wProductID);
992	999
993		− snprintf(hid->name, sizeof(hid->name), "%s %04hX:%04hX",
994		− client->name, hid->vendor, hid->product);
	1000	+ snprintf(hid->name, sizeof(hid->name), "%s %04X:%04X",
	1001	+ client->name, (u16)hid->vendor, (u16)hid->product);
995	1002	strlcpy(hid->phys, dev_name(&client->dev), sizeof(hid->phys));
996	1003
997	1004	ihid->quirks = i2c_hid_lookup_quirk(hid->vendor, hid->product);

drivers/hid/intel-ish-hid/ipc/hw-ish.h+2−0

@@ -28,6 +28,8 @@ · 27 unmodified lines
28	28	#define EHL_Ax_DEVICE_ID 0x4BB3
29	29	#define TGL_LP_DEVICE_ID 0xA0FC
30	30	#define TGL_H_DEVICE_ID 0x43FC
	31	+#define ADL_S_DEVICE_ID 0x7AF8
	32	+#define ADL_P_DEVICE_ID 0x51FC
31	33
32	34	#define REVISION_ID_CHT_A0 0x6
33	35	#define REVISION_ID_CHT_Ax_SI 0x0

drivers/hid/intel-ish-hid/ipc/pci-ish.c+2−0

@@ -39,6 +39,8 @@ static const struct pci_device_id ish_pci_tbl[] = { · 38 unmodified lines
39	39	{PCI_DEVICE(PCI_VENDOR_ID_INTEL, EHL_Ax_DEVICE_ID)},
40	40	{PCI_DEVICE(PCI_VENDOR_ID_INTEL, TGL_LP_DEVICE_ID)},
41	41	{PCI_DEVICE(PCI_VENDOR_ID_INTEL, TGL_H_DEVICE_ID)},
	42	+ {PCI_DEVICE(PCI_VENDOR_ID_INTEL, ADL_S_DEVICE_ID)},
	43	+ {PCI_DEVICE(PCI_VENDOR_ID_INTEL, ADL_P_DEVICE_ID)},
42	44	{0, }
43	45	};
44	46	MODULE_DEVICE_TABLE(pci, ish_pci_tbl);

drivers/hid/surface-hid/surface_hid_core.c+3−3

@@ -168,9 +168,9 @@ int surface_hid_device_add(struct surface_hid_device *shid) · 167 unmodified lines
168	168
169	169	shid->hid->dev.parent = shid->dev;
170	170	shid->hid->bus = BUS_HOST;
171		− shid->hid->vendor = cpu_to_le16(shid->attrs.vendor);
172		− shid->hid->product = cpu_to_le16(shid->attrs.product);
173		− shid->hid->version = cpu_to_le16(shid->hid_desc.hid_version);
	171	+ shid->hid->vendor = get_unaligned_le16(&shid->attrs.vendor);
	172	+ shid->hid->product = get_unaligned_le16(&shid->attrs.product);
	173	+ shid->hid->version = get_unaligned_le16(&shid->hid_desc.hid_version);
174	174	shid->hid->country = shid->hid_desc.country_code;
175	175
176	176	snprintf(shid->hid->name, sizeof(shid->hid->name), "Microsoft Surface %04X:%04X",

drivers/hid/usbhid/hid-core.c+1−1

@@ -374,7 +374,7 @@ static int hid_submit_ctrl(struct hid_device *hid) · 373 unmodified lines
374	374	raw_report = usbhid->ctrl[usbhid->ctrltail].raw_report;
375	375	dir = usbhid->ctrl[usbhid->ctrltail].dir;
376	376
377		− len = ((report->size - 1) >> 3) + 1 + (report->id > 0);
	377	+ len = hid_report_len(report);
378	378	if (dir == USB_DIR_OUT) {
379	379	usbhid->urbctrl->pipe = usb_sndctrlpipe(hid_to_usb_dev(hid), 0);
380	380	usbhid->urbctrl->transfer_buffer_length = len;

drivers/hid/usbhid/hid-pidff.c+1−0

@@ -1292,6 +1292,7 @@ int hid_pidff_init(struct hid_device *hid) · 1291 unmodified lines
1292	1292
1293	1293	if (pidff->pool[PID_DEVICE_MANAGED_POOL].value &&
1294	1294	pidff->pool[PID_DEVICE_MANAGED_POOL].value[0] == 0) {
	1295	+ error = -EPERM;
1295	1296	hid_notice(hid,
1296	1297	"device does not support device managed pool\n");
1297	1298	goto fail;

drivers/hwmon/corsair-psu.c+14−0

@@ -771,6 +771,16 @@ static int corsairpsu_raw_event(struct hid_device hdev, struct hid_report repo · 770 unmodified lines
771	771	return 0;
772	772	}
773	773
	774	+#ifdef CONFIG_PM
	775	+static int corsairpsu_resume(struct hid_device *hdev)
	776	+{
	777	+ struct corsairpsu_data *priv = hid_get_drvdata(hdev);
	778	+
	779	+ /* some PSUs turn off the microcontroller during standby, so a reinit is required */
	780	+ return corsairpsu_init(priv);
	781	+}
	782	+#endif
	783	+
774	784	static const struct hid_device_id corsairpsu_idtable[] = {
775	785	{ HID_USB_DEVICE(0x1b1c, 0x1c03) }, /* Corsair HX550i */
776	786	{ HID_USB_DEVICE(0x1b1c, 0x1c04) }, /* Corsair HX650i */
@@ -793,6 +803,10 @@ static struct hid_driver corsairpsu_driver = { · 16 unmodified lines
793	803	.probe = corsairpsu_probe,
794	804	.remove = corsairpsu_remove,
795	805	.raw_event = corsairpsu_raw_event,
	806	+#ifdef CONFIG_PM
	807	+ .resume = corsairpsu_resume,
	808	+ .reset_resume = corsairpsu_resume,
	809	+#endif
796	810	};
797	811	module_hid_driver(corsairpsu_driver);
798	812

drivers/hwmon/dell-smm-hwmon.c+2−2

@@ -838,10 +838,10 @@ static struct attribute *i8k_attrs[] = { · 837 unmodified lines
838	838	static umode_t i8k_is_visible(struct kobject kobj, struct attribute attr,
839	839	int index)
840	840	{
841		− if (disallow_fan_support && index >= 8)
	841	+ if (disallow_fan_support && index >= 20)
842	842	return 0;
843	843	if (disallow_fan_type_call &&
844		− (index == 9 \|\| index == 12 \|\| index == 15))
	844	+ (index == 21 \|\| index == 25 \|\| index == 28))
845	845	return 0;
846	846	if (index >= 0 && index <= 1 &&
847	847	!(i8k_hwmon_flags & I8K_HWMON_HAVE_TEMP1))

drivers/hwmon/pmbus/fsp-3y.c+25−7

@@ -37,6 +37,8 @@ struct fsp3y_data { · 36 unmodified lines
37	37	struct pmbus_driver_info info;
38	38	int chip;
39	39	int page;
	40	+
	41	+ bool vout_linear_11;
40	42	};
41	43
42	44	#define to_fsp3y_data(x) container_of(x, struct fsp3y_data, info)
@@ -108,11 +110,9 @@ static int fsp3y_read_byte_data(struct i2c_client *client, int page, int reg) · 65 unmodified lines
108	110	int rv;
109	111
110	112	/*
111		− * YH5151-E outputs vout in linear11. The conversion is done when
112		− * reading. Here, we have to inject pmbus_core with the correct
113		− * exponent (it is -6).
	113	+ * Inject an exponent for non-compliant YH5151-E.
114	114	*/
115		− if (data->chip == yh5151e && reg == PMBUS_VOUT_MODE)
	115	+ if (data->vout_linear_11 && reg == PMBUS_VOUT_MODE)
116	116	return 0x1A;
117	117
118	118	rv = set_page(client, page);
@@ -161,10 +161,9 @@ static int fsp3y_read_word_data(struct i2c_client *client, int page, int phase, · 42 unmodified lines
161	161	return rv;
162	162
163	163	/*
164		− * YH-5151E is non-compliant and outputs output voltages in linear11
165		− * instead of linear16.
	164	+ * Handle YH-5151E non-compliant linear11 vout voltage.
166	165	*/
167		− if (data->chip == yh5151e && reg == PMBUS_READ_VOUT)
	166	+ if (data->vout_linear_11 && reg == PMBUS_READ_VOUT)
168	167	rv = sign_extend32(rv, 10) & 0xffff;
169	168
170	169	return rv;
@@ -256,6 +255,25 @@ static int fsp3y_probe(struct i2c_client *client) · 85 unmodified lines
256	255
257	256	data->info = fsp3y_info[data->chip];
258	257
	258	+ /*
	259	+ * YH-5151E sometimes reports vout in linear11 and sometimes in
	260	+ * linear16. This depends on the exact individual piece of hardware. One
	261	+ * YH-5151E can use linear16 and another might use linear11 instead.
	262	+ *
	263	+ * The format can be recognized by reading VOUT_MODE - if it doesn't
	264	+ * report a valid exponent, then vout uses linear11. Otherwise, the
	265	+ * device is compliant and uses linear16.
	266	+ */
	267	+ data->vout_linear_11 = false;
	268	+ if (data->chip == yh5151e) {
	269	+ rv = i2c_smbus_read_byte_data(client, PMBUS_VOUT_MODE);
	270	+ if (rv < 0)
	271	+ return rv;
	272	+
	273	+ if (rv == 0xFF)
	274	+ data->vout_linear_11 = true;
	275	+ }
	276	+
259	277	return pmbus_do_probe(client, &data->info);
260	278	}
261	279

drivers/hwmon/pmbus/isl68137.c+2−2

@@ -244,8 +244,8 @@ static int isl68137_probe(struct i2c_client *client) · 243 unmodified lines
244	244	info->read_word_data = raa_dmpvr2_read_word_data;
245	245	break;
246	246	case raa_dmpvr2_2rail_nontc:
247		− info->func[0] &= ~PMBUS_HAVE_TEMP;
248		− info->func[1] &= ~PMBUS_HAVE_TEMP;
	247	+ info->func[0] &= ~PMBUS_HAVE_TEMP3;
	248	+ info->func[1] &= ~PMBUS_HAVE_TEMP3;
249	249	fallthrough;
250	250	case raa_dmpvr2_2rail:
251	251	info->pages = 2;

drivers/hwmon/pmbus/q54sj108a2.c+1−1

@@ -299,7 +299,7 @@ static int q54sj108a2_probe(struct i2c_client *client) · 298 unmodified lines
299	299	dev_err(&client->dev, "Failed to read Manufacturer ID\n");
300	300	return ret;
301	301	}
302		− if (ret != 5 \|\| strncmp(buf, "DELTA", 5)) {
	302	+ if (ret != 6 \|\| strncmp(buf, "DELTA", 5)) {
303	303	buf[ret] = '\0';
304	304	dev_err(dev, "Unsupported Manufacturer ID '%s'\n", buf);
305	305	return -ENODEV;

drivers/hwmon/scpi-hwmon.c+9−0

@@ -99,6 +99,15 @@ scpi_show_sensor(struct device dev, struct device_attribute attr, char *buf) · 98 unmodified lines
99	99
100	100	scpi_scale_reading(&value, sensor);
101	101
	102	+ /*
	103	+ * Temperature sensor values are treated as signed values based on
	104	+ * observation even though that is not explicitly specified, and
	105	+ * because an unsigned u64 temperature does not really make practical
	106	+ * sense especially when the temperature is below zero degrees Celsius.
	107	+ */
	108	+ if (sensor->info.class == TEMPERATURE)
	109	+ return sprintf(buf, "%lld\n", (s64)value);
	110	+
102	111	return sprintf(buf, "%llu\n", value);
103	112	}
104	113

drivers/hwmon/tps23861.c+15−2

@@ -99,11 +99,14 @@ · 98 unmodified lines
99	99	#define POWER_ENABLE 0x19
100	100	#define TPS23861_NUM_PORTS 4
101	101
	102	+#define TPS23861_GENERAL_MASK_1 0x17
	103	+#define TPS23861_CURRENT_SHUNT_MASK BIT(0)
	104	+
102	105	#define TEMPERATURE_LSB 652 /* 0.652 degrees Celsius */
103	106	#define VOLTAGE_LSB 3662 /* 3.662 mV */
104	107	#define SHUNT_RESISTOR_DEFAULT 255000 /* 255 mOhm */
105		−#define CURRENT_LSB_255 62260 /* 62.260 uA */
106		−#define CURRENT_LSB_250 61039 /* 61.039 uA */
	108	+#define CURRENT_LSB_250 62260 /* 62.260 uA */
	109	+#define CURRENT_LSB_255 61039 /* 61.039 uA */
107	110	#define RESISTANCE_LSB 110966 /* 11.0966 Ohm*/
108	111	#define RESISTANCE_LSB_LOW 157216 /* 15.7216 Ohm*/
109	112
@@ -117,6 +120,7 @@ struct tps23861_data { · 7 unmodified lines
117	120	static struct regmap_config tps23861_regmap_config = {
118	121	.reg_bits = 8,
119	122	.val_bits = 8,
	123	+ .max_register = 0x6f,
120	124	};
121	125
122	126	static int tps23861_read_temp(struct tps23861_data data, long val)
@@ -560,6 +564,15 @@ static int tps23861_probe(struct i2c_client *client) · 437 unmodified lines
560	564	else
561	565	data->shunt_resistor = SHUNT_RESISTOR_DEFAULT;
562	566
	567	+ if (data->shunt_resistor == SHUNT_RESISTOR_DEFAULT)
	568	+ regmap_clear_bits(data->regmap,
	569	+ TPS23861_GENERAL_MASK_1,
	570	+ TPS23861_CURRENT_SHUNT_MASK);
	571	+ else
	572	+ regmap_set_bits(data->regmap,
	573	+ TPS23861_GENERAL_MASK_1,
	574	+ TPS23861_CURRENT_SHUNT_MASK);
	575	+
563	576	hwmon_dev = devm_hwmon_device_register_with_info(dev, client->name,
564	577	data, &tps23861_chip_info,
565	578	NULL);

drivers/i2c/busses/Kconfig+1−1

@@ -647,7 +647,7 @@ config I2C_HIGHLANDER · 646 unmodified lines
647	647
648	648	config I2C_HISI
649	649	tristate "HiSilicon I2C controller"
650		− depends on ARM64 \|\| COMPILE_TEST
	650	+ depends on (ARM64 && ACPI) \|\| COMPILE_TEST
651	651	help
652	652	Say Y here if you want to have Hisilicon I2C controller support
653	653	available on the Kunpeng Server.

drivers/i2c/busses/i2c-ali1563.c+1−1

@@ -1,5 +1,5 @@
1	1	// SPDX-License-Identifier: GPL-2.0-only
2		−/**
	2	+/*
3	3	* i2c-ali1563.c - i2c driver for the ALi 1563 Southbridge
4	4	*
5	5	* Copyright (C) 2004 Patrick Mochel

drivers/i2c/busses/i2c-altera.c+4−5

@@ -55,7 +55,7 @@ · 54 unmodified lines
55	55	#define ALTR_I2C_XFER_TIMEOUT (msecs_to_jiffies(250))
56	56
57	57	/**
58		− * altr_i2c_dev - I2C device context
	58	+ * struct altr_i2c_dev - I2C device context
59	59	* @base: pointer to register struct
60	60	* @msg: pointer to current message
61	61	* @msg_len: number of bytes transferred in msg
@@ -172,7 +172,7 @@ static void altr_i2c_init(struct altr_i2c_dev *idev) · 110 unmodified lines
172	172	altr_i2c_int_enable(idev, ALTR_I2C_ALL_IRQ, false);
173	173	}
174	174
175		−/**
	175	+/*
176	176	* altr_i2c_transfer - On the last byte to be transmitted, send
177	177	* a Stop bit on the last byte.
178	178	*/
@@ -185,7 +185,7 @@ static void altr_i2c_transfer(struct altr_i2c_dev *idev, u32 data) · 6 unmodified lines
185	185	writel(data, idev->base + ALTR_I2C_TFR_CMD);
186	186	}
187	187
188		−/**
	188	+/*
189	189	* altr_i2c_empty_rx_fifo - Fetch data from RX FIFO until end of
190	190	* transfer. Send a Stop bit on the last byte.
191	191	*/
@@ -201,9 +201,8 @@ static void altr_i2c_empty_rx_fifo(struct altr_i2c_dev *idev) · 9 unmodified lines
201	201	}
202	202	}
203	203
204		−/**
	204	+/*
205	205	* altr_i2c_fill_tx_fifo - Fill TX FIFO from current message buffer.
206		− * @return: Number of bytes left to transfer.
207	206	*/
208	207	static int altr_i2c_fill_tx_fifo(struct altr_i2c_dev *idev)
209	208	{

drivers/i2c/busses/i2c-cadence.c+1−1

@@ -144,7 +144,7 @@ enum cdns_i2c_mode { · 143 unmodified lines
144	144	};
145	145
146	146	/**
147		− * enum cdns_i2c_slave_mode - Slave state when I2C is operating in slave mode
	147	+ * enum cdns_i2c_slave_state - Slave state when I2C is operating in slave mode
148	148	*
149	149	* @CDNS_I2C_SLAVE_STATE_IDLE: I2C slave idle
150	150	* @CDNS_I2C_SLAVE_STATE_SEND: I2C slave sending data to master

drivers/i2c/busses/i2c-designware-master.c+1−1

@@ -165,7 +165,7 @@ static int i2c_dw_set_timings_master(struct dw_i2c_dev *dev) · 164 unmodified lines
165	165	}
166	166
167	167	/**
168		− * i2c_dw_init() - Initialize the designware I2C master hardware
	168	+ * i2c_dw_init_master() - Initialize the designware I2C master hardware
169	169	* @dev: device private data
170	170	*
171	171	* This functions configures and enables the I2C master.

drivers/i2c/busses/i2c-eg20t.c+2−1

@@ -148,7 +148,7 @@ struct i2c_algo_pch_data { · 147 unmodified lines
148	148
149	149	/**
150	150	* struct adapter_info - This structure holds the adapter information for the
151		− PCH i2c controller
	151	+ * PCH i2c controller
152	152	* @pch_data: stores a list of i2c_algo_pch_data
153	153	* @pch_i2c_suspended: specifies whether the system is suspended or not
154	154	* perhaps with more lines and words.
@@ -358,6 +358,7 @@ static void pch_i2c_repstart(struct i2c_algo_pch_data *adap) · 203 unmodified lines
358	358	/**
359	359	* pch_i2c_writebytes() - write data to I2C bus in normal mode
360	360	* @i2c_adap: Pointer to the struct i2c_adapter.
	361	+ * @msgs: Pointer to the i2c message structure.
361	362	* @last: specifies whether last message or not.
362	363	* In the case of compound mode it will be 1 for last message,
363	364	* otherwise 0.

drivers/i2c/busses/i2c-i801.c+2−4

@@ -395,11 +395,9 @@ static int i801_check_post(struct i801_priv *priv, int status) · 394 unmodified lines
395	395	dev_err(&priv->pci_dev->dev, "Transaction timeout\n");
396	396	/* try to stop the current command */
397	397	dev_dbg(&priv->pci_dev->dev, "Terminating the current operation\n");
398		− outb_p(inb_p(SMBHSTCNT(priv)) \| SMBHSTCNT_KILL,
399		− SMBHSTCNT(priv));
	398	+ outb_p(SMBHSTCNT_KILL, SMBHSTCNT(priv));
400	399	usleep_range(1000, 2000);
401		− outb_p(inb_p(SMBHSTCNT(priv)) & (~SMBHSTCNT_KILL),
402		− SMBHSTCNT(priv));
	400	+ outb_p(0, SMBHSTCNT(priv));
403	401
404	402	/* Check if it worked */
405	403	status = inb_p(SMBHSTSTS(priv));

drivers/i2c/busses/i2c-icy.c+0−1

@@ -123,7 +123,6 @@ static int icy_probe(struct zorro_dev *z, · 122 unmodified lines
123	123	{
124	124	struct icy_i2c *i2c;
125	125	struct i2c_algo_pcf_data *algo_data;
126		− struct fwnode_handle *new_fwnode;
127	126	struct i2c_board_info ltc2990_info = {
128	127	.type = "ltc2990",
129	128	.swnode = &icy_ltc2990_node,

drivers/i2c/busses/i2c-mpc.c+79−2

@@ -19,6 +19,7 @@ · 18 unmodified lines
19	19
20	20	#include <linux/clk.h>
21	21	#include <linux/io.h>
	22	+#include <linux/iopoll.h>
22	23	#include <linux/fsl_devices.h>
23	24	#include <linux/i2c.h>
24	25	#include <linux/interrupt.h>
@@ -45,6 +46,7 @@ · 20 unmodified lines
45	46	#define CCR_MTX 0x10
46	47	#define CCR_TXAK 0x08
47	48	#define CCR_RSTA 0x04
	49	+#define CCR_RSVD 0x02
48	50
49	51	#define CSR_MCF 0x80
50	52	#define CSR_MAAS 0x40
@@ -97,7 +99,7 @@ struct mpc_i2c { · 46 unmodified lines
97	99	u32 block;
98	100	int rc;
99	101	int expect_rxack;
100		−
	102	+ bool has_errata_A004447;
101	103	};
102	104
103	105	struct mpc_i2c_divider {
@@ -136,6 +138,75 @@ static void mpc_i2c_fixup(struct mpc_i2c *i2c) · 32 unmodified lines
136	138	}
137	139	}
138	140
	141	+static int i2c_mpc_wait_sr(struct mpc_i2c *i2c, int mask)
	142	+{
	143	+ void __iomem *addr = i2c->base + MPC_I2C_SR;
	144	+ u8 val;
	145	+
	146	+ return readb_poll_timeout(addr, val, val & mask, 0, 100);
	147	+}
	148	+
	149	+/*
	150	+ * Workaround for Erratum A004447. From the P2040CE Rev Q
	151	+ *
	152	+ * 1. Set up the frequency divider and sampling rate.
	153	+ * 2. I2CCR - a0h
	154	+ * 3. Poll for I2CSR[MBB] to get set.
	155	+ * 4. If I2CSR[MAL] is set (an indication that SDA is stuck low), then go to
	156	+ * step 5. If MAL is not set, then go to step 13.
	157	+ * 5. I2CCR - 00h
	158	+ * 6. I2CCR - 22h
	159	+ * 7. I2CCR - a2h
	160	+ * 8. Poll for I2CSR[MBB] to get set.
	161	+ * 9. Issue read to I2CDR.
	162	+ * 10. Poll for I2CSR[MIF] to be set.
	163	+ * 11. I2CCR - 82h
	164	+ * 12. Workaround complete. Skip the next steps.
	165	+ * 13. Issue read to I2CDR.
	166	+ * 14. Poll for I2CSR[MIF] to be set.
	167	+ * 15. I2CCR - 80h
	168	+ */
	169	+static void mpc_i2c_fixup_A004447(struct mpc_i2c *i2c)
	170	+{
	171	+ int ret;
	172	+ u32 val;
	173	+
	174	+ writeccr(i2c, CCR_MEN \| CCR_MSTA);
	175	+ ret = i2c_mpc_wait_sr(i2c, CSR_MBB);
	176	+ if (ret) {
	177	+ dev_err(i2c->dev, "timeout waiting for CSR_MBB\n");
	178	+ return;
	179	+ }
	180	+
	181	+ val = readb(i2c->base + MPC_I2C_SR);
	182	+
	183	+ if (val & CSR_MAL) {
	184	+ writeccr(i2c, 0x00);
	185	+ writeccr(i2c, CCR_MSTA \| CCR_RSVD);
	186	+ writeccr(i2c, CCR_MEN \| CCR_MSTA \| CCR_RSVD);
	187	+ ret = i2c_mpc_wait_sr(i2c, CSR_MBB);
	188	+ if (ret) {
	189	+ dev_err(i2c->dev, "timeout waiting for CSR_MBB\n");
	190	+ return;
	191	+ }
	192	+ val = readb(i2c->base + MPC_I2C_DR);
	193	+ ret = i2c_mpc_wait_sr(i2c, CSR_MIF);
	194	+ if (ret) {
	195	+ dev_err(i2c->dev, "timeout waiting for CSR_MIF\n");
	196	+ return;
	197	+ }
	198	+ writeccr(i2c, CCR_MEN \| CCR_RSVD);
	199	+ } else {
	200	+ val = readb(i2c->base + MPC_I2C_DR);
	201	+ ret = i2c_mpc_wait_sr(i2c, CSR_MIF);
	202	+ if (ret) {
	203	+ dev_err(i2c->dev, "timeout waiting for CSR_MIF\n");
	204	+ return;
	205	+ }
	206	+ writeccr(i2c, CCR_MEN);
	207	+ }
	208	+}
	209	+
139	210	#if defined(CONFIG_PPC_MPC52xx) \|\| defined(CONFIG_PPC_MPC512x)
140	211	static const struct mpc_i2c_divider mpc_i2c_dividers_52xx[] = {
141	212	{20, 0x20}, {22, 0x21}, {24, 0x22}, {26, 0x23},
@@ -670,7 +741,10 @@ static int fsl_i2c_bus_recovery(struct i2c_adapter *adap) · 528 unmodified lines
670	741	{
671	742	struct mpc_i2c *i2c = i2c_get_adapdata(adap);
672	743
673		− mpc_i2c_fixup(i2c);
	744	+ if (i2c->has_errata_A004447)
	745	+ mpc_i2c_fixup_A004447(i2c);
	746	+ else
	747	+ mpc_i2c_fixup(i2c);
674	748
675	749	return 0;
676	750	}
@@ -767,6 +841,9 @@ static int fsl_i2c_probe(struct platform_device *op) · 90 unmodified lines
767	841	}
768	842	dev_info(i2c->dev, "timeout %u us\n", mpc_ops.timeout * 1000000 / HZ);
769	843
	844	+ if (of_property_read_bool(op->dev.of_node, "fsl,i2c-erratum-a004447"))
	845	+ i2c->has_errata_A004447 = true;
	846	+
770	847	i2c->adap = mpc_ops;
771	848	scnprintf(i2c->adap.name, sizeof(i2c->adap.name),
772	849	"MPC adapter (%s)", of_node_full_name(op->dev.of_node));

drivers/i2c/busses/i2c-mt65xx.c+5−0

@@ -479,6 +479,11 @@ static void mtk_i2c_clock_disable(struct mtk_i2c *i2c) · 478 unmodified lines
479	479	static void mtk_i2c_init_hw(struct mtk_i2c *i2c)
480	480	{
481	481	u16 control_reg;
	482	+ u16 intr_stat_reg;
	483	+
	484	+ mtk_i2c_writew(i2c, I2C_CHN_CLR_FLAG, OFFSET_START);
	485	+ intr_stat_reg = mtk_i2c_readw(i2c, OFFSET_INTR_STAT);
	486	+ mtk_i2c_writew(i2c, intr_stat_reg, OFFSET_INTR_STAT);
482	487
483	488	if (i2c->dev_comp->apdma_sync) {
484	489	writel(I2C_DMA_WARM_RST, i2c->pdmabase + OFFSET_RST);

drivers/i2c/busses/i2c-nomadik.c+1−1

@@ -159,7 +159,7 @@ struct i2c_nmk_client { · 158 unmodified lines
159	159	* @clk_freq: clock frequency for the operation mode
160	160	* @tft: Tx FIFO Threshold in bytes
161	161	* @rft: Rx FIFO Threshold in bytes
162		− * @timeout Slave response timeout (ms)
	162	+ * @timeout: Slave response timeout (ms)
163	163	* @sm: speed mode
164	164	* @stop: stop condition.
165	165	* @xfer_complete: acknowledge completion for a I2C message.

drivers/i2c/busses/i2c-ocores.c+4−4

@@ -250,7 +250,7 @@ static irqreturn_t ocores_isr(int irq, void *dev_id) · 249 unmodified lines
250	250	}
251	251
252	252	/**
253		− * Process timeout event
	253	+ * ocores_process_timeout() - Process timeout event
254	254	* @i2c: ocores I2C device instance
255	255	*/
256	256	static void ocores_process_timeout(struct ocores_i2c *i2c)
@@ -264,7 +264,7 @@ static void ocores_process_timeout(struct ocores_i2c *i2c) · 7 unmodified lines
264	264	}
265	265
266	266	/**
267		− * Wait until something change in a given register
	267	+ * ocores_wait() - Wait until something change in a given register
268	268	* @i2c: ocores I2C device instance
269	269	* @reg: register to query
270	270	* @mask: bitmask to apply on register value
@@ -296,7 +296,7 @@ static int ocores_wait(struct ocores_i2c *i2c, · 25 unmodified lines
296	296	}
297	297
298	298	/**
299		− * Wait until is possible to process some data
	299	+ * ocores_poll_wait() - Wait until is possible to process some data
300	300	* @i2c: ocores I2C device instance
301	301	*
302	302	* Used when the device is in polling mode (interrupts disabled).
@@ -334,7 +334,7 @@ static int ocores_poll_wait(struct ocores_i2c *i2c) · 31 unmodified lines
334	334	}
335	335
336	336	/**
337		− * It handles an IRQ-less transfer
	337	+ * ocores_process_polling() - It handles an IRQ-less transfer
338	338	* @i2c: ocores I2C device instance
339	339	*
340	340	* Even if IRQ are disabled, the I2C OpenCore IP behavior is exactly the same

drivers/i2c/busses/i2c-pnx.c+4−4

@@ -138,7 +138,7 @@ static inline void i2c_pnx_arm_timer(struct i2c_pnx_algo_data *alg_data) · 137 unmodified lines
138	138	/**
139	139	* i2c_pnx_start - start a device
140	140	* @slave_addr: slave address
141		− * @adap: pointer to adapter structure
	141	+ * @alg_data: pointer to local driver data structure
142	142	*
143	143	* Generate a START signal in the desired mode.
144	144	*/
@@ -194,7 +194,7 @@ static int i2c_pnx_start(unsigned char slave_addr, · 49 unmodified lines
194	194
195	195	/**
196	196	* i2c_pnx_stop - stop a device
197		− * @adap: pointer to I2C adapter structure
	197	+ * @alg_data: pointer to local driver data structure
198	198	*
199	199	* Generate a STOP signal to terminate the master transaction.
200	200	*/
@@ -223,7 +223,7 @@ static void i2c_pnx_stop(struct i2c_pnx_algo_data *alg_data) · 22 unmodified lines
223	223
224	224	/**
225	225	* i2c_pnx_master_xmit - transmit data to slave
226		− * @adap: pointer to I2C adapter structure
	226	+ * @alg_data: pointer to local driver data structure
227	227	*
228	228	* Sends one byte of data to the slave
229	229	*/
@@ -293,7 +293,7 @@ static int i2c_pnx_master_xmit(struct i2c_pnx_algo_data *alg_data) · 63 unmodified lines
293	293
294	294	/**
295	295	* i2c_pnx_master_rcv - receive data from slave
296		− * @adap: pointer to I2C adapter structure
	296	+ * @alg_data: pointer to local driver data structure
297	297	*
298	298	* Reads one byte data from the slave
299	299	*/

drivers/i2c/busses/i2c-qcom-geni.c+21−2

@@ -100,7 +100,7 @@ static const struct geni_i2c_err_log gi2c_log[] = { · 99 unmodified lines
100	100	[GP_IRQ0] = {-EIO, "Unknown I2C err GP_IRQ0"},
101	101	[NACK] = {-ENXIO, "NACK: slv unresponsive, check its power/reset-ln"},
102	102	[GP_IRQ2] = {-EIO, "Unknown I2C err GP IRQ2"},
103		− [BUS_PROTO] = {-EPROTO, "Bus proto err, noisy/unepxected start/stop"},
	103	+ [BUS_PROTO] = {-EPROTO, "Bus proto err, noisy/unexpected start/stop"},
104	104	[ARB_LOST] = {-EAGAIN, "Bus arbitration lost, clock line undriveable"},
105	105	[GP_IRQ5] = {-EIO, "Unknown I2C err GP IRQ5"},
106	106	[GENI_OVERRUN] = {-EIO, "Cmd overrun, check GENI cmd-state machine"},
@@ -650,6 +650,14 @@ static int geni_i2c_remove(struct platform_device *pdev) · 543 unmodified lines
650	650	return 0;
651	651	}
652	652
	653	+static void geni_i2c_shutdown(struct platform_device *pdev)
	654	+{
	655	+ struct geni_i2c_dev *gi2c = platform_get_drvdata(pdev);
	656	+
	657	+ /* Make client i2c transfers start failing */
	658	+ i2c_mark_adapter_suspended(&gi2c->adap);
	659	+}
	660	+
653	661	static int __maybe_unused geni_i2c_runtime_suspend(struct device *dev)
654	662	{
655	663	int ret;
@@ -690,6 +698,8 @@ static int __maybe_unused geni_i2c_suspend_noirq(struct device *dev) · 34 unmodified lines
690	698	{
691	699	struct geni_i2c_dev *gi2c = dev_get_drvdata(dev);
692	700
	701	+ i2c_mark_adapter_suspended(&gi2c->adap);
	702	+
693	703	if (!gi2c->suspended) {
694	704	geni_i2c_runtime_suspend(dev);
695	705	pm_runtime_disable(dev);
@@ -699,8 +709,16 @@ static int __maybe_unused geni_i2c_suspend_noirq(struct device *dev) · 3 unmodified lines
699	709	return 0;
700	710	}
701	711
	712	+static int __maybe_unused geni_i2c_resume_noirq(struct device *dev)
	713	+{
	714	+ struct geni_i2c_dev *gi2c = dev_get_drvdata(dev);
	715	+
	716	+ i2c_mark_adapter_resumed(&gi2c->adap);
	717	+ return 0;
	718	+}
	719	+
702	720	static const struct dev_pm_ops geni_i2c_pm_ops = {
703		− SET_NOIRQ_SYSTEM_SLEEP_PM_OPS(geni_i2c_suspend_noirq, NULL)
	721	+ SET_NOIRQ_SYSTEM_SLEEP_PM_OPS(geni_i2c_suspend_noirq, geni_i2c_resume_noirq)
704	722	SET_RUNTIME_PM_OPS(geni_i2c_runtime_suspend, geni_i2c_runtime_resume,
705	723	NULL)
706	724	};
@@ -714,6 +732,7 @@ MODULE_DEVICE_TABLE(of, geni_i2c_dt_match); · 7 unmodified lines
714	732	static struct platform_driver geni_i2c_driver = {
715	733	.probe = geni_i2c_probe,
716	734	.remove = geni_i2c_remove,
	735	+ .shutdown = geni_i2c_shutdown,
717	736	.driver = {
718	737	.name = "geni_i2c",
719	738	.pm = &geni_i2c_pm_ops,

drivers/i2c/busses/i2c-s3c2410.c+3−0

@@ -480,7 +480,10 @@ static int i2c_s3c_irq_nextbyte(struct s3c24xx_i2c *i2c, unsigned long iicstat) · 479 unmodified lines
480	480	* forces us to send a new START
481	481	* when we change direction
482	482	*/
	483	+ dev_dbg(i2c->dev,
	484	+ "missing START before write->read\n");
483	485	s3c24xx_i2c_stop(i2c, -EINVAL);
	486	+ break;
484	487	}
485	488
486	489	goto retry_write;

drivers/i2c/busses/i2c-sh_mobile.c+1−1

@@ -807,7 +807,7 @@ static const struct sh_mobile_dt_config r8a7740_dt_config = { · 806 unmodified lines
807	807	static const struct of_device_id sh_mobile_i2c_dt_ids[] = {
808	808	{ .compatible = "renesas,iic-r8a73a4", .data = &fast_clock_dt_config },
809	809	{ .compatible = "renesas,iic-r8a7740", .data = &r8a7740_dt_config },
810		− { .compatible = "renesas,iic-r8a774c0", .data = &fast_clock_dt_config },
	810	+ { .compatible = "renesas,iic-r8a774c0", .data = &v2_freq_calc_dt_config },
811	811	{ .compatible = "renesas,iic-r8a7790", .data = &v2_freq_calc_dt_config },
812	812	{ .compatible = "renesas,iic-r8a7791", .data = &v2_freq_calc_dt_config },
813	813	{ .compatible = "renesas,iic-r8a7792", .data = &v2_freq_calc_dt_config },

drivers/i2c/busses/i2c-st.c+2−2

@@ -524,7 +524,7 @@ static void st_i2c_handle_write(struct st_i2c_dev *i2c_dev) · 523 unmodified lines
524	524	}
525	525
526	526	/**
527		− * st_i2c_handle_write() - Handle FIFO enmpty interrupt in case of read
	527	+ * st_i2c_handle_read() - Handle FIFO empty interrupt in case of read
528	528	* @i2c_dev: Controller's private data
529	529	*/
530	530	static void st_i2c_handle_read(struct st_i2c_dev *i2c_dev)
@@ -558,7 +558,7 @@ static void st_i2c_handle_read(struct st_i2c_dev *i2c_dev) · 27 unmodified lines
558	558	}
559	559
560	560	/**
561		− * st_i2c_isr() - Interrupt routine
	561	+ * st_i2c_isr_thread() - Interrupt routine
562	562	* @irq: interrupt number
563	563	* @data: Controller's private data
564	564	*/

drivers/i2c/busses/i2c-stm32f4.c+1−1

@@ -313,7 +313,7 @@ static int stm32f4_i2c_wait_free_bus(struct stm32f4_i2c_dev *i2c_dev) · 312 unmodified lines
313	313	}
314	314
315	315	/**
316		− * stm32f4_i2c_write_ byte() - Write a byte in the data register
	316	+ * stm32f4_i2c_write_byte() - Write a byte in the data register
317	317	* @i2c_dev: Controller's private data
318	318	* @byte: Data to write in the register
319	319	*/

drivers/i2c/busses/i2c-tegra-bpmp.c+2−2

@@ -65,7 +65,7 @@ static void tegra_bpmp_xlate_flags(u16 flags, u16 *out) · 64 unmodified lines
65	65	*out \|= SERIALI2C_RECV_LEN;
66	66	}
67	67
68		−/**
	68	+/*
69	69	* The serialized I2C format is simply the following:
70	70	* [addr little-endian][flags little-endian][len little-endian][data if write]
71	71	* [addr little-endian][flags little-endian][len little-endian][data if write]
@@ -109,7 +109,7 @@ static void tegra_bpmp_serialize_i2c_msg(struct tegra_bpmp_i2c *i2c, · 37 unmodified lines
109	109	request->xfer.data_size = pos;
110	110	}
111	111
112		−/**
	112	+/*
113	113	* The data in the BPMP -> CPU direction is composed of sequential blocks for
114	114	* those messages that have I2C_M_RD. So, for example, if you have:
115	115	*

drivers/i2c/muxes/i2c-arb-gpio-challenge.c+2−2

@@ -34,7 +34,7 @@ struct i2c_arbitrator_data { · 33 unmodified lines
34	34	};
35	35
36	36
37		−/**
	37	+/*
38	38	* i2c_arbitrator_select - claim the I2C bus
39	39	*
40	40	* Use the GPIO-based signalling protocol; return -EBUSY if we fail.
@@ -77,7 +77,7 @@ static int i2c_arbitrator_select(struct i2c_mux_core *muxc, u32 chan) · 36 unmodified lines
77	77	return -EBUSY;
78	78	}
79	79
80		−/**
	80	+/*
81	81	* i2c_arbitrator_deselect - release the I2C bus
82	82	*
83	83	* Release the I2C bus using the GPIO-based signalling protocol.

drivers/iio/adc/ad7124.c+20−16

@@ -771,6 +771,13 @@ static int ad7124_of_parse_channel_config(struct iio_dev *indio_dev, · 770 unmodified lines
771	771	if (ret)
772	772	goto err;
773	773
	774	+ if (channel >= indio_dev->num_channels) {
	775	+ dev_err(indio_dev->dev.parent,
	776	+ "Channel index >= number of channels\n");
	777	+ ret = -EINVAL;
	778	+ goto err;
	779	+ }
	780	+
774	781	ret = of_property_read_u32_array(child, "diff-channels",
775	782	ain, 2);
776	783	if (ret)
@@ -850,6 +857,11 @@ static int ad7124_setup(struct ad7124_state *st) · 73 unmodified lines
850	857	return ret;
851	858	}
852	859
	860	+static void ad7124_reg_disable(void *r)
	861	+{
	862	+ regulator_disable(r);
	863	+}
	864	+
853	865	static int ad7124_probe(struct spi_device *spi)
854	866	{
855	867	const struct ad7124_chip_info *info;
@@ -895,17 +907,20 @@ static int ad7124_probe(struct spi_device *spi) · 39 unmodified lines
895	907	ret = regulator_enable(st->vref[i]);
896	908	if (ret)
897	909	return ret;
	910	+
	911	+ ret = devm_add_action_or_reset(&spi->dev, ad7124_reg_disable,
	912	+ st->vref[i]);
	913	+ if (ret)
	914	+ return ret;
898	915	}
899	916
900	917	st->mclk = devm_clk_get(&spi->dev, "mclk");
901		− if (IS_ERR(st->mclk)) {
902		− ret = PTR_ERR(st->mclk);
903		− goto error_regulator_disable;
904		− }
	918	+ if (IS_ERR(st->mclk))
	919	+ return PTR_ERR(st->mclk);
905	920
906	921	ret = clk_prepare_enable(st->mclk);
907	922	if (ret < 0)
908		− goto error_regulator_disable;
	923	+ return ret;
909	924
910	925	ret = ad7124_soft_reset(st);
911	926	if (ret < 0)
@@ -935,11 +950,6 @@ error_remove_trigger: · 23 unmodified lines
935	950	ad_sd_cleanup_buffer_and_trigger(indio_dev);
936	951	error_clk_disable_unprepare:
937	952	clk_disable_unprepare(st->mclk);
938		−error_regulator_disable:
939		− for (i = ARRAY_SIZE(st->vref) - 1; i >= 0; i--) {
940		− if (!IS_ERR_OR_NULL(st->vref[i]))
941		− regulator_disable(st->vref[i]);
942		− }
943	953
944	954	return ret;
945	955	}
@@ -948,17 +958,11 @@ static int ad7124_remove(struct spi_device *spi) · 2 unmodified lines
948	958	{
949	959	struct iio_dev *indio_dev = spi_get_drvdata(spi);
950	960	struct ad7124_state *st = iio_priv(indio_dev);
951		− int i;
952	961
953	962	iio_device_unregister(indio_dev);
954	963	ad_sd_cleanup_buffer_and_trigger(indio_dev);
955	964	clk_disable_unprepare(st->mclk);
956	965
957		− for (i = ARRAY_SIZE(st->vref) - 1; i >= 0; i--) {
958		− if (!IS_ERR_OR_NULL(st->vref[i]))
959		− regulator_disable(st->vref[i]);
960		− }
961		−
962	966	return 0;
963	967	}
964	968

drivers/iio/adc/ad7192.c+10−9

@@ -912,7 +912,7 @@ static int ad7192_probe(struct spi_device *spi) · 911 unmodified lines
912	912	{
913	913	struct ad7192_state *st;
914	914	struct iio_dev *indio_dev;
915		− int ret, voltage_uv = 0;
	915	+ int ret;
916	916
917	917	if (!spi->irq) {
918	918	dev_err(&spi->dev, "no IRQ?\n");
@@ -949,15 +949,12 @@ static int ad7192_probe(struct spi_device *spi) · 30 unmodified lines
949	949	goto error_disable_avdd;
950	950	}
951	951
952		− voltage_uv = regulator_get_voltage(st->avdd);
953		−
954		− if (voltage_uv > 0) {
955		− st->int_vref_mv = voltage_uv / 1000;
956		− } else {
957		− ret = voltage_uv;
	952	+ ret = regulator_get_voltage(st->avdd);
	953	+ if (ret < 0) {
958	954	dev_err(&spi->dev, "Device tree error, reference voltage undefined\n");
959	955	goto error_disable_avdd;
960	956	}
	957	+ st->int_vref_mv = ret / 1000;
961	958
962	959	spi_set_drvdata(spi, indio_dev);
963	960	st->chip_info = of_device_get_match_data(&spi->dev);
@@ -1014,7 +1011,9 @@ static int ad7192_probe(struct spi_device *spi) · 50 unmodified lines
1014	1011	return 0;
1015	1012
1016	1013	error_disable_clk:
1017		− clk_disable_unprepare(st->mclk);
	1014	+ if (st->clock_sel == AD7192_CLK_EXT_MCLK1_2 \|\|
	1015	+ st->clock_sel == AD7192_CLK_EXT_MCLK2)
	1016	+ clk_disable_unprepare(st->mclk);
1018	1017	error_remove_trigger:
1019	1018	ad_sd_cleanup_buffer_and_trigger(indio_dev);
1020	1019	error_disable_dvdd:
@@ -1031,7 +1030,9 @@ static int ad7192_remove(struct spi_device *spi) · 10 unmodified lines
1031	1030	struct ad7192_state *st = iio_priv(indio_dev);
1032	1031
1033	1032	iio_device_unregister(indio_dev);
1034		− clk_disable_unprepare(st->mclk);
	1033	+ if (st->clock_sel == AD7192_CLK_EXT_MCLK1_2 \|\|
	1034	+ st->clock_sel == AD7192_CLK_EXT_MCLK2)
	1035	+ clk_disable_unprepare(st->mclk);
1035	1036	ad_sd_cleanup_buffer_and_trigger(indio_dev);
1036	1037
1037	1038	regulator_disable(st->dvdd);

drivers/iio/adc/ad7768-1.c+6−2

@@ -167,6 +167,10 @@ struct ad7768_state { · 166 unmodified lines
167	167	* transfer buffers to live in their own cache lines.
168	168	*/
169	169	union {
	170	+ struct {
	171	+ __be32 chan;
	172	+ s64 timestamp;
	173	+ } scan;
170	174	__be32 d32;
171	175	u8 d8[2];
172	176	} data ____cacheline_aligned;
@@ -469,11 +473,11 @@ static irqreturn_t ad7768_trigger_handler(int irq, void *p) · 296 unmodified lines
469	473
470	474	mutex_lock(&st->lock);
471	475
472		− ret = spi_read(st->spi, &st->data.d32, 3);
	476	+ ret = spi_read(st->spi, &st->data.scan.chan, 3);
473	477	if (ret < 0)
474	478	goto err_unlock;
475	479
476		− iio_push_to_buffers_with_timestamp(indio_dev, &st->data.d32,
	480	+ iio_push_to_buffers_with_timestamp(indio_dev, &st->data.scan,
477	481	iio_get_time_ns(indio_dev));
478	482
479	483	iio_trigger_notify_done(indio_dev->trig);

drivers/iio/adc/ad7793.c+1−0

@@ -279,6 +279,7 @@ static int ad7793_setup(struct iio_dev *indio_dev, · 278 unmodified lines
279	279	id &= AD7793_ID_MASK;
280	280
281	281	if (id != st->chip_info->id) {
	282	+ ret = -ENODEV;
282	283	dev_err(&st->sd.spi->dev, "device ID query failed\n");
283	284	goto out;
284	285	}

drivers/iio/adc/ad7923.c+3−1

@@ -59,8 +59,10 @@ struct ad7923_state { · 58 unmodified lines
59	59	/*
60	60	* DMA (thus cache coherency maintenance) requires the
61	61	* transfer buffers to live in their own cache lines.
	62	+ * Ensure rx_buf can be directly used in iio_push_to_buffers_with_timetamp
	63	+ * Length = 8 channels + 4 extra for 8 byte timestamp
62	64	*/
63		− __be16 rx_buf[4] ____cacheline_aligned;
	65	+ __be16 rx_buf[12] ____cacheline_aligned;
64	66	__be16 tx_buf[4];
65	67	};
66	68

drivers/iio/dac/ad5770r.c+11−5

@@ -524,23 +524,29 @@ static int ad5770r_channel_config(struct ad5770r_state *st) · 523 unmodified lines
524	524	device_for_each_child_node(&st->spi->dev, child) {
525	525	ret = fwnode_property_read_u32(child, "num", &num);
526	526	if (ret)
527		− return ret;
528		− if (num >= AD5770R_MAX_CHANNELS)
529		− return -EINVAL;
	527	+ goto err_child_out;
	528	+ if (num >= AD5770R_MAX_CHANNELS) {
	529	+ ret = -EINVAL;
	530	+ goto err_child_out;
	531	+ }
530	532
531	533	ret = fwnode_property_read_u32_array(child,
532	534	"adi,range-microamp",
533	535	tmp, 2);
534	536	if (ret)
535		− return ret;
	537	+ goto err_child_out;
536	538
537	539	min = tmp[0] / 1000;
538	540	max = tmp[1] / 1000;
539	541	ret = ad5770r_store_output_range(st, min, max, num);
540	542	if (ret)
541		− return ret;
	543	+ goto err_child_out;
542	544	}
543	545
	546	+ return 0;
	547	+
	548	+err_child_out:
	549	+ fwnode_handle_put(child);
544	550	return ret;
545	551	}
546	552

drivers/iio/gyro/fxas21002c_core.c+2−0

@@ -399,6 +399,7 @@ static int fxas21002c_temp_get(struct fxas21002c_data data, int val) · 398 unmodified lines
399	399	ret = regmap_field_read(data->regmap_fields[F_TEMP], &temp);
400	400	if (ret < 0) {
401	401	dev_err(dev, "failed to read temp: %d\n", ret);
	402	+ fxas21002c_pm_put(data);
402	403	goto data_unlock;
403	404	}
404	405
@@ -432,6 +433,7 @@ static int fxas21002c_axis_get(struct fxas21002c_data *data, · 27 unmodified lines
432	433	&axis_be, sizeof(axis_be));
433	434	if (ret < 0) {
434	435	dev_err(dev, "failed to read axis: %d: %d\n", index, ret);
	436	+ fxas21002c_pm_put(data);
435	437	goto data_unlock;
436	438	}
437	439

drivers/infiniband/core/uverbs_cmd.c+5−0

@@ -3248,6 +3248,11 @@ static int ib_uverbs_ex_create_flow(struct uverbs_attr_bundle *attrs) · 3247 unmodified lines
3248	3248	goto err_free_attr;
3249	3249	}
3250	3250
	3251	+ if (!rdma_is_port_valid(uobj->context->device, cmd.flow_attr.port)) {
	3252	+ err = -EINVAL;
	3253	+ goto err_uobj;
	3254	+ }
	3255	+
3251	3256	qp = uobj_get_obj_read(qp, UVERBS_OBJECT_QP, cmd.qp_handle, attrs);
3252	3257	if (!qp) {
3253	3258	err = -EINVAL;

drivers/infiniband/hw/mlx4/main.c+1−7

@@ -581,12 +581,9 @@ static int mlx4_ib_query_device(struct ib_device *ibdev, · 580 unmodified lines
581	581	props->cq_caps.max_cq_moderation_count = MLX4_MAX_CQ_COUNT;
582	582	props->cq_caps.max_cq_moderation_period = MLX4_MAX_CQ_PERIOD;
583	583
584		− if (!mlx4_is_slave(dev->dev))
585		− err = mlx4_get_internal_clock_params(dev->dev, &clock_params);
586		−
587	584	if (uhw->outlen >= resp.response_length + sizeof(resp.hca_core_clock_offset)) {
588	585	resp.response_length += sizeof(resp.hca_core_clock_offset);
589		− if (!err && !mlx4_is_slave(dev->dev)) {
	586	+ if (!mlx4_get_internal_clock_params(dev->dev, &clock_params)) {
590	587	resp.comp_mask \|= MLX4_IB_QUERY_DEV_RESP_MASK_CORE_CLOCK_OFFSET;
591	588	resp.hca_core_clock_offset = clock_params.offset % PAGE_SIZE;
592	589	}
@@ -1702,9 +1699,6 @@ static struct ib_flow mlx4_ib_create_flow(struct ib_qp qp, · 1109 unmodified lines
1702	1699	struct mlx4_dev *dev = (to_mdev(qp->device))->dev;
1703	1700	int is_bonded = mlx4_is_bonded(dev);
1704	1701
1705		− if (!rdma_is_port_valid(qp->device, flow_attr->port))
1706		− return ERR_PTR(-EINVAL);
1707		−
1708	1702	if (flow_attr->flags & ~IB_FLOW_ATTR_FLAGS_DONT_TRAP)
1709	1703	return ERR_PTR(-EOPNOTSUPP);
1710	1704

drivers/infiniband/hw/mlx5/cq.c+4−5

@@ -849,15 +849,14 @@ static void destroy_cq_user(struct mlx5_ib_cq cq, struct ib_udata udata) · 848 unmodified lines
849	849	ib_umem_release(cq->buf.umem);
850	850	}
851	851
852		−static void init_cq_frag_buf(struct mlx5_ib_cq *cq,
853		− struct mlx5_ib_cq_buf *buf)
	852	+static void init_cq_frag_buf(struct mlx5_ib_cq_buf *buf)
854	853	{
855	854	int i;
856	855	void *cqe;
857	856	struct mlx5_cqe64 *cqe64;
858	857
859	858	for (i = 0; i < buf->nent; i++) {
860		− cqe = get_cqe(cq, i);
	859	+ cqe = mlx5_frag_buf_get_wqe(&buf->fbc, i);
861	860	cqe64 = buf->cqe_size == 64 ? cqe : cqe + 64;
862	861	cqe64->op_own = MLX5_CQE_INVALID << 4;
863	862	}
@@ -883,7 +882,7 @@ static int create_cq_kernel(struct mlx5_ib_dev dev, struct mlx5_ib_cq cq, · 19 unmodified lines
883	882	if (err)
884	883	goto err_db;
885	884
886		− init_cq_frag_buf(cq, &cq->buf);
	885	+ init_cq_frag_buf(&cq->buf);
887	886
888	887	*inlen = MLX5_ST_SZ_BYTES(create_cq_in) +
889	888	MLX5_FLD_SZ_BYTES(create_cq_in, pas[0]) *
@@ -1184,7 +1183,7 @@ static int resize_kernel(struct mlx5_ib_dev dev, struct mlx5_ib_cq cq, · 294 unmodified lines
1184	1183	if (err)
1185	1184	goto ex;
1186	1185
1187		− init_cq_frag_buf(cq, cq->resize_buf);
	1186	+ init_cq_frag_buf(cq->resize_buf);
1188	1187
1189	1188	return 0;
1190	1189

drivers/infiniband/hw/mlx5/doorbell.c+6−1

@@ -41,6 +41,7 @@ struct mlx5_ib_user_db_page { · 40 unmodified lines
41	41	struct ib_umem *umem;
42	42	unsigned long user_virt;
43	43	int refcnt;
	44	+ struct mm_struct *mm;
44	45	};
45	46
46	47	int mlx5_ib_db_map_user(struct mlx5_ib_ucontext *context,
@@ -53,7 +54,8 @@ int mlx5_ib_db_map_user(struct mlx5_ib_ucontext *context, · 6 unmodified lines
53	54	mutex_lock(&context->db_page_mutex);
54	55
55	56	list_for_each_entry(page, &context->db_page_list, list)
56		− if (page->user_virt == (virt & PAGE_MASK))
	57	+ if ((current->mm == page->mm) &&
	58	+ (page->user_virt == (virt & PAGE_MASK)))
57	59	goto found;
58	60
59	61	page = kmalloc(sizeof(*page), GFP_KERNEL);
@@ -71,6 +73,8 @@ int mlx5_ib_db_map_user(struct mlx5_ib_ucontext *context, · 11 unmodified lines
71	73	kfree(page);
72	74	goto out;
73	75	}
	76	+ mmgrab(current->mm);
	77	+ page->mm = current->mm;
74	78
75	79	list_add(&page->list, &context->db_page_list);
76	80
@@ -91,6 +95,7 @@ void mlx5_ib_db_unmap_user(struct mlx5_ib_ucontext context, struct mlx5_db db) · 14 unmodified lines
91	95
92	96	if (!--db->u.user_page->refcnt) {
93	97	list_del(&db->u.user_page->list);
	98	+ mmdrop(db->u.user_page->mm);
94	99	ib_umem_release(db->u.user_page->umem);
95	100	kfree(db->u.user_page);
96	101	}

drivers/infiniband/hw/mlx5/fs.c+8−3

@@ -1194,9 +1194,8 @@ static struct ib_flow mlx5_ib_create_flow(struct ib_qp qp, · 1193 unmodified lines
1194	1194	goto free_ucmd;
1195	1195	}
1196	1196
1197		− if (flow_attr->port > dev->num_ports \|\|
1198		− (flow_attr->flags &
1199		− ~(IB_FLOW_ATTR_FLAGS_DONT_TRAP \| IB_FLOW_ATTR_FLAGS_EGRESS))) {
	1197	+ if (flow_attr->flags &
	1198	+ ~(IB_FLOW_ATTR_FLAGS_DONT_TRAP \| IB_FLOW_ATTR_FLAGS_EGRESS)) {
1200	1199	err = -EINVAL;
1201	1200	goto free_ucmd;
1202	1201	}
@@ -2134,6 +2133,12 @@ static int UVERBS_HANDLER(MLX5_IB_METHOD_FLOW_MATCHER_CREATE)( · 931 unmodified lines
2134	2133	if (err)
2135	2134	goto end;
2136	2135
	2136	+ if (obj->ns_type == MLX5_FLOW_NAMESPACE_FDB &&
	2137	+ mlx5_eswitch_mode(dev->mdev) != MLX5_ESWITCH_OFFLOADS) {
	2138	+ err = -EINVAL;
	2139	+ goto end;
	2140	+ }
	2141	+
2137	2142	uobj->object = obj;
2138	2143	obj->mdev = dev->mdev;
2139	2144	atomic_set(&obj->usecnt, 0);

drivers/infiniband/hw/mlx5/mr.c+4−4

@@ -743,10 +743,10 @@ int mlx5_mr_cache_init(struct mlx5_ib_dev *dev) · 742 unmodified lines
743	743	ent->xlt = (1 << ent->order) * sizeof(struct mlx5_mtt) /
744	744	MLX5_IB_UMR_OCTOWORD;
745	745	ent->access_mode = MLX5_MKC_ACCESS_MODE_MTT;
746		− if ((dev->mdev->profile->mask & MLX5_PROF_MASK_MR_CACHE) &&
	746	+ if ((dev->mdev->profile.mask & MLX5_PROF_MASK_MR_CACHE) &&
747	747	!dev->is_rep && mlx5_core_is_pf(dev->mdev) &&
748	748	mlx5_ib_can_load_pas_with_umr(dev, 0))
749		− ent->limit = dev->mdev->profile->mr_cache[i].limit;
	749	+ ent->limit = dev->mdev->profile.mr_cache[i].limit;
750	750	else
751	751	ent->limit = 0;
752	752	spin_lock_irq(&ent->lock);
@@ -1940,8 +1940,8 @@ int mlx5_ib_dereg_mr(struct ib_mr ibmr, struct ib_udata udata) · 1187 unmodified lines
1940	1940	mlx5r_deref_wait_odp_mkey(&mr->mmkey);
1941	1941
1942	1942	if (ibmr->type == IB_MR_TYPE_INTEGRITY) {
1943		− xa_cmpxchg(&dev->sig_mrs, mlx5_base_mkey(mr->mmkey.key), ibmr,
1944		− NULL, GFP_KERNEL);
	1943	+ xa_cmpxchg(&dev->sig_mrs, mlx5_base_mkey(mr->mmkey.key),
	1944	+ mr->sig, NULL, GFP_KERNEL);
1945	1945
1946	1946	if (mr->mtt_mr) {
1947	1947	rc = mlx5_ib_dereg_mr(&mr->mtt_mr->ibmr, NULL);

drivers/infiniband/ulp/ipoib/ipoib_netlink.c+1−0

@@ -163,6 +163,7 @@ static size_t ipoib_get_size(const struct net_device *dev) · 162 unmodified lines
163	163
164	164	static struct rtnl_link_ops ipoib_link_ops __read_mostly = {
165	165	.kind = "ipoib",
	166	+ .netns_refund = true,
166	167	.maxtype = IFLA_IPOIB_MAX,
167	168	.policy = ipoib_policy,
168	169	.priv_size = sizeof(struct ipoib_dev_priv),

drivers/interconnect/qcom/bcm-voter.c+3−1

@@ -1,6 +1,6 @@
1	1	// SPDX-License-Identifier: GPL-2.0
2	2	/*
3		− * Copyright (c) 2020, The Linux Foundation. All rights reserved.
	3	+ * Copyright (c) 2020-2021, The Linux Foundation. All rights reserved.
4	4	*/
5	5
6	6	#include <asm/div64.h>
@@ -205,6 +205,7 @@ struct bcm_voter of_bcm_voter_get(struct device dev, const char *name) · 198 unmodified lines
205	205	}
206	206	mutex_unlock(&bcm_voter_lock);
207	207
	208	+ of_node_put(node);
208	209	return voter;
209	210	}
210	211	EXPORT_SYMBOL_GPL(of_bcm_voter_get);
@@ -362,6 +363,7 @@ static const struct of_device_id bcm_voter_of_match[] = { · 151 unmodified lines
362	363	{ .compatible = "qcom,bcm-voter" },
363	364	{ }
364	365	};
	366	+MODULE_DEVICE_TABLE(of, bcm_voter_of_match);
365	367
366	368	static struct platform_driver qcom_icc_bcm_voter_driver = {
367	369	.probe = qcom_icc_bcm_voter_probe,

drivers/iommu/amd/iommu.c+3−1

@@ -884,7 +884,7 @@ static inline u64 build_inv_address(u64 address, size_t size) · 883 unmodified lines
884	884	* The msb-bit must be clear on the address. Just set all the
885	885	* lower bits.
886	886	*/
887		− address \|= 1ull << (msb_diff - 1);
	887	+ address \|= (1ull << msb_diff) - 1;
888	888	}
889	889
890	890	/* Clear bits 11:0 */
@@ -1714,6 +1714,8 @@ static void amd_iommu_probe_finalize(struct device *dev) · 823 unmodified lines
1714	1714	domain = iommu_get_domain_for_dev(dev);
1715	1715	if (domain->type == IOMMU_DOMAIN_DMA)
1716	1716	iommu_setup_dma_ops(dev, IOVA_START_PFN << PAGE_SHIFT, 0);
	1717	+ else
	1718	+ set_dma_ops(dev, NULL);
1717	1719	}
1718	1720
1719	1721	static void amd_iommu_release_device(struct device *dev)

drivers/iommu/intel/dmar.c+3−1

@@ -1142,7 +1142,7 @@ static int alloc_iommu(struct dmar_drhd_unit *drhd) · 1141 unmodified lines
1142	1142
1143	1143	err = iommu_device_register(&iommu->iommu, &intel_iommu_ops, NULL);
1144	1144	if (err)
1145		− goto err_unmap;
	1145	+ goto err_sysfs;
1146	1146	}
1147	1147
1148	1148	drhd->iommu = iommu;
@@ -1150,6 +1150,8 @@ static int alloc_iommu(struct dmar_drhd_unit *drhd) · 1 unmodified line
1150	1150
1151	1151	return 0;
1152	1152
	1153	+err_sysfs:
	1154	+ iommu_device_sysfs_remove(&iommu->iommu);
1153	1155	err_unmap:
1154	1156	unmap_iommu(iommu);
1155	1157	error_free_seq_id:

drivers/iommu/intel/iommu.c+7−2

@@ -2525,9 +2525,9 @@ static int domain_setup_first_level(struct intel_iommu *iommu, · 2524 unmodified lines
2525	2525	struct device *dev,
2526	2526	u32 pasid)
2527	2527	{
2528		− int flags = PASID_FLAG_SUPERVISOR_MODE;
2529	2528	struct dma_pte *pgd = domain->pgd;
2530	2529	int agaw, level;
	2530	+ int flags = 0;
2531	2531
2532	2532	/*
2533	2533	* Skip top levels of page tables for iommu which has
@@ -2543,7 +2543,10 @@ static int domain_setup_first_level(struct intel_iommu *iommu, · 9 unmodified lines
2543	2543	if (level != 4 && level != 5)
2544	2544	return -EINVAL;
2545	2545
2546		− flags \|= (level == 5) ? PASID_FLAG_FL5LP : 0;
	2546	+ if (pasid != PASID_RID2PASID)
	2547	+ flags \|= PASID_FLAG_SUPERVISOR_MODE;
	2548	+ if (level == 5)
	2549	+ flags \|= PASID_FLAG_FL5LP;
2547	2550
2548	2551	if (domain->domain.type == IOMMU_DOMAIN_UNMANAGED)
2549	2552	flags \|= PASID_FLAG_PAGE_SNOOP;
@@ -4606,6 +4609,8 @@ static int auxiliary_link_device(struct dmar_domain *domain, · 2056 unmodified lines
4606	4609
4607	4610	if (!sinfo) {
4608	4611	sinfo = kzalloc(sizeof(*sinfo), GFP_ATOMIC);
	4612	+ if (!sinfo)
	4613	+ return -ENOMEM;
4609	4614	sinfo->domain = domain;
4610	4615	sinfo->pdev = dev;
4611	4616	list_add(&sinfo->link_phys, &info->subdevices);

drivers/iommu/intel/pasid.c+2−1

@@ -699,7 +699,8 @@ int intel_pasid_setup_second_level(struct intel_iommu *iommu, · 698 unmodified lines
699	699	* Since it is a second level only translation setup, we should
700	700	* set SRE bit as well (addresses are expected to be GPAs).
701	701	*/
702		− pasid_set_sre(pte);
	702	+ if (pasid != PASID_RID2PASID)
	703	+ pasid_set_sre(pte);
703	704	pasid_set_present(pte);
704	705	pasid_flush_caches(iommu, pte, pasid, did);
705	706

drivers/iommu/virtio-iommu.c+1−0

@@ -1136,6 +1136,7 @@ static struct virtio_device_id id_table[] = { · 1135 unmodified lines
1136	1136	{ VIRTIO_ID_IOMMU, VIRTIO_DEV_ANY_ID },
1137	1137	{ 0 },
1138	1138	};
	1139	+MODULE_DEVICE_TABLE(virtio, id_table);
1139	1140
1140	1141	static struct virtio_driver virtio_iommu_drv = {
1141	1142	.driver.name = KBUILD_MODNAME,

drivers/irqchip/irq-gic-v3.c+35−1

@@ -642,11 +642,45 @@ static inline void gic_handle_nmi(u32 irqnr, struct pt_regs *regs) · 641 unmodified lines
642	642	nmi_exit();
643	643	}
644	644
	645	+static u32 do_read_iar(struct pt_regs *regs)
	646	+{
	647	+ u32 iar;
	648	+
	649	+ if (gic_supports_nmi() && unlikely(!interrupts_enabled(regs))) {
	650	+ u64 pmr;
	651	+
	652	+ /*
	653	+ * We were in a context with IRQs disabled. However, the
	654	+ * entry code has set PMR to a value that allows any
	655	+ * interrupt to be acknowledged, and not just NMIs. This can
	656	+ * lead to surprising effects if the NMI has been retired in
	657	+ * the meantime, and that there is an IRQ pending. The IRQ
	658	+ * would then be taken in NMI context, something that nobody
	659	+ * wants to debug twice.
	660	+ *
	661	+ * Until we sort this, drop PMR again to a level that will
	662	+ * actually only allow NMIs before reading IAR, and then
	663	+ * restore it to what it was.
	664	+ */
	665	+ pmr = gic_read_pmr();
	666	+ gic_pmr_mask_irqs();
	667	+ isb();
	668	+
	669	+ iar = gic_read_iar();
	670	+
	671	+ gic_write_pmr(pmr);
	672	+ } else {
	673	+ iar = gic_read_iar();
	674	+ }
	675	+
	676	+ return iar;
	677	+}
	678	+
645	679	static asmlinkage void __exception_irq_entry gic_handle_irq(struct pt_regs *regs)
646	680	{
647	681	u32 irqnr;
648	682
649		− irqnr = gic_read_iar();
	683	+ irqnr = do_read_iar(regs);
650	684
651	685	/* Check for special IDs first */
652	686	if ((irqnr >= 1020 && irqnr <= 1023))

drivers/isdn/hardware/mISDN/netjet.c+0−1

@@ -1100,7 +1100,6 @@ nj_probe(struct pci_dev pdev, const struct pci_device_id ent) · 1099 unmodified lines
1100	1100	card->typ = NETJET_S_TJ300;
1101	1101
1102	1102	card->base = pci_resource_start(pdev, 0);
1103		− card->irq = pdev->irq;
1104	1103	pci_set_drvdata(pdev, card);
1105	1104	err = setup_instance(card);
1106	1105	if (err)

drivers/md/bcache/bcache.h+0−1

@@ -364,7 +364,6 @@ struct cached_dev { · 363 unmodified lines
364	364
365	365	/* The rest of this all shows up in sysfs */
366	366	unsigned int sequential_cutoff;
367		− unsigned int readahead;
368	367
369	368	unsigned int io_disable:1;
370	369	unsigned int verify:1;

drivers/md/bcache/request.c+7−13

@@ -880,9 +880,9 @@ static int cached_dev_cache_miss(struct btree b, struct search s, · 879 unmodified lines
880	880	struct bio *bio, unsigned int sectors)
881	881	{
882	882	int ret = MAP_CONTINUE;
883		− unsigned int reada = 0;
884	883	struct cached_dev *dc = container_of(s->d, struct cached_dev, disk);
885	884	struct bio miss, cache_bio;
	885	+ unsigned int size_limit;
886	886
887	887	s->cache_missed = 1;
888	888
@@ -892,14 +892,10 @@ static int cached_dev_cache_miss(struct btree b, struct search s, · 3 unmodified lines
892	892	goto out_submit;
893	893	}
894	894
895		− if (!(bio->bi_opf & REQ_RAHEAD) &&
896		− !(bio->bi_opf & (REQ_META\|REQ_PRIO)) &&
897		− s->iop.c->gc_stats.in_use < CUTOFF_CACHE_READA)
898		− reada = min_t(sector_t, dc->readahead >> 9,
899		− get_capacity(bio->bi_bdev->bd_disk) -
900		− bio_end_sector(bio));
901		−
902		− s->insert_bio_sectors = min(sectors, bio_sectors(bio) + reada);
	895	+ /* Limitation for valid replace key size and cache_bio bvecs number */
	896	+ size_limit = min_t(unsigned int, BIO_MAX_VECS * PAGE_SECTORS,
	897	+ (1 << KEY_SIZE_BITS) - 1);
	898	+ s->insert_bio_sectors = min3(size_limit, sectors, bio_sectors(bio));
903	899
904	900	s->iop.replace_key = KEY(s->iop.inode,
905	901	bio->bi_iter.bi_sector + s->insert_bio_sectors,
@@ -911,7 +907,8 @@ static int cached_dev_cache_miss(struct btree b, struct search s, · 5 unmodified lines
911	907
912	908	s->iop.replace = true;
913	909
914		− miss = bio_next_split(bio, sectors, GFP_NOIO, &s->d->bio_split);
	910	+ miss = bio_next_split(bio, s->insert_bio_sectors, GFP_NOIO,
	911	+ &s->d->bio_split);
915	912
916	913	/* btree_search_recurse()'s btree iterator is no good anymore */
917	914	ret = miss == bio ? MAP_DONE : -EINTR;
@@ -933,9 +930,6 @@ static int cached_dev_cache_miss(struct btree b, struct search s, · 15 unmodified lines
933	930	if (bch_bio_alloc_pages(cache_bio, __GFP_NOWARN\|GFP_NOIO))
934	931	goto out_put;
935	932
936		− if (reada)
937		− bch_mark_cache_readahead(s->iop.c, s->d);
938		−
939	933	s->cache_miss = miss;
940	934	s->iop.bio = cache_bio;
941	935	bio_get(cache_bio);

drivers/md/bcache/stats.c+0−14

@@ -46,7 +46,6 @@ read_attribute(cache_misses); · 45 unmodified lines
46	46	read_attribute(cache_bypass_hits);
47	47	read_attribute(cache_bypass_misses);
48	48	read_attribute(cache_hit_ratio);
49		−read_attribute(cache_readaheads);
50	49	read_attribute(cache_miss_collisions);
51	50	read_attribute(bypassed);
52	51
@@ -64,7 +63,6 @@ SHOW(bch_stats) · 11 unmodified lines
64	63	DIV_SAFE(var(cache_hits) * 100,
65	64	var(cache_hits) + var(cache_misses)));
66	65
67		− var_print(cache_readaheads);
68	66	var_print(cache_miss_collisions);
69	67	sysfs_hprint(bypassed, var(sectors_bypassed) << 9);
70	68	#undef var
@@ -86,7 +84,6 @@ static struct attribute *bch_stats_files[] = { · 15 unmodified lines
86	84	&sysfs_cache_bypass_hits,
87	85	&sysfs_cache_bypass_misses,
88	86	&sysfs_cache_hit_ratio,
89		− &sysfs_cache_readaheads,
90	87	&sysfs_cache_miss_collisions,
91	88	&sysfs_bypassed,
92	89	NULL
@@ -113,7 +110,6 @@ void bch_cache_accounting_clear(struct cache_accounting *acc) · 20 unmodified lines
113	110	acc->total.cache_misses = 0;
114	111	acc->total.cache_bypass_hits = 0;
115	112	acc->total.cache_bypass_misses = 0;
116		− acc->total.cache_readaheads = 0;
117	113	acc->total.cache_miss_collisions = 0;
118	114	acc->total.sectors_bypassed = 0;
119	115	}
@@ -145,7 +141,6 @@ static void scale_stats(struct cache_stats *stats, unsigned long rescale_at) · 25 unmodified lines
145	141	scale_stat(&stats->cache_misses);
146	142	scale_stat(&stats->cache_bypass_hits);
147	143	scale_stat(&stats->cache_bypass_misses);
148		− scale_stat(&stats->cache_readaheads);
149	144	scale_stat(&stats->cache_miss_collisions);
150	145	scale_stat(&stats->sectors_bypassed);
151	146	}
@@ -168,7 +163,6 @@ static void scale_accounting(struct timer_list *t) · 16 unmodified lines
168	163	move_stat(cache_misses);
169	164	move_stat(cache_bypass_hits);
170	165	move_stat(cache_bypass_misses);
171		− move_stat(cache_readaheads);
172	166	move_stat(cache_miss_collisions);
173	167	move_stat(sectors_bypassed);
174	168
@@ -209,14 +203,6 @@ void bch_mark_cache_accounting(struct cache_set c, struct bcache_device d, · 34 unmodified lines
209	203	mark_cache_stats(&c->accounting.collector, hit, bypass);
210	204	}
211	205
212		−void bch_mark_cache_readahead(struct cache_set c, struct bcache_device d)
213		−{
214		− struct cached_dev *dc = container_of(d, struct cached_dev, disk);
215		−
216		− atomic_inc(&dc->accounting.collector.cache_readaheads);
217		− atomic_inc(&c->accounting.collector.cache_readaheads);
218		−}
219		−
220	206	void bch_mark_cache_miss_collision(struct cache_set c, struct bcache_device d)
221	207	{
222	208	struct cached_dev *dc = container_of(d, struct cached_dev, disk);

drivers/md/bcache/stats.h+0−1

@@ -7,7 +7,6 @@ struct cache_stat_collector { · 6 unmodified lines
7	7	atomic_t cache_misses;
8	8	atomic_t cache_bypass_hits;
9	9	atomic_t cache_bypass_misses;
10		− atomic_t cache_readaheads;
11	10	atomic_t cache_miss_collisions;
12	11	atomic_t sectors_bypassed;
13	12	};

drivers/md/bcache/sysfs.c+0−4

@@ -137,7 +137,6 @@ rw_attribute(io_disable); · 136 unmodified lines
137	137	rw_attribute(discard);
138	138	rw_attribute(running);
139	139	rw_attribute(label);
140		−rw_attribute(readahead);
141	140	rw_attribute(errors);
142	141	rw_attribute(io_error_limit);
143	142	rw_attribute(io_error_halflife);
@@ -260,7 +259,6 @@ SHOW(__bch_cached_dev) · 116 unmodified lines
260	259	var_printf(partial_stripes_expensive, "%u");
261	260
262	261	var_hprint(sequential_cutoff);
263		− var_hprint(readahead);
264	262
265	263	sysfs_print(running, atomic_read(&dc->running));
266	264	sysfs_print(state, states[BDEV_STATE(&dc->sb)]);
@@ -365,7 +363,6 @@ STORE(__cached_dev) · 98 unmodified lines
365	363	sysfs_strtoul_clamp(sequential_cutoff,
366	364	dc->sequential_cutoff,
367	365	0, UINT_MAX);
368		− d_strtoi_h(readahead);
369	366
370	367	if (attr == &sysfs_clear_stats)
371	368	bch_cache_accounting_clear(&dc->accounting);
@@ -538,7 +535,6 @@ static struct attribute *bch_cached_dev_files[] = { · 166 unmodified lines
538	535	&sysfs_running,
539	536	&sysfs_state,
540	537	&sysfs_label,
541		− &sysfs_readahead,
542	538	#ifdef CONFIG_BCACHE_DEBUG
543	539	&sysfs_verify,
544	540	&sysfs_bypass_torture_test,

drivers/md/dm-snap.c+2−1

@@ -859,7 +859,8 @@ static uint32_t __minimum_chunk_size(struct origin *o) · 858 unmodified lines
859	859
860	860	if (o)
861	861	list_for_each_entry(snap, &o->snapshots, list)
862		− chunk_size = min(chunk_size, snap->store->chunk_size);
	862	+ chunk_size = min_not_zero(chunk_size,
	863	+ snap->store->chunk_size);
863	864
864	865	return (uint32_t) chunk_size;
865	866	}

drivers/md/dm-verity-verify-sig.c+1−1

@@ -15,7 +15,7 @@ · 14 unmodified lines
15	15	#define DM_VERITY_VERIFY_ERR(s) DM_VERITY_ROOT_HASH_VERIFICATION " " s
16	16
17	17	static bool require_signatures;
18		−module_param(require_signatures, bool, false);
	18	+module_param(require_signatures, bool, 0444);
19	19	MODULE_PARM_DESC(require_signatures,
20	20	"Verify the roothash of dm-verity hash tree");
21	21

drivers/md/raid5.c+0−2

@@ -5311,8 +5311,6 @@ static int in_chunk_boundary(struct mddev mddev, struct bio bio) · 5310 unmodified lines
5311	5311	unsigned int chunk_sectors;
5312	5312	unsigned int bio_sectors = bio_sectors(bio);
5313	5313
5314		− WARN_ON_ONCE(bio->bi_bdev->bd_partno);
5315		−
5316	5314	chunk_sectors = min(conf->chunk_sectors, conf->prev_chunk_sectors);
5317	5315	return chunk_sectors >=
5318	5316	((sector & (chunk_sectors - 1)) + bio_sectors);

drivers/misc/cardreader/rtl8411.c+1−0

@@ -468,6 +468,7 @@ static void rtl8411_init_common_params(struct rtsx_pcr *pcr) · 467 unmodified lines
468	468	pcr->sd30_drive_sel_1v8 = DRIVER_TYPE_B;
469	469	pcr->sd30_drive_sel_3v3 = DRIVER_TYPE_D;
470	470	pcr->aspm_en = ASPM_L1_EN;
	471	+ pcr->aspm_mode = ASPM_MODE_CFG;
471	472	pcr->tx_initial_phase = SET_CLOCK_PHASE(23, 7, 14);
472	473	pcr->rx_initial_phase = SET_CLOCK_PHASE(4, 3, 10);
473	474	pcr->ic_version = rtl8411_get_ic_version(pcr);

drivers/misc/cardreader/rts5209.c+1−0

@@ -255,6 +255,7 @@ void rts5209_init_params(struct rtsx_pcr *pcr) · 254 unmodified lines
255	255	pcr->sd30_drive_sel_1v8 = DRIVER_TYPE_B;
256	256	pcr->sd30_drive_sel_3v3 = DRIVER_TYPE_D;
257	257	pcr->aspm_en = ASPM_L1_EN;
	258	+ pcr->aspm_mode = ASPM_MODE_CFG;
258	259	pcr->tx_initial_phase = SET_CLOCK_PHASE(27, 27, 16);
259	260	pcr->rx_initial_phase = SET_CLOCK_PHASE(24, 6, 5);
260	261

drivers/misc/cardreader/rts5227.c+2−0

@@ -358,6 +358,7 @@ void rts5227_init_params(struct rtsx_pcr *pcr) · 357 unmodified lines
358	358	pcr->sd30_drive_sel_1v8 = CFG_DRIVER_TYPE_B;
359	359	pcr->sd30_drive_sel_3v3 = CFG_DRIVER_TYPE_B;
360	360	pcr->aspm_en = ASPM_L1_EN;
	361	+ pcr->aspm_mode = ASPM_MODE_CFG;
361	362	pcr->tx_initial_phase = SET_CLOCK_PHASE(27, 27, 15);
362	363	pcr->rx_initial_phase = SET_CLOCK_PHASE(30, 7, 7);
363	364
@@ -483,6 +484,7 @@ void rts522a_init_params(struct rtsx_pcr *pcr) · 119 unmodified lines
483	484
484	485	rts5227_init_params(pcr);
485	486	pcr->ops = &rts522a_pcr_ops;
	487	+ pcr->aspm_mode = ASPM_MODE_REG;
486	488	pcr->tx_initial_phase = SET_CLOCK_PHASE(20, 20, 11);
487	489	pcr->reg_pm_ctrl3 = RTS522A_PM_CTRL3;
488	490

drivers/misc/cardreader/rts5228.c+1−0

@@ -718,6 +718,7 @@ void rts5228_init_params(struct rtsx_pcr *pcr) · 717 unmodified lines
718	718	pcr->sd30_drive_sel_1v8 = CFG_DRIVER_TYPE_B;
719	719	pcr->sd30_drive_sel_3v3 = CFG_DRIVER_TYPE_B;
720	720	pcr->aspm_en = ASPM_L1_EN;
	721	+ pcr->aspm_mode = ASPM_MODE_REG;
721	722	pcr->tx_initial_phase = SET_CLOCK_PHASE(28, 27, 11);
722	723	pcr->rx_initial_phase = SET_CLOCK_PHASE(24, 6, 5);
723	724

drivers/misc/cardreader/rts5229.c+1−0

@@ -246,6 +246,7 @@ void rts5229_init_params(struct rtsx_pcr *pcr) · 245 unmodified lines
246	246	pcr->sd30_drive_sel_1v8 = DRIVER_TYPE_B;
247	247	pcr->sd30_drive_sel_3v3 = DRIVER_TYPE_D;
248	248	pcr->aspm_en = ASPM_L1_EN;
	249	+ pcr->aspm_mode = ASPM_MODE_CFG;
249	250	pcr->tx_initial_phase = SET_CLOCK_PHASE(27, 27, 15);
250	251	pcr->rx_initial_phase = SET_CLOCK_PHASE(30, 6, 6);
251	252

drivers/misc/cardreader/rts5249.c+3−0

@@ -566,6 +566,7 @@ void rts5249_init_params(struct rtsx_pcr *pcr) · 565 unmodified lines
566	566	pcr->sd30_drive_sel_1v8 = CFG_DRIVER_TYPE_B;
567	567	pcr->sd30_drive_sel_3v3 = CFG_DRIVER_TYPE_B;
568	568	pcr->aspm_en = ASPM_L1_EN;
	569	+ pcr->aspm_mode = ASPM_MODE_CFG;
569	570	pcr->tx_initial_phase = SET_CLOCK_PHASE(1, 29, 16);
570	571	pcr->rx_initial_phase = SET_CLOCK_PHASE(24, 6, 5);
571	572
@@ -729,6 +730,7 @@ static const struct pcr_ops rts524a_pcr_ops = { · 157 unmodified lines
729	730	void rts524a_init_params(struct rtsx_pcr *pcr)
730	731	{
731	732	rts5249_init_params(pcr);
	733	+ pcr->aspm_mode = ASPM_MODE_REG;
732	734	pcr->tx_initial_phase = SET_CLOCK_PHASE(27, 29, 11);
733	735	pcr->option.ltr_l1off_sspwrgate = LTR_L1OFF_SSPWRGATE_5250_DEF;
734	736	pcr->option.ltr_l1off_snooze_sspwrgate =
@@ -845,6 +847,7 @@ static const struct pcr_ops rts525a_pcr_ops = { · 110 unmodified lines
845	847	void rts525a_init_params(struct rtsx_pcr *pcr)
846	848	{
847	849	rts5249_init_params(pcr);
	850	+ pcr->aspm_mode = ASPM_MODE_REG;
848	851	pcr->tx_initial_phase = SET_CLOCK_PHASE(25, 29, 11);
849	852	pcr->option.ltr_l1off_sspwrgate = LTR_L1OFF_SSPWRGATE_5250_DEF;
850	853	pcr->option.ltr_l1off_snooze_sspwrgate =

drivers/misc/cardreader/rts5260.c+1−0

@@ -628,6 +628,7 @@ void rts5260_init_params(struct rtsx_pcr *pcr) · 627 unmodified lines
628	628	pcr->sd30_drive_sel_1v8 = CFG_DRIVER_TYPE_B;
629	629	pcr->sd30_drive_sel_3v3 = CFG_DRIVER_TYPE_B;
630	630	pcr->aspm_en = ASPM_L1_EN;
	631	+ pcr->aspm_mode = ASPM_MODE_REG;
631	632	pcr->tx_initial_phase = SET_CLOCK_PHASE(27, 29, 11);
632	633	pcr->rx_initial_phase = SET_CLOCK_PHASE(24, 6, 5);
633	634

drivers/misc/cardreader/rts5261.c+1−0

@@ -783,6 +783,7 @@ void rts5261_init_params(struct rtsx_pcr *pcr) · 782 unmodified lines
783	783	pcr->sd30_drive_sel_1v8 = 0x00;
784	784	pcr->sd30_drive_sel_3v3 = 0x00;
785	785	pcr->aspm_en = ASPM_L1_EN;
	786	+ pcr->aspm_mode = ASPM_MODE_REG;
786	787	pcr->tx_initial_phase = SET_CLOCK_PHASE(27, 27, 11);
787	788	pcr->rx_initial_phase = SET_CLOCK_PHASE(24, 6, 5);
788	789

drivers/misc/cardreader/rtsx_pcr.c+31−13

@@ -85,12 +85,18 @@ static void rtsx_comm_set_aspm(struct rtsx_pcr *pcr, bool enable) · 84 unmodified lines
85	85	if (pcr->aspm_enabled == enable)
86	86	return;
87	87
88		− if (pcr->aspm_en & 0x02)
89		− rtsx_pci_write_register(pcr, ASPM_FORCE_CTL, FORCE_ASPM_CTL0 \|
90		− FORCE_ASPM_CTL1, enable ? 0 : FORCE_ASPM_CTL0 \| FORCE_ASPM_CTL1);
91		− else
92		− rtsx_pci_write_register(pcr, ASPM_FORCE_CTL, FORCE_ASPM_CTL0 \|
93		− FORCE_ASPM_CTL1, FORCE_ASPM_CTL0 \| FORCE_ASPM_CTL1);
	88	+ if (pcr->aspm_mode == ASPM_MODE_CFG) {
	89	+ pcie_capability_clear_and_set_word(pcr->pci, PCI_EXP_LNKCTL,
	90	+ PCI_EXP_LNKCTL_ASPMC,
	91	+ enable ? pcr->aspm_en : 0);
	92	+ } else if (pcr->aspm_mode == ASPM_MODE_REG) {
	93	+ if (pcr->aspm_en & 0x02)
	94	+ rtsx_pci_write_register(pcr, ASPM_FORCE_CTL, FORCE_ASPM_CTL0 \|
	95	+ FORCE_ASPM_CTL1, enable ? 0 : FORCE_ASPM_CTL0 \| FORCE_ASPM_CTL1);
	96	+ else
	97	+ rtsx_pci_write_register(pcr, ASPM_FORCE_CTL, FORCE_ASPM_CTL0 \|
	98	+ FORCE_ASPM_CTL1, FORCE_ASPM_CTL0 \| FORCE_ASPM_CTL1);
	99	+ }
94	100
95	101	if (!enable && (pcr->aspm_en & 0x02))
96	102	mdelay(10);
@@ -1394,7 +1400,8 @@ static int rtsx_pci_init_hw(struct rtsx_pcr *pcr) · 1297 unmodified lines
1394	1400	return err;
1395	1401	}
1396	1402
1397		− rtsx_pci_write_register(pcr, ASPM_FORCE_CTL, 0x30, 0x30);
	1403	+ if (pcr->aspm_mode == ASPM_MODE_REG)
	1404	+ rtsx_pci_write_register(pcr, ASPM_FORCE_CTL, 0x30, 0x30);
1398	1405
1399	1406	/* No CD interrupt if probing driver with card inserted.
1400	1407	* So we need to initialize pcr->card_exist here.
@@ -1410,6 +1417,8 @@ static int rtsx_pci_init_hw(struct rtsx_pcr *pcr) · 9 unmodified lines
1410	1417	static int rtsx_pci_init_chip(struct rtsx_pcr *pcr)
1411	1418	{
1412	1419	int err;
	1420	+ u16 cfg_val;
	1421	+ u8 val;
1413	1422
1414	1423	spin_lock_init(&pcr->lock);
1415	1424	mutex_init(&pcr->pcr_mutex);
@@ -1477,6 +1486,21 @@ static int rtsx_pci_init_chip(struct rtsx_pcr *pcr) · 61 unmodified lines
1477	1486	if (!pcr->slots)
1478	1487	return -ENOMEM;
1479	1488
	1489	+ if (pcr->aspm_mode == ASPM_MODE_CFG) {
	1490	+ pcie_capability_read_word(pcr->pci, PCI_EXP_LNKCTL, &cfg_val);
	1491	+ if (cfg_val & PCI_EXP_LNKCTL_ASPM_L1)
	1492	+ pcr->aspm_enabled = true;
	1493	+ else
	1494	+ pcr->aspm_enabled = false;
	1495	+
	1496	+ } else if (pcr->aspm_mode == ASPM_MODE_REG) {
	1497	+ rtsx_pci_read_register(pcr, ASPM_FORCE_CTL, &val);
	1498	+ if (val & FORCE_ASPM_CTL0 && val & FORCE_ASPM_CTL1)
	1499	+ pcr->aspm_enabled = false;
	1500	+ else
	1501	+ pcr->aspm_enabled = true;
	1502	+ }
	1503	+
1480	1504	if (pcr->ops->fetch_vendor_settings)
1481	1505	pcr->ops->fetch_vendor_settings(pcr);
1482	1506
@@ -1506,7 +1530,6 @@ static int rtsx_pci_probe(struct pci_dev *pcidev, · 23 unmodified lines
1506	1530	struct pcr_handle *handle;
1507	1531	u32 base, len;
1508	1532	int ret, i, bar = 0;
1509		− u8 val;
1510	1533
1511	1534	dev_dbg(&(pcidev->dev),
1512	1535	": Realtek PCI-E Card Reader found at %s [%04x:%04x] (rev %x)\n",
@@ -1572,11 +1595,6 @@ static int rtsx_pci_probe(struct pci_dev *pcidev, · 59 unmodified lines
1572	1595	pcr->host_cmds_addr = pcr->rtsx_resv_buf_addr;
1573	1596	pcr->host_sg_tbl_ptr = pcr->rtsx_resv_buf + HOST_CMDS_BUF_LEN;
1574	1597	pcr->host_sg_tbl_addr = pcr->rtsx_resv_buf_addr + HOST_CMDS_BUF_LEN;
1575		− rtsx_pci_read_register(pcr, ASPM_FORCE_CTL, &val);
1576		− if (val & FORCE_ASPM_CTL0 && val & FORCE_ASPM_CTL1)
1577		− pcr->aspm_enabled = false;
1578		− else
1579		− pcr->aspm_enabled = true;
1580	1598	pcr->card_inserted = 0;
1581	1599	pcr->card_removed = 0;
1582	1600	INIT_DELAYED_WORK(&pcr->carddet_work, rtsx_pci_card_detect);

drivers/misc/kgdbts.c+2−1

@@ -101,8 +101,9 @@ · 100 unmodified lines
101	101	printk(KERN_INFO a); \
102	102	} while (0)
103	103	#define v2printk(a...) do { \
104		− if (verbose > 1) \
	104	+ if (verbose > 1) { \
105	105	printk(KERN_INFO a); \
	106	+ } \
106	107	touch_nmi_watchdog(); \
107	108	} while (0)
108	109	#define eprintk(a...) do { \

drivers/misc/mei/interrupt.c+3−0

@@ -277,6 +277,9 @@ static int mei_cl_irq_read(struct mei_cl cl, struct mei_cl_cb cb, · 276 unmodified lines
277	277	return ret;
278	278	}
279	279
	280	+ pm_runtime_mark_last_busy(dev->dev);
	281	+ pm_request_autosuspend(dev->dev);
	282	+
280	283	list_move_tail(&cb->list, &cl->rd_pending);
281	284
282	285	return 0;

drivers/mmc/host/renesas_sdhi_core.c+7−2

@@ -692,14 +692,19 @@ static int renesas_sdhi_execute_tuning(struct mmc_host *mmc, u32 opcode) · 691 unmodified lines
692	692
693	693	/* Issue CMD19 twice for each tap */
694	694	for (i = 0; i < 2 * priv->tap_num; i++) {
	695	+ int cmd_error;
	696	+
695	697	/* Set sampling clock position */
696	698	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_TAPSET, i % priv->tap_num);
697	699
698		− if (mmc_send_tuning(mmc, opcode, NULL) == 0)
	700	+ if (mmc_send_tuning(mmc, opcode, &cmd_error) == 0)
699	701	set_bit(i, priv->taps);
700	702
701	703	if (sd_scc_read32(host, priv, SH_MOBILE_SDHI_SCC_SMPCMP) == 0)
702	704	set_bit(i, priv->smpcmp);
	705	+
	706	+ if (cmd_error)
	707	+ mmc_abort_tuning(mmc, opcode);
703	708	}
704	709
705	710	ret = renesas_sdhi_select_tuning(host);
@@ -939,7 +944,7 @@ static const struct soc_device_attribute sdhi_quirks_match[] = { · 233 unmodified lines
939	944	{ .soc_id = "r8a7795", .revision = "ES3.*", .data = &sdhi_quirks_bad_taps2367 },
940	945	{ .soc_id = "r8a7796", .revision = "ES1.[012]", .data = &sdhi_quirks_4tap_nohs400 },
941	946	{ .soc_id = "r8a7796", .revision = "ES1.*", .data = &sdhi_quirks_r8a7796_es13 },
942		− { .soc_id = "r8a7796", .revision = "ES3.*", .data = &sdhi_quirks_bad_taps1357 },
	947	+ { .soc_id = "r8a77961", .data = &sdhi_quirks_bad_taps1357 },
943	948	{ .soc_id = "r8a77965", .data = &sdhi_quirks_r8a77965 },
944	949	{ .soc_id = "r8a77980", .data = &sdhi_quirks_nohs400 },
945	950	{ .soc_id = "r8a77990", .data = &sdhi_quirks_r8a77990 },

drivers/mtd/nand/raw/cs553x_nand.c+11−1

@@ -18,6 +18,7 @@ · 17 unmodified lines
18	18	#include <linux/module.h>
19	19	#include <linux/delay.h>
20	20	#include <linux/mtd/mtd.h>
	21	+#include <linux/mtd/nand-ecc-sw-hamming.h>
21	22	#include <linux/mtd/rawnand.h>
22	23	#include <linux/mtd/partitions.h>
23	24	#include <linux/iopoll.h>
@@ -240,6 +241,15 @@ static int cs_calculate_ecc(struct nand_chip this, const u_char dat, · 216 unmodified lines
240	241	return 0;
241	242	}
242	243
	244	+static int cs553x_ecc_correct(struct nand_chip *chip,
	245	+ unsigned char *buf,
	246	+ unsigned char *read_ecc,
	247	+ unsigned char *calc_ecc)
	248	+{
	249	+ return ecc_sw_hamming_correct(buf, read_ecc, calc_ecc,
	250	+ chip->ecc.size, false);
	251	+}
	252	+
243	253	static struct cs553x_nand_controller *controllers[4];
244	254
245	255	static int cs553x_attach_chip(struct nand_chip *chip)
@@ -251,7 +261,7 @@ static int cs553x_attach_chip(struct nand_chip *chip) · 5 unmodified lines
251	261	chip->ecc.bytes = 3;
252	262	chip->ecc.hwctl = cs_enable_hwecc;
253	263	chip->ecc.calculate = cs_calculate_ecc;
254		− chip->ecc.correct = rawnand_sw_hamming_correct;
	264	+ chip->ecc.correct = cs553x_ecc_correct;
255	265	chip->ecc.strength = 1;
256	266
257	267	return 0;

drivers/mtd/nand/raw/fsmc_nand.c+11−1

@@ -25,6 +25,7 @@ · 24 unmodified lines
25	25	#include <linux/sched.h>
26	26	#include <linux/types.h>
27	27	#include <linux/mtd/mtd.h>
	28	+#include <linux/mtd/nand-ecc-sw-hamming.h>
28	29	#include <linux/mtd/rawnand.h>
29	30	#include <linux/platform_device.h>
30	31	#include <linux/of.h>
@@ -432,6 +433,15 @@ static int fsmc_read_hwecc_ecc1(struct nand_chip chip, const u8 data, · 401 unmodified lines
432	433	return 0;
433	434	}
434	435
	436	+static int fsmc_correct_ecc1(struct nand_chip *chip,
	437	+ unsigned char *buf,
	438	+ unsigned char *read_ecc,
	439	+ unsigned char *calc_ecc)
	440	+{
	441	+ return ecc_sw_hamming_correct(buf, read_ecc, calc_ecc,
	442	+ chip->ecc.size, false);
	443	+}
	444	+
435	445	/* Count the number of 0's in buff upto a max of max_bits */
436	446	static int count_written_bits(u8 *buff, int size, int max_bits)
437	447	{
@@ -917,7 +927,7 @@ static int fsmc_nand_attach_chip(struct nand_chip *nand) · 479 unmodified lines
917	927	case NAND_ECC_ENGINE_TYPE_ON_HOST:
918	928	dev_info(host->dev, "Using 1-bit HW ECC scheme\n");
919	929	nand->ecc.calculate = fsmc_read_hwecc_ecc1;
920		− nand->ecc.correct = rawnand_sw_hamming_correct;
	930	+ nand->ecc.correct = fsmc_correct_ecc1;
921	931	nand->ecc.hwctl = fsmc_enable_hwecc;
922	932	nand->ecc.bytes = 3;
923	933	nand->ecc.strength = 1;

drivers/mtd/nand/raw/lpc32xx_slc.c+14−1

@@ -27,6 +27,7 @@ · 26 unmodified lines
27	27	#include <linux/of.h>
28	28	#include <linux/of_gpio.h>
29	29	#include <linux/mtd/lpc32xx_slc.h>
	30	+#include <linux/mtd/nand-ecc-sw-hamming.h>
30	31
31	32	#define LPC32XX_MODNAME "lpc32xx-nand"
32	33
@@ -344,6 +345,18 @@ static int lpc32xx_nand_ecc_calculate(struct nand_chip *chip, · 311 unmodified lines
344	345	return 0;
345	346	}
346	347
	348	+/*
	349	+ * Corrects the data
	350	+ */
	351	+static int lpc32xx_nand_ecc_correct(struct nand_chip *chip,
	352	+ unsigned char *buf,
	353	+ unsigned char *read_ecc,
	354	+ unsigned char *calc_ecc)
	355	+{
	356	+ return ecc_sw_hamming_correct(buf, read_ecc, calc_ecc,
	357	+ chip->ecc.size, false);
	358	+}
	359	+
347	360	/*
348	361	* Read a single byte from NAND device
349	362	*/
@@ -802,7 +815,7 @@ static int lpc32xx_nand_attach_chip(struct nand_chip *chip) · 452 unmodified lines
802	815	chip->ecc.write_oob = lpc32xx_nand_write_oob_syndrome;
803	816	chip->ecc.read_oob = lpc32xx_nand_read_oob_syndrome;
804	817	chip->ecc.calculate = lpc32xx_nand_ecc_calculate;
805		− chip->ecc.correct = rawnand_sw_hamming_correct;
	818	+ chip->ecc.correct = lpc32xx_nand_ecc_correct;
806	819	chip->ecc.hwctl = lpc32xx_nand_ecc_enable;
807	820
808	821	/*

drivers/mtd/nand/raw/ndfc.c+11−1

@@ -22,6 +22,7 @@ · 21 unmodified lines
22	22	#include <linux/mtd/ndfc.h>
23	23	#include <linux/slab.h>
24	24	#include <linux/mtd/mtd.h>
	25	+#include <linux/mtd/nand-ecc-sw-hamming.h>
25	26	#include <linux/of_address.h>
26	27	#include <linux/of_platform.h>
27	28	#include <asm/io.h>
@@ -100,6 +101,15 @@ static int ndfc_calculate_ecc(struct nand_chip *chip, · 72 unmodified lines
100	101	return 0;
101	102	}
102	103
	104	+static int ndfc_correct_ecc(struct nand_chip *chip,
	105	+ unsigned char *buf,
	106	+ unsigned char *read_ecc,
	107	+ unsigned char *calc_ecc)
	108	+{
	109	+ return ecc_sw_hamming_correct(buf, read_ecc, calc_ecc,
	110	+ chip->ecc.size, false);
	111	+}
	112	+
103	113	/*
104	114	* Speedups for buffer read/write/verify
105	115	*
@@ -145,7 +155,7 @@ static int ndfc_chip_init(struct ndfc_controller *ndfc, · 39 unmodified lines
145	155	chip->controller = &ndfc->ndfc_control;
146	156	chip->legacy.read_buf = ndfc_read_buf;
147	157	chip->legacy.write_buf = ndfc_write_buf;
148		− chip->ecc.correct = rawnand_sw_hamming_correct;
	158	+ chip->ecc.correct = ndfc_correct_ecc;
149	159	chip->ecc.hwctl = ndfc_enable_hwecc;
150	160	chip->ecc.calculate = ndfc_calculate_ecc;
151	161	chip->ecc.engine_type = NAND_ECC_ENGINE_TYPE_ON_HOST;

drivers/mtd/nand/raw/sharpsl.c+11−1

@@ -11,6 +11,7 @@ · 10 unmodified lines
11	11	#include <linux/module.h>
12	12	#include <linux/delay.h>
13	13	#include <linux/mtd/mtd.h>
	14	+#include <linux/mtd/nand-ecc-sw-hamming.h>
14	15	#include <linux/mtd/rawnand.h>
15	16	#include <linux/mtd/partitions.h>
16	17	#include <linux/mtd/sharpsl.h>
@@ -96,6 +97,15 @@ static int sharpsl_nand_calculate_ecc(struct nand_chip *chip, · 79 unmodified lines
96	97	return readb(sharpsl->io + ECCCNTR) != 0;
97	98	}
98	99
	100	+static int sharpsl_nand_correct_ecc(struct nand_chip *chip,
	101	+ unsigned char *buf,
	102	+ unsigned char *read_ecc,
	103	+ unsigned char *calc_ecc)
	104	+{
	105	+ return ecc_sw_hamming_correct(buf, read_ecc, calc_ecc,
	106	+ chip->ecc.size, false);
	107	+}
	108	+
99	109	static int sharpsl_attach_chip(struct nand_chip *chip)
100	110	{
101	111	if (chip->ecc.engine_type != NAND_ECC_ENGINE_TYPE_ON_HOST)
@@ -106,7 +116,7 @@ static int sharpsl_attach_chip(struct nand_chip *chip) · 4 unmodified lines
106	116	chip->ecc.strength = 1;
107	117	chip->ecc.hwctl = sharpsl_nand_enable_hwecc;
108	118	chip->ecc.calculate = sharpsl_nand_calculate_ecc;
109		− chip->ecc.correct = rawnand_sw_hamming_correct;
	119	+ chip->ecc.correct = sharpsl_nand_correct_ecc;
110	120
111	121	return 0;
112	122	}

drivers/mtd/nand/raw/tmio_nand.c+5−3

@@ -34,6 +34,7 @@ · 33 unmodified lines
34	34	#include <linux/interrupt.h>
35	35	#include <linux/ioport.h>
36	36	#include <linux/mtd/mtd.h>
	37	+#include <linux/mtd/nand-ecc-sw-hamming.h>
37	38	#include <linux/mtd/rawnand.h>
38	39	#include <linux/mtd/partitions.h>
39	40	#include <linux/slab.h>
@@ -292,11 +293,12 @@ static int tmio_nand_correct_data(struct nand_chip chip, unsigned char buf, · 252 unmodified lines
292	293	int r0, r1;
293	294
294	295	/* assume ecc.size = 512 and ecc.bytes = 6 */
295		− r0 = rawnand_sw_hamming_correct(chip, buf, read_ecc, calc_ecc);
	296	+ r0 = ecc_sw_hamming_correct(buf, read_ecc, calc_ecc,
	297	+ chip->ecc.size, false);
296	298	if (r0 < 0)
297	299	return r0;
298		− r1 = rawnand_sw_hamming_correct(chip, buf + 256, read_ecc + 3,
299		− calc_ecc + 3);
	300	+ r1 = ecc_sw_hamming_correct(buf + 256, read_ecc + 3, calc_ecc + 3,
	301	+ chip->ecc.size, false);
300	302	if (r1 < 0)
301	303	return r1;
302	304	return r0 + r1;

drivers/mtd/nand/raw/txx9ndfmc.c+3−2

@@ -13,6 +13,7 @@ · 12 unmodified lines
13	13	#include <linux/platform_device.h>
14	14	#include <linux/delay.h>
15	15	#include <linux/mtd/mtd.h>
	16	+#include <linux/mtd/nand-ecc-sw-hamming.h>
16	17	#include <linux/mtd/rawnand.h>
17	18	#include <linux/mtd/partitions.h>
18	19	#include <linux/io.h>
@@ -193,8 +194,8 @@ static int txx9ndfmc_correct_data(struct nand_chip chip, unsigned char buf, · 174 unmodified lines
193	194	int stat;
194	195
195	196	for (eccsize = chip->ecc.size; eccsize > 0; eccsize -= 256) {
196		− stat = rawnand_sw_hamming_correct(chip, buf, read_ecc,
197		− calc_ecc);
	197	+ stat = ecc_sw_hamming_correct(buf, read_ecc, calc_ecc,
	198	+ chip->ecc.size, false);
198	199	if (stat < 0)
199	200	return stat;
200	201	corrected += stat;

drivers/mtd/parsers/ofpart_core.c+14−12

@@ -57,20 +57,22 @@ static int parse_fixed_partitions(struct mtd_info *master, · 56 unmodified lines
57	57	if (!mtd_node)
58	58	return 0;
59	59
60		− ofpart_node = of_get_child_by_name(mtd_node, "partitions");
61		− if (!ofpart_node && !master->parent) {
62		− /*
63		− * We might get here even when ofpart isn't used at all (e.g.,
64		− * when using another parser), so don't be louder than
65		− * KERN_DEBUG
66		− */
67		− pr_debug("%s: 'partitions' subnode not found on %pOF. Trying to parse direct subnodes as partitions.\n",
68		− master->name, mtd_node);
	60	+ if (!master->parent) { /* Master */
	61	+ ofpart_node = of_get_child_by_name(mtd_node, "partitions");
	62	+ if (!ofpart_node) {
	63	+ /*
	64	+ * We might get here even when ofpart isn't used at all (e.g.,
	65	+ * when using another parser), so don't be louder than
	66	+ * KERN_DEBUG
	67	+ */
	68	+ pr_debug("%s: 'partitions' subnode not found on %pOF. Trying to parse direct subnodes as partitions.\n",
	69	+ master->name, mtd_node);
	70	+ ofpart_node = mtd_node;
	71	+ dedicated = false;
	72	+ }
	73	+ } else { /* Partition */
69	74	ofpart_node = mtd_node;
70		− dedicated = false;
71	75	}
72		− if (!ofpart_node)
73		− return 0;
74	76
75	77	of_id = of_match_node(parse_ofpart_match_table, ofpart_node);
76	78	if (dedicated && !of_id) {

drivers/net/appletalk/cops.c+2−2

@@ -327,6 +327,8 @@ static int __init cops_probe1(struct net_device *dev, int ioaddr) · 326 unmodified lines
327	327	break;
328	328	}
329	329
	330	+ dev->base_addr = ioaddr;
	331	+
330	332	/* Reserve any actual interrupt. */
331	333	if (dev->irq) {
332	334	retval = request_irq(dev->irq, cops_interrupt, 0, dev->name, dev);
@@ -334,8 +336,6 @@ static int __init cops_probe1(struct net_device *dev, int ioaddr) · 1 unmodified line
334	336	goto err_out;
335	337	}
336	338
337		− dev->base_addr = ioaddr;
338		−
339	339	lp = netdev_priv(dev);
340	340	spin_lock_init(&lp->lock);
341	341

drivers/net/bonding/bond_main.c+1−1

@@ -1526,6 +1526,7 @@ static struct slave bond_alloc_slave(struct bonding bond, · 1525 unmodified lines
1526	1526
1527	1527	slave->bond = bond;
1528	1528	slave->dev = slave_dev;
	1529	+ INIT_DELAYED_WORK(&slave->notify_work, bond_netdev_notify_work);
1529	1530
1530	1531	if (bond_kobj_init(slave))
1531	1532	return NULL;
@@ -1538,7 +1539,6 @@ static struct slave bond_alloc_slave(struct bonding bond, · 6 unmodified lines
1538	1539	return NULL;
1539	1540	}
1540	1541	}
1541		− INIT_DELAYED_WORK(&slave->notify_work, bond_netdev_notify_work);
1542	1542
1543	1543	return slave;
1544	1544	}

drivers/net/caif/caif_serial.c+1−0

@@ -350,6 +350,7 @@ static int ldisc_open(struct tty_struct *tty) · 349 unmodified lines
350	350	rtnl_lock();
351	351	result = register_netdevice(dev);
352	352	if (result) {
	353	+ tty_kref_put(tty);
353	354	rtnl_unlock();
354	355	free_netdev(dev);
355	356	return -ENODEV;

drivers/net/can/usb/mcba_usb.c+15−2

@@ -82,6 +82,8 @@ struct mcba_priv { · 81 unmodified lines
82	82	bool can_ka_first_pass;
83	83	bool can_speed_check;
84	84	atomic_t free_ctx_cnt;
	85	+ void *rxbuf[MCBA_MAX_RX_URBS];
	86	+ dma_addr_t rxbuf_dma[MCBA_MAX_RX_URBS];
85	87	};
86	88
87	89	/* CAN frame */
@@ -633,6 +635,7 @@ static int mcba_usb_start(struct mcba_priv *priv) · 545 unmodified lines
633	635	for (i = 0; i < MCBA_MAX_RX_URBS; i++) {
634	636	struct urb *urb = NULL;
635	637	u8 *buf;
	638	+ dma_addr_t buf_dma;
636	639
637	640	/* create a URB, and a buffer for it */
638	641	urb = usb_alloc_urb(0, GFP_KERNEL);
@@ -642,7 +645,7 @@ static int mcba_usb_start(struct mcba_priv *priv) · 3 unmodified lines
642	645	}
643	646
644	647	buf = usb_alloc_coherent(priv->udev, MCBA_USB_RX_BUFF_SIZE,
645		− GFP_KERNEL, &urb->transfer_dma);
	648	+ GFP_KERNEL, &buf_dma);
646	649	if (!buf) {
647	650	netdev_err(netdev, "No memory left for USB buffer\n");
648	651	usb_free_urb(urb);
@@ -661,11 +664,14 @@ static int mcba_usb_start(struct mcba_priv *priv) · 12 unmodified lines
661	664	if (err) {
662	665	usb_unanchor_urb(urb);
663	666	usb_free_coherent(priv->udev, MCBA_USB_RX_BUFF_SIZE,
664		− buf, urb->transfer_dma);
	667	+ buf, buf_dma);
665	668	usb_free_urb(urb);
666	669	break;
667	670	}
668	671
	672	+ priv->rxbuf[i] = buf;
	673	+ priv->rxbuf_dma[i] = buf_dma;
	674	+
669	675	/* Drop reference, USB core will take care of freeing it */
670	676	usb_free_urb(urb);
671	677	}
@@ -708,7 +714,14 @@ static int mcba_usb_open(struct net_device *netdev) · 36 unmodified lines
708	714
709	715	static void mcba_urb_unlink(struct mcba_priv *priv)
710	716	{
	717	+ int i;
	718	+
711	719	usb_kill_anchored_urbs(&priv->rx_submitted);
	720	+
	721	+ for (i = 0; i < MCBA_MAX_RX_URBS; ++i)
	722	+ usb_free_coherent(priv->udev, MCBA_USB_RX_BUFF_SIZE,
	723	+ priv->rxbuf[i], priv->rxbuf_dma[i]);
	724	+
712	725	usb_kill_anchored_urbs(&priv->tx_submitted);
713	726	}
714	727

drivers/net/dsa/bcm_sf2.c+2−3

@@ -821,11 +821,9 @@ static void bcm_sf2_sw_mac_link_up(struct dsa_switch *ds, int port, · 820 unmodified lines
821	821	bcm_sf2_sw_mac_link_set(ds, port, interface, true);
822	822
823	823	if (port != core_readl(priv, CORE_IMP0_PRT_ID)) {
824		− u32 reg_rgmii_ctrl;
	824	+ u32 reg_rgmii_ctrl = 0;
825	825	u32 reg, offset;
826	826
827		− reg_rgmii_ctrl = bcm_sf2_reg_rgmii_cntrl(priv, port);
828		−
829	827	if (priv->type == BCM4908_DEVICE_ID \|\|
830	828	priv->type == BCM7445_DEVICE_ID)
831	829	offset = CORE_STS_OVERRIDE_GMIIP_PORT(port);
@@ -836,6 +834,7 @@ static void bcm_sf2_sw_mac_link_up(struct dsa_switch *ds, int port, · 4 unmodified lines
836	834	interface == PHY_INTERFACE_MODE_RGMII_TXID \|\|
837	835	interface == PHY_INTERFACE_MODE_MII \|\|
838	836	interface == PHY_INTERFACE_MODE_REVMII) {
	837	+ reg_rgmii_ctrl = bcm_sf2_reg_rgmii_cntrl(priv, port);
839	838	reg = reg_readl(priv, reg_rgmii_ctrl);
840	839	reg &= ~(RX_PAUSE_EN \| TX_PAUSE_EN);
841	840

drivers/net/dsa/microchip/ksz9477.c+1−0

@@ -1530,6 +1530,7 @@ static const struct ksz_chip_data ksz9477_switch_chips[] = { · 1529 unmodified lines
1530	1530	.num_statics = 16,
1531	1531	.cpu_ports = 0x7F, /* can be configured as cpu port */
1532	1532	.port_cnt = 7, /* total physical port count */
	1533	+ .phy_errata_9477 = true,
1533	1534	},
1534	1535	};
1535	1536

drivers/net/dsa/mt7530.c+0−8

@@ -1262,14 +1262,6 @@ mt7530_port_set_vlan_aware(struct dsa_switch *ds, int port) · 1261 unmodified lines
1262	1262	{
1263	1263	struct mt7530_priv *priv = ds->priv;
1264	1264
1265		− /* The real fabric path would be decided on the membership in the
1266		− * entry of VLAN table. PCR_MATRIX set up here with ALL_MEMBERS
1267		− * means potential VLAN can be consisting of certain subset of all
1268		− * ports.
1269		− */
1270		− mt7530_rmw(priv, MT7530_PCR_P(port),
1271		− PCR_MATRIX_MASK, PCR_MATRIX(MT7530_ALL_MEMBERS));
1272		−
1273	1265	/* Trapped into security mode allows packet forwarding through VLAN
1274	1266	* table lookup. CPU port is set to fallback mode to let untagged
1275	1267	* frames pass through.

drivers/net/dsa/ocelot/felix_vsc9959.c+10−5

@@ -1227,12 +1227,17 @@ static int vsc9959_qos_port_tas_set(struct ocelot *ocelot, int port, · 1226 unmodified lines
1227	1227	if (taprio->num_entries > VSC9959_TAS_GCL_ENTRY_MAX)
1228	1228	return -ERANGE;
1229	1229
1230		− /* Set port num and disable ALWAYS_GUARD_BAND_SCH_Q, which means set
1231		− * guard band to be implemented for nonschedule queues to schedule
1232		− * queues transition.
	1230	+ /* Enable guard band. The switch will schedule frames without taking
	1231	+ * their length into account. Thus we'll always need to enable the
	1232	+ * guard band which reserves the time of a maximum sized frame at the
	1233	+ * end of the time window.
	1234	+ *
	1235	+ * Although the ALWAYS_GUARD_BAND_SCH_Q bit is global for all ports, we
	1236	+ * need to set PORT_NUM, because subsequent writes to PARAM_CFG_REG_n
	1237	+ * operate on the port number.
1233	1238	*/
1234		− ocelot_rmw(ocelot,
1235		− QSYS_TAS_PARAM_CFG_CTRL_PORT_NUM(port),
	1239	+ ocelot_rmw(ocelot, QSYS_TAS_PARAM_CFG_CTRL_PORT_NUM(port) \|
	1240	+ QSYS_TAS_PARAM_CFG_CTRL_ALWAYS_GUARD_BAND_SCH_Q,
1236	1241	QSYS_TAS_PARAM_CFG_CTRL_PORT_NUM_M \|
1237	1242	QSYS_TAS_PARAM_CFG_CTRL_ALWAYS_GUARD_BAND_SCH_Q,
1238	1243	QSYS_TAS_PARAM_CFG_CTRL);

drivers/net/dsa/sja1105/sja1105_dynamic_config.c+19−4

@@ -167,9 +167,10 @@ enum sja1105_hostcmd { · 166 unmodified lines
167	167	SJA1105_HOSTCMD_INVALIDATE = 4,
168	168	};
169	169
	170	+/* Command and entry overlap */
170	171	static void
171		−sja1105_vl_lookup_cmd_packing(void buf, struct sja1105_dyn_cmd cmd,
172		− enum packing_op op)
	172	+sja1105et_vl_lookup_cmd_packing(void buf, struct sja1105_dyn_cmd cmd,
	173	+ enum packing_op op)
173	174	{
174	175	const int size = SJA1105_SIZE_DYN_CMD;
175	176
@@ -179,6 +180,20 @@ sja1105_vl_lookup_cmd_packing(void buf, struct sja1105_dyn_cmd cmd, · 3 unmodified lines
179	180	sja1105_packing(buf, &cmd->index, 9, 0, size, op);
180	181	}
181	182
	183	+/* Command and entry are separate */
	184	+static void
	185	+sja1105pqrs_vl_lookup_cmd_packing(void buf, struct sja1105_dyn_cmd cmd,
	186	+ enum packing_op op)
	187	+{
	188	+ u8 *p = buf + SJA1105_SIZE_VL_LOOKUP_ENTRY;
	189	+ const int size = SJA1105_SIZE_DYN_CMD;
	190	+
	191	+ sja1105_packing(p, &cmd->valid, 31, 31, size, op);
	192	+ sja1105_packing(p, &cmd->errors, 30, 30, size, op);
	193	+ sja1105_packing(p, &cmd->rdwrset, 29, 29, size, op);
	194	+ sja1105_packing(p, &cmd->index, 9, 0, size, op);
	195	+}
	196	+
182	197	static size_t sja1105et_vl_lookup_entry_packing(void buf, void entry_ptr,
183	198	enum packing_op op)
184	199	{
@@ -641,7 +656,7 @@ static size_t sja1105pqrs_cbs_entry_packing(void buf, void entry_ptr, · 456 unmodified lines
641	656	const struct sja1105_dynamic_table_ops sja1105et_dyn_ops[BLK_IDX_MAX_DYN] = {
642	657	[BLK_IDX_VL_LOOKUP] = {
643	658	.entry_packing = sja1105et_vl_lookup_entry_packing,
644		− .cmd_packing = sja1105_vl_lookup_cmd_packing,
	659	+ .cmd_packing = sja1105et_vl_lookup_cmd_packing,
645	660	.access = OP_WRITE,
646	661	.max_entry_count = SJA1105_MAX_VL_LOOKUP_COUNT,
647	662	.packed_size = SJA1105ET_SIZE_VL_LOOKUP_DYN_CMD,
@@ -725,7 +740,7 @@ const struct sja1105_dynamic_table_ops sja1105et_dyn_ops[BLK_IDX_MAX_DYN] = { · 77 unmodified lines
725	740	const struct sja1105_dynamic_table_ops sja1105pqrs_dyn_ops[BLK_IDX_MAX_DYN] = {
726	741	[BLK_IDX_VL_LOOKUP] = {
727	742	.entry_packing = sja1105_vl_lookup_entry_packing,
728		− .cmd_packing = sja1105_vl_lookup_cmd_packing,
	743	+ .cmd_packing = sja1105pqrs_vl_lookup_cmd_packing,
729	744	.access = (OP_READ \| OP_WRITE),
730	745	.max_entry_count = SJA1105_MAX_VL_LOOKUP_COUNT,
731	746	.packed_size = SJA1105PQRS_SIZE_VL_LOOKUP_DYN_CMD,

drivers/net/dsa/sja1105/sja1105_main.c+51−23

@@ -26,6 +26,7 @@ · 25 unmodified lines
26	26	#include "sja1105_tas.h"
27	27
28	28	#define SJA1105_UNKNOWN_MULTICAST 0x010000000000ull
	29	+#define SJA1105_DEFAULT_VLAN (VLAN_N_VID - 1)
29	30
30	31	static const struct dsa_switch_ops sja1105_switch_ops;
31	32
@@ -207,6 +208,7 @@ static int sja1105_init_mii_settings(struct sja1105_private *priv, · 175 unmodified lines
207	208	default:
208	209	dev_err(dev, "Unsupported PHY mode %s!\n",
209	210	phy_modes(ports[i].phy_mode));
	211	+ return -EINVAL;
210	212	}
211	213
212	214	/* Even though the SerDes port is able to drive SGMII autoneg
@@ -321,6 +323,13 @@ static int sja1105_init_l2_lookup_params(struct sja1105_private *priv) · 108 unmodified lines
321	323	return 0;
322	324	}
323	325
	326	+/* Set up a default VLAN for untagged traffic injected from the CPU
	327	+ * using management routes (e.g. STP, PTP) as opposed to tag_8021q.
	328	+ * All DT-defined ports are members of this VLAN, and there are no
	329	+ * restrictions on forwarding (since the CPU selects the destination).
	330	+ * Frames from this VLAN will always be transmitted as untagged, and
	331	+ * neither the bridge nor the 8021q module cannot create this VLAN ID.
	332	+ */
324	333	static int sja1105_init_static_vlan(struct sja1105_private *priv)
325	334	{
326	335	struct sja1105_table *table;
@@ -330,17 +339,13 @@ static int sja1105_init_static_vlan(struct sja1105_private *priv) · 3 unmodified lines
330	339	.vmemb_port = 0,
331	340	.vlan_bc = 0,
332	341	.tag_port = 0,
333		− .vlanid = 1,
	342	+ .vlanid = SJA1105_DEFAULT_VLAN,
334	343	};
335	344	struct dsa_switch *ds = priv->ds;
336	345	int port;
337	346
338	347	table = &priv->static_config.tables[BLK_IDX_VLAN_LOOKUP];
339	348
340		− /* The static VLAN table will only contain the initial pvid of 1.
341		− * All other VLANs are to be configured through dynamic entries,
342		− * and kept in the static configuration table as backing memory.
343		− */
344	349	if (table->entry_count) {
345	350	kfree(table->entries);
346	351	table->entry_count = 0;
@@ -353,9 +358,6 @@ static int sja1105_init_static_vlan(struct sja1105_private *priv) · 6 unmodified lines
353	358
354	359	table->entry_count = 1;
355	360
356		− /* VLAN 1: all DT-defined ports are members; no restrictions on
357		− * forwarding; always transmit as untagged.
358		− */
359	361	for (port = 0; port < ds->num_ports; port++) {
360	362	struct sja1105_bridge_vlan *v;
361	363
@@ -366,15 +368,12 @@ static int sja1105_init_static_vlan(struct sja1105_private *priv) · 4 unmodified lines
366	368	pvid.vlan_bc \|= BIT(port);
367	369	pvid.tag_port &= ~BIT(port);
368	370
369		− /* Let traffic that don't need dsa_8021q (e.g. STP, PTP) be
370		− * transmitted as untagged.
371		− */
372	371	v = kzalloc(sizeof(*v), GFP_KERNEL);
373	372	if (!v)
374	373	return -ENOMEM;
375	374
376	375	v->port = port;
377		− v->vid = 1;
	376	+ v->vid = SJA1105_DEFAULT_VLAN;
378	377	v->untagged = true;
379	378	if (dsa_is_cpu_port(ds, port))
380	379	v->pvid = true;
@@ -2817,11 +2816,22 @@ static int sja1105_vlan_add_one(struct dsa_switch *ds, int port, u16 vid, · 2436 unmodified lines
2817	2816	bool pvid = flags & BRIDGE_VLAN_INFO_PVID;
2818	2817	struct sja1105_bridge_vlan *v;
2819	2818
2820		− list_for_each_entry(v, vlan_list, list)
2821		− if (v->port == port && v->vid == vid &&
2822		− v->untagged == untagged && v->pvid == pvid)
	2819	+ list_for_each_entry(v, vlan_list, list) {
	2820	+ if (v->port == port && v->vid == vid) {
2823	2821	/* Already added */
2824		− return 0;
	2822	+ if (v->untagged == untagged && v->pvid == pvid)
	2823	+ /* Nothing changed */
	2824	+ return 0;
	2825	+
	2826	+ /* It's the same VLAN, but some of the flags changed
	2827	+ * and the user did not bother to delete it first.
	2828	+ * Update it and trigger sja1105_build_vlan_table.
	2829	+ */
	2830	+ v->untagged = untagged;
	2831	+ v->pvid = pvid;
	2832	+ return 1;
	2833	+ }
	2834	+ }
2825	2835
2826	2836	v = kzalloc(sizeof(*v), GFP_KERNEL);
2827	2837	if (!v) {
@@ -2976,13 +2986,13 @@ static int sja1105_setup(struct dsa_switch *ds) · 148 unmodified lines
2976	2986	rc = sja1105_static_config_load(priv, ports);
2977	2987	if (rc < 0) {
2978	2988	dev_err(ds->dev, "Failed to load static config: %d\n", rc);
2979		− return rc;
	2989	+ goto out_ptp_clock_unregister;
2980	2990	}
2981	2991	/* Configure the CGU (PHY link modes and speeds) */
2982	2992	rc = sja1105_clocking_setup(priv);
2983	2993	if (rc < 0) {
2984	2994	dev_err(ds->dev, "Failed to configure MII clocking: %d\n", rc);
2985		− return rc;
	2995	+ goto out_static_config_free;
2986	2996	}
2987	2997	/* On SJA1105, VLAN filtering per se is always enabled in hardware.
2988	2998	* The only thing we can do to disable it is lie about what the 802.1Q
@@ -3003,7 +3013,7 @@ static int sja1105_setup(struct dsa_switch *ds) · 14 unmodified lines
3003	3013
3004	3014	rc = sja1105_devlink_setup(ds);
3005	3015	if (rc < 0)
3006		− return rc;
	3016	+ goto out_static_config_free;
3007	3017
3008	3018	/* The DSA/switchdev model brings up switch ports in standalone mode by
3009	3019	* default, and that means vlan_filtering is 0 since they're not under
@@ -3012,6 +3022,17 @@ static int sja1105_setup(struct dsa_switch *ds) · 2 unmodified lines
3012	3022	rtnl_lock();
3013	3023	rc = sja1105_setup_8021q_tagging(ds, true);
3014	3024	rtnl_unlock();
	3025	+ if (rc)
	3026	+ goto out_devlink_teardown;
	3027	+
	3028	+ return 0;
	3029	+
	3030	+out_devlink_teardown:
	3031	+ sja1105_devlink_teardown(ds);
	3032	+out_ptp_clock_unregister:
	3033	+ sja1105_ptp_clock_unregister(ds);
	3034	+out_static_config_free:
	3035	+ sja1105_static_config_free(&priv->static_config);
3015	3036
3016	3037	return rc;
3017	3038	}
@@ -3646,8 +3667,10 @@ static int sja1105_probe(struct spi_device *spi) · 628 unmodified lines
3646	3667	priv->cbs = devm_kcalloc(dev, priv->info->num_cbs_shapers,
3647	3668	sizeof(struct sja1105_cbs_entry),
3648	3669	GFP_KERNEL);
3649		− if (!priv->cbs)
3650		− return -ENOMEM;
	3670	+ if (!priv->cbs) {
	3671	+ rc = -ENOMEM;
	3672	+ goto out_unregister_switch;
	3673	+ }
3651	3674	}
3652	3675
3653	3676	/* Connections between dsa_port and sja1105_port */
@@ -3672,7 +3695,7 @@ static int sja1105_probe(struct spi_device *spi) · 18 unmodified lines
3672	3695	dev_err(ds->dev,
3673	3696	"failed to create deferred xmit thread: %d\n",
3674	3697	rc);
3675		− goto out;
	3698	+ goto out_destroy_workers;
3676	3699	}
3677	3700	skb_queue_head_init(&sp->xmit_queue);
3678	3701	sp->xmit_tpid = ETH_P_SJA1105;
@@ -3682,7 +3705,8 @@ static int sja1105_probe(struct spi_device *spi) · 3 unmodified lines
3682	3705	}
3683	3706
3684	3707	return 0;
3685		−out:
	3708	+
	3709	+out_destroy_workers:
3686	3710	while (port-- > 0) {
3687	3711	struct sja1105_port *sp = &priv->ports[port];
3688	3712
@@ -3691,6 +3715,10 @@ out: · 2 unmodified lines
3691	3715
3692	3716	kthread_destroy_worker(sp->xmit_worker);
3693	3717	}
	3718	+
	3719	+out_unregister_switch:
	3720	+ dsa_unregister_switch(ds);
	3721	+
3694	3722	return rc;
3695	3723	}
3696	3724

drivers/net/ethernet/amazon/ena/ena_netdev.c+28−26

@@ -236,36 +236,48 @@ static int ena_xdp_io_poll(struct napi_struct *napi, int budget) · 235 unmodified lines
236	236	static int ena_xdp_tx_map_frame(struct ena_ring *xdp_ring,
237	237	struct ena_tx_buffer *tx_info,
238	238	struct xdp_frame *xdpf,
239		− void **push_hdr,
240		− u32 *push_len)
	239	+ struct ena_com_tx_ctx *ena_tx_ctx)
241	240	{
242	241	struct ena_adapter *adapter = xdp_ring->adapter;
243	242	struct ena_com_buf *ena_buf;
244		− dma_addr_t dma = 0;
	243	+ int push_len = 0;
	244	+ dma_addr_t dma;
	245	+ void *data;
245	246	u32 size;
246	247
247	248	tx_info->xdpf = xdpf;
	249	+ data = tx_info->xdpf->data;
248	250	size = tx_info->xdpf->len;
249		− ena_buf = tx_info->bufs;
250	251
251		− /* llq push buffer */
252		− *push_len = min_t(u32, size, xdp_ring->tx_max_header_size);
253		− *push_hdr = tx_info->xdpf->data;
	252	+ if (xdp_ring->tx_mem_queue_type == ENA_ADMIN_PLACEMENT_POLICY_DEV) {
	253	+ /* Designate part of the packet for LLQ */
	254	+ push_len = min_t(u32, size, xdp_ring->tx_max_header_size);
	255	+
	256	+ ena_tx_ctx->push_header = data;
	257	+
	258	+ size -= push_len;
	259	+ data += push_len;
	260	+ }
	261	+
	262	+ ena_tx_ctx->header_len = push_len;
254	263
255		− if (size - *push_len > 0) {
	264	+ if (size > 0) {
256	265	dma = dma_map_single(xdp_ring->dev,
257		− push_hdr + push_len,
258		− size - *push_len,
	266	+ data,
	267	+ size,
259	268	DMA_TO_DEVICE);
260	269	if (unlikely(dma_mapping_error(xdp_ring->dev, dma)))
261	270	goto error_report_dma_error;
262	271
263		− tx_info->map_linear_data = 1;
264		− tx_info->num_of_bufs = 1;
265		− }
	272	+ tx_info->map_linear_data = 0;
266	273
267		− ena_buf->paddr = dma;
268		− ena_buf->len = size;
	274	+ ena_buf = tx_info->bufs;
	275	+ ena_buf->paddr = dma;
	276	+ ena_buf->len = size;
	277	+
	278	+ ena_tx_ctx->ena_bufs = ena_buf;
	279	+ ena_tx_ctx->num_bufs = tx_info->num_of_bufs = 1;
	280	+ }
269	281
270	282	return 0;
271	283
@@ -274,10 +286,6 @@ error_report_dma_error: · 2 unmodified lines
274	286	&xdp_ring->syncp);
275	287	netif_warn(adapter, tx_queued, adapter->netdev, "Failed to map xdp buff\n");
276	288
277		− xdp_return_frame_rx_napi(tx_info->xdpf);
278		− tx_info->xdpf = NULL;
279		− tx_info->num_of_bufs = 0;
280		−
281	289	return -EINVAL;
282	290	}
283	291
@@ -289,8 +297,6 @@ static int ena_xdp_xmit_frame(struct ena_ring *xdp_ring, · 5 unmodified lines
289	297	struct ena_com_tx_ctx ena_tx_ctx = {};
290	298	struct ena_tx_buffer *tx_info;
291	299	u16 next_to_use, req_id;
292		− void *push_hdr;
293		− u32 push_len;
294	300	int rc;
295	301
296	302	next_to_use = xdp_ring->next_to_use;
@@ -298,15 +304,11 @@ static int ena_xdp_xmit_frame(struct ena_ring *xdp_ring, · 1 unmodified line
298	304	tx_info = &xdp_ring->tx_buffer_info[req_id];
299	305	tx_info->num_of_bufs = 0;
300	306
301		− rc = ena_xdp_tx_map_frame(xdp_ring, tx_info, xdpf, &push_hdr, &push_len);
	307	+ rc = ena_xdp_tx_map_frame(xdp_ring, tx_info, xdpf, &ena_tx_ctx);
302	308	if (unlikely(rc))
303	309	return rc;
304	310
305		− ena_tx_ctx.ena_bufs = tx_info->bufs;
306		− ena_tx_ctx.push_header = push_hdr;
307		− ena_tx_ctx.num_bufs = tx_info->num_of_bufs;
308	311	ena_tx_ctx.req_id = req_id;
309		− ena_tx_ctx.header_len = push_len;
310	312
311	313	rc = ena_xmit_common(dev,
312	314	xdp_ring,

drivers/net/ethernet/atheros/alx/main.c+1−0

@@ -1849,6 +1849,7 @@ out_free_netdev: · 1848 unmodified lines
1849	1849	free_netdev(netdev);
1850	1850	out_pci_release:
1851	1851	pci_release_mem_regions(pdev);
	1852	+ pci_disable_pcie_error_reporting(pdev);
1852	1853	out_pci_disable:
1853	1854	pci_disable_device(pdev);
1854	1855	return err;

drivers/net/ethernet/broadcom/bnx2.c+1−1

@@ -8247,9 +8247,9 @@ bnx2_init_board(struct pci_dev pdev, struct net_device dev) · 8246 unmodified lines
8247	8247	BNX2_WR(bp, PCI_COMMAND, reg);
8248	8248	} else if ((BNX2_CHIP_ID(bp) == BNX2_CHIP_ID_5706_A1) &&
8249	8249	!(bp->flags & BNX2_FLAG_PCIX)) {
8250		−
8251	8250	dev_err(&pdev->dev,
8252	8251	"5706 A1 can only be used in a PCIX bus, aborting\n");
	8252	+ rc = -EPERM;
8253	8253	goto err_out_unmap;
8254	8254	}
8255	8255

drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c+3−1

@@ -1223,8 +1223,10 @@ int bnx2x_iov_init_one(struct bnx2x *bp, int int_mode_param, · 1222 unmodified lines
1223	1223	goto failed;
1224	1224
1225	1225	/* SR-IOV capability was enabled but there are no VFs*/
1226		− if (iov->total == 0)
	1226	+ if (iov->total == 0) {
	1227	+ err = -EINVAL;
1227	1228	goto failed;
	1229	+ }
1228	1230
1229	1231	iov->nr_virtfn = min_t(u16, iov->total, num_vfs_param);
1230	1232

drivers/net/ethernet/broadcom/bnxt/bnxt.c+117−29

@@ -282,7 +282,8 @@ static bool bnxt_vf_pciid(enum board_idx idx) · 281 unmodified lines
282	282	{
283	283	return (idx == NETXTREME_C_VF \|\| idx == NETXTREME_E_VF \|\|
284	284	idx == NETXTREME_S_VF \|\| idx == NETXTREME_C_VF_HV \|\|
285		− idx == NETXTREME_E_VF_HV \|\| idx == NETXTREME_E_P5_VF);
	285	+ idx == NETXTREME_E_VF_HV \|\| idx == NETXTREME_E_P5_VF \|\|
	286	+ idx == NETXTREME_E_P5_VF_HV);
286	287	}
287	288
288	289	#define DB_CP_REARM_FLAGS (DB_KEY_CP \| DB_IDX_VALID)
@@ -6932,17 +6933,10 @@ ctx_err: · 6643 unmodified lines
6932	6933	static void bnxt_hwrm_set_pg_attr(struct bnxt_ring_mem_info rmem, u8 pg_attr,
6933	6934	__le64 *pg_dir)
6934	6935	{
6935		− u8 pg_size = 0;
6936		−
6937	6936	if (!rmem->nr_pages)
6938	6937	return;
6939	6938
6940		− if (BNXT_PAGE_SHIFT == 13)
6941		− pg_size = 1 << 4;
6942		− else if (BNXT_PAGE_SIZE == 16)
6943		− pg_size = 2 << 4;
6944		−
6945		− *pg_attr = pg_size;
	6939	+ BNXT_SET_CTX_PAGE_ATTR(*pg_attr);
6946	6940	if (rmem->depth >= 1) {
6947	6941	if (rmem->depth == 2)
6948	6942	*pg_attr \|= 2;
@@ -7314,7 +7308,7 @@ skip_rdma: · 365 unmodified lines
7314	7308	entries_sp = ctx->vnic_max_vnic_entries + ctx->qp_max_l2_entries +
7315	7309	2 * (extra_qps + ctx->qp_min_qp1_entries) + min;
7316	7310	entries_sp = roundup(entries_sp, ctx->tqm_entries_multiple);
7317		− entries = ctx->qp_max_l2_entries + extra_qps + ctx->qp_min_qp1_entries;
	7311	+ entries = ctx->qp_max_l2_entries + 2 * (extra_qps + ctx->qp_min_qp1_entries);
7318	7312	entries = roundup(entries, ctx->tqm_entries_multiple);
7319	7313	entries = clamp_t(u32, entries, min, ctx->tqm_max_entries_per_ring);
7320	7314	for (i = 0; i < ctx->tqm_fp_rings_count + 1; i++) {
@@ -10785,37 +10779,125 @@ static int bnxt_set_features(struct net_device *dev, netdev_features_t features) · 3464 unmodified lines
10785	10779	return rc;
10786	10780	}
10787	10781
	10782	+static bool bnxt_exthdr_check(struct bnxt bp, struct sk_buff skb, int nw_off,
	10783	+ u8 **nextp)
	10784	+{
	10785	+ struct ipv6hdr ip6h = (struct ipv6hdr )(skb->data + nw_off);
	10786	+ int hdr_count = 0;
	10787	+ u8 *nexthdr;
	10788	+ int start;
	10789	+
	10790	+ /* Check that there are at most 2 IPv6 extension headers, no
	10791	+ * fragment header, and each is <= 64 bytes.
	10792	+ */
	10793	+ start = nw_off + sizeof(*ip6h);
	10794	+ nexthdr = &ip6h->nexthdr;
	10795	+ while (ipv6_ext_hdr(*nexthdr)) {
	10796	+ struct ipv6_opt_hdr *hp;
	10797	+ int hdrlen;
	10798	+
	10799	+ if (hdr_count >= 3 \|\| *nexthdr == NEXTHDR_NONE \|\|
	10800	+ *nexthdr == NEXTHDR_FRAGMENT)
	10801	+ return false;
	10802	+ hp = __skb_header_pointer(NULL, start, sizeof(*hp), skb->data,
	10803	+ skb_headlen(skb), NULL);
	10804	+ if (!hp)
	10805	+ return false;
	10806	+ if (*nexthdr == NEXTHDR_AUTH)
	10807	+ hdrlen = ipv6_authlen(hp);
	10808	+ else
	10809	+ hdrlen = ipv6_optlen(hp);
	10810	+
	10811	+ if (hdrlen > 64)
	10812	+ return false;
	10813	+ nexthdr = &hp->nexthdr;
	10814	+ start += hdrlen;
	10815	+ hdr_count++;
	10816	+ }
	10817	+ if (nextp) {
	10818	+ /* Caller will check inner protocol */
	10819	+ if (skb->encapsulation) {
	10820	+ *nextp = nexthdr;
	10821	+ return true;
	10822	+ }
	10823	+ *nextp = NULL;
	10824	+ }
	10825	+ /* Only support TCP/UDP for non-tunneled ipv6 and inner ipv6 */
	10826	+ return nexthdr == IPPROTO_TCP \|\| nexthdr == IPPROTO_UDP;
	10827	+}
	10828	+
	10829	+/* For UDP, we can only handle 1 Vxlan port and 1 Geneve port. */
	10830	+static bool bnxt_udp_tunl_check(struct bnxt bp, struct sk_buff skb)
	10831	+{
	10832	+ struct udphdr *uh = udp_hdr(skb);
	10833	+ __be16 udp_port = uh->dest;
	10834	+
	10835	+ if (udp_port != bp->vxlan_port && udp_port != bp->nge_port)
	10836	+ return false;
	10837	+ if (skb->inner_protocol_type == ENCAP_TYPE_ETHER) {
	10838	+ struct ethhdr *eh = inner_eth_hdr(skb);
	10839	+
	10840	+ switch (eh->h_proto) {
	10841	+ case htons(ETH_P_IP):
	10842	+ return true;
	10843	+ case htons(ETH_P_IPV6):
	10844	+ return bnxt_exthdr_check(bp, skb,
	10845	+ skb_inner_network_offset(skb),
	10846	+ NULL);
	10847	+ }
	10848	+ }
	10849	+ return false;
	10850	+}
	10851	+
	10852	+static bool bnxt_tunl_check(struct bnxt bp, struct sk_buff skb, u8 l4_proto)
	10853	+{
	10854	+ switch (l4_proto) {
	10855	+ case IPPROTO_UDP:
	10856	+ return bnxt_udp_tunl_check(bp, skb);
	10857	+ case IPPROTO_IPIP:
	10858	+ return true;
	10859	+ case IPPROTO_GRE: {
	10860	+ switch (skb->inner_protocol) {
	10861	+ default:
	10862	+ return false;
	10863	+ case htons(ETH_P_IP):
	10864	+ return true;
	10865	+ case htons(ETH_P_IPV6):
	10866	+ fallthrough;
	10867	+ }
	10868	+ }
	10869	+ case IPPROTO_IPV6:
	10870	+ /* Check ext headers of inner ipv6 */
	10871	+ return bnxt_exthdr_check(bp, skb, skb_inner_network_offset(skb),
	10872	+ NULL);
	10873	+ }
	10874	+ return false;
	10875	+}
	10876	+
10788	10877	static netdev_features_t bnxt_features_check(struct sk_buff *skb,
10789	10878	struct net_device *dev,
10790	10879	netdev_features_t features)
10791	10880	{
10792		− struct bnxt *bp;
10793		− __be16 udp_port;
10794		− u8 l4_proto = 0;
	10881	+ struct bnxt *bp = netdev_priv(dev);
	10882	+ u8 *l4_proto;
10795	10883
10796	10884	features = vlan_features_check(skb, features);
10797		− if (!skb->encapsulation)
10798		− return features;
10799		−
10800	10885	switch (vlan_get_protocol(skb)) {
10801	10886	case htons(ETH_P_IP):
10802		− l4_proto = ip_hdr(skb)->protocol;
	10887	+ if (!skb->encapsulation)
	10888	+ return features;
	10889	+ l4_proto = &ip_hdr(skb)->protocol;
	10890	+ if (bnxt_tunl_check(bp, skb, *l4_proto))
	10891	+ return features;
10803	10892	break;
10804	10893	case htons(ETH_P_IPV6):
10805		− l4_proto = ipv6_hdr(skb)->nexthdr;
	10894	+ if (!bnxt_exthdr_check(bp, skb, skb_network_offset(skb),
	10895	+ &l4_proto))
	10896	+ break;
	10897	+ if (!l4_proto \|\| bnxt_tunl_check(bp, skb, *l4_proto))
	10898	+ return features;
10806	10899	break;
10807		− default:
10808		− return features;
10809	10900	}
10810		−
10811		− if (l4_proto != IPPROTO_UDP)
10812		− return features;
10813		−
10814		− bp = netdev_priv(dev);
10815		− /* For UDP, we can only handle 1 Vxlan port and 1 Geneve port. */
10816		− udp_port = udp_hdr(skb)->dest;
10817		− if (udp_port == bp->vxlan_port \|\| udp_port == bp->nge_port)
10818		− return features;
10819	10901	return features & ~(NETIF_F_CSUM_MASK \| NETIF_F_GSO_MASK);
10820	10902	}
10821	10903
@@ -11668,6 +11750,8 @@ static void bnxt_fw_init_one_p3(struct bnxt *bp) · 846 unmodified lines
11668	11750	bnxt_hwrm_coal_params_qcaps(bp);
11669	11751	}
11670	11752
	11753	+static int bnxt_probe_phy(struct bnxt *bp, bool fw_dflt);
	11754	+
11671	11755	static int bnxt_fw_init_one(struct bnxt *bp)
11672	11756	{
11673	11757	int rc;
@@ -11682,6 +11766,9 @@ static int bnxt_fw_init_one(struct bnxt *bp) · 8 unmodified lines
11682	11766	netdev_err(bp->dev, "Firmware init phase 2 failed\n");
11683	11767	return rc;
11684	11768	}
	11769	+ rc = bnxt_probe_phy(bp, false);
	11770	+ if (rc)
	11771	+ return rc;
11685	11772	rc = bnxt_approve_mac(bp, bp->dev->dev_addr, false);
11686	11773	if (rc)
11687	11774	return rc;
@@ -13073,6 +13160,7 @@ init_err_pci_clean: · 1385 unmodified lines
13073	13160	bnxt_hwrm_func_drv_unrgtr(bp);
13074	13161	bnxt_free_hwrm_short_cmd_req(bp);
13075	13162	bnxt_free_hwrm_resources(bp);
	13163	+ bnxt_ethtool_free(bp);
13076	13164	kfree(bp->fw_health);
13077	13165	bp->fw_health = NULL;
13078	13166	bnxt_cleanup_pci(bp);

drivers/net/ethernet/broadcom/bnxt/bnxt.h+10−0

@@ -1457,6 +1457,16 @@ struct bnxt_ctx_pg_info { · 1456 unmodified lines
1457	1457
1458	1458	#define BNXT_BACKING_STORE_CFG_LEGACY_LEN 256
1459	1459
	1460	+#define BNXT_SET_CTX_PAGE_ATTR(attr) \
	1461	+do { \
	1462	+ if (BNXT_PAGE_SIZE == 0x2000) \
	1463	+ attr = FUNC_BACKING_STORE_CFG_REQ_SRQ_PG_SIZE_PG_8K; \
	1464	+ else if (BNXT_PAGE_SIZE == 0x10000) \
	1465	+ attr = FUNC_BACKING_STORE_CFG_REQ_QPC_PG_SIZE_PG_64K; \
	1466	+ else \
	1467	+ attr = FUNC_BACKING_STORE_CFG_REQ_QPC_PG_SIZE_PG_4K; \
	1468	+} while (0)
	1469	+
1460	1470	struct bnxt_ctx_mem_info {
1461	1471	u32 qp_max_entries;
1462	1472	u16 qp_min_qp1_entries;

drivers/net/ethernet/cadence/macb_main.c+3−0

@@ -2867,6 +2867,9 @@ static struct net_device_stats gem_get_stats(struct macb bp) · 2866 unmodified lines
2867	2867	struct gem_stats *hwstat = &bp->hw_stats.gem;
2868	2868	struct net_device_stats *nstat = &bp->dev->stats;
2869	2869
	2870	+ if (!netif_running(bp->dev))
	2871	+ return nstat;
	2872	+
2870	2873	gem_update_stats(bp);
2871	2874
2872	2875	nstat->rx_errors = (hwstat->rx_frame_check_sequence_errors +

drivers/net/ethernet/chelsio/cxgb4/cxgb4.h+0−2

@@ -2177,8 +2177,6 @@ int cxgb4_update_mac_filt(struct port_info *pi, unsigned int viid, · 2176 unmodified lines
2177	2177	bool persistent, u8 *smt_idx);
2178	2178	int cxgb4_get_msix_idx_from_bmap(struct adapter *adap);
2179	2179	void cxgb4_free_msix_idx_in_bmap(struct adapter *adap, u32 msix_idx);
2180		−int cxgb_open(struct net_device *dev);
2181		−int cxgb_close(struct net_device *dev);
2182	2180	void cxgb4_enable_rx(struct adapter adap, struct sge_rspq q);
2183	2181	void cxgb4_quiesce_rx(struct sge_rspq *q);
2184	2182	int cxgb4_port_mirror_alloc(struct net_device *dev);

drivers/net/ethernet/chelsio/cxgb4/cxgb4_ethtool.c+35−13

@@ -1337,13 +1337,27 @@ static int cxgb4_ethtool_flash_phy(struct net_device *netdev, · 1336 unmodified lines
1337	1337	return ret;
1338	1338	}
1339	1339
1340		− spin_lock_bh(&adap->win0_lock);
	1340	+ /* We have to RESET the chip/firmware because we need the
	1341	+ * chip in uninitialized state for loading new PHY image.
	1342	+ * Otherwise, the running firmware will only store the PHY
	1343	+ * image in local RAM which will be lost after next reset.
	1344	+ */
	1345	+ ret = t4_fw_reset(adap, adap->mbox, PIORSTMODE_F \| PIORST_F);
	1346	+ if (ret < 0) {
	1347	+ dev_err(adap->pdev_dev,
	1348	+ "Set FW to RESET for flashing PHY FW failed. ret: %d\n",
	1349	+ ret);
	1350	+ return ret;
	1351	+ }
	1352	+
1341	1353	ret = t4_load_phy_fw(adap, MEMWIN_NIC, NULL, data, size);
1342		− spin_unlock_bh(&adap->win0_lock);
1343		− if (ret)
1344		− dev_err(adap->pdev_dev, "Failed to load PHY FW\n");
	1354	+ if (ret < 0) {
	1355	+ dev_err(adap->pdev_dev, "Failed to load PHY FW. ret: %d\n",
	1356	+ ret);
	1357	+ return ret;
	1358	+ }
1345	1359
1346		− return ret;
	1360	+ return 0;
1347	1361	}
1348	1362
1349	1363	static int cxgb4_ethtool_flash_fw(struct net_device *netdev,
@@ -1610,16 +1624,14 @@ static struct filter_entry cxgb4_get_filter_entry(struct adapter adap, · 260 unmodified lines
1610	1624	u32 ftid)
1611	1625	{
1612	1626	struct tid_info *t = &adap->tids;
1613		− struct filter_entry *f;
1614	1627
1615		− if (ftid < t->nhpftids)
1616		− f = &adap->tids.hpftid_tab[ftid];
1617		− else if (ftid < t->nftids)
1618		− f = &adap->tids.ftid_tab[ftid - t->nhpftids];
1619		− else
1620		− f = lookup_tid(&adap->tids, ftid);
	1628	+ if (ftid >= t->hpftid_base && ftid < t->hpftid_base + t->nhpftids)
	1629	+ return &t->hpftid_tab[ftid - t->hpftid_base];
1621	1630
1622		− return f;
	1631	+ if (ftid >= t->ftid_base && ftid < t->ftid_base + t->nftids)
	1632	+ return &t->ftid_tab[ftid - t->ftid_base];
	1633	+
	1634	+ return lookup_tid(t, ftid);
1623	1635	}
1624	1636
1625	1637	static void cxgb4_fill_filter_rule(struct ethtool_rx_flow_spec *fs,
@@ -1826,6 +1838,11 @@ static int cxgb4_ntuple_del_filter(struct net_device *dev, · 200 unmodified lines
1826	1838	filter_id = filter_info->loc_array[cmd->fs.location];
1827	1839	f = cxgb4_get_filter_entry(adapter, filter_id);
1828	1840
	1841	+ if (f->fs.prio)
	1842	+ filter_id -= adapter->tids.hpftid_base;
	1843	+ else if (!f->fs.hash)
	1844	+ filter_id -= (adapter->tids.ftid_base - adapter->tids.nhpftids);
	1845	+
1829	1846	ret = cxgb4_flow_rule_destroy(dev, f->fs.tc_prio, &f->fs, filter_id);
1830	1847	if (ret)
1831	1848	goto err;
@@ -1885,6 +1902,11 @@ static int cxgb4_ntuple_set_filter(struct net_device *netdev, · 53 unmodified lines
1885	1902
1886	1903	filter_info = &adapter->ethtool_filters->port[pi->port_id];
1887	1904
	1905	+ if (fs.prio)
	1906	+ tid += adapter->tids.hpftid_base;
	1907	+ else if (!fs.hash)
	1908	+ tid += (adapter->tids.ftid_base - adapter->tids.nhpftids);
	1909	+
1888	1910	filter_info->loc_array[cmd->fs.location] = tid;
1889	1911	set_bit(cmd->fs.location, filter_info->bmap);
1890	1912	filter_info->in_use++;

drivers/net/ethernet/chelsio/cxgb4/cxgb4_filter.c+2−2

@@ -198,7 +198,7 @@ static void set_nat_params(struct adapter adap, struct filter_entry f, · 197 unmodified lines
198	198	WORD_MASK, f->fs.nat_lip[3] \|
199	199	f->fs.nat_lip[2] << 8 \|
200	200	f->fs.nat_lip[1] << 16 \|
201		− (u64)f->fs.nat_lip[0] << 25, 1);
	201	+ (u64)f->fs.nat_lip[0] << 24, 1);
202	202	}
203	203	}
204	204
@@ -1042,7 +1042,7 @@ void clear_all_filters(struct adapter *adapter) · 837 unmodified lines
1042	1042	cxgb4_del_filter(dev, f->tid, &f->fs);
1043	1043	}
1044	1044
1045		− sb = t4_read_reg(adapter, LE_DB_SRVR_START_INDEX_A);
	1045	+ sb = adapter->tids.stid_base;
1046	1046	for (i = 0; i < sb; i++) {
1047	1047	f = (struct filter_entry *)adapter->tids.tid_tab[i];
1048	1048

drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c+3−5

@@ -2834,7 +2834,7 @@ static void cxgb_down(struct adapter *adapter) · 2833 unmodified lines
2834	2834	/*
2835	2835	* net_device operations
2836	2836	*/
2837		−int cxgb_open(struct net_device *dev)
	2837	+static int cxgb_open(struct net_device *dev)
2838	2838	{
2839	2839	struct port_info *pi = netdev_priv(dev);
2840	2840	struct adapter *adapter = pi->adapter;
@@ -2882,7 +2882,7 @@ out_unlock: · 41 unmodified lines
2882	2882	return err;
2883	2883	}
2884	2884
2885		−int cxgb_close(struct net_device *dev)
	2885	+static int cxgb_close(struct net_device *dev)
2886	2886	{
2887	2887	struct port_info *pi = netdev_priv(dev);
2888	2888	struct adapter *adapter = pi->adapter;
@@ -4424,10 +4424,8 @@ static int adap_init0_phy(struct adapter *adap) · 1535 unmodified lines
4424	4424
4425	4425	/* Load PHY Firmware onto adapter.
4426	4426	*/
4427		− spin_lock_bh(&adap->win0_lock);
4428	4427	ret = t4_load_phy_fw(adap, MEMWIN_NIC, phy_info->phy_fw_version,
4429	4428	(u8 *)phyf->data, phyf->size);
4430		− spin_unlock_bh(&adap->win0_lock);
4431	4429	if (ret < 0)
4432	4430	dev_err(adap->pdev_dev, "PHY Firmware transfer error %d\n",
4433	4431	-ret);
@@ -6480,9 +6478,9 @@ static void cxgb4_ktls_dev_del(struct net_device *netdev, · 2046 unmodified lines
6480	6478
6481	6479	adap->uld[CXGB4_ULD_KTLS].tlsdev_ops->tls_dev_del(netdev, tls_ctx,
6482	6480	direction);
6483		− cxgb4_set_ktls_feature(adap, FW_PARAMS_PARAM_DEV_KTLS_HW_DISABLE);
6484	6481
6485	6482	out_unlock:
	6483	+ cxgb4_set_ktls_feature(adap, FW_PARAMS_PARAM_DEV_KTLS_HW_DISABLE);
6486	6484	mutex_unlock(&uld_mutex);
6487	6485	}
6488	6486

drivers/net/ethernet/chelsio/cxgb4/cxgb4_tc_flower.c+5−9

@@ -997,20 +997,16 @@ int cxgb4_tc_flower_destroy(struct net_device *dev, · 996 unmodified lines
997	997	if (!ch_flower)
998	998	return -ENOENT;
999	999
	1000	+ rhashtable_remove_fast(&adap->flower_tbl, &ch_flower->node,
	1001	+ adap->flower_ht_params);
	1002	+
1000	1003	ret = cxgb4_flow_rule_destroy(dev, ch_flower->fs.tc_prio,
1001	1004	&ch_flower->fs, ch_flower->filter_id);
1002	1005	if (ret)
1003		− goto err;
	1006	+ netdev_err(dev, "Flow rule destroy failed for tid: %u, ret: %d",
	1007	+ ch_flower->filter_id, ret);
1004	1008
1005		− ret = rhashtable_remove_fast(&adap->flower_tbl, &ch_flower->node,
1006		− adap->flower_ht_params);
1007		− if (ret) {
1008		− netdev_err(dev, "Flow remove from rhashtable failed");
1009		− goto err;
1010		− }
1011	1009	kfree_rcu(ch_flower, rcu);
1012		−
1013		−err:
1014	1010	return ret;
1015	1011	}
1016	1012

drivers/net/ethernet/chelsio/cxgb4/cxgb4_tc_mqprio.c+6−3

@@ -589,7 +589,8 @@ int cxgb4_setup_tc_mqprio(struct net_device *dev, · 588 unmodified lines
589	589	* down before configuring tc params.
590	590	*/
591	591	if (netif_running(dev)) {
592		− cxgb_close(dev);
	592	+ netif_tx_stop_all_queues(dev);
	593	+ netif_carrier_off(dev);
593	594	needs_bring_up = true;
594	595	}
595	596
@@ -615,8 +616,10 @@ int cxgb4_setup_tc_mqprio(struct net_device *dev, · 19 unmodified lines
615	616	}
616	617
617	618	out:
618		− if (needs_bring_up)
619		− cxgb_open(dev);
	619	+ if (needs_bring_up) {
	620	+ netif_tx_start_all_queues(dev);
	621	+ netif_carrier_on(dev);
	622	+ }
620	623
621	624	mutex_unlock(&adap->tc_mqprio->mqprio_mutex);
622	625	return ret;

drivers/net/ethernet/chelsio/cxgb4/sge.c+6−0

@@ -2556,6 +2556,12 @@ int cxgb4_ethofld_send_flowc(struct net_device *dev, u32 eotid, u32 tc) · 2555 unmodified lines
2556	2556	if (!eosw_txq)
2557	2557	return -ENOMEM;
2558	2558
	2559	+ if (!(adap->flags & CXGB4_FW_OK)) {
	2560	+ /* Don't stall caller when access to FW is lost */
	2561	+ complete(&eosw_txq->completion);
	2562	+ return -EIO;
	2563	+ }
	2564	+
2559	2565	skb = alloc_skb(len, GFP_KERNEL);
2560	2566	if (!skb)
2561	2567	return -ENOMEM;

drivers/net/ethernet/chelsio/cxgb4/t4_hw.c+29−17

@@ -3060,16 +3060,19 @@ int t4_read_flash(struct adapter *adapter, unsigned int addr, · 3059 unmodified lines
3060	3060	* @addr: the start address to write
3061	3061	* @n: length of data to write in bytes
3062	3062	* @data: the data to write
	3063	+ * @byte_oriented: whether to store data as bytes or as words
3063	3064	*
3064	3065	* Writes up to a page of data (256 bytes) to the serial flash starting
3065	3066	* at the given address. All the data must be written to the same page.
	3067	+ * If @byte_oriented is set the write data is stored as byte stream
	3068	+ * (i.e. matches what on disk), otherwise in big-endian.
3066	3069	*/
3067	3070	static int t4_write_flash(struct adapter *adapter, unsigned int addr,
3068		− unsigned int n, const u8 *data)
	3071	+ unsigned int n, const u8 *data, bool byte_oriented)
3069	3072	{
3070		− int ret;
3071		− u32 buf[64];
3072	3073	unsigned int i, c, left, val, offset = addr & 0xff;
	3074	+ u32 buf[64];
	3075	+ int ret;
3073	3076
3074	3077	if (addr >= adapter->params.sf_size \|\| offset + n > SF_PAGE_SIZE)
3075	3078	return -EINVAL;
@@ -3080,10 +3083,14 @@ static int t4_write_flash(struct adapter *adapter, unsigned int addr, · 4 unmodified lines
3080	3083	(ret = sf1_write(adapter, 4, 1, 1, val)) != 0)
3081	3084	goto unlock;
3082	3085
3083		− for (left = n; left; left -= c) {
	3086	+ for (left = n; left; left -= c, data += c) {
3084	3087	c = min(left, 4U);
3085		− for (val = 0, i = 0; i < c; ++i)
3086		− val = (val << 8) + *data++;
	3088	+ for (val = 0, i = 0; i < c; ++i) {
	3089	+ if (byte_oriented)
	3090	+ val = (val << 8) + data[i];
	3091	+ else
	3092	+ val = (val << 8) + data[c - i - 1];
	3093	+ }
3087	3094
3088	3095	ret = sf1_write(adapter, c, c != left, 1, val);
3089	3096	if (ret)
@@ -3096,7 +3103,8 @@ static int t4_write_flash(struct adapter *adapter, unsigned int addr, · 6 unmodified lines
3096	3103	t4_write_reg(adapter, SF_OP_A, 0); /* unlock SF */
3097	3104
3098	3105	/* Read the page to verify the write succeeded */
3099		− ret = t4_read_flash(adapter, addr & ~0xff, ARRAY_SIZE(buf), buf, 1);
	3106	+ ret = t4_read_flash(adapter, addr & ~0xff, ARRAY_SIZE(buf), buf,
	3107	+ byte_oriented);
3100	3108	if (ret)
3101	3109	return ret;
3102	3110
@@ -3692,7 +3700,7 @@ int t4_load_fw(struct adapter adap, const u8 fw_data, unsigned int size) · 589 unmodified lines
3692	3700	*/
3693	3701	memcpy(first_page, fw_data, SF_PAGE_SIZE);
3694	3702	((struct fw_hdr *)first_page)->fw_ver = cpu_to_be32(0xffffffff);
3695		− ret = t4_write_flash(adap, fw_start, SF_PAGE_SIZE, first_page);
	3703	+ ret = t4_write_flash(adap, fw_start, SF_PAGE_SIZE, first_page, true);
3696	3704	if (ret)
3697	3705	goto out;
3698	3706
@@ -3700,14 +3708,14 @@ int t4_load_fw(struct adapter adap, const u8 fw_data, unsigned int size) · 1 unmodified line
3700	3708	for (size -= SF_PAGE_SIZE; size; size -= SF_PAGE_SIZE) {
3701	3709	addr += SF_PAGE_SIZE;
3702	3710	fw_data += SF_PAGE_SIZE;
3703		− ret = t4_write_flash(adap, addr, SF_PAGE_SIZE, fw_data);
	3711	+ ret = t4_write_flash(adap, addr, SF_PAGE_SIZE, fw_data, true);
3704	3712	if (ret)
3705	3713	goto out;
3706	3714	}
3707	3715
3708		− ret = t4_write_flash(adap,
3709		− fw_start + offsetof(struct fw_hdr, fw_ver),
3710		− sizeof(hdr->fw_ver), (const u8 *)&hdr->fw_ver);
	3716	+ ret = t4_write_flash(adap, fw_start + offsetof(struct fw_hdr, fw_ver),
	3717	+ sizeof(hdr->fw_ver), (const u8 *)&hdr->fw_ver,
	3718	+ true);
3711	3719	out:
3712	3720	if (ret)
3713	3721	dev_err(adap->pdev_dev, "firmware download failed, error %d\n",
@@ -3812,9 +3820,11 @@ int t4_load_phy_fw(struct adapter *adap, int win, · 98 unmodified lines
3812	3820	/* Copy the supplied PHY Firmware image to the adapter memory location
3813	3821	* allocated by the adapter firmware.
3814	3822	*/
	3823	+ spin_lock_bh(&adap->win0_lock);
3815	3824	ret = t4_memory_rw(adap, win, mtype, maddr,
3816	3825	phy_fw_size, (__be32 *)phy_fw_data,
3817	3826	T4_MEMORY_WRITE);
	3827	+ spin_unlock_bh(&adap->win0_lock);
3818	3828	if (ret)
3819	3829	return ret;
3820	3830
@@ -10208,7 +10218,7 @@ int t4_load_cfg(struct adapter adap, const u8 cfg_data, unsigned int size) · 6387 unmodified lines
10208	10218	n = size - i;
10209	10219	else
10210	10220	n = SF_PAGE_SIZE;
10211		− ret = t4_write_flash(adap, addr, n, cfg_data);
	10221	+ ret = t4_write_flash(adap, addr, n, cfg_data, true);
10212	10222	if (ret)
10213	10223	goto out;
10214	10224
@@ -10677,13 +10687,14 @@ int t4_load_boot(struct adapter adap, u8 boot_data, · 462 unmodified lines
10677	10687	for (size -= SF_PAGE_SIZE; size; size -= SF_PAGE_SIZE) {
10678	10688	addr += SF_PAGE_SIZE;
10679	10689	boot_data += SF_PAGE_SIZE;
10680		− ret = t4_write_flash(adap, addr, SF_PAGE_SIZE, boot_data);
	10690	+ ret = t4_write_flash(adap, addr, SF_PAGE_SIZE, boot_data,
	10691	+ false);
10681	10692	if (ret)
10682	10693	goto out;
10683	10694	}
10684	10695
10685	10696	ret = t4_write_flash(adap, boot_sector, SF_PAGE_SIZE,
10686		− (const u8 *)header);
	10697	+ (const u8 *)header, false);
10687	10698
10688	10699	out:
10689	10700	if (ret)
@@ -10758,7 +10769,7 @@ int t4_load_bootcfg(struct adapter adap, const u8 cfg_data, unsigned int size) · 68 unmodified lines
10758	10769	for (i = 0; i < size; i += SF_PAGE_SIZE) {
10759	10770	n = min_t(u32, size - i, SF_PAGE_SIZE);
10760	10771
10761		− ret = t4_write_flash(adap, addr, n, cfg_data);
	10772	+ ret = t4_write_flash(adap, addr, n, cfg_data, false);
10762	10773	if (ret)
10763	10774	goto out;
10764	10775
@@ -10770,7 +10781,8 @@ int t4_load_bootcfg(struct adapter adap, const u8 cfg_data, unsigned int size) · 5 unmodified lines
10770	10781	for (i = 0; i < npad; i++) {
10771	10782	u8 data = 0;
10772	10783
10773		− ret = t4_write_flash(adap, cfg_addr + size + i, 1, &data);
	10784	+ ret = t4_write_flash(adap, cfg_addr + size + i, 1, &data,
	10785	+ false);
10774	10786	if (ret)
10775	10787	goto out;
10776	10788	}

drivers/net/ethernet/chelsio/inline_crypto/ch_ktls/chcr_ktls.c+79−1

@@ -59,6 +59,7 @@ static int chcr_get_nfrags_to_send(struct sk_buff *skb, u32 start, u32 len) · 58 unmodified lines
59	59	}
60	60
61	61	static int chcr_init_tcb_fields(struct chcr_ktls_info *tx_info);
	62	+static void clear_conn_resources(struct chcr_ktls_info *tx_info);
62	63	/*
63	64	* chcr_ktls_save_keys: calculate and save crypto keys.
64	65	* @tx_info - driver specific tls info.
@@ -364,10 +365,14 @@ static void chcr_ktls_dev_del(struct net_device *netdev, · 299 unmodified lines
364	365	chcr_get_ktls_tx_context(tls_ctx);
365	366	struct chcr_ktls_info *tx_info = tx_ctx->chcr_info;
366	367	struct ch_ktls_port_stats_debug *port_stats;
	368	+ struct chcr_ktls_uld_ctx *u_ctx;
367	369
368	370	if (!tx_info)
369	371	return;
370	372
	373	+ u_ctx = tx_info->adap->uld[CXGB4_ULD_KTLS].handle;
	374	+ if (u_ctx && u_ctx->detach)
	375	+ return;
371	376	/* clear l2t entry */
372	377	if (tx_info->l2te)
373	378	cxgb4_l2t_release(tx_info->l2te);
@@ -384,6 +389,8 @@ static void chcr_ktls_dev_del(struct net_device *netdev, · 10 unmodified lines
384	389	if (tx_info->tid != -1) {
385	390	cxgb4_remove_tid(&tx_info->adap->tids, tx_info->tx_chan,
386	391	tx_info->tid, tx_info->ip_family);
	392	+
	393	+ xa_erase(&u_ctx->tid_list, tx_info->tid);
387	394	}
388	395
389	396	port_stats = &tx_info->adap->ch_ktls_stats.ktls_port[tx_info->port_id];
@@ -411,6 +418,7 @@ static int chcr_ktls_dev_add(struct net_device netdev, struct sock sk, · 21 unmodified lines
411	418	struct tls_context *tls_ctx = tls_get_ctx(sk);
412	419	struct ch_ktls_port_stats_debug *port_stats;
413	420	struct chcr_ktls_ofld_ctx_tx *tx_ctx;
	421	+ struct chcr_ktls_uld_ctx *u_ctx;
414	422	struct chcr_ktls_info *tx_info;
415	423	struct dst_entry *dst;
416	424	struct adapter *adap;
@@ -425,6 +433,7 @@ static int chcr_ktls_dev_add(struct net_device netdev, struct sock sk, · 8 unmodified lines
425	433	adap = pi->adapter;
426	434	port_stats = &adap->ch_ktls_stats.ktls_port[pi->port_id];
427	435	atomic64_inc(&port_stats->ktls_tx_connection_open);
	436	+ u_ctx = adap->uld[CXGB4_ULD_KTLS].handle;
428	437
429	438	if (direction == TLS_OFFLOAD_CTX_DIR_RX) {
430	439	pr_err("not expecting for RX direction\n");
@@ -434,6 +443,9 @@ static int chcr_ktls_dev_add(struct net_device netdev, struct sock sk, · 3 unmodified lines
434	443	if (tx_ctx->chcr_info)
435	444	goto out;
436	445
	446	+ if (u_ctx && u_ctx->detach)
	447	+ goto out;
	448	+
437	449	tx_info = kvzalloc(sizeof(*tx_info), GFP_KERNEL);
438	450	if (!tx_info)
439	451	goto out;
@@ -569,6 +581,8 @@ free_tid: · 129 unmodified lines
569	581	cxgb4_remove_tid(&tx_info->adap->tids, tx_info->tx_chan,
570	582	tx_info->tid, tx_info->ip_family);
571	583
	584	+ xa_erase(&u_ctx->tid_list, tx_info->tid);
	585	+
572	586	put_module:
573	587	/* release module refcount */
574	588	module_put(THIS_MODULE);
@@ -633,8 +647,12 @@ static int chcr_ktls_cpl_act_open_rpl(struct adapter *adap, · 58 unmodified lines
633	647	{
634	648	const struct cpl_act_open_rpl p = (void )input;
635	649	struct chcr_ktls_info *tx_info = NULL;
	650	+ struct chcr_ktls_ofld_ctx_tx *tx_ctx;
	651	+ struct chcr_ktls_uld_ctx *u_ctx;
636	652	unsigned int atid, tid, status;
	653	+ struct tls_context *tls_ctx;
637	654	struct tid_info *t;
	655	+ int ret = 0;
638	656
639	657	tid = GET_TID(p);
640	658	status = AOPEN_STATUS_G(ntohl(p->atid_status));
@@ -666,14 +684,29 @@ static int chcr_ktls_cpl_act_open_rpl(struct adapter *adap, · 25 unmodified lines
666	684	if (!status) {
667	685	tx_info->tid = tid;
668	686	cxgb4_insert_tid(t, tx_info, tx_info->tid, tx_info->ip_family);
	687	+ /* Adding tid */
	688	+ tls_ctx = tls_get_ctx(tx_info->sk);
	689	+ tx_ctx = chcr_get_ktls_tx_context(tls_ctx);
	690	+ u_ctx = adap->uld[CXGB4_ULD_KTLS].handle;
	691	+ if (u_ctx) {
	692	+ ret = xa_insert_bh(&u_ctx->tid_list, tid, tx_ctx,
	693	+ GFP_NOWAIT);
	694	+ if (ret < 0) {
	695	+ pr_err("%s: Failed to allocate tid XA entry = %d\n",
	696	+ __func__, tx_info->tid);
	697	+ tx_info->open_state = CH_KTLS_OPEN_FAILURE;
	698	+ goto out;
	699	+ }
	700	+ }
669	701	tx_info->open_state = CH_KTLS_OPEN_SUCCESS;
670	702	} else {
671	703	tx_info->open_state = CH_KTLS_OPEN_FAILURE;
672	704	}
	705	+out:
673	706	spin_unlock(&tx_info->lock);
674	707
675	708	complete(&tx_info->completion);
676		− return 0;
	709	+ return ret;
677	710	}
678	711
679	712	/*
@@ -2090,6 +2123,8 @@ static void chcr_ktls_uld_add(const struct cxgb4_lld_info lldi) · 1410 unmodified lines
2090	2123	goto out;
2091	2124	}
2092	2125	u_ctx->lldi = *lldi;
	2126	+ u_ctx->detach = false;
	2127	+ xa_init_flags(&u_ctx->tid_list, XA_FLAGS_LOCK_BH);
2093	2128	out:
2094	2129	return u_ctx;
2095	2130	}
@@ -2123,6 +2158,45 @@ static int chcr_ktls_uld_rx_handler(void handle, const __be64 rsp, · 27 unmodified lines
2123	2158	return 0;
2124	2159	}
2125	2160
	2161	+static void clear_conn_resources(struct chcr_ktls_info *tx_info)
	2162	+{
	2163	+ /* clear l2t entry */
	2164	+ if (tx_info->l2te)
	2165	+ cxgb4_l2t_release(tx_info->l2te);
	2166	+
	2167	+#if IS_ENABLED(CONFIG_IPV6)
	2168	+ /* clear clip entry */
	2169	+ if (tx_info->ip_family == AF_INET6)
	2170	+ cxgb4_clip_release(tx_info->netdev, (const u32 *)
	2171	+ &tx_info->sk->sk_v6_rcv_saddr,
	2172	+ 1);
	2173	+#endif
	2174	+
	2175	+ /* clear tid */
	2176	+ if (tx_info->tid != -1)
	2177	+ cxgb4_remove_tid(&tx_info->adap->tids, tx_info->tx_chan,
	2178	+ tx_info->tid, tx_info->ip_family);
	2179	+}
	2180	+
	2181	+static void ch_ktls_reset_all_conn(struct chcr_ktls_uld_ctx *u_ctx)
	2182	+{
	2183	+ struct ch_ktls_port_stats_debug *port_stats;
	2184	+ struct chcr_ktls_ofld_ctx_tx *tx_ctx;
	2185	+ struct chcr_ktls_info *tx_info;
	2186	+ unsigned long index;
	2187	+
	2188	+ xa_for_each(&u_ctx->tid_list, index, tx_ctx) {
	2189	+ tx_info = tx_ctx->chcr_info;
	2190	+ clear_conn_resources(tx_info);
	2191	+ port_stats = &tx_info->adap->ch_ktls_stats.ktls_port[tx_info->port_id];
	2192	+ atomic64_inc(&port_stats->ktls_tx_connection_close);
	2193	+ kvfree(tx_info);
	2194	+ tx_ctx->chcr_info = NULL;
	2195	+ /* release module refcount */
	2196	+ module_put(THIS_MODULE);
	2197	+ }
	2198	+}
	2199	+
2126	2200	static int chcr_ktls_uld_state_change(void *handle, enum cxgb4_state new_state)
2127	2201	{
2128	2202	struct chcr_ktls_uld_ctx *u_ctx = handle;
@@ -2139,7 +2213,10 @@ static int chcr_ktls_uld_state_change(void *handle, enum cxgb4_state new_state) · 10 unmodified lines
2139	2213	case CXGB4_STATE_DETACH:
2140	2214	pr_info("%s: Down\n", pci_name(u_ctx->lldi.pdev));
2141	2215	mutex_lock(&dev_mutex);
	2216	+ u_ctx->detach = true;
2142	2217	list_del(&u_ctx->entry);
	2218	+ ch_ktls_reset_all_conn(u_ctx);
	2219	+ xa_destroy(&u_ctx->tid_list);
2143	2220	mutex_unlock(&dev_mutex);
2144	2221	break;
2145	2222	default:
@@ -2178,6 +2255,7 @@ static void __exit chcr_ktls_exit(void) · 32 unmodified lines
2178	2255	adap = pci_get_drvdata(u_ctx->lldi.pdev);
2179	2256	memset(&adap->ch_ktls_stats, 0, sizeof(adap->ch_ktls_stats));
2180	2257	list_del(&u_ctx->entry);
	2258	+ xa_destroy(&u_ctx->tid_list);
2181	2259	kfree(u_ctx);
2182	2260	}
2183	2261	mutex_unlock(&dev_mutex);

drivers/net/ethernet/chelsio/inline_crypto/ch_ktls/chcr_ktls.h+2−0

@@ -75,6 +75,8 @@ struct chcr_ktls_ofld_ctx_tx { · 74 unmodified lines
75	75	struct chcr_ktls_uld_ctx {
76	76	struct list_head entry;
77	77	struct cxgb4_lld_info lldi;
	78	+ struct xarray tid_list;
	79	+ bool detach;
78	80	};
79	81
80	82	static inline struct chcr_ktls_ofld_ctx_tx *

drivers/net/ethernet/chelsio/inline_crypto/chtls/chtls_io.c+4−2

@@ -1564,8 +1564,10 @@ found_ok_skb: · 1563 unmodified lines
1564	1564	cerr = put_cmsg(msg, SOL_TLS, TLS_GET_RECORD_TYPE,
1565	1565	sizeof(thdr->type), &thdr->type);
1566	1566
1567		− if (cerr && thdr->type != TLS_RECORD_TYPE_DATA)
1568		− return -EIO;
	1567	+ if (cerr && thdr->type != TLS_RECORD_TYPE_DATA) {
	1568	+ copied = -EIO;
	1569	+ break;
	1570	+ }
1569	1571	/* don't send tls header, skip copy */
1570	1572	goto skip_copy;
1571	1573	}

drivers/net/ethernet/ec_bhf.c+3−1

@@ -576,10 +576,12 @@ static void ec_bhf_remove(struct pci_dev *dev) · 575 unmodified lines
576	576	struct ec_bhf_priv *priv = netdev_priv(net_dev);
577	577
578	578	unregister_netdev(net_dev);
579		− free_netdev(net_dev);
580	579
581	580	pci_iounmap(dev, priv->dma_io);
582	581	pci_iounmap(dev, priv->io);
	582	+
	583	+ free_netdev(net_dev);
	584	+
583	585	pci_release_regions(dev);
584	586	pci_clear_master(dev);
585	587	pci_disable_device(dev);

drivers/net/ethernet/emulex/benet/be_main.c+1−0

@@ -5897,6 +5897,7 @@ drv_cleanup: · 5896 unmodified lines
5897	5897	unmap_bars:
5898	5898	be_unmap_pci_bars(adapter);
5899	5899	free_netdev:
	5900	+ pci_disable_pcie_error_reporting(pdev);
5900	5901	free_netdev(netdev);
5901	5902	rel_reg:
5902	5903	pci_release_regions(pdev);

drivers/net/ethernet/freescale/fec_main.c+19−5

@@ -1662,7 +1662,7 @@ static int fec_enet_rx_napi(struct napi_struct *napi, int budget) · 1661 unmodified lines
1662	1662	}
1663	1663
1664	1664	/* ------------------------------------------------------------------------- */
1665		−static void fec_get_mac(struct net_device *ndev)
	1665	+static int fec_get_mac(struct net_device *ndev)
1666	1666	{
1667	1667	struct fec_enet_private *fep = netdev_priv(ndev);
1668	1668	unsigned char *iap, tmpaddr[ETH_ALEN];
@@ -1685,6 +1685,8 @@ static void fec_get_mac(struct net_device *ndev) · 16 unmodified lines
1685	1685	ret = of_get_mac_address(np, tmpaddr);
1686	1686	if (!ret)
1687	1687	iap = tmpaddr;
	1688	+ else if (ret == -EPROBE_DEFER)
	1689	+ return ret;
1688	1690	}
1689	1691	}
1690	1692
@@ -1723,7 +1725,7 @@ static void fec_get_mac(struct net_device *ndev) · 32 unmodified lines
1723	1725	eth_hw_addr_random(ndev);
1724	1726	dev_info(&fep->pdev->dev, "Using random MAC address: %pM\n",
1725	1727	ndev->dev_addr);
1726		− return;
	1728	+ return 0;
1727	1729	}
1728	1730
1729	1731	memcpy(ndev->dev_addr, iap, ETH_ALEN);
@@ -1731,6 +1733,8 @@ static void fec_get_mac(struct net_device *ndev) · 1 unmodified line
1731	1733	/* Adjust MAC if using macaddr */
1732	1734	if (iap == macaddr)
1733	1735	ndev->dev_addr[ETH_ALEN-1] = macaddr[ETH_ALEN-1] + fep->dev_id;
	1736	+
	1737	+ return 0;
1734	1738	}
1735	1739
1736	1740	/* ------------------------------------------------------------------------- */
@@ -3290,7 +3294,9 @@ static int fec_enet_init(struct net_device *ndev) · 1553 unmodified lines
3290	3294	return ret;
3291	3295	}
3292	3296
3293		− fec_enet_alloc_queue(ndev);
	3297	+ ret = fec_enet_alloc_queue(ndev);
	3298	+ if (ret)
	3299	+ return ret;
3294	3300
3295	3301	bd_size = (fep->total_tx_ring_size + fep->total_rx_ring_size) * dsize;
3296	3302
@@ -3298,11 +3304,15 @@ static int fec_enet_init(struct net_device *ndev) · 1 unmodified line
3298	3304	cbd_base = dmam_alloc_coherent(&fep->pdev->dev, bd_size, &bd_dma,
3299	3305	GFP_KERNEL);
3300	3306	if (!cbd_base) {
3301		− return -ENOMEM;
	3307	+ ret = -ENOMEM;
	3308	+ goto free_queue_mem;
3302	3309	}
3303	3310
3304	3311	/* Get the Ethernet address */
3305		− fec_get_mac(ndev);
	3312	+ ret = fec_get_mac(ndev);
	3313	+ if (ret)
	3314	+ goto free_queue_mem;
	3315	+
3306	3316	/* make sure MAC we just acquired is programmed into the hw */
3307	3317	fec_set_mac_address(ndev, NULL);
3308	3318
@@ -3376,6 +3386,10 @@ static int fec_enet_init(struct net_device *ndev) · 67 unmodified lines
3376	3386	fec_enet_update_ethtool_stats(ndev);
3377	3387
3378	3388	return 0;
	3389	+
	3390	+free_queue_mem:
	3391	+ fec_enet_free_queue(ndev);
	3392	+ return ret;
3379	3393	}
3380	3394
3381	3395	#ifdef CONFIG_OF

drivers/net/ethernet/freescale/fec_ptp.c+5−3

@@ -215,15 +215,13 @@ static u64 fec_ptp_read(const struct cyclecounter *cc) · 214 unmodified lines
215	215	{
216	216	struct fec_enet_private *fep =
217	217	container_of(cc, struct fec_enet_private, cc);
218		− const struct platform_device_id *id_entry =
219		− platform_get_device_id(fep->pdev);
220	218	u32 tempval;
221	219
222	220	tempval = readl(fep->hwp + FEC_ATIME_CTRL);
223	221	tempval \|= FEC_T_CTRL_CAPTURE;
224	222	writel(tempval, fep->hwp + FEC_ATIME_CTRL);
225	223
226		− if (id_entry->driver_data & FEC_QUIRK_BUG_CAPTURE)
	224	+ if (fep->quirks & FEC_QUIRK_BUG_CAPTURE)
227	225	udelay(1);
228	226
229	227	return readl(fep->hwp + FEC_ATIME);
@@ -604,6 +602,10 @@ void fec_ptp_init(struct platform_device *pdev, int irq_idx) · 374 unmodified lines
604	602	fep->ptp_caps.enable = fec_ptp_enable;
605	603
606	604	fep->cycle_speed = clk_get_rate(fep->clk_ptp);
	605	+ if (!fep->cycle_speed) {
	606	+ fep->cycle_speed = NSEC_PER_SEC;
	607	+ dev_err(&fep->pdev->dev, "clk_ptp clock rate is zero\n");
	608	+ }
607	609	fep->ptp_inc = NSEC_PER_SEC / fep->cycle_speed;
608	610
609	611	spin_lock_init(&fep->tmreg_lock);

drivers/net/ethernet/google/gve/gve_main.c+12−9

@@ -180,7 +180,7 @@ static int gve_napi_poll(struct napi_struct *napi, int budget) · 179 unmodified lines
180	180	/* Double check we have no extra work.
181	181	* Ensure unmask synchronizes with checking for work.
182	182	*/
183		− dma_rmb();
	183	+ mb();
184	184	if (block->tx)
185	185	reschedule \|= gve_tx_poll(block, -1);
186	186	if (block->rx)
@@ -220,6 +220,7 @@ static int gve_alloc_notify_blocks(struct gve_priv *priv) · 33 unmodified lines
220	220	int vecs_left = new_num_ntfy_blks % 2;
221	221
222	222	priv->num_ntfy_blks = new_num_ntfy_blks;
	223	+ priv->mgmt_msix_idx = priv->num_ntfy_blks;
223	224	priv->tx_cfg.max_queues = min_t(int, priv->tx_cfg.max_queues,
224	225	vecs_per_type);
225	226	priv->rx_cfg.max_queues = min_t(int, priv->rx_cfg.max_queues,
@@ -300,20 +301,22 @@ static void gve_free_notify_blocks(struct gve_priv *priv) · 74 unmodified lines
300	301	{
301	302	int i;
302	303
303		− /* Free the irqs */
304		− for (i = 0; i < priv->num_ntfy_blks; i++) {
305		− struct gve_notify_block *block = &priv->ntfy_blocks[i];
306		− int msix_idx = i;
	304	+ if (priv->msix_vectors) {
	305	+ /* Free the irqs */
	306	+ for (i = 0; i < priv->num_ntfy_blks; i++) {
	307	+ struct gve_notify_block *block = &priv->ntfy_blocks[i];
	308	+ int msix_idx = i;
307	309
308		− irq_set_affinity_hint(priv->msix_vectors[msix_idx].vector,
309		− NULL);
310		− free_irq(priv->msix_vectors[msix_idx].vector, block);
	310	+ irq_set_affinity_hint(priv->msix_vectors[msix_idx].vector,
	311	+ NULL);
	312	+ free_irq(priv->msix_vectors[msix_idx].vector, block);
	313	+ }
	314	+ free_irq(priv->msix_vectors[priv->mgmt_msix_idx].vector, priv);
311	315	}
312	316	dma_free_coherent(&priv->pdev->dev,
313	317	priv->num_ntfy_blks * sizeof(*priv->ntfy_blocks),
314	318	priv->ntfy_blocks, priv->ntfy_block_bus);
315	319	priv->ntfy_blocks = NULL;
316		− free_irq(priv->msix_vectors[priv->mgmt_msix_idx].vector, priv);
317	320	pci_disable_msix(priv->pdev);
318	321	kvfree(priv->msix_vectors);
319	322	priv->msix_vectors = NULL;

drivers/net/ethernet/google/gve/gve_tx.c+7−3

@@ -212,10 +212,11 @@ static int gve_tx_alloc_ring(struct gve_priv *priv, int idx) · 211 unmodified lines
212	212	tx->dev = &priv->pdev->dev;
213	213	if (!tx->raw_addressing) {
214	214	tx->tx_fifo.qpl = gve_assign_tx_qpl(priv);
215		−
	215	+ if (!tx->tx_fifo.qpl)
	216	+ goto abort_with_desc;
216	217	/* map Tx FIFO */
217	218	if (gve_tx_fifo_init(priv, &tx->tx_fifo))
218		− goto abort_with_desc;
	219	+ goto abort_with_qpl;
219	220	}
220	221
221	222	tx->q_resources =
@@ -236,6 +237,9 @@ static int gve_tx_alloc_ring(struct gve_priv *priv, int idx) · 14 unmodified lines
236	237	abort_with_fifo:
237	238	if (!tx->raw_addressing)
238	239	gve_tx_fifo_release(priv, &tx->tx_fifo);
	240	+abort_with_qpl:
	241	+ if (!tx->raw_addressing)
	242	+ gve_unassign_qpl(priv, tx->tx_fifo.qpl->id);
239	243	abort_with_desc:
240	244	dma_free_coherent(hdev, bytes, tx->desc, tx->bus);
241	245	tx->desc = NULL;
@@ -589,7 +593,7 @@ netdev_tx_t gve_tx(struct sk_buff skb, struct net_device dev) · 347 unmodified lines
589	593	struct gve_tx_ring *tx;
590	594	int nsegs;
591	595
592		− WARN(skb_get_queue_mapping(skb) > priv->tx_cfg.num_queues,
	596	+ WARN(skb_get_queue_mapping(skb) >= priv->tx_cfg.num_queues,
593	597	"skb queue index out of range");
594	598	tx = &priv->tx[skb_get_queue_mapping(skb)];
595	599	if (unlikely(gve_maybe_stop_tx(tx, skb))) {

drivers/net/ethernet/hisilicon/hns/hns_ethtool.c+3−3

@@ -192,7 +192,7 @@ static int hns_nic_get_link_ksettings(struct net_device *net_dev, · 191 unmodified lines
192	192	}
193	193
194	194	/**
195		− *hns_nic_set_link_settings - implement ethtool set link ksettings
	195	+ *hns_nic_set_link_ksettings - implement ethtool set link ksettings
196	196	*@net_dev: net_device
197	197	*@cmd: ethtool_link_ksettings
198	198	*retuen 0 - success , negative --fail
@@ -827,7 +827,7 @@ hns_get_channels(struct net_device net_dev, struct ethtool_channels ch) · 628 unmodified lines
827	827	}
828	828
829	829	/**
830		− * get_ethtool_stats - get detail statistics.
	830	+ * hns_get_ethtool_stats - get detail statistics.
831	831	* @netdev: net device
832	832	* @stats: statistics info.
833	833	* @data: statistics data.
@@ -885,7 +885,7 @@ static void hns_get_ethtool_stats(struct net_device *netdev, · 51 unmodified lines
885	885	}
886	886
887	887	/**
888		− * get_strings: Return a set of strings that describe the requested objects
	888	+ * hns_get_strings: Return a set of strings that describe the requested objects
889	889	* @netdev: net device
890	890	* @stringset: string set ID.
891	891	* @data: objects data.

drivers/net/ethernet/hisilicon/hns3/hns3_enet.c+51−59

@@ -264,22 +264,17 @@ static void hns3_vector_coalesce_init(struct hns3_enet_tqp_vector *tqp_vector, · 263 unmodified lines
264	264	struct hnae3_ae_dev *ae_dev = pci_get_drvdata(priv->ae_handle->pdev);
265	265	struct hns3_enet_coalesce *tx_coal = &tqp_vector->tx_group.coal;
266	266	struct hns3_enet_coalesce *rx_coal = &tqp_vector->rx_group.coal;
	267	+ struct hns3_enet_coalesce *ptx_coal = &priv->tx_coal;
	268	+ struct hns3_enet_coalesce *prx_coal = &priv->rx_coal;
267	269
268		− /* initialize the configuration for interrupt coalescing.
269		− * 1. GL (Interrupt Gap Limiter)
270		− * 2. RL (Interrupt Rate Limiter)
271		− * 3. QL (Interrupt Quantity Limiter)
272		− *
273		− * Default: enable interrupt coalescing self-adaptive and GL
274		− */
275		− tx_coal->adapt_enable = 1;
276		− rx_coal->adapt_enable = 1;
	270	+ tx_coal->adapt_enable = ptx_coal->adapt_enable;
	271	+ rx_coal->adapt_enable = prx_coal->adapt_enable;
277	272
278		− tx_coal->int_gl = HNS3_INT_GL_50K;
279		− rx_coal->int_gl = HNS3_INT_GL_50K;
	273	+ tx_coal->int_gl = ptx_coal->int_gl;
	274	+ rx_coal->int_gl = prx_coal->int_gl;
280	275
281		− rx_coal->flow_level = HNS3_FLOW_LOW;
282		− tx_coal->flow_level = HNS3_FLOW_LOW;
	276	+ rx_coal->flow_level = prx_coal->flow_level;
	277	+ tx_coal->flow_level = ptx_coal->flow_level;
283	278
284	279	/* device version above V3(include V3), GL can configure 1us
285	280	* unit, so uses 1us unit.
@@ -294,8 +289,8 @@ static void hns3_vector_coalesce_init(struct hns3_enet_tqp_vector *tqp_vector, · 8 unmodified lines
294	289	rx_coal->ql_enable = 1;
295	290	tx_coal->int_ql_max = ae_dev->dev_specs.int_ql_max;
296	291	rx_coal->int_ql_max = ae_dev->dev_specs.int_ql_max;
297		− tx_coal->int_ql = HNS3_INT_QL_DEFAULT_CFG;
298		− rx_coal->int_ql = HNS3_INT_QL_DEFAULT_CFG;
	292	+ tx_coal->int_ql = ptx_coal->int_ql;
	293	+ rx_coal->int_ql = prx_coal->int_ql;
299	294	}
300	295	}
301	296
@@ -846,8 +841,6 @@ static bool hns3_tunnel_csum_bug(struct sk_buff *skb) · 544 unmodified lines
846	841	l4.udp->dest == htons(4790))))
847	842	return false;
848	843
849		− skb_checksum_help(skb);
850		−
851	844	return true;
852	845	}
853	846
@@ -924,8 +917,7 @@ static int hns3_set_l2l3l4(struct sk_buff *skb, u8 ol4_proto, · 70 unmodified lines
924	917	/* the stack computes the IP header already,
925	918	* driver calculate l4 checksum when not TSO.
926	919	*/
927		− skb_checksum_help(skb);
928		− return 0;
	920	+ return skb_checksum_help(skb);
929	921	}
930	922
931	923	hns3_set_outer_l2l3l4(skb, ol4_proto, ol_type_vlan_len_msec);
@@ -970,7 +962,7 @@ static int hns3_set_l2l3l4(struct sk_buff *skb, u8 ol4_proto, · 38 unmodified lines
970	962	break;
971	963	case IPPROTO_UDP:
972	964	if (hns3_tunnel_csum_bug(skb))
973		− break;
	965	+ return skb_checksum_help(skb);
974	966
975	967	hns3_set_field(*type_cs_vlan_tso, HNS3_TXD_L4CS_B, 1);
976	968	hns3_set_field(*type_cs_vlan_tso, HNS3_TXD_L4T_S,
@@ -995,8 +987,7 @@ static int hns3_set_l2l3l4(struct sk_buff *skb, u8 ol4_proto, · 18 unmodified lines
995	987	/* the stack computes the IP header already,
996	988	* driver calculate l4 checksum when not TSO.
997	989	*/
998		− skb_checksum_help(skb);
999		− return 0;
	990	+ return skb_checksum_help(skb);
1000	991	}
1001	992
1002	993	return 0;
@@ -3844,6 +3835,34 @@ map_ring_fail: · 2841 unmodified lines
3844	3835	return ret;
3845	3836	}
3846	3837
	3838	+static void hns3_nic_init_coal_cfg(struct hns3_nic_priv *priv)
	3839	+{
	3840	+ struct hnae3_ae_dev *ae_dev = pci_get_drvdata(priv->ae_handle->pdev);
	3841	+ struct hns3_enet_coalesce *tx_coal = &priv->tx_coal;
	3842	+ struct hns3_enet_coalesce *rx_coal = &priv->rx_coal;
	3843	+
	3844	+ /* initialize the configuration for interrupt coalescing.
	3845	+ * 1. GL (Interrupt Gap Limiter)
	3846	+ * 2. RL (Interrupt Rate Limiter)
	3847	+ * 3. QL (Interrupt Quantity Limiter)
	3848	+ *
	3849	+ * Default: enable interrupt coalescing self-adaptive and GL
	3850	+ */
	3851	+ tx_coal->adapt_enable = 1;
	3852	+ rx_coal->adapt_enable = 1;
	3853	+
	3854	+ tx_coal->int_gl = HNS3_INT_GL_50K;
	3855	+ rx_coal->int_gl = HNS3_INT_GL_50K;
	3856	+
	3857	+ rx_coal->flow_level = HNS3_FLOW_LOW;
	3858	+ tx_coal->flow_level = HNS3_FLOW_LOW;
	3859	+
	3860	+ if (ae_dev->dev_specs.int_ql_max) {
	3861	+ tx_coal->int_ql = HNS3_INT_QL_DEFAULT_CFG;
	3862	+ rx_coal->int_ql = HNS3_INT_QL_DEFAULT_CFG;
	3863	+ }
	3864	+}
	3865	+
3847	3866	static int hns3_nic_alloc_vector_data(struct hns3_nic_priv *priv)
3848	3867	{
3849	3868	struct hnae3_handle *h = priv->ae_handle;
@@ -4295,6 +4314,8 @@ static int hns3_client_init(struct hnae3_handle *handle) · 445 unmodified lines
4295	4314	goto out_get_ring_cfg;
4296	4315	}
4297	4316
	4317	+ hns3_nic_init_coal_cfg(priv);
	4318	+
4298	4319	ret = hns3_nic_alloc_vector_data(priv);
4299	4320	if (ret) {
4300	4321	ret = -ENOMEM;
@@ -4317,12 +4338,6 @@ static int hns3_client_init(struct hnae3_handle *handle) · 16 unmodified lines
4317	4338	if (ret)
4318	4339	goto out_init_phy;
4319	4340
4320		− ret = register_netdev(netdev);
4321		− if (ret) {
4322		− dev_err(priv->dev, "probe register netdev fail!\n");
4323		− goto out_reg_netdev_fail;
4324		− }
4325		−
4326	4341	/* the device can work without cpu rmap, only aRFS needs it */
4327	4342	ret = hns3_set_rx_cpu_rmap(netdev);
4328	4343	if (ret)
@@ -4355,17 +4370,23 @@ static int hns3_client_init(struct hnae3_handle *handle) · 26 unmodified lines
4355	4370	if (ae_dev->dev_version >= HNAE3_DEVICE_VERSION_V3)
4356	4371	set_bit(HNAE3_PFLAG_LIMIT_PROMISC, &handle->supported_pflags);
4357	4372
	4373	+ ret = register_netdev(netdev);
	4374	+ if (ret) {
	4375	+ dev_err(priv->dev, "probe register netdev fail!\n");
	4376	+ goto out_reg_netdev_fail;
	4377	+ }
	4378	+
4358	4379	if (netif_msg_drv(handle))
4359	4380	hns3_info_show(priv);
4360	4381
4361	4382	return ret;
4362	4383
	4384	+out_reg_netdev_fail:
	4385	+ hns3_dbg_uninit(handle);
4363	4386	out_client_start:
4364	4387	hns3_free_rx_cpu_rmap(netdev);
4365	4388	hns3_nic_uninit_irq(priv);
4366	4389	out_init_irq_fail:
4367		− unregister_netdev(netdev);
4368		−out_reg_netdev_fail:
4369	4390	hns3_uninit_phy(netdev);
4370	4391	out_init_phy:
4371	4392	hns3_uninit_all_ring(priv);
@@ -4571,31 +4592,6 @@ int hns3_nic_reset_all_ring(struct hnae3_handle *h) · 199 unmodified lines
4571	4592	return 0;
4572	4593	}
4573	4594
4574		−static void hns3_store_coal(struct hns3_nic_priv *priv)
4575		−{
4576		− /* ethtool only support setting and querying one coal
4577		− * configuration for now, so save the vector 0' coal
4578		− * configuration here in order to restore it.
4579		− */
4580		− memcpy(&priv->tx_coal, &priv->tqp_vector[0].tx_group.coal,
4581		− sizeof(struct hns3_enet_coalesce));
4582		− memcpy(&priv->rx_coal, &priv->tqp_vector[0].rx_group.coal,
4583		− sizeof(struct hns3_enet_coalesce));
4584		−}
4585		−
4586		−static void hns3_restore_coal(struct hns3_nic_priv *priv)
4587		−{
4588		− u16 vector_num = priv->vector_num;
4589		− int i;
4590		−
4591		− for (i = 0; i < vector_num; i++) {
4592		− memcpy(&priv->tqp_vector[i].tx_group.coal, &priv->tx_coal,
4593		− sizeof(struct hns3_enet_coalesce));
4594		− memcpy(&priv->tqp_vector[i].rx_group.coal, &priv->rx_coal,
4595		− sizeof(struct hns3_enet_coalesce));
4596		− }
4597		−}
4598		−
4599	4595	static int hns3_reset_notify_down_enet(struct hnae3_handle *handle)
4600	4596	{
4601	4597	struct hnae3_knic_private_info *kinfo = &handle->kinfo;
@@ -4654,8 +4650,6 @@ static int hns3_reset_notify_init_enet(struct hnae3_handle *handle) · 52 unmodified lines
4654	4650	if (ret)
4655	4651	goto err_put_ring;
4656	4652
4657		− hns3_restore_coal(priv);
4658		−
4659	4653	ret = hns3_nic_init_vector_data(priv);
4660	4654	if (ret)
4661	4655	goto err_dealloc_vector;
@@ -4721,8 +4715,6 @@ static int hns3_reset_notify_uninit_enet(struct hnae3_handle *handle) · 59 unmodified lines
4721	4715
4722	4716	hns3_nic_uninit_vector_data(priv);
4723	4717
4724		− hns3_store_coal(priv);
4725		−
4726	4718	hns3_nic_dealloc_vector_data(priv);
4727	4719
4728	4720	hns3_uninit_all_ring(priv);

drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c+23−41

@@ -1134,50 +1134,32 @@ static void hns3_get_channels(struct net_device *netdev, · 1133 unmodified lines
1134	1134	h->ae_algo->ops->get_channels(h, ch);
1135	1135	}
1136	1136
1137		−static int hns3_get_coalesce_per_queue(struct net_device *netdev, u32 queue,
1138		− struct ethtool_coalesce *cmd)
	1137	+static int hns3_get_coalesce(struct net_device *netdev,
	1138	+ struct ethtool_coalesce *cmd)
1139	1139	{
1140		− struct hns3_enet_tqp_vector tx_vector, rx_vector;
1141	1140	struct hns3_nic_priv *priv = netdev_priv(netdev);
	1141	+ struct hns3_enet_coalesce *tx_coal = &priv->tx_coal;
	1142	+ struct hns3_enet_coalesce *rx_coal = &priv->rx_coal;
1142	1143	struct hnae3_handle *h = priv->ae_handle;
1143		− u16 queue_num = h->kinfo.num_tqps;
1144	1144
1145	1145	if (hns3_nic_resetting(netdev))
1146	1146	return -EBUSY;
1147	1147
1148		− if (queue >= queue_num) {
1149		− netdev_err(netdev,
1150		− "Invalid queue value %u! Queue max id=%u\n",
1151		− queue, queue_num - 1);
1152		− return -EINVAL;
1153		− }
1154		−
1155		− tx_vector = priv->ring[queue].tqp_vector;
1156		− rx_vector = priv->ring[queue_num + queue].tqp_vector;
	1148	+ cmd->use_adaptive_tx_coalesce = tx_coal->adapt_enable;
	1149	+ cmd->use_adaptive_rx_coalesce = rx_coal->adapt_enable;
1157	1150
1158		− cmd->use_adaptive_tx_coalesce =
1159		− tx_vector->tx_group.coal.adapt_enable;
1160		− cmd->use_adaptive_rx_coalesce =
1161		− rx_vector->rx_group.coal.adapt_enable;
1162		−
1163		− cmd->tx_coalesce_usecs = tx_vector->tx_group.coal.int_gl;
1164		− cmd->rx_coalesce_usecs = rx_vector->rx_group.coal.int_gl;
	1151	+ cmd->tx_coalesce_usecs = tx_coal->int_gl;
	1152	+ cmd->rx_coalesce_usecs = rx_coal->int_gl;
1165	1153
1166	1154	cmd->tx_coalesce_usecs_high = h->kinfo.int_rl_setting;
1167	1155	cmd->rx_coalesce_usecs_high = h->kinfo.int_rl_setting;
1168	1156
1169		− cmd->tx_max_coalesced_frames = tx_vector->tx_group.coal.int_ql;
1170		− cmd->rx_max_coalesced_frames = rx_vector->rx_group.coal.int_ql;
	1157	+ cmd->tx_max_coalesced_frames = tx_coal->int_ql;
	1158	+ cmd->rx_max_coalesced_frames = rx_coal->int_ql;
1171	1159
1172	1160	return 0;
1173	1161	}
1174	1162
1175		−static int hns3_get_coalesce(struct net_device *netdev,
1176		− struct ethtool_coalesce *cmd)
1177		−{
1178		− return hns3_get_coalesce_per_queue(netdev, 0, cmd);
1179		−}
1180		−
1181	1163	static int hns3_check_gl_coalesce_para(struct net_device *netdev,
1182	1164	struct ethtool_coalesce *cmd)
1183	1165	{
@@ -1292,19 +1274,7 @@ static int hns3_check_coalesce_para(struct net_device *netdev, · 108 unmodified lines
1292	1274	return ret;
1293	1275	}
1294	1276
1295		− ret = hns3_check_ql_coalesce_param(netdev, cmd);
1296		− if (ret)
1297		− return ret;
1298		−
1299		− if (cmd->use_adaptive_tx_coalesce == 1 \|\|
1300		− cmd->use_adaptive_rx_coalesce == 1) {
1301		− netdev_info(netdev,
1302		− "adaptive-tx=%u and adaptive-rx=%u, tx_usecs or rx_usecs will changed dynamically.\n",
1303		− cmd->use_adaptive_tx_coalesce,
1304		− cmd->use_adaptive_rx_coalesce);
1305		− }
1306		−
1307		− return 0;
	1277	+ return hns3_check_ql_coalesce_param(netdev, cmd);
1308	1278	}
1309	1279
1310	1280	static void hns3_set_coalesce_per_queue(struct net_device *netdev,
@@ -1350,6 +1320,9 @@ static int hns3_set_coalesce(struct net_device *netdev, · 39 unmodified lines
1350	1320	struct ethtool_coalesce *cmd)
1351	1321	{
1352	1322	struct hnae3_handle *h = hns3_get_handle(netdev);
	1323	+ struct hns3_nic_priv *priv = netdev_priv(netdev);
	1324	+ struct hns3_enet_coalesce *tx_coal = &priv->tx_coal;
	1325	+ struct hns3_enet_coalesce *rx_coal = &priv->rx_coal;
1353	1326	u16 queue_num = h->kinfo.num_tqps;
1354	1327	int ret;
1355	1328	int i;
@@ -1364,6 +1337,15 @@ static int hns3_set_coalesce(struct net_device *netdev, · 8 unmodified lines
1364	1337	h->kinfo.int_rl_setting =
1365	1338	hns3_rl_round_down(cmd->rx_coalesce_usecs_high);
1366	1339
	1340	+ tx_coal->adapt_enable = cmd->use_adaptive_tx_coalesce;
	1341	+ rx_coal->adapt_enable = cmd->use_adaptive_rx_coalesce;
	1342	+
	1343	+ tx_coal->int_gl = cmd->tx_coalesce_usecs;
	1344	+ rx_coal->int_gl = cmd->rx_coalesce_usecs;
	1345	+
	1346	+ tx_coal->int_ql = cmd->tx_max_coalesced_frames;
	1347	+ rx_coal->int_ql = cmd->rx_max_coalesced_frames;
	1348	+
1367	1349	for (i = 0; i < queue_num; i++)
1368	1350	hns3_set_coalesce_per_queue(netdev, cmd, i);
1369	1351

drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_mbx.c+3−1

@@ -710,7 +710,6 @@ void hclge_mbx_handler(struct hclge_dev *hdev) · 709 unmodified lines
710	710	unsigned int flag;
711	711	int ret = 0;
712	712
713		− memset(&resp_msg, 0, sizeof(resp_msg));
714	713	/* handle all the mailbox requests in the queue */
715	714	while (!hclge_cmd_crq_empty(&hdev->hw)) {
716	715	if (test_bit(HCLGE_STATE_CMD_DISABLE, &hdev->state)) {
@@ -738,6 +737,9 @@ void hclge_mbx_handler(struct hclge_dev *hdev) · 21 unmodified lines
738	737
739	738	trace_hclge_pf_mbx_get(hdev, req);
740	739
	740	+ /* clear the resp_msg before processing every mailbox message */
	741	+ memset(&resp_msg, 0, sizeof(resp_msg));
	742	+
741	743	switch (req->msg.code) {
742	744	case HCLGE_MBX_MAP_RING_TO_VECTOR:
743	745	ret = hclge_map_unmap_ring_to_vf_vector(vport, true,

drivers/net/ethernet/intel/i40e/i40e_txrx.c+6−1

@@ -2313,15 +2313,20 @@ static int i40e_run_xdp(struct i40e_ring rx_ring, struct xdp_buff xdp) · 2312 unmodified lines
2313	2313	case XDP_TX:
2314	2314	xdp_ring = rx_ring->vsi->xdp_rings[rx_ring->queue_index];
2315	2315	result = i40e_xmit_xdp_tx_ring(xdp, xdp_ring);
	2316	+ if (result == I40E_XDP_CONSUMED)
	2317	+ goto out_failure;
2316	2318	break;
2317	2319	case XDP_REDIRECT:
2318	2320	err = xdp_do_redirect(rx_ring->netdev, xdp, xdp_prog);
2319		− result = !err ? I40E_XDP_REDIR : I40E_XDP_CONSUMED;
	2321	+ if (err)
	2322	+ goto out_failure;
	2323	+ result = I40E_XDP_REDIR;
2320	2324	break;
2321	2325	default:
2322	2326	bpf_warn_invalid_xdp_action(act);
2323	2327	fallthrough;
2324	2328	case XDP_ABORTED:
	2329	+out_failure:
2325	2330	trace_xdp_exception(rx_ring->netdev, xdp_prog, act);
2326	2331	fallthrough; /* handle aborts by dropping packet */
2327	2332	case XDP_DROP:

drivers/net/ethernet/intel/i40e/i40e_xsk.c+6−2

@@ -162,9 +162,10 @@ static int i40e_run_xdp_zc(struct i40e_ring rx_ring, struct xdp_buff xdp) · 161 unmodified lines
162	162
163	163	if (likely(act == XDP_REDIRECT)) {
164	164	err = xdp_do_redirect(rx_ring->netdev, xdp, xdp_prog);
165		− result = !err ? I40E_XDP_REDIR : I40E_XDP_CONSUMED;
	165	+ if (err)
	166	+ goto out_failure;
166	167	rcu_read_unlock();
167		− return result;
	168	+ return I40E_XDP_REDIR;
168	169	}
169	170
170	171	switch (act) {
@@ -173,11 +174,14 @@ static int i40e_run_xdp_zc(struct i40e_ring rx_ring, struct xdp_buff xdp) · 2 unmodified lines
173	174	case XDP_TX:
174	175	xdp_ring = rx_ring->vsi->xdp_rings[rx_ring->queue_index];
175	176	result = i40e_xmit_xdp_tx_ring(xdp, xdp_ring);
	177	+ if (result == I40E_XDP_CONSUMED)
	178	+ goto out_failure;
176	179	break;
177	180	default:
178	181	bpf_warn_invalid_xdp_action(act);
179	182	fallthrough;
180	183	case XDP_ABORTED:
	184	+out_failure:
181	185	trace_xdp_exception(rx_ring->netdev, xdp_prog, act);
182	186	fallthrough; /* handle aborts by dropping packet */
183	187	case XDP_DROP:

drivers/net/ethernet/intel/ice/ice.h+5−3

@@ -335,6 +335,7 @@ struct ice_vsi { · 334 unmodified lines
335	335	struct ice_tc_cfg tc_cfg;
336	336	struct bpf_prog *xdp_prog;
337	337	struct ice_ring *xdp_rings; / XDP ring array */
	338	+ unsigned long af_xdp_zc_qps; / tracks AF_XDP ZC enabled qps */
338	339	u16 num_xdp_txq; /* Used XDP queues */
339	340	u8 xdp_mapping_mode; /* ICE_MAP_MODE_[CONTIG\|SCATTER] */
340	341
@@ -547,15 +548,16 @@ static inline void ice_set_ring_xdp(struct ice_ring *ring) · 206 unmodified lines
547	548	*/
548	549	static inline struct xsk_buff_pool ice_xsk_pool(struct ice_ring ring)
549	550	{
	551	+ struct ice_vsi *vsi = ring->vsi;
550	552	u16 qid = ring->q_index;
551	553
552	554	if (ice_ring_is_xdp(ring))
553		− qid -= ring->vsi->num_xdp_txq;
	555	+ qid -= vsi->num_xdp_txq;
554	556
555		− if (!ice_is_xdp_ena_vsi(ring->vsi))
	557	+ if (!ice_is_xdp_ena_vsi(vsi) \|\| !test_bit(qid, vsi->af_xdp_zc_qps))
556	558	return NULL;
557	559
558		− return xsk_get_pool_from_qid(ring->vsi->netdev, qid);
	560	+ return xsk_get_pool_from_qid(vsi->netdev, qid);
559	561	}
560	562
561	563	/**

drivers/net/ethernet/intel/ice/ice_ethtool.c+6−45

@@ -1773,49 +1773,6 @@ ice_phy_type_to_ethtool(struct net_device *netdev, · 1772 unmodified lines
1773	1773	ice_ethtool_advertise_link_mode(ICE_AQ_LINK_SPEED_100GB,
1774	1774	100000baseKR4_Full);
1775	1775	}
1776		−
1777		− /* Autoneg PHY types */
1778		− if (phy_types_low & ICE_PHY_TYPE_LOW_100BASE_TX \|\|
1779		− phy_types_low & ICE_PHY_TYPE_LOW_1000BASE_T \|\|
1780		− phy_types_low & ICE_PHY_TYPE_LOW_1000BASE_KX \|\|
1781		− phy_types_low & ICE_PHY_TYPE_LOW_2500BASE_T \|\|
1782		− phy_types_low & ICE_PHY_TYPE_LOW_2500BASE_KX \|\|
1783		− phy_types_low & ICE_PHY_TYPE_LOW_5GBASE_T \|\|
1784		− phy_types_low & ICE_PHY_TYPE_LOW_5GBASE_KR \|\|
1785		− phy_types_low & ICE_PHY_TYPE_LOW_10GBASE_T \|\|
1786		− phy_types_low & ICE_PHY_TYPE_LOW_10GBASE_KR_CR1 \|\|
1787		− phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_T \|\|
1788		− phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_CR \|\|
1789		− phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_CR_S \|\|
1790		− phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_CR1 \|\|
1791		− phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_KR \|\|
1792		− phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_KR_S \|\|
1793		− phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_KR1 \|\|
1794		− phy_types_low & ICE_PHY_TYPE_LOW_40GBASE_CR4 \|\|
1795		− phy_types_low & ICE_PHY_TYPE_LOW_40GBASE_KR4) {
1796		− ethtool_link_ksettings_add_link_mode(ks, supported,
1797		− Autoneg);
1798		− ethtool_link_ksettings_add_link_mode(ks, advertising,
1799		− Autoneg);
1800		− }
1801		− if (phy_types_low & ICE_PHY_TYPE_LOW_50GBASE_CR2 \|\|
1802		− phy_types_low & ICE_PHY_TYPE_LOW_50GBASE_KR2 \|\|
1803		− phy_types_low & ICE_PHY_TYPE_LOW_50GBASE_CP \|\|
1804		− phy_types_low & ICE_PHY_TYPE_LOW_50GBASE_KR_PAM4) {
1805		− ethtool_link_ksettings_add_link_mode(ks, supported,
1806		− Autoneg);
1807		− ethtool_link_ksettings_add_link_mode(ks, advertising,
1808		− Autoneg);
1809		− }
1810		− if (phy_types_low & ICE_PHY_TYPE_LOW_100GBASE_CR4 \|\|
1811		− phy_types_low & ICE_PHY_TYPE_LOW_100GBASE_KR4 \|\|
1812		− phy_types_low & ICE_PHY_TYPE_LOW_100GBASE_KR_PAM4 \|\|
1813		− phy_types_low & ICE_PHY_TYPE_LOW_100GBASE_CP2) {
1814		− ethtool_link_ksettings_add_link_mode(ks, supported,
1815		− Autoneg);
1816		− ethtool_link_ksettings_add_link_mode(ks, advertising,
1817		− Autoneg);
1818		− }
1819	1776	}
1820	1777
1821	1778	#define TEST_SET_BITS_TIMEOUT 50
@@ -1972,9 +1929,7 @@ ice_get_link_ksettings(struct net_device *netdev, · 150 unmodified lines
1972	1929	ks->base.port = PORT_TP;
1973	1930	break;
1974	1931	case ICE_MEDIA_BACKPLANE:
1975		− ethtool_link_ksettings_add_link_mode(ks, supported, Autoneg);
1976	1932	ethtool_link_ksettings_add_link_mode(ks, supported, Backplane);
1977		− ethtool_link_ksettings_add_link_mode(ks, advertising, Autoneg);
1978	1933	ethtool_link_ksettings_add_link_mode(ks, advertising,
1979	1934	Backplane);
1980	1935	ks->base.port = PORT_NONE;
@@ -2049,6 +2004,12 @@ ice_get_link_ksettings(struct net_device *netdev, · 68 unmodified lines
2049	2004	if (caps->link_fec_options & ICE_AQC_PHY_FEC_25G_RS_CLAUSE91_EN)
2050	2005	ethtool_link_ksettings_add_link_mode(ks, supported, FEC_RS);
2051	2006
	2007	+ /* Set supported and advertised autoneg */
	2008	+ if (ice_is_phy_caps_an_enabled(caps)) {
	2009	+ ethtool_link_ksettings_add_link_mode(ks, supported, Autoneg);
	2010	+ ethtool_link_ksettings_add_link_mode(ks, advertising, Autoneg);
	2011	+ }
	2012	+
2052	2013	done:
2053	2014	kfree(caps);
2054	2015	return err;

drivers/net/ethernet/intel/ice/ice_hw_autogen.h+1−0

@@ -31,6 +31,7 @@ · 30 unmodified lines
31	31	#define PF_FW_ATQLEN_ATQOVFL_M BIT(29)
32	32	#define PF_FW_ATQLEN_ATQCRIT_M BIT(30)
33	33	#define VF_MBX_ARQLEN(_VF) (0x0022BC00 + ((_VF) * 4))
	34	+#define VF_MBX_ATQLEN(_VF) (0x0022A800 + ((_VF) * 4))
34	35	#define PF_FW_ATQLEN_ATQENABLE_M BIT(31)
35	36	#define PF_FW_ATQT 0x00080400
36	37	#define PF_MBX_ARQBAH 0x0022E400

drivers/net/ethernet/intel/ice/ice_lib.c+22−8

@@ -105,8 +105,14 @@ static int ice_vsi_alloc_arrays(struct ice_vsi *vsi) · 104 unmodified lines
105	105	if (!vsi->q_vectors)
106	106	goto err_vectors;
107	107
	108	+ vsi->af_xdp_zc_qps = bitmap_zalloc(max_t(int, vsi->alloc_txq, vsi->alloc_rxq), GFP_KERNEL);
	109	+ if (!vsi->af_xdp_zc_qps)
	110	+ goto err_zc_qps;
	111	+
108	112	return 0;
109	113
	114	+err_zc_qps:
	115	+ devm_kfree(dev, vsi->q_vectors);
110	116	err_vectors:
111	117	devm_kfree(dev, vsi->rxq_map);
112	118	err_rxq_map:
@@ -194,6 +200,8 @@ static void ice_vsi_set_num_qs(struct ice_vsi *vsi, u16 vf_id) · 81 unmodified lines
194	200	break;
195	201	case ICE_VSI_VF:
196	202	vf = &pf->vf[vsi->vf_id];
	203	+ if (vf->num_req_qs)
	204	+ vf->num_vf_qs = vf->num_req_qs;
197	205	vsi->alloc_txq = vf->num_vf_qs;
198	206	vsi->alloc_rxq = vf->num_vf_qs;
199	207	/* pf->num_msix_per_vf includes (VF miscellaneous vector +
@@ -288,6 +296,10 @@ static void ice_vsi_free_arrays(struct ice_vsi *vsi) · 88 unmodified lines
288	296
289	297	dev = ice_pf_to_dev(pf);
290	298
	299	+ if (vsi->af_xdp_zc_qps) {
	300	+ bitmap_free(vsi->af_xdp_zc_qps);
	301	+ vsi->af_xdp_zc_qps = NULL;
	302	+ }
291	303	/* free the ring and vector containers */
292	304	if (vsi->q_vectors) {
293	305	devm_kfree(dev, vsi->q_vectors);
@@ -1705,12 +1717,13 @@ setup_rings: · 1411 unmodified lines
1705	1717	* ice_vsi_cfg_txqs - Configure the VSI for Tx
1706	1718	* @vsi: the VSI being configured
1707	1719	* @rings: Tx ring array to be configured
	1720	+ * @count: number of Tx ring array elements
1708	1721	*
1709	1722	* Return 0 on success and a negative value on error
1710	1723	* Configure the Tx VSI for operation.
1711	1724	*/
1712	1725	static int
1713		−ice_vsi_cfg_txqs(struct ice_vsi vsi, struct ice_ring *rings)
	1726	+ice_vsi_cfg_txqs(struct ice_vsi vsi, struct ice_ring *rings, u16 count)
1714	1727	{
1715	1728	struct ice_aqc_add_tx_qgrp *qg_buf;
1716	1729	u16 q_idx = 0;
@@ -1722,7 +1735,7 @@ ice_vsi_cfg_txqs(struct ice_vsi vsi, struct ice_ring *rings) · 5 unmodified lines
1722	1735
1723	1736	qg_buf->num_txqs = 1;
1724	1737
1725		− for (q_idx = 0; q_idx < vsi->num_txq; q_idx++) {
	1738	+ for (q_idx = 0; q_idx < count; q_idx++) {
1726	1739	err = ice_vsi_cfg_txq(vsi, rings[q_idx], qg_buf);
1727	1740	if (err)
1728	1741	goto err_cfg_txqs;
@@ -1742,7 +1755,7 @@ err_cfg_txqs: · 13 unmodified lines
1742	1755	*/
1743	1756	int ice_vsi_cfg_lan_txqs(struct ice_vsi *vsi)
1744	1757	{
1745		− return ice_vsi_cfg_txqs(vsi, vsi->tx_rings);
	1758	+ return ice_vsi_cfg_txqs(vsi, vsi->tx_rings, vsi->num_txq);
1746	1759	}
1747	1760
1748	1761	/**
@@ -1757,7 +1770,7 @@ int ice_vsi_cfg_xdp_txqs(struct ice_vsi *vsi) · 8 unmodified lines
1757	1770	int ret;
1758	1771	int i;
1759	1772
1760		− ret = ice_vsi_cfg_txqs(vsi, vsi->xdp_rings);
	1773	+ ret = ice_vsi_cfg_txqs(vsi, vsi->xdp_rings, vsi->num_xdp_txq);
1761	1774	if (ret)
1762	1775	return ret;
1763	1776
@@ -1997,17 +2010,18 @@ int ice_vsi_stop_all_rx_rings(struct ice_vsi *vsi) · 233 unmodified lines
1997	2010	* @rst_src: reset source
1998	2011	* @rel_vmvf_num: Relative ID of VF/VM
1999	2012	* @rings: Tx ring array to be stopped
	2013	+ * @count: number of Tx ring array elements
2000	2014	*/
2001	2015	static int
2002	2016	ice_vsi_stop_tx_rings(struct ice_vsi *vsi, enum ice_disq_rst_src rst_src,
2003		− u16 rel_vmvf_num, struct ice_ring **rings)
	2017	+ u16 rel_vmvf_num, struct ice_ring **rings, u16 count)
2004	2018	{
2005	2019	u16 q_idx;
2006	2020
2007	2021	if (vsi->num_txq > ICE_LAN_TXQ_MAX_QDIS)
2008	2022	return -EINVAL;
2009	2023
2010		− for (q_idx = 0; q_idx < vsi->num_txq; q_idx++) {
	2024	+ for (q_idx = 0; q_idx < count; q_idx++) {
2011	2025	struct ice_txq_meta txq_meta = { };
2012	2026	int status;
2013	2027
@@ -2035,7 +2049,7 @@ int · 21 unmodified lines
2035	2049	ice_vsi_stop_lan_tx_rings(struct ice_vsi *vsi, enum ice_disq_rst_src rst_src,
2036	2050	u16 rel_vmvf_num)
2037	2051	{
2038		− return ice_vsi_stop_tx_rings(vsi, rst_src, rel_vmvf_num, vsi->tx_rings);
	2052	+ return ice_vsi_stop_tx_rings(vsi, rst_src, rel_vmvf_num, vsi->tx_rings, vsi->num_txq);
2039	2053	}
2040	2054
2041	2055	/**
@@ -2044,7 +2058,7 @@ ice_vsi_stop_lan_tx_rings(struct ice_vsi *vsi, enum ice_disq_rst_src rst_src, · 2 unmodified lines
2044	2058	*/
2045	2059	int ice_vsi_stop_xdp_tx_rings(struct ice_vsi *vsi)
2046	2060	{
2047		− return ice_vsi_stop_tx_rings(vsi, ICE_NO_RESET, 0, vsi->xdp_rings);
	2061	+ return ice_vsi_stop_tx_rings(vsi, ICE_NO_RESET, 0, vsi->xdp_rings, vsi->num_xdp_txq);
2048	2062	}
2049	2063
2050	2064	/**

drivers/net/ethernet/intel/ice/ice_main.c+15−0

@@ -2555,6 +2555,20 @@ ice_xdp_setup_prog(struct ice_vsi vsi, struct bpf_prog prog, · 2554 unmodified lines
2555	2555	return (ret \|\| xdp_ring_err) ? -ENOMEM : 0;
2556	2556	}
2557	2557
	2558	+/**
	2559	+ * ice_xdp_safe_mode - XDP handler for safe mode
	2560	+ * @dev: netdevice
	2561	+ * @xdp: XDP command
	2562	+ */
	2563	+static int ice_xdp_safe_mode(struct net_device __always_unused *dev,
	2564	+ struct netdev_bpf *xdp)
	2565	+{
	2566	+ NL_SET_ERR_MSG_MOD(xdp->extack,
	2567	+ "Please provide working DDP firmware package in order to use XDP\n"
	2568	+ "Refer to Documentation/networking/device_drivers/ethernet/intel/ice.rst");
	2569	+ return -EOPNOTSUPP;
	2570	+}
	2571	+
2558	2572	/**
2559	2573	* ice_xdp - implements XDP handler
2560	2574	* @dev: netdevice
@@ -6937,6 +6951,7 @@ static const struct net_device_ops ice_netdev_safe_mode_ops = { · 4376 unmodified lines
6937	6951	.ndo_change_mtu = ice_change_mtu,
6938	6952	.ndo_get_stats64 = ice_get_stats64,
6939	6953	.ndo_tx_timeout = ice_tx_timeout,
	6954	+ .ndo_bpf = ice_xdp_safe_mode,
6940	6955	};
6941	6956
6942	6957	static const struct net_device_ops ice_netdev_ops = {

drivers/net/ethernet/intel/ice/ice_txrx.c+13−4

@@ -523,7 +523,7 @@ ice_run_xdp(struct ice_ring rx_ring, struct xdp_buff xdp, · 522 unmodified lines
523	523	struct bpf_prog *xdp_prog)
524	524	{
525	525	struct ice_ring *xdp_ring;
526		− int err;
	526	+ int err, result;
527	527	u32 act;
528	528
529	529	act = bpf_prog_run_xdp(xdp_prog, xdp);
@@ -532,14 +532,20 @@ ice_run_xdp(struct ice_ring rx_ring, struct xdp_buff xdp, · 2 unmodified lines
532	532	return ICE_XDP_PASS;
533	533	case XDP_TX:
534	534	xdp_ring = rx_ring->vsi->xdp_rings[smp_processor_id()];
535		− return ice_xmit_xdp_buff(xdp, xdp_ring);
	535	+ result = ice_xmit_xdp_buff(xdp, xdp_ring);
	536	+ if (result == ICE_XDP_CONSUMED)
	537	+ goto out_failure;
	538	+ return result;
536	539	case XDP_REDIRECT:
537	540	err = xdp_do_redirect(rx_ring->netdev, xdp, xdp_prog);
538		− return !err ? ICE_XDP_REDIR : ICE_XDP_CONSUMED;
	541	+ if (err)
	542	+ goto out_failure;
	543	+ return ICE_XDP_REDIR;
539	544	default:
540	545	bpf_warn_invalid_xdp_action(act);
541	546	fallthrough;
542	547	case XDP_ABORTED:
	548	+out_failure:
543	549	trace_xdp_exception(rx_ring->netdev, xdp_prog, act);
544	550	fallthrough;
545	551	case XDP_DROP:
@@ -2143,6 +2149,7 @@ ice_xmit_frame_ring(struct sk_buff skb, struct ice_ring tx_ring) · 1597 unmodified lines
2143	2149	struct ice_tx_offload_params offload = { 0 };
2144	2150	struct ice_vsi *vsi = tx_ring->vsi;
2145	2151	struct ice_tx_buf *first;
	2152	+ struct ethhdr *eth;
2146	2153	unsigned int count;
2147	2154	int tso, csum;
2148	2155
@@ -2189,7 +2196,9 @@ ice_xmit_frame_ring(struct sk_buff skb, struct ice_ring tx_ring) · 40 unmodified lines
2189	2196	goto out_drop;
2190	2197
2191	2198	/* allow CONTROL frames egress from main VSI if FW LLDP disabled */
2192		− if (unlikely(skb->priority == TC_PRIO_CONTROL &&
	2199	+ eth = (struct ethhdr *)skb_mac_header(skb);
	2200	+ if (unlikely((skb->priority == TC_PRIO_CONTROL \|\|
	2201	+ eth->h_proto == htons(ETH_P_LLDP)) &&
2193	2202	vsi->type == ICE_VSI_PF &&
2194	2203	vsi->port_info->qos_cfg.is_sw_lldp))
2195	2204	offload.cd_qw1 \|= (u64)(ICE_TX_DESC_DTYPE_CTX \|

drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c+13−6

@@ -713,13 +713,15 @@ static void ice_trigger_vf_reset(struct ice_vf *vf, bool is_vflr, bool is_pfr) · 712 unmodified lines
713	713	*/
714	714	clear_bit(ICE_VF_STATE_INIT, vf->vf_states);
715	715
716		− /* VF_MBX_ARQLEN is cleared by PFR, so the driver needs to clear it
717		− * in the case of VFR. If this is done for PFR, it can mess up VF
718		− * resets because the VF driver may already have started cleanup
719		− * by the time we get here.
	716	+ /* VF_MBX_ARQLEN and VF_MBX_ATQLEN are cleared by PFR, so the driver
	717	+ * needs to clear them in the case of VFR/VFLR. If this is done for
	718	+ * PFR, it can mess up VF resets because the VF driver may already
	719	+ * have started cleanup by the time we get here.
720	720	*/
721		− if (!is_pfr)
	721	+ if (!is_pfr) {
722	722	wr32(hw, VF_MBX_ARQLEN(vf->vf_id), 0);
	723	+ wr32(hw, VF_MBX_ATQLEN(vf->vf_id), 0);
	724	+ }
723	725
724	726	/* In the case of a VFLR, the HW has already reset the VF and we
725	727	* just need to clean up, so don't hit the VFRTRIG register.
@@ -1698,7 +1700,12 @@ bool ice_reset_vf(struct ice_vf *vf, bool is_vflr) · 972 unmodified lines
1698	1700	ice_vf_ctrl_vsi_release(vf);
1699	1701
1700	1702	ice_vf_pre_vsi_rebuild(vf);
1701		− ice_vf_rebuild_vsi_with_release(vf);
	1703	+
	1704	+ if (ice_vf_rebuild_vsi_with_release(vf)) {
	1705	+ dev_err(dev, "Failed to release and setup the VF%u's VSI\n", vf->vf_id);
	1706	+ return false;
	1707	+ }
	1708	+
1702	1709	ice_vf_post_vsi_rebuild(vf);
1703	1710
1704	1711	/* if the VF has been reset allow it to come up again */

drivers/net/ethernet/intel/ice/ice_xsk.c+9−2

@@ -270,6 +270,7 @@ static int ice_xsk_pool_disable(struct ice_vsi *vsi, u16 qid) · 269 unmodified lines
270	270	if (!pool)
271	271	return -EINVAL;
272	272
	273	+ clear_bit(qid, vsi->af_xdp_zc_qps);
273	274	xsk_pool_dma_unmap(pool, ICE_RX_DMA_ATTR);
274	275
275	276	return 0;
@@ -300,6 +301,8 @@ ice_xsk_pool_enable(struct ice_vsi vsi, struct xsk_buff_pool pool, u16 qid) · 24 unmodified lines
300	301	if (err)
301	302	return err;
302	303
	304	+ set_bit(qid, vsi->af_xdp_zc_qps);
	305	+
303	306	return 0;
304	307	}
305	308
@@ -473,9 +476,10 @@ ice_run_xdp_zc(struct ice_ring rx_ring, struct xdp_buff xdp) · 167 unmodified lines
473	476
474	477	if (likely(act == XDP_REDIRECT)) {
475	478	err = xdp_do_redirect(rx_ring->netdev, xdp, xdp_prog);
476		− result = !err ? ICE_XDP_REDIR : ICE_XDP_CONSUMED;
	479	+ if (err)
	480	+ goto out_failure;
477	481	rcu_read_unlock();
478		− return result;
	482	+ return ICE_XDP_REDIR;
479	483	}
480	484
481	485	switch (act) {
@@ -484,11 +488,14 @@ ice_run_xdp_zc(struct ice_ring rx_ring, struct xdp_buff xdp) · 2 unmodified lines
484	488	case XDP_TX:
485	489	xdp_ring = rx_ring->vsi->xdp_rings[rx_ring->q_index];
486	490	result = ice_xmit_xdp_buff(xdp, xdp_ring);
	491	+ if (result == ICE_XDP_CONSUMED)
	492	+ goto out_failure;
487	493	break;
488	494	default:
489	495	bpf_warn_invalid_xdp_action(act);
490	496	fallthrough;
491	497	case XDP_ABORTED:
	498	+out_failure:
492	499	trace_xdp_exception(rx_ring->netdev, xdp_prog, act);
493	500	fallthrough;
494	501	case XDP_DROP:

drivers/net/ethernet/intel/igb/igb.h+1−1

@@ -749,7 +749,7 @@ void igb_ptp_rx_hang(struct igb_adapter *adapter); · 748 unmodified lines
749	749	void igb_ptp_tx_hang(struct igb_adapter *adapter);
750	750	void igb_ptp_rx_rgtstamp(struct igb_q_vector q_vector, struct sk_buff skb);
751	751	int igb_ptp_rx_pktstamp(struct igb_q_vector q_vector, void va,
752		− struct sk_buff *skb);
	752	+ ktime_t *timestamp);
753	753	int igb_ptp_set_ts_config(struct net_device netdev, struct ifreq ifr);
754	754	int igb_ptp_get_ts_config(struct net_device netdev, struct ifreq ifr);
755	755	void igb_set_flag_queue_pairs(struct igb_adapter *, const u32);

drivers/net/ethernet/intel/igb/igb_main.c+32−23

@@ -8280,7 +8280,7 @@ static void igb_add_rx_frag(struct igb_ring *rx_ring, · 8279 unmodified lines
8280	8280	static struct sk_buff igb_construct_skb(struct igb_ring rx_ring,
8281	8281	struct igb_rx_buffer *rx_buffer,
8282	8282	struct xdp_buff *xdp,
8283		− union e1000_adv_rx_desc *rx_desc)
	8283	+ ktime_t timestamp)
8284	8284	{
8285	8285	#if (PAGE_SIZE < 8192)
8286	8286	unsigned int truesize = igb_rx_pg_size(rx_ring) / 2;
@@ -8300,12 +8300,8 @@ static struct sk_buff igb_construct_skb(struct igb_ring rx_ring, · 13 unmodified lines
8300	8300	if (unlikely(!skb))
8301	8301	return NULL;
8302	8302
8303		− if (unlikely(igb_test_staterr(rx_desc, E1000_RXDADV_STAT_TSIP))) {
8304		− if (!igb_ptp_rx_pktstamp(rx_ring->q_vector, xdp->data, skb)) {
8305		− xdp->data += IGB_TS_HDR_LEN;
8306		− size -= IGB_TS_HDR_LEN;
8307		− }
8308		− }
	8303	+ if (timestamp)
	8304	+ skb_hwtstamps(skb)->hwtstamp = timestamp;
8309	8305
8310	8306	/* Determine available headroom for copy */
8311	8307	headlen = size;
@@ -8336,7 +8332,7 @@ static struct sk_buff igb_construct_skb(struct igb_ring rx_ring, · 24 unmodified lines
8336	8332	static struct sk_buff igb_build_skb(struct igb_ring rx_ring,
8337	8333	struct igb_rx_buffer *rx_buffer,
8338	8334	struct xdp_buff *xdp,
8339		− union e1000_adv_rx_desc *rx_desc)
	8335	+ ktime_t timestamp)
8340	8336	{
8341	8337	#if (PAGE_SIZE < 8192)
8342	8338	unsigned int truesize = igb_rx_pg_size(rx_ring) / 2;
@@ -8363,11 +8359,8 @@ static struct sk_buff igb_build_skb(struct igb_ring rx_ring, · 20 unmodified lines
8363	8359	if (metasize)
8364	8360	skb_metadata_set(skb, metasize);
8365	8361
8366		− /* pull timestamp out of packet data */
8367		− if (igb_test_staterr(rx_desc, E1000_RXDADV_STAT_TSIP)) {
8368		− if (!igb_ptp_rx_pktstamp(rx_ring->q_vector, skb->data, skb))
8369		− __skb_pull(skb, IGB_TS_HDR_LEN);
8370		− }
	8362	+ if (timestamp)
	8363	+ skb_hwtstamps(skb)->hwtstamp = timestamp;
8371	8364
8372	8365	/* update buffer offset */
8373	8366	#if (PAGE_SIZE < 8192)
@@ -8401,18 +8394,20 @@ static struct sk_buff igb_run_xdp(struct igb_adapter adapter, · 27 unmodified lines
8401	8394	break;
8402	8395	case XDP_TX:
8403	8396	result = igb_xdp_xmit_back(adapter, xdp);
	8397	+ if (result == IGB_XDP_CONSUMED)
	8398	+ goto out_failure;
8404	8399	break;
8405	8400	case XDP_REDIRECT:
8406	8401	err = xdp_do_redirect(adapter->netdev, xdp, xdp_prog);
8407		− if (!err)
8408		− result = IGB_XDP_REDIR;
8409		− else
8410		− result = IGB_XDP_CONSUMED;
	8402	+ if (err)
	8403	+ goto out_failure;
	8404	+ result = IGB_XDP_REDIR;
8411	8405	break;
8412	8406	default:
8413	8407	bpf_warn_invalid_xdp_action(act);
8414	8408	fallthrough;
8415	8409	case XDP_ABORTED:
	8410	+out_failure:
8416	8411	trace_xdp_exception(rx_ring->netdev, xdp_prog, act);
8417	8412	fallthrough;
8418	8413	case XDP_DROP:
@@ -8682,7 +8677,10 @@ static int igb_clean_rx_irq(struct igb_q_vector *q_vector, const int budget) · 263 unmodified lines
8682	8677	while (likely(total_packets < budget)) {
8683	8678	union e1000_adv_rx_desc *rx_desc;
8684	8679	struct igb_rx_buffer *rx_buffer;
	8680	+ ktime_t timestamp = 0;
	8681	+ int pkt_offset = 0;
8685	8682	unsigned int size;
	8683	+ void *pktbuf;
8686	8684
8687	8685	/* return some buffers to hardware, one at a time is too slow */
8688	8686	if (cleaned_count >= IGB_RX_BUFFER_WRITE) {
@@ -8702,14 +8700,24 @@ static int igb_clean_rx_irq(struct igb_q_vector *q_vector, const int budget) · 13 unmodified lines
8702	8700	dma_rmb();
8703	8701
8704	8702	rx_buffer = igb_get_rx_buffer(rx_ring, size, &rx_buf_pgcnt);
	8703	+ pktbuf = page_address(rx_buffer->page) + rx_buffer->page_offset;
	8704	+
	8705	+ /* pull rx packet timestamp if available and valid */
	8706	+ if (igb_test_staterr(rx_desc, E1000_RXDADV_STAT_TSIP)) {
	8707	+ int ts_hdr_len;
	8708	+
	8709	+ ts_hdr_len = igb_ptp_rx_pktstamp(rx_ring->q_vector,
	8710	+ pktbuf, &timestamp);
	8711	+
	8712	+ pkt_offset += ts_hdr_len;
	8713	+ size -= ts_hdr_len;
	8714	+ }
8705	8715
8706	8716	/* retrieve a buffer from the ring */
8707	8717	if (!skb) {
8708		− unsigned int offset = igb_rx_offset(rx_ring);
8709		− unsigned char *hard_start;
	8718	+ unsigned char *hard_start = pktbuf - igb_rx_offset(rx_ring);
	8719	+ unsigned int offset = pkt_offset + igb_rx_offset(rx_ring);
8710	8720
8711		− hard_start = page_address(rx_buffer->page) +
8712		− rx_buffer->page_offset - offset;
8713	8721	xdp_prepare_buff(&xdp, hard_start, offset, size, true);
8714	8722	#if (PAGE_SIZE > 4096)
8715	8723	/* At larger PAGE_SIZE, frame_sz depend on len size */
@@ -8732,10 +8740,11 @@ static int igb_clean_rx_irq(struct igb_q_vector *q_vector, const int budget) · 16 unmodified lines
8732	8740	} else if (skb)
8733	8741	igb_add_rx_frag(rx_ring, rx_buffer, skb, size);
8734	8742	else if (ring_uses_build_skb(rx_ring))
8735		− skb = igb_build_skb(rx_ring, rx_buffer, &xdp, rx_desc);
	8743	+ skb = igb_build_skb(rx_ring, rx_buffer, &xdp,
	8744	+ timestamp);
8736	8745	else
8737	8746	skb = igb_construct_skb(rx_ring, rx_buffer,
8738		− &xdp, rx_desc);
	8747	+ &xdp, timestamp);
8739	8748
8740	8749	/* exit if we failed to retrieve a buffer */
8741	8750	if (!skb) {

drivers/net/ethernet/intel/igb/igb_ptp.c+10−13

@@ -856,30 +856,28 @@ static void igb_ptp_tx_hwtstamp(struct igb_adapter *adapter) · 855 unmodified lines
856	856	dev_kfree_skb_any(skb);
857	857	}
858	858
859		−#define IGB_RET_PTP_DISABLED 1
860		−#define IGB_RET_PTP_INVALID 2
861		−
862	859	/**
863	860	* igb_ptp_rx_pktstamp - retrieve Rx per packet timestamp
864	861	* @q_vector: Pointer to interrupt specific structure
865	862	* @va: Pointer to address containing Rx buffer
866		− * @skb: Buffer containing timestamp and packet
	863	+ * @timestamp: Pointer where timestamp will be stored
867	864	*
868	865	* This function is meant to retrieve a timestamp from the first buffer of an
869	866	* incoming frame. The value is stored in little endian format starting on
870	867	* byte 8
871	868	*
872		− * Returns: 0 if success, nonzero if failure
	869	+ * Returns: The timestamp header length or 0 if not available
873	870	**/
874	871	int igb_ptp_rx_pktstamp(struct igb_q_vector q_vector, void va,
875		− struct sk_buff *skb)
	872	+ ktime_t *timestamp)
876	873	{
877	874	struct igb_adapter *adapter = q_vector->adapter;
	875	+ struct skb_shared_hwtstamps ts;
878	876	__le64 regval = (__le64 )va;
879	877	int adjust = 0;
880	878
881	879	if (!(adapter->ptp_flags & IGB_PTP_ENABLED))
882		− return IGB_RET_PTP_DISABLED;
	880	+ return 0;
883	881
884	882	/* The timestamp is recorded in little endian format.
885	883	* DWORD: 0 1 2 3
@@ -888,10 +886,9 @@ int igb_ptp_rx_pktstamp(struct igb_q_vector q_vector, void va, · 2 unmodified lines
888	886
889	887	/* check reserved dwords are zero, be/le doesn't matter for zero */
890	888	if (regval[0])
891		− return IGB_RET_PTP_INVALID;
	889	+ return 0;
892	890
893		− igb_ptp_systim_to_hwtstamp(adapter, skb_hwtstamps(skb),
894		− le64_to_cpu(regval[1]));
	891	+ igb_ptp_systim_to_hwtstamp(adapter, &ts, le64_to_cpu(regval[1]));
895	892
896	893	/* adjust timestamp for the RX latency based on link speed */
897	894	if (adapter->hw.mac.type == e1000_i210) {
@@ -907,10 +904,10 @@ int igb_ptp_rx_pktstamp(struct igb_q_vector q_vector, void va, · 9 unmodified lines
907	904	break;
908	905	}
909	906	}
910		− skb_hwtstamps(skb)->hwtstamp =
911		− ktime_sub_ns(skb_hwtstamps(skb)->hwtstamp, adjust);
912	907
913		− return 0;
	908	+ *timestamp = ktime_sub_ns(ts.hwtstamp, adjust);
	909	+
	910	+ return IGB_TS_HDR_LEN;
914	911	}
915	912
916	913	/**

drivers/net/ethernet/intel/igc/igc_main.c+5−6

@@ -2047,20 +2047,19 @@ static struct sk_buff igc_xdp_run_prog(struct igc_adapter adapter, · 2046 unmodified lines
2047	2047	break;
2048	2048	case XDP_TX:
2049	2049	if (igc_xdp_xmit_back(adapter, xdp) < 0)
2050		− res = IGC_XDP_CONSUMED;
2051		− else
2052		− res = IGC_XDP_TX;
	2050	+ goto out_failure;
	2051	+ res = IGC_XDP_TX;
2053	2052	break;
2054	2053	case XDP_REDIRECT:
2055	2054	if (xdp_do_redirect(adapter->netdev, xdp, prog) < 0)
2056		− res = IGC_XDP_CONSUMED;
2057		− else
2058		− res = IGC_XDP_REDIRECT;
	2055	+ goto out_failure;
	2056	+ res = IGC_XDP_REDIRECT;
2059	2057	break;
2060	2058	default:
2061	2059	bpf_warn_invalid_xdp_action(act);
2062	2060	fallthrough;
2063	2061	case XDP_ABORTED:
	2062	+out_failure:
2064	2063	trace_xdp_exception(adapter->netdev, prog, act);
2065	2064	fallthrough;
2066	2065	case XDP_DROP:

drivers/net/ethernet/intel/ixgbe/ixgbe_main.c+8−8

@@ -2213,23 +2213,23 @@ static struct sk_buff ixgbe_run_xdp(struct ixgbe_adapter adapter, · 2212 unmodified lines
2213	2213	break;
2214	2214	case XDP_TX:
2215	2215	xdpf = xdp_convert_buff_to_frame(xdp);
2216		− if (unlikely(!xdpf)) {
2217		− result = IXGBE_XDP_CONSUMED;
2218		− break;
2219		− }
	2216	+ if (unlikely(!xdpf))
	2217	+ goto out_failure;
2220	2218	result = ixgbe_xmit_xdp_ring(adapter, xdpf);
	2219	+ if (result == IXGBE_XDP_CONSUMED)
	2220	+ goto out_failure;
2221	2221	break;
2222	2222	case XDP_REDIRECT:
2223	2223	err = xdp_do_redirect(adapter->netdev, xdp, xdp_prog);
2224		− if (!err)
2225		− result = IXGBE_XDP_REDIR;
2226		− else
2227		− result = IXGBE_XDP_CONSUMED;
	2224	+ if (err)
	2225	+ goto out_failure;
	2226	+ result = IXGBE_XDP_REDIR;
2228	2227	break;
2229	2228	default:
2230	2229	bpf_warn_invalid_xdp_action(act);
2231	2230	fallthrough;
2232	2231	case XDP_ABORTED:
	2232	+out_failure:
2233	2233	trace_xdp_exception(rx_ring->netdev, xdp_prog, act);
2234	2234	fallthrough; /* handle aborts by dropping packet */
2235	2235	case XDP_DROP:

drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c+7−9

@@ -467,12 +467,16 @@ static int ixgbe_set_vf_vlan(struct ixgbe_adapter *adapter, int add, int vid, · 466 unmodified lines
467	467	return err;
468	468	}
469	469
470		−static s32 ixgbe_set_vf_lpe(struct ixgbe_adapter adapter, u32 msgbuf, u32 vf)
	470	+static int ixgbe_set_vf_lpe(struct ixgbe_adapter *adapter, u32 max_frame, u32 vf)
471	471	{
472	472	struct ixgbe_hw *hw = &adapter->hw;
473		− int max_frame = msgbuf[1];
474	473	u32 max_frs;
475	474
	475	+ if (max_frame < ETH_MIN_MTU \|\| max_frame > IXGBE_MAX_JUMBO_FRAME_SIZE) {
	476	+ e_err(drv, "VF max_frame %d out of range\n", max_frame);
	477	+ return -EINVAL;
	478	+ }
	479	+
476	480	/*
477	481	* For 82599EB we have to keep all PFs and VFs operating with
478	482	* the same max_frame value in order to avoid sending an oversize
@@ -533,12 +537,6 @@ static s32 ixgbe_set_vf_lpe(struct ixgbe_adapter adapter, u32 msgbuf, u32 vf) · 54 unmodified lines
533	537	}
534	538	}
535	539
536		− /* MTU < 68 is an error and causes problems on some kernels */
537		− if (max_frame > IXGBE_MAX_JUMBO_FRAME_SIZE) {
538		− e_err(drv, "VF max_frame %d out of range\n", max_frame);
539		− return -EINVAL;
540		− }
541		−
542	540	/* pull current max frame size from hardware */
543	541	max_frs = IXGBE_READ_REG(hw, IXGBE_MAXFRS);
544	542	max_frs &= IXGBE_MHADD_MFS_MASK;
@@ -1249,7 +1247,7 @@ static int ixgbe_rcv_msg_from_vf(struct ixgbe_adapter *adapter, u32 vf) · 704 unmodified lines
1249	1247	retval = ixgbe_set_vf_vlan_msg(adapter, msgbuf, vf);
1250	1248	break;
1251	1249	case IXGBE_VF_SET_LPE:
1252		− retval = ixgbe_set_vf_lpe(adapter, msgbuf, vf);
	1250	+ retval = ixgbe_set_vf_lpe(adapter, msgbuf[1], vf);
1253	1251	break;
1254	1252	case IXGBE_VF_SET_MACVLAN:
1255	1253	retval = ixgbe_set_vf_macvlan_msg(adapter, msgbuf, vf);

drivers/net/ethernet/intel/ixgbe/ixgbe_xsk.c+8−6

@@ -106,9 +106,10 @@ static int ixgbe_run_xdp_zc(struct ixgbe_adapter *adapter, · 105 unmodified lines
106	106
107	107	if (likely(act == XDP_REDIRECT)) {
108	108	err = xdp_do_redirect(rx_ring->netdev, xdp, xdp_prog);
109		− result = !err ? IXGBE_XDP_REDIR : IXGBE_XDP_CONSUMED;
	109	+ if (err)
	110	+ goto out_failure;
110	111	rcu_read_unlock();
111		− return result;
	112	+ return IXGBE_XDP_REDIR;
112	113	}
113	114
114	115	switch (act) {
@@ -116,16 +117,17 @@ static int ixgbe_run_xdp_zc(struct ixgbe_adapter *adapter, · 1 unmodified line
116	117	break;
117	118	case XDP_TX:
118	119	xdpf = xdp_convert_buff_to_frame(xdp);
119		− if (unlikely(!xdpf)) {
120		− result = IXGBE_XDP_CONSUMED;
121		− break;
122		− }
	120	+ if (unlikely(!xdpf))
	121	+ goto out_failure;
123	122	result = ixgbe_xmit_xdp_ring(adapter, xdpf);
	123	+ if (result == IXGBE_XDP_CONSUMED)
	124	+ goto out_failure;
124	125	break;
125	126	default:
126	127	bpf_warn_invalid_xdp_action(act);
127	128	fallthrough;
128	129	case XDP_ABORTED:
	130	+out_failure:
129	131	trace_xdp_exception(rx_ring->netdev, xdp_prog, act);
130	132	fallthrough; /* handle aborts by dropping packet */
131	133	case XDP_DROP:

drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c+3−0

@@ -1067,11 +1067,14 @@ static struct sk_buff ixgbevf_run_xdp(struct ixgbevf_adapter adapter, · 1066 unmodified lines
1067	1067	case XDP_TX:
1068	1068	xdp_ring = adapter->xdp_ring[rx_ring->queue_index];
1069	1069	result = ixgbevf_xmit_xdp_ring(xdp_ring, xdp);
	1070	+ if (result == IXGBEVF_XDP_CONSUMED)
	1071	+ goto out_failure;
1070	1072	break;
1071	1073	default:
1072	1074	bpf_warn_invalid_xdp_action(act);
1073	1075	fallthrough;
1074	1076	case XDP_ABORTED:
	1077	+out_failure:
1075	1078	trace_xdp_exception(rx_ring->netdev, xdp_prog, act);
1076	1079	fallthrough; /* handle aborts by dropping packet */
1077	1080	case XDP_DROP:

drivers/net/ethernet/korina.c+6−6

@@ -1315,23 +1315,23 @@ static int korina_probe(struct platform_device *pdev) · 1314 unmodified lines
1315	1315	lp->tx_irq = platform_get_irq_byname(pdev, "tx");
1316	1316
1317	1317	p = devm_platform_ioremap_resource_byname(pdev, "emac");
1318		− if (!p) {
	1318	+ if (IS_ERR(p)) {
1319	1319	printk(KERN_ERR DRV_NAME ": cannot remap registers\n");
1320		− return -ENOMEM;
	1320	+ return PTR_ERR(p);
1321	1321	}
1322	1322	lp->eth_regs = p;
1323	1323
1324	1324	p = devm_platform_ioremap_resource_byname(pdev, "dma_rx");
1325		− if (!p) {
	1325	+ if (IS_ERR(p)) {
1326	1326	printk(KERN_ERR DRV_NAME ": cannot remap Rx DMA registers\n");
1327		− return -ENOMEM;
	1327	+ return PTR_ERR(p);
1328	1328	}
1329	1329	lp->rx_dma_regs = p;
1330	1330
1331	1331	p = devm_platform_ioremap_resource_byname(pdev, "dma_tx");
1332		− if (!p) {
	1332	+ if (IS_ERR(p)) {
1333	1333	printk(KERN_ERR DRV_NAME ": cannot remap Tx DMA registers\n");
1334		− return -ENOMEM;
	1334	+ return PTR_ERR(p);
1335	1335	}
1336	1336	lp->tx_dma_regs = p;
1337	1337

drivers/net/ethernet/lantiq_xrx200.c+11−6

@@ -154,6 +154,8 @@ static int xrx200_close(struct net_device *net_dev) · 153 unmodified lines
154	154
155	155	static int xrx200_alloc_skb(struct xrx200_chan *ch)
156	156	{
	157	+ struct sk_buff *skb = ch->skb[ch->dma.desc];
	158	+ dma_addr_t mapping;
157	159	int ret = 0;
158	160
159	161	ch->skb[ch->dma.desc] = netdev_alloc_skb_ip_align(ch->priv->net_dev,
@@ -163,16 +165,18 @@ static int xrx200_alloc_skb(struct xrx200_chan *ch) · 3 unmodified lines
163	165	goto skip;
164	166	}
165	167
166		− ch->dma.desc_base[ch->dma.desc].addr = dma_map_single(ch->priv->dev,
167		− ch->skb[ch->dma.desc]->data, XRX200_DMA_DATA_LEN,
168		− DMA_FROM_DEVICE);
169		− if (unlikely(dma_mapping_error(ch->priv->dev,
170		− ch->dma.desc_base[ch->dma.desc].addr))) {
	168	+ mapping = dma_map_single(ch->priv->dev, ch->skb[ch->dma.desc]->data,
	169	+ XRX200_DMA_DATA_LEN, DMA_FROM_DEVICE);
	170	+ if (unlikely(dma_mapping_error(ch->priv->dev, mapping))) {
171	171	dev_kfree_skb_any(ch->skb[ch->dma.desc]);
	172	+ ch->skb[ch->dma.desc] = skb;
172	173	ret = -ENOMEM;
173	174	goto skip;
174	175	}
175	176
	177	+ ch->dma.desc_base[ch->dma.desc].addr = mapping;
	178	+ /* Make sure the address is written before we give it to HW */
	179	+ wmb();
176	180	skip:
177	181	ch->dma.desc_base[ch->dma.desc].ctl =
178	182	LTQ_DMA_OWN \| LTQ_DMA_RX_OFFSET(NET_IP_ALIGN) \|
@@ -196,6 +200,7 @@ static int xrx200_hw_receive(struct xrx200_chan *ch) · 17 unmodified lines
196	200	ch->dma.desc %= LTQ_DESC_NUM;
197	201
198	202	if (ret) {
	203	+ net_dev->stats.rx_dropped++;
199	204	netdev_err(net_dev, "failed to allocate new rx buffer\n");
200	205	return ret;
201	206	}
@@ -348,8 +353,8 @@ static irqreturn_t xrx200_dma_irq(int irq, void *ptr) · 146 unmodified lines
348	353	struct xrx200_chan *ch = ptr;
349	354
350	355	if (napi_schedule_prep(&ch->napi)) {
351		− __napi_schedule(&ch->napi);
352	356	ltq_dma_disable_irq(&ch->dma);
	357	+ __napi_schedule(&ch->napi);
353	358	}
354	359
355	360	ltq_dma_ack_irq(&ch->dma);

drivers/net/ethernet/marvell/mvpp2/mvpp2.h+22−0

@@ -993,6 +993,14 @@ enum mvpp22_ptp_packet_format { · 992 unmodified lines
993	993
994	994	#define MVPP2_DESC_DMA_MASK DMA_BIT_MASK(40)
995	995
	996	+/* Buffer header info bits */
	997	+#define MVPP2_B_HDR_INFO_MC_ID_MASK 0xfff
	998	+#define MVPP2_B_HDR_INFO_MC_ID(info) ((info) & MVPP2_B_HDR_INFO_MC_ID_MASK)
	999	+#define MVPP2_B_HDR_INFO_LAST_OFFS 12
	1000	+#define MVPP2_B_HDR_INFO_LAST_MASK BIT(12)
	1001	+#define MVPP2_B_HDR_INFO_IS_LAST(info) \
	1002	+ (((info) & MVPP2_B_HDR_INFO_LAST_MASK) >> MVPP2_B_HDR_INFO_LAST_OFFS)
	1003	+
996	1004	struct mvpp2_tai;
997	1005
998	1006	/* Definitions */
@@ -1002,6 +1010,20 @@ struct mvpp2_rss_table { · 3 unmodified lines
1002	1010	u32 indir[MVPP22_RSS_TABLE_ENTRIES];
1003	1011	};
1004	1012
	1013	+struct mvpp2_buff_hdr {
	1014	+ __le32 next_phys_addr;
	1015	+ __le32 next_dma_addr;
	1016	+ __le16 byte_count;
	1017	+ __le16 info;
	1018	+ __le16 reserved1; /* bm_qset (for future use, BM) */
	1019	+ u8 next_phys_addr_high;
	1020	+ u8 next_dma_addr_high;
	1021	+ __le16 reserved2;
	1022	+ __le16 reserved3;
	1023	+ __le16 reserved4;
	1024	+ __le16 reserved5;
	1025	+};
	1026	+
1005	1027	/* Shared Packet Processor resources */
1006	1028	struct mvpp2 {
1007	1029	/* Shared registers' base addresses */

drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c+45−9

@@ -3839,6 +3839,35 @@ mvpp2_run_xdp(struct mvpp2_port port, struct mvpp2_rx_queue rxq, · 3838 unmodified lines
3839	3839	return ret;
3840	3840	}
3841	3841
	3842	+static void mvpp2_buff_hdr_pool_put(struct mvpp2_port port, struct mvpp2_rx_desc rx_desc,
	3843	+ int pool, u32 rx_status)
	3844	+{
	3845	+ phys_addr_t phys_addr, phys_addr_next;
	3846	+ dma_addr_t dma_addr, dma_addr_next;
	3847	+ struct mvpp2_buff_hdr *buff_hdr;
	3848	+
	3849	+ phys_addr = mvpp2_rxdesc_dma_addr_get(port, rx_desc);
	3850	+ dma_addr = mvpp2_rxdesc_cookie_get(port, rx_desc);
	3851	+
	3852	+ do {
	3853	+ buff_hdr = (struct mvpp2_buff_hdr *)phys_to_virt(phys_addr);
	3854	+
	3855	+ phys_addr_next = le32_to_cpu(buff_hdr->next_phys_addr);
	3856	+ dma_addr_next = le32_to_cpu(buff_hdr->next_dma_addr);
	3857	+
	3858	+ if (port->priv->hw_version >= MVPP22) {
	3859	+ phys_addr_next \|= ((u64)buff_hdr->next_phys_addr_high << 32);
	3860	+ dma_addr_next \|= ((u64)buff_hdr->next_dma_addr_high << 32);
	3861	+ }
	3862	+
	3863	+ mvpp2_bm_pool_put(port, pool, dma_addr, phys_addr);
	3864	+
	3865	+ phys_addr = phys_addr_next;
	3866	+ dma_addr = dma_addr_next;
	3867	+
	3868	+ } while (!MVPP2_B_HDR_INFO_IS_LAST(le16_to_cpu(buff_hdr->info)));
	3869	+}
	3870	+
3842	3871	/* Main rx processing */
3843	3872	static int mvpp2_rx(struct mvpp2_port port, struct napi_struct napi,
3844	3873	int rx_todo, struct mvpp2_rx_queue *rxq)
@@ -3885,14 +3914,6 @@ static int mvpp2_rx(struct mvpp2_port port, struct napi_struct napi, · 40 unmodified lines
3885	3914	MVPP2_RXD_BM_POOL_ID_OFFS;
3886	3915	bm_pool = &port->priv->bm_pools[pool];
3887	3916
3888		− /* In case of an error, release the requested buffer pointer
3889		− * to the Buffer Manager. This request process is controlled
3890		− * by the hardware, and the information about the buffer is
3891		− * comprised by the RX descriptor.
3892		− */
3893		− if (rx_status & MVPP2_RXD_ERR_SUMMARY)
3894		− goto err_drop_frame;
3895		−
3896	3917	if (port->priv->percpu_pools) {
3897	3918	pp = port->priv->page_pool[pool];
3898	3919	dma_dir = page_pool_get_dma_dir(pp);
@@ -3904,6 +3925,18 @@ static int mvpp2_rx(struct mvpp2_port port, struct napi_struct napi, · 5 unmodified lines
3904	3925	rx_bytes + MVPP2_MH_SIZE,
3905	3926	dma_dir);
3906	3927
	3928	+ /* Buffer header not supported */
	3929	+ if (rx_status & MVPP2_RXD_BUF_HDR)
	3930	+ goto err_drop_frame;
	3931	+
	3932	+ /* In case of an error, release the requested buffer pointer
	3933	+ * to the Buffer Manager. This request process is controlled
	3934	+ * by the hardware, and the information about the buffer is
	3935	+ * comprised by the RX descriptor.
	3936	+ */
	3937	+ if (rx_status & MVPP2_RXD_ERR_SUMMARY)
	3938	+ goto err_drop_frame;
	3939	+
3907	3940	/* Prefetch header */
3908	3941	prefetch(data);
3909	3942
@@ -3985,7 +4018,10 @@ err_drop_frame: · 75 unmodified lines
3985	4018	dev->stats.rx_errors++;
3986	4019	mvpp2_rx_error(port, rx_desc);
3987	4020	/* Return the buffer to the pool */
3988		− mvpp2_bm_pool_put(port, pool, dma_addr, phys_addr);
	4021	+ if (rx_status & MVPP2_RXD_BUF_HDR)
	4022	+ mvpp2_buff_hdr_pool_put(port, rx_desc, pool, rx_status);
	4023	+ else
	4024	+ mvpp2_bm_pool_put(port, pool, dma_addr, phys_addr);
3989	4025	}
3990	4026
3991	4027	rcu_read_unlock();

drivers/net/ethernet/marvell/octeontx2/nic/otx2_ethtool.c+4−0

@@ -786,6 +786,10 @@ static int otx2_set_rxfh_context(struct net_device dev, const u32 indir, · 785 unmodified lines
786	786	if (hfunc != ETH_RSS_HASH_NO_CHANGE && hfunc != ETH_RSS_HASH_TOP)
787	787	return -EOPNOTSUPP;
788	788
	789	+ if (*rss_context != ETH_RXFH_CONTEXT_ALLOC &&
	790	+ *rss_context >= MAX_RSS_GROUPS)
	791	+ return -EINVAL;
	792	+
789	793	rss = &pfvf->hw.rss_info;
790	794
791	795	if (!rss->enable) {

drivers/net/ethernet/mediatek/mtk_eth_soc.c+52−25

@@ -681,32 +681,53 @@ static int mtk_set_mac_address(struct net_device dev, void p) · 680 unmodified lines
681	681	void mtk_stats_update_mac(struct mtk_mac *mac)
682	682	{
683	683	struct mtk_hw_stats *hw_stats = mac->hw_stats;
684		− unsigned int base = MTK_GDM1_TX_GBCNT;
685		− u64 stats;
686		−
687		− base += hw_stats->reg_offset;
	684	+ struct mtk_eth *eth = mac->hw;
688	685
689	686	u64_stats_update_begin(&hw_stats->syncp);
690	687
691		− hw_stats->rx_bytes += mtk_r32(mac->hw, base);
692		− stats = mtk_r32(mac->hw, base + 0x04);
693		− if (stats)
694		− hw_stats->rx_bytes += (stats << 32);
695		− hw_stats->rx_packets += mtk_r32(mac->hw, base + 0x08);
696		− hw_stats->rx_overflow += mtk_r32(mac->hw, base + 0x10);
697		− hw_stats->rx_fcs_errors += mtk_r32(mac->hw, base + 0x14);
698		− hw_stats->rx_short_errors += mtk_r32(mac->hw, base + 0x18);
699		− hw_stats->rx_long_errors += mtk_r32(mac->hw, base + 0x1c);
700		− hw_stats->rx_checksum_errors += mtk_r32(mac->hw, base + 0x20);
701		− hw_stats->rx_flow_control_packets +=
702		− mtk_r32(mac->hw, base + 0x24);
703		− hw_stats->tx_skip += mtk_r32(mac->hw, base + 0x28);
704		− hw_stats->tx_collisions += mtk_r32(mac->hw, base + 0x2c);
705		− hw_stats->tx_bytes += mtk_r32(mac->hw, base + 0x30);
706		− stats = mtk_r32(mac->hw, base + 0x34);
707		− if (stats)
708		− hw_stats->tx_bytes += (stats << 32);
709		− hw_stats->tx_packets += mtk_r32(mac->hw, base + 0x38);
	688	+ if (MTK_HAS_CAPS(eth->soc->caps, MTK_SOC_MT7628)) {
	689	+ hw_stats->tx_packets += mtk_r32(mac->hw, MT7628_SDM_TPCNT);
	690	+ hw_stats->tx_bytes += mtk_r32(mac->hw, MT7628_SDM_TBCNT);
	691	+ hw_stats->rx_packets += mtk_r32(mac->hw, MT7628_SDM_RPCNT);
	692	+ hw_stats->rx_bytes += mtk_r32(mac->hw, MT7628_SDM_RBCNT);
	693	+ hw_stats->rx_checksum_errors +=
	694	+ mtk_r32(mac->hw, MT7628_SDM_CS_ERR);
	695	+ } else {
	696	+ unsigned int offs = hw_stats->reg_offset;
	697	+ u64 stats;
	698	+
	699	+ hw_stats->rx_bytes += mtk_r32(mac->hw,
	700	+ MTK_GDM1_RX_GBCNT_L + offs);
	701	+ stats = mtk_r32(mac->hw, MTK_GDM1_RX_GBCNT_H + offs);
	702	+ if (stats)
	703	+ hw_stats->rx_bytes += (stats << 32);
	704	+ hw_stats->rx_packets +=
	705	+ mtk_r32(mac->hw, MTK_GDM1_RX_GPCNT + offs);
	706	+ hw_stats->rx_overflow +=
	707	+ mtk_r32(mac->hw, MTK_GDM1_RX_OERCNT + offs);
	708	+ hw_stats->rx_fcs_errors +=
	709	+ mtk_r32(mac->hw, MTK_GDM1_RX_FERCNT + offs);
	710	+ hw_stats->rx_short_errors +=
	711	+ mtk_r32(mac->hw, MTK_GDM1_RX_SERCNT + offs);
	712	+ hw_stats->rx_long_errors +=
	713	+ mtk_r32(mac->hw, MTK_GDM1_RX_LENCNT + offs);
	714	+ hw_stats->rx_checksum_errors +=
	715	+ mtk_r32(mac->hw, MTK_GDM1_RX_CERCNT + offs);
	716	+ hw_stats->rx_flow_control_packets +=
	717	+ mtk_r32(mac->hw, MTK_GDM1_RX_FCCNT + offs);
	718	+ hw_stats->tx_skip +=
	719	+ mtk_r32(mac->hw, MTK_GDM1_TX_SKIPCNT + offs);
	720	+ hw_stats->tx_collisions +=
	721	+ mtk_r32(mac->hw, MTK_GDM1_TX_COLCNT + offs);
	722	+ hw_stats->tx_bytes +=
	723	+ mtk_r32(mac->hw, MTK_GDM1_TX_GBCNT_L + offs);
	724	+ stats = mtk_r32(mac->hw, MTK_GDM1_TX_GBCNT_H + offs);
	725	+ if (stats)
	726	+ hw_stats->tx_bytes += (stats << 32);
	727	+ hw_stats->tx_packets +=
	728	+ mtk_r32(mac->hw, MTK_GDM1_TX_GPCNT + offs);
	729	+ }
	730	+
710	731	u64_stats_update_end(&hw_stats->syncp);
711	732	}
712	733
@@ -2423,7 +2444,8 @@ static void mtk_dim_rx(struct work_struct *work) · 1710 unmodified lines
2423	2444	val \|= cur << MTK_PDMA_DELAY_RX_PINT_SHIFT;
2424	2445
2425	2446	mtk_w32(eth, val, MTK_PDMA_DELAY_INT);
2426		− mtk_w32(eth, val, MTK_QDMA_DELAY_INT);
	2447	+ if (MTK_HAS_CAPS(eth->soc->caps, MTK_QDMA))
	2448	+ mtk_w32(eth, val, MTK_QDMA_DELAY_INT);
2427	2449
2428	2450	spin_unlock_bh(&eth->dim_lock);
2429	2451
@@ -2452,7 +2474,8 @@ static void mtk_dim_tx(struct work_struct *work) · 22 unmodified lines
2452	2474	val \|= cur << MTK_PDMA_DELAY_TX_PINT_SHIFT;
2453	2475
2454	2476	mtk_w32(eth, val, MTK_PDMA_DELAY_INT);
2455		− mtk_w32(eth, val, MTK_QDMA_DELAY_INT);
	2477	+ if (MTK_HAS_CAPS(eth->soc->caps, MTK_QDMA))
	2478	+ mtk_w32(eth, val, MTK_QDMA_DELAY_INT);
2456	2479
2457	2480	spin_unlock_bh(&eth->dim_lock);
2458	2481
@@ -2480,6 +2503,10 @@ static int mtk_hw_init(struct mtk_eth *eth) · 21 unmodified lines
2480	2503	goto err_disable_pm;
2481	2504	}
2482	2505
	2506	+ /* set interrupt delays based on current Net DIM sample */
	2507	+ mtk_dim_rx(&eth->rx_dim.work);
	2508	+ mtk_dim_tx(&eth->tx_dim.work);
	2509	+
2483	2510	/* disable delay and normal interrupt */
2484	2511	mtk_tx_irq_disable(eth, ~0);
2485	2512	mtk_rx_irq_disable(eth, ~0);

drivers/net/ethernet/mediatek/mtk_eth_soc.h+22−2

@@ -278,8 +278,21 @@ · 277 unmodified lines
278	278	/* QDMA FQ Free Page Buffer Length Register */
279	279	#define MTK_QDMA_FQ_BLEN 0x1B2C
280	280
281		−/* GMA1 Received Good Byte Count Register */
282		−#define MTK_GDM1_TX_GBCNT 0x2400
	281	+/* GMA1 counter / statics register */
	282	+#define MTK_GDM1_RX_GBCNT_L 0x2400
	283	+#define MTK_GDM1_RX_GBCNT_H 0x2404
	284	+#define MTK_GDM1_RX_GPCNT 0x2408
	285	+#define MTK_GDM1_RX_OERCNT 0x2410
	286	+#define MTK_GDM1_RX_FERCNT 0x2414
	287	+#define MTK_GDM1_RX_SERCNT 0x2418
	288	+#define MTK_GDM1_RX_LENCNT 0x241c
	289	+#define MTK_GDM1_RX_CERCNT 0x2420
	290	+#define MTK_GDM1_RX_FCCNT 0x2424
	291	+#define MTK_GDM1_TX_SKIPCNT 0x2428
	292	+#define MTK_GDM1_TX_COLCNT 0x242c
	293	+#define MTK_GDM1_TX_GBCNT_L 0x2430
	294	+#define MTK_GDM1_TX_GBCNT_H 0x2434
	295	+#define MTK_GDM1_TX_GPCNT 0x2438
283	296	#define MTK_STAT_OFFSET 0x40
284	297
285	298	/* QDMA descriptor txd4 */
@@ -502,6 +515,13 @@ · 216 unmodified lines
502	515	#define MT7628_SDM_MAC_ADRL (MT7628_SDM_OFFSET + 0x0c)
503	516	#define MT7628_SDM_MAC_ADRH (MT7628_SDM_OFFSET + 0x10)
504	517
	518	+/* Counter / stat register */
	519	+#define MT7628_SDM_TPCNT (MT7628_SDM_OFFSET + 0x100)
	520	+#define MT7628_SDM_TBCNT (MT7628_SDM_OFFSET + 0x104)
	521	+#define MT7628_SDM_RPCNT (MT7628_SDM_OFFSET + 0x108)
	522	+#define MT7628_SDM_RBCNT (MT7628_SDM_OFFSET + 0x10c)
	523	+#define MT7628_SDM_CS_ERR (MT7628_SDM_OFFSET + 0x110)
	524	+
505	525	struct mtk_rx_dma {
506	526	unsigned int rxd1;
507	527	unsigned int rxd2;

drivers/net/ethernet/mellanox/mlx4/en_ethtool.c+1−3

@@ -2027,8 +2027,6 @@ static int mlx4_en_set_tunable(struct net_device *dev, · 2026 unmodified lines
2027	2027	return ret;
2028	2028	}
2029	2029
2030		−#define MLX4_EEPROM_PAGE_LEN 256
2031		−
2032	2030	static int mlx4_en_get_module_info(struct net_device *dev,
2033	2031	struct ethtool_modinfo *modinfo)
2034	2032	{
@@ -2063,7 +2061,7 @@ static int mlx4_en_get_module_info(struct net_device *dev, · 28 unmodified lines
2063	2061	break;
2064	2062	case MLX4_MODULE_ID_SFP:
2065	2063	modinfo->type = ETH_MODULE_SFF_8472;
2066		− modinfo->eeprom_len = MLX4_EEPROM_PAGE_LEN;
	2064	+ modinfo->eeprom_len = ETH_MODULE_SFF_8472_LEN;
2067	2065	break;
2068	2066	default:
2069	2067	return -EINVAL;

drivers/net/ethernet/mellanox/mlx4/fw.c+3−0

@@ -823,6 +823,7 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev dev, struct mlx4_dev_cap dev_cap) · 822 unmodified lines
823	823	#define QUERY_DEV_CAP_MAD_DEMUX_OFFSET 0xb0
824	824	#define QUERY_DEV_CAP_DMFS_HIGH_RATE_QPN_BASE_OFFSET 0xa8
825	825	#define QUERY_DEV_CAP_DMFS_HIGH_RATE_QPN_RANGE_OFFSET 0xac
	826	+#define QUERY_DEV_CAP_MAP_CLOCK_TO_USER 0xc1
826	827	#define QUERY_DEV_CAP_QP_RATE_LIMIT_NUM_OFFSET 0xcc
827	828	#define QUERY_DEV_CAP_QP_RATE_LIMIT_MAX_OFFSET 0xd0
828	829	#define QUERY_DEV_CAP_QP_RATE_LIMIT_MIN_OFFSET 0xd2
@@ -841,6 +842,8 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev dev, struct mlx4_dev_cap dev_cap) · 12 unmodified lines
841	842
842	843	if (mlx4_is_mfunc(dev))
843	844	disable_unsupported_roce_caps(outbox);
	845	+ MLX4_GET(field, outbox, QUERY_DEV_CAP_MAP_CLOCK_TO_USER);
	846	+ dev_cap->map_clock_to_user = field & 0x80;
844	847	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_QP_OFFSET);
845	848	dev_cap->reserved_qps = 1 << (field & 0xf);
846	849	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_QP_OFFSET);

drivers/net/ethernet/mellanox/mlx4/fw.h+1−0

@@ -131,6 +131,7 @@ struct mlx4_dev_cap { · 130 unmodified lines
131	131	u32 health_buffer_addrs;
132	132	struct mlx4_port_cap port_cap[MLX4_MAX_PORTS + 1];
133	133	bool wol_port[MLX4_MAX_PORTS + 1];
	134	+ bool map_clock_to_user;
134	135	};
135	136
136	137	struct mlx4_func_cap {

drivers/net/ethernet/mellanox/mlx4/main.c+6−0

@@ -498,6 +498,7 @@ static int mlx4_dev_cap(struct mlx4_dev dev, struct mlx4_dev_cap dev_cap) · 497 unmodified lines
498	498	}
499	499	}
500	500
	501	+ dev->caps.map_clock_to_user = dev_cap->map_clock_to_user;
501	502	dev->caps.uar_page_size = PAGE_SIZE;
502	503	dev->caps.num_uars = dev_cap->uar_size / PAGE_SIZE;
503	504	dev->caps.local_ca_ack_delay = dev_cap->local_ca_ack_delay;
@@ -1948,6 +1949,11 @@ int mlx4_get_internal_clock_params(struct mlx4_dev *dev, · 1444 unmodified lines
1948	1949	if (mlx4_is_slave(dev))
1949	1950	return -EOPNOTSUPP;
1950	1951
	1952	+ if (!dev->caps.map_clock_to_user) {
	1953	+ mlx4_dbg(dev, "Map clock to user is not supported.\n");
	1954	+ return -EOPNOTSUPP;
	1955	+ }
	1956	+
1951	1957	if (!params)
1952	1958	return -EINVAL;
1953	1959

drivers/net/ethernet/mellanox/mlx4/port.c+103−4

@@ -1973,6 +1973,7 @@ EXPORT_SYMBOL(mlx4_get_roce_gid_from_slave); · 1972 unmodified lines
1973	1973	#define I2C_ADDR_LOW 0x50
1974	1974	#define I2C_ADDR_HIGH 0x51
1975	1975	#define I2C_PAGE_SIZE 256
	1976	+#define I2C_HIGH_PAGE_SIZE 128
1976	1977
1977	1978	/* Module Info Data */
1978	1979	struct mlx4_cable_info {
@@ -2026,6 +2027,88 @@ static inline const char *cable_info_mad_err_str(u16 mad_status) · 47 unmodified lines
2026	2027	return "Unknown Error";
2027	2028	}
2028	2029
	2030	+static int mlx4_get_module_id(struct mlx4_dev dev, u8 port, u8 module_id)
	2031	+{
	2032	+ struct mlx4_cmd_mailbox inbox, outbox;
	2033	+ struct mlx4_mad_ifc inmad, outmad;
	2034	+ struct mlx4_cable_info *cable_info;
	2035	+ int ret;
	2036	+
	2037	+ inbox = mlx4_alloc_cmd_mailbox(dev);
	2038	+ if (IS_ERR(inbox))
	2039	+ return PTR_ERR(inbox);
	2040	+
	2041	+ outbox = mlx4_alloc_cmd_mailbox(dev);
	2042	+ if (IS_ERR(outbox)) {
	2043	+ mlx4_free_cmd_mailbox(dev, inbox);
	2044	+ return PTR_ERR(outbox);
	2045	+ }
	2046	+
	2047	+ inmad = (struct mlx4_mad_ifc *)(inbox->buf);
	2048	+ outmad = (struct mlx4_mad_ifc *)(outbox->buf);
	2049	+
	2050	+ inmad->method = 0x1; /* Get */
	2051	+ inmad->class_version = 0x1;
	2052	+ inmad->mgmt_class = 0x1;
	2053	+ inmad->base_version = 0x1;
	2054	+ inmad->attr_id = cpu_to_be16(0xFF60); /* Module Info */
	2055	+
	2056	+ cable_info = (struct mlx4_cable_info *)inmad->data;
	2057	+ cable_info->dev_mem_address = 0;
	2058	+ cable_info->page_num = 0;
	2059	+ cable_info->i2c_addr = I2C_ADDR_LOW;
	2060	+ cable_info->size = cpu_to_be16(1);
	2061	+
	2062	+ ret = mlx4_cmd_box(dev, inbox->dma, outbox->dma, port, 3,
	2063	+ MLX4_CMD_MAD_IFC, MLX4_CMD_TIME_CLASS_C,
	2064	+ MLX4_CMD_NATIVE);
	2065	+ if (ret)
	2066	+ goto out;
	2067	+
	2068	+ if (be16_to_cpu(outmad->status)) {
	2069	+ /* Mad returned with bad status */
	2070	+ ret = be16_to_cpu(outmad->status);
	2071	+ mlx4_warn(dev,
	2072	+ "MLX4_CMD_MAD_IFC Get Module ID attr(%x) port(%d) i2c_addr(%x) offset(%d) size(%d): Response Mad Status(%x) - %s\n",
	2073	+ 0xFF60, port, I2C_ADDR_LOW, 0, 1, ret,
	2074	+ cable_info_mad_err_str(ret));
	2075	+ ret = -ret;
	2076	+ goto out;
	2077	+ }
	2078	+ cable_info = (struct mlx4_cable_info *)outmad->data;
	2079	+ *module_id = cable_info->data[0];
	2080	+out:
	2081	+ mlx4_free_cmd_mailbox(dev, inbox);
	2082	+ mlx4_free_cmd_mailbox(dev, outbox);
	2083	+ return ret;
	2084	+}
	2085	+
	2086	+static void mlx4_sfp_eeprom_params_set(u8 i2c_addr, u8 page_num, u16 *offset)
	2087	+{
	2088	+ *i2c_addr = I2C_ADDR_LOW;
	2089	+ *page_num = 0;
	2090	+
	2091	+ if (*offset < I2C_PAGE_SIZE)
	2092	+ return;
	2093	+
	2094	+ *i2c_addr = I2C_ADDR_HIGH;
	2095	+ *offset -= I2C_PAGE_SIZE;
	2096	+}
	2097	+
	2098	+static void mlx4_qsfp_eeprom_params_set(u8 i2c_addr, u8 page_num, u16 *offset)
	2099	+{
	2100	+ /* Offsets 0-255 belong to page 0.
	2101	+ * Offsets 256-639 belong to pages 01, 02, 03.
	2102	+ * For example, offset 400 is page 02: 1 + (400 - 256) / 128 = 2
	2103	+ */
	2104	+ if (*offset < I2C_PAGE_SIZE)
	2105	+ *page_num = 0;
	2106	+ else
	2107	+ page_num = 1 + (offset - I2C_PAGE_SIZE) / I2C_HIGH_PAGE_SIZE;
	2108	+ *i2c_addr = I2C_ADDR_LOW;
	2109	+ offset -= page_num * I2C_HIGH_PAGE_SIZE;
	2110	+}
	2111	+
2029	2112	/**
2030	2113	* mlx4_get_module_info - Read cable module eeprom data
2031	2114	* @dev: mlx4_dev.
@@ -2045,12 +2128,30 @@ int mlx4_get_module_info(struct mlx4_dev *dev, u8 port, · 13 unmodified lines
2045	2128	struct mlx4_cmd_mailbox inbox, outbox;
2046	2129	struct mlx4_mad_ifc inmad, outmad;
2047	2130	struct mlx4_cable_info *cable_info;
2048		− u16 i2c_addr;
	2131	+ u8 module_id, i2c_addr, page_num;
2049	2132	int ret;
2050	2133
2051	2134	if (size > MODULE_INFO_MAX_READ)
2052	2135	size = MODULE_INFO_MAX_READ;
2053	2136
	2137	+ ret = mlx4_get_module_id(dev, port, &module_id);
	2138	+ if (ret)
	2139	+ return ret;
	2140	+
	2141	+ switch (module_id) {
	2142	+ case MLX4_MODULE_ID_SFP:
	2143	+ mlx4_sfp_eeprom_params_set(&i2c_addr, &page_num, &offset);
	2144	+ break;
	2145	+ case MLX4_MODULE_ID_QSFP:
	2146	+ case MLX4_MODULE_ID_QSFP_PLUS:
	2147	+ case MLX4_MODULE_ID_QSFP28:
	2148	+ mlx4_qsfp_eeprom_params_set(&i2c_addr, &page_num, &offset);
	2149	+ break;
	2150	+ default:
	2151	+ mlx4_err(dev, "Module ID not recognized: %#x\n", module_id);
	2152	+ return -EINVAL;
	2153	+ }
	2154	+
2054	2155	inbox = mlx4_alloc_cmd_mailbox(dev);
2055	2156	if (IS_ERR(inbox))
2056	2157	return PTR_ERR(inbox);
@@ -2076,11 +2177,9 @@ int mlx4_get_module_info(struct mlx4_dev *dev, u8 port, · 19 unmodified lines
2076	2177	*/
2077	2178	size -= offset + size - I2C_PAGE_SIZE;
2078	2179
2079		− i2c_addr = I2C_ADDR_LOW;
2080		−
2081	2180	cable_info = (struct mlx4_cable_info *)inmad->data;
2082	2181	cable_info->dev_mem_address = cpu_to_be16(offset);
2083		− cable_info->page_num = 0;
	2182	+ cable_info->page_num = page_num;
2084	2183	cable_info->i2c_addr = i2c_addr;
2085	2184	cable_info->size = cpu_to_be16(size);
2086	2185

drivers/net/ethernet/mellanox/mlx5/core/dev.c+19−0

@@ -303,6 +303,7 @@ int mlx5_attach_device(struct mlx5_core_dev *dev) · 302 unmodified lines
303	303	int ret = 0, i;
304	304
305	305	mutex_lock(&mlx5_intf_mutex);
	306	+ priv->flags &= ~MLX5_PRIV_FLAGS_DETACH;
306	307	for (i = 0; i < ARRAY_SIZE(mlx5_adev_devices); i++) {
307	308	if (!priv->adev[i]) {
308	309	bool is_supported = false;
@@ -320,6 +321,16 @@ int mlx5_attach_device(struct mlx5_core_dev *dev) · 11 unmodified lines
320	321	}
321	322	} else {
322	323	adev = &priv->adev[i]->adev;
	324	+
	325	+ /* Pay attention that this is not PCI driver that
	326	+ * mlx5_core_dev is connected, but auxiliary driver.
	327	+ *
	328	+ * Here we can race of module unload with devlink
	329	+ * reload, but we don't need to take extra lock because
	330	+ * we are holding global mlx5_intf_mutex.
	331	+ */
	332	+ if (!adev->dev.driver)
	333	+ continue;
323	334	adrv = to_auxiliary_drv(adev->dev.driver);
324	335
325	336	if (adrv->resume)
@@ -350,6 +361,10 @@ void mlx5_detach_device(struct mlx5_core_dev *dev) · 24 unmodified lines
350	361	continue;
351	362
352	363	adev = &priv->adev[i]->adev;
	364	+ /* Auxiliary driver was unbind manually through sysfs */
	365	+ if (!adev->dev.driver)
	366	+ goto skip_suspend;
	367	+
353	368	adrv = to_auxiliary_drv(adev->dev.driver);
354	369
355	370	if (adrv->suspend) {
@@ -357,9 +372,11 @@ void mlx5_detach_device(struct mlx5_core_dev *dev) · 1 unmodified line
357	372	continue;
358	373	}
359	374
	375	+skip_suspend:
360	376	del_adev(&priv->adev[i]->adev);
361	377	priv->adev[i] = NULL;
362	378	}
	379	+ priv->flags \|= MLX5_PRIV_FLAGS_DETACH;
363	380	mutex_unlock(&mlx5_intf_mutex);
364	381	}
365	382
@@ -448,6 +465,8 @@ int mlx5_rescan_drivers_locked(struct mlx5_core_dev *dev) · 82 unmodified lines
448	465	struct mlx5_priv *priv = &dev->priv;
449	466
450	467	lockdep_assert_held(&mlx5_intf_mutex);
	468	+ if (priv->flags & MLX5_PRIV_FLAGS_DETACH)
	469	+ return 0;
451	470
452	471	delete_drivers(dev);
453	472	if (priv->flags & MLX5_PRIV_FLAGS_DISABLE_ALL_ADEV)

drivers/net/ethernet/mellanox/mlx5/core/en/devlink.c+2−0

@@ -64,6 +64,8 @@ struct devlink_port mlx5e_get_devlink_port(struct net_device dev) · 63 unmodified lines
64	64	struct mlx5e_priv *priv = netdev_priv(dev);
65	65	struct devlink_port *port;
66	66
	67	+ if (!netif_device_present(dev))
	68	+ return NULL;
67	69	port = mlx5e_devlink_get_dl_port(priv);
68	70	if (port->registered)
69	71	return port;

drivers/net/ethernet/mellanox/mlx5/core/en/ptp.c+0−1

@@ -1,7 +1,6 @@
1	1	// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB
2	2	// Copyright (c) 2020 Mellanox Technologies
3	3
4		−#include <linux/ptp_classify.h>
5	4	#include "en/ptp.h"
6	5	#include "en/txrx.h"
7	6	#include "en/params.h"

drivers/net/ethernet/mellanox/mlx5/core/en/ptp.h+22−0

@@ -6,6 +6,7 @@ · 5 unmodified lines
6	6
7	7	#include "en.h"
8	8	#include "en_stats.h"
	9	+#include <linux/ptp_classify.h>
9	10
10	11	struct mlx5e_ptpsq {
11	12	struct mlx5e_txqsq txqsq;
@@ -43,6 +44,27 @@ struct mlx5e_ptp { · 31 unmodified lines
43	44	DECLARE_BITMAP(state, MLX5E_PTP_STATE_NUM_STATES);
44	45	};
45	46
	47	+static inline bool mlx5e_use_ptpsq(struct sk_buff *skb)
	48	+{
	49	+ struct flow_keys fk;
	50	+
	51	+ if (!(skb_shinfo(skb)->tx_flags & SKBTX_HW_TSTAMP))
	52	+ return false;
	53	+
	54	+ if (!skb_flow_dissect_flow_keys(skb, &fk, 0))
	55	+ return false;
	56	+
	57	+ if (fk.basic.n_proto == htons(ETH_P_1588))
	58	+ return true;
	59	+
	60	+ if (fk.basic.n_proto != htons(ETH_P_IP) &&
	61	+ fk.basic.n_proto != htons(ETH_P_IPV6))
	62	+ return false;
	63	+
	64	+ return (fk.basic.ip_proto == IPPROTO_UDP &&
	65	+ fk.ports.dst == htons(PTP_EV_PORT));
	66	+}
	67	+
46	68	int mlx5e_ptp_open(struct mlx5e_priv priv, struct mlx5e_params params,
47	69	u8 lag_port, struct mlx5e_ptp **cp);
48	70	void mlx5e_ptp_close(struct mlx5e_ptp *c);

drivers/net/ethernet/mellanox/mlx5/core/en/rep/bond.c+2−0

@@ -223,6 +223,8 @@ static void mlx5e_rep_changelowerstate_event(struct net_device netdev, void pt · 222 unmodified lines
223	223	rpriv = priv->ppriv;
224	224	fwd_vport_num = rpriv->rep->vport;
225	225	lag_dev = netdev_master_upper_dev_get(netdev);
	226	+ if (!lag_dev)
	227	+ return;
226	228
227	229	netdev_dbg(netdev, "lag_dev(%s)'s slave vport(%d) is txable(%d)\n",
228	230	lag_dev->name, fwd_vport_num, net_lag_port_dev_txable(netdev));

drivers/net/ethernet/mellanox/mlx5/core/en/rep/neigh.c+6−9

@@ -129,10 +129,9 @@ static void mlx5e_rep_neigh_update(struct work_struct *work) · 128 unmodified lines
129	129	work);
130	130	struct mlx5e_neigh_hash_entry *nhe = update_work->nhe;
131	131	struct neighbour *n = update_work->n;
	132	+ struct mlx5e_encap_entry *e = NULL;
132	133	bool neigh_connected, same_dev;
133		− struct mlx5e_encap_entry *e;
134	134	unsigned char ha[ETH_ALEN];
135		− struct mlx5e_priv *priv;
136	135	u8 nud_state, dead;
137	136
138	137	rtnl_lock();
@@ -156,14 +155,12 @@ static void mlx5e_rep_neigh_update(struct work_struct *work) · 17 unmodified lines
156	155	if (!same_dev)
157	156	goto out;
158	157
159		− list_for_each_entry(e, &nhe->encap_list, encap_list) {
160		− if (!mlx5e_encap_take(e))
161		− continue;
	158	+ /* mlx5e_get_next_init_encap() releases previous encap before returning
	159	+ * the next one.
	160	+ */
	161	+ while ((e = mlx5e_get_next_init_encap(nhe, e)) != NULL)
	162	+ mlx5e_rep_update_flows(netdev_priv(e->out_dev), e, neigh_connected, ha);
162	163
163		− priv = netdev_priv(e->out_dev);
164		− mlx5e_rep_update_flows(priv, e, neigh_connected, ha);
165		− mlx5e_encap_put(priv, e);
166		− }
167	164	out:
168	165	rtnl_unlock();
169	166	mlx5e_release_neigh_update_work(update_work);

drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.c+2−6

@@ -94,13 +94,9 @@ void mlx5e_rep_update_flows(struct mlx5e_priv *priv, · 93 unmodified lines
94	94
95	95	ASSERT_RTNL();
96	96
97		− /* wait for encap to be fully initialized */
98		− wait_for_completion(&e->res_ready);
99		−
100	97	mutex_lock(&esw->offloads.encap_tbl_lock);
101	98	encap_connected = !!(e->flags & MLX5_ENCAP_ENTRY_VALID);
102		− if (e->compl_result < 0 \|\| (encap_connected == neigh_connected &&
103		− ether_addr_equal(e->h_dest, ha)))
	99	+ if (encap_connected == neigh_connected && ether_addr_equal(e->h_dest, ha))
104	100	goto unlock;
105	101
106	102	mlx5e_take_all_encap_flows(e, &flow_list);
@@ -626,7 +622,7 @@ static bool mlx5e_restore_skb(struct sk_buff *skb, u32 chain, u32 reg_c1, · 519 unmodified lines
626	622	struct mlx5_eswitch *esw;
627	623	u32 zone_restore_id;
628	624
629		− tc_skb_ext = skb_ext_add(skb, TC_SKB_EXT);
	625	+ tc_skb_ext = tc_skb_ext_alloc(skb);
630	626	if (!tc_skb_ext) {
631	627	WARN_ON(1);
632	628	return false;

drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c+31−4

@@ -251,9 +251,12 @@ static void mlx5e_take_all_route_decap_flows(struct mlx5e_route_entry *r, · 250 unmodified lines
251	251	mlx5e_take_tmp_flow(flow, flow_list, 0);
252	252	}
253	253
	254	+typedef bool (match_cb)(struct mlx5e_encap_entry *);
	255	+
254	256	static struct mlx5e_encap_entry *
255		−mlx5e_get_next_valid_encap(struct mlx5e_neigh_hash_entry *nhe,
256		− struct mlx5e_encap_entry *e)
	257	+mlx5e_get_next_matching_encap(struct mlx5e_neigh_hash_entry *nhe,
	258	+ struct mlx5e_encap_entry *e,
	259	+ match_cb match)
257	260	{
258	261	struct mlx5e_encap_entry *next = NULL;
259	262
@@ -288,7 +291,7 @@ retry: · 28 unmodified lines
288	291	/* wait for encap to be fully initialized */
289	292	wait_for_completion(&next->res_ready);
290	293	/* continue searching if encap entry is not in valid state after completion */
291		− if (!(next->flags & MLX5_ENCAP_ENTRY_VALID)) {
	294	+ if (!match(next)) {
292	295	e = next;
293	296	goto retry;
294	297	}
@@ -296,6 +299,30 @@ retry: · 1 unmodified line
296	299	return next;
297	300	}
298	301
	302	+static bool mlx5e_encap_valid(struct mlx5e_encap_entry *e)
	303	+{
	304	+ return e->flags & MLX5_ENCAP_ENTRY_VALID;
	305	+}
	306	+
	307	+static struct mlx5e_encap_entry *
	308	+mlx5e_get_next_valid_encap(struct mlx5e_neigh_hash_entry *nhe,
	309	+ struct mlx5e_encap_entry *e)
	310	+{
	311	+ return mlx5e_get_next_matching_encap(nhe, e, mlx5e_encap_valid);
	312	+}
	313	+
	314	+static bool mlx5e_encap_initialized(struct mlx5e_encap_entry *e)
	315	+{
	316	+ return e->compl_result >= 0;
	317	+}
	318	+
	319	+struct mlx5e_encap_entry *
	320	+mlx5e_get_next_init_encap(struct mlx5e_neigh_hash_entry *nhe,
	321	+ struct mlx5e_encap_entry *e)
	322	+{
	323	+ return mlx5e_get_next_matching_encap(nhe, e, mlx5e_encap_initialized);
	324	+}
	325	+
299	326	void mlx5e_tc_update_neigh_used_value(struct mlx5e_neigh_hash_entry *nhe)
300	327	{
301	328	struct mlx5e_neigh *m_neigh = &nhe->m_neigh;
@@ -1505,7 +1532,7 @@ mlx5e_init_fib_work_ipv4(struct mlx5e_priv *priv, · 1203 unmodified lines
1505	1532
1506	1533	fen_info = container_of(info, struct fib_entry_notifier_info, info);
1507	1534	fib_dev = fib_info_nh(fen_info->fi, 0)->fib_nh_dev;
1508		− if (fib_dev->netdev_ops != &mlx5e_netdev_ops \|\|
	1535	+ if (!fib_dev \|\| fib_dev->netdev_ops != &mlx5e_netdev_ops \|\|
1509	1536	fen_info->dst_len != 32)
1510	1537	return NULL;
1511	1538

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c+0−3

@@ -532,9 +532,6 @@ void mlx5e_ipsec_build_netdev(struct mlx5e_priv *priv) · 531 unmodified lines
532	532	struct mlx5_core_dev *mdev = priv->mdev;
533	533	struct net_device *netdev = priv->netdev;
534	534
535		− if (!priv->ipsec)
536		− return;
537		−
538	535	if (!(mlx5_accel_ipsec_device_caps(mdev) & MLX5_ACCEL_IPSEC_CAP_ESP) \|\|
539	536	!MLX5_CAP_ETH(mdev, swp)) {
540	537	mlx5_core_dbg(mdev, "mlx5e: ESP and SWP offload not supported\n");

drivers/net/ethernet/mellanox/mlx5/core/en_arfs.c+1−1

@@ -356,7 +356,7 @@ err: · 355 unmodified lines
356	356
357	357	int mlx5e_arfs_create_tables(struct mlx5e_priv *priv)
358	358	{
359		− int err = 0;
	359	+ int err = -ENOMEM;
360	360	int i;
361	361
362	362	if (!(priv->netdev->hw_features & NETIF_F_NTUPLE))

drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c+10−2

@@ -1624,12 +1624,13 @@ static int mlx5e_set_fecparam(struct net_device *netdev, · 1623 unmodified lines
1624	1624	{
1625	1625	struct mlx5e_priv *priv = netdev_priv(netdev);
1626	1626	struct mlx5_core_dev *mdev = priv->mdev;
	1627	+ unsigned long fec_bitmap;
1627	1628	u16 fec_policy = 0;
1628	1629	int mode;
1629	1630	int err;
1630	1631
1631		− if (bitmap_weight((unsigned long *)&fecparam->fec,
1632		− ETHTOOL_FEC_LLRS_BIT + 1) > 1)
	1632	+ bitmap_from_arr32(&fec_bitmap, &fecparam->fec, sizeof(fecparam->fec) * BITS_PER_BYTE);
	1633	+ if (bitmap_weight(&fec_bitmap, ETHTOOL_FEC_LLRS_BIT + 1) > 1)
1633	1634	return -EOPNOTSUPP;
1634	1635
1635	1636	for (mode = 0; mode < ARRAY_SIZE(pplm_fec_2_ethtool); mode++) {
@@ -1893,6 +1894,13 @@ int mlx5e_modify_rx_cqe_compression_locked(struct mlx5e_priv *priv, bool new_val · 257 unmodified lines
1893	1894	if (curr_val == new_val)
1894	1895	return 0;
1895	1896
	1897	+ if (new_val && !priv->profile->rx_ptp_support &&
	1898	+ priv->tstamp.rx_filter != HWTSTAMP_FILTER_NONE) {
	1899	+ netdev_err(priv->netdev,
	1900	+ "Profile doesn't support enabling of CQE compression while hardware time-stamping is enabled.\n");
	1901	+ return -EINVAL;
	1902	+ }
	1903	+
1896	1904	new_params = priv->channels.params;
1897	1905	MLX5E_SET_PFLAG(&new_params, MLX5E_PFLAG_RX_CQE_COMPRESS, new_val);
1898	1906	if (priv->tstamp.rx_filter != HWTSTAMP_FILTER_NONE)

drivers/net/ethernet/mellanox/mlx5/core/en_fs.c+1−0

@@ -35,6 +35,7 @@ · 34 unmodified lines
35	35	#include <linux/ipv6.h>
36	36	#include <linux/tcp.h>
37	37	#include <linux/mlx5/fs.h>
	38	+#include <linux/mlx5/mpfs.h>
38	39	#include "en.h"
39	40	#include "en_rep.h"
40	41	#include "lib/mpfs.h"

drivers/net/ethernet/mellanox/mlx5/core/en_main.c+81−33

@@ -889,10 +889,13 @@ err_free_rq: · 888 unmodified lines
889	889	void mlx5e_activate_rq(struct mlx5e_rq *rq)
890	890	{
891	891	set_bit(MLX5E_RQ_STATE_ENABLED, &rq->state);
892		− if (rq->icosq)
	892	+ if (rq->icosq) {
893	893	mlx5e_trigger_irq(rq->icosq);
894		− else
	894	+ } else {
	895	+ local_bh_disable();
895	896	napi_schedule(rq->cq.napi);
	897	+ local_bh_enable();
	898	+ }
896	899	}
897	900
898	901	void mlx5e_deactivate_rq(struct mlx5e_rq *rq)
@@ -2697,13 +2700,11 @@ static int mlx5e_update_netdev_queues(struct mlx5e_priv *priv) · 1798 unmodified lines
2697	2700	int err;
2698	2701
2699	2702	old_num_txqs = netdev->real_num_tx_queues;
2700		− old_ntc = netdev->num_tc;
	2703	+ old_ntc = netdev->num_tc ? : 1;
2701	2704
2702	2705	nch = priv->channels.params.num_channels;
2703	2706	ntc = priv->channels.params.num_tc;
2704	2707	num_rxqs = nch * priv->profile->rq_groups;
2705		− if (priv->channels.params.ptp_rx)
2706		− num_rxqs++;
2707	2708
2708	2709	mlx5e_netdev_set_tcs(netdev, nch, ntc);
2709	2710
@@ -3855,6 +3856,16 @@ static netdev_features_t mlx5e_fix_features(struct net_device *netdev, · 1145 unmodified lines
3855	3856	netdev_warn(netdev, "Disabling rxhash, not supported when CQE compress is active\n");
3856	3857	}
3857	3858
	3859	+ if (mlx5e_is_uplink_rep(priv)) {
	3860	+ features &= ~NETIF_F_HW_TLS_RX;
	3861	+ if (netdev->features & NETIF_F_HW_TLS_RX)
	3862	+ netdev_warn(netdev, "Disabling hw_tls_rx, not supported in switchdev mode\n");
	3863	+
	3864	+ features &= ~NETIF_F_HW_TLS_TX;
	3865	+ if (netdev->features & NETIF_F_HW_TLS_TX)
	3866	+ netdev_warn(netdev, "Disabling hw_tls_tx, not supported in switchdev mode\n");
	3867	+ }
	3868	+
3858	3869	mutex_unlock(&priv->state_lock);
3859	3870
3860	3871	return features;
@@ -3971,11 +3982,45 @@ int mlx5e_ptp_rx_manage_fs_ctx(struct mlx5e_priv priv, void ctx) · 110 unmodified lines
3971	3982	return mlx5e_ptp_rx_manage_fs(priv, set);
3972	3983	}
3973	3984
3974		−int mlx5e_hwstamp_set(struct mlx5e_priv priv, struct ifreq ifr)
	3985	+static int mlx5e_hwstamp_config_no_ptp_rx(struct mlx5e_priv *priv, bool rx_filter)
	3986	+{
	3987	+ bool rx_cqe_compress_def = priv->channels.params.rx_cqe_compress_def;
	3988	+ int err;
	3989	+
	3990	+ if (!rx_filter)
	3991	+ /* Reset CQE compression to Admin default */
	3992	+ return mlx5e_modify_rx_cqe_compression_locked(priv, rx_cqe_compress_def);
	3993	+
	3994	+ if (!MLX5E_GET_PFLAG(&priv->channels.params, MLX5E_PFLAG_RX_CQE_COMPRESS))
	3995	+ return 0;
	3996	+
	3997	+ /* Disable CQE compression */
	3998	+ netdev_warn(priv->netdev, "Disabling RX cqe compression\n");
	3999	+ err = mlx5e_modify_rx_cqe_compression_locked(priv, false);
	4000	+ if (err)
	4001	+ netdev_err(priv->netdev, "Failed disabling cqe compression err=%d\n", err);
	4002	+
	4003	+ return err;
	4004	+}
	4005	+
	4006	+static int mlx5e_hwstamp_config_ptp_rx(struct mlx5e_priv *priv, bool ptp_rx)
3975	4007	{
3976	4008	struct mlx5e_params new_params;
	4009	+
	4010	+ if (ptp_rx == priv->channels.params.ptp_rx)
	4011	+ return 0;
	4012	+
	4013	+ new_params = priv->channels.params;
	4014	+ new_params.ptp_rx = ptp_rx;
	4015	+ return mlx5e_safe_switch_params(priv, &new_params, mlx5e_ptp_rx_manage_fs_ctx,
	4016	+ &new_params.ptp_rx, true);
	4017	+}
	4018	+
	4019	+int mlx5e_hwstamp_set(struct mlx5e_priv priv, struct ifreq ifr)
	4020	+{
3977	4021	struct hwtstamp_config config;
3978	4022	bool rx_cqe_compress_def;
	4023	+ bool ptp_rx;
3979	4024	int err;
3980	4025
3981	4026	if (!MLX5_CAP_GEN(priv->mdev, device_frequency_khz) \|\|
@@ -3995,13 +4040,12 @@ int mlx5e_hwstamp_set(struct mlx5e_priv priv, struct ifreq ifr) · 13 unmodified lines
3995	4040	}
3996	4041
3997	4042	mutex_lock(&priv->state_lock);
3998		− new_params = priv->channels.params;
3999	4043	rx_cqe_compress_def = priv->channels.params.rx_cqe_compress_def;
4000	4044
4001	4045	/* RX HW timestamp */
4002	4046	switch (config.rx_filter) {
4003	4047	case HWTSTAMP_FILTER_NONE:
4004		− new_params.ptp_rx = false;
	4048	+ ptp_rx = false;
4005	4049	break;
4006	4050	case HWTSTAMP_FILTER_ALL:
4007	4051	case HWTSTAMP_FILTER_SOME:
@@ -4018,24 +4062,25 @@ int mlx5e_hwstamp_set(struct mlx5e_priv priv, struct ifreq ifr) · 10 unmodified lines
4018	4062	case HWTSTAMP_FILTER_PTP_V2_SYNC:
4019	4063	case HWTSTAMP_FILTER_PTP_V2_DELAY_REQ:
4020	4064	case HWTSTAMP_FILTER_NTP_ALL:
4021		− new_params.ptp_rx = rx_cqe_compress_def;
4022	4065	config.rx_filter = HWTSTAMP_FILTER_ALL;
	4066	+ /* ptp_rx is set if both HW TS is set and CQE
	4067	+ * compression is set
	4068	+ */
	4069	+ ptp_rx = rx_cqe_compress_def;
4023	4070	break;
4024	4071	default:
4025		− mutex_unlock(&priv->state_lock);
4026		− return -ERANGE;
	4072	+ err = -ERANGE;
	4073	+ goto err_unlock;
4027	4074	}
4028	4075
4029		− if (new_params.ptp_rx == priv->channels.params.ptp_rx)
4030		− goto out;
	4076	+ if (!priv->profile->rx_ptp_support)
	4077	+ err = mlx5e_hwstamp_config_no_ptp_rx(priv,
	4078	+ config.rx_filter != HWTSTAMP_FILTER_NONE);
	4079	+ else
	4080	+ err = mlx5e_hwstamp_config_ptp_rx(priv, ptp_rx);
	4081	+ if (err)
	4082	+ goto err_unlock;
4031	4083
4032		− err = mlx5e_safe_switch_params(priv, &new_params, mlx5e_ptp_rx_manage_fs_ctx,
4033		− &new_params.ptp_rx, true);
4034		− if (err) {
4035		− mutex_unlock(&priv->state_lock);
4036		− return err;
4037		− }
4038		−out:
4039	4084	memcpy(&priv->tstamp, &config, sizeof(config));
4040	4085	mutex_unlock(&priv->state_lock);
4041	4086
@@ -4044,6 +4089,9 @@ out: · 2 unmodified lines
4044	4089
4045	4090	return copy_to_user(ifr->ifr_data, &config,
4046	4091	sizeof(config)) ? -EFAULT : 0;
	4092	+err_unlock:
	4093	+ mutex_unlock(&priv->state_lock);
	4094	+ return err;
4047	4095	}
4048	4096
4049	4097	int mlx5e_hwstamp_get(struct mlx5e_priv priv, struct ifreq ifr)
@@ -4774,22 +4822,15 @@ static void mlx5e_build_nic_netdev(struct net_device *netdev) · 724 unmodified lines
4774	4822	}
4775	4823
4776	4824	if (mlx5_vxlan_allowed(mdev->vxlan) \|\| mlx5_geneve_tx_allowed(mdev)) {
4777		− netdev->hw_features \|= NETIF_F_GSO_UDP_TUNNEL \|
4778		− NETIF_F_GSO_UDP_TUNNEL_CSUM;
4779		− netdev->hw_enc_features \|= NETIF_F_GSO_UDP_TUNNEL \|
4780		− NETIF_F_GSO_UDP_TUNNEL_CSUM;
4781		− netdev->gso_partial_features = NETIF_F_GSO_UDP_TUNNEL_CSUM;
4782		− netdev->vlan_features \|= NETIF_F_GSO_UDP_TUNNEL \|
4783		− NETIF_F_GSO_UDP_TUNNEL_CSUM;
	4825	+ netdev->hw_features \|= NETIF_F_GSO_UDP_TUNNEL;
	4826	+ netdev->hw_enc_features \|= NETIF_F_GSO_UDP_TUNNEL;
	4827	+ netdev->vlan_features \|= NETIF_F_GSO_UDP_TUNNEL;
4784	4828	}
4785	4829
4786	4830	if (mlx5e_tunnel_proto_supported_tx(mdev, IPPROTO_GRE)) {
4787		− netdev->hw_features \|= NETIF_F_GSO_GRE \|
4788		− NETIF_F_GSO_GRE_CSUM;
4789		− netdev->hw_enc_features \|= NETIF_F_GSO_GRE \|
4790		− NETIF_F_GSO_GRE_CSUM;
4791		− netdev->gso_partial_features \|= NETIF_F_GSO_GRE \|
4792		− NETIF_F_GSO_GRE_CSUM;
	4831	+ netdev->hw_features \|= NETIF_F_GSO_GRE;
	4832	+ netdev->hw_enc_features \|= NETIF_F_GSO_GRE;
	4833	+ netdev->gso_partial_features \|= NETIF_F_GSO_GRE;
4793	4834	}
4794	4835
4795	4836	if (mlx5e_tunnel_proto_supported_tx(mdev, IPPROTO_IPIP)) {
@@ -5229,6 +5270,11 @@ static void mlx5e_update_features(struct net_device *netdev) · 433 unmodified lines
5229	5270	rtnl_unlock();
5230	5271	}
5231	5272
	5273	+static void mlx5e_reset_channels(struct net_device *netdev)
	5274	+{
	5275	+ netdev_reset_tc(netdev);
	5276	+}
	5277	+
5232	5278	int mlx5e_attach_netdev(struct mlx5e_priv *priv)
5233	5279	{
5234	5280	const bool take_rtnl = priv->netdev->reg_state == NETREG_REGISTERED;
@@ -5283,6 +5329,7 @@ err_cleanup_tx: · 48 unmodified lines
5283	5329	profile->cleanup_tx(priv);
5284	5330
5285	5331	out:
	5332	+ mlx5e_reset_channels(priv->netdev);
5286	5333	set_bit(MLX5E_STATE_DESTROYING, &priv->state);
5287	5334	cancel_work_sync(&priv->update_stats_work);
5288	5335	return err;
@@ -5300,6 +5347,7 @@ void mlx5e_detach_netdev(struct mlx5e_priv *priv) · 11 unmodified lines
5300	5347
5301	5348	profile->cleanup_rx(priv);
5302	5349	profile->cleanup_tx(priv);
	5350	+ mlx5e_reset_channels(priv->netdev);
5303	5351	cancel_work_sync(&priv->update_stats_work);
5304	5352	}
5305	5353

drivers/net/ethernet/mellanox/mlx5/core/en_tc.c+32−7

@@ -1322,10 +1322,10 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv, · 1321 unmodified lines
1322	1322	struct netlink_ext_ack *extack)
1323	1323	{
1324	1324	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
1325		− struct net_device out_dev, encap_dev = NULL;
1326	1325	struct mlx5e_tc_flow_parse_attr *parse_attr;
1327	1326	struct mlx5_flow_attr *attr = flow->attr;
1328	1327	bool vf_tun = false, encap_valid = true;
	1328	+ struct net_device *encap_dev = NULL;
1329	1329	struct mlx5_esw_flow_attr *esw_attr;
1330	1330	struct mlx5_fc *counter = NULL;
1331	1331	struct mlx5e_rep_priv *rpriv;
@@ -1371,16 +1371,22 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv, · 39 unmodified lines
1371	1371	esw_attr = attr->esw_attr;
1372	1372
1373	1373	for (out_index = 0; out_index < MLX5_MAX_FLOW_FWD_VPORTS; out_index++) {
	1374	+ struct net_device *out_dev;
1374	1375	int mirred_ifindex;
1375	1376
1376	1377	if (!(esw_attr->dests[out_index].flags & MLX5_ESW_DEST_ENCAP))
1377	1378	continue;
1378	1379
1379	1380	mirred_ifindex = parse_attr->mirred_ifindex[out_index];
1380		− out_dev = __dev_get_by_index(dev_net(priv->netdev),
1381		− mirred_ifindex);
	1381	+ out_dev = dev_get_by_index(dev_net(priv->netdev), mirred_ifindex);
	1382	+ if (!out_dev) {
	1383	+ NL_SET_ERR_MSG_MOD(extack, "Requested mirred device not found");
	1384	+ err = -ENODEV;
	1385	+ goto err_out;
	1386	+ }
1382	1387	err = mlx5e_attach_encap(priv, flow, out_dev, out_index,
1383	1388	extack, &encap_dev, &encap_valid);
	1389	+ dev_put(out_dev);
1384	1390	if (err)
1385	1391	goto err_out;
1386	1392
@@ -1393,6 +1399,12 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv, · 6 unmodified lines
1393	1399	esw_attr->dests[out_index].mdev = out_priv->mdev;
1394	1400	}
1395	1401
	1402	+ if (vf_tun && esw_attr->out_count > 1) {
	1403	+ NL_SET_ERR_MSG_MOD(extack, "VF tunnel encap with mirroring is not supported");
	1404	+ err = -EOPNOTSUPP;
	1405	+ goto err_out;
	1406	+ }
	1407	+
1396	1408	err = mlx5_eswitch_add_vlan_action(esw, attr);
1397	1409	if (err)
1398	1410	goto err_out;
@@ -2003,11 +2015,13 @@ static int __parse_cls_flower(struct mlx5e_priv *priv, · 604 unmodified lines
2003	2015	misc_parameters_3);
2004	2016	struct flow_rule *rule = flow_cls_offload_flow_rule(f);
2005	2017	struct flow_dissector *dissector = rule->match.dissector;
	2018	+ enum fs_flow_table_type fs_type;
2006	2019	u16 addr_type = 0;
2007	2020	u8 ip_proto = 0;
2008	2021	u8 *match_level;
2009	2022	int err;
2010	2023
	2024	+ fs_type = mlx5e_is_eswitch_flow(flow) ? FS_FT_FDB : FS_FT_NIC_RX;
2011	2025	match_level = outer_match_level;
2012	2026
2013	2027	if (dissector->used_keys &
@@ -2133,6 +2147,13 @@ static int __parse_cls_flower(struct mlx5e_priv *priv, · 119 unmodified lines
2133	2147	if (match.mask->vlan_id \|\|
2134	2148	match.mask->vlan_priority \|\|
2135	2149	match.mask->vlan_tpid) {
	2150	+ if (!MLX5_CAP_FLOWTABLE_TYPE(priv->mdev, ft_field_support.outer_second_vid,
	2151	+ fs_type)) {
	2152	+ NL_SET_ERR_MSG_MOD(extack,
	2153	+ "Matching on CVLAN is not supported");
	2154	+ return -EOPNOTSUPP;
	2155	+ }
	2156	+
2136	2157	if (match.key->vlan_tpid == htons(ETH_P_8021AD)) {
2137	2158	MLX5_SET(fte_match_set_misc, misc_c,
2138	2159	outer_second_svlan_tag, 1);
@@ -3526,8 +3547,12 @@ static int add_vlan_push_action(struct mlx5e_priv *priv, · 1387 unmodified lines
3526	3547	if (err)
3527	3548	return err;
3528	3549
3529		− *out_dev = dev_get_by_index_rcu(dev_net(vlan_dev),
3530		− dev_get_iflink(vlan_dev));
	3550	+ rcu_read_lock();
	3551	+ *out_dev = dev_get_by_index_rcu(dev_net(vlan_dev), dev_get_iflink(vlan_dev));
	3552	+ rcu_read_unlock();
	3553	+ if (!*out_dev)
	3554	+ return -ENODEV;
	3555	+
3531	3556	if (is_vlan_dev(*out_dev))
3532	3557	err = add_vlan_push_action(priv, attr, out_dev, action);
3533	3558
@@ -4740,7 +4765,7 @@ static void mlx5e_tc_hairpin_update_dead_peer(struct mlx5e_priv *priv, · 1206 unmodified lines
4740	4765	list_for_each_entry_safe(hpe, tmp, &init_wait_list, dead_peer_wait_list) {
4741	4766	wait_for_completion(&hpe->res_ready);
4742	4767	if (!IS_ERR_OR_NULL(hpe->hp) && hpe->peer_vhca_id == peer_vhca_id)
4743		− hpe->hp->pair->peer_gone = true;
	4768	+ mlx5_core_hairpin_clear_dead_peer(hpe->hp->pair);
4744	4769
4745	4770	mlx5e_hairpin_put(priv, hpe);
4746	4771	}
@@ -5074,7 +5099,7 @@ bool mlx5e_tc_update_skb(struct mlx5_cqe64 *cqe, · 327 unmodified lines
5074	5099
5075	5100	if (mapped_obj.type == MLX5_MAPPED_OBJ_CHAIN) {
5076	5101	chain = mapped_obj.chain;
5077		− tc_skb_ext = skb_ext_add(skb, TC_SKB_EXT);
	5102	+ tc_skb_ext = tc_skb_ext_alloc(skb);
5078	5103	if (WARN_ON(!tc_skb_ext))
5079	5104	return false;
5080	5105

drivers/net/ethernet/mellanox/mlx5/core/en_tc.h+3−0

@@ -178,6 +178,9 @@ void mlx5e_take_all_encap_flows(struct mlx5e_encap_entry e, struct list_head f · 177 unmodified lines
178	178	void mlx5e_put_flow_list(struct mlx5e_priv priv, struct list_head flow_list);
179	179
180	180	struct mlx5e_neigh_hash_entry;
	181	+struct mlx5e_encap_entry *
	182	+mlx5e_get_next_init_encap(struct mlx5e_neigh_hash_entry *nhe,
	183	+ struct mlx5e_encap_entry *e);
181	184	void mlx5e_tc_update_neigh_used_value(struct mlx5e_neigh_hash_entry *nhe);
182	185
183	186	void mlx5e_tc_reoffload_flows_work(struct work_struct *work);

drivers/net/ethernet/mellanox/mlx5/core/en_tx.c+3−22

@@ -32,7 +32,6 @@ · 31 unmodified lines
32	32
33	33	#include <linux/tcp.h>
34	34	#include <linux/if_vlan.h>
35		−#include <linux/ptp_classify.h>
36	35	#include <net/geneve.h>
37	36	#include <net/dsfield.h>
38	37	#include "en.h"
@@ -67,24 +66,6 @@ static inline int mlx5e_get_dscp_up(struct mlx5e_priv priv, struct sk_buff skb · 28 unmodified lines
67	66	}
68	67	#endif
69	68
70		−static bool mlx5e_use_ptpsq(struct sk_buff *skb)
71		−{
72		− struct flow_keys fk;
73		−
74		− if (!skb_flow_dissect_flow_keys(skb, &fk, 0))
75		− return false;
76		−
77		− if (fk.basic.n_proto == htons(ETH_P_1588))
78		− return true;
79		−
80		− if (fk.basic.n_proto != htons(ETH_P_IP) &&
81		− fk.basic.n_proto != htons(ETH_P_IPV6))
82		− return false;
83		−
84		− return (fk.basic.ip_proto == IPPROTO_UDP &&
85		− fk.ports.dst == htons(PTP_EV_PORT));
86		−}
87		−
88	69	static u16 mlx5e_select_ptpsq(struct net_device dev, struct sk_buff skb)
89	70	{
90	71	struct mlx5e_priv *priv = netdev_priv(dev);
@@ -145,9 +126,9 @@ u16 mlx5e_select_queue(struct net_device dev, struct sk_buff skb, · 54 unmodified lines
145	126	}
146	127
147	128	ptp_channel = READ_ONCE(priv->channels.ptp);
148		− if (unlikely(ptp_channel) &&
149		− test_bit(MLX5E_PTP_STATE_TX, ptp_channel->state) &&
150		− mlx5e_use_ptpsq(skb))
	129	+ if (unlikely(ptp_channel &&
	130	+ test_bit(MLX5E_PTP_STATE_TX, ptp_channel->state) &&
	131	+ mlx5e_use_ptpsq(skb)))
151	132	return mlx5e_select_ptpsq(dev, skb);
152	133
153	134	txq_ix = netdev_pick_tx(dev, skb, NULL);

drivers/net/ethernet/mellanox/mlx5/core/eq.c+4−2

@@ -136,7 +136,7 @@ static int mlx5_eq_comp_int(struct notifier_block *nb, · 135 unmodified lines
136	136
137	137	eqe = next_eqe_sw(eq);
138	138	if (!eqe)
139		− return 0;
	139	+ goto out;
140	140
141	141	do {
142	142	struct mlx5_core_cq *cq;
@@ -161,6 +161,8 @@ static int mlx5_eq_comp_int(struct notifier_block *nb, · 18 unmodified lines
161	161	++eq->cons_index;
162	162
163	163	} while ((++num_eqes < MLX5_EQ_POLLING_BUDGET) && (eqe = next_eqe_sw(eq)));
	164	+
	165	+out:
164	166	eq_update_ci(eq, 1);
165	167
166	168	if (cqn != -1)
@@ -248,9 +250,9 @@ static int mlx5_eq_async_int(struct notifier_block *nb, · 81 unmodified lines
248	250	++eq->cons_index;
249	251
250	252	} while ((++num_eqes < MLX5_EQ_POLLING_BUDGET) && (eqe = next_eqe_sw(eq)));
251		− eq_update_ci(eq, 1);
252	253
253	254	out:
	255	+ eq_update_ci(eq, 1);
254	256	mlx5_eq_async_int_unlock(eq_async, recovery, &flags);
255	257
256	258	return unlikely(recovery) ? num_eqes : 0;

drivers/net/ethernet/mellanox/mlx5/core/eswitch.c+7−0

@@ -35,6 +35,7 @@ · 34 unmodified lines
35	35	#include <linux/mlx5/mlx5_ifc.h>
36	36	#include <linux/mlx5/vport.h>
37	37	#include <linux/mlx5/fs.h>
	38	+#include <linux/mlx5/mpfs.h>
38	39	#include "esw/acl/lgcy.h"
39	40	#include "esw/legacy.h"
40	41	#include "mlx5_core.h"
@@ -1053,6 +1054,12 @@ int mlx5_esw_vport_enable(struct mlx5_eswitch *esw, u16 vport_num, · 1012 unmodified lines
1053	1054	goto err_vhca_mapping;
1054	1055	}
1055	1056
	1057	+ /* External controller host PF has factory programmed MAC.
	1058	+ * Read it from the device.
	1059	+ */
	1060	+ if (mlx5_core_is_ecpf(esw->dev) && vport_num == MLX5_VPORT_PF)
	1061	+ mlx5_query_nic_vport_mac_address(esw->dev, vport_num, true, vport->info.mac);
	1062	+
1056	1063	esw_vport_change_handle_locked(vport);
1057	1064
1058	1065	esw->enabled_vports++;

drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c+2−1

@@ -219,7 +219,8 @@ esw_setup_slow_path_dest(struct mlx5_flow_destination *dest, · 218 unmodified lines
219	219	struct mlx5_fs_chains *chains,
220	220	int i)
221	221	{
222		− flow_act->flags \|= FLOW_ACT_IGNORE_FLOW_LEVEL;
	222	+ if (mlx5_chains_ignore_flow_level_supported(chains))
	223	+ flow_act->flags \|= FLOW_ACT_IGNORE_FLOW_LEVEL;
223	224	dest[i].type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
224	225	dest[i].ft = mlx5_chains_get_tc_end_ft(chains);
225	226	}

drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_termtbl.c+26−35

@@ -65,7 +65,7 @@ mlx5_eswitch_termtbl_create(struct mlx5_core_dev *dev, · 64 unmodified lines
65	65	{
66	66	struct mlx5_flow_table_attr ft_attr = {};
67	67	struct mlx5_flow_namespace *root_ns;
68		− int err;
	68	+ int err, err2;
69	69
70	70	root_ns = mlx5_get_flow_namespace(dev, MLX5_FLOW_NAMESPACE_FDB);
71	71	if (!root_ns) {
@@ -76,33 +76,34 @@ mlx5_eswitch_termtbl_create(struct mlx5_core_dev *dev, · 4 unmodified lines
76	76	/* As this is the terminating action then the termination table is the
77	77	* same prio as the slow path
78	78	*/
79		− ft_attr.flags = MLX5_FLOW_TABLE_TERMINATION \|
	79	+ ft_attr.flags = MLX5_FLOW_TABLE_TERMINATION \| MLX5_FLOW_TABLE_UNMANAGED \|
80	80	MLX5_FLOW_TABLE_TUNNEL_EN_REFORMAT;
81		− ft_attr.prio = FDB_SLOW_PATH;
	81	+ ft_attr.prio = FDB_TC_OFFLOAD;
82	82	ft_attr.max_fte = 1;
	83	+ ft_attr.level = 1;
83	84	ft_attr.autogroup.max_num_groups = 1;
84	85	tt->termtbl = mlx5_create_auto_grouped_flow_table(root_ns, &ft_attr);
85	86	if (IS_ERR(tt->termtbl)) {
86		− esw_warn(dev, "Failed to create termination table (error %d)\n",
87		− IS_ERR(tt->termtbl));
88		− return -EOPNOTSUPP;
	87	+ err = PTR_ERR(tt->termtbl);
	88	+ esw_warn(dev, "Failed to create termination table, err %pe\n", tt->termtbl);
	89	+ return err;
89	90	}
90	91
91	92	tt->rule = mlx5_add_flow_rules(tt->termtbl, NULL, flow_act,
92	93	&tt->dest, 1);
93	94	if (IS_ERR(tt->rule)) {
94		− esw_warn(dev, "Failed to create termination table rule (error %d)\n",
95		− IS_ERR(tt->rule));
	95	+ err = PTR_ERR(tt->rule);
	96	+ esw_warn(dev, "Failed to create termination table rule, err %pe\n", tt->rule);
96	97	goto add_flow_err;
97	98	}
98	99	return 0;
99	100
100	101	add_flow_err:
101		− err = mlx5_destroy_flow_table(tt->termtbl);
102		− if (err)
103		− esw_warn(dev, "Failed to destroy termination table\n");
	102	+ err2 = mlx5_destroy_flow_table(tt->termtbl);
	103	+ if (err2)
	104	+ esw_warn(dev, "Failed to destroy termination table, err %d\n", err2);
104	105
105		− return -EOPNOTSUPP;
	106	+ return err;
106	107	}
107	108
108	109	static struct mlx5_termtbl_handle *
@@ -172,19 +173,6 @@ mlx5_eswitch_termtbl_put(struct mlx5_eswitch *esw, · 63 unmodified lines
172	173	}
173	174	}
174	175
175		−static bool mlx5_eswitch_termtbl_is_encap_reformat(struct mlx5_pkt_reformat *rt)
176		−{
177		− switch (rt->reformat_type) {
178		− case MLX5_REFORMAT_TYPE_L2_TO_VXLAN:
179		− case MLX5_REFORMAT_TYPE_L2_TO_NVGRE:
180		− case MLX5_REFORMAT_TYPE_L2_TO_L2_TUNNEL:
181		− case MLX5_REFORMAT_TYPE_L2_TO_L3_TUNNEL:
182		− return true;
183		− default:
184		− return false;
185		− }
186		−}
187		−
188	176	static void
189	177	mlx5_eswitch_termtbl_actions_move(struct mlx5_flow_act *src,
190	178	struct mlx5_flow_act *dst)
@@ -202,14 +190,6 @@ mlx5_eswitch_termtbl_actions_move(struct mlx5_flow_act *src, · 11 unmodified lines
202	190	memset(&src->vlan[1], 0, sizeof(src->vlan[1]));
203	191	}
204	192	}
205		−
206		− if (src->action & MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT &&
207		− mlx5_eswitch_termtbl_is_encap_reformat(src->pkt_reformat)) {
208		− src->action &= ~MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT;
209		− dst->action \|= MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT;
210		− dst->pkt_reformat = src->pkt_reformat;
211		− src->pkt_reformat = NULL;
212		− }
213	193	}
214	194
215	195	static bool mlx5_eswitch_offload_is_uplink_port(const struct mlx5_eswitch *esw,
@@ -238,6 +218,7 @@ mlx5_eswitch_termtbl_required(struct mlx5_eswitch *esw, · 22 unmodified lines
238	218	int i;
239	219
240	220	if (!MLX5_CAP_ESW_FLOWTABLE_FDB(esw->dev, termination_table) \|\|
	221	+ !MLX5_CAP_ESW_FLOWTABLE_FDB(esw->dev, ignore_flow_level) \|\|
241	222	attr->flags & MLX5_ESW_ATTR_FLAG_SLOW_PATH \|\|
242	223	!mlx5_eswitch_offload_is_uplink_port(esw, spec))
243	224	return false;
@@ -279,12 +260,19 @@ mlx5_eswitch_add_termtbl_rule(struct mlx5_eswitch *esw, · 35 unmodified lines
279	260	if (dest[i].type != MLX5_FLOW_DESTINATION_TYPE_VPORT)
280	261	continue;
281	262
	263	+ if (attr->dests[num_vport_dests].flags & MLX5_ESW_DEST_ENCAP) {
	264	+ term_tbl_act.action \|= MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT;
	265	+ term_tbl_act.pkt_reformat = attr->dests[num_vport_dests].pkt_reformat;
	266	+ } else {
	267	+ term_tbl_act.action &= ~MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT;
	268	+ term_tbl_act.pkt_reformat = NULL;
	269	+ }
	270	+
282	271	/* get the terminating table for the action list */
283	272	tt = mlx5_eswitch_termtbl_get_create(esw, &term_tbl_act,
284	273	&dest[i], attr);
285	274	if (IS_ERR(tt)) {
286		− esw_warn(esw->dev, "Failed to get termination table (error %d)\n",
287		− IS_ERR(tt));
	275	+ esw_warn(esw->dev, "Failed to get termination table, err %pe\n", tt);
288	276	goto revert_changes;
289	277	}
290	278	attr->dests[num_vport_dests].termtbl = tt;
@@ -301,6 +289,9 @@ mlx5_eswitch_add_termtbl_rule(struct mlx5_eswitch *esw, · 10 unmodified lines
301	289	goto revert_changes;
302	290
303	291	/* create the FTE */
	292	+ flow_act->action &= ~MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT;
	293	+ flow_act->pkt_reformat = NULL;
	294	+ flow_act->flags \|= FLOW_ACT_IGNORE_FLOW_LEVEL;
304	295	rule = mlx5_add_flow_rules(fdb, spec, flow_act, dest, num_dest);
305	296	if (IS_ERR(rule))
306	297	goto revert_changes;

drivers/net/ethernet/mellanox/mlx5/core/fw_reset.c+3−0

@@ -349,6 +349,9 @@ static void mlx5_sync_reset_abort_event(struct work_struct *work) · 348 unmodified lines
349	349	reset_abort_work);
350	350	struct mlx5_core_dev *dev = fw_reset->dev;
351	351
	352	+ if (!test_bit(MLX5_FW_RESET_FLAGS_RESET_REQUESTED, &fw_reset->reset_flags))
	353	+ return;
	354	+
352	355	mlx5_sync_reset_clear_reset_requested(dev, true);
353	356	mlx5_core_warn(dev, "PCI Sync FW Update Reset Aborted.\n");
354	357	}

drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c+6−0

@@ -307,6 +307,11 @@ int mlx5_lag_mp_init(struct mlx5_lag *ldev) · 306 unmodified lines
307	307	struct lag_mp *mp = &ldev->lag_mp;
308	308	int err;
309	309
	310	+ /* always clear mfi, as it might become stale when a route delete event
	311	+ * has been missed
	312	+ */
	313	+ mp->mfi = NULL;
	314	+
310	315	if (mp->fib_nb.notifier_call)
311	316	return 0;
312	317
@@ -335,4 +340,5 @@ void mlx5_lag_mp_cleanup(struct mlx5_lag *ldev) · 22 unmodified lines
335	340	unregister_fib_notifier(&init_net, &mp->fib_nb);
336	341	destroy_workqueue(mp->wq);
337	342	mp->fib_nb.notifier_call = NULL;
	343	+ mp->mfi = NULL;
338	344	}

drivers/net/ethernet/mellanox/mlx5/core/lib/fs_chains.c+1−1

@@ -107,7 +107,7 @@ bool mlx5_chains_prios_supported(struct mlx5_fs_chains *chains) · 106 unmodified lines
107	107	return chains->flags & MLX5_CHAINS_AND_PRIOS_SUPPORTED;
108	108	}
109	109
110		−static bool mlx5_chains_ignore_flow_level_supported(struct mlx5_fs_chains *chains)
	110	+bool mlx5_chains_ignore_flow_level_supported(struct mlx5_fs_chains *chains)
111	111	{
112	112	return chains->flags & MLX5_CHAINS_IGNORE_FLOW_LEVEL_SUPPORTED;
113	113	}

drivers/net/ethernet/mellanox/mlx5/core/lib/fs_chains.h+5−0

@@ -28,6 +28,7 @@ struct mlx5_chains_attr { · 27 unmodified lines
28	28
29	29	bool
30	30	mlx5_chains_prios_supported(struct mlx5_fs_chains *chains);
	31	+bool mlx5_chains_ignore_flow_level_supported(struct mlx5_fs_chains *chains);
31	32	bool
32	33	mlx5_chains_backwards_supported(struct mlx5_fs_chains *chains);
33	34	u32
@@ -70,6 +71,10 @@ mlx5_chains_set_end_ft(struct mlx5_fs_chains *chains, · 36 unmodified lines
70	71
71	72	#else /* CONFIG_MLX5_CLS_ACT */
72	73
	74	+static inline bool
	75	+mlx5_chains_ignore_flow_level_supported(struct mlx5_fs_chains *chains)
	76	+{ return false; }
	77	+
73	78	static inline struct mlx5_flow_table *
74	79	mlx5_chains_get_table(struct mlx5_fs_chains *chains, u32 chain, u32 prio,
75	80	u32 level) { return ERR_PTR(-EOPNOTSUPP); }

drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.c+3−0

@@ -33,6 +33,7 @@ · 32 unmodified lines
33	33	#include <linux/etherdevice.h>
34	34	#include <linux/mlx5/driver.h>
35	35	#include <linux/mlx5/mlx5_ifc.h>
	36	+#include <linux/mlx5/mpfs.h>
36	37	#include <linux/mlx5/eswitch.h>
37	38	#include "mlx5_core.h"
38	39	#include "lib/mpfs.h"
@@ -175,6 +176,7 @@ out: · 136 unmodified lines
175	176	mutex_unlock(&mpfs->lock);
176	177	return err;
177	178	}
	179	+EXPORT_SYMBOL(mlx5_mpfs_add_mac);
178	180
179	181	int mlx5_mpfs_del_mac(struct mlx5_core_dev dev, u8 mac)
180	182	{
@@ -206,3 +208,4 @@ unlock: · 25 unmodified lines
206	208	mutex_unlock(&mpfs->lock);
207	209	return err;
208	210	}
	211	+EXPORT_SYMBOL(mlx5_mpfs_del_mac);

drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.h+1−4

@@ -84,12 +84,9 @@ struct l2addr_node { · 83 unmodified lines
84	84	#ifdef CONFIG_MLX5_MPFS
85	85	int mlx5_mpfs_init(struct mlx5_core_dev *dev);
86	86	void mlx5_mpfs_cleanup(struct mlx5_core_dev *dev);
87		−int mlx5_mpfs_add_mac(struct mlx5_core_dev dev, u8 mac);
88		−int mlx5_mpfs_del_mac(struct mlx5_core_dev dev, u8 mac);
89	87	#else /* #ifndef CONFIG_MLX5_MPFS */
90	88	static inline int mlx5_mpfs_init(struct mlx5_core_dev *dev) { return 0; }
91	89	static inline void mlx5_mpfs_cleanup(struct mlx5_core_dev *dev) {}
92		−static inline int mlx5_mpfs_add_mac(struct mlx5_core_dev dev, u8 mac) { return 0; }
93		−static inline int mlx5_mpfs_del_mac(struct mlx5_core_dev dev, u8 mac) { return 0; }
94	90	#endif
	91	+
95	92	#endif

drivers/net/ethernet/mellanox/mlx5/core/main.c+7−7

@@ -503,7 +503,7 @@ static int handle_hca_cap_odp(struct mlx5_core_dev dev, void set_ctx) · 502 unmodified lines
503	503
504	504	static int handle_hca_cap(struct mlx5_core_dev dev, void set_ctx)
505	505	{
506		− struct mlx5_profile *prof = dev->profile;
	506	+ struct mlx5_profile *prof = &dev->profile;
507	507	void *set_hca_cap;
508	508	int err;
509	509
@@ -524,11 +524,11 @@ static int handle_hca_cap(struct mlx5_core_dev dev, void set_ctx) · 14 unmodified lines
524	524	to_fw_pkey_sz(dev, 128));
525	525
526	526	/* Check log_max_qp from HCA caps to set in current profile */
527		− if (MLX5_CAP_GEN_MAX(dev, log_max_qp) < profile[prof_sel].log_max_qp) {
	527	+ if (MLX5_CAP_GEN_MAX(dev, log_max_qp) < prof->log_max_qp) {
528	528	mlx5_core_warn(dev, "log_max_qp value in current profile is %d, changing it to HCA capability limit (%d)\n",
529		− profile[prof_sel].log_max_qp,
	529	+ prof->log_max_qp,
530	530	MLX5_CAP_GEN_MAX(dev, log_max_qp));
531		− profile[prof_sel].log_max_qp = MLX5_CAP_GEN_MAX(dev, log_max_qp);
	531	+ prof->log_max_qp = MLX5_CAP_GEN_MAX(dev, log_max_qp);
532	532	}
533	533	if (prof->mask & MLX5_PROF_MASK_QP_SIZE)
534	534	MLX5_SET(cmd_hca_cap, set_hca_cap, log_max_qp,
@@ -1161,7 +1161,7 @@ static int mlx5_load(struct mlx5_core_dev *dev) · 626 unmodified lines
1161	1161	err = mlx5_core_set_hca_defaults(dev);
1162	1162	if (err) {
1163	1163	mlx5_core_err(dev, "Failed to set hca defaults\n");
1164		− goto err_sriov;
	1164	+ goto err_set_hca;
1165	1165	}
1166	1166
1167	1167	mlx5_vhca_event_start(dev);
@@ -1194,6 +1194,7 @@ err_ec: · 26 unmodified lines
1194	1194	mlx5_sf_hw_table_destroy(dev);
1195	1195	err_vhca:
1196	1196	mlx5_vhca_event_stop(dev);
	1197	+err_set_hca:
1197	1198	mlx5_cleanup_fs(dev);
1198	1199	err_fs:
1199	1200	mlx5_accel_tls_cleanup(dev);
@@ -1381,8 +1382,7 @@ int mlx5_mdev_init(struct mlx5_core_dev *dev, int profile_idx) · 181 unmodified lines
1381	1382	struct mlx5_priv *priv = &dev->priv;
1382	1383	int err;
1383	1384
1384		− dev->profile = &profile[profile_idx];
1385		−
	1385	+ memcpy(&dev->profile, &profile[profile_idx], sizeof(dev->profile));
1386	1386	INIT_LIST_HEAD(&priv->ctx_list);
1387	1387	spin_lock_init(&priv->ctx_lock);
1388	1388	mutex_init(&dev->intf_state_mutex);

drivers/net/ethernet/mellanox/mlx5/core/mr.c+1−1

@@ -54,7 +54,7 @@ int mlx5_core_create_mkey(struct mlx5_core_dev *dev, · 53 unmodified lines
54	54	mkey_index = MLX5_GET(create_mkey_out, lout, mkey_index);
55	55	mkey->iova = MLX5_GET64(mkc, mkc, start_addr);
56	56	mkey->size = MLX5_GET64(mkc, mkc, len);
57		− mkey->key \|= mlx5_idx_to_mkey(mkey_index);
	57	+ mkey->key = (u32)mlx5_mkey_variant(mkey->key) \| mlx5_idx_to_mkey(mkey_index);
58	58	mkey->pd = MLX5_GET(mkc, mkc, pd);
59	59	init_waitqueue_head(&mkey->wait);
60	60

drivers/net/ethernet/mellanox/mlx5/core/pci_irq.c+17−5

@@ -95,9 +95,10 @@ int mlx5_get_default_msix_vec_count(struct mlx5_core_dev *dev, int num_vfs) · 94 unmodified lines
95	95	int mlx5_set_msix_vec_count(struct mlx5_core_dev *dev, int function_id,
96	96	int msix_vec_count)
97	97	{
98		− int sz = MLX5_ST_SZ_BYTES(set_hca_cap_in);
	98	+ int query_sz = MLX5_ST_SZ_BYTES(query_hca_cap_out);
	99	+ int set_sz = MLX5_ST_SZ_BYTES(set_hca_cap_in);
	100	+ void hca_cap = NULL, query_cap = NULL, *cap;
99	101	int num_vf_msix, min_msix, max_msix;
100		− void hca_cap, cap;
101	102	int ret;
102	103
103	104	num_vf_msix = MLX5_CAP_GEN_MAX(dev, num_total_dynamic_vf_msix);
@@ -116,11 +117,20 @@ int mlx5_set_msix_vec_count(struct mlx5_core_dev *dev, int function_id, · 12 unmodified lines
116	117	if (msix_vec_count > max_msix)
117	118	return -EOVERFLOW;
118	119
119		− hca_cap = kzalloc(sz, GFP_KERNEL);
120		− if (!hca_cap)
121		− return -ENOMEM;
	120	+ query_cap = kzalloc(query_sz, GFP_KERNEL);
	121	+ hca_cap = kzalloc(set_sz, GFP_KERNEL);
	122	+ if (!hca_cap \|\| !query_cap) {
	123	+ ret = -ENOMEM;
	124	+ goto out;
	125	+ }
	126	+
	127	+ ret = mlx5_vport_get_other_func_cap(dev, function_id, query_cap);
	128	+ if (ret)
	129	+ goto out;
122	130
123	131	cap = MLX5_ADDR_OF(set_hca_cap_in, hca_cap, capability);
	132	+ memcpy(cap, MLX5_ADDR_OF(query_hca_cap_out, query_cap, capability),
	133	+ MLX5_UN_SZ_BYTES(hca_cap_union));
124	134	MLX5_SET(cmd_hca_cap, cap, dynamic_msix_table_size, msix_vec_count);
125	135
126	136	MLX5_SET(set_hca_cap_in, hca_cap, opcode, MLX5_CMD_OP_SET_HCA_CAP);
@@ -130,7 +140,9 @@ int mlx5_set_msix_vec_count(struct mlx5_core_dev *dev, int function_id, · 3 unmodified lines
130	140	MLX5_SET(set_hca_cap_in, hca_cap, op_mod,
131	141	MLX5_SET_HCA_CAP_OP_MOD_GENERAL_DEVICE << 1);
132	142	ret = mlx5_cmd_exec_in(dev, set_hca_cap, hca_cap);
	143	+out:
133	144	kfree(hca_cap);
	145	+ kfree(query_cap);
134	146	return ret;
135	147	}
136	148

drivers/net/ethernet/mellanox/mlx5/core/rdma.c+3−0

@@ -156,6 +156,9 @@ void mlx5_rdma_enable_roce(struct mlx5_core_dev *dev) · 155 unmodified lines
156	156	{
157	157	int err;
158	158
	159	+ if (!MLX5_CAP_GEN(dev, roce))
	160	+ return;
	161	+
159	162	err = mlx5_nic_vport_enable_roce(dev);
160	163	if (err) {
161	164	mlx5_core_err(dev, "Failed to enable RoCE: %d\n", err);

drivers/net/ethernet/mellanox/mlx5/core/sf/dev/dev.c+1−0

@@ -163,6 +163,7 @@ mlx5_sf_dev_state_change_handler(struct notifier_block *nb, unsigned long event_ · 162 unmodified lines
163	163	sf_index = event->function_id - base_id;
164	164	sf_dev = xa_load(&table->devices, sf_index);
165	165	switch (event->new_vhca_state) {
	166	+ case MLX5_VHCA_STATE_INVALID:
166	167	case MLX5_VHCA_STATE_ALLOCATED:
167	168	if (sf_dev)
168	169	mlx5_sf_dev_del(table->dev, sf_dev, sf_index);

drivers/net/ethernet/mellanox/mlx5/core/sf/devlink.c+11−7

@@ -136,10 +136,10 @@ static enum devlink_port_fn_state mlx5_sf_to_devlink_state(u8 hw_state) · 135 unmodified lines
136	136	switch (hw_state) {
137	137	case MLX5_VHCA_STATE_ACTIVE:
138	138	case MLX5_VHCA_STATE_IN_USE:
139		− case MLX5_VHCA_STATE_TEARDOWN_REQUEST:
140	139	return DEVLINK_PORT_FN_STATE_ACTIVE;
141	140	case MLX5_VHCA_STATE_INVALID:
142	141	case MLX5_VHCA_STATE_ALLOCATED:
	142	+ case MLX5_VHCA_STATE_TEARDOWN_REQUEST:
143	143	default:
144	144	return DEVLINK_PORT_FN_STATE_INACTIVE;
145	145	}
@@ -192,14 +192,17 @@ sf_err: · 46 unmodified lines
192	192	return err;
193	193	}
194	194
195		−static int mlx5_sf_activate(struct mlx5_core_dev dev, struct mlx5_sf sf)
	195	+static int mlx5_sf_activate(struct mlx5_core_dev dev, struct mlx5_sf sf,
	196	+ struct netlink_ext_ack *extack)
196	197	{
197	198	int err;
198	199
199	200	if (mlx5_sf_is_active(sf))
200	201	return 0;
201		− if (sf->hw_state != MLX5_VHCA_STATE_ALLOCATED)
202		− return -EINVAL;
	202	+ if (sf->hw_state != MLX5_VHCA_STATE_ALLOCATED) {
	203	+ NL_SET_ERR_MSG_MOD(extack, "SF is inactivated but it is still attached");
	204	+ return -EBUSY;
	205	+ }
203	206
204	207	err = mlx5_cmd_sf_enable_hca(dev, sf->hw_fn_id);
205	208	if (err)
@@ -226,7 +229,8 @@ static int mlx5_sf_deactivate(struct mlx5_core_dev dev, struct mlx5_sf sf) · 20 unmodified lines
226	229
227	230	static int mlx5_sf_state_set(struct mlx5_core_dev dev, struct mlx5_sf_table table,
228	231	struct mlx5_sf *sf,
229		− enum devlink_port_fn_state state)
	232	+ enum devlink_port_fn_state state,
	233	+ struct netlink_ext_ack *extack)
230	234	{
231	235	int err = 0;
232	236
@@ -234,7 +238,7 @@ static int mlx5_sf_state_set(struct mlx5_core_dev dev, struct mlx5_sf_table ta · 1 unmodified line
234	238	if (state == mlx5_sf_to_devlink_state(sf->hw_state))
235	239	goto out;
236	240	if (state == DEVLINK_PORT_FN_STATE_ACTIVE)
237		− err = mlx5_sf_activate(dev, sf);
	241	+ err = mlx5_sf_activate(dev, sf, extack);
238	242	else if (state == DEVLINK_PORT_FN_STATE_INACTIVE)
239	243	err = mlx5_sf_deactivate(dev, sf);
240	244	else
@@ -265,7 +269,7 @@ int mlx5_devlink_sf_port_fn_state_set(struct devlink *devlink, struct devlink_po · 24 unmodified lines
265	269	goto out;
266	270	}
267	271
268		− err = mlx5_sf_state_set(dev, table, sf, state);
	272	+ err = mlx5_sf_state_set(dev, table, sf, state, extack);
269	273	out:
270	274	mlx5_sf_table_put(table);
271	275	return err;

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_fw.c+2−1

@@ -112,7 +112,8 @@ int mlx5dr_fw_create_md_tbl(struct mlx5dr_domain *dmn, · 111 unmodified lines
112	112	int ret;
113	113
114	114	ft_attr.table_type = MLX5_FLOW_TABLE_TYPE_FDB;
115		− ft_attr.level = dmn->info.caps.max_ft_level - 2;
	115	+ ft_attr.level = min_t(int, dmn->info.caps.max_ft_level - 2,
	116	+ MLX5_FT_MAX_MULTIPATH_LEVEL);
116	117	ft_attr.reformat_en = reformat_req;
117	118	ft_attr.decap_en = reformat_req;
118	119

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste_v1.c+16−10

@@ -694,7 +694,11 @@ static int dr_ste_v1_set_action_decap_l3_list(void *data, · 693 unmodified lines
694	694	if (hw_action_sz / DR_STE_ACTION_DOUBLE_SZ < DR_STE_DECAP_L3_ACTION_NUM)
695	695	return -EINVAL;
696	696
697		− memcpy(padded_data, data, data_sz);
	697	+ inline_data_sz =
	698	+ MLX5_FLD_SZ_BYTES(ste_double_action_insert_with_inline_v1, inline_data);
	699	+
	700	+ /* Add an alignment padding */
	701	+ memcpy(padded_data + data_sz % inline_data_sz, data, data_sz);
698	702
699	703	/* Remove L2L3 outer headers */
700	704	MLX5_SET(ste_single_action_remove_header_v1, hw_action, action_id,
@@ -706,32 +710,34 @@ static int dr_ste_v1_set_action_decap_l3_list(void *data, · 5 unmodified lines
706	710	hw_action += DR_STE_ACTION_DOUBLE_SZ;
707	711	used_actions++; /* Remove and NOP are a single double action */
708	712
709		− inline_data_sz =
710		− MLX5_FLD_SZ_BYTES(ste_double_action_insert_with_inline_v1, inline_data);
	713	+ /* Point to the last dword of the header */
	714	+ data_ptr += (data_sz / inline_data_sz) * inline_data_sz;
711	715
712		− /* Add the new header inline + 2 extra bytes */
	716	+ /* Add the new header using inline action 4Byte at a time, the header
	717	+ * is added in reversed order to the beginning of the packet to avoid
	718	+ * incorrect parsing by the HW. Since header is 14B or 18B an extra
	719	+ * two bytes are padded and later removed.
	720	+ */
713	721	for (i = 0; i < data_sz / inline_data_sz + 1; i++) {
714	722	void *addr_inline;
715	723
716	724	MLX5_SET(ste_double_action_insert_with_inline_v1, hw_action, action_id,
717	725	DR_STE_V1_ACTION_ID_INSERT_INLINE);
718	726	/* The hardware expects here offset to words (2 bytes) */
719		− MLX5_SET(ste_double_action_insert_with_inline_v1, hw_action, start_offset,
720		− i * 2);
	727	+ MLX5_SET(ste_double_action_insert_with_inline_v1, hw_action, start_offset, 0);
721	728
722	729	/* Copy bytes one by one to avoid endianness problem */
723	730	addr_inline = MLX5_ADDR_OF(ste_double_action_insert_with_inline_v1,
724	731	hw_action, inline_data);
725		− memcpy(addr_inline, data_ptr, inline_data_sz);
	732	+ memcpy(addr_inline, data_ptr - i * inline_data_sz, inline_data_sz);
726	733	hw_action += DR_STE_ACTION_DOUBLE_SZ;
727		− data_ptr += inline_data_sz;
728	734	used_actions++;
729	735	}
730	736
731		− /* Remove 2 extra bytes */
	737	+ /* Remove first 2 extra bytes */
732	738	MLX5_SET(ste_single_action_remove_header_size_v1, hw_action, action_id,
733	739	DR_STE_V1_ACTION_ID_REMOVE_BY_SIZE);
734		− MLX5_SET(ste_single_action_remove_header_size_v1, hw_action, start_offset, data_sz / 2);
	740	+ MLX5_SET(ste_single_action_remove_header_size_v1, hw_action, start_offset, 0);
735	741	/* The hardware expects here size in words (2 bytes) */
736	742	MLX5_SET(ste_single_action_remove_header_size_v1, hw_action, remove_size, 1);
737	743	used_actions++;

drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5dr.h+5−4

@@ -124,10 +124,11 @@ int mlx5dr_action_destroy(struct mlx5dr_action *action); · 123 unmodified lines
124	124	static inline bool
125	125	mlx5dr_is_supported(struct mlx5_core_dev *dev)
126	126	{
127		− return MLX5_CAP_ESW_FLOWTABLE_FDB(dev, sw_owner) \|\|
128		− (MLX5_CAP_ESW_FLOWTABLE_FDB(dev, sw_owner_v2) &&
129		− (MLX5_CAP_GEN(dev, steering_format_version) <=
130		− MLX5_STEERING_FORMAT_CONNECTX_6DX));
	127	+ return MLX5_CAP_GEN(dev, roce) &&
	128	+ (MLX5_CAP_ESW_FLOWTABLE_FDB(dev, sw_owner) \|\|
	129	+ (MLX5_CAP_ESW_FLOWTABLE_FDB(dev, sw_owner_v2) &&
	130	+ (MLX5_CAP_GEN(dev, steering_format_version) <=
	131	+ MLX5_STEERING_FORMAT_CONNECTX_6DX)));
131	132	}
132	133
133	134	/* buddy functions & structure */

drivers/net/ethernet/mellanox/mlx5/core/transobj.c+24−6

@@ -424,6 +424,15 @@ err_modify_sq: · 423 unmodified lines
424	424	return err;
425	425	}
426	426
	427	+static void mlx5_hairpin_unpair_peer_sq(struct mlx5_hairpin *hp)
	428	+{
	429	+ int i;
	430	+
	431	+ for (i = 0; i < hp->num_channels; i++)
	432	+ mlx5_hairpin_modify_sq(hp->peer_mdev, hp->sqn[i], MLX5_SQC_STATE_RDY,
	433	+ MLX5_SQC_STATE_RST, 0, 0);
	434	+}
	435	+
427	436	static void mlx5_hairpin_unpair_queues(struct mlx5_hairpin *hp)
428	437	{
429	438	int i;
@@ -432,13 +441,9 @@ static void mlx5_hairpin_unpair_queues(struct mlx5_hairpin *hp) · 2 unmodified lines
432	441	for (i = 0; i < hp->num_channels; i++)
433	442	mlx5_hairpin_modify_rq(hp->func_mdev, hp->rqn[i], MLX5_RQC_STATE_RDY,
434	443	MLX5_RQC_STATE_RST, 0, 0);
435		−
436	444	/* unset peer SQs */
437		− if (hp->peer_gone)
438		− return;
439		− for (i = 0; i < hp->num_channels; i++)
440		− mlx5_hairpin_modify_sq(hp->peer_mdev, hp->sqn[i], MLX5_SQC_STATE_RDY,
441		− MLX5_SQC_STATE_RST, 0, 0);
	445	+ if (!hp->peer_gone)
	446	+ mlx5_hairpin_unpair_peer_sq(hp);
442	447	}
443	448
444	449	struct mlx5_hairpin *
@@ -485,3 +490,16 @@ void mlx5_core_hairpin_destroy(struct mlx5_hairpin *hp) · 40 unmodified lines
485	490	mlx5_hairpin_destroy_queues(hp);
486	491	kfree(hp);
487	492	}
	493	+
	494	+void mlx5_core_hairpin_clear_dead_peer(struct mlx5_hairpin *hp)
	495	+{
	496	+ int i;
	497	+
	498	+ mlx5_hairpin_unpair_peer_sq(hp);
	499	+
	500	+ /* destroy peer SQ */
	501	+ for (i = 0; i < hp->num_channels; i++)
	502	+ mlx5_core_destroy_sq(hp->peer_mdev, hp->sqn[i]);
	503	+
	504	+ hp->peer_gone = true;
	505	+}

drivers/net/ethernet/mellanox/mlx5/core/vport.c+0−2

@@ -465,8 +465,6 @@ int mlx5_modify_nic_vport_node_guid(struct mlx5_core_dev *mdev, · 464 unmodified lines
465	465	void *in;
466	466	int err;
467	467
468		− if (!vport)
469		− return -EINVAL;
470	468	if (!MLX5_CAP_GEN(mdev, vport_group_manager))
471	469	return -EACCES;
472	470

drivers/net/ethernet/mellanox/mlxsw/core_thermal.c+4−2

@@ -693,7 +693,8 @@ mlxsw_thermal_module_tz_init(struct mlxsw_thermal_module *module_tz) · 692 unmodified lines
693	693	MLXSW_THERMAL_TRIP_MASK,
694	694	module_tz,
695	695	&mlxsw_thermal_module_ops,
696		− NULL, 0, 0);
	696	+ NULL, 0,
	697	+ module_tz->parent->polling_delay);
697	698	if (IS_ERR(module_tz->tzdev)) {
698	699	err = PTR_ERR(module_tz->tzdev);
699	700	return err;
@@ -815,7 +816,8 @@ mlxsw_thermal_gearbox_tz_init(struct mlxsw_thermal_module *gearbox_tz) · 115 unmodified lines
815	816	MLXSW_THERMAL_TRIP_MASK,
816	817	gearbox_tz,
817	818	&mlxsw_thermal_gearbox_ops,
818		− NULL, 0, 0);
	819	+ NULL, 0,
	820	+ gearbox_tz->parent->polling_delay);
819	821	if (IS_ERR(gearbox_tz->tzdev))
820	822	return PTR_ERR(gearbox_tz->tzdev);
821	823

drivers/net/ethernet/mellanox/mlxsw/reg.h+1−1

@@ -3907,7 +3907,7 @@ MLXSW_ITEM32(reg, qeec, max_shaper_bs, 0x1C, 0, 6); · 3906 unmodified lines
3907	3907	#define MLXSW_REG_QEEC_HIGHEST_SHAPER_BS 25
3908	3908	#define MLXSW_REG_QEEC_LOWEST_SHAPER_BS_SP1 5
3909	3909	#define MLXSW_REG_QEEC_LOWEST_SHAPER_BS_SP2 11
3910		−#define MLXSW_REG_QEEC_LOWEST_SHAPER_BS_SP3 5
	3910	+#define MLXSW_REG_QEEC_LOWEST_SHAPER_BS_SP3 11
3911	3911
3912	3912	static inline void mlxsw_reg_qeec_pack(char *payload, u8 local_port,
3913	3913	enum mlxsw_reg_qeec_hr hr, u8 index,

drivers/net/ethernet/mellanox/mlxsw/spectrum_qdisc.c+4−1

@@ -1332,6 +1332,7 @@ __mlxsw_sp_qdisc_ets_graft(struct mlxsw_sp_port *mlxsw_sp_port, · 1331 unmodified lines
1332	1332	u8 band, u32 child_handle)
1333	1333	{
1334	1334	struct mlxsw_sp_qdisc *old_qdisc;
	1335	+ u32 parent;
1335	1336
1336	1337	if (band < mlxsw_sp_qdisc->num_classes &&
1337	1338	mlxsw_sp_qdisc->qdiscs[band].handle == child_handle)
@@ -1352,7 +1353,9 @@ __mlxsw_sp_qdisc_ets_graft(struct mlxsw_sp_port *mlxsw_sp_port, · 14 unmodified lines
1352	1353	if (old_qdisc)
1353	1354	mlxsw_sp_qdisc_destroy(mlxsw_sp_port, old_qdisc);
1354	1355
1355		− mlxsw_sp_qdisc = mlxsw_sp_qdisc->ops->find_class(mlxsw_sp_qdisc, band);
	1356	+ parent = TC_H_MAKE(mlxsw_sp_qdisc->handle, band + 1);
	1357	+ mlxsw_sp_qdisc = mlxsw_sp_qdisc->ops->find_class(mlxsw_sp_qdisc,
	1358	+ parent);
1356	1359	if (!WARN_ON(!mlxsw_sp_qdisc))
1357	1360	mlxsw_sp_qdisc_destroy(mlxsw_sp_port, mlxsw_sp_qdisc);
1358	1361

drivers/net/ethernet/microchip/encx24j600.c+1−1

@@ -1,5 +1,5 @@
1	1	// SPDX-License-Identifier: GPL-2.0-or-later
2		−/**
	2	+/*
3	3	* Microchip ENCX24J600 ethernet driver
4	4	*
5	5	* Copyright (C) 2015 Gridpoint

drivers/net/ethernet/microchip/encx24j600_hw.h+1−1

@@ -1,5 +1,5 @@
1	1	/* SPDX-License-Identifier: GPL-2.0 */
2		−/**
	2	+/*
3	3	* encx24j600_hw.h: Register definitions
4	4	*
5	5	*/

drivers/net/ethernet/mscc/ocelot.c+5−0

@@ -379,6 +379,7 @@ static u32 ocelot_read_eq_avail(struct ocelot *ocelot, int port) · 378 unmodified lines
379	379
380	380	int ocelot_port_flush(struct ocelot *ocelot, int port)
381	381	{
	382	+ unsigned int pause_ena;
382	383	int err, val;
383	384
384	385	/* Disable dequeuing from the egress queues */
@@ -387,6 +388,7 @@ int ocelot_port_flush(struct ocelot *ocelot, int port) · 2 unmodified lines
387	388	QSYS_PORT_MODE, port);
388	389
389	390	/* Disable flow control */
	391	+ ocelot_fields_read(ocelot, port, SYS_PAUSE_CFG_PAUSE_ENA, &pause_ena);
390	392	ocelot_fields_write(ocelot, port, SYS_PAUSE_CFG_PAUSE_ENA, 0);
391	393
392	394	/* Disable priority flow control */
@@ -422,6 +424,9 @@ int ocelot_port_flush(struct ocelot *ocelot, int port) · 29 unmodified lines
422	424	/* Clear flushing again. */
423	425	ocelot_rmw_gix(ocelot, 0, REW_PORT_CFG_FLUSH_ENA, REW_PORT_CFG, port);
424	426
	427	+ /* Re-enable flow control */
	428	+ ocelot_fields_write(ocelot, port, SYS_PAUSE_CFG_PAUSE_ENA, pause_ena);
	429	+
425	430	return err;
426	431	}
427	432	EXPORT_SYMBOL(ocelot_port_flush);

drivers/net/ethernet/myricom/myri10ge/myri10ge.c+1−0

@@ -3815,6 +3815,7 @@ static int myri10ge_probe(struct pci_dev pdev, const struct pci_device_id ent) · 3814 unmodified lines
3815	3815	dev_err(&pdev->dev,
3816	3816	"invalid sram_size %dB or board span %ldB\n",
3817	3817	mgp->sram_size, mgp->board_span);
	3818	+ status = -EINVAL;
3818	3819	goto abort_with_ioremap;
3819	3820	}
3820	3821	memcpy_fromio(mgp->eeprom_strings,

drivers/net/ethernet/pensando/Kconfig+1−0

@@ -20,6 +20,7 @@ if NET_VENDOR_PENSANDO · 19 unmodified lines
20	20	config IONIC
21	21	tristate "Pensando Ethernet IONIC Support"
22	22	depends on 64BIT && PCI
	23	+ depends on PTP_1588_CLOCK \|\| !PTP_1588_CLOCK
23	24	select NET_DEVLINK
24	25	select DIMLIB
25	26	help

drivers/net/ethernet/qlogic/netxen/netxen_nic_main.c+2−0

@@ -1602,6 +1602,8 @@ err_out_free_netdev: · 1601 unmodified lines
1602	1602	free_netdev(netdev);
1603	1603
1604	1604	err_out_free_res:
	1605	+ if (NX_IS_REVISION_P3(pdev->revision))
	1606	+ pci_disable_pcie_error_reporting(pdev);
1605	1607	pci_release_regions(pdev);
1606	1608
1607	1609	err_out_disable_pdev:

drivers/net/ethernet/qlogic/qed/qed_dcbx.c+3−1

@@ -1266,9 +1266,11 @@ int qed_dcbx_get_config_params(struct qed_hwfn *p_hwfn, · 1265 unmodified lines
1266	1266	p_hwfn->p_dcbx_info->set.ver_num \|= DCBX_CONFIG_VERSION_STATIC;
1267	1267
1268	1268	p_hwfn->p_dcbx_info->set.enabled = dcbx_info->operational.enabled;
	1269	+ BUILD_BUG_ON(sizeof(dcbx_info->operational.params) !=
	1270	+ sizeof(p_hwfn->p_dcbx_info->set.config.params));
1269	1271	memcpy(&p_hwfn->p_dcbx_info->set.config.params,
1270	1272	&dcbx_info->operational.params,
1271		− sizeof(struct qed_dcbx_admin_params));
	1273	+ sizeof(p_hwfn->p_dcbx_info->set.config.params));
1272	1274	p_hwfn->p_dcbx_info->set.config.valid = true;
1273	1275
1274	1276	memcpy(params, &p_hwfn->p_dcbx_info->set, sizeof(struct qed_dcbx_set));

drivers/net/ethernet/qlogic/qla3xxx.c+1−1

@@ -114,7 +114,7 @@ static int ql_sem_spinlock(struct ql3_adapter *qdev, · 113 unmodified lines
114	114	value = readl(&port_regs->CommonRegs.semaphoreReg);
115	115	if ((value & (sem_mask >> 16)) == sem_bits)
116	116	return 0;
117		− ssleep(1);
	117	+ mdelay(1000);
118	118	} while (--seconds);
119	119	return -1;
120	120	}

drivers/net/ethernet/qlogic/qlcnic/qlcnic_main.c+1−0

@@ -2690,6 +2690,7 @@ err_out_free_hw_res: · 2689 unmodified lines
2690	2690	kfree(ahw);
2691	2691
2692	2692	err_out_free_res:
	2693	+ pci_disable_pcie_error_reporting(pdev);
2693	2694	pci_release_regions(pdev);
2694	2695
2695	2696	err_out_disable_pdev:

drivers/net/ethernet/qualcomm/rmnet/rmnet_vnd.c+9−9

@@ -126,24 +126,24 @@ static void rmnet_get_stats64(struct net_device *dev, · 125 unmodified lines
126	126	struct rtnl_link_stats64 *s)
127	127	{
128	128	struct rmnet_priv *priv = netdev_priv(dev);
129		− struct rmnet_vnd_stats total_stats;
	129	+ struct rmnet_vnd_stats total_stats = { };
130	130	struct rmnet_pcpu_stats *pcpu_ptr;
	131	+ struct rmnet_vnd_stats snapshot;
131	132	unsigned int cpu, start;
132	133
133		− memset(&total_stats, 0, sizeof(struct rmnet_vnd_stats));
134		−
135	134	for_each_possible_cpu(cpu) {
136	135	pcpu_ptr = per_cpu_ptr(priv->pcpu_stats, cpu);
137	136
138	137	do {
139	138	start = u64_stats_fetch_begin_irq(&pcpu_ptr->syncp);
140		− total_stats.rx_pkts += pcpu_ptr->stats.rx_pkts;
141		− total_stats.rx_bytes += pcpu_ptr->stats.rx_bytes;
142		− total_stats.tx_pkts += pcpu_ptr->stats.tx_pkts;
143		− total_stats.tx_bytes += pcpu_ptr->stats.tx_bytes;
	139	+ snapshot = pcpu_ptr->stats; /* struct assignment */
144	140	} while (u64_stats_fetch_retry_irq(&pcpu_ptr->syncp, start));
145	141
146		− total_stats.tx_drops += pcpu_ptr->stats.tx_drops;
	142	+ total_stats.rx_pkts += snapshot.rx_pkts;
	143	+ total_stats.rx_bytes += snapshot.rx_bytes;
	144	+ total_stats.tx_pkts += snapshot.tx_pkts;
	145	+ total_stats.tx_bytes += snapshot.tx_bytes;
	146	+ total_stats.tx_drops += snapshot.tx_drops;
147	147	}
148	148
149	149	s->rx_packets = total_stats.rx_pkts;
@@ -354,4 +354,4 @@ int rmnet_vnd_update_dev_mtu(struct rmnet_port *port, · 204 unmodified lines
354	354	}
355	355
356	356	return 0;
357		−}
	357	+}

drivers/net/ethernet/realtek/r8169_main.c+1−1

@@ -1671,7 +1671,7 @@ static void rtl8169_get_strings(struct net_device dev, u32 stringset, u8 data) · 1670 unmodified lines
1671	1671	{
1672	1672	switch(stringset) {
1673	1673	case ETH_SS_STATS:
1674		− memcpy(data, *rtl8169_gstrings, sizeof(rtl8169_gstrings));
	1674	+ memcpy(data, rtl8169_gstrings, sizeof(rtl8169_gstrings));
1675	1675	break;
1676	1676	}
1677	1677	}

drivers/net/ethernet/renesas/sh_eth.c+1−1

@@ -2287,7 +2287,7 @@ static void sh_eth_get_strings(struct net_device ndev, u32 stringset, u8 data) · 2286 unmodified lines
2287	2287	{
2288	2288	switch (stringset) {
2289	2289	case ETH_SS_STATS:
2290		− memcpy(data, *sh_eth_gstrings_stats,
	2290	+ memcpy(data, sh_eth_gstrings_stats,
2291	2291	sizeof(sh_eth_gstrings_stats));
2292	2292	break;
2293	2293	}

drivers/net/ethernet/sfc/nic.c+1−0

@@ -90,6 +90,7 @@ int efx_nic_init_interrupt(struct efx_nic *efx) · 89 unmodified lines
90	90	efx->pci_dev->irq);
91	91	goto fail1;
92	92	}
	93	+ efx->irqs_hooked = true;
93	94	return 0;
94	95	}
95	96

drivers/net/ethernet/stmicro/stmmac/dwmac1000.h+4−4

@@ -76,10 +76,10 @@ enum power_event { · 75 unmodified lines
76	76	#define LPI_CTRL_STATUS_TLPIEN 0x00000001 /* Transmit LPI Entry */
77	77
78	78	/* GMAC HW ADDR regs */
79		−#define GMAC_ADDR_HIGH(reg) (((reg > 15) ? 0x00000800 : 0x00000040) + \
80		− (reg * 8))
81		−#define GMAC_ADDR_LOW(reg) (((reg > 15) ? 0x00000804 : 0x00000044) + \
82		− (reg * 8))
	79	+#define GMAC_ADDR_HIGH(reg) ((reg > 15) ? 0x00000800 + (reg - 16) * 8 : \
	80	+ 0x00000040 + (reg * 8))
	81	+#define GMAC_ADDR_LOW(reg) ((reg > 15) ? 0x00000804 + (reg - 16) * 8 : \
	82	+ 0x00000044 + (reg * 8))
83	83	#define GMAC_MAX_PERFECT_ADDRESSES 1
84	84
85	85	#define GMAC_PCS_BASE 0x000000c0 /* PCS register base */

drivers/net/ethernet/stmicro/stmmac/stmmac_main.c+25−13

@@ -1196,7 +1196,6 @@ static void stmmac_check_pcs_mode(struct stmmac_priv *priv) · 1195 unmodified lines
1196	1196	*/
1197	1197	static int stmmac_init_phy(struct net_device *dev)
1198	1198	{
1199		− struct ethtool_wolinfo wol = { .cmd = ETHTOOL_GWOL };
1200	1199	struct stmmac_priv *priv = netdev_priv(dev);
1201	1200	struct device_node *node;
1202	1201	int ret;
@@ -1222,8 +1221,12 @@ static int stmmac_init_phy(struct net_device *dev) · 19 unmodified lines
1222	1221	ret = phylink_connect_phy(priv->phylink, phydev);
1223	1222	}
1224	1223
1225		− phylink_ethtool_get_wol(priv->phylink, &wol);
1226		− device_set_wakeup_capable(priv->device, !!wol.supported);
	1224	+ if (!priv->plat->pmt) {
	1225	+ struct ethtool_wolinfo wol = { .cmd = ETHTOOL_GWOL };
	1226	+
	1227	+ phylink_ethtool_get_wol(priv->phylink, &wol);
	1228	+ device_set_wakeup_capable(priv->device, !!wol.supported);
	1229	+ }
1227	1230
1228	1231	return ret;
1229	1232	}
@@ -1237,8 +1240,9 @@ static int stmmac_phy_setup(struct stmmac_priv *priv) · 7 unmodified lines
1237	1240	priv->phylink_config.dev = &priv->dev->dev;
1238	1241	priv->phylink_config.type = PHYLINK_NETDEV;
1239	1242	priv->phylink_config.pcs_poll = true;
1240		− priv->phylink_config.ovr_an_inband =
1241		− priv->plat->mdio_bus_data->xpcs_an_inband;
	1243	+ if (priv->plat->mdio_bus_data)
	1244	+ priv->phylink_config.ovr_an_inband =
	1245	+ priv->plat->mdio_bus_data->xpcs_an_inband;
1242	1246
1243	1247	if (!fwnode)
1244	1248	fwnode = dev_fwnode(priv->device);
@@ -5888,12 +5892,21 @@ static int stmmac_set_mac_address(struct net_device ndev, void addr) · 4643 unmodified lines
5888	5892	struct stmmac_priv *priv = netdev_priv(ndev);
5889	5893	int ret = 0;
5890	5894
	5895	+ ret = pm_runtime_get_sync(priv->device);
	5896	+ if (ret < 0) {
	5897	+ pm_runtime_put_noidle(priv->device);
	5898	+ return ret;
	5899	+ }
	5900	+
5891	5901	ret = eth_mac_addr(ndev, addr);
5892	5902	if (ret)
5893		− return ret;
	5903	+ goto set_mac_error;
5894	5904
5895	5905	stmmac_set_umac_addr(priv, priv->hw, ndev->dev_addr, 0);
5896	5906
	5907	+set_mac_error:
	5908	+ pm_runtime_put(priv->device);
	5909	+
5897	5910	return ret;
5898	5911	}
5899	5912
@@ -6188,12 +6201,6 @@ static int stmmac_vlan_rx_add_vid(struct net_device *ndev, __be16 proto, u16 vid · 288 unmodified lines
6188	6201	bool is_double = false;
6189	6202	int ret;
6190	6203
6191		− ret = pm_runtime_get_sync(priv->device);
6192		− if (ret < 0) {
6193		− pm_runtime_put_noidle(priv->device);
6194		− return ret;
6195		− }
6196		−
6197	6204	if (be16_to_cpu(proto) == ETH_P_8021AD)
6198	6205	is_double = true;
6199	6206
@@ -6219,6 +6226,12 @@ static int stmmac_vlan_rx_kill_vid(struct net_device *ndev, __be16 proto, u16 vi · 19 unmodified lines
6219	6226	bool is_double = false;
6220	6227	int ret;
6221	6228
	6229	+ ret = pm_runtime_get_sync(priv->device);
	6230	+ if (ret < 0) {
	6231	+ pm_runtime_put_noidle(priv->device);
	6232	+ return ret;
	6233	+ }
	6234	+
6222	6235	if (be16_to_cpu(proto) == ETH_P_8021AD)
6223	6236	is_double = true;
6224	6237
@@ -7036,7 +7049,6 @@ error_mdio_register: · 811 unmodified lines
7036	7049	stmmac_napi_del(ndev);
7037	7050	error_hw_init:
7038	7051	destroy_workqueue(priv->wq);
7039		− stmmac_bus_clks_config(priv, false);
7040	7052	bitmap_free(priv->af_xdp_zc_qps);
7041	7053
7042	7054	return ret;

drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c+2−0

@@ -622,6 +622,8 @@ error_pclk_get: · 621 unmodified lines
622	622	void stmmac_remove_config_dt(struct platform_device *pdev,
623	623	struct plat_stmmacenet_data *plat)
624	624	{
	625	+ clk_disable_unprepare(plat->stmmac_clk);
	626	+ clk_disable_unprepare(plat->pclk);
625	627	of_node_put(plat->phy_node);
626	628	of_node_put(plat->mdio_node);
627	629	}

drivers/net/ethernet/ti/netcp_core.c+2−2

@@ -1350,8 +1350,8 @@ int netcp_txpipe_open(struct netcp_tx_pipe *tx_pipe) · 1349 unmodified lines
1350	1350	tx_pipe->dma_queue = knav_queue_open(name, tx_pipe->dma_queue_id,
1351	1351	KNAV_QUEUE_SHARED);
1352	1352	if (IS_ERR(tx_pipe->dma_queue)) {
1353		− dev_err(dev, "Could not open DMA queue for channel \"%s\": %d\n",
1354		− name, ret);
	1353	+ dev_err(dev, "Could not open DMA queue for channel \"%s\": %pe\n",
	1354	+ name, tx_pipe->dma_queue);
1355	1355	ret = PTR_ERR(tx_pipe->dma_queue);
1356	1356	goto err;
1357	1357	}

drivers/net/ethernet/xilinx/ll_temac_main.c+24−3

@@ -774,12 +774,15 @@ static void temac_start_xmit_done(struct net_device *ndev) · 773 unmodified lines
774	774	stat = be32_to_cpu(cur_p->app0);
775	775
776	776	while (stat & STS_CTRL_APP0_CMPLT) {
	777	+ /* Make sure that the other fields are read after bd is
	778	+ * released by dma
	779	+ */
	780	+ rmb();
777	781	dma_unmap_single(ndev->dev.parent, be32_to_cpu(cur_p->phys),
778	782	be32_to_cpu(cur_p->len), DMA_TO_DEVICE);
779	783	skb = (struct sk_buff *)ptr_from_txbd(cur_p);
780	784	if (skb)
781	785	dev_consume_skb_irq(skb);
782		− cur_p->app0 = 0;
783	786	cur_p->app1 = 0;
784	787	cur_p->app2 = 0;
785	788	cur_p->app3 = 0;
@@ -788,6 +791,12 @@ static void temac_start_xmit_done(struct net_device *ndev) · 2 unmodified lines
788	791	ndev->stats.tx_packets++;
789	792	ndev->stats.tx_bytes += be32_to_cpu(cur_p->len);
790	793
	794	+ /* app0 must be visible last, as it is used to flag
	795	+ * availability of the bd
	796	+ */
	797	+ smp_mb();
	798	+ cur_p->app0 = 0;
	799	+
791	800	lp->tx_bd_ci++;
792	801	if (lp->tx_bd_ci >= lp->tx_bd_num)
793	802	lp->tx_bd_ci = 0;
@@ -814,6 +823,9 @@ static inline int temac_check_tx_bd_space(struct temac_local *lp, int num_frag) · 20 unmodified lines
814	823	if (cur_p->app0)
815	824	return NETDEV_TX_BUSY;
816	825
	826	+ /* Make sure to read next bd app0 after this one */
	827	+ rmb();
	828	+
817	829	tail++;
818	830	if (tail >= lp->tx_bd_num)
819	831	tail = 0;
@@ -849,7 +861,7 @@ temac_start_xmit(struct sk_buff skb, struct net_device ndev) · 29 unmodified lines
849	861	smp_mb();
850	862
851	863	/* Space might have just been freed - check again */
852		− if (temac_check_tx_bd_space(lp, num_frag))
	864	+ if (temac_check_tx_bd_space(lp, num_frag + 1))
853	865	return NETDEV_TX_BUSY;
854	866
855	867	netif_wake_queue(ndev);
@@ -876,7 +888,6 @@ temac_start_xmit(struct sk_buff skb, struct net_device ndev) · 20 unmodified lines
876	888	return NETDEV_TX_OK;
877	889	}
878	890	cur_p->phys = cpu_to_be32(skb_dma_addr);
879		− ptr_to_txbd((void *)skb, cur_p);
880	891
881	892	for (ii = 0; ii < num_frag; ii++) {
882	893	if (++lp->tx_bd_tail >= lp->tx_bd_num)
@@ -915,6 +926,11 @@ temac_start_xmit(struct sk_buff skb, struct net_device ndev) · 32 unmodified lines
915	926	}
916	927	cur_p->app0 \|= cpu_to_be32(STS_CTRL_APP0_EOP);
917	928
	929	+ /* Mark last fragment with skb address, so it can be consumed
	930	+ * in temac_start_xmit_done()
	931	+ */
	932	+ ptr_to_txbd((void *)skb, cur_p);
	933	+
918	934	tail_p = lp->tx_bd_p + sizeof(lp->tx_bd_v) lp->tx_bd_tail;
919	935	lp->tx_bd_tail++;
920	936	if (lp->tx_bd_tail >= lp->tx_bd_num)
@@ -926,6 +942,11 @@ temac_start_xmit(struct sk_buff skb, struct net_device ndev) · 5 unmodified lines
926	942	wmb();
927	943	lp->dma_out(lp, TX_TAILDESC_PTR, tail_p); /* DMA start */
928	944
	945	+ if (temac_check_tx_bd_space(lp, MAX_SKB_FRAGS + 1)) {
	946	+ netdev_info(ndev, "%s -> netif_stop_queue\n", __func__);
	947	+ netif_stop_queue(ndev);
	948	+ }
	949	+
929	950	return NETDEV_TX_OK;
930	951	}
931	952

drivers/net/hamradio/mkiss.c+1−0

@@ -799,6 +799,7 @@ static void mkiss_close(struct tty_struct *tty) · 798 unmodified lines
799	799	ax->tty = NULL;
800	800
801	801	unregister_netdev(ax->dev);
	802	+ free_netdev(ax->dev);
802	803	}
803	804
804	805	/* Perform I/O control on an active ax25 channel. */

drivers/net/ieee802154/mrf24j40.c+2−2

@@ -8,8 +8,8 @@ · 7 unmodified lines
8	8
9	9	#include <linux/spi/spi.h>
10	10	#include <linux/interrupt.h>
	11	+#include <linux/mod_devicetable.h>
11	12	#include <linux/module.h>
12		−#include <linux/of.h>
13	13	#include <linux/regmap.h>
14	14	#include <linux/ieee802154.h>
15	15	#include <linux/irq.h>
@@ -1388,7 +1388,7 @@ MODULE_DEVICE_TABLE(spi, mrf24j40_ids); · 1372 unmodified lines
1388	1388
1389	1389	static struct spi_driver mrf24j40_driver = {
1390	1390	.driver = {
1391		− .of_match_table = of_match_ptr(mrf24j40_of_match),
	1391	+ .of_match_table = mrf24j40_of_match,
1392	1392	.name = "mrf24j40",
1393	1393	},
1394	1394	.id_table = mrf24j40_ids,

drivers/net/ipa/ipa.h+2−0

@@ -58,6 +58,7 @@ enum ipa_flag { · 57 unmodified lines
58	58	* @mem_virt: Virtual address of IPA-local memory space
59	59	* @mem_offset: Offset from @mem_virt used for access to IPA memory
60	60	* @mem_size: Total size (bytes) of memory at @mem_virt
	61	+ * @mem_count: Number of entries in the mem array
61	62	* @mem: Array of IPA-local memory region descriptors
62	63	* @imem_iova: I/O virtual address of IPA region in IMEM
63	64	* @imem_size: Size of IMEM region
@@ -103,6 +104,7 @@ struct ipa { · 39 unmodified lines
103	104	void *mem_virt;
104	105	u32 mem_offset;
105	106	u32 mem_size;
	107	+ u32 mem_count;
106	108	const struct ipa_mem *mem;
107	109
108	110	unsigned long imem_iova;

drivers/net/ipa/ipa_mem.c+2−1

@@ -180,7 +180,7 @@ int ipa_mem_config(struct ipa *ipa) · 179 unmodified lines
180	180	* for the region, write "canary" values in the space prior to
181	181	* the region's base address.
182	182	*/
183		− for (mem_id = 0; mem_id < IPA_MEM_COUNT; mem_id++) {
	183	+ for (mem_id = 0; mem_id < ipa->mem_count; mem_id++) {
184	184	const struct ipa_mem *mem = &ipa->mem[mem_id];
185	185	u16 canary_count;
186	186	__le32 *canary;
@@ -487,6 +487,7 @@ int ipa_mem_init(struct ipa ipa, const struct ipa_mem_data mem_data) · 300 unmodified lines
487	487	ipa->mem_size = resource_size(res);
488	488
489	489	/* The ipa->mem[] array is indexed by enum ipa_mem_id values */
	490	+ ipa->mem_count = mem_data->local_count;
490	491	ipa->mem = mem_data->local;
491	492
492	493	ret = ipa_imem_init(ipa, mem_data->imem_addr, mem_data->imem_size);

drivers/net/mdio/mdio-octeon.c+0−2

@@ -71,7 +71,6 @@ static int octeon_mdiobus_probe(struct platform_device *pdev) · 70 unmodified lines
71	71
72	72	return 0;
73	73	fail_register:
74		− mdiobus_free(bus->mii_bus);
75	74	smi_en.u64 = 0;
76	75	oct_mdio_writeq(smi_en.u64, bus->register_base + SMI_EN);
77	76	return err;
@@ -85,7 +84,6 @@ static int octeon_mdiobus_remove(struct platform_device *pdev) · 7 unmodified lines
85	84	bus = platform_get_drvdata(pdev);
86	85
87	86	mdiobus_unregister(bus->mii_bus);
88		− mdiobus_free(bus->mii_bus);
89	87	smi_en.u64 = 0;
90	88	oct_mdio_writeq(smi_en.u64, bus->register_base + SMI_EN);
91	89	return 0;

drivers/net/mdio/mdio-thunder.c+0−1

@@ -126,7 +126,6 @@ static void thunder_mdiobus_pci_remove(struct pci_dev *pdev) · 125 unmodified lines
126	126	continue;
127	127
128	128	mdiobus_unregister(bus->mii_bus);
129		− mdiobus_free(bus->mii_bus);
130	129	oct_mdio_writeq(0, bus->register_base + SMI_EN);
131	130	}
132	131	pci_release_regions(pdev);

drivers/net/mhi/net.c+1−1

@@ -49,7 +49,7 @@ static int mhi_ndo_stop(struct net_device *ndev) · 48 unmodified lines
49	49	return 0;
50	50	}
51	51
52		−static int mhi_ndo_xmit(struct sk_buff skb, struct net_device ndev)
	52	+static netdev_tx_t mhi_ndo_xmit(struct sk_buff skb, struct net_device ndev)
53	53	{
54	54	struct mhi_net_dev *mhi_netdev = netdev_priv(ndev);
55	55	const struct mhi_net_proto *proto = mhi_netdev->proto;

drivers/net/phy/dp83867.c+1−5

@@ -826,16 +826,12 @@ static int dp83867_phy_reset(struct phy_device *phydev) · 825 unmodified lines
826	826	{
827	827	int err;
828	828
829		− err = phy_write(phydev, DP83867_CTRL, DP83867_SW_RESET);
	829	+ err = phy_write(phydev, DP83867_CTRL, DP83867_SW_RESTART);
830	830	if (err < 0)
831	831	return err;
832	832
833	833	usleep_range(10, 20);
834	834
835		− /* After reset FORCE_LINK_GOOD bit is set. Although the
836		− * default value should be unset. Disable FORCE_LINK_GOOD
837		− * for the phy to work properly.
838		− */
839	835	return phy_modify(phydev, MII_DP83867_PHYCTRL,
840	836	DP83867_PHYCR_FORCE_LINK_GOOD, 0);
841	837	}

drivers/net/phy/mdio_bus.c+2−1

@@ -607,7 +607,8 @@ void mdiobus_unregister(struct mii_bus *bus) · 606 unmodified lines
607	607	struct mdio_device *mdiodev;
608	608	int i;
609	609
610		− BUG_ON(bus->state != MDIOBUS_REGISTERED);
	610	+ if (WARN_ON_ONCE(bus->state != MDIOBUS_REGISTERED))
	611	+ return;
611	612	bus->state = MDIOBUS_UNREGISTERED;
612	613
613	614	for (i = 0; i < PHY_MAX_ADDR; i++) {

drivers/net/usb/cdc_eem.c+2−2

@@ -26,7 +26,7 @@ · 25 unmodified lines
26	26	* for transport over USB using a simpler USB device model than the
27	27	* previous CDC "Ethernet Control Model" (ECM, or "CDC Ethernet").
28	28	*
29		− * For details, see www.usb.org/developers/devclass_docs/CDC_EEM10.pdf
	29	+ * For details, see https://usb.org/sites/default/files/CDC_EEM10.pdf
30	30	*
31	31	* This version has been tested with GIGAntIC WuaoW SIM Smart Card on 2.6.24,
32	32	* 2.6.27 and 2.6.30rc2 kernel.
@@ -123,10 +123,10 @@ static struct sk_buff eem_tx_fixup(struct usbnet dev, struct sk_buff *skb, · 90 unmodified lines
123	123	}
124	124
125	125	skb2 = skb_copy_expand(skb, EEM_HEAD, ETH_FCS_LEN + padlen, flags);
	126	+ dev_kfree_skb_any(skb);
126	127	if (!skb2)
127	128	return NULL;
128	129
129		− dev_kfree_skb_any(skb);
130	130	skb = skb2;
131	131
132	132	done:

drivers/net/usb/cdc_ncm.c+1−1

@@ -1880,7 +1880,7 @@ static void cdc_ncm_status(struct usbnet dev, struct urb urb) · 1879 unmodified lines
1880	1880	static const struct driver_info cdc_ncm_info = {
1881	1881	.description = "CDC NCM",
1882	1882	.flags = FLAG_POINTTOPOINT \| FLAG_NO_SETINT \| FLAG_MULTI_PACKET
1883		− \| FLAG_LINK_INTR,
	1883	+ \| FLAG_LINK_INTR \| FLAG_ETHER,
1884	1884	.bind = cdc_ncm_bind,
1885	1885	.unbind = cdc_ncm_unbind,
1886	1886	.manage_power = usbnet_manage_power,

drivers/net/usb/hso.c+22−23

@@ -1689,7 +1689,7 @@ static int hso_serial_tiocmset(struct tty_struct *tty, · 1688 unmodified lines
1689	1689	spin_unlock_irqrestore(&serial->serial_lock, flags);
1690	1690
1691	1691	return usb_control_msg(serial->parent->usb,
1692		− usb_rcvctrlpipe(serial->parent->usb, 0), 0x22,
	1692	+ usb_sndctrlpipe(serial->parent->usb, 0), 0x22,
1693	1693	0x21, val, if_num, NULL, 0,
1694	1694	USB_CTRL_SET_TIMEOUT);
1695	1695	}
@@ -2436,7 +2436,7 @@ static int hso_rfkill_set_block(void *data, bool blocked) · 740 unmodified lines
2436	2436	if (hso_dev->usb_gone)
2437	2437	rv = 0;
2438	2438	else
2439		− rv = usb_control_msg(hso_dev->usb, usb_rcvctrlpipe(hso_dev->usb, 0),
	2439	+ rv = usb_control_msg(hso_dev->usb, usb_sndctrlpipe(hso_dev->usb, 0),
2440	2440	enabled ? 0x82 : 0x81, 0x40, 0, 0, NULL, 0,
2441	2441	USB_CTRL_SET_TIMEOUT);
2442	2442	mutex_unlock(&hso_dev->mutex);
@@ -2618,32 +2618,31 @@ static struct hso_device *hso_create_bulk_serial_device( · 175 unmodified lines
2618	2618	num_urbs = 2;
2619	2619	serial->tiocmget = kzalloc(sizeof(struct hso_tiocmget),
2620	2620	GFP_KERNEL);
	2621	+ if (!serial->tiocmget)
	2622	+ goto exit;
2621	2623	serial->tiocmget->serial_state_notification
2622	2624	= kzalloc(sizeof(struct hso_serial_state_notification),
2623	2625	GFP_KERNEL);
2624		− /* it isn't going to break our heart if serial->tiocmget
2625		− * allocation fails don't bother checking this.
2626		− */
2627		− if (serial->tiocmget && serial->tiocmget->serial_state_notification) {
2628		− tiocmget = serial->tiocmget;
2629		− tiocmget->endp = hso_get_ep(interface,
2630		− USB_ENDPOINT_XFER_INT,
2631		− USB_DIR_IN);
2632		− if (!tiocmget->endp) {
2633		− dev_err(&interface->dev, "Failed to find INT IN ep\n");
2634		− goto exit;
2635		− }
2636		−
2637		− tiocmget->urb = usb_alloc_urb(0, GFP_KERNEL);
2638		− if (tiocmget->urb) {
2639		− mutex_init(&tiocmget->mutex);
2640		− init_waitqueue_head(&tiocmget->waitq);
2641		− } else
2642		− hso_free_tiomget(serial);
	2626	+ if (!serial->tiocmget->serial_state_notification)
	2627	+ goto exit;
	2628	+ tiocmget = serial->tiocmget;
	2629	+ tiocmget->endp = hso_get_ep(interface,
	2630	+ USB_ENDPOINT_XFER_INT,
	2631	+ USB_DIR_IN);
	2632	+ if (!tiocmget->endp) {
	2633	+ dev_err(&interface->dev, "Failed to find INT IN ep\n");
	2634	+ goto exit;
2643	2635	}
2644		− }
2645		− else
	2636	+
	2637	+ tiocmget->urb = usb_alloc_urb(0, GFP_KERNEL);
	2638	+ if (!tiocmget->urb)
	2639	+ goto exit;
	2640	+
	2641	+ mutex_init(&tiocmget->mutex);
	2642	+ init_waitqueue_head(&tiocmget->waitq);
	2643	+ } else {
2646	2644	num_urbs = 1;
	2645	+ }
2647	2646
2648	2647	if (hso_serial_common_create(serial, num_urbs, BULK_URB_RX_SIZE,
2649	2648	BULK_URB_TX_SIZE))

drivers/net/usb/lan78xx.c+1−0

@@ -1645,6 +1645,7 @@ static const struct ethtool_ops lan78xx_ethtool_ops = { · 1644 unmodified lines
1645	1645	.get_strings = lan78xx_get_strings,
1646	1646	.get_wol = lan78xx_get_wol,
1647	1647	.set_wol = lan78xx_set_wol,
	1648	+ .get_ts_info = ethtool_op_get_ts_info,
1648	1649	.get_eee = lan78xx_get_eee,
1649	1650	.set_eee = lan78xx_set_eee,
1650	1651	.get_pauseparam = lan78xx_get_pause,

drivers/net/usb/qmi_wwan.c+1−1

@@ -575,7 +575,7 @@ static int qmi_wwan_rx_fixup(struct usbnet dev, struct sk_buff skb) · 574 unmodified lines
575	575
576	576	if (info->flags & QMI_WWAN_FLAG_PASS_THROUGH) {
577	577	skb->protocol = htons(ETH_P_MAP);
578		− return (netif_rx(skb) == NET_RX_SUCCESS);
	578	+ return 1;
579	579	}
580	580
581	581	switch (skb->data[0] & 0xf0) {

drivers/net/usb/r8152.c+38−6

@@ -8107,6 +8107,37 @@ static void r8156b_init(struct r8152 *tp) · 8106 unmodified lines
8107	8107	tp->coalesce = 15000; /* 15 us */
8108	8108	}
8109	8109
	8110	+static bool rtl_check_vendor_ok(struct usb_interface *intf)
	8111	+{
	8112	+ struct usb_host_interface *alt = intf->cur_altsetting;
	8113	+ struct usb_endpoint_descriptor in, out, *intr;
	8114	+
	8115	+ if (usb_find_common_endpoints(alt, &in, &out, &intr, NULL) < 0) {
	8116	+ dev_err(&intf->dev, "Expected endpoints are not found\n");
	8117	+ return false;
	8118	+ }
	8119	+
	8120	+ /* Check Rx endpoint address */
	8121	+ if (usb_endpoint_num(in) != 1) {
	8122	+ dev_err(&intf->dev, "Invalid Rx endpoint address\n");
	8123	+ return false;
	8124	+ }
	8125	+
	8126	+ /* Check Tx endpoint address */
	8127	+ if (usb_endpoint_num(out) != 2) {
	8128	+ dev_err(&intf->dev, "Invalid Tx endpoint address\n");
	8129	+ return false;
	8130	+ }
	8131	+
	8132	+ /* Check interrupt endpoint address */
	8133	+ if (usb_endpoint_num(intr) != 3) {
	8134	+ dev_err(&intf->dev, "Invalid interrupt endpoint address\n");
	8135	+ return false;
	8136	+ }
	8137	+
	8138	+ return true;
	8139	+}
	8140	+
8110	8141	static bool rtl_vendor_mode(struct usb_interface *intf)
8111	8142	{
8112	8143	struct usb_host_interface *alt = intf->cur_altsetting;
@@ -8115,12 +8146,15 @@ static bool rtl_vendor_mode(struct usb_interface *intf) · 2 unmodified lines
8115	8146	int i, num_configs;
8116	8147
8117	8148	if (alt->desc.bInterfaceClass == USB_CLASS_VENDOR_SPEC)
8118		− return true;
	8149	+ return rtl_check_vendor_ok(intf);
8119	8150
8120	8151	/* The vendor mode is not always config #1, so to find it out. */
8121	8152	udev = interface_to_usbdev(intf);
8122	8153	c = udev->config;
8123	8154	num_configs = udev->descriptor.bNumConfigurations;
	8155	+ if (num_configs < 2)
	8156	+ return false;
	8157	+
8124	8158	for (i = 0; i < num_configs; (i++, c++)) {
8125	8159	struct usb_interface_descriptor *desc = NULL;
8126	8160
@@ -8135,7 +8169,8 @@ static bool rtl_vendor_mode(struct usb_interface *intf) · 8 unmodified lines
8135	8169	}
8136	8170	}
8137	8171
8138		− WARN_ON_ONCE(i == num_configs);
	8172	+ if (i == num_configs)
	8173	+ dev_err(&intf->dev, "Unexpected Device\n");
8139	8174
8140	8175	return false;
8141	8176	}
@@ -8643,7 +8678,7 @@ static void rtl8152_get_strings(struct net_device dev, u32 stringset, u8 data) · 501 unmodified lines
8643	8678	{
8644	8679	switch (stringset) {
8645	8680	case ETH_SS_STATS:
8646		− memcpy(data, *rtl8152_gstrings, sizeof(rtl8152_gstrings));
	8681	+ memcpy(data, rtl8152_gstrings, sizeof(rtl8152_gstrings));
8647	8682	break;
8648	8683	}
8649	8684	}
@@ -9381,9 +9416,6 @@ static int rtl8152_probe(struct usb_interface *intf, · 731 unmodified lines
9381	9416	if (!rtl_vendor_mode(intf))
9382	9417	return -ENODEV;
9383	9418
9384		− if (intf->cur_altsetting->desc.bNumEndpoints < 3)
9385		− return -ENODEV;
9386		−
9387	9419	usb_reset_device(udev);
9388	9420	netdev = alloc_etherdev(sizeof(struct r8152));
9389	9421	if (!netdev) {

drivers/net/usb/smsc75xx.c+9−3

@@ -1483,7 +1483,7 @@ static int smsc75xx_bind(struct usbnet dev, struct usb_interface intf) · 1482 unmodified lines
1483	1483	ret = smsc75xx_wait_ready(dev, 0);
1484	1484	if (ret < 0) {
1485	1485	netdev_warn(dev->net, "device not ready in smsc75xx_bind\n");
1486		− return ret;
	1486	+ goto free_pdata;
1487	1487	}
1488	1488
1489	1489	smsc75xx_init_mac_address(dev);
@@ -1492,7 +1492,7 @@ static int smsc75xx_bind(struct usbnet dev, struct usb_interface intf) · 2 unmodified lines
1492	1492	ret = smsc75xx_reset(dev);
1493	1493	if (ret < 0) {
1494	1494	netdev_warn(dev->net, "smsc75xx_reset error %d\n", ret);
1495		− return ret;
	1495	+ goto cancel_work;
1496	1496	}
1497	1497
1498	1498	dev->net->netdev_ops = &smsc75xx_netdev_ops;
@@ -1502,6 +1502,13 @@ static int smsc75xx_bind(struct usbnet dev, struct usb_interface intf) · 3 unmodified lines
1502	1502	dev->hard_mtu = dev->net->mtu + dev->net->hard_header_len;
1503	1503	dev->net->max_mtu = MAX_SINGLE_PACKET_SIZE;
1504	1504	return 0;
	1505	+
	1506	+cancel_work:
	1507	+ cancel_work_sync(&pdata->set_multicast);
	1508	+free_pdata:
	1509	+ kfree(pdata);
	1510	+ dev->data[0] = 0;
	1511	+ return ret;
1505	1512	}
1506	1513
1507	1514	static void smsc75xx_unbind(struct usbnet dev, struct usb_interface intf)
@@ -1511,7 +1518,6 @@ static void smsc75xx_unbind(struct usbnet dev, struct usb_interface intf) · 3 unmodified lines
1511	1518	cancel_work_sync(&pdata->set_multicast);
1512	1519	netif_dbg(dev, ifdown, dev->net, "free pdata\n");
1513	1520	kfree(pdata);
1514		− pdata = NULL;
1515	1521	dev->data[0] = 0;
1516	1522	}
1517	1523	}

drivers/net/virtio_net.c+8−12

@@ -401,18 +401,13 @@ static struct sk_buff page_to_skb(struct virtnet_info vi, · 400 unmodified lines
401	401	/* If headroom is not 0, there is an offset between the beginning of the
402	402	* data and the allocated space, otherwise the data and the allocated
403	403	* space are aligned.
	404	+ *
	405	+ * Buffers with headroom use PAGE_SIZE as alloc size, see
	406	+ * add_recvbuf_mergeable() + get_mergeable_buf_len()
404	407	*/
405		− if (headroom) {
406		− /* Buffers with headroom use PAGE_SIZE as alloc size,
407		− * see add_recvbuf_mergeable() + get_mergeable_buf_len()
408		− */
409		− truesize = PAGE_SIZE;
410		− tailroom = truesize - len - offset;
411		− buf = page_address(page);
412		− } else {
413		− tailroom = truesize - len;
414		− buf = p;
415		− }
	408	+ truesize = headroom ? PAGE_SIZE : truesize;
	409	+ tailroom = truesize - len - headroom - (hdr_padded_len - hdr_len);
	410	+ buf = p - headroom;
416	411
417	412	len -= hdr_len;
418	413	offset += hdr_padded_len;
@@ -958,7 +953,8 @@ static struct sk_buff receive_mergeable(struct net_device dev, · 539 unmodified lines
958	953	put_page(page);
959	954	head_skb = page_to_skb(vi, rq, xdp_page, offset,
960	955	len, PAGE_SIZE, false,
961		− metasize, headroom);
	956	+ metasize,
	957	+ VIRTIO_XDP_HEADROOM);
962	958	return head_skb;
963	959	}
964	960	break;

drivers/net/vrf.c+2−4

@@ -1183,9 +1183,6 @@ static int vrf_dev_init(struct net_device *dev) · 1182 unmodified lines
1183	1183
1184	1184	dev->flags = IFF_MASTER \| IFF_NOARP;
1185	1185
1186		− /* MTU is irrelevant for VRF device; set to 64k similar to lo */
1187		− dev->mtu = 64 * 1024;
1188		−
1189	1186	/* similarly, oper state is irrelevant; set to up to avoid confusion */
1190	1187	dev->operstate = IF_OPER_UP;
1191	1188	netdev_lockdep_set_classes(dev);
@@ -1685,7 +1682,8 @@ static void vrf_setup(struct net_device *dev) · 493 unmodified lines
1685	1682	* which breaks networking.
1686	1683	*/
1687	1684	dev->min_mtu = IPV6_MIN_MTU;
1688		− dev->max_mtu = ETH_MAX_MTU;
	1685	+ dev->max_mtu = IP6_MAX_MTU;
	1686	+ dev->mtu = dev->max_mtu;
1689	1687	}
1690	1688
1691	1689	static int vrf_validate(struct nlattr tb[], struct nlattr data[],

drivers/net/wireguard/Makefile+1−2

@@ -1,5 +1,4 @@
1		−ccflags-y := -O3
2		−ccflags-y += -D'pr_fmt(fmt)=KBUILD_MODNAME ": " fmt'
	1	+ccflags-y := -D'pr_fmt(fmt)=KBUILD_MODNAME ": " fmt'
3	2	ccflags-$(CONFIG_WIREGUARD_DEBUG) += -DDEBUG
4	3	wireguard-y := main.o
5	4	wireguard-y += noise.o

drivers/net/wireguard/allowedips.c+99−90

@@ -6,6 +6,8 @@ · 5 unmodified lines
6	6	#include "allowedips.h"
7	7	#include "peer.h"
8	8
	9	+static struct kmem_cache *node_cache;
	10	+
9	11	static void swap_endian(u8 dst, const u8 src, u8 bits)
10	12	{
11	13	if (bits == 32) {
@@ -28,8 +30,11 @@ static void copy_and_assign_cidr(struct allowedips_node node, const u8 src, · 16 unmodified lines
28	30	node->bitlen = bits;
29	31	memcpy(node->bits, src, bits / 8U);
30	32	}
31		−#define CHOOSE_NODE(parent, key) \
32		− parent->bit[(key[parent->bit_at_a] >> parent->bit_at_b) & 1]
	33	+
	34	+static inline u8 choose(struct allowedips_node node, const u8 key)
	35	+{
	36	+ return (key[node->bit_at_a] >> node->bit_at_b) & 1;
	37	+}
33	38
34	39	static void push_rcu(struct allowedips_node **stack,
35	40	struct allowedips_node __rcu p, unsigned int len)
@@ -40,6 +45,11 @@ static void push_rcu(struct allowedips_node **stack, · 4 unmodified lines
40	45	}
41	46	}
42	47
	48	+static void node_free_rcu(struct rcu_head *rcu)
	49	+{
	50	+ kmem_cache_free(node_cache, container_of(rcu, struct allowedips_node, rcu));
	51	+}
	52	+
43	53	static void root_free_rcu(struct rcu_head *rcu)
44	54	{
45	55	struct allowedips_node node, stack[128] = {
@@ -49,7 +59,7 @@ static void root_free_rcu(struct rcu_head *rcu) · 3 unmodified lines
49	59	while (len > 0 && (node = stack[--len])) {
50	60	push_rcu(stack, node->bit[0], &len);
51	61	push_rcu(stack, node->bit[1], &len);
52		− kfree(node);
	62	+ kmem_cache_free(node_cache, node);
53	63	}
54	64	}
55	65
@@ -66,60 +76,6 @@ static void root_remove_peer_lists(struct allowedips_node *root) · 10 unmodified lines
66	76	}
67	77	}
68	78
69		−static void walk_remove_by_peer(struct allowedips_node __rcu **top,
70		− struct wg_peer peer, struct mutex lock)
71		−{
72		−#define REF(p) rcu_access_pointer(p)
73		−#define DEREF(p) rcu_dereference_protected(*(p), lockdep_is_held(lock))
74		−#define PUSH(p) ({ \
75		− WARN_ON(IS_ENABLED(DEBUG) && len >= 128); \
76		− stack[len++] = p; \
77		− })
78		−
79		− struct allowedips_node __rcu stack[128], nptr;
80		− struct allowedips_node node, prev;
81		− unsigned int len;
82		−
83		− if (unlikely(!peer \|\| !REF(*top)))
84		− return;
85		−
86		− for (prev = NULL, len = 0, PUSH(top); len > 0; prev = node) {
87		− nptr = stack[len - 1];
88		− node = DEREF(nptr);
89		− if (!node) {
90		− --len;
91		− continue;
92		− }
93		− if (!prev \|\| REF(prev->bit[0]) == node \|\|
94		− REF(prev->bit[1]) == node) {
95		− if (REF(node->bit[0]))
96		− PUSH(&node->bit[0]);
97		− else if (REF(node->bit[1]))
98		− PUSH(&node->bit[1]);
99		− } else if (REF(node->bit[0]) == prev) {
100		− if (REF(node->bit[1]))
101		− PUSH(&node->bit[1]);
102		− } else {
103		− if (rcu_dereference_protected(node->peer,
104		− lockdep_is_held(lock)) == peer) {
105		− RCU_INIT_POINTER(node->peer, NULL);
106		− list_del_init(&node->peer_list);
107		− if (!node->bit[0] \|\| !node->bit[1]) {
108		− rcu_assign_pointer(*nptr, DEREF(
109		− &node->bit[!REF(node->bit[0])]));
110		− kfree_rcu(node, rcu);
111		− node = DEREF(nptr);
112		− }
113		− }
114		− --len;
115		− }
116		− }
117		−
118		−#undef REF
119		−#undef DEREF
120		−#undef PUSH
121		−}
122		−
123	79	static unsigned int fls128(u64 a, u64 b)
124	80	{
125	81	return a ? fls64(a) + 64U : fls64(b);
@@ -159,7 +115,7 @@ static struct allowedips_node find_node(struct allowedips_node trie, u8 bits, · 33 unmodified lines
159	115	found = node;
160	116	if (node->cidr == bits)
161	117	break;
162		− node = rcu_dereference_bh(CHOOSE_NODE(node, key));
	118	+ node = rcu_dereference_bh(node->bit[choose(node, key)]);
163	119	}
164	120	return found;
165	121	}
@@ -191,8 +147,7 @@ static bool node_placement(struct allowedips_node __rcu trie, const u8 key, · 25 unmodified lines
191	147	u8 cidr, u8 bits, struct allowedips_node **rnode,
192	148	struct mutex *lock)
193	149	{
194		− struct allowedips_node *node = rcu_dereference_protected(trie,
195		− lockdep_is_held(lock));
	150	+ struct allowedips_node *node = rcu_dereference_protected(trie, lockdep_is_held(lock));
196	151	struct allowedips_node *parent = NULL;
197	152	bool exact = false;
198	153
@@ -202,13 +157,24 @@ static bool node_placement(struct allowedips_node __rcu trie, const u8 key, · 3 unmodified lines
202	157	exact = true;
203	158	break;
204	159	}
205		− node = rcu_dereference_protected(CHOOSE_NODE(parent, key),
206		− lockdep_is_held(lock));
	160	+ node = rcu_dereference_protected(parent->bit[choose(parent, key)], lockdep_is_held(lock));
207	161	}
208	162	*rnode = parent;
209	163	return exact;
210	164	}
211	165
	166	+static inline void connect_node(struct allowedips_node *parent, u8 bit, struct allowedips_node node)
	167	+{
	168	+ node->parent_bit_packed = (unsigned long)parent \| bit;
	169	+ rcu_assign_pointer(*parent, node);
	170	+}
	171	+
	172	+static inline void choose_and_connect_node(struct allowedips_node parent, struct allowedips_node node)
	173	+{
	174	+ u8 bit = choose(parent, node->bits);
	175	+ connect_node(&parent->bit[bit], bit, node);
	176	+}
	177	+
212	178	static int add(struct allowedips_node __rcu *trie, u8 bits, const u8 key,
213	179	u8 cidr, struct wg_peer peer, struct mutex lock)
214	180	{
@@ -218,13 +184,13 @@ static int add(struct allowedips_node __rcu *trie, u8 bits, const u8 key, · 3 unmodified lines
218	184	return -EINVAL;
219	185
220	186	if (!rcu_access_pointer(*trie)) {
221		− node = kzalloc(sizeof(*node), GFP_KERNEL);
	187	+ node = kmem_cache_zalloc(node_cache, GFP_KERNEL);
222	188	if (unlikely(!node))
223	189	return -ENOMEM;
224	190	RCU_INIT_POINTER(node->peer, peer);
225	191	list_add_tail(&node->peer_list, &peer->allowedips_list);
226	192	copy_and_assign_cidr(node, key, cidr, bits);
227		− rcu_assign_pointer(*trie, node);
	193	+ connect_node(trie, 2, node);
228	194	return 0;
229	195	}
230	196	if (node_placement(*trie, key, cidr, bits, &node, lock)) {
@@ -233,7 +199,7 @@ static int add(struct allowedips_node __rcu *trie, u8 bits, const u8 key, · 2 unmodified lines
233	199	return 0;
234	200	}
235	201
236		− newnode = kzalloc(sizeof(*newnode), GFP_KERNEL);
	202	+ newnode = kmem_cache_zalloc(node_cache, GFP_KERNEL);
237	203	if (unlikely(!newnode))
238	204	return -ENOMEM;
239	205	RCU_INIT_POINTER(newnode->peer, peer);
@@ -243,10 +209,10 @@ static int add(struct allowedips_node __rcu *trie, u8 bits, const u8 key, · 3 unmodified lines
243	209	if (!node) {
244	210	down = rcu_dereference_protected(*trie, lockdep_is_held(lock));
245	211	} else {
246		− down = rcu_dereference_protected(CHOOSE_NODE(node, key),
247		− lockdep_is_held(lock));
	212	+ const u8 bit = choose(node, key);
	213	+ down = rcu_dereference_protected(node->bit[bit], lockdep_is_held(lock));
248	214	if (!down) {
249		− rcu_assign_pointer(CHOOSE_NODE(node, key), newnode);
	215	+ connect_node(&node->bit[bit], bit, newnode);
250	216	return 0;
251	217	}
252	218	}
@@ -254,30 +220,29 @@ static int add(struct allowedips_node __rcu *trie, u8 bits, const u8 key, · 1 unmodified line
254	220	parent = node;
255	221
256	222	if (newnode->cidr == cidr) {
257		− rcu_assign_pointer(CHOOSE_NODE(newnode, down->bits), down);
	223	+ choose_and_connect_node(newnode, down);
258	224	if (!parent)
259		− rcu_assign_pointer(*trie, newnode);
	225	+ connect_node(trie, 2, newnode);
260	226	else
261		− rcu_assign_pointer(CHOOSE_NODE(parent, newnode->bits),
262		− newnode);
263		− } else {
264		− node = kzalloc(sizeof(*node), GFP_KERNEL);
265		− if (unlikely(!node)) {
266		− list_del(&newnode->peer_list);
267		− kfree(newnode);
268		− return -ENOMEM;
269		− }
270		− INIT_LIST_HEAD(&node->peer_list);
271		− copy_and_assign_cidr(node, newnode->bits, cidr, bits);
	227	+ choose_and_connect_node(parent, newnode);
	228	+ return 0;
	229	+ }
272	230
273		− rcu_assign_pointer(CHOOSE_NODE(node, down->bits), down);
274		− rcu_assign_pointer(CHOOSE_NODE(node, newnode->bits), newnode);
275		− if (!parent)
276		− rcu_assign_pointer(*trie, node);
277		− else
278		− rcu_assign_pointer(CHOOSE_NODE(parent, node->bits),
279		− node);
	231	+ node = kmem_cache_zalloc(node_cache, GFP_KERNEL);
	232	+ if (unlikely(!node)) {
	233	+ list_del(&newnode->peer_list);
	234	+ kmem_cache_free(node_cache, newnode);
	235	+ return -ENOMEM;
280	236	}
	237	+ INIT_LIST_HEAD(&node->peer_list);
	238	+ copy_and_assign_cidr(node, newnode->bits, cidr, bits);
	239	+
	240	+ choose_and_connect_node(node, down);
	241	+ choose_and_connect_node(node, newnode);
	242	+ if (!parent)
	243	+ connect_node(trie, 2, node);
	244	+ else
	245	+ choose_and_connect_node(parent, node);
281	246	return 0;
282	247	}
283	248
@@ -335,9 +300,41 @@ int wg_allowedips_insert_v6(struct allowedips table, const struct in6_addr ip, · 51 unmodified lines
335	300	void wg_allowedips_remove_by_peer(struct allowedips *table,
336	301	struct wg_peer peer, struct mutex lock)
337	302	{
	303	+ struct allowedips_node node, child, *parent_bit, parent, *tmp;
	304	+ bool free_parent;
	305	+
	306	+ if (list_empty(&peer->allowedips_list))
	307	+ return;
338	308	++table->seq;
339		− walk_remove_by_peer(&table->root4, peer, lock);
340		− walk_remove_by_peer(&table->root6, peer, lock);
	309	+ list_for_each_entry_safe(node, tmp, &peer->allowedips_list, peer_list) {
	310	+ list_del_init(&node->peer_list);
	311	+ RCU_INIT_POINTER(node->peer, NULL);
	312	+ if (node->bit[0] && node->bit[1])
	313	+ continue;
	314	+ child = rcu_dereference_protected(node->bit[!rcu_access_pointer(node->bit[0])],
	315	+ lockdep_is_held(lock));
	316	+ if (child)
	317	+ child->parent_bit_packed = node->parent_bit_packed;
	318	+ parent_bit = (struct allowedips_node **)(node->parent_bit_packed & ~3UL);
	319	+ *parent_bit = child;
	320	+ parent = (void *)parent_bit -
	321	+ offsetof(struct allowedips_node, bit[node->parent_bit_packed & 1]);
	322	+ free_parent = !rcu_access_pointer(node->bit[0]) &&
	323	+ !rcu_access_pointer(node->bit[1]) &&
	324	+ (node->parent_bit_packed & 3) <= 1 &&
	325	+ !rcu_access_pointer(parent->peer);
	326	+ if (free_parent)
	327	+ child = rcu_dereference_protected(
	328	+ parent->bit[!(node->parent_bit_packed & 1)],
	329	+ lockdep_is_held(lock));
	330	+ call_rcu(&node->rcu, node_free_rcu);
	331	+ if (!free_parent)
	332	+ continue;
	333	+ if (child)
	334	+ child->parent_bit_packed = parent->parent_bit_packed;
	335	+ (struct allowedips_node *)(parent->parent_bit_packed & ~3UL) = child;
	336	+ call_rcu(&parent->rcu, node_free_rcu);
	337	+ }
341	338	}
342	339
343	340	int wg_allowedips_read_node(struct allowedips_node node, u8 ip[16], u8 cidr)
@@ -374,4 +371,16 @@ struct wg_peer wg_allowedips_lookup_src(struct allowedips table, · 30 unmodified lines
374	371	return NULL;
375	372	}
376	373
	374	+int __init wg_allowedips_slab_init(void)
	375	+{
	376	+ node_cache = KMEM_CACHE(allowedips_node, 0);
	377	+ return node_cache ? 0 : -ENOMEM;
	378	+}
	379	+
	380	+void wg_allowedips_slab_uninit(void)
	381	+{
	382	+ rcu_barrier();
	383	+ kmem_cache_destroy(node_cache);
	384	+}
	385	+
377	386	#include "selftest/allowedips.c"

drivers/net/wireguard/allowedips.h+7−7

@@ -15,14 +15,11 @@ struct wg_peer; · 14 unmodified lines
15	15	struct allowedips_node {
16	16	struct wg_peer __rcu *peer;
17	17	struct allowedips_node __rcu *bit[2];
18		− /* While it may seem scandalous that we waste space for v4,
19		− * we're alloc'ing to the nearest power of 2 anyway, so this
20		− * doesn't actually make a difference.
21		− */
22		− u8 bits[16] __aligned(__alignof(u64));
23	18	u8 cidr, bit_at_a, bit_at_b, bitlen;
	19	+ u8 bits[16] __aligned(__alignof(u64));
24	20
25		− /* Keep rarely used list at bottom to be beyond cache line. */
	21	+ /* Keep rarely used members at bottom to be beyond cache line. */
	22	+ unsigned long parent_bit_packed;
26	23	union {
27	24	struct list_head peer_list;
28	25	struct rcu_head rcu;
@@ -33,7 +30,7 @@ struct allowedips { · 4 unmodified lines
33	30	struct allowedips_node __rcu *root4;
34	31	struct allowedips_node __rcu *root6;
35	32	u64 seq;
36		−};
	33	+} __aligned(4); /* We pack the lower 2 bits of &root, but m68k only gives 16-bit alignment. */
37	34
38	35	void wg_allowedips_init(struct allowedips *table);
39	36	void wg_allowedips_free(struct allowedips table, struct mutex mutex);
@@ -56,4 +53,7 @@ struct wg_peer wg_allowedips_lookup_src(struct allowedips table, · 16 unmodified lines
56	53	bool wg_allowedips_selftest(void);
57	54	#endif
58	55
	56	+int wg_allowedips_slab_init(void);
	57	+void wg_allowedips_slab_uninit(void);
	58	+
59	59	#endif /* _WG_ALLOWEDIPS_H */

drivers/net/wireguard/main.c+16−1

@@ -21,13 +21,22 @@ static int __init mod_init(void) · 20 unmodified lines
21	21	{
22	22	int ret;
23	23
	24	+ ret = wg_allowedips_slab_init();
	25	+ if (ret < 0)
	26	+ goto err_allowedips;
	27	+
24	28	#ifdef DEBUG
	29	+ ret = -ENOTRECOVERABLE;
25	30	if (!wg_allowedips_selftest() \|\| !wg_packet_counter_selftest() \|\|
26	31	!wg_ratelimiter_selftest())
27		− return -ENOTRECOVERABLE;
	32	+ goto err_peer;
28	33	#endif
29	34	wg_noise_init();
30	35
	36	+ ret = wg_peer_init();
	37	+ if (ret < 0)
	38	+ goto err_peer;
	39	+
31	40	ret = wg_device_init();
32	41	if (ret < 0)
33	42	goto err_device;
@@ -44,6 +53,10 @@ static int __init mod_init(void) · 10 unmodified lines
44	53	err_netlink:
45	54	wg_device_uninit();
46	55	err_device:
	56	+ wg_peer_uninit();
	57	+err_peer:
	58	+ wg_allowedips_slab_uninit();
	59	+err_allowedips:
47	60	return ret;
48	61	}
49	62
@@ -51,6 +64,8 @@ static void __exit mod_exit(void) · 1 unmodified line
51	64	{
52	65	wg_genetlink_uninit();
53	66	wg_device_uninit();
	67	+ wg_peer_uninit();
	68	+ wg_allowedips_slab_uninit();
54	69	}
55	70
56	71	module_init(mod_init);

drivers/net/wireguard/peer.c+20−7

@@ -15,6 +15,7 @@ · 14 unmodified lines
15	15	#include <linux/rcupdate.h>
16	16	#include <linux/list.h>
17	17
	18	+static struct kmem_cache *peer_cache;
18	19	static atomic64_t peer_counter = ATOMIC64_INIT(0);
19	20
20	21	struct wg_peer wg_peer_create(struct wg_device wg,
@@ -29,10 +30,10 @@ struct wg_peer wg_peer_create(struct wg_device wg, · 8 unmodified lines
29	30	if (wg->num_peers >= MAX_PEERS_PER_DEVICE)
30	31	return ERR_PTR(ret);
31	32
32		− peer = kzalloc(sizeof(*peer), GFP_KERNEL);
	33	+ peer = kmem_cache_zalloc(peer_cache, GFP_KERNEL);
33	34	if (unlikely(!peer))
34	35	return ERR_PTR(ret);
35		− if (dst_cache_init(&peer->endpoint_cache, GFP_KERNEL))
	36	+ if (unlikely(dst_cache_init(&peer->endpoint_cache, GFP_KERNEL)))
36	37	goto err;
37	38
38	39	peer->device = wg;
@@ -64,7 +65,7 @@ struct wg_peer wg_peer_create(struct wg_device wg, · 25 unmodified lines
64	65	return peer;
65	66
66	67	err:
67		− kfree(peer);
	68	+ kmem_cache_free(peer_cache, peer);
68	69	return ERR_PTR(ret);
69	70	}
70	71
@@ -88,7 +89,7 @@ static void peer_make_dead(struct wg_peer *peer) · 17 unmodified lines
88	89	/* Mark as dead, so that we don't allow jumping contexts after. */
89	90	WRITE_ONCE(peer->is_dead, true);
90	91
91		− /* The caller must now synchronize_rcu() for this to take effect. */
	92	+ /* The caller must now synchronize_net() for this to take effect. */
92	93	}
93	94
94	95	static void peer_remove_after_dead(struct wg_peer *peer)
@@ -160,7 +161,7 @@ void wg_peer_remove(struct wg_peer *peer) · 65 unmodified lines
160	161	lockdep_assert_held(&peer->device->device_update_lock);
161	162
162	163	peer_make_dead(peer);
163		− synchronize_rcu();
	164	+ synchronize_net();
164	165	peer_remove_after_dead(peer);
165	166	}
166	167
@@ -178,7 +179,7 @@ void wg_peer_remove_all(struct wg_device *wg) · 11 unmodified lines
178	179	peer_make_dead(peer);
179	180	list_add_tail(&peer->peer_list, &dead_peers);
180	181	}
181		− synchronize_rcu();
	182	+ synchronize_net();
182	183	list_for_each_entry_safe(peer, temp, &dead_peers, peer_list)
183	184	peer_remove_after_dead(peer);
184	185	}
@@ -193,7 +194,8 @@ static void rcu_release(struct rcu_head *rcu) · 8 unmodified lines
193	194	/* The final zeroing takes care of clearing any remaining handshake key
194	195	* material and other potentially sensitive information.
195	196	*/
196		− kfree_sensitive(peer);
	197	+ memzero_explicit(peer, sizeof(*peer));
	198	+ kmem_cache_free(peer_cache, peer);
197	199	}
198	200
199	201	static void kref_release(struct kref *refcount)
@@ -225,3 +227,14 @@ void wg_peer_put(struct wg_peer *peer) · 25 unmodified lines
225	227	return;
226	228	kref_put(&peer->refcount, kref_release);
227	229	}
	230	+
	231	+int __init wg_peer_init(void)
	232	+{
	233	+ peer_cache = KMEM_CACHE(wg_peer, 0);
	234	+ return peer_cache ? 0 : -ENOMEM;
	235	+}
	236	+
	237	+void wg_peer_uninit(void)
	238	+{
	239	+ kmem_cache_destroy(peer_cache);
	240	+}

drivers/net/wireguard/peer.h+3−0

@@ -80,4 +80,7 @@ void wg_peer_put(struct wg_peer *peer); · 79 unmodified lines
80	80	void wg_peer_remove(struct wg_peer *peer);
81	81	void wg_peer_remove_all(struct wg_device *wg);
82	82
	83	+int wg_peer_init(void);
	84	+void wg_peer_uninit(void);
	85	+
83	86	#endif /* _WG_PEER_H */

drivers/net/wireguard/selftest/allowedips.c+79−86

@@ -19,32 +19,22 @@ · 18 unmodified lines
19	19
20	20	#include <linux/siphash.h>
21	21
22		−static __init void swap_endian_and_apply_cidr(u8 dst, const u8 src, u8 bits,
23		− u8 cidr)
24		−{
25		− swap_endian(dst, src, bits);
26		− memset(dst + (cidr + 7) / 8, 0, bits / 8 - (cidr + 7) / 8);
27		− if (cidr)
28		− dst[(cidr + 7) / 8 - 1] &= ~0U << ((8 - (cidr % 8)) % 8);
29		−}
30		−
31	22	static __init void print_node(struct allowedips_node *node, u8 bits)
32	23	{
33	24	char *fmt_connection = KERN_DEBUG "\t\"%p/%d\" -> \"%p/%d\";\n";
34		− char *fmt_declaration = KERN_DEBUG
35		− "\t\"%p/%d\"[style=%s, color=\"#%06x\"];\n";
	25	+ char *fmt_declaration = KERN_DEBUG "\t\"%p/%d\"[style=%s, color=\"#%06x\"];\n";
	26	+ u8 ip1[16], ip2[16], cidr1, cidr2;
36	27	char *style = "dotted";
37		− u8 ip1[16], ip2[16];
38	28	u32 color = 0;
39	29
	30	+ if (node == NULL)
	31	+ return;
40	32	if (bits == 32) {
41	33	fmt_connection = KERN_DEBUG "\t\"%pI4/%d\" -> \"%pI4/%d\";\n";
42		− fmt_declaration = KERN_DEBUG
43		− "\t\"%pI4/%d\"[style=%s, color=\"#%06x\"];\n";
	34	+ fmt_declaration = KERN_DEBUG "\t\"%pI4/%d\"[style=%s, color=\"#%06x\"];\n";
44	35	} else if (bits == 128) {
45	36	fmt_connection = KERN_DEBUG "\t\"%pI6/%d\" -> \"%pI6/%d\";\n";
46		− fmt_declaration = KERN_DEBUG
47		− "\t\"%pI6/%d\"[style=%s, color=\"#%06x\"];\n";
	37	+ fmt_declaration = KERN_DEBUG "\t\"%pI6/%d\"[style=%s, color=\"#%06x\"];\n";
48	38	}
49	39	if (node->peer) {
50	40	hsiphash_key_t key = { { 0 } };
@@ -55,24 +45,20 @@ static __init void print_node(struct allowedips_node *node, u8 bits) · 4 unmodified lines
55	45	hsiphash_1u32(0xabad1dea, &key) % 200;
56	46	style = "bold";
57	47	}
58		− swap_endian_and_apply_cidr(ip1, node->bits, bits, node->cidr);
59		− printk(fmt_declaration, ip1, node->cidr, style, color);
	48	+ wg_allowedips_read_node(node, ip1, &cidr1);
	49	+ printk(fmt_declaration, ip1, cidr1, style, color);
60	50	if (node->bit[0]) {
61		− swap_endian_and_apply_cidr(ip2,
62		− rcu_dereference_raw(node->bit[0])->bits, bits,
63		− node->cidr);
64		− printk(fmt_connection, ip1, node->cidr, ip2,
65		− rcu_dereference_raw(node->bit[0])->cidr);
66		− print_node(rcu_dereference_raw(node->bit[0]), bits);
	51	+ wg_allowedips_read_node(rcu_dereference_raw(node->bit[0]), ip2, &cidr2);
	52	+ printk(fmt_connection, ip1, cidr1, ip2, cidr2);
67	53	}
68	54	if (node->bit[1]) {
69		− swap_endian_and_apply_cidr(ip2,
70		− rcu_dereference_raw(node->bit[1])->bits,
71		− bits, node->cidr);
72		− printk(fmt_connection, ip1, node->cidr, ip2,
73		− rcu_dereference_raw(node->bit[1])->cidr);
74		− print_node(rcu_dereference_raw(node->bit[1]), bits);
	55	+ wg_allowedips_read_node(rcu_dereference_raw(node->bit[1]), ip2, &cidr2);
	56	+ printk(fmt_connection, ip1, cidr1, ip2, cidr2);
75	57	}
	58	+ if (node->bit[0])
	59	+ print_node(rcu_dereference_raw(node->bit[0]), bits);
	60	+ if (node->bit[1])
	61	+ print_node(rcu_dereference_raw(node->bit[1]), bits);
76	62	}
77	63
78	64	static __init void print_tree(struct allowedips_node __rcu *top, u8 bits)
@@ -121,8 +107,8 @@ static __init inline union nf_inet_addr horrible_cidr_to_mask(u8 cidr) · 42 unmodified lines
121	107	{
122	108	union nf_inet_addr mask;
123	109
124		− memset(&mask, 0x00, 128 / 8);
125		− memset(&mask, 0xff, cidr / 8);
	110	+ memset(&mask, 0, sizeof(mask));
	111	+ memset(&mask.all, 0xff, cidr / 8);
126	112	if (cidr % 32)
127	113	mask.all[cidr / 32] = (__force u32)htonl(
128	114	(0xFFFFFFFFUL << (32 - (cidr % 32))) & 0xFFFFFFFFUL);
@@ -149,42 +135,36 @@ horrible_mask_self(struct horrible_allowedips_node *node) · 20 unmodified lines
149	135	}
150	136
151	137	static __init inline bool
152		−horrible_match_v4(const struct horrible_allowedips_node *node,
153		− struct in_addr *ip)
	138	+horrible_match_v4(const struct horrible_allowedips_node node, struct in_addr ip)
154	139	{
155	140	return (ip->s_addr & node->mask.ip) == node->ip.ip;
156	141	}
157	142
158	143	static __init inline bool
159		−horrible_match_v6(const struct horrible_allowedips_node *node,
160		− struct in6_addr *ip)
	144	+horrible_match_v6(const struct horrible_allowedips_node node, struct in6_addr ip)
161	145	{
162		− return (ip->in6_u.u6_addr32[0] & node->mask.ip6[0]) ==
163		− node->ip.ip6[0] &&
164		− (ip->in6_u.u6_addr32[1] & node->mask.ip6[1]) ==
165		− node->ip.ip6[1] &&
166		− (ip->in6_u.u6_addr32[2] & node->mask.ip6[2]) ==
167		− node->ip.ip6[2] &&
	146	+ return (ip->in6_u.u6_addr32[0] & node->mask.ip6[0]) == node->ip.ip6[0] &&
	147	+ (ip->in6_u.u6_addr32[1] & node->mask.ip6[1]) == node->ip.ip6[1] &&
	148	+ (ip->in6_u.u6_addr32[2] & node->mask.ip6[2]) == node->ip.ip6[2] &&
168	149	(ip->in6_u.u6_addr32[3] & node->mask.ip6[3]) == node->ip.ip6[3];
169	150	}
170	151
171	152	static __init void
172		−horrible_insert_ordered(struct horrible_allowedips *table,
173		− struct horrible_allowedips_node *node)
	153	+horrible_insert_ordered(struct horrible_allowedips table, struct horrible_allowedips_node node)
174	154	{
175	155	struct horrible_allowedips_node other = NULL, where = NULL;
176	156	u8 my_cidr = horrible_mask_to_cidr(node->mask);
177	157
178	158	hlist_for_each_entry(other, &table->head, table) {
179		− if (!memcmp(&other->mask, &node->mask,
180		− sizeof(union nf_inet_addr)) &&
181		− !memcmp(&other->ip, &node->ip,
182		− sizeof(union nf_inet_addr)) &&
183		− other->ip_version == node->ip_version) {
	159	+ if (other->ip_version == node->ip_version &&
	160	+ !memcmp(&other->mask, &node->mask, sizeof(union nf_inet_addr)) &&
	161	+ !memcmp(&other->ip, &node->ip, sizeof(union nf_inet_addr))) {
184	162	other->value = node->value;
185	163	kfree(node);
186	164	return;
187	165	}
	166	+ }
	167	+ hlist_for_each_entry(other, &table->head, table) {
188	168	where = other;
189	169	if (horrible_mask_to_cidr(other->mask) <= my_cidr)
190	170	break;
@@ -201,8 +181,7 @@ static __init int · 10 unmodified lines
201	181	horrible_allowedips_insert_v4(struct horrible_allowedips *table,
202	182	struct in_addr ip, u8 cidr, void value)
203	183	{
204		− struct horrible_allowedips_node node = kzalloc(sizeof(node),
205		− GFP_KERNEL);
	184	+ struct horrible_allowedips_node node = kzalloc(sizeof(node), GFP_KERNEL);
206	185
207	186	if (unlikely(!node))
208	187	return -ENOMEM;
@@ -219,8 +198,7 @@ static __init int · 10 unmodified lines
219	198	horrible_allowedips_insert_v6(struct horrible_allowedips *table,
220	199	struct in6_addr ip, u8 cidr, void value)
221	200	{
222		− struct horrible_allowedips_node node = kzalloc(sizeof(node),
223		− GFP_KERNEL);
	201	+ struct horrible_allowedips_node node = kzalloc(sizeof(node), GFP_KERNEL);
224	202
225	203	if (unlikely(!node))
226	204	return -ENOMEM;
@@ -234,39 +212,43 @@ horrible_allowedips_insert_v6(struct horrible_allowedips *table, · 7 unmodified lines
234	212	}
235	213
236	214	static __init void *
237		−horrible_allowedips_lookup_v4(struct horrible_allowedips *table,
238		− struct in_addr *ip)
	215	+horrible_allowedips_lookup_v4(struct horrible_allowedips table, struct in_addr ip)
239	216	{
240	217	struct horrible_allowedips_node *node;
241		− void *ret = NULL;
242	218
243	219	hlist_for_each_entry(node, &table->head, table) {
244		− if (node->ip_version != 4)
245		− continue;
246		− if (horrible_match_v4(node, ip)) {
247		− ret = node->value;
248		− break;
249		− }
	220	+ if (node->ip_version == 4 && horrible_match_v4(node, ip))
	221	+ return node->value;
250	222	}
251		− return ret;
	223	+ return NULL;
252	224	}
253	225
254	226	static __init void *
255		−horrible_allowedips_lookup_v6(struct horrible_allowedips *table,
256		− struct in6_addr *ip)
	227	+horrible_allowedips_lookup_v6(struct horrible_allowedips table, struct in6_addr ip)
257	228	{
258	229	struct horrible_allowedips_node *node;
259		− void *ret = NULL;
260	230
261	231	hlist_for_each_entry(node, &table->head, table) {
262		− if (node->ip_version != 6)
	232	+ if (node->ip_version == 6 && horrible_match_v6(node, ip))
	233	+ return node->value;
	234	+ }
	235	+ return NULL;
	236	+}
	237	+
	238	+
	239	+static __init void
	240	+horrible_allowedips_remove_by_value(struct horrible_allowedips table, void value)
	241	+{
	242	+ struct horrible_allowedips_node *node;
	243	+ struct hlist_node *h;
	244	+
	245	+ hlist_for_each_entry_safe(node, h, &table->head, table) {
	246	+ if (node->value != value)
263	247	continue;
264		− if (horrible_match_v6(node, ip)) {
265		− ret = node->value;
266		− break;
267		− }
	248	+ hlist_del(&node->table);
	249	+ kfree(node);
268	250	}
269		− return ret;
	251	+
270	252	}
271	253
272	254	static __init bool randomized_test(void)
@@ -296,6 +278,7 @@ static __init bool randomized_test(void) · 23 unmodified lines
296	278	goto free;
297	279	}
298	280	kref_init(&peers[i]->refcount);
	281	+ INIT_LIST_HEAD(&peers[i]->allowedips_list);
299	282	}
300	283
301	284	mutex_lock(&mutex);
@@ -333,7 +316,7 @@ static __init bool randomized_test(void) · 31 unmodified lines
333	316	if (wg_allowedips_insert_v4(&t,
334	317	(struct in_addr *)mutated,
335	318	cidr, peer, &mutex) < 0) {
336		− pr_err("allowedips random malloc: FAIL\n");
	319	+ pr_err("allowedips random self-test malloc: FAIL\n");
337	320	goto free_locked;
338	321	}
339	322	if (horrible_allowedips_insert_v4(&h,
@@ -396,23 +379,33 @@ static __init bool randomized_test(void) · 56 unmodified lines
396	379	print_tree(t.root6, 128);
397	380	}
398	381
399		− for (i = 0; i < NUM_QUERIES; ++i) {
400		− prandom_bytes(ip, 4);
401		− if (lookup(t.root4, 32, ip) !=
402		− horrible_allowedips_lookup_v4(&h, (struct in_addr *)ip)) {
403		− pr_err("allowedips random self-test: FAIL\n");
404		− goto free;
	382	+ for (j = 0;; ++j) {
	383	+ for (i = 0; i < NUM_QUERIES; ++i) {
	384	+ prandom_bytes(ip, 4);
	385	+ if (lookup(t.root4, 32, ip) != horrible_allowedips_lookup_v4(&h, (struct in_addr *)ip)) {
	386	+ horrible_allowedips_lookup_v4(&h, (struct in_addr *)ip);
	387	+ pr_err("allowedips random v4 self-test: FAIL\n");
	388	+ goto free;
	389	+ }
	390	+ prandom_bytes(ip, 16);
	391	+ if (lookup(t.root6, 128, ip) != horrible_allowedips_lookup_v6(&h, (struct in6_addr *)ip)) {
	392	+ pr_err("allowedips random v6 self-test: FAIL\n");
	393	+ goto free;
	394	+ }
405	395	}
	396	+ if (j >= NUM_PEERS)
	397	+ break;
	398	+ mutex_lock(&mutex);
	399	+ wg_allowedips_remove_by_peer(&t, peers[j], &mutex);
	400	+ mutex_unlock(&mutex);
	401	+ horrible_allowedips_remove_by_value(&h, peers[j]);
406	402	}
407	403
408		− for (i = 0; i < NUM_QUERIES; ++i) {
409		− prandom_bytes(ip, 16);
410		− if (lookup(t.root6, 128, ip) !=
411		− horrible_allowedips_lookup_v6(&h, (struct in6_addr *)ip)) {
412		− pr_err("allowedips random self-test: FAIL\n");
413		− goto free;
414		− }
	404	+ if (t.root4 \|\| t.root6) {
	405	+ pr_err("allowedips random self-test removal: FAIL\n");
	406	+ goto free;
415	407	}
	408	+
416	409	ret = true;
417	410
418	411	free:

drivers/net/wireguard/socket.c+1−1

@@ -430,7 +430,7 @@ void wg_socket_reinit(struct wg_device wg, struct sock new4, · 429 unmodified lines
430	430	if (new4)
431	431	wg->incoming_port = ntohs(inet_sk(new4)->inet_sport);
432	432	mutex_unlock(&wg->socket_update_lock);
433		− synchronize_rcu();
	433	+ synchronize_net();
434	434	sock_free(old4);
435	435	sock_free(old6);
436	436	}

drivers/net/wireless/ath/ath10k/htt.h+1−0

@@ -845,6 +845,7 @@ enum htt_security_types { · 844 unmodified lines
845	845
846	846	#define ATH10K_HTT_TXRX_PEER_SECURITY_MAX 2
847	847	#define ATH10K_TXRX_NUM_EXT_TIDS 19
	848	+#define ATH10K_TXRX_NON_QOS_TID 16
848	849
849	850	enum htt_security_flags {
850	851	#define HTT_SECURITY_TYPE_MASK 0x7F

drivers/net/wireless/ath/ath10k/htt_rx.c+191−10

@@ -1746,16 +1746,97 @@ static void ath10k_htt_rx_h_csum_offload(struct sk_buff *msdu) · 1745 unmodified lines
1746	1746	msdu->ip_summed = ath10k_htt_rx_get_csum_state(msdu);
1747	1747	}
1748	1748
	1749	+static u64 ath10k_htt_rx_h_get_pn(struct ath10k ar, struct sk_buff skb,
	1750	+ u16 offset,
	1751	+ enum htt_rx_mpdu_encrypt_type enctype)
	1752	+{
	1753	+ struct ieee80211_hdr *hdr;
	1754	+ u64 pn = 0;
	1755	+ u8 *ehdr;
	1756	+
	1757	+ hdr = (struct ieee80211_hdr *)(skb->data + offset);
	1758	+ ehdr = skb->data + offset + ieee80211_hdrlen(hdr->frame_control);
	1759	+
	1760	+ if (enctype == HTT_RX_MPDU_ENCRYPT_AES_CCM_WPA2) {
	1761	+ pn = ehdr[0];
	1762	+ pn \|= (u64)ehdr[1] << 8;
	1763	+ pn \|= (u64)ehdr[4] << 16;
	1764	+ pn \|= (u64)ehdr[5] << 24;
	1765	+ pn \|= (u64)ehdr[6] << 32;
	1766	+ pn \|= (u64)ehdr[7] << 40;
	1767	+ }
	1768	+ return pn;
	1769	+}
	1770	+
	1771	+static bool ath10k_htt_rx_h_frag_multicast_check(struct ath10k *ar,
	1772	+ struct sk_buff *skb,
	1773	+ u16 offset)
	1774	+{
	1775	+ struct ieee80211_hdr *hdr;
	1776	+
	1777	+ hdr = (struct ieee80211_hdr *)(skb->data + offset);
	1778	+ return !is_multicast_ether_addr(hdr->addr1);
	1779	+}
	1780	+
	1781	+static bool ath10k_htt_rx_h_frag_pn_check(struct ath10k *ar,
	1782	+ struct sk_buff *skb,
	1783	+ u16 peer_id,
	1784	+ u16 offset,
	1785	+ enum htt_rx_mpdu_encrypt_type enctype)
	1786	+{
	1787	+ struct ath10k_peer *peer;
	1788	+ union htt_rx_pn_t *last_pn, new_pn = {0};
	1789	+ struct ieee80211_hdr *hdr;
	1790	+ bool more_frags;
	1791	+ u8 tid, frag_number;
	1792	+ u32 seq;
	1793	+
	1794	+ peer = ath10k_peer_find_by_id(ar, peer_id);
	1795	+ if (!peer) {
	1796	+ ath10k_dbg(ar, ATH10K_DBG_HTT, "invalid peer for frag pn check\n");
	1797	+ return false;
	1798	+ }
	1799	+
	1800	+ hdr = (struct ieee80211_hdr *)(skb->data + offset);
	1801	+ if (ieee80211_is_data_qos(hdr->frame_control))
	1802	+ tid = ieee80211_get_tid(hdr);
	1803	+ else
	1804	+ tid = ATH10K_TXRX_NON_QOS_TID;
	1805	+
	1806	+ last_pn = &peer->frag_tids_last_pn[tid];
	1807	+ new_pn.pn48 = ath10k_htt_rx_h_get_pn(ar, skb, offset, enctype);
	1808	+ more_frags = ieee80211_has_morefrags(hdr->frame_control);
	1809	+ frag_number = le16_to_cpu(hdr->seq_ctrl) & IEEE80211_SCTL_FRAG;
	1810	+ seq = (__le16_to_cpu(hdr->seq_ctrl) & IEEE80211_SCTL_SEQ) >> 4;
	1811	+
	1812	+ if (frag_number == 0) {
	1813	+ last_pn->pn48 = new_pn.pn48;
	1814	+ peer->frag_tids_seq[tid] = seq;
	1815	+ } else {
	1816	+ if (seq != peer->frag_tids_seq[tid])
	1817	+ return false;
	1818	+
	1819	+ if (new_pn.pn48 != last_pn->pn48 + 1)
	1820	+ return false;
	1821	+
	1822	+ last_pn->pn48 = new_pn.pn48;
	1823	+ }
	1824	+
	1825	+ return true;
	1826	+}
	1827	+
1749	1828	static void ath10k_htt_rx_h_mpdu(struct ath10k *ar,
1750	1829	struct sk_buff_head *amsdu,
1751	1830	struct ieee80211_rx_status *status,
1752	1831	bool fill_crypt_header,
1753	1832	u8 *rx_hdr,
1754		− enum ath10k_pkt_rx_err *err)
	1833	+ enum ath10k_pkt_rx_err *err,
	1834	+ u16 peer_id,
	1835	+ bool frag)
1755	1836	{
1756	1837	struct sk_buff *first;
1757	1838	struct sk_buff *last;
1758		− struct sk_buff *msdu;
	1839	+ struct sk_buff msdu, temp;
1759	1840	struct htt_rx_desc *rxd;
1760	1841	struct ieee80211_hdr *hdr;
1761	1842	enum htt_rx_mpdu_encrypt_type enctype;
@@ -1768,6 +1849,7 @@ static void ath10k_htt_rx_h_mpdu(struct ath10k *ar, · 6 unmodified lines
1768	1849	bool is_decrypted;
1769	1850	bool is_mgmt;
1770	1851	u32 attention;
	1852	+ bool frag_pn_check = true, multicast_check = true;
1771	1853
1772	1854	if (skb_queue_empty(amsdu))
1773	1855	return;
@@ -1866,7 +1948,37 @@ static void ath10k_htt_rx_h_mpdu(struct ath10k *ar, · 92 unmodified lines
1866	1948	}
1867	1949
1868	1950	skb_queue_walk(amsdu, msdu) {
	1951	+ if (frag && !fill_crypt_header && is_decrypted &&
	1952	+ enctype == HTT_RX_MPDU_ENCRYPT_AES_CCM_WPA2)
	1953	+ frag_pn_check = ath10k_htt_rx_h_frag_pn_check(ar,
	1954	+ msdu,
	1955	+ peer_id,
	1956	+ 0,
	1957	+ enctype);
	1958	+
	1959	+ if (frag)
	1960	+ multicast_check = ath10k_htt_rx_h_frag_multicast_check(ar,
	1961	+ msdu,
	1962	+ 0);
	1963	+
	1964	+ if (!frag_pn_check \|\| !multicast_check) {
	1965	+ /* Discard the fragment with invalid PN or multicast DA
	1966	+ */
	1967	+ temp = msdu->prev;
	1968	+ __skb_unlink(msdu, amsdu);
	1969	+ dev_kfree_skb_any(msdu);
	1970	+ msdu = temp;
	1971	+ frag_pn_check = true;
	1972	+ multicast_check = true;
	1973	+ continue;
	1974	+ }
	1975	+
1869	1976	ath10k_htt_rx_h_csum_offload(msdu);
	1977	+
	1978	+ if (frag && !fill_crypt_header &&
	1979	+ enctype == HTT_RX_MPDU_ENCRYPT_TKIP_WPA)
	1980	+ status->flag &= ~RX_FLAG_MMIC_STRIPPED;
	1981	+
1870	1982	ath10k_htt_rx_h_undecap(ar, msdu, status, first_hdr, enctype,
1871	1983	is_decrypted);
1872	1984
@@ -1884,6 +1996,11 @@ static void ath10k_htt_rx_h_mpdu(struct ath10k *ar, · 11 unmodified lines
1884	1996
1885	1997	hdr = (void *)msdu->data;
1886	1998	hdr->frame_control &= ~__cpu_to_le16(IEEE80211_FCTL_PROTECTED);
	1999	+
	2000	+ if (frag && !fill_crypt_header &&
	2001	+ enctype == HTT_RX_MPDU_ENCRYPT_TKIP_WPA)
	2002	+ status->flag &= ~RX_FLAG_IV_STRIPPED &
	2003	+ ~RX_FLAG_MMIC_STRIPPED;
1887	2004	}
1888	2005	}
1889	2006
@@ -1991,14 +2108,62 @@ static void ath10k_htt_rx_h_unchain(struct ath10k *ar, · 101 unmodified lines
1991	2108	ath10k_unchain_msdu(amsdu, unchain_cnt);
1992	2109	}
1993	2110
	2111	+static bool ath10k_htt_rx_validate_amsdu(struct ath10k *ar,
	2112	+ struct sk_buff_head *amsdu)
	2113	+{
	2114	+ u8 *subframe_hdr;
	2115	+ struct sk_buff *first;
	2116	+ bool is_first, is_last;
	2117	+ struct htt_rx_desc *rxd;
	2118	+ struct ieee80211_hdr *hdr;
	2119	+ size_t hdr_len, crypto_len;
	2120	+ enum htt_rx_mpdu_encrypt_type enctype;
	2121	+ int bytes_aligned = ar->hw_params.decap_align_bytes;
	2122	+
	2123	+ first = skb_peek(amsdu);
	2124	+
	2125	+ rxd = (void )first->data - sizeof(rxd);
	2126	+ hdr = (void *)rxd->rx_hdr_status;
	2127	+
	2128	+ is_first = !!(rxd->msdu_end.common.info0 &
	2129	+ __cpu_to_le32(RX_MSDU_END_INFO0_FIRST_MSDU));
	2130	+ is_last = !!(rxd->msdu_end.common.info0 &
	2131	+ __cpu_to_le32(RX_MSDU_END_INFO0_LAST_MSDU));
	2132	+
	2133	+ /* Return in case of non-aggregated msdu */
	2134	+ if (is_first && is_last)
	2135	+ return true;
	2136	+
	2137	+ /* First msdu flag is not set for the first msdu of the list */
	2138	+ if (!is_first)
	2139	+ return false;
	2140	+
	2141	+ enctype = MS(__le32_to_cpu(rxd->mpdu_start.info0),
	2142	+ RX_MPDU_START_INFO0_ENCRYPT_TYPE);
	2143	+
	2144	+ hdr_len = ieee80211_hdrlen(hdr->frame_control);
	2145	+ crypto_len = ath10k_htt_rx_crypto_param_len(ar, enctype);
	2146	+
	2147	+ subframe_hdr = (u8 *)hdr + round_up(hdr_len, bytes_aligned) +
	2148	+ crypto_len;
	2149	+
	2150	+ /* Validate if the amsdu has a proper first subframe.
	2151	+ * There are chances a single msdu can be received as amsdu when
	2152	+ * the unauthenticated amsdu flag of a QoS header
	2153	+ * gets flipped in non-SPP AMSDU's, in such cases the first
	2154	+ * subframe has llc/snap header in place of a valid da.
	2155	+ * return false if the da matches rfc1042 pattern
	2156	+ */
	2157	+ if (ether_addr_equal(subframe_hdr, rfc1042_header))
	2158	+ return false;
	2159	+
	2160	+ return true;
	2161	+}
	2162	+
1994	2163	static bool ath10k_htt_rx_amsdu_allowed(struct ath10k *ar,
1995	2164	struct sk_buff_head *amsdu,
1996	2165	struct ieee80211_rx_status *rx_status)
1997	2166	{
1998		− /* FIXME: It might be a good idea to do some fuzzy-testing to drop
1999		− * invalid/dangerous frames.
2000		− */
2001		−
2002	2167	if (!rx_status->freq) {
2003	2168	ath10k_dbg(ar, ATH10K_DBG_HTT, "no channel configured; ignoring frame(s)!\n");
2004	2169	return false;
@@ -2009,6 +2174,11 @@ static bool ath10k_htt_rx_amsdu_allowed(struct ath10k *ar, · 4 unmodified lines
2009	2174	return false;
2010	2175	}
2011	2176
	2177	+ if (!ath10k_htt_rx_validate_amsdu(ar, amsdu)) {
	2178	+ ath10k_dbg(ar, ATH10K_DBG_HTT, "invalid amsdu received\n");
	2179	+ return false;
	2180	+ }
	2181	+
2012	2182	return true;
2013	2183	}
2014	2184
@@ -2071,7 +2241,8 @@ static int ath10k_htt_rx_handle_amsdu(struct ath10k_htt *htt) · 56 unmodified lines
2071	2241	ath10k_htt_rx_h_unchain(ar, &amsdu, &drop_cnt, &unchain_cnt);
2072	2242
2073	2243	ath10k_htt_rx_h_filter(ar, &amsdu, rx_status, &drop_cnt_filter);
2074		− ath10k_htt_rx_h_mpdu(ar, &amsdu, rx_status, true, first_hdr, &err);
	2244	+ ath10k_htt_rx_h_mpdu(ar, &amsdu, rx_status, true, first_hdr, &err, 0,
	2245	+ false);
2075	2246	msdus_to_queue = skb_queue_len(&amsdu);
2076	2247	ath10k_htt_rx_h_enqueue(ar, &amsdu, rx_status);
2077	2248
@@ -2204,6 +2375,11 @@ static bool ath10k_htt_rx_proc_rx_ind_hl(struct ath10k_htt *htt, · 126 unmodified lines
2204	2375	fw_desc = &rx->fw_desc;
2205	2376	rx_desc_len = fw_desc->len;
2206	2377
	2378	+ if (fw_desc->u.bits.discard) {
	2379	+ ath10k_dbg(ar, ATH10K_DBG_HTT, "htt discard mpdu\n");
	2380	+ goto err;
	2381	+ }
	2382	+
2207	2383	/* I have not yet seen any case where num_mpdu_ranges > 1.
2208	2384	* qcacld does not seem handle that case either, so we introduce the
2209	2385	* same limitiation here as well.
@@ -2509,6 +2685,13 @@ static bool ath10k_htt_rx_proc_rx_frag_ind_hl(struct ath10k_htt *htt, · 299 unmodified lines
2509	2685	rx_desc = (struct htt_hl_rx_desc *)(skb->data + tot_hdr_len);
2510	2686	rx_desc_info = __le32_to_cpu(rx_desc->info);
2511	2687
	2688	+ hdr = (struct ieee80211_hdr )((u8 )rx_desc + rx_hl->fw_desc.len);
	2689	+
	2690	+ if (is_multicast_ether_addr(hdr->addr1)) {
	2691	+ /* Discard the fragment with multicast DA */
	2692	+ goto err;
	2693	+ }
	2694	+
2512	2695	if (!MS(rx_desc_info, HTT_RX_DESC_HL_INFO_ENCRYPTED)) {
2513	2696	spin_unlock_bh(&ar->data_lock);
2514	2697	return ath10k_htt_rx_proc_rx_ind_hl(htt, &resp->rx_ind_hl, skb,
@@ -2516,8 +2699,6 @@ static bool ath10k_htt_rx_proc_rx_frag_ind_hl(struct ath10k_htt *htt, · 1 unmodified line
2516	2699	HTT_RX_NON_TKIP_MIC);
2517	2700	}
2518	2701
2519		− hdr = (struct ieee80211_hdr )((u8 )rx_desc + rx_hl->fw_desc.len);
2520		−
2521	2702	if (ieee80211_has_retry(hdr->frame_control))
2522	2703	goto err;
2523	2704
@@ -3027,7 +3208,7 @@ static int ath10k_htt_rx_in_ord_ind(struct ath10k ar, struct sk_buff skb) · 503 unmodified lines
3027	3208	ath10k_htt_rx_h_ppdu(ar, &amsdu, status, vdev_id);
3028	3209	ath10k_htt_rx_h_filter(ar, &amsdu, status, NULL);
3029	3210	ath10k_htt_rx_h_mpdu(ar, &amsdu, status, false, NULL,
3030		− NULL);
	3211	+ NULL, peer_id, frag);
3031	3212	ath10k_htt_rx_h_enqueue(ar, &amsdu, status);
3032	3213	break;
3033	3214	case -EAGAIN:

drivers/net/wireless/ath/ath10k/rx_desc.h+13−1

@@ -1282,7 +1282,19 @@ struct fw_rx_desc_base { · 1281 unmodified lines
1282	1282	#define FW_RX_DESC_UDP (1 << 6)
1283	1283
1284	1284	struct fw_rx_desc_hl {
1285		− u8 info0;
	1285	+ union {
	1286	+ struct {
	1287	+ u8 discard:1,
	1288	+ forward:1,
	1289	+ any_err:1,
	1290	+ dup_err:1,
	1291	+ reserved:1,
	1292	+ inspect:1,
	1293	+ extension:2;
	1294	+ } bits;
	1295	+ u8 info0;
	1296	+ } u;
	1297	+
1286	1298	u8 version;
1287	1299	u8 len;
1288	1300	u8 flags;

drivers/net/wireless/ath/ath11k/dp_rx.c+34−0

@@ -260,6 +260,16 @@ static void ath11k_dp_rxdesc_set_msdu_len(struct ath11k_base *ab, · 259 unmodified lines
260	260	ab->hw_params.hw_ops->rx_desc_set_msdu_len(desc, len);
261	261	}
262	262
	263	+static bool ath11k_dp_rx_h_attn_is_mcbc(struct ath11k_base *ab,
	264	+ struct hal_rx_desc *desc)
	265	+{
	266	+ struct rx_attention *attn = ath11k_dp_rx_get_attention(ab, desc);
	267	+
	268	+ return ath11k_dp_rx_h_msdu_end_first_msdu(ab, desc) &&
	269	+ (!!FIELD_GET(RX_ATTENTION_INFO1_MCAST_BCAST,
	270	+ __le32_to_cpu(attn->info1)));
	271	+}
	272	+
263	273	static void ath11k_dp_service_mon_ring(struct timer_list *t)
264	274	{
265	275	struct ath11k_base *ab = from_timer(ab, t, mon_reap_timer);
@@ -852,6 +862,24 @@ static void ath11k_dp_rx_frags_cleanup(struct dp_rx_tid *rx_tid, bool rel_link_d · 586 unmodified lines
852	862	__skb_queue_purge(&rx_tid->rx_frags);
853	863	}
854	864
	865	+void ath11k_peer_frags_flush(struct ath11k ar, struct ath11k_peer peer)
	866	+{
	867	+ struct dp_rx_tid *rx_tid;
	868	+ int i;
	869	+
	870	+ lockdep_assert_held(&ar->ab->base_lock);
	871	+
	872	+ for (i = 0; i <= IEEE80211_NUM_TIDS; i++) {
	873	+ rx_tid = &peer->rx_tid[i];
	874	+
	875	+ spin_unlock_bh(&ar->ab->base_lock);
	876	+ del_timer_sync(&rx_tid->frag_timer);
	877	+ spin_lock_bh(&ar->ab->base_lock);
	878	+
	879	+ ath11k_dp_rx_frags_cleanup(rx_tid, true);
	880	+ }
	881	+}
	882	+
855	883	void ath11k_peer_rx_tid_cleanup(struct ath11k ar, struct ath11k_peer peer)
856	884	{
857	885	struct dp_rx_tid *rx_tid;
@@ -3450,6 +3478,7 @@ static int ath11k_dp_rx_frag_h_mpdu(struct ath11k *ar, · 2592 unmodified lines
3450	3478	u8 tid;
3451	3479	int ret = 0;
3452	3480	bool more_frags;
	3481	+ bool is_mcbc;
3453	3482
3454	3483	rx_desc = (struct hal_rx_desc *)msdu->data;
3455	3484	peer_id = ath11k_dp_rx_h_mpdu_start_peer_id(ar->ab, rx_desc);
@@ -3457,6 +3486,11 @@ static int ath11k_dp_rx_frag_h_mpdu(struct ath11k *ar, · 1 unmodified line
3457	3486	seqno = ath11k_dp_rx_h_mpdu_start_seq_no(ar->ab, rx_desc);
3458	3487	frag_no = ath11k_dp_rx_h_mpdu_start_frag_no(ar->ab, msdu);
3459	3488	more_frags = ath11k_dp_rx_h_mpdu_start_more_frags(ar->ab, msdu);
	3489	+ is_mcbc = ath11k_dp_rx_h_attn_is_mcbc(ar->ab, rx_desc);
	3490	+
	3491	+ /* Multicast/Broadcast fragments are not expected */
	3492	+ if (is_mcbc)
	3493	+ return -EINVAL;
3460	3494
3461	3495	if (!ath11k_dp_rx_h_mpdu_start_seq_ctrl_valid(ar->ab, rx_desc) \|\|
3462	3496	!ath11k_dp_rx_h_mpdu_start_fc_valid(ar->ab, rx_desc) \|\|

drivers/net/wireless/ath/ath11k/dp_rx.h+1−0

@@ -49,6 +49,7 @@ int ath11k_dp_peer_rx_pn_replay_config(struct ath11k_vif *arvif, · 48 unmodified lines
49	49	const u8 *peer_addr,
50	50	enum set_key_cmd key_cmd,
51	51	struct ieee80211_key_conf *key);
	52	+void ath11k_peer_frags_flush(struct ath11k ar, struct ath11k_peer peer);
52	53	void ath11k_peer_rx_tid_cleanup(struct ath11k ar, struct ath11k_peer peer);
53	54	void ath11k_peer_rx_tid_delete(struct ath11k *ar,
54	55	struct ath11k_peer *peer, u8 tid);

drivers/net/wireless/ath/ath11k/mac.c+6−0

@@ -2779,6 +2779,12 @@ static int ath11k_mac_op_set_key(struct ieee80211_hw *hw, enum set_key_cmd cmd, · 2778 unmodified lines
2779	2779	*/
2780	2780	spin_lock_bh(&ab->base_lock);
2781	2781	peer = ath11k_peer_find(ab, arvif->vdev_id, peer_addr);
	2782	+
	2783	+ /* flush the fragments cache during key (re)install to
	2784	+ * ensure all frags in the new frag list belong to the same key.
	2785	+ */
	2786	+ if (peer && cmd == SET_KEY)
	2787	+ ath11k_peer_frags_flush(ar, peer);
2782	2788	spin_unlock_bh(&ab->base_lock);
2783	2789
2784	2790	if (!peer) {

drivers/net/wireless/mac80211_hwsim.c+5−0

@@ -1693,8 +1693,13 @@ static int mac80211_hwsim_start(struct ieee80211_hw *hw) · 1692 unmodified lines
1693	1693	static void mac80211_hwsim_stop(struct ieee80211_hw *hw)
1694	1694	{
1695	1695	struct mac80211_hwsim_data *data = hw->priv;
	1696	+
1696	1697	data->started = false;
1697	1698	hrtimer_cancel(&data->beacon_timer);
	1699	+
	1700	+ while (!skb_queue_empty(&data->pending))
	1701	+ ieee80211_free_txskb(hw, skb_dequeue(&data->pending));
	1702	+
1698	1703	wiphy_dbg(hw->wiphy, "%s\n", __func__);
1699	1704	}
1700	1705

drivers/net/wireless/mediatek/mt76/mac80211.c+26−0

@@ -514,10 +514,36 @@ EXPORT_SYMBOL_GPL(mt76_free_device); · 513 unmodified lines
514	514	static void mt76_rx_release_amsdu(struct mt76_phy *phy, enum mt76_rxq_id q)
515	515	{
516	516	struct sk_buff *skb = phy->rx_amsdu[q].head;
	517	+ struct mt76_rx_status status = (struct mt76_rx_status )skb->cb;
517	518	struct mt76_dev *dev = phy->dev;
518	519
519	520	phy->rx_amsdu[q].head = NULL;
520	521	phy->rx_amsdu[q].tail = NULL;
	522	+
	523	+ /*
	524	+ * Validate if the amsdu has a proper first subframe.
	525	+ * A single MSDU can be parsed as A-MSDU when the unauthenticated A-MSDU
	526	+ * flag of the QoS header gets flipped. In such cases, the first
	527	+ * subframe has a LLC/SNAP header in the location of the destination
	528	+ * address.
	529	+ */
	530	+ if (skb_shinfo(skb)->frag_list) {
	531	+ int offset = 0;
	532	+
	533	+ if (!(status->flag & RX_FLAG_8023)) {
	534	+ offset = ieee80211_get_hdrlen_from_skb(skb);
	535	+
	536	+ if ((status->flag &
	537	+ (RX_FLAG_DECRYPTED \| RX_FLAG_IV_STRIPPED)) ==
	538	+ RX_FLAG_DECRYPTED)
	539	+ offset += 8;
	540	+ }
	541	+
	542	+ if (ether_addr_equal(skb->data + offset, rfc1042_header)) {
	543	+ dev_kfree_skb(skb);
	544	+ return;
	545	+ }
	546	+ }
521	547	__skb_queue_tail(&dev->rx_skb[q], skb);
522	548	}
523	549

drivers/net/wireless/mediatek/mt76/mt7615/init.c+0−1

@@ -510,7 +510,6 @@ void mt7615_init_device(struct mt7615_dev *dev) · 509 unmodified lines
510	510	mutex_init(&dev->pm.mutex);
511	511	init_waitqueue_head(&dev->pm.wait);
512	512	spin_lock_init(&dev->pm.txq_lock);
513		− set_bit(MT76_STATE_PM, &dev->mphy.state);
514	513	INIT_DELAYED_WORK(&dev->mphy.mac_work, mt7615_mac_work);
515	514	INIT_DELAYED_WORK(&dev->phy.scan_work, mt7615_scan_work);
516	515	INIT_DELAYED_WORK(&dev->coredump.work, mt7615_coredump_work);

drivers/net/wireless/mediatek/mt76/mt7615/mac.c+3−2

@@ -1912,8 +1912,9 @@ void mt7615_pm_wake_work(struct work_struct *work) · 1911 unmodified lines
1912	1912	napi_schedule(&dev->mt76.napi[i]);
1913	1913	mt76_connac_pm_dequeue_skbs(mphy, &dev->pm);
1914	1914	mt76_queue_tx_cleanup(dev, dev->mt76.q_mcu[MT_MCUQ_WM], false);
1915		− ieee80211_queue_delayed_work(mphy->hw, &mphy->mac_work,
1916		− MT7615_WATCHDOG_TIME);
	1915	+ if (test_bit(MT76_STATE_RUNNING, &mphy->state))
	1916	+ ieee80211_queue_delayed_work(mphy->hw, &mphy->mac_work,
	1917	+ MT7615_WATCHDOG_TIME);
1917	1918	}
1918	1919
1919	1920	ieee80211_wake_queues(mphy->hw);

drivers/net/wireless/mediatek/mt76/mt7615/sdio_mcu.c+12−7

@@ -51,16 +51,13 @@ mt7663s_mcu_send_message(struct mt76_dev mdev, struct sk_buff skb, · 50 unmodified lines
51	51	return ret;
52	52	}
53	53
54		−static int mt7663s_mcu_drv_pmctrl(struct mt7615_dev *dev)
	54	+static int __mt7663s_mcu_drv_pmctrl(struct mt7615_dev *dev)
55	55	{
56	56	struct sdio_func *func = dev->mt76.sdio.func;
57	57	struct mt76_phy *mphy = &dev->mt76.phy;
58	58	u32 status;
59	59	int ret;
60	60
61		− if (!test_and_clear_bit(MT76_STATE_PM, &mphy->state))
62		− goto out;
63		−
64	61	sdio_claim_host(func);
65	62
66	63	sdio_writel(func, WHLPCR_FW_OWN_REQ_CLR, MCR_WHLPCR, NULL);
@@ -76,13 +73,21 @@ static int mt7663s_mcu_drv_pmctrl(struct mt7615_dev *dev) · 9 unmodified lines
76	73	}
77	74
78	75	sdio_release_host(func);
79		−
80		−out:
81	76	dev->pm.last_activity = jiffies;
82	77
83	78	return 0;
84	79	}
85	80
	81	+static int mt7663s_mcu_drv_pmctrl(struct mt7615_dev *dev)
	82	+{
	83	+ struct mt76_phy *mphy = &dev->mt76.phy;
	84	+
	85	+ if (test_and_clear_bit(MT76_STATE_PM, &mphy->state))
	86	+ return __mt7663s_mcu_drv_pmctrl(dev);
	87	+
	88	+ return 0;
	89	+}
	90	+
86	91	static int mt7663s_mcu_fw_pmctrl(struct mt7615_dev *dev)
87	92	{
88	93	struct sdio_func *func = dev->mt76.sdio.func;
@@ -123,7 +128,7 @@ int mt7663s_mcu_init(struct mt7615_dev *dev) · 34 unmodified lines
123	128	struct mt7615_mcu_ops *mcu_ops;
124	129	int ret;
125	130
126		− ret = mt7663s_mcu_drv_pmctrl(dev);
	131	+ ret = __mt7663s_mcu_drv_pmctrl(dev);
127	132	if (ret)
128	133	return ret;
129	134

drivers/net/wireless/mediatek/mt76/mt7615/usb_mcu.c+0−3

@@ -55,10 +55,7 @@ int mt7663u_mcu_init(struct mt7615_dev *dev) · 54 unmodified lines
55	55
56	56	dev->mt76.mcu_ops = &mt7663u_mcu_ops,
57	57
58		− /* usb does not support runtime-pm */
59		− clear_bit(MT76_STATE_PM, &dev->mphy.state);
60	58	mt76_set(dev, MT_UDMA_TX_QSEL, MT_FW_DL_EN);
61		−
62	59	if (test_and_clear_bit(MT76_STATE_POWER_OFF, &dev->mphy.state)) {
63	60	mt7615_mcu_restart(&dev->mt76);
64	61	if (!mt76_poll_msec(dev, MT_CONN_ON_MISC,

drivers/net/wireless/mediatek/mt76/mt76_connac_mcu.c+4−0

@@ -721,6 +721,10 @@ void mt76_connac_mcu_sta_tlv(struct mt76_phy mphy, struct sk_buff skb, · 720 unmodified lines
721	721	phy->phy_type = mt76_connac_get_phy_mode_v2(mphy, vif, band, sta);
722	722	phy->basic_rate = cpu_to_le16((u16)vif->bss_conf.basic_rates);
723	723	phy->rcpi = rcpi;
	724	+ phy->ampdu = FIELD_PREP(IEEE80211_HT_AMPDU_PARM_FACTOR,
	725	+ sta->ht_cap.ampdu_factor) \|
	726	+ FIELD_PREP(IEEE80211_HT_AMPDU_PARM_DENSITY,
	727	+ sta->ht_cap.ampdu_density);
724	728
725	729	tlv = mt76_connac_mcu_add_tlv(skb, STA_REC_RA, sizeof(*ra_info));
726	730	ra_info = (struct sta_rec_ra_info *)tlv;

drivers/net/wireless/mediatek/mt76/mt76x0/pci.c+77−4

@@ -87,7 +87,7 @@ static const struct ieee80211_ops mt76x0e_ops = { · 86 unmodified lines
87	87	.reconfig_complete = mt76x02_reconfig_complete,
88	88	};
89	89
90		−static int mt76x0e_register_device(struct mt76x02_dev *dev)
	90	+static int mt76x0e_init_hardware(struct mt76x02_dev *dev, bool resume)
91	91	{
92	92	int err;
93	93
@@ -100,9 +100,11 @@ static int mt76x0e_register_device(struct mt76x02_dev *dev) · 6 unmodified lines
100	100	if (err < 0)
101	101	return err;
102	102
103		− err = mt76x02_dma_init(dev);
104		− if (err < 0)
105		− return err;
	103	+ if (!resume) {
	104	+ err = mt76x02_dma_init(dev);
	105	+ if (err < 0)
	106	+ return err;
	107	+ }
106	108
107	109	err = mt76x0_init_hardware(dev);
108	110	if (err < 0)
@@ -123,6 +125,17 @@ static int mt76x0e_register_device(struct mt76x02_dev *dev) · 14 unmodified lines
123	125	mt76_clear(dev, 0x110, BIT(9));
124	126	mt76_set(dev, MT_MAX_LEN_CFG, BIT(13));
125	127
	128	+ return 0;
	129	+}
	130	+
	131	+static int mt76x0e_register_device(struct mt76x02_dev *dev)
	132	+{
	133	+ int err;
	134	+
	135	+ err = mt76x0e_init_hardware(dev, false);
	136	+ if (err < 0)
	137	+ return err;
	138	+
126	139	err = mt76x0_register_device(dev);
127	140	if (err < 0)
128	141	return err;
@@ -167,6 +180,8 @@ mt76x0e_probe(struct pci_dev pdev, const struct pci_device_id id) · 38 unmodified lines
167	180	if (ret)
168	181	return ret;
169	182
	183	+ mt76_pci_disable_aspm(pdev);
	184	+
170	185	mdev = mt76_alloc_device(&pdev->dev, sizeof(*dev), &mt76x0e_ops,
171	186	&drv_ops);
172	187	if (!mdev)
@@ -220,6 +235,60 @@ mt76x0e_remove(struct pci_dev *pdev) · 47 unmodified lines
220	235	mt76_free_device(mdev);
221	236	}
222	237
	238	+#ifdef CONFIG_PM
	239	+static int mt76x0e_suspend(struct pci_dev *pdev, pm_message_t state)
	240	+{
	241	+ struct mt76_dev *mdev = pci_get_drvdata(pdev);
	242	+ struct mt76x02_dev *dev = container_of(mdev, struct mt76x02_dev, mt76);
	243	+ int i;
	244	+
	245	+ mt76_worker_disable(&mdev->tx_worker);
	246	+ for (i = 0; i < ARRAY_SIZE(mdev->phy.q_tx); i++)
	247	+ mt76_queue_tx_cleanup(dev, mdev->phy.q_tx[i], true);
	248	+ for (i = 0; i < ARRAY_SIZE(mdev->q_mcu); i++)
	249	+ mt76_queue_tx_cleanup(dev, mdev->q_mcu[i], true);
	250	+ napi_disable(&mdev->tx_napi);
	251	+
	252	+ mt76_for_each_q_rx(mdev, i)
	253	+ napi_disable(&mdev->napi[i]);
	254	+
	255	+ mt76x02_dma_disable(dev);
	256	+ mt76x02_mcu_cleanup(dev);
	257	+ mt76x0_chip_onoff(dev, false, false);
	258	+
	259	+ pci_enable_wake(pdev, pci_choose_state(pdev, state), true);
	260	+ pci_save_state(pdev);
	261	+
	262	+ return pci_set_power_state(pdev, pci_choose_state(pdev, state));
	263	+}
	264	+
	265	+static int mt76x0e_resume(struct pci_dev *pdev)
	266	+{
	267	+ struct mt76_dev *mdev = pci_get_drvdata(pdev);
	268	+ struct mt76x02_dev *dev = container_of(mdev, struct mt76x02_dev, mt76);
	269	+ int err, i;
	270	+
	271	+ err = pci_set_power_state(pdev, PCI_D0);
	272	+ if (err)
	273	+ return err;
	274	+
	275	+ pci_restore_state(pdev);
	276	+
	277	+ mt76_worker_enable(&mdev->tx_worker);
	278	+
	279	+ mt76_for_each_q_rx(mdev, i) {
	280	+ mt76_queue_rx_reset(dev, i);
	281	+ napi_enable(&mdev->napi[i]);
	282	+ napi_schedule(&mdev->napi[i]);
	283	+ }
	284	+
	285	+ napi_enable(&mdev->tx_napi);
	286	+ napi_schedule(&mdev->tx_napi);
	287	+
	288	+ return mt76x0e_init_hardware(dev, true);
	289	+}
	290	+#endif /* CONFIG_PM */
	291	+
223	292	static const struct pci_device_id mt76x0e_device_table[] = {
224	293	{ PCI_DEVICE(PCI_VENDOR_ID_MEDIATEK, 0x7610) },
225	294	{ PCI_DEVICE(PCI_VENDOR_ID_MEDIATEK, 0x7630) },
@@ -237,6 +306,10 @@ static struct pci_driver mt76x0e_driver = { · 11 unmodified lines
237	306	.id_table = mt76x0e_device_table,
238	307	.probe = mt76x0e_probe,
239	308	.remove = mt76x0e_remove,
	309	+#ifdef CONFIG_PM
	310	+ .suspend = mt76x0e_suspend,
	311	+ .resume = mt76x0e_resume,
	312	+#endif /* CONFIG_PM */
240	313	};
241	314
242	315	module_pci_driver(mt76x0e_driver);

drivers/net/wireless/mediatek/mt76/mt7921/init.c+2−2

@@ -76,8 +76,8 @@ mt7921_init_wiphy(struct ieee80211_hw *hw) · 75 unmodified lines
76	76	struct wiphy *wiphy = hw->wiphy;
77	77
78	78	hw->queues = 4;
79		− hw->max_rx_aggregation_subframes = IEEE80211_MAX_AMPDU_BUF;
80		− hw->max_tx_aggregation_subframes = IEEE80211_MAX_AMPDU_BUF;
	79	+ hw->max_rx_aggregation_subframes = 64;
	80	+ hw->max_tx_aggregation_subframes = 128;
81	81
82	82	hw->radiotap_timestamp.units_pos =
83	83	IEEE80211_RADIOTAP_TIMESTAMP_UNIT_US;

drivers/net/wireless/mediatek/mt76/mt7921/mac.c+3−2

@@ -1404,8 +1404,9 @@ void mt7921_pm_wake_work(struct work_struct *work) · 1403 unmodified lines
1404	1404	napi_schedule(&dev->mt76.napi[i]);
1405	1405	mt76_connac_pm_dequeue_skbs(mphy, &dev->pm);
1406	1406	mt7921_tx_cleanup(dev);
1407		− ieee80211_queue_delayed_work(mphy->hw, &mphy->mac_work,
1408		− MT7921_WATCHDOG_TIME);
	1407	+ if (test_bit(MT76_STATE_RUNNING, &mphy->state))
	1408	+ ieee80211_queue_delayed_work(mphy->hw, &mphy->mac_work,
	1409	+ MT7921_WATCHDOG_TIME);
1409	1410	}
1410	1411
1411	1412	ieee80211_wake_queues(mphy->hw);

drivers/net/wireless/mediatek/mt76/mt7921/main.c+1−2

@@ -74,8 +74,7 @@ mt7921_init_he_caps(struct mt7921_phy *phy, enum nl80211_band band, · 73 unmodified lines
74	74	IEEE80211_HE_PHY_CAP0_CHANNEL_WIDTH_SET_40MHZ_IN_2G;
75	75	else if (band == NL80211_BAND_5GHZ)
76	76	he_cap_elem->phy_cap_info[0] =
77		− IEEE80211_HE_PHY_CAP0_CHANNEL_WIDTH_SET_40MHZ_80MHZ_IN_5G \|
78		− IEEE80211_HE_PHY_CAP0_CHANNEL_WIDTH_SET_80PLUS80_MHZ_IN_5G;
	77	+ IEEE80211_HE_PHY_CAP0_CHANNEL_WIDTH_SET_40MHZ_80MHZ_IN_5G;
79	78
80	79	he_cap_elem->phy_cap_info[1] =
81	80	IEEE80211_HE_PHY_CAP1_LDPC_CODING_IN_PAYLOAD;

drivers/net/wireless/mediatek/mt76/mt7921/mcu.c+10−7

@@ -402,20 +402,22 @@ static void · 401 unmodified lines
402	402	mt7921_mcu_tx_rate_report(struct mt7921_dev dev, struct sk_buff skb,
403	403	u16 wlan_idx)
404	404	{
405		− struct mt7921_mcu_wlan_info_event *wtbl_info =
406		− (struct mt7921_mcu_wlan_info_event *)(skb->data);
407		− struct rate_info rate = {};
408		− u8 curr_idx = wtbl_info->rate_info.rate_idx;
409		− u16 curr = le16_to_cpu(wtbl_info->rate_info.rate[curr_idx]);
410		− struct mt7921_mcu_peer_cap peer = wtbl_info->peer_cap;
	405	+ struct mt7921_mcu_wlan_info_event *wtbl_info;
411	406	struct mt76_phy *mphy = &dev->mphy;
412	407	struct mt7921_sta_stats *stats;
	408	+ struct rate_info rate = {};
413	409	struct mt7921_sta *msta;
414	410	struct mt76_wcid *wcid;
	411	+ u8 idx;
415	412
416	413	if (wlan_idx >= MT76_N_WCIDS)
417	414	return;
418	415
	416	+ wtbl_info = (struct mt7921_mcu_wlan_info_event *)skb->data;
	417	+ idx = wtbl_info->rate_info.rate_idx;
	418	+ if (idx >= ARRAY_SIZE(wtbl_info->rate_info.rate))
	419	+ return;
	420	+
419	421	rcu_read_lock();
420	422
421	423	wcid = rcu_dereference(dev->mt76.wcid[wlan_idx]);
@@ -426,7 +428,8 @@ mt7921_mcu_tx_rate_report(struct mt7921_dev dev, struct sk_buff skb, · 4 unmodified lines
426	428	stats = &msta->stats;
427	429
428	430	/* current rate */
429		− mt7921_mcu_tx_rate_parse(mphy, &peer, &rate, curr);
	431	+ mt7921_mcu_tx_rate_parse(mphy, &wtbl_info->peer_cap, &rate,
	432	+ le16_to_cpu(wtbl_info->rate_info.rate[idx]));
430	433	stats->tx_rate = rate;
431	434	out:
432	435	rcu_read_unlock();

drivers/net/xen-netback/interface.c+6−0

@@ -684,6 +684,7 @@ static void xenvif_disconnect_queue(struct xenvif_queue *queue) · 683 unmodified lines
684	684	{
685	685	if (queue->task) {
686	686	kthread_stop(queue->task);
	687	+ put_task_struct(queue->task);
687	688	queue->task = NULL;
688	689	}
689	690
@@ -745,6 +746,11 @@ int xenvif_connect_data(struct xenvif_queue *queue, · 55 unmodified lines
745	746	if (IS_ERR(task))
746	747	goto kthread_err;
747	748	queue->task = task;
	749	+ /*
	750	+ * Take a reference to the task in order to prevent it from being freed
	751	+ * if the thread function returns before kthread_stop is called.
	752	+ */
	753	+ get_task_struct(task);
748	754
749	755	task = kthread_run(xenvif_dealloc_kthread, queue,
750	756	"%s-dealloc", queue->name);

drivers/nfc/nfcmrvl/fw_dnld.h+1−1

@@ -1,4 +1,4 @@
1		−/**
	1	+/*
2	2	* Marvell NFC driver: Firmware downloader
3	3	*
4	4	* Copyright (C) 2015, Marvell International Ltd.

drivers/nfc/nfcmrvl/i2c.c+1−1

@@ -1,4 +1,4 @@
1		−/**
	1	+/*
2	2	* Marvell NFC-over-I2C driver: I2C interface related functions
3	3	*
4	4	* Copyright (C) 2015, Marvell International Ltd.

drivers/nfc/nfcmrvl/nfcmrvl.h+1−1

@@ -1,4 +1,4 @@
1		−/**
	1	+/*
2	2	* Marvell NFC driver
3	3	*
4	4	* Copyright (C) 2014-2015, Marvell International Ltd.

drivers/nfc/nfcmrvl/spi.c+1−1

@@ -1,4 +1,4 @@
1		−/**
	1	+/*
2	2	* Marvell NFC-over-SPI driver: SPI interface related functions
3	3	*
4	4	* Copyright (C) 2015, Marvell International Ltd.

drivers/nfc/nfcmrvl/uart.c+1−1

@@ -1,4 +1,4 @@
1		−/**
	1	+/*
2	2	* Marvell NFC-over-UART driver
3	3	*
4	4	* Copyright (C) 2015, Marvell International Ltd.

drivers/nfc/nfcmrvl/usb.c+1−1

@@ -1,4 +1,4 @@
1		−/**
	1	+/*
2	2	* Marvell NFC-over-USB driver: USB interface related functions
3	3	*
4	4	* Copyright (C) 2014, Marvell International Ltd.

drivers/nvme/host/Kconfig+2−1

@@ -71,7 +71,8 @@ config NVME_FC · 70 unmodified lines
71	71	config NVME_TCP
72	72	tristate "NVM Express over Fabrics TCP host driver"
73	73	depends on INET
74		− depends on BLK_DEV_NVME
	74	+ depends on BLOCK
	75	+ select NVME_CORE
75	76	select NVME_FABRICS
76	77	select CRYPTO
77	78	select CRYPTO_CRC32C

drivers/nvme/host/core.c+3−1

@@ -3485,8 +3485,10 @@ int nvme_cdev_add(struct cdev cdev, struct device cdev_device, · 3484 unmodified lines
3485	3485	cdev_init(cdev, fops);
3486	3486	cdev->owner = owner;
3487	3487	ret = cdev_device_add(cdev, cdev_device);
3488		− if (ret)
	3488	+ if (ret) {
	3489	+ put_device(cdev_device);
3489	3490	ida_simple_remove(&nvme_ns_chr_minor_ida, minor);
	3491	+ }
3490	3492	return ret;
3491	3493	}
3492	3494

drivers/nvme/host/fabrics.c+5−0

@@ -336,6 +336,11 @@ static void nvmf_log_connect_error(struct nvme_ctrl *ctrl, · 335 unmodified lines
336	336	cmd->connect.recfmt);
337	337	break;
338	338
	339	+ case NVME_SC_HOST_PATH_ERROR:
	340	+ dev_err(ctrl->device,
	341	+ "Connect command failed: host path error\n");
	342	+ break;
	343	+
339	344	default:
340	345	dev_err(ctrl->device,
341	346	"Connect command failed, error wo/DNR bit: %d\n",

drivers/nvme/host/fc.c+17−8

@@ -3107,6 +3107,7 @@ nvme_fc_create_association(struct nvme_fc_ctrl *ctrl) · 3106 unmodified lines
3107	3107	if (ctrl->ctrl.icdoff) {
3108	3108	dev_err(ctrl->ctrl.device, "icdoff %d is not supported!\n",
3109	3109	ctrl->ctrl.icdoff);
	3110	+ ret = NVME_SC_INVALID_FIELD \| NVME_SC_DNR;
3110	3111	goto out_disconnect_admin_queue;
3111	3112	}
3112	3113
@@ -3114,6 +3115,7 @@ nvme_fc_create_association(struct nvme_fc_ctrl *ctrl) · 1 unmodified line
3114	3115	if (!(ctrl->ctrl.sgls & ((1 << 0) \| (1 << 1)))) {
3115	3116	dev_err(ctrl->ctrl.device,
3116	3117	"Mandatory sgls are not supported!\n");
	3118	+ ret = NVME_SC_INVALID_FIELD \| NVME_SC_DNR;
3117	3119	goto out_disconnect_admin_queue;
3118	3120	}
3119	3121
@@ -3280,11 +3282,13 @@ nvme_fc_reconnect_or_delete(struct nvme_fc_ctrl *ctrl, int status) · 160 unmodified lines
3280	3282	if (ctrl->ctrl.state != NVME_CTRL_CONNECTING)
3281	3283	return;
3282	3284
3283		− if (portptr->port_state == FC_OBJSTATE_ONLINE)
	3285	+ if (portptr->port_state == FC_OBJSTATE_ONLINE) {
3284	3286	dev_info(ctrl->ctrl.device,
3285	3287	"NVME-FC{%d}: reset: Reconnect attempt failed (%d)\n",
3286	3288	ctrl->cnum, status);
3287		− else if (time_after_eq(jiffies, rport->dev_loss_end))
	3289	+ if (status > 0 && (status & NVME_SC_DNR))
	3290	+ recon = false;
	3291	+ } else if (time_after_eq(jiffies, rport->dev_loss_end))
3288	3292	recon = false;
3289	3293
3290	3294	if (recon && nvmf_should_reconnect(&ctrl->ctrl)) {
@@ -3298,12 +3302,17 @@ nvme_fc_reconnect_or_delete(struct nvme_fc_ctrl *ctrl, int status) · 7 unmodified lines
3298	3302
3299	3303	queue_delayed_work(nvme_wq, &ctrl->connect_work, recon_delay);
3300	3304	} else {
3301		− if (portptr->port_state == FC_OBJSTATE_ONLINE)
3302		− dev_warn(ctrl->ctrl.device,
3303		− "NVME-FC{%d}: Max reconnect attempts (%d) "
3304		− "reached.\n",
3305		− ctrl->cnum, ctrl->ctrl.nr_reconnects);
3306		− else
	3305	+ if (portptr->port_state == FC_OBJSTATE_ONLINE) {
	3306	+ if (status > 0 && (status & NVME_SC_DNR))
	3307	+ dev_warn(ctrl->ctrl.device,
	3308	+ "NVME-FC{%d}: reconnect failure\n",
	3309	+ ctrl->cnum);
	3310	+ else
	3311	+ dev_warn(ctrl->ctrl.device,
	3312	+ "NVME-FC{%d}: Max reconnect attempts "
	3313	+ "(%d) reached.\n",
	3314	+ ctrl->cnum, ctrl->ctrl.nr_reconnects);
	3315	+ } else
3307	3316	dev_warn(ctrl->ctrl.device,
3308	3317	"NVME-FC{%d}: dev_loss_tmo (%d) expired "
3309	3318	"while waiting for remoteport connectivity.\n",

drivers/nvme/host/rdma.c+3−2

@@ -1320,16 +1320,17 @@ static int nvme_rdma_map_sg_inline(struct nvme_rdma_queue *queue, · 1319 unmodified lines
1320	1320	int count)
1321	1321	{
1322	1322	struct nvme_sgl_desc *sg = &c->common.dptr.sgl;
1323		− struct scatterlist *sgl = req->data_sgl.sg_table.sgl;
1324	1323	struct ib_sge *sge = &req->sge[1];
	1324	+ struct scatterlist *sgl;
1325	1325	u32 len = 0;
1326	1326	int i;
1327	1327
1328		− for (i = 0; i < count; i++, sgl++, sge++) {
	1328	+ for_each_sg(req->data_sgl.sg_table.sgl, sgl, count, i) {
1329	1329	sge->addr = sg_dma_address(sgl);
1330	1330	sge->length = sg_dma_len(sgl);
1331	1331	sge->lkey = queue->device->pd->local_dma_lkey;
1332	1332	len += sge->length;
	1333	+ sge++;
1333	1334	}
1334	1335
1335	1336	sg->addr = cpu_to_le64(queue->ctrl->ctrl.icdoff);

drivers/nvme/target/core.c+27−21

@@ -388,10 +388,10 @@ static void nvmet_keep_alive_timer(struct work_struct *work) · 387 unmodified lines
388	388	{
389	389	struct nvmet_ctrl *ctrl = container_of(to_delayed_work(work),
390	390	struct nvmet_ctrl, ka_work);
391		− bool cmd_seen = ctrl->cmd_seen;
	391	+ bool reset_tbkas = ctrl->reset_tbkas;
392	392
393		− ctrl->cmd_seen = false;
394		− if (cmd_seen) {
	393	+ ctrl->reset_tbkas = false;
	394	+ if (reset_tbkas) {
395	395	pr_debug("ctrl %d reschedule traffic based keep-alive timer\n",
396	396	ctrl->cntlid);
397	397	schedule_delayed_work(&ctrl->ka_work, ctrl->kato * HZ);
@@ -804,6 +804,13 @@ void nvmet_sq_destroy(struct nvmet_sq *sq) · 406 unmodified lines
804	804	percpu_ref_exit(&sq->ref);
805	805
806	806	if (ctrl) {
	807	+ /*
	808	+ * The teardown flow may take some time, and the host may not
	809	+ * send us keep-alive during this period, hence reset the
	810	+ * traffic based keep-alive timer so we don't trigger a
	811	+ * controller teardown as a result of a keep-alive expiration.
	812	+ */
	813	+ ctrl->reset_tbkas = true;
807	814	nvmet_ctrl_put(ctrl);
808	815	sq->ctrl = NULL; /* allows reusing the queue later */
809	816	}
@@ -952,7 +959,7 @@ bool nvmet_req_init(struct nvmet_req req, struct nvmet_cq cq, · 142 unmodified lines
952	959	}
953	960
954	961	if (sq->ctrl)
955		− sq->ctrl->cmd_seen = true;
	962	+ sq->ctrl->reset_tbkas = true;
956	963
957	964	return true;
958	965
@@ -998,19 +1005,23 @@ static unsigned int nvmet_data_transfer_len(struct nvmet_req *req) · 39 unmodified lines
998	1005	return req->transfer_len - req->metadata_len;
999	1006	}
1000	1007
1001		−static int nvmet_req_alloc_p2pmem_sgls(struct nvmet_req *req)
	1008	+static int nvmet_req_alloc_p2pmem_sgls(struct pci_dev *p2p_dev,
	1009	+ struct nvmet_req *req)
1002	1010	{
1003		− req->sg = pci_p2pmem_alloc_sgl(req->p2p_dev, &req->sg_cnt,
	1011	+ req->sg = pci_p2pmem_alloc_sgl(p2p_dev, &req->sg_cnt,
1004	1012	nvmet_data_transfer_len(req));
1005	1013	if (!req->sg)
1006	1014	goto out_err;
1007	1015
1008	1016	if (req->metadata_len) {
1009		− req->metadata_sg = pci_p2pmem_alloc_sgl(req->p2p_dev,
	1017	+ req->metadata_sg = pci_p2pmem_alloc_sgl(p2p_dev,
1010	1018	&req->metadata_sg_cnt, req->metadata_len);
1011	1019	if (!req->metadata_sg)
1012	1020	goto out_free_sg;
1013	1021	}
	1022	+
	1023	+ req->p2p_dev = p2p_dev;
	1024	+
1014	1025	return 0;
1015	1026	out_free_sg:
1016	1027	pci_p2pmem_free_sgl(req->p2p_dev, req->sg);
@@ -1018,25 +1029,19 @@ out_err: · 1 unmodified line
1018	1029	return -ENOMEM;
1019	1030	}
1020	1031
1021		−static bool nvmet_req_find_p2p_dev(struct nvmet_req *req)
	1032	+static struct pci_dev nvmet_req_find_p2p_dev(struct nvmet_req req)
1022	1033	{
1023		− if (!IS_ENABLED(CONFIG_PCI_P2PDMA))
1024		− return false;
1025		−
1026		− if (req->sq->ctrl && req->sq->qid && req->ns) {
1027		− req->p2p_dev = radix_tree_lookup(&req->sq->ctrl->p2p_ns_map,
1028		− req->ns->nsid);
1029		− if (req->p2p_dev)
1030		− return true;
1031		− }
1032		−
1033		− req->p2p_dev = NULL;
1034		− return false;
	1034	+ if (!IS_ENABLED(CONFIG_PCI_P2PDMA) \|\|
	1035	+ !req->sq->ctrl \|\| !req->sq->qid \|\| !req->ns)
	1036	+ return NULL;
	1037	+ return radix_tree_lookup(&req->sq->ctrl->p2p_ns_map, req->ns->nsid);
1035	1038	}
1036	1039
1037	1040	int nvmet_req_alloc_sgls(struct nvmet_req *req)
1038	1041	{
1039		− if (nvmet_req_find_p2p_dev(req) && !nvmet_req_alloc_p2pmem_sgls(req))
	1042	+ struct pci_dev *p2p_dev = nvmet_req_find_p2p_dev(req);
	1043	+
	1044	+ if (p2p_dev && !nvmet_req_alloc_p2pmem_sgls(p2p_dev, req))
1040	1045	return 0;
1041	1046
1042	1047	req->sg = sgl_alloc(nvmet_data_transfer_len(req), GFP_KERNEL,
@@ -1065,6 +1070,7 @@ void nvmet_req_free_sgls(struct nvmet_req *req) · 22 unmodified lines
1065	1070	pci_p2pmem_free_sgl(req->p2p_dev, req->sg);
1066	1071	if (req->metadata_sg)
1067	1072	pci_p2pmem_free_sgl(req->p2p_dev, req->metadata_sg);
	1073	+ req->p2p_dev = NULL;
1068	1074	} else {
1069	1075	sgl_free(req->sg);
1070	1076	if (req->metadata_sg)

drivers/nvme/target/loop.c+8−3

@@ -263,7 +263,8 @@ static const struct blk_mq_ops nvme_loop_admin_mq_ops = { · 262 unmodified lines
263	263
264	264	static void nvme_loop_destroy_admin_queue(struct nvme_loop_ctrl *ctrl)
265	265	{
266		− clear_bit(NVME_LOOP_Q_LIVE, &ctrl->queues[0].flags);
	266	+ if (!test_and_clear_bit(NVME_LOOP_Q_LIVE, &ctrl->queues[0].flags))
	267	+ return;
267	268	nvmet_sq_destroy(&ctrl->queues[0].nvme_sq);
268	269	blk_cleanup_queue(ctrl->ctrl.admin_q);
269	270	blk_cleanup_queue(ctrl->ctrl.fabrics_q);
@@ -299,6 +300,7 @@ static void nvme_loop_destroy_io_queues(struct nvme_loop_ctrl *ctrl) · 29 unmodified lines
299	300	clear_bit(NVME_LOOP_Q_LIVE, &ctrl->queues[i].flags);
300	301	nvmet_sq_destroy(&ctrl->queues[i].nvme_sq);
301	302	}
	303	+ ctrl->ctrl.queue_count = 1;
302	304	}
303	305
304	306	static int nvme_loop_init_io_queues(struct nvme_loop_ctrl *ctrl)
@@ -405,6 +407,7 @@ static int nvme_loop_configure_admin_queue(struct nvme_loop_ctrl *ctrl) · 100 unmodified lines
405	407	return 0;
406	408
407	409	out_cleanup_queue:
	410	+ clear_bit(NVME_LOOP_Q_LIVE, &ctrl->queues[0].flags);
408	411	blk_cleanup_queue(ctrl->ctrl.admin_q);
409	412	out_cleanup_fabrics_q:
410	413	blk_cleanup_queue(ctrl->ctrl.fabrics_q);
@@ -462,8 +465,10 @@ static void nvme_loop_reset_ctrl_work(struct work_struct *work) · 51 unmodified lines
462	465	nvme_loop_shutdown_ctrl(ctrl);
463	466
464	467	if (!nvme_change_ctrl_state(&ctrl->ctrl, NVME_CTRL_CONNECTING)) {
465		− /* state change failure should never happen */
466		− WARN_ON_ONCE(1);
	468	+ if (ctrl->ctrl.state != NVME_CTRL_DELETING &&
	469	+ ctrl->ctrl.state != NVME_CTRL_DELETING_NOIO)
	470	+ /* state change failure for non-deleted ctrl? */
	471	+ WARN_ON_ONCE(1);
467	472	return;
468	473	}
469	474

drivers/nvme/target/nvmet.h+1−1

@@ -167,7 +167,7 @@ struct nvmet_ctrl { · 166 unmodified lines
167	167	struct nvmet_subsys *subsys;
168	168	struct nvmet_sq **sqs;
169	169
170		− bool cmd_seen;
	170	+ bool reset_tbkas;
171	171
172	172	struct mutex lock;
173	173	u64 cap;

drivers/nvme/target/tcp.c+1−1

@@ -550,7 +550,7 @@ static void nvmet_tcp_queue_response(struct nvmet_req *req) · 549 unmodified lines
550	550	* nvmet_req_init is completed.
551	551	*/
552	552	if (queue->rcv_state == NVMET_TCP_RECV_PDU &&
553		− len && len < cmd->req.port->inline_data_size &&
	553	+ len && len <= cmd->req.port->inline_data_size &&
554	554	nvme_is_write(cmd->req.cmd))
555	555	return;
556	556	}

drivers/pci/controller/dwc/Makefile+2−1

@@ -18,6 +18,7 @@ obj-$(CONFIG_PCIE_INTEL_GW) += pcie-intel-gw.o · 17 unmodified lines
18	18	obj-$(CONFIG_PCIE_KIRIN) += pcie-kirin.o
19	19	obj-$(CONFIG_PCIE_HISI_STB) += pcie-histb.o
20	20	obj-$(CONFIG_PCI_MESON) += pci-meson.o
	21	+obj-$(CONFIG_PCIE_TEGRA194) += pcie-tegra194.o
21	22	obj-$(CONFIG_PCIE_UNIPHIER) += pcie-uniphier.o
22	23	obj-$(CONFIG_PCIE_UNIPHIER_EP) += pcie-uniphier-ep.o
23	24
@@ -38,6 +39,6 @@ ifdef CONFIG_ACPI · 14 unmodified lines
38	39	ifdef CONFIG_PCI_QUIRKS
39	40	obj-$(CONFIG_ARM64) += pcie-al.o
40	41	obj-$(CONFIG_ARM64) += pcie-hisi.o
41		−obj-$(CONFIG_ARM64) += pcie-tegra194.o
	42	+obj-$(CONFIG_ARM64) += pcie-tegra194-acpi.o
42	43	endif
43	44	endif

drivers/pci/controller/dwc/pcie-tegra194-acpi.c+108−0

@@ -0,0 +1,108 @@
	1	+// SPDX-License-Identifier: GPL-2.0+
	2	+/*
	3	+ * ACPI quirks for Tegra194 PCIe host controller
	4	+ *
	5	+ * Copyright (C) 2021 NVIDIA Corporation.
	6	+ *
	7	+ * Author: Vidya Sagar <vidyas@nvidia.com>
	8	+ */
	9	+
	10	+#include <linux/pci.h>
	11	+#include <linux/pci-acpi.h>
	12	+#include <linux/pci-ecam.h>
	13	+
	14	+#include "pcie-designware.h"
	15	+
	16	+struct tegra194_pcie_ecam {
	17	+ void __iomem *config_base;
	18	+ void __iomem *iatu_base;
	19	+ void __iomem *dbi_base;
	20	+};
	21	+
	22	+static int tegra194_acpi_init(struct pci_config_window *cfg)
	23	+{
	24	+ struct device *dev = cfg->parent;
	25	+ struct tegra194_pcie_ecam *pcie_ecam;
	26	+
	27	+ pcie_ecam = devm_kzalloc(dev, sizeof(*pcie_ecam), GFP_KERNEL);
	28	+ if (!pcie_ecam)
	29	+ return -ENOMEM;
	30	+
	31	+ pcie_ecam->config_base = cfg->win;
	32	+ pcie_ecam->iatu_base = cfg->win + SZ_256K;
	33	+ pcie_ecam->dbi_base = cfg->win + SZ_512K;
	34	+ cfg->priv = pcie_ecam;
	35	+
	36	+ return 0;
	37	+}
	38	+
	39	+static void atu_reg_write(struct tegra194_pcie_ecam *pcie_ecam, int index,
	40	+ u32 val, u32 reg)
	41	+{
	42	+ u32 offset = PCIE_GET_ATU_OUTB_UNR_REG_OFFSET(index);
	43	+
	44	+ writel(val, pcie_ecam->iatu_base + offset + reg);
	45	+}
	46	+
	47	+static void program_outbound_atu(struct tegra194_pcie_ecam *pcie_ecam,
	48	+ int index, int type, u64 cpu_addr,
	49	+ u64 pci_addr, u64 size)
	50	+{
	51	+ atu_reg_write(pcie_ecam, index, lower_32_bits(cpu_addr),
	52	+ PCIE_ATU_LOWER_BASE);
	53	+ atu_reg_write(pcie_ecam, index, upper_32_bits(cpu_addr),
	54	+ PCIE_ATU_UPPER_BASE);
	55	+ atu_reg_write(pcie_ecam, index, lower_32_bits(pci_addr),
	56	+ PCIE_ATU_LOWER_TARGET);
	57	+ atu_reg_write(pcie_ecam, index, lower_32_bits(cpu_addr + size - 1),
	58	+ PCIE_ATU_LIMIT);
	59	+ atu_reg_write(pcie_ecam, index, upper_32_bits(pci_addr),
	60	+ PCIE_ATU_UPPER_TARGET);
	61	+ atu_reg_write(pcie_ecam, index, type, PCIE_ATU_CR1);
	62	+ atu_reg_write(pcie_ecam, index, PCIE_ATU_ENABLE, PCIE_ATU_CR2);
	63	+}
	64	+
	65	+static void __iomem tegra194_map_bus(struct pci_bus bus,
	66	+ unsigned int devfn, int where)
	67	+{
	68	+ struct pci_config_window *cfg = bus->sysdata;
	69	+ struct tegra194_pcie_ecam *pcie_ecam = cfg->priv;
	70	+ u32 busdev;
	71	+ int type;
	72	+
	73	+ if (bus->number < cfg->busr.start \|\| bus->number > cfg->busr.end)
	74	+ return NULL;
	75	+
	76	+ if (bus->number == cfg->busr.start) {
	77	+ if (PCI_SLOT(devfn) == 0)
	78	+ return pcie_ecam->dbi_base + where;
	79	+ else
	80	+ return NULL;
	81	+ }
	82	+
	83	+ busdev = PCIE_ATU_BUS(bus->number) \| PCIE_ATU_DEV(PCI_SLOT(devfn)) \|
	84	+ PCIE_ATU_FUNC(PCI_FUNC(devfn));
	85	+
	86	+ if (bus->parent->number == cfg->busr.start) {
	87	+ if (PCI_SLOT(devfn) == 0)
	88	+ type = PCIE_ATU_TYPE_CFG0;
	89	+ else
	90	+ return NULL;
	91	+ } else {
	92	+ type = PCIE_ATU_TYPE_CFG1;
	93	+ }
	94	+
	95	+ program_outbound_atu(pcie_ecam, 0, type, cfg->res.start, busdev,
	96	+ SZ_256K);
	97	+
	98	+ return pcie_ecam->config_base + where;
	99	+}
	100	+
	101	+const struct pci_ecam_ops tegra194_pcie_ops = {
	102	+ .init = tegra194_acpi_init,
	103	+ .pci_ops = {
	104	+ .map_bus = tegra194_map_bus,
	105	+ .read = pci_generic_config_read,
	106	+ .write = pci_generic_config_write,
	107	+ }
	108	+};

drivers/pci/controller/dwc/pcie-tegra194.c+18−120

@@ -22,8 +22,6 @@ · 21 unmodified lines
22	22	#include <linux/of_irq.h>
23	23	#include <linux/of_pci.h>
24	24	#include <linux/pci.h>
25		−#include <linux/pci-acpi.h>
26		−#include <linux/pci-ecam.h>
27	25	#include <linux/phy/phy.h>
28	26	#include <linux/pinctrl/consumer.h>
29	27	#include <linux/platform_device.h>
@@ -247,24 +245,6 @@ static const unsigned int pcie_gen_freq[] = { · 217 unmodified lines
247	245	GEN4_CORE_CLK_FREQ
248	246	};
249	247
250		−static const u32 event_cntr_ctrl_offset[] = {
251		− 0x1d8,
252		− 0x1a8,
253		− 0x1a8,
254		− 0x1a8,
255		− 0x1c4,
256		− 0x1d8
257		−};
258		−
259		−static const u32 event_cntr_data_offset[] = {
260		− 0x1dc,
261		− 0x1ac,
262		− 0x1ac,
263		− 0x1ac,
264		− 0x1c8,
265		− 0x1dc
266		−};
267		−
268	248	struct tegra_pcie_dw {
269	249	struct device *dev;
270	250	struct resource *appl_res;
@@ -313,104 +293,6 @@ struct tegra_pcie_dw_of_data { · 42 unmodified lines
313	293	enum dw_pcie_device_mode mode;
314	294	};
315	295
316		−#if defined(CONFIG_ACPI) && defined(CONFIG_PCI_QUIRKS)
317		−struct tegra194_pcie_ecam {
318		− void __iomem *config_base;
319		− void __iomem *iatu_base;
320		− void __iomem *dbi_base;
321		−};
322		−
323		−static int tegra194_acpi_init(struct pci_config_window *cfg)
324		−{
325		− struct device *dev = cfg->parent;
326		− struct tegra194_pcie_ecam *pcie_ecam;
327		−
328		− pcie_ecam = devm_kzalloc(dev, sizeof(*pcie_ecam), GFP_KERNEL);
329		− if (!pcie_ecam)
330		− return -ENOMEM;
331		−
332		− pcie_ecam->config_base = cfg->win;
333		− pcie_ecam->iatu_base = cfg->win + SZ_256K;
334		− pcie_ecam->dbi_base = cfg->win + SZ_512K;
335		− cfg->priv = pcie_ecam;
336		−
337		− return 0;
338		−}
339		−
340		−static void atu_reg_write(struct tegra194_pcie_ecam *pcie_ecam, int index,
341		− u32 val, u32 reg)
342		−{
343		− u32 offset = PCIE_GET_ATU_OUTB_UNR_REG_OFFSET(index);
344		−
345		− writel(val, pcie_ecam->iatu_base + offset + reg);
346		−}
347		−
348		−static void program_outbound_atu(struct tegra194_pcie_ecam *pcie_ecam,
349		− int index, int type, u64 cpu_addr,
350		− u64 pci_addr, u64 size)
351		−{
352		− atu_reg_write(pcie_ecam, index, lower_32_bits(cpu_addr),
353		− PCIE_ATU_LOWER_BASE);
354		− atu_reg_write(pcie_ecam, index, upper_32_bits(cpu_addr),
355		− PCIE_ATU_UPPER_BASE);
356		− atu_reg_write(pcie_ecam, index, lower_32_bits(pci_addr),
357		− PCIE_ATU_LOWER_TARGET);
358		− atu_reg_write(pcie_ecam, index, lower_32_bits(cpu_addr + size - 1),
359		− PCIE_ATU_LIMIT);
360		− atu_reg_write(pcie_ecam, index, upper_32_bits(pci_addr),
361		− PCIE_ATU_UPPER_TARGET);
362		− atu_reg_write(pcie_ecam, index, type, PCIE_ATU_CR1);
363		− atu_reg_write(pcie_ecam, index, PCIE_ATU_ENABLE, PCIE_ATU_CR2);
364		−}
365		−
366		−static void __iomem tegra194_map_bus(struct pci_bus bus,
367		− unsigned int devfn, int where)
368		−{
369		− struct pci_config_window *cfg = bus->sysdata;
370		− struct tegra194_pcie_ecam *pcie_ecam = cfg->priv;
371		− u32 busdev;
372		− int type;
373		−
374		− if (bus->number < cfg->busr.start \|\| bus->number > cfg->busr.end)
375		− return NULL;
376		−
377		− if (bus->number == cfg->busr.start) {
378		− if (PCI_SLOT(devfn) == 0)
379		− return pcie_ecam->dbi_base + where;
380		− else
381		− return NULL;
382		− }
383		−
384		− busdev = PCIE_ATU_BUS(bus->number) \| PCIE_ATU_DEV(PCI_SLOT(devfn)) \|
385		− PCIE_ATU_FUNC(PCI_FUNC(devfn));
386		−
387		− if (bus->parent->number == cfg->busr.start) {
388		− if (PCI_SLOT(devfn) == 0)
389		− type = PCIE_ATU_TYPE_CFG0;
390		− else
391		− return NULL;
392		− } else {
393		− type = PCIE_ATU_TYPE_CFG1;
394		− }
395		−
396		− program_outbound_atu(pcie_ecam, 0, type, cfg->res.start, busdev,
397		− SZ_256K);
398		−
399		− return pcie_ecam->config_base + where;
400		−}
401		−
402		−const struct pci_ecam_ops tegra194_pcie_ops = {
403		− .init = tegra194_acpi_init,
404		− .pci_ops = {
405		− .map_bus = tegra194_map_bus,
406		− .read = pci_generic_config_read,
407		− .write = pci_generic_config_write,
408		− }
409		−};
410		−#endif /* defined(CONFIG_ACPI) && defined(CONFIG_PCI_QUIRKS) */
411		−
412		−#ifdef CONFIG_PCIE_TEGRA194
413		−
414	296	static inline struct tegra_pcie_dw to_tegra_pcie(struct dw_pcie pci)
415	297	{
416	298	return container_of(pci, struct tegra_pcie_dw, pci);
@@ -694,6 +576,24 @@ static struct pci_ops tegra_pci_ops = { · 277 unmodified lines
694	576	};
695	577
696	578	#if defined(CONFIG_PCIEASPM)
	579	+static const u32 event_cntr_ctrl_offset[] = {
	580	+ 0x1d8,
	581	+ 0x1a8,
	582	+ 0x1a8,
	583	+ 0x1a8,
	584	+ 0x1c4,
	585	+ 0x1d8
	586	+};
	587	+
	588	+static const u32 event_cntr_data_offset[] = {
	589	+ 0x1dc,
	590	+ 0x1ac,
	591	+ 0x1ac,
	592	+ 0x1ac,
	593	+ 0x1c8,
	594	+ 0x1dc
	595	+};
	596	+
697	597	static void disable_aspm_l11(struct tegra_pcie_dw *pcie)
698	598	{
699	599	u32 val;
@@ -2411,5 +2311,3 @@ MODULE_DEVICE_TABLE(of, tegra_pcie_dw_of_match); · 1711 unmodified lines
2411	2311	MODULE_AUTHOR("Vidya Sagar <vidyas@nvidia.com>");
2412	2312	MODULE_DESCRIPTION("NVIDIA PCIe host controller driver");
2413	2313	MODULE_LICENSE("GPL v2");
2414		−
2415		−#endif /* CONFIG_PCIE_TEGRA194 */

drivers/pci/controller/pci-aardvark.c+40−9

@@ -514,7 +514,7 @@ static int advk_pcie_wait_pio(struct advk_pcie *pcie) · 513 unmodified lines
514	514	udelay(PIO_RETRY_DELAY);
515	515	}
516	516
517		− dev_err(dev, "config read/write timed out\n");
	517	+ dev_err(dev, "PIO read/write transfer time out\n");
518	518	return -ETIMEDOUT;
519	519	}
520	520
@@ -657,6 +657,35 @@ static bool advk_pcie_valid_device(struct advk_pcie pcie, struct pci_bus bus, · 136 unmodified lines
657	657	return true;
658	658	}
659	659
	660	+static bool advk_pcie_pio_is_running(struct advk_pcie *pcie)
	661	+{
	662	+ struct device *dev = &pcie->pdev->dev;
	663	+
	664	+ /*
	665	+ * Trying to start a new PIO transfer when previous has not completed
	666	+ * cause External Abort on CPU which results in kernel panic:
	667	+ *
	668	+ * SError Interrupt on CPU0, code 0xbf000002 -- SError
	669	+ * Kernel panic - not syncing: Asynchronous SError Interrupt
	670	+ *
	671	+ * Functions advk_pcie_rd_conf() and advk_pcie_wr_conf() are protected
	672	+ * by raw_spin_lock_irqsave() at pci_lock_config() level to prevent
	673	+ * concurrent calls at the same time. But because PIO transfer may take
	674	+ * about 1.5s when link is down or card is disconnected, it means that
	675	+ * advk_pcie_wait_pio() does not always have to wait for completion.
	676	+ *
	677	+ * Some versions of ARM Trusted Firmware handles this External Abort at
	678	+ * EL3 level and mask it to prevent kernel panic. Relevant TF-A commit:
	679	+ * https://git.trustedfirmware.org/TF-A/trusted-firmware-a.git/commit/?id=3c7dcdac5c50
	680	+ */
	681	+ if (advk_readl(pcie, PIO_START)) {
	682	+ dev_err(dev, "Previous PIO read/write transfer is still running\n");
	683	+ return true;
	684	+ }
	685	+
	686	+ return false;
	687	+}
	688	+
660	689	static int advk_pcie_rd_conf(struct pci_bus *bus, u32 devfn,
661	690	int where, int size, u32 *val)
662	691	{
@@ -673,9 +702,10 @@ static int advk_pcie_rd_conf(struct pci_bus *bus, u32 devfn, · 10 unmodified lines
673	702	return pci_bridge_emul_conf_read(&pcie->bridge, where,
674	703	size, val);
675	704
676		− /* Start PIO */
677		− advk_writel(pcie, 0, PIO_START);
678		− advk_writel(pcie, 1, PIO_ISR);
	705	+ if (advk_pcie_pio_is_running(pcie)) {
	706	+ *val = 0xffffffff;
	707	+ return PCIBIOS_SET_FAILED;
	708	+ }
679	709
680	710	/* Program the control register */
681	711	reg = advk_readl(pcie, PIO_CTRL);
@@ -694,7 +724,8 @@ static int advk_pcie_rd_conf(struct pci_bus *bus, u32 devfn, · 12 unmodified lines
694	724	/* Program the data strobe */
695	725	advk_writel(pcie, 0xf, PIO_WR_DATA_STRB);
696	726
697		− /* Start the transfer */
	727	+ /* Clear PIO DONE ISR and start the transfer */
	728	+ advk_writel(pcie, 1, PIO_ISR);
698	729	advk_writel(pcie, 1, PIO_START);
699	730
700	731	ret = advk_pcie_wait_pio(pcie);
@@ -734,9 +765,8 @@ static int advk_pcie_wr_conf(struct pci_bus *bus, u32 devfn, · 33 unmodified lines
734	765	if (where % size)
735	766	return PCIBIOS_SET_FAILED;
736	767
737		− /* Start PIO */
738		− advk_writel(pcie, 0, PIO_START);
739		− advk_writel(pcie, 1, PIO_ISR);
	768	+ if (advk_pcie_pio_is_running(pcie))
	769	+ return PCIBIOS_SET_FAILED;
740	770
741	771	/* Program the control register */
742	772	reg = advk_readl(pcie, PIO_CTRL);
@@ -763,7 +793,8 @@ static int advk_pcie_wr_conf(struct pci_bus *bus, u32 devfn, · 20 unmodified lines
763	793	/* Program the data strobe */
764	794	advk_writel(pcie, data_strobe, PIO_WR_DATA_STRB);
765	795
766		− /* Start the transfer */
	796	+ /* Clear PIO DONE ISR and start the transfer */
	797	+ advk_writel(pcie, 1, PIO_ISR);
767	798	advk_writel(pcie, 1, PIO_START);
768	799
769	800	ret = advk_pcie_wait_pio(pcie);

drivers/pci/of.c+9−0

@@ -103,6 +103,13 @@ struct irq_domain pci_host_bridge_of_msi_domain(struct pci_bus bus) · 102 unmodified lines
103	103	#endif
104	104	}
105	105
	106	+bool pci_host_of_has_msi_map(struct device *dev)
	107	+{
	108	+ if (dev && dev->of_node)
	109	+ return of_get_property(dev->of_node, "msi-map", NULL);
	110	+ return false;
	111	+}
	112	+
106	113	static inline int __of_pci_pci_compare(struct device_node *node,
107	114	unsigned int data)
108	115	{
@@ -346,6 +353,8 @@ static int devm_of_pci_get_host_bridge_resources(struct device *dev, · 237 unmodified lines
346	353	dev_warn(dev, "More than one I/O resource converted for %pOF. CPU base address for old range lost!\n",
347	354	dev_node);
348	355	*io_base = range.cpu_addr;
	356	+ } else if (resource_type(res) == IORESOURCE_MEM) {
	357	+ res->flags &= ~IORESOURCE_MEM_64;
349	358	}
350	359
351	360	pci_add_resource_offset(resources, res, res->start - range.pci_addr);

drivers/pci/probe.c+2−1

@@ -925,7 +925,8 @@ static int pci_register_host_bridge(struct pci_host_bridge *bridge) · 924 unmodified lines
925	925	device_enable_async_suspend(bus->bridge);
926	926	pci_set_bus_of_node(bus);
927	927	pci_set_bus_msi_domain(bus);
928		− if (bridge->msi_domain && !dev_get_msi_domain(&bus->dev))
	928	+ if (bridge->msi_domain && !dev_get_msi_domain(&bus->dev) &&
	929	+ !pci_host_of_has_msi_map(parent))
929	930	bus->bus_flags \|= PCI_BUS_FLAGS_NO_MSI;
930	931
931	932	if (!parent)

drivers/pci/quirks.c+92−1

@@ -3546,6 +3546,18 @@ static void quirk_no_bus_reset(struct pci_dev *dev) · 3545 unmodified lines
3546	3546	dev->dev_flags \|= PCI_DEV_FLAGS_NO_BUS_RESET;
3547	3547	}
3548	3548
	3549	+/*
	3550	+ * Some NVIDIA GPU devices do not work with bus reset, SBR needs to be
	3551	+ * prevented for those affected devices.
	3552	+ */
	3553	+static void quirk_nvidia_no_bus_reset(struct pci_dev *dev)
	3554	+{
	3555	+ if ((dev->device & 0xffc0) == 0x2340)
	3556	+ quirk_no_bus_reset(dev);
	3557	+}
	3558	+DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_NVIDIA, PCI_ANY_ID,
	3559	+ quirk_nvidia_no_bus_reset);
	3560	+
3549	3561	/*
3550	3562	* Some Atheros AR9xxx and QCA988x chips do not behave after a bus reset.
3551	3563	* The device will throw a Link Down error on AER-capable systems and
@@ -3566,6 +3578,16 @@ DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_ATHEROS, 0x0034, quirk_no_bus_reset); · 14 unmodified lines
3566	3578	*/
3567	3579	DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_CAVIUM, 0xa100, quirk_no_bus_reset);
3568	3580
	3581	+/*
	3582	+ * Some TI KeyStone C667X devices do not support bus/hot reset. The PCIESS
	3583	+ * automatically disables LTSSM when Secondary Bus Reset is received and
	3584	+ * the device stops working. Prevent bus reset for these devices. With
	3585	+ * this change, the device can be assigned to VMs with VFIO, but it will
	3586	+ * leak state between VMs. Reference
	3587	+ * https://e2e.ti.com/support/processors/f/791/t/954382
	3588	+ */
	3589	+DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_TI, 0xb005, quirk_no_bus_reset);
	3590	+
3569	3591	static void quirk_no_pm_reset(struct pci_dev *dev)
3570	3592	{
3571	3593	/*
@@ -3901,6 +3923,69 @@ static int delay_250ms_after_flr(struct pci_dev *dev, int probe) · 329 unmodified lines
3901	3923	return 0;
3902	3924	}
3903	3925
	3926	+#define PCI_DEVICE_ID_HINIC_VF 0x375E
	3927	+#define HINIC_VF_FLR_TYPE 0x1000
	3928	+#define HINIC_VF_FLR_CAP_BIT (1UL << 30)
	3929	+#define HINIC_VF_OP 0xE80
	3930	+#define HINIC_VF_FLR_PROC_BIT (1UL << 18)
	3931	+#define HINIC_OPERATION_TIMEOUT 15000 /* 15 seconds */
	3932	+
	3933	+/* Device-specific reset method for Huawei Intelligent NIC virtual functions */
	3934	+static int reset_hinic_vf_dev(struct pci_dev *pdev, int probe)
	3935	+{
	3936	+ unsigned long timeout;
	3937	+ void __iomem *bar;
	3938	+ u32 val;
	3939	+
	3940	+ if (probe)
	3941	+ return 0;
	3942	+
	3943	+ bar = pci_iomap(pdev, 0, 0);
	3944	+ if (!bar)
	3945	+ return -ENOTTY;
	3946	+
	3947	+ /* Get and check firmware capabilities */
	3948	+ val = ioread32be(bar + HINIC_VF_FLR_TYPE);
	3949	+ if (!(val & HINIC_VF_FLR_CAP_BIT)) {
	3950	+ pci_iounmap(pdev, bar);
	3951	+ return -ENOTTY;
	3952	+ }
	3953	+
	3954	+ /* Set HINIC_VF_FLR_PROC_BIT for the start of FLR */
	3955	+ val = ioread32be(bar + HINIC_VF_OP);
	3956	+ val = val \| HINIC_VF_FLR_PROC_BIT;
	3957	+ iowrite32be(val, bar + HINIC_VF_OP);
	3958	+
	3959	+ pcie_flr(pdev);
	3960	+
	3961	+ /*
	3962	+ * The device must recapture its Bus and Device Numbers after FLR
	3963	+ * in order generate Completions. Issue a config write to let the
	3964	+ * device capture this information.
	3965	+ */
	3966	+ pci_write_config_word(pdev, PCI_VENDOR_ID, 0);
	3967	+
	3968	+ /* Firmware clears HINIC_VF_FLR_PROC_BIT when reset is complete */
	3969	+ timeout = jiffies + msecs_to_jiffies(HINIC_OPERATION_TIMEOUT);
	3970	+ do {
	3971	+ val = ioread32be(bar + HINIC_VF_OP);
	3972	+ if (!(val & HINIC_VF_FLR_PROC_BIT))
	3973	+ goto reset_complete;
	3974	+ msleep(20);
	3975	+ } while (time_before(jiffies, timeout));
	3976	+
	3977	+ val = ioread32be(bar + HINIC_VF_OP);
	3978	+ if (!(val & HINIC_VF_FLR_PROC_BIT))
	3979	+ goto reset_complete;
	3980	+
	3981	+ pci_warn(pdev, "Reset dev timeout, FLR ack reg: %#010x\n", val);
	3982	+
	3983	+reset_complete:
	3984	+ pci_iounmap(pdev, bar);
	3985	+
	3986	+ return 0;
	3987	+}
	3988	+
3904	3989	static const struct pci_dev_reset_methods pci_dev_reset_methods[] = {
3905	3990	{ PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_82599_SFP_VF,
3906	3991	reset_intel_82599_sfp_virtfn },
@@ -3913,6 +3998,8 @@ static const struct pci_dev_reset_methods pci_dev_reset_methods[] = { · 6 unmodified lines
3913	3998	{ PCI_VENDOR_ID_INTEL, 0x0a54, delay_250ms_after_flr },
3914	3999	{ PCI_VENDOR_ID_CHELSIO, PCI_ANY_ID,
3915	4000	reset_chelsio_generic_dev },
	4001	+ { PCI_VENDOR_ID_HUAWEI, PCI_DEVICE_ID_HINIC_VF,
	4002	+ reset_hinic_vf_dev },
3916	4003	{ 0 }
3917	4004	};
3918	4005
@@ -4753,6 +4840,8 @@ static const struct pci_dev_acs_enabled { · 834 unmodified lines
4753	4840	{ PCI_VENDOR_ID_AMPERE, 0xE00A, pci_quirk_xgene_acs },
4754	4841	{ PCI_VENDOR_ID_AMPERE, 0xE00B, pci_quirk_xgene_acs },
4755	4842	{ PCI_VENDOR_ID_AMPERE, 0xE00C, pci_quirk_xgene_acs },
	4843	+ /* Broadcom multi-function device */
	4844	+ { PCI_VENDOR_ID_BROADCOM, 0x16D7, pci_quirk_mf_endpoint_acs },
4756	4845	{ PCI_VENDOR_ID_BROADCOM, 0xD714, pci_quirk_brcm_acs },
4757	4846	/* Amazon Annapurna Labs */
4758	4847	{ PCI_VENDOR_ID_AMAZON_ANNAPURNA_LABS, 0x0031, pci_quirk_al_acs },
@@ -5154,7 +5243,8 @@ DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_SERVERWORKS, 0x0422, quirk_no_ext_tags); · 395 unmodified lines
5154	5243	static void quirk_amd_harvest_no_ats(struct pci_dev *pdev)
5155	5244	{
5156	5245	if ((pdev->device == 0x7312 && pdev->revision != 0x00) \|\|
5157		− (pdev->device == 0x7340 && pdev->revision != 0xc5))
	5246	+ (pdev->device == 0x7340 && pdev->revision != 0xc5) \|\|
	5247	+ (pdev->device == 0x7341 && pdev->revision != 0x00))
5158	5248	return;
5159	5249
5160	5250	if (pdev->device == 0x15d8) {
@@ -5181,6 +5271,7 @@ DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_ATI, 0x6900, quirk_amd_harvest_no_ats); · 20 unmodified lines
5181	5271	DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_ATI, 0x7312, quirk_amd_harvest_no_ats);
5182	5272	/* AMD Navi14 dGPU */
5183	5273	DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_ATI, 0x7340, quirk_amd_harvest_no_ats);
	5274	+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_ATI, 0x7341, quirk_amd_harvest_no_ats);
5184	5275	/* AMD Raven platform iGPU */
5185	5276	DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_ATI, 0x15d8, quirk_amd_harvest_no_ats);
5186	5277	#endif /* CONFIG_PCI_ATS */

drivers/phy/broadcom/phy-brcm-usb-init.h+2−2

@@ -78,7 +78,7 @@ static inline u32 brcm_usb_readl(void __iomem *addr) · 77 unmodified lines
78	78	* Other architectures (e.g., ARM) either do not support big endian, or
79	79	* else leave I/O in little endian mode.
80	80	*/
81		− if (IS_ENABLED(CONFIG_MIPS) && IS_ENABLED(__BIG_ENDIAN))
	81	+ if (IS_ENABLED(CONFIG_MIPS) && IS_ENABLED(CONFIG_CPU_BIG_ENDIAN))
82	82	return __raw_readl(addr);
83	83	else
84	84	return readl_relaxed(addr);
@@ -87,7 +87,7 @@ static inline u32 brcm_usb_readl(void __iomem *addr) · 2 unmodified lines
87	87	static inline void brcm_usb_writel(u32 val, void __iomem *addr)
88	88	{
89	89	/* See brcmnand_readl() comments */
90		− if (IS_ENABLED(CONFIG_MIPS) && IS_ENABLED(__BIG_ENDIAN))
	90	+ if (IS_ENABLED(CONFIG_MIPS) && IS_ENABLED(CONFIG_CPU_BIG_ENDIAN))
91	91	__raw_writel(val, addr);
92	92	else
93	93	writel_relaxed(val, addr);

drivers/phy/cadence/phy-cadence-sierra.c+1−0

@@ -940,6 +940,7 @@ static int cdns_sierra_phy_probe(struct platform_device *pdev) · 939 unmodified lines
940	940	sp->nsubnodes = node;
941	941
942	942	if (sp->num_lanes > SIERRA_MAX_LANES) {
	943	+ ret = -EINVAL;
943	944	dev_err(dev, "Invalid lane configuration\n");
944	945	goto put_child2;
945	946	}

drivers/phy/mediatek/phy-mtk-tphy.c+2−0

@@ -949,6 +949,8 @@ static int mtk_phy_init(struct phy *phy) · 948 unmodified lines
949	949	break;
950	950	default:
951	951	dev_err(tphy->dev, "incompatible PHY type\n");
	952	+ clk_disable_unprepare(instance->ref_clk);
	953	+ clk_disable_unprepare(instance->da_ref_clk);
952	954	return -EINVAL;
953	955	}
954	956

drivers/phy/microchip/sparx5_serdes.c+4−0

@@ -2470,6 +2470,10 @@ static int sparx5_serdes_probe(struct platform_device *pdev) · 2469 unmodified lines
2470	2470	priv->coreclock = clock;
2471	2471
2472	2472	iores = platform_get_resource(pdev, IORESOURCE_MEM, 0);
	2473	+ if (!iores) {
	2474	+ dev_err(priv->dev, "Invalid resource\n");
	2475	+ return -EINVAL;
	2476	+ }
2473	2477	iomem = devm_ioremap(priv->dev, iores->start, resource_size(iores));
2474	2478	if (IS_ERR(iomem)) {
2475	2479	dev_err(priv->dev, "Unable to get serdes registers: %s\n",

drivers/phy/ralink/phy-mt7621-pci.c+1−1

@@ -341,7 +341,7 @@ static struct platform_driver mt7621_pci_phy_driver = { · 340 unmodified lines
341	341	.probe = mt7621_pci_phy_probe,
342	342	.driver = {
343	343	.name = "mt7621-pci-phy",
344		− .of_match_table = of_match_ptr(mt7621_pci_phy_ids),
	344	+ .of_match_table = mt7621_pci_phy_ids,
345	345	},
346	346	};
347	347

drivers/phy/ti/phy-j721e-wiz.c+1−0

@@ -1212,6 +1212,7 @@ static int wiz_probe(struct platform_device *pdev) · 1211 unmodified lines
1212	1212
1213	1213	if (wiz->typec_dir_delay < WIZ_TYPEC_DIR_DEBOUNCE_MIN \|\|
1214	1214	wiz->typec_dir_delay > WIZ_TYPEC_DIR_DEBOUNCE_MAX) {
	1215	+ ret = -EINVAL;
1215	1216	dev_err(dev, "Invalid typec-dir-debounce property\n");
1216	1217	goto err_addr_to_resource;
1217	1218	}

drivers/pinctrl/aspeed/pinctrl-aspeed-g5.c+2−2

@@ -2702,8 +2702,8 @@ static int aspeed_g5_sig_expr_eval(struct aspeed_pinmux_data *ctx, · 2701 unmodified lines
2702	2702	}
2703	2703
2704	2704	/**
2705		− * Configure a pin's signal by applying an expression's descriptor state for
2706		− * all descriptors in the expression.
	2705	+ * aspeed_g5_sig_expr_set() - Configure a pin's signal by applying an
	2706	+ * expression's descriptor state for all descriptors in the expression.
2707	2707	*
2708	2708	* @ctx: The pinmux context
2709	2709	* @expr: The expression associated with the function whose signal is to be

drivers/pinctrl/aspeed/pinctrl-aspeed-g6.c+2−2

@@ -2611,8 +2611,8 @@ static struct aspeed_pin_config aspeed_g6_configs[] = { · 2610 unmodified lines
2611	2611	};
2612	2612
2613	2613	/**
2614		− * Configure a pin's signal by applying an expression's descriptor state for
2615		− * all descriptors in the expression.
	2614	+ * aspeed_g6_sig_expr_set() - Configure a pin's signal by applying an
	2615	+ * expression's descriptor state for all descriptors in the expression.
2616	2616	*
2617	2617	* @ctx: The pinmux context
2618	2618	* @expr: The expression associated with the function whose signal is to be

drivers/pinctrl/aspeed/pinctrl-aspeed.c+2−1

@@ -108,7 +108,8 @@ static int aspeed_sig_expr_disable(struct aspeed_pinmux_data *ctx, · 107 unmodified lines
108	108	}
109	109
110	110	/**
111		− * Disable a signal on a pin by disabling all provided signal expressions.
	111	+ * aspeed_disable_sig() - Disable a signal on a pin by disabling all provided
	112	+ * signal expressions.
112	113	*
113	114	* @ctx: The pinmux context
114	115	* @exprs: The list of signal expressions (from a priority level on a pin)

drivers/pinctrl/aspeed/pinmux-aspeed.c+2−1

@@ -21,7 +21,8 @@ static inline void aspeed_sig_desc_print_val( · 20 unmodified lines
21	21	}
22	22
23	23	/**
24		− * Query the enabled or disabled state of a signal descriptor
	24	+ * aspeed_sig_desc_eval() - Query the enabled or disabled state of a signal
	25	+ * descriptor.
25	26	*
26	27	* @desc: The signal descriptor of interest
27	28	* @enabled: True to query the enabled state, false to query disabled state

drivers/pinctrl/qcom/Kconfig+1−1

@@ -223,7 +223,7 @@ config PINCTRL_SC7280 · 222 unmodified lines
223	223	config PINCTRL_SC8180X
224	224	tristate "Qualcomm Technologies Inc SC8180x pin controller driver"
225	225	depends on GPIOLIB && (OF \|\| ACPI)
226		− select PINCTRL_MSM
	226	+ depends on PINCTRL_MSM
227	227	help
228	228	This is the pinctrl, pinmux, pinconf and gpiolib driver for the
229	229	Qualcomm Technologies Inc TLMM block found on the Qualcomm

drivers/pinctrl/qcom/pinctrl-sdx55.c+9−9

@@ -410,15 +410,15 @@ static const char * const gpio_groups[] = { · 409 unmodified lines
410	410	"gpio29", "gpio30", "gpio31", "gpio32", "gpio33", "gpio34", "gpio35",
411	411	"gpio36", "gpio37", "gpio38", "gpio39", "gpio40", "gpio41", "gpio42",
412	412	"gpio43", "gpio44", "gpio45", "gpio46", "gpio47", "gpio48", "gpio49",
413		− "gpio50", "gpio51", "gpio52", "gpio52", "gpio53", "gpio53", "gpio54",
414		− "gpio55", "gpio56", "gpio57", "gpio58", "gpio59", "gpio60", "gpio61",
415		− "gpio62", "gpio63", "gpio64", "gpio65", "gpio66", "gpio67", "gpio68",
416		− "gpio69", "gpio70", "gpio71", "gpio72", "gpio73", "gpio74", "gpio75",
417		− "gpio76", "gpio77", "gpio78", "gpio79", "gpio80", "gpio81", "gpio82",
418		− "gpio83", "gpio84", "gpio85", "gpio86", "gpio87", "gpio88", "gpio89",
419		− "gpio90", "gpio91", "gpio92", "gpio93", "gpio94", "gpio95", "gpio96",
420		− "gpio97", "gpio98", "gpio99", "gpio100", "gpio101", "gpio102",
421		− "gpio103", "gpio104", "gpio105", "gpio106", "gpio107",
	413	+ "gpio50", "gpio51", "gpio52", "gpio53", "gpio54", "gpio55", "gpio56",
	414	+ "gpio57", "gpio58", "gpio59", "gpio60", "gpio61", "gpio62", "gpio63",
	415	+ "gpio64", "gpio65", "gpio66", "gpio67", "gpio68", "gpio69", "gpio70",
	416	+ "gpio71", "gpio72", "gpio73", "gpio74", "gpio75", "gpio76", "gpio77",
	417	+ "gpio78", "gpio79", "gpio80", "gpio81", "gpio82", "gpio83", "gpio84",
	418	+ "gpio85", "gpio86", "gpio87", "gpio88", "gpio89", "gpio90", "gpio91",
	419	+ "gpio92", "gpio93", "gpio94", "gpio95", "gpio96", "gpio97", "gpio98",
	420	+ "gpio99", "gpio100", "gpio101", "gpio102", "gpio103", "gpio104",
	421	+ "gpio105", "gpio106", "gpio107",
422	422	};
423	423
424	424	static const char * const qdss_stm_groups[] = {

drivers/pinctrl/ralink/pinctrl-rt2880.c+1−1

@@ -127,7 +127,7 @@ static int rt2880_pmx_group_enable(struct pinctrl_dev *pctrldev, · 126 unmodified lines
127	127	if (p->groups[group].enabled) {
128	128	dev_err(p->dev, "%s is already enabled\n",
129	129	p->groups[group].name);
130		− return -EBUSY;
	130	+ return 0;
131	131	}
132	132
133	133	p->groups[group].enabled = 1;

drivers/platform/mellanox/mlxreg-hotplug.c+2−2

@@ -683,13 +683,13 @@ static int mlxreg_hotplug_probe(struct platform_device *pdev) · 682 unmodified lines
683	683
684	684	err = devm_request_irq(&pdev->dev, priv->irq,
685	685	mlxreg_hotplug_irq_handler, IRQF_TRIGGER_FALLING
686		− \| IRQF_SHARED \| IRQF_NO_AUTOEN,
687		− "mlxreg-hotplug", priv);
	686	+ \| IRQF_SHARED, "mlxreg-hotplug", priv);
688	687	if (err) {
689	688	dev_err(&pdev->dev, "Failed to request irq: %d\n", err);
690	689	return err;
691	690	}
692	691
	692	+ disable_irq(priv->irq);
693	693	spin_lock_init(&priv->lock);
694	694	INIT_DELAYED_WORK(&priv->dwork_irq, mlxreg_hotplug_work_handler);
695	695	dev_set_drvdata(&pdev->dev, priv);

drivers/platform/surface/aggregator/controller.c+1−1

@@ -1907,7 +1907,7 @@ static int ssam_ssh_event_disable(struct ssam_controller *ctrl, · 1906 unmodified lines
1907	1907	{
1908	1908	int status;
1909	1909
1910		− status = __ssam_ssh_event_request(ctrl, reg, reg.cid_enable, id, flags);
	1910	+ status = __ssam_ssh_event_request(ctrl, reg, reg.cid_disable, id, flags);
1911	1911
1912	1912	if (status < 0 && status != -EINVAL) {
1913	1913	ssam_err(ctrl,

drivers/platform/surface/surface_aggregator_registry.c+5−2

@@ -156,7 +156,7 @@ static const struct software_node *ssam_node_group_sl2[] = { · 155 unmodified lines
156	156	NULL,
157	157	};
158	158
159		−/* Devices for Surface Laptop 3. */
	159	+/* Devices for Surface Laptop 3 and 4. */
160	160	static const struct software_node *ssam_node_group_sl3[] = {
161	161	&ssam_node_root,
162	162	&ssam_node_bat_ac,
@@ -521,9 +521,12 @@ static const struct acpi_device_id ssam_platform_hub_match[] = { · 358 unmodified lines
521	521	/* Surface Laptop 3 (13", Intel) */
522	522	{ "MSHW0114", (unsigned long)ssam_node_group_sl3 },
523	523
524		− /* Surface Laptop 3 (15", AMD) */
	524	+ /* Surface Laptop 3 (15", AMD) and 4 (15", AMD) */
525	525	{ "MSHW0110", (unsigned long)ssam_node_group_sl3 },
526	526
	527	+ /* Surface Laptop 4 (13", Intel) */
	528	+ { "MSHW0250", (unsigned long)ssam_node_group_sl3 },
	529	+
527	530	/* Surface Laptop Go 1 */
528	531	{ "MSHW0118", (unsigned long)ssam_node_group_slg1 },
529	532

drivers/platform/surface/surface_dtx.c+1−0

@@ -427,6 +427,7 @@ static int surface_dtx_open(struct inode inode, struct file file) · 426 unmodified lines
427	427	*/
428	428	if (test_bit(SDTX_DEVICE_SHUTDOWN_BIT, &ddev->flags)) {
429	429	up_write(&ddev->client_lock);
	430	+ mutex_destroy(&client->read_lock);
430	431	sdtx_device_put(client->ddev);
431	432	kfree(client);
432	433	return -ENODEV;

drivers/platform/x86/thinkpad_acpi.c+1−0

@@ -8853,6 +8853,7 @@ static const struct tpacpi_quirk fan_quirk_table[] __initconst = { · 8852 unmodified lines
8853	8853	TPACPI_Q_LNV3('N', '2', 'O', TPACPI_FAN_2CTL), /* P1 / X1 Extreme (2nd gen) */
8854	8854	TPACPI_Q_LNV3('N', '2', 'V', TPACPI_FAN_2CTL), /* P1 / X1 Extreme (3nd gen) */
8855	8855	TPACPI_Q_LNV3('N', '3', '0', TPACPI_FAN_2CTL), /* P15 (1st gen) / P15v (1st gen) */
	8856	+ TPACPI_Q_LNV3('N', '3', '2', TPACPI_FAN_2CTL), /* X1 Carbon (9th gen) */
8856	8857	};
8857	8858
8858	8859	static int __init fan_init(struct ibm_init_struct *iibm)

drivers/ptp/ptp_clock.c+3−3

@@ -63,7 +63,7 @@ static void enqueue_external_timestamp(struct timestamp_event_queue *queue, · 62 unmodified lines
63	63	spin_unlock_irqrestore(&queue->lock, flags);
64	64	}
65	65
66		−s32 scaled_ppm_to_ppb(long ppm)
	66	+long scaled_ppm_to_ppb(long ppm)
67	67	{
68	68	/*
69	69	* The 'freq' field in the 'struct timex' is in parts per
@@ -80,7 +80,7 @@ s32 scaled_ppm_to_ppb(long ppm) · 10 unmodified lines
80	80	s64 ppb = 1 + ppm;
81	81	ppb *= 125;
82	82	ppb >>= 13;
83		− return (s32) ppb;
	83	+ return (long) ppb;
84	84	}
85	85	EXPORT_SYMBOL(scaled_ppm_to_ppb);
86	86
@@ -138,7 +138,7 @@ static int ptp_clock_adjtime(struct posix_clock pc, struct __kernel_timex tx) · 51 unmodified lines
138	138	delta = ktime_to_ns(kt);
139	139	err = ops->adjtime(ops, delta);
140	140	} else if (tx->modes & ADJ_FREQUENCY) {
141		− s32 ppb = scaled_ppm_to_ppb(tx->freq);
	141	+ long ppb = scaled_ppm_to_ppb(tx->freq);
142	142	if (ppb > ops->max_adj \|\| ppb < -ops->max_adj)
143	143	return -ERANGE;
144	144	if (ops->adjfine)

drivers/ptp/ptp_ocp.c+3−1

@@ -324,7 +324,7 @@ ptp_ocp_probe(struct pci_dev pdev, const struct pci_device_id id) · 323 unmodified lines
324	324	if (!bp->base) {
325	325	dev_err(&pdev->dev, "io_remap bar0\n");
326	326	err = -ENOMEM;
327		− goto out;
	327	+ goto out_release_regions;
328	328	}
329	329	bp->reg = bp->base + OCP_REGISTER_OFFSET;
330	330	bp->tod = bp->base + TOD_REGISTER_OFFSET;
@@ -347,6 +347,8 @@ ptp_ocp_probe(struct pci_dev pdev, const struct pci_device_id id) · 16 unmodified lines
347	347	return 0;
348	348
349	349	out:
	350	+ pci_iounmap(pdev, bp->base);
	351	+out_release_regions:
350	352	pci_release_regions(pdev);
351	353	out_disable:
352	354	pci_disable_device(pdev);

drivers/regulator/Kconfig+1−1

@@ -1031,7 +1031,7 @@ config REGULATOR_RT5033 · 1030 unmodified lines
1031	1031	current source, LDO and Buck.
1032	1032
1033	1033	config REGULATOR_RTMV20
1034		− tristate "RTMV20 Laser Diode Regulator"
	1034	+ tristate "Richtek RTMV20 Laser Diode Regulator"
1035	1035	depends on I2C
1036	1036	select REGMAP_I2C
1037	1037	help

drivers/regulator/atc260x-regulator.c+10−9

@@ -28,16 +28,16 @@ static const struct linear_range atc2609a_dcdc_voltage_ranges[] = { · 27 unmodified lines
28	28
29	29	static const struct linear_range atc2609a_ldo_voltage_ranges0[] = {
30	30	REGULATOR_LINEAR_RANGE(700000, 0, 15, 100000),
31		− REGULATOR_LINEAR_RANGE(2100000, 16, 28, 100000),
	31	+ REGULATOR_LINEAR_RANGE(2100000, 0, 12, 100000),
32	32	};
33	33
34	34	static const struct linear_range atc2609a_ldo_voltage_ranges1[] = {
35	35	REGULATOR_LINEAR_RANGE(850000, 0, 15, 100000),
36		− REGULATOR_LINEAR_RANGE(2100000, 16, 27, 100000),
	36	+ REGULATOR_LINEAR_RANGE(2100000, 0, 11, 100000),
37	37	};
38	38
39	39	static const unsigned int atc260x_ldo_voltage_range_sel[] = {
40		− 0x0, 0x1,
	40	+ 0x0, 0x20,
41	41	};
42	42
43	43	static int atc260x_dcdc_set_voltage_time_sel(struct regulator_dev *rdev,
@@ -411,7 +411,7 @@ enum atc2609a_reg_ids { · 367 unmodified lines
411	411	.owner = THIS_MODULE, \
412	412	}
413	413
414		−#define atc2609a_reg_desc_ldo_range_pick(num, n_range) { \
	414	+#define atc2609a_reg_desc_ldo_range_pick(num, n_range, n_volt) { \
415	415	.name = "LDO"#num, \
416	416	.supply_name = "ldo"#num, \
417	417	.of_match = of_match_ptr("ldo"#num), \
@@ -421,6 +421,7 @@ enum atc2609a_reg_ids { · 3 unmodified lines
421	421	.type = REGULATOR_VOLTAGE, \
422	422	.linear_ranges = atc2609a_ldo_voltage_ranges##n_range, \
423	423	.n_linear_ranges = ARRAY_SIZE(atc2609a_ldo_voltage_ranges##n_range), \
	424	+ .n_voltages = n_volt, \
424	425	.vsel_reg = ATC2609A_PMU_LDO##num##_CTL0, \
425	426	.vsel_mask = GENMASK(4, 1), \
426	427	.vsel_range_reg = ATC2609A_PMU_LDO##num##_CTL0, \
@@ -458,12 +459,12 @@ static const struct regulator_desc atc2609a_reg[] = { · 31 unmodified lines
458	459	atc2609a_reg_desc_ldo_bypass(0),
459	460	atc2609a_reg_desc_ldo_bypass(1),
460	461	atc2609a_reg_desc_ldo_bypass(2),
461		− atc2609a_reg_desc_ldo_range_pick(3, 0),
462		− atc2609a_reg_desc_ldo_range_pick(4, 0),
	462	+ atc2609a_reg_desc_ldo_range_pick(3, 0, 29),
	463	+ atc2609a_reg_desc_ldo_range_pick(4, 0, 29),
463	464	atc2609a_reg_desc_ldo(5),
464		− atc2609a_reg_desc_ldo_range_pick(6, 1),
465		− atc2609a_reg_desc_ldo_range_pick(7, 0),
466		− atc2609a_reg_desc_ldo_range_pick(8, 0),
	465	+ atc2609a_reg_desc_ldo_range_pick(6, 1, 28),
	466	+ atc2609a_reg_desc_ldo_range_pick(7, 0, 29),
	467	+ atc2609a_reg_desc_ldo_range_pick(8, 0, 29),
467	468	atc2609a_reg_desc_ldo_fixed(9),
468	469	};
469	470

drivers/regulator/bd718x7-regulator.c+1−1

@@ -334,7 +334,7 @@ BD718XX_OPS(bd71837_buck_regulator_ops, regulator_list_voltage_linear_range, · 333 unmodified lines
334	334	NULL);
335	335
336	336	BD718XX_OPS(bd71837_buck_regulator_nolinear_ops, regulator_list_voltage_table,
337		− regulator_map_voltage_ascend, bd718xx_set_voltage_sel_restricted,
	337	+ regulator_map_voltage_ascend, bd71837_set_voltage_sel_restricted,
338	338	regulator_get_voltage_sel_regmap, regulator_set_voltage_time_sel,
339	339	NULL);
340	340	/*

drivers/regulator/core.c+6−0

@@ -1425,6 +1425,12 @@ static int set_machine_constraints(struct regulator_dev *rdev) · 1424 unmodified lines
1425	1425	* and we have control then make sure it is enabled.
1426	1426	*/
1427	1427	if (rdev->constraints->always_on \|\| rdev->constraints->boot_on) {
	1428	+ /* If we want to enable this regulator, make sure that we know
	1429	+ * the supplying regulator.
	1430	+ */
	1431	+ if (rdev->supply_name && !rdev->supply)
	1432	+ return -EPROBE_DEFER;
	1433	+
1428	1434	if (rdev->supply) {
1429	1435	ret = regulator_enable(rdev->supply);
1430	1436	if (ret < 0) {

drivers/regulator/cros-ec-regulator.c+2−1

@@ -225,8 +225,9 @@ static int cros_ec_regulator_probe(struct platform_device *pdev) · 224 unmodified lines
225	225
226	226	drvdata->dev = devm_regulator_register(dev, &drvdata->desc, &cfg);
227	227	if (IS_ERR(drvdata->dev)) {
	228	+ ret = PTR_ERR(drvdata->dev);
228	229	dev_err(&pdev->dev, "Failed to register regulator: %d\n", ret);
229		− return PTR_ERR(drvdata->dev);
	230	+ return ret;
230	231	}
231	232
232	233	platform_set_drvdata(pdev, drvdata);

drivers/regulator/da9121-regulator.c+7−3

@@ -280,7 +280,7 @@ static unsigned int da9121_map_mode(unsigned int mode) · 279 unmodified lines
280	280	case DA9121_BUCK_MODE_FORCE_PFM:
281	281	return REGULATOR_MODE_STANDBY;
282	282	default:
283		− return -EINVAL;
	283	+ return REGULATOR_MODE_INVALID;
284	284	}
285	285	}
286	286
@@ -317,7 +317,7 @@ static unsigned int da9121_buck_get_mode(struct regulator_dev *rdev) · 30 unmodified lines
317	317	{
318	318	struct da9121 *chip = rdev_get_drvdata(rdev);
319	319	int id = rdev_get_id(rdev);
320		− unsigned int val;
	320	+ unsigned int val, mode;
321	321	int ret = 0;
322	322
323	323	ret = regmap_read(chip->regmap, da9121_mode_field[id].reg, &val);
@@ -326,7 +326,11 @@ static unsigned int da9121_buck_get_mode(struct regulator_dev *rdev) · 2 unmodified lines
326	326	return -EINVAL;
327	327	}
328	328
329		− return da9121_map_mode(val & da9121_mode_field[id].msk);
	329	+ mode = da9121_map_mode(val & da9121_mode_field[id].msk);
	330	+ if (mode == REGULATOR_MODE_INVALID)
	331	+ return -EINVAL;
	332	+
	333	+ return mode;
330	334	}
331	335
332	336	static const struct regulator_ops da9121_buck_ops = {

drivers/regulator/fan53555.c+1−2

@@ -55,7 +55,6 @@ · 54 unmodified lines
55	55
56	56	#define FAN53555_NVOLTAGES 64 /* Numbers of voltages */
57	57	#define FAN53526_NVOLTAGES 128
58		−#define TCS4525_NVOLTAGES 127 /* Numbers of voltages */
59	58
60	59	#define TCS_VSEL_NSEL_MASK 0x7f
61	60	#define TCS_VSEL0_MODE (1 << 7)
@@ -376,7 +375,7 @@ static int fan53555_voltages_setup_tcs(struct fan53555_device_info *di) · 314 unmodified lines
376	375	/* Init voltage range and step */
377	376	di->vsel_min = 600000;
378	377	di->vsel_step = 6250;
379		− di->vsel_count = TCS4525_NVOLTAGES;
	378	+ di->vsel_count = FAN53526_NVOLTAGES;
380	379
381	380	return 0;
382	381	}

drivers/regulator/fan53880.c+3−0

@@ -51,6 +51,7 @@ static const struct regulator_ops fan53880_ops = { · 50 unmodified lines
51	51	REGULATOR_LINEAR_RANGE(800000, 0xf, 0x73, 25000), \
52	52	}, \
53	53	.n_linear_ranges = 2, \
	54	+ .n_voltages = 0x74, \
54	55	.vsel_reg = FAN53880_LDO ## _num ## VOUT, \
55	56	.vsel_mask = 0x7f, \
56	57	.enable_reg = FAN53880_ENABLE, \
@@ -76,6 +77,7 @@ static const struct regulator_desc fan53880_regulators[] = { · 19 unmodified lines
76	77	REGULATOR_LINEAR_RANGE(600000, 0x1f, 0xf7, 12500),
77	78	},
78	79	.n_linear_ranges = 2,
	80	+ .n_voltages = 0xf8,
79	81	.vsel_reg = FAN53880_BUCKVOUT,
80	82	.vsel_mask = 0x7f,
81	83	.enable_reg = FAN53880_ENABLE,
@@ -95,6 +97,7 @@ static const struct regulator_desc fan53880_regulators[] = { · 13 unmodified lines
95	97	REGULATOR_LINEAR_RANGE(3000000, 0x4, 0x70, 25000),
96	98	},
97	99	.n_linear_ranges = 2,
	100	+ .n_voltages = 0x71,
98	101	.vsel_reg = FAN53880_BOOSTVOUT,
99	102	.vsel_mask = 0x7f,
100	103	.enable_reg = FAN53880_ENABLE_BOOST,

drivers/regulator/fixed.c+6−1

@@ -88,10 +88,15 @@ static int reg_domain_disable(struct regulator_dev *rdev) · 87 unmodified lines
88	88	{
89	89	struct fixed_voltage_data *priv = rdev_get_drvdata(rdev);
90	90	struct device *dev = rdev->dev.parent;
	91	+ int ret;
	92	+
	93	+ ret = dev_pm_genpd_set_performance_state(dev, 0);
	94	+ if (ret)
	95	+ return ret;
91	96
92	97	priv->enable_counter--;
93	98
94		− return dev_pm_genpd_set_performance_state(dev, 0);
	99	+ return 0;
95	100	}
96	101
97	102	static int reg_is_enabled(struct regulator_dev *rdev)

drivers/regulator/helpers.c+1−1

@@ -948,7 +948,7 @@ int regulator_set_ramp_delay_regmap(struct regulator_dev *rdev, int ramp_delay) · 947 unmodified lines
948	948	int ret;
949	949	unsigned int sel;
950	950
951		− if (!rdev->desc->n_ramp_values)
	951	+ if (WARN_ON(!rdev->desc->n_ramp_values \|\| !rdev->desc->ramp_delay_table))
952	952	return -EINVAL;
953	953
954	954	ret = find_closest_bigger(ramp_delay, rdev->desc->ramp_delay_table,

drivers/regulator/hi6421v600-regulator.c+2−2

@@ -3,7 +3,7 @@ · 2 unmodified lines
3	3	// Device driver for regulators in Hisi IC
4	4	//
5	5	// Copyright (c) 2013 Linaro Ltd.
6		−// Copyright (c) 2011 Hisilicon.
	6	+// Copyright (c) 2011 HiSilicon Ltd.
7	7	// Copyright (c) 2020-2021 Huawei Technologies Co., Ltd
8	8	//
9	9	// Guodong Xu <guodong.xu@linaro.org>
@@ -83,7 +83,7 @@ static const unsigned int ldo34_voltages[] = { · 73 unmodified lines
83	83	.owner = THIS_MODULE, \
84	84	.volt_table = vtable, \
85	85	.n_voltages = ARRAY_SIZE(vtable), \
86		− .vsel_mask = (1 << (ARRAY_SIZE(vtable) - 1)) - 1, \
	86	+ .vsel_mask = ARRAY_SIZE(vtable) - 1, \
87	87	.vsel_reg = vreg, \
88	88	.enable_reg = ereg, \
89	89	.enable_mask = emask, \

drivers/regulator/hi655x-regulator.c+1−1

@@ -2,7 +2,7 @@ · 1 unmodified line
2	2	//
3	3	// Device driver for regulators in Hi655x IC
4	4	//
5		−// Copyright (c) 2016 Hisilicon.
	5	+// Copyright (c) 2016 HiSilicon Ltd.
6	6	//
7	7	// Authors:
8	8	// Chen Feng <puck.chen@hisilicon.com>

drivers/regulator/max77620-regulator.c+11−6

@@ -814,6 +814,13 @@ static int max77620_regulator_probe(struct platform_device *pdev) · 813 unmodified lines
814	814	config.dev = dev;
815	815	config.driver_data = pmic;
816	816
	817	+ /*
	818	+ * Set of_node_reuse flag to prevent driver core from attempting to
	819	+ * claim any pinmux resources already claimed by the parent device.
	820	+ * Otherwise PMIC driver will fail to re-probe.
	821	+ */
	822	+ device_set_of_node_from_dev(&pdev->dev, pdev->dev.parent);
	823	+
817	824	for (id = 0; id < MAX77620_NUM_REGS; id++) {
818	825	struct regulator_dev *rdev;
819	826	struct regulator_desc *rdesc;
@@ -839,12 +846,10 @@ static int max77620_regulator_probe(struct platform_device *pdev) · 19 unmodified lines
839	846	return ret;
840	847
841	848	rdev = devm_regulator_register(dev, rdesc, &config);
842		− if (IS_ERR(rdev)) {
843		− ret = PTR_ERR(rdev);
844		− dev_err(dev, "Regulator registration %s failed: %d\n",
845		− rdesc->name, ret);
846		− return ret;
847		− }
	849	+ if (IS_ERR(rdev))
	850	+ return dev_err_probe(dev, PTR_ERR(rdev),
	851	+ "Regulator registration %s failed\n",
	852	+ rdesc->name);
848	853	}
849	854
850	855	return 0;

drivers/regulator/mt6315-regulator.c+1−1

@@ -59,7 +59,7 @@ static const struct linear_range mt_volt_range1[] = { · 58 unmodified lines
59	59	REGULATOR_LINEAR_RANGE(0, 0, 0xbf, 6250),
60	60	};
61	61
62		−static unsigned int mt6315_map_mode(u32 mode)
	62	+static unsigned int mt6315_map_mode(unsigned int mode)
63	63	{
64	64	switch (mode) {
65	65	case MT6315_BUCK_MODE_AUTO:

drivers/regulator/rt4801-regulator.c+2−2

@@ -66,7 +66,7 @@ static int rt4801_enable(struct regulator_dev *rdev) · 65 unmodified lines
66	66	struct gpio_descs *gpios = priv->enable_gpios;
67	67	int id = rdev_get_id(rdev), ret;
68	68
69		− if (gpios->ndescs <= id) {
	69	+ if (!gpios \|\| gpios->ndescs <= id) {
70	70	dev_warn(&rdev->dev, "no dedicated gpio can control\n");
71	71	goto bypass_gpio;
72	72	}
@@ -88,7 +88,7 @@ static int rt4801_disable(struct regulator_dev *rdev) · 15 unmodified lines
88	88	struct gpio_descs *gpios = priv->enable_gpios;
89	89	int id = rdev_get_id(rdev);
90	90
91		− if (gpios->ndescs <= id) {
	91	+ if (!gpios \|\| gpios->ndescs <= id) {
92	92	dev_warn(&rdev->dev, "no dedicated gpio can control\n");
93	93	goto bypass_gpio;
94	94	}

drivers/regulator/rtmv20-regulator.c+42−2

@@ -27,6 +27,7 @@ · 26 unmodified lines
27	27	#define RTMV20_REG_LDIRQ 0x30
28	28	#define RTMV20_REG_LDSTAT 0x40
29	29	#define RTMV20_REG_LDMASK 0x50
	30	+#define RTMV20_MAX_REGS (RTMV20_REG_LDMASK + 1)
30	31
31	32	#define RTMV20_VID_MASK GENMASK(7, 4)
32	33	#define RICHTEK_VID 0x80
@@ -103,9 +104,47 @@ static int rtmv20_lsw_disable(struct regulator_dev *rdev) · 70 unmodified lines
103	104	return 0;
104	105	}
105	106
	107	+static int rtmv20_lsw_set_current_limit(struct regulator_dev *rdev, int min_uA,
	108	+ int max_uA)
	109	+{
	110	+ int sel;
	111	+
	112	+ if (min_uA > RTMV20_LSW_MAXUA \|\| max_uA < RTMV20_LSW_MINUA)
	113	+ return -EINVAL;
	114	+
	115	+ if (max_uA > RTMV20_LSW_MAXUA)
	116	+ max_uA = RTMV20_LSW_MAXUA;
	117	+
	118	+ sel = (max_uA - RTMV20_LSW_MINUA) / RTMV20_LSW_STEPUA;
	119	+
	120	+ /* Ensure the selected setting is still in range */
	121	+ if ((sel * RTMV20_LSW_STEPUA + RTMV20_LSW_MINUA) < min_uA)
	122	+ return -EINVAL;
	123	+
	124	+ sel <<= ffs(rdev->desc->csel_mask) - 1;
	125	+
	126	+ return regmap_update_bits(rdev->regmap, rdev->desc->csel_reg,
	127	+ rdev->desc->csel_mask, sel);
	128	+}
	129	+
	130	+static int rtmv20_lsw_get_current_limit(struct regulator_dev *rdev)
	131	+{
	132	+ unsigned int val;
	133	+ int ret;
	134	+
	135	+ ret = regmap_read(rdev->regmap, rdev->desc->csel_reg, &val);
	136	+ if (ret)
	137	+ return ret;
	138	+
	139	+ val &= rdev->desc->csel_mask;
	140	+ val >>= ffs(rdev->desc->csel_mask) - 1;
	141	+
	142	+ return val * RTMV20_LSW_STEPUA + RTMV20_LSW_MINUA;
	143	+}
	144	+
106	145	static const struct regulator_ops rtmv20_regulator_ops = {
107		− .set_current_limit = regulator_set_current_limit_regmap,
108		− .get_current_limit = regulator_get_current_limit_regmap,
	146	+ .set_current_limit = rtmv20_lsw_set_current_limit,
	147	+ .get_current_limit = rtmv20_lsw_get_current_limit,
109	148	.enable = rtmv20_lsw_enable,
110	149	.disable = rtmv20_lsw_disable,
111	150	.is_enabled = regulator_is_enabled_regmap,
@@ -275,6 +314,7 @@ static const struct regmap_config rtmv20_regmap_config = { · 163 unmodified lines
275	314	.val_bits = 8,
276	315	.cache_type = REGCACHE_RBTREE,
277	316	.max_register = RTMV20_REG_LDMASK,
	317	+ .num_reg_defaults_raw = RTMV20_MAX_REGS,
278	318
279	319	.writeable_reg = rtmv20_is_accessible_reg,
280	320	.readable_reg = rtmv20_is_accessible_reg,

drivers/regulator/scmi-regulator.c+1−1

@@ -173,7 +173,7 @@ scmi_config_linear_regulator_mappings(struct scmi_regulator *sreg, · 172 unmodified lines
173	173	sreg->desc.uV_step =
174	174	vinfo->levels_uv[SCMI_VOLTAGE_SEGMENT_STEP];
175	175	sreg->desc.linear_min_sel = 0;
176		− sreg->desc.n_voltages = delta_uV / sreg->desc.uV_step;
	176	+ sreg->desc.n_voltages = (delta_uV / sreg->desc.uV_step) + 1;
177	177	sreg->desc.ops = &scmi_reg_linear_ops;
178	178	}
179	179

drivers/s390/block/dasd_diag.c+7−1

@@ -642,12 +642,18 @@ static void dasd_diag_setup_blk_queue(struct dasd_block *block) · 641 unmodified lines
642	642	blk_queue_segment_boundary(q, PAGE_SIZE - 1);
643	643	}
644	644
	645	+static int dasd_diag_pe_handler(struct dasd_device *device,
	646	+ __u8 tbvpm, __u8 fcsecpm)
	647	+{
	648	+ return dasd_generic_verify_path(device, tbvpm);
	649	+}
	650	+
645	651	static struct dasd_discipline dasd_diag_discipline = {
646	652	.owner = THIS_MODULE,
647	653	.name = "DIAG",
648	654	.ebcname = "DIAG",
649	655	.check_device = dasd_diag_check_device,
650		− .verify_path = dasd_generic_verify_path,
	656	+ .pe_handler = dasd_diag_pe_handler,
651	657	.fill_geometry = dasd_diag_fill_geometry,
652	658	.setup_blk_queue = dasd_diag_setup_blk_queue,
653	659	.start_IO = dasd_start_diag,

drivers/s390/block/dasd_fba.c+7−1

@@ -794,13 +794,19 @@ static void dasd_fba_setup_blk_queue(struct dasd_block *block) · 793 unmodified lines
794	794	blk_queue_flag_set(QUEUE_FLAG_DISCARD, q);
795	795	}
796	796
	797	+static int dasd_fba_pe_handler(struct dasd_device *device,
	798	+ __u8 tbvpm, __u8 fcsecpm)
	799	+{
	800	+ return dasd_generic_verify_path(device, tbvpm);
	801	+}
	802	+
797	803	static struct dasd_discipline dasd_fba_discipline = {
798	804	.owner = THIS_MODULE,
799	805	.name = "FBA ",
800	806	.ebcname = "FBA ",
801	807	.check_device = dasd_fba_check_characteristics,
802	808	.do_analysis = dasd_fba_do_analysis,
803		− .verify_path = dasd_generic_verify_path,
	809	+ .pe_handler = dasd_fba_pe_handler,
804	810	.setup_blk_queue = dasd_fba_setup_blk_queue,
805	811	.fill_geometry = dasd_fba_fill_geometry,
806	812	.start_IO = dasd_start_IO,

drivers/s390/block/dasd_int.h+0−1

@@ -297,7 +297,6 @@ struct dasd_discipline { · 296 unmodified lines
297	297	* e.g. verify that new path is compatible with the current
298	298	* configuration.
299	299	*/
300		− int (verify_path)(struct dasd_device , __u8);
301	300	int (pe_handler)(struct dasd_device , __u8, __u8);
302	301
303	302	/*

drivers/s390/cio/vfio_ccw_cp.c+4−0

@@ -638,6 +638,10 @@ int cp_init(struct channel_program cp, struct device mdev, union orb *orb) · 637 unmodified lines
638	638	static DEFINE_RATELIMIT_STATE(ratelimit_state, 5 * HZ, 1);
639	639	int ret;
640	640
	641	+ /* this is an error in the caller */
	642	+ if (cp->initialized)
	643	+ return -EBUSY;
	644	+
641	645	/*
642	646	* We only support prefetching the channel program. We assume all channel
643	647	* programs executed by supported guests likewise support prefetching.

drivers/s390/cio/vfio_ccw_drv.c+10−2

@@ -86,6 +86,7 @@ static void vfio_ccw_sch_io_todo(struct work_struct *work) · 85 unmodified lines
86	86	struct vfio_ccw_private *private;
87	87	struct irb *irb;
88	88	bool is_final;
	89	+ bool cp_is_finished = false;
89	90
90	91	private = container_of(work, struct vfio_ccw_private, io_work);
91	92	irb = &private->irb;
@@ -94,14 +95,21 @@ static void vfio_ccw_sch_io_todo(struct work_struct *work) · 2 unmodified lines
94	95	(SCSW_ACTL_DEVACT \| SCSW_ACTL_SCHACT));
95	96	if (scsw_is_solicited(&irb->scsw)) {
96	97	cp_update_scsw(&private->cp, &irb->scsw);
97		− if (is_final && private->state == VFIO_CCW_STATE_CP_PENDING)
	98	+ if (is_final && private->state == VFIO_CCW_STATE_CP_PENDING) {
98	99	cp_free(&private->cp);
	100	+ cp_is_finished = true;
	101	+ }
99	102	}
100	103	mutex_lock(&private->io_mutex);
101	104	memcpy(private->io_region->irb_area, irb, sizeof(*irb));
102	105	mutex_unlock(&private->io_mutex);
103	106
104		− if (private->mdev && is_final)
	107	+ /*
	108	+ * Reset to IDLE only if processing of a channel program
	109	+ * has finished. Do not overwrite a possible processing
	110	+ * state if the final interrupt was for HSCH or CSCH.
	111	+ */
	112	+ if (private->mdev && cp_is_finished)
105	113	private->state = VFIO_CCW_STATE_IDLE;
106	114
107	115	if (private->io_trigger)

drivers/s390/cio/vfio_ccw_fsm.c+1−0

@@ -318,6 +318,7 @@ static void fsm_io_request(struct vfio_ccw_private *private, · 317 unmodified lines
318	318	}
319	319
320	320	err_out:
	321	+ private->state = VFIO_CCW_STATE_IDLE;
321	322	trace_vfio_ccw_fsm_io_request(scsw->cmd.fctl, schid,
322	323	io_region->ret_code, errstr);
323	324	}

drivers/s390/cio/vfio_ccw_ops.c+0−2

@@ -279,8 +279,6 @@ static ssize_t vfio_ccw_mdev_write_io_region(struct vfio_ccw_private *private, · 278 unmodified lines
279	279	}
280	280
281	281	vfio_ccw_fsm_event(private, VFIO_CCW_EVENT_IO_REQ);
282		− if (region->ret_code != 0)
283		− private->state = VFIO_CCW_STATE_IDLE;
284	282	ret = (region->ret_code != 0) ? region->ret_code : count;
285	283
286	284	out_unlock:

drivers/s390/crypto/ap_queue.c+9−2

@@ -135,12 +135,13 @@ static struct ap_queue_status ap_sm_recv(struct ap_queue *aq) · 134 unmodified lines
135	135	{
136	136	struct ap_queue_status status;
137	137	struct ap_message *ap_msg;
	138	+ bool found = false;
138	139
139	140	status = ap_dqap(aq->qid, &aq->reply->psmid,
140	141	aq->reply->msg, aq->reply->len);
141	142	switch (status.response_code) {
142	143	case AP_RESPONSE_NORMAL:
143		− aq->queue_count--;
	144	+ aq->queue_count = max_t(int, 0, aq->queue_count - 1);
144	145	if (aq->queue_count > 0)
145	146	mod_timer(&aq->timeout,
146	147	jiffies + aq->request_timeout);
@@ -150,8 +151,14 @@ static struct ap_queue_status ap_sm_recv(struct ap_queue *aq) · 3 unmodified lines
150	151	list_del_init(&ap_msg->list);
151	152	aq->pendingq_count--;
152	153	ap_msg->receive(aq, ap_msg, aq->reply);
	154	+ found = true;
153	155	break;
154	156	}
	157	+ if (!found) {
	158	+ AP_DBF_WARN("%s unassociated reply psmid=0x%016llx on 0x%02x.%04x\n",
	159	+ __func__, aq->reply->psmid,
	160	+ AP_QID_CARD(aq->qid), AP_QID_QUEUE(aq->qid));
	161	+ }
155	162	fallthrough;
156	163	case AP_RESPONSE_NO_PENDING_REPLY:
157	164	if (!status.queue_empty \|\| aq->queue_count <= 0)
@@ -232,7 +239,7 @@ static enum ap_sm_wait ap_sm_write(struct ap_queue *aq) · 74 unmodified lines
232	239	ap_msg->flags & AP_MSG_FLAG_SPECIAL);
233	240	switch (status.response_code) {
234	241	case AP_RESPONSE_NORMAL:
235		− aq->queue_count++;
	242	+ aq->queue_count = max_t(int, 1, aq->queue_count + 1);
236	243	if (aq->queue_count == 1)
237	244	mod_timer(&aq->timeout, jiffies + aq->request_timeout);
238	245	list_move_tail(&ap_msg->list, &aq->pendingq);

drivers/scsi/aic7xxx/aicasm/aicasm_gram.y+0−1

@@ -58,7 +58,6 @@ · 57 unmodified lines
58	58	#include "aicasm_symbol.h"
59	59	#include "aicasm_insformat.h"
60	60
61		−int yylineno;
62	61	char *yyfilename;
63	62	char stock_prefix[] = "aic_";
64	63	char *prefix = stock_prefix;

drivers/scsi/aic7xxx/aicasm/aicasm_symbol.h+1−1

@@ -108,7 +108,7 @@ struct macro_arg { · 107 unmodified lines
108	108	regex_t arg_regex;
109	109	char *replacement_text;
110	110	};
111		−STAILQ_HEAD(macro_arg_list, macro_arg) args;
	111	+STAILQ_HEAD(macro_arg_list, macro_arg);
112	112
113	113	struct macro_info {
114	114	struct macro_arg_list args;

drivers/scsi/aic7xxx/scsi_message.h+11−0

@@ -3,6 +3,17 @@ · 2 unmodified lines
3	3	* $FreeBSD: src/sys/cam/scsi/scsi_message.h,v 1.2 2000/05/01 20:21:29 peter Exp $
4	4	*/
5	5
	6	+/* Messages (1 byte) / / I/T (M)andatory or (O)ptional */
	7	+#define MSG_SAVEDATAPOINTER 0x02 /* O/O */
	8	+#define MSG_RESTOREPOINTERS 0x03 /* O/O */
	9	+#define MSG_DISCONNECT 0x04 /* O/O */
	10	+#define MSG_MESSAGE_REJECT 0x07 /* M/M */
	11	+#define MSG_NOOP 0x08 /* M/M */
	12	+
	13	+/* Messages (2 byte) */
	14	+#define MSG_SIMPLE_Q_TAG 0x20 /* O/O */
	15	+#define MSG_IGN_WIDE_RESIDUE 0x23 /* O/O */
	16	+
6	17	/* Identify message / / M/M */
7	18	#define MSG_IDENTIFYFLAG 0x80
8	19	#define MSG_IDENTIFY_DISCFLAG 0x40

drivers/scsi/bnx2fc/bnx2fc_io.c+1−0

@@ -1220,6 +1220,7 @@ int bnx2fc_eh_abort(struct scsi_cmnd *sc_cmd) · 1219 unmodified lines
1220	1220	was a result from the ABTS request rather than the CLEANUP
1221	1221	request */
1222	1222	set_bit(BNX2FC_FLAG_IO_CLEANUP, &io_req->req_flags);
	1223	+ rc = FAILED;
1223	1224	goto done;
1224	1225	}
1225	1226

drivers/scsi/hisi_sas/hisi_sas_v3_hw.c+4−4

@@ -4811,14 +4811,14 @@ hisi_sas_v3_destroy_irqs(struct pci_dev pdev, struct hisi_hba hisi_hba) · 4810 unmodified lines
4811	4811	{
4812	4812	int i;
4813	4813
4814		− free_irq(pci_irq_vector(pdev, 1), hisi_hba);
4815		− free_irq(pci_irq_vector(pdev, 2), hisi_hba);
4816		− free_irq(pci_irq_vector(pdev, 11), hisi_hba);
	4814	+ devm_free_irq(&pdev->dev, pci_irq_vector(pdev, 1), hisi_hba);
	4815	+ devm_free_irq(&pdev->dev, pci_irq_vector(pdev, 2), hisi_hba);
	4816	+ devm_free_irq(&pdev->dev, pci_irq_vector(pdev, 11), hisi_hba);
4817	4817	for (i = 0; i < hisi_hba->cq_nvecs; i++) {
4818	4818	struct hisi_sas_cq *cq = &hisi_hba->cq[i];
4819	4819	int nr = hisi_sas_intr_conv ? 16 : 16 + i;
4820	4820
4821		− free_irq(pci_irq_vector(pdev, nr), cq);
	4821	+ devm_free_irq(&pdev->dev, pci_irq_vector(pdev, nr), cq);
4822	4822	}
4823	4823	pci_free_irq_vectors(pdev);
4824	4824	}

drivers/scsi/hosts.c+26−21

@@ -254,12 +254,11 @@ int scsi_add_host_with_dma(struct Scsi_Host shost, struct device dev, · 253 unmodified lines
254	254
255	255	device_enable_async_suspend(&shost->shost_dev);
256	256
	257	+ get_device(&shost->shost_gendev);
257	258	error = device_add(&shost->shost_dev);
258	259	if (error)
259	260	goto out_del_gendev;
260	261
261		− get_device(&shost->shost_gendev);
262		−
263	262	if (shost->transportt->host_size) {
264	263	shost->shost_data = kzalloc(shost->transportt->host_size,
265	264	GFP_KERNEL);
@@ -278,33 +277,36 @@ int scsi_add_host_with_dma(struct Scsi_Host shost, struct device dev, · 12 unmodified lines
278	277
279	278	if (!shost->work_q) {
280	279	error = -EINVAL;
281		− goto out_free_shost_data;
	280	+ goto out_del_dev;
282	281	}
283	282	}
284	283
285	284	error = scsi_sysfs_add_host(shost);
286	285	if (error)
287		− goto out_destroy_host;
	286	+ goto out_del_dev;
288	287
289	288	scsi_proc_host_add(shost);
290	289	scsi_autopm_put_host(shost);
291	290	return error;
292	291
293		− out_destroy_host:
294		− if (shost->work_q)
295		− destroy_workqueue(shost->work_q);
296		− out_free_shost_data:
297		− kfree(shost->shost_data);
	292	+ /*
	293	+ * Any host allocation in this function will be freed in
	294	+ * scsi_host_dev_release().
	295	+ */
298	296	out_del_dev:
299	297	device_del(&shost->shost_dev);
300	298	out_del_gendev:
	299	+ /*
	300	+ * Host state is SHOST_RUNNING so we have to explicitly release
	301	+ * ->shost_dev.
	302	+ */
	303	+ put_device(&shost->shost_dev);
301	304	device_del(&shost->shost_gendev);
302	305	out_disable_runtime_pm:
303	306	device_disable_async_suspend(&shost->shost_gendev);
304	307	pm_runtime_disable(&shost->shost_gendev);
305	308	pm_runtime_set_suspended(&shost->shost_gendev);
306	309	pm_runtime_put_noidle(&shost->shost_gendev);
307		− scsi_mq_destroy_tags(shost);
308	310	fail:
309	311	return error;
310	312	}
@@ -345,7 +347,7 @@ static void scsi_host_dev_release(struct device *dev) · 34 unmodified lines
345	347
346	348	ida_simple_remove(&host_index_ida, shost->host_no);
347	349
348		− if (parent)
	350	+ if (shost->shost_state != SHOST_CREATED)
349	351	put_device(parent);
350	352	kfree(shost);
351	353	}
@@ -388,8 +390,10 @@ struct Scsi_Host scsi_host_alloc(struct scsi_host_template sht, int privsize) · 36 unmodified lines
388	390	mutex_init(&shost->scan_mutex);
389	391
390	392	index = ida_simple_get(&host_index_ida, 0, 0, GFP_KERNEL);
391		− if (index < 0)
392		− goto fail_kfree;
	393	+ if (index < 0) {
	394	+ kfree(shost);
	395	+ return NULL;
	396	+ }
393	397	shost->host_no = index;
394	398
395	399	shost->dma_channel = 0xff;
@@ -481,7 +485,7 @@ struct Scsi_Host scsi_host_alloc(struct scsi_host_template sht, int privsize) · 85 unmodified lines
481	485	shost_printk(KERN_WARNING, shost,
482	486	"error handler thread failed to spawn, error = %ld\n",
483	487	PTR_ERR(shost->ehandler));
484		− goto fail_index_remove;
	488	+ goto fail;
485	489	}
486	490
487	491	shost->tmf_work_q = alloc_workqueue("scsi_tmf_%d",
@@ -490,17 +494,18 @@ struct Scsi_Host scsi_host_alloc(struct scsi_host_template sht, int privsize) · 2 unmodified lines
490	494	if (!shost->tmf_work_q) {
491	495	shost_printk(KERN_WARNING, shost,
492	496	"failed to create tmf workq\n");
493		− goto fail_kthread;
	497	+ goto fail;
494	498	}
495	499	scsi_proc_hostdir_add(shost->hostt);
496	500	return shost;
	501	+ fail:
	502	+ /*
	503	+ * Host state is still SHOST_CREATED and that is enough to release
	504	+ * ->shost_gendev. scsi_host_dev_release() will free
	505	+ * dev_name(&shost->shost_dev).
	506	+ */
	507	+ put_device(&shost->shost_gendev);
497	508
498		− fail_kthread:
499		− kthread_stop(shost->ehandler);
500		− fail_index_remove:
501		− ida_simple_remove(&host_index_ida, shost->host_no);
502		− fail_kfree:
503		− kfree(shost);
504	509	return NULL;
505	510	}
506	511	EXPORT_SYMBOL(scsi_host_alloc);

drivers/scsi/libsas/sas_port.c+2−2

@@ -25,7 +25,7 @@ static bool phy_is_wideport_member(struct asd_sas_port *port, struct asd_sas_phy · 24 unmodified lines
25	25
26	26	static void sas_resume_port(struct asd_sas_phy *phy)
27	27	{
28		− struct domain_device *dev;
	28	+ struct domain_device dev, n;
29	29	struct asd_sas_port *port = phy->port;
30	30	struct sas_ha_struct *sas_ha = phy->ha;
31	31	struct sas_internal *si = to_sas_internal(sas_ha->core.shost->transportt);
@@ -44,7 +44,7 @@ static void sas_resume_port(struct asd_sas_phy *phy) · 12 unmodified lines
44	44	* 1/ presume every device came back
45	45	* 2/ force the next revalidation to check all expander phys
46	46	*/
47		− list_for_each_entry(dev, &port->dev_list, dev_list_node) {
	47	+ list_for_each_entry_safe(dev, n, &port->dev_list, dev_list_node) {
48	48	int i, rc;
49	49
50	50	rc = sas_notify_lldd_dev_found(dev);

drivers/scsi/lpfc/lpfc_sli.c+1−3

@@ -20589,10 +20589,8 @@ lpfc_sli4_issue_abort_iotag(struct lpfc_hba phba, struct lpfc_iocbq cmdiocb, · 20588 unmodified lines
20589	20589	abtswqe = &abtsiocb->wqe;
20590	20590	memset(abtswqe, 0, sizeof(*abtswqe));
20591	20591
20592		− if (lpfc_is_link_up(phba))
	20592	+ if (!lpfc_is_link_up(phba))
20593	20593	bf_set(abort_cmd_ia, &abtswqe->abort_cmd, 1);
20594		− else
20595		− bf_set(abort_cmd_ia, &abtswqe->abort_cmd, 0);
20596	20594	bf_set(abort_cmd_criteria, &abtswqe->abort_cmd, T_XRI_TAG);
20597	20595	abtswqe->abort_cmd.rsrvd5 = 0;
20598	20596	abtswqe->abort_cmd.wqe_com.abort_tag = xritag;

drivers/scsi/qedf/qedf_main.c+9−11

@@ -1827,22 +1827,20 @@ static int qedf_vport_create(struct fc_vport *vport, bool disabled) · 1826 unmodified lines
1827	1827	fcoe_wwn_to_str(vport->port_name, buf, sizeof(buf));
1828	1828	QEDF_WARN(&(base_qedf->dbg_ctx), "Failed to create vport, "
1829	1829	"WWPN (0x%s) already exists.\n", buf);
1830		− goto err1;
	1830	+ return rc;
1831	1831	}
1832	1832
1833	1833	if (atomic_read(&base_qedf->link_state) != QEDF_LINK_UP) {
1834	1834	QEDF_WARN(&(base_qedf->dbg_ctx), "Cannot create vport "
1835	1835	"because link is not up.\n");
1836		− rc = -EIO;
1837		− goto err1;
	1836	+ return -EIO;
1838	1837	}
1839	1838
1840	1839	vn_port = libfc_vport_create(vport, sizeof(struct qedf_ctx));
1841	1840	if (!vn_port) {
1842	1841	QEDF_WARN(&(base_qedf->dbg_ctx), "Could not create lport "
1843	1842	"for vport.\n");
1844		− rc = -ENOMEM;
1845		− goto err1;
	1843	+ return -ENOMEM;
1846	1844	}
1847	1845
1848	1846	fcoe_wwn_to_str(vport->port_name, buf, sizeof(buf));
@@ -1866,7 +1864,7 @@ static int qedf_vport_create(struct fc_vport *vport, bool disabled) · 17 unmodified lines
1866	1864	if (rc) {
1867	1865	QEDF_ERR(&(base_qedf->dbg_ctx), "Could not allocate memory "
1868	1866	"for lport stats.\n");
1869		− goto err2;
	1867	+ goto err;
1870	1868	}
1871	1869
1872	1870	fc_set_wwnn(vn_port, vport->node_name);
@@ -1884,7 +1882,7 @@ static int qedf_vport_create(struct fc_vport *vport, bool disabled) · 11 unmodified lines
1884	1882	if (rc) {
1885	1883	QEDF_WARN(&base_qedf->dbg_ctx,
1886	1884	"Error adding Scsi_Host rc=0x%x.\n", rc);
1887		− goto err2;
	1885	+ goto err;
1888	1886	}
1889	1887
1890	1888	/* Set default dev_loss_tmo based on module parameter */
@@ -1925,9 +1923,10 @@ static int qedf_vport_create(struct fc_vport *vport, bool disabled) · 34 unmodified lines
1925	1923	vport_qedf->dbg_ctx.host_no = vn_port->host->host_no;
1926	1924	vport_qedf->dbg_ctx.pdev = base_qedf->pdev;
1927	1925
1928		−err2:
	1926	+ return 0;
	1927	+
	1928	+err:
1929	1929	scsi_host_put(vn_port->host);
1930		−err1:
1931	1930	return rc;
1932	1931	}
1933	1932
@@ -1968,8 +1967,7 @@ static int qedf_vport_destroy(struct fc_vport *vport) · 34 unmodified lines
1968	1967	fc_lport_free_stats(vn_port);
1969	1968
1970	1969	/* Release Scsi_Host */
1971		− if (vn_port->host)
1972		− scsi_host_put(vn_port->host);
	1970	+ scsi_host_put(vn_port->host);
1973	1971
1974	1972	out:
1975	1973	return 0;

drivers/scsi/qla2xxx/qla_target.c+2−0

@@ -1563,10 +1563,12 @@ void qlt_stop_phase2(struct qla_tgt *tgt) · 1562 unmodified lines
1563	1563	return;
1564	1564	}
1565	1565
	1566	+ mutex_lock(&tgt->ha->optrom_mutex);
1566	1567	mutex_lock(&vha->vha_tgt.tgt_mutex);
1567	1568	tgt->tgt_stop = 0;
1568	1569	tgt->tgt_stopped = 1;
1569	1570	mutex_unlock(&vha->vha_tgt.tgt_mutex);
	1571	+ mutex_unlock(&tgt->ha->optrom_mutex);
1570	1572
1571	1573	ql_dbg(ql_dbg_tgt_mgt, vha, 0xf00c, "Stop of tgt %p finished\n",
1572	1574	tgt);

drivers/scsi/scsi_devinfo.c+1−0

@@ -184,6 +184,7 @@ static struct { · 183 unmodified lines
184	184	{"HP", "C3323-300", "4269", BLIST_NOTQ},
185	185	{"HP", "C5713A", NULL, BLIST_NOREPORTLUN},
186	186	{"HP", "DISK-SUBSYSTEM", "*", BLIST_REPORTLUN2},
	187	+ {"HPE", "OPEN-", "*", BLIST_REPORTLUN2 \| BLIST_TRY_VPD_PAGES},
187	188	{"IBM", "AuSaV1S2", NULL, BLIST_FORCELUN},
188	189	{"IBM", "ProFibre 4000R", "*", BLIST_SPARSELUN \| BLIST_LARGELUN},
189	190	{"IBM", "2105", NULL, BLIST_RETRY_HWERROR},

drivers/scsi/ufs/ufs-mediatek.c+14−1

@@ -603,11 +603,23 @@ static void ufs_mtk_get_controller_version(struct ufs_hba *hba) · 602 unmodified lines
603	603
604	604	ret = ufshcd_dme_get(hba, UIC_ARG_MIB(PA_LOCALVERINFO), &ver);
605	605	if (!ret) {
606		− if (ver >= UFS_UNIPRO_VER_1_8)
	606	+ if (ver >= UFS_UNIPRO_VER_1_8) {
607	607	host->hw_ver.major = 3;
	608	+ /*
	609	+ * Fix HCI version for some platforms with
	610	+ * incorrect version
	611	+ */
	612	+ if (hba->ufs_version < ufshci_version(3, 0))
	613	+ hba->ufs_version = ufshci_version(3, 0);
	614	+ }
608	615	}
609	616	}
610	617
	618	+static u32 ufs_mtk_get_ufs_hci_version(struct ufs_hba *hba)
	619	+{
	620	+ return hba->ufs_version;
	621	+}
	622	+
611	623	/**
612	624	* ufs_mtk_init - find other essential mmio bases
613	625	* @hba: host controller instance
@@ -1048,6 +1060,7 @@ static void ufs_mtk_event_notify(struct ufs_hba *hba, · 434 unmodified lines
1048	1060	static const struct ufs_hba_variant_ops ufs_hba_mtk_vops = {
1049	1061	.name = "mediatek.ufshci",
1050	1062	.init = ufs_mtk_init,
	1063	+ .get_ufs_hci_version = ufs_mtk_get_ufs_hci_version,
1051	1064	.setup_clocks = ufs_mtk_setup_clocks,
1052	1065	.hce_enable_notify = ufs_mtk_hce_enable_notify,
1053	1066	.link_startup_notify = ufs_mtk_link_startup_notify,

drivers/scsi/vmw_pvscsi.c+7−1

@@ -585,7 +585,13 @@ static void pvscsi_complete_request(struct pvscsi_adapter *adapter, · 584 unmodified lines
585	585	case BTSTAT_SUCCESS:
586	586	case BTSTAT_LINKED_COMMAND_COMPLETED:
587	587	case BTSTAT_LINKED_COMMAND_COMPLETED_WITH_FLAG:
588		− /* If everything went fine, let's move on.. */
	588	+ /*
	589	+ * Commands like INQUIRY may transfer less data than
	590	+ * requested by the initiator via bufflen. Set residual
	591	+ * count to make upper layer aware of the actual amount
	592	+ * of data returned.
	593	+ */
	594	+ scsi_set_resid(cmd, scsi_bufflen(cmd) - e->dataLen);
589	595	cmd->result = (DID_OK << 16);
590	596	break;
591	597

drivers/soc/amlogic/meson-clk-measure.c+1−3

@@ -626,10 +626,8 @@ static int meson_msr_probe(struct platform_device *pdev) · 625 unmodified lines
626	626
627	627	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
628	628	base = devm_ioremap_resource(&pdev->dev, res);
629		− if (IS_ERR(base)) {
630		− dev_err(&pdev->dev, "io resource mapping failed\n");
	629	+ if (IS_ERR(base))
631	630	return PTR_ERR(base);
632		− }
633	631
634	632	priv->regmap = devm_regmap_init_mmio(&pdev->dev, base,
635	633	&meson_clk_msr_regmap_config);

drivers/soundwire/qcom.c+10−2

@@ -1150,8 +1150,16 @@ static int qcom_swrm_get_port_config(struct qcom_swrm_ctrl *ctrl) · 1149 unmodified lines
1150	1150
1151	1151	ret = of_property_read_u8_array(np, "qcom,ports-block-pack-mode",
1152	1152	bp_mode, nports);
1153		− if (ret)
1154		− return ret;
	1153	+ if (ret) {
	1154	+ u32 version;
	1155	+
	1156	+ ctrl->reg_read(ctrl, SWRM_COMP_HW_VERSION, &version);
	1157	+
	1158	+ if (version <= 0x01030000)
	1159	+ memset(bp_mode, SWR_INVALID_PARAM, QCOM_SDW_MAX_PORTS);
	1160	+ else
	1161	+ return ret;
	1162	+ }
1155	1163
1156	1164	memset(hstart, SWR_INVALID_PARAM, QCOM_SDW_MAX_PORTS);
1157	1165	of_property_read_u8_array(np, "qcom,ports-hstart", hstart, nports);

drivers/spi/Kconfig+1−1

@@ -65,7 +65,7 @@ config SPI_ALTERA · 64 unmodified lines
65	65	This is the driver for the Altera SPI Controller.
66	66
67	67	config SPI_ALTERA_CORE
68		− tristate "Altera SPI Controller core code"
	68	+ tristate "Altera SPI Controller core code" if COMPILE_TEST
69	69	select REGMAP
70	70	help
71	71	"The core code for the Altera SPI Controller"

drivers/spi/spi-bcm2835.c+8−2

@@ -68,7 +68,7 @@ · 67 unmodified lines
68	68	#define BCM2835_SPI_FIFO_SIZE 64
69	69	#define BCM2835_SPI_FIFO_SIZE_3_4 48
70	70	#define BCM2835_SPI_DMA_MIN_LENGTH 96
71		−#define BCM2835_SPI_NUM_CS 4 /* raise as necessary */
	71	+#define BCM2835_SPI_NUM_CS 24 /* raise as necessary */
72	72	#define BCM2835_SPI_MODE_BITS (SPI_CPOL \| SPI_CPHA \| SPI_CS_HIGH \
73	73	\| SPI_NO_CS \| SPI_3WIRE)
74	74
@@ -1195,6 +1195,12 @@ static int bcm2835_spi_setup(struct spi_device *spi) · 1120 unmodified lines
1195	1195	struct gpio_chip *chip;
1196	1196	u32 cs;
1197	1197
	1198	+ if (spi->chip_select >= BCM2835_SPI_NUM_CS) {
	1199	+ dev_err(&spi->dev, "only %d chip-selects supported\n",
	1200	+ BCM2835_SPI_NUM_CS - 1);
	1201	+ return -EINVAL;
	1202	+ }
	1203	+
1198	1204	/*
1199	1205	* Precalculate SPI slave's CS register value for ->prepare_message():
1200	1206	* The driver always uses software-controlled GPIO chip select, hence
@@ -1288,7 +1294,7 @@ static int bcm2835_spi_probe(struct platform_device *pdev) · 87 unmodified lines
1288	1294	ctlr->use_gpio_descriptors = true;
1289	1295	ctlr->mode_bits = BCM2835_SPI_MODE_BITS;
1290	1296	ctlr->bits_per_word_mask = SPI_BPW_MASK(8);
1291		− ctlr->num_chipselect = BCM2835_SPI_NUM_CS;
	1297	+ ctlr->num_chipselect = 3;
1292	1298	ctlr->setup = bcm2835_spi_setup;
1293	1299	ctlr->transfer_one = bcm2835_spi_transfer_one;
1294	1300	ctlr->handle_err = bcm2835_spi_handle_err;

drivers/spi/spi-bitbang.c+14−4

@@ -184,6 +184,8 @@ int spi_bitbang_setup(struct spi_device *spi) · 183 unmodified lines
184	184	{
185	185	struct spi_bitbang_cs *cs = spi->controller_state;
186	186	struct spi_bitbang *bitbang;
	187	+ bool initial_setup = false;
	188	+ int retval;
187	189
188	190	bitbang = spi_master_get_devdata(spi->master);
189	191
@@ -192,22 +194,30 @@ int spi_bitbang_setup(struct spi_device *spi) · 2 unmodified lines
192	194	if (!cs)
193	195	return -ENOMEM;
194	196	spi->controller_state = cs;
	197	+ initial_setup = true;
195	198	}
196	199
197	200	/* per-word shift register access, in hardware or bitbanging */
198	201	cs->txrx_word = bitbang->txrx_word[spi->mode & (SPI_CPOL\|SPI_CPHA)];
199		− if (!cs->txrx_word)
200		− return -EINVAL;
	202	+ if (!cs->txrx_word) {
	203	+ retval = -EINVAL;
	204	+ goto err_free;
	205	+ }
201	206
202	207	if (bitbang->setup_transfer) {
203		− int retval = bitbang->setup_transfer(spi, NULL);
	208	+ retval = bitbang->setup_transfer(spi, NULL);
204	209	if (retval < 0)
205		− return retval;
	210	+ goto err_free;
206	211	}
207	212
208	213	dev_dbg(&spi->dev, "%s, %u nsec/bit\n", __func__, 2 * cs->nsecs);
209	214
210	215	return 0;
	216	+
	217	+err_free:
	218	+ if (initial_setup)
	219	+ kfree(cs);
	220	+ return retval;
211	221	}
212	222	EXPORT_SYMBOL_GPL(spi_bitbang_setup);
213	223

drivers/spi/spi-fsl-dspi.c+3−1

@@ -1375,11 +1375,13 @@ poll_mode: · 1374 unmodified lines
1375	1375	ret = spi_register_controller(ctlr);
1376	1376	if (ret != 0) {
1377	1377	dev_err(&pdev->dev, "Problem registering DSPI ctlr\n");
1378		− goto out_free_irq;
	1378	+ goto out_release_dma;
1379	1379	}
1380	1380
1381	1381	return ret;
1382	1382
	1383	+out_release_dma:
	1384	+ dspi_release_dma(dspi);
1383	1385	out_free_irq:
1384	1386	if (dspi->irq)
1385	1387	free_irq(dspi->irq, dspi);

drivers/spi/spi-fsl-spi.c+4−0

@@ -440,6 +440,7 @@ static int fsl_spi_setup(struct spi_device *spi) · 439 unmodified lines
440	440	{
441	441	struct mpc8xxx_spi *mpc8xxx_spi;
442	442	struct fsl_spi_reg __iomem *reg_base;
	443	+ bool initial_setup = false;
443	444	int retval;
444	445	u32 hw_mode;
445	446	struct spi_mpc8xxx_cs *cs = spi_get_ctldata(spi);
@@ -452,6 +453,7 @@ static int fsl_spi_setup(struct spi_device *spi) · 6 unmodified lines
452	453	if (!cs)
453	454	return -ENOMEM;
454	455	spi_set_ctldata(spi, cs);
	456	+ initial_setup = true;
455	457	}
456	458	mpc8xxx_spi = spi_master_get_devdata(spi->master);
457	459
@@ -475,6 +477,8 @@ static int fsl_spi_setup(struct spi_device *spi) · 17 unmodified lines
475	477	retval = fsl_spi_setup_transfer(spi, NULL);
476	478	if (retval < 0) {
477	479	cs->hw_mode = hw_mode; /* Restore settings */
	480	+ if (initial_setup)
	481	+ kfree(cs);
478	482	return retval;
479	483	}
480	484

drivers/spi/spi-omap-uwire.c+8−1

@@ -424,15 +424,22 @@ done: · 423 unmodified lines
424	424	static int uwire_setup(struct spi_device *spi)
425	425	{
426	426	struct uwire_state *ust = spi->controller_state;
	427	+ bool initial_setup = false;
	428	+ int status;
427	429
428	430	if (ust == NULL) {
429	431	ust = kzalloc(sizeof(*ust), GFP_KERNEL);
430	432	if (ust == NULL)
431	433	return -ENOMEM;
432	434	spi->controller_state = ust;
	435	+ initial_setup = true;
433	436	}
434	437
435		− return uwire_setup_transfer(spi, NULL);
	438	+ status = uwire_setup_transfer(spi, NULL);
	439	+ if (status && initial_setup)
	440	+ kfree(ust);
	441	+
	442	+ return status;
436	443	}
437	444
438	445	static void uwire_cleanup(struct spi_device *spi)

drivers/spi/spi-omap2-mcspi.c+20−13

@@ -1032,8 +1032,22 @@ static void omap2_mcspi_release_dma(struct spi_master *master) · 1031 unmodified lines
1032	1032	}
1033	1033	}
1034	1034
	1035	+static void omap2_mcspi_cleanup(struct spi_device *spi)
	1036	+{
	1037	+ struct omap2_mcspi_cs *cs;
	1038	+
	1039	+ if (spi->controller_state) {
	1040	+ /* Unlink controller state from context save list */
	1041	+ cs = spi->controller_state;
	1042	+ list_del(&cs->node);
	1043	+
	1044	+ kfree(cs);
	1045	+ }
	1046	+}
	1047	+
1035	1048	static int omap2_mcspi_setup(struct spi_device *spi)
1036	1049	{
	1050	+ bool initial_setup = false;
1037	1051	int ret;
1038	1052	struct omap2_mcspi *mcspi = spi_master_get_devdata(spi->master);
1039	1053	struct omap2_mcspi_regs *ctx = &mcspi->ctx;
@@ -1051,35 +1065,28 @@ static int omap2_mcspi_setup(struct spi_device *spi) · 11 unmodified lines
1051	1065	spi->controller_state = cs;
1052	1066	/* Link this to context save list */
1053	1067	list_add_tail(&cs->node, &ctx->cs);
	1068	+ initial_setup = true;
1054	1069	}
1055	1070
1056	1071	ret = pm_runtime_get_sync(mcspi->dev);
1057	1072	if (ret < 0) {
1058	1073	pm_runtime_put_noidle(mcspi->dev);
	1074	+ if (initial_setup)
	1075	+ omap2_mcspi_cleanup(spi);
1059	1076
1060	1077	return ret;
1061	1078	}
1062	1079
1063	1080	ret = omap2_mcspi_setup_transfer(spi, NULL);
	1081	+ if (ret && initial_setup)
	1082	+ omap2_mcspi_cleanup(spi);
	1083	+
1064	1084	pm_runtime_mark_last_busy(mcspi->dev);
1065	1085	pm_runtime_put_autosuspend(mcspi->dev);
1066	1086
1067	1087	return ret;
1068	1088	}
1069	1089
1070		−static void omap2_mcspi_cleanup(struct spi_device *spi)
1071		−{
1072		− struct omap2_mcspi_cs *cs;
1073		−
1074		− if (spi->controller_state) {
1075		− /* Unlink controller state from context save list */
1076		− cs = spi->controller_state;
1077		− list_del(&cs->node);
1078		−
1079		− kfree(cs);
1080		− }
1081		−}
1082		−
1083	1090	static irqreturn_t omap2_mcspi_irq_handler(int irq, void *data)
1084	1091	{
1085	1092	struct omap2_mcspi *mcspi = data;

drivers/spi/spi-pxa2xx.c+8−1

@@ -1254,6 +1254,8 @@ static int setup_cs(struct spi_device spi, struct chip_data chip, · 1253 unmodified lines
1254	1254	chip->gpio_cs_inverted = spi->mode & SPI_CS_HIGH;
1255	1255
1256	1256	err = gpiod_direction_output(gpiod, !chip->gpio_cs_inverted);
	1257	+ if (err)
	1258	+ gpiod_put(chip->gpiod_cs);
1257	1259	}
1258	1260
1259	1261	return err;
@@ -1267,6 +1269,7 @@ static int setup(struct spi_device *spi) · 7 unmodified lines
1267	1269	struct driver_data *drv_data =
1268	1270	spi_controller_get_devdata(spi->controller);
1269	1271	uint tx_thres, tx_hi_thres, rx_thres;
	1272	+ int err;
1270	1273
1271	1274	switch (drv_data->ssp_type) {
1272	1275	case QUARK_X1000_SSP:
@@ -1413,7 +1416,11 @@ static int setup(struct spi_device *spi) · 140 unmodified lines
1413	1416	if (drv_data->ssp_type == CE4100_SSP)
1414	1417	return 0;
1415	1418
1416		− return setup_cs(spi, chip, chip_info);
	1419	+ err = setup_cs(spi, chip, chip_info);
	1420	+ if (err)
	1421	+ kfree(chip);
	1422	+
	1423	+ return err;
1417	1424	}
1418	1425
1419	1426	static void cleanup(struct spi_device *spi)

drivers/spi/spi-sc18is602.c+8−1

@@ -174,7 +174,7 @@ static int sc18is602_setup_transfer(struct sc18is602 *hw, u32 hz, u8 mode) · 173 unmodified lines
174	174	static int sc18is602_check_transfer(struct spi_device *spi,
175	175	struct spi_transfer *t, int tlen)
176	176	{
177		− if (t && t->len + tlen > SC18IS602_BUFSIZ)
	177	+ if (t && t->len + tlen > SC18IS602_BUFSIZ + 1)
178	178	return -EINVAL;
179	179
180	180	return 0;
@@ -219,6 +219,11 @@ static int sc18is602_transfer_one(struct spi_master *master, · 38 unmodified lines
219	219	return status;
220	220	}
221	221
	222	+static size_t sc18is602_max_transfer_size(struct spi_device *spi)
	223	+{
	224	+ return SC18IS602_BUFSIZ;
	225	+}
	226	+
222	227	static int sc18is602_setup(struct spi_device *spi)
223	228	{
224	229	struct sc18is602 *hw = spi_master_get_devdata(spi->master);
@@ -293,6 +298,8 @@ static int sc18is602_probe(struct i2c_client *client, · 68 unmodified lines
293	298	master->bits_per_word_mask = SPI_BPW_MASK(8);
294	299	master->setup = sc18is602_setup;
295	300	master->transfer_one_message = sc18is602_transfer_one;
	301	+ master->max_transfer_size = sc18is602_max_transfer_size;
	302	+ master->max_message_size = sc18is602_max_transfer_size;
296	303	master->dev.of_node = np;
297	304	master->min_speed_hz = hw->freq / 128;
298	305	master->max_speed_hz = hw->freq / 4;

drivers/spi/spi-sprd.c+1−0

@@ -1068,6 +1068,7 @@ static const struct of_device_id sprd_spi_of_match[] = { · 1067 unmodified lines
1068	1068	{ .compatible = "sprd,sc9860-spi", },
1069	1069	{ /* sentinel */ }
1070	1070	};
	1071	+MODULE_DEVICE_TABLE(of, sprd_spi_of_match);
1071	1072
1072	1073	static struct platform_driver sprd_spi_driver = {
1073	1074	.driver = {

drivers/spi/spi-stm32-qspi.c+4−1

@@ -294,7 +294,7 @@ static int stm32_qspi_wait_cmd(struct stm32_qspi *qspi, · 293 unmodified lines
294	294	int err = 0;
295	295
296	296	if (!op->data.nbytes)
297		− return stm32_qspi_wait_nobusy(qspi);
	297	+ goto wait_nobusy;
298	298
299	299	if (readl_relaxed(qspi->io_base + QSPI_SR) & SR_TCF)
300	300	goto out;
@@ -315,6 +315,9 @@ static int stm32_qspi_wait_cmd(struct stm32_qspi *qspi, · 14 unmodified lines
315	315	out:
316	316	/* clear flags */
317	317	writel_relaxed(FCR_CTCF \| FCR_CTEF, qspi->io_base + QSPI_FCR);
	318	+wait_nobusy:
	319	+ if (!err)
	320	+ err = stm32_qspi_wait_nobusy(qspi);
318	321
319	322	return err;
320	323	}

drivers/spi/spi-zynq-qspi.c+8−8

@@ -367,7 +367,7 @@ static int zynq_qspi_config_op(struct zynq_qspi xqspi, struct spi_device spi) · 366 unmodified lines
367	367	}
368	368
369	369	/**
370		− * zynq_qspi_setup - Configure the QSPI controller
	370	+ * zynq_qspi_setup_op - Configure the QSPI controller
371	371	* @spi: Pointer to the spi_device structure
372	372	*
373	373	* Sets the operational mode of QSPI controller for the next QSPI transfer, baud
@@ -528,18 +528,17 @@ static int zynq_qspi_exec_mem_op(struct spi_mem *mem, · 154 unmodified lines
528	528	struct zynq_qspi *xqspi = spi_controller_get_devdata(mem->spi->master);
529	529	int err = 0, i;
530	530	u8 *tmpbuf;
531		− u8 opcode = op->cmd.opcode;
532	531
533	532	dev_dbg(xqspi->dev, "cmd:%#x mode:%d.%d.%d.%d\n",
534		− opcode, op->cmd.buswidth, op->addr.buswidth,
	533	+ op->cmd.opcode, op->cmd.buswidth, op->addr.buswidth,
535	534	op->dummy.buswidth, op->data.buswidth);
536	535
537	536	zynq_qspi_chipselect(mem->spi, true);
538	537	zynq_qspi_config_op(xqspi, mem->spi);
539	538
540		− if (op->cmd.nbytes) {
	539	+ if (op->cmd.opcode) {
541	540	reinit_completion(&xqspi->data_completion);
542		− xqspi->txbuf = &opcode;
	541	+ xqspi->txbuf = (u8 *)&op->cmd.opcode;
543	542	xqspi->rxbuf = NULL;
544	543	xqspi->tx_bytes = op->cmd.nbytes;
545	544	xqspi->rx_bytes = op->cmd.nbytes;
@@ -679,14 +678,14 @@ static int zynq_qspi_probe(struct platform_device *pdev) · 133 unmodified lines
679	678	xqspi->irq = platform_get_irq(pdev, 0);
680	679	if (xqspi->irq <= 0) {
681	680	ret = -ENXIO;
682		− goto remove_master;
	681	+ goto clk_dis_all;
683	682	}
684	683	ret = devm_request_irq(&pdev->dev, xqspi->irq, zynq_qspi_irq,
685	684	0, pdev->name, xqspi);
686	685	if (ret != 0) {
687	686	ret = -ENXIO;
688	687	dev_err(&pdev->dev, "request_irq failed\n");
689		− goto remove_master;
	688	+ goto clk_dis_all;
690	689	}
691	690
692	691	ret = of_property_read_u32(np, "num-cs",
@@ -694,8 +693,9 @@ static int zynq_qspi_probe(struct platform_device *pdev) · 1 unmodified line
694	693	if (ret < 0) {
695	694	ctlr->num_chipselect = 1;
696	695	} else if (num_cs > ZYNQ_QSPI_MAX_NUM_CS) {
	696	+ ret = -EINVAL;
697	697	dev_err(&pdev->dev, "only 2 chip selects are available\n");
698		− goto remove_master;
	698	+ goto clk_dis_all;
699	699	} else {
700	700	ctlr->num_chipselect = num_cs;
701	701	}

drivers/spi/spi.c+38−13

@@ -47,10 +47,6 @@ static void spidev_release(struct device *dev) · 46 unmodified lines
47	47	{
48	48	struct spi_device *spi = to_spi_device(dev);
49	49
50		− /* spi controllers may cleanup for released devices */
51		− if (spi->controller->cleanup)
52		− spi->controller->cleanup(spi);
53		−
54	50	spi_controller_put(spi->controller);
55	51	kfree(spi->driver_override);
56	52	kfree(spi);
@@ -558,6 +554,12 @@ static int spi_dev_check(struct device dev, void data) · 501 unmodified lines
558	554	return 0;
559	555	}
560	556
	557	+static void spi_cleanup(struct spi_device *spi)
	558	+{
	559	+ if (spi->controller->cleanup)
	560	+ spi->controller->cleanup(spi);
	561	+}
	562	+
561	563	/**
562	564	* spi_add_device - Add spi_device allocated with spi_alloc_device
563	565	* @spi: spi_device to register
@@ -622,11 +624,13 @@ int spi_add_device(struct spi_device *spi) · 58 unmodified lines
622	624
623	625	/* Device may be bound to an active driver when this returns */
624	626	status = device_add(&spi->dev);
625		− if (status < 0)
	627	+ if (status < 0) {
626	628	dev_err(dev, "can't add %s, status %d\n",
627	629	dev_name(&spi->dev), status);
628		− else
	630	+ spi_cleanup(spi);
	631	+ } else {
629	632	dev_dbg(dev, "registered child %s\n", dev_name(&spi->dev));
	633	+ }
630	634
631	635	done:
632	636	mutex_unlock(&spi_add_lock);
@@ -717,7 +721,9 @@ void spi_unregister_device(struct spi_device *spi) · 84 unmodified lines
717	721	if (ACPI_COMPANION(&spi->dev))
718	722	acpi_device_clear_enumerated(ACPI_COMPANION(&spi->dev));
719	723	device_remove_software_node(&spi->dev);
720		− device_unregister(&spi->dev);
	724	+ device_del(&spi->dev);
	725	+ spi_cleanup(spi);
	726	+ put_device(&spi->dev);
721	727	}
722	728	EXPORT_SYMBOL_GPL(spi_unregister_device);
723	729
@@ -814,15 +820,29 @@ static void spi_set_cs(struct spi_device *spi, bool enable, bool force) · 90 unmodified lines
814	820
815	821	if (spi->cs_gpiod \|\| gpio_is_valid(spi->cs_gpio)) {
816	822	if (!(spi->mode & SPI_NO_CS)) {
817		− if (spi->cs_gpiod)
818		− /* polarity handled by gpiolib */
819		− gpiod_set_value_cansleep(spi->cs_gpiod, activate);
820		− else
	823	+ if (spi->cs_gpiod) {
	824	+ /*
	825	+ * Historically ACPI has no means of the GPIO polarity and
	826	+ * thus the SPISerialBus() resource defines it on the per-chip
	827	+ * basis. In order to avoid a chain of negations, the GPIO
	828	+ * polarity is considered being Active High. Even for the cases
	829	+ * when _DSD() is involved (in the updated versions of ACPI)
	830	+ * the GPIO CS polarity must be defined Active High to avoid
	831	+ * ambiguity. That's why we use enable, that takes SPI_CS_HIGH
	832	+ * into account.
	833	+ */
	834	+ if (has_acpi_companion(&spi->dev))
	835	+ gpiod_set_value_cansleep(spi->cs_gpiod, !enable);
	836	+ else
	837	+ /* Polarity handled by GPIO library */
	838	+ gpiod_set_value_cansleep(spi->cs_gpiod, activate);
	839	+ } else {
821	840	/*
822	841	* invert the enable line, as active low is
823	842	* default for SPI.
824	843	*/
825	844	gpio_set_value_cansleep(spi->cs_gpio, !enable);
	845	+ }
826	846	}
827	847	/* Some SPI masters need both GPIO CS & slave_select */
828	848	if ((spi->controller->flags & SPI_MASTER_GPIO_SS) &&
@@ -3451,9 +3471,12 @@ int spi_set_cs_timing(struct spi_device spi, struct spi_delay setup, · 2622 unmodified lines
3451	3471
3452	3472	if (spi->controller->set_cs_timing &&
3453	3473	!(spi->cs_gpiod \|\| gpio_is_valid(spi->cs_gpio))) {
	3474	+ mutex_lock(&spi->controller->io_mutex);
	3475	+
3454	3476	if (spi->controller->auto_runtime_pm) {
3455	3477	status = pm_runtime_get_sync(parent);
3456	3478	if (status < 0) {
	3479	+ mutex_unlock(&spi->controller->io_mutex);
3457	3480	pm_runtime_put_noidle(parent);
3458	3481	dev_err(&spi->controller->dev, "Failed to power device: %d\n",
3459	3482	status);
@@ -3464,11 +3487,13 @@ int spi_set_cs_timing(struct spi_device spi, struct spi_delay setup, · 4 unmodified lines
3464	3487	hold, inactive);
3465	3488	pm_runtime_mark_last_busy(parent);
3466	3489	pm_runtime_put_autosuspend(parent);
3467		− return status;
3468	3490	} else {
3469		− return spi->controller->set_cs_timing(spi, setup, hold,
	3491	+ status = spi->controller->set_cs_timing(spi, setup, hold,
3470	3492	inactive);
3471	3493	}
	3494	+
	3495	+ mutex_unlock(&spi->controller->io_mutex);
	3496	+ return status;
3472	3497	}
3473	3498
3474	3499	if ((setup && setup->unit == SPI_DELAY_UNIT_SCK) \|\|

drivers/staging/emxx_udc/emxx_udc.c+2−2

@@ -2064,7 +2064,7 @@ static int _nbu2ss_nuke(struct nbu2ss_udc *udc, · 2063 unmodified lines
2064	2064	struct nbu2ss_ep *ep,
2065	2065	int status)
2066	2066	{
2067		− struct nbu2ss_req *req;
	2067	+ struct nbu2ss_req req, n;
2068	2068
2069	2069	/* Endpoint Disable */
2070	2070	_nbu2ss_epn_exit(udc, ep);
@@ -2076,7 +2076,7 @@ static int _nbu2ss_nuke(struct nbu2ss_udc *udc, · 5 unmodified lines
2076	2076	return 0;
2077	2077
2078	2078	/* called with irqs blocked */
2079		− list_for_each_entry(req, &ep->queue, queue) {
	2079	+ list_for_each_entry_safe(req, n, &ep->queue, queue) {
2080	2080	_nbu2ss_ep_done(ep, req, status);
2081	2081	}
2082	2082

drivers/staging/iio/cdc/ad7746.c+0−1

@@ -700,7 +700,6 @@ static int ad7746_probe(struct i2c_client *client, · 699 unmodified lines
700	700	indio_dev->num_channels = ARRAY_SIZE(ad7746_channels);
701	701	else
702	702	indio_dev->num_channels = ARRAY_SIZE(ad7746_channels) - 2;
703		− indio_dev->num_channels = ARRAY_SIZE(ad7746_channels);
704	703	indio_dev->modes = INDIO_DIRECT_MODE;
705	704
706	705	if (pdata) {

drivers/staging/ralink-gdma/ralink-gdma.c+0−2

@@ -1,6 +1,5 @@
1	1	// SPDX-License-Identifier: GPL-2.0+
2	2	/*
3		− * Copyright (C) 2013, Lars-Peter Clausen <lars@metafoo.de>
4	3	* GDMA4740 DMAC support
5	4	*/
6	5
@@ -914,6 +913,5 @@ static struct platform_driver gdma_dma_driver = { · 907 unmodified lines
914	913	};
915	914	module_platform_driver(gdma_dma_driver);
916	915
917		−MODULE_AUTHOR("Lars-Peter Clausen <lars@metafoo.de>");
918	916	MODULE_DESCRIPTION("Ralink/MTK DMA driver");
919	917	MODULE_LICENSE("GPL v2");

drivers/staging/rtl8723bs/os_dep/ioctl_cfg80211.c+3−3

@@ -2091,7 +2091,7 @@ void rtw_cfg80211_indicate_sta_assoc(struct adapter padapter, u8 pmgmt_frame, · 2090 unmodified lines
2091	2091	struct net_device *ndev = padapter->pnetdev;
2092	2092
2093	2093	{
2094		− struct station_info sinfo;
	2094	+ struct station_info sinfo = {};
2095	2095	u8 ie_offset;
2096	2096	if (GetFrameSubType(pmgmt_frame) == WIFI_ASSOCREQ)
2097	2097	ie_offset = _ASOCREQ_IE_OFFSET_;
@@ -2284,7 +2284,7 @@ static int rtw_cfg80211_add_monitor_if(struct adapter padapter, char name, str · 186 unmodified lines
2284	2284	mon_wdev->iftype = NL80211_IFTYPE_MONITOR;
2285	2285	mon_ndev->ieee80211_ptr = mon_wdev;
2286	2286
2287		− ret = register_netdevice(mon_ndev);
	2287	+ ret = cfg80211_register_netdevice(mon_ndev);
2288	2288	if (ret) {
2289	2289	goto out;
2290	2290	}
@@ -2360,7 +2360,7 @@ static int cfg80211_rtw_del_virtual_intf(struct wiphy *wiphy, · 69 unmodified lines
2360	2360	adapter = rtw_netdev_priv(ndev);
2361	2361	pwdev_priv = adapter_wdev_data(adapter);
2362	2362
2363		− unregister_netdevice(ndev);
	2363	+ cfg80211_unregister_netdevice(ndev);
2364	2364
2365	2365	if (ndev == pwdev_priv->pmon_ndev) {
2366	2366	pwdev_priv->pmon_ndev = NULL;

drivers/target/target_core_iblock.c+2−2

@@ -204,11 +204,11 @@ static struct se_dev_plug iblock_plug_device(struct se_device se_dev) · 203 unmodified lines
204	204	struct iblock_dev_plug *ib_dev_plug;
205	205
206	206	/*
207		− * Each se_device has a per cpu work this can be run from. Wwe
	207	+ * Each se_device has a per cpu work this can be run from. We
208	208	* shouldn't have multiple threads on the same cpu calling this
209	209	* at the same time.
210	210	*/
211		− ib_dev_plug = &ib_dev->ibd_plug[smp_processor_id()];
	211	+ ib_dev_plug = &ib_dev->ibd_plug[raw_smp_processor_id()];
212	212	if (test_and_set_bit(IBD_PLUGF_PLUGGED, &ib_dev_plug->flags))
213	213	return NULL;
214	214

drivers/target/target_core_transport.c+2−4

@@ -1416,7 +1416,7 @@ void __target_init_cmd( · 1415 unmodified lines
1416	1416	cmd->orig_fe_lun = unpacked_lun;
1417	1417
1418	1418	if (!(cmd->se_cmd_flags & SCF_USE_CPUID))
1419		− cmd->cpuid = smp_processor_id();
	1419	+ cmd->cpuid = raw_smp_processor_id();
1420	1420
1421	1421	cmd->state_active = false;
1422	1422	}
@@ -3121,9 +3121,7 @@ __transport_wait_for_tasks(struct se_cmd *cmd, bool fabric_stop, · 1698 unmodified lines
3121	3121	__releases(&cmd->t_state_lock)
3122	3122	__acquires(&cmd->t_state_lock)
3123	3123	{
3124		−
3125		− assert_spin_locked(&cmd->t_state_lock);
3126		− WARN_ON_ONCE(!irqs_disabled());
	3124	+ lockdep_assert_held(&cmd->t_state_lock);
3127	3125
3128	3126	if (fabric_stop)
3129	3127	cmd->transport_state \|= CMD_T_FABRIC_STOP;

drivers/target/target_core_user.c+7−5

@@ -516,8 +516,10 @@ static inline int tcmu_get_empty_block(struct tcmu_dev *udev, · 515 unmodified lines
516	516	dpi = dbi * udev->data_pages_per_blk;
517	517	/* Count the number of already allocated pages */
518	518	xas_set(&xas, dpi);
	519	+ rcu_read_lock();
519	520	for (cnt = 0; xas_next(&xas) && cnt < page_cnt;)
520	521	cnt++;
	522	+ rcu_read_unlock();
521	523
522	524	for (i = cnt; i < page_cnt; i++) {
523	525	/* try to get new page from the mm */
@@ -699,11 +701,10 @@ static inline void tcmu_copy_data(struct tcmu_dev *udev, · 175 unmodified lines
699	701	struct scatterlist *sg, unsigned int sg_nents,
700	702	struct iovec **iov, size_t data_len)
701	703	{
702		− XA_STATE(xas, &udev->data_pages, 0);
703	704	/* start value of dbi + 1 must not be a valid dbi */
704	705	int dbi = -2;
705	706	size_t page_remaining, cp_len;
706		− int page_cnt, page_inx;
	707	+ int page_cnt, page_inx, dpi;
707	708	struct sg_mapping_iter sg_iter;
708	709	unsigned int sg_flags;
709	710	struct page *page;
@@ -726,9 +727,10 @@ static inline void tcmu_copy_data(struct tcmu_dev *udev, · 16 unmodified lines
726	727	if (page_cnt > udev->data_pages_per_blk)
727	728	page_cnt = udev->data_pages_per_blk;
728	729
729		− xas_set(&xas, dbi * udev->data_pages_per_blk);
730		− for (page_inx = 0; page_inx < page_cnt && data_len; page_inx++) {
731		− page = xas_next(&xas);
	730	+ dpi = dbi * udev->data_pages_per_blk;
	731	+ for (page_inx = 0; page_inx < page_cnt && data_len;
	732	+ page_inx++, dpi++) {
	733	+ page = xa_load(&udev->data_pages, dpi);
732	734
733	735	if (direction == TCMU_DATA_AREA_TO_SG)
734	736	flush_dcache_page(page);

drivers/tee/optee/call.c+4−2

@@ -220,6 +220,7 @@ int optee_open_session(struct tee_context *ctx, · 219 unmodified lines
220	220	struct optee_msg_arg *msg_arg;
221	221	phys_addr_t msg_parg;
222	222	struct optee_session *sess = NULL;
	223	+ uuid_t client_uuid;
223	224
224	225	/* +2 for the meta parameters added below */
225	226	shm = get_msg_arg(ctx, arg->num_params + 2, &msg_arg, &msg_parg);
@@ -240,10 +241,11 @@ int optee_open_session(struct tee_context *ctx, · 14 unmodified lines
240	241	memcpy(&msg_arg->params[0].u.value, arg->uuid, sizeof(arg->uuid));
241	242	msg_arg->params[1].u.value.c = arg->clnt_login;
242	243
243		− rc = tee_session_calc_client_uuid((uuid_t *)&msg_arg->params[1].u.value,
244		− arg->clnt_login, arg->clnt_uuid);
	244	+ rc = tee_session_calc_client_uuid(&client_uuid, arg->clnt_login,
	245	+ arg->clnt_uuid);
245	246	if (rc)
246	247	goto out;
	248	+ export_uuid(msg_arg->params[1].u.octets, &client_uuid);
247	249
248	250	rc = optee_to_msg_param(msg_arg->params + 2, arg->num_params, param);
249	251	if (rc)

drivers/tee/optee/optee_msg.h+4−2

@@ -9,7 +9,7 @@ · 8 unmodified lines
9	9	#include <linux/types.h>
10	10
11	11	/*
12		− * This file defines the OP-TEE message protocol used to communicate
	12	+ * This file defines the OP-TEE message protocol (ABI) used to communicate
13	13	* with an instance of OP-TEE running in secure world.
14	14	*
15	15	* This file is divided into two sections.
@@ -144,9 +144,10 @@ struct optee_msg_param_value { · 128 unmodified lines
144	144	* @tmem: parameter by temporary memory reference
145	145	* @rmem: parameter by registered memory reference
146	146	* @value: parameter by opaque value
	147	+ * @octets: parameter by octet string
147	148	*
148	149	* @attr & OPTEE_MSG_ATTR_TYPE_MASK indicates if tmem, rmem or value is used in
149		− * the union. OPTEE_MSG_ATTR_TYPE_VALUE_* indicates value,
	150	+ * the union. OPTEE_MSG_ATTR_TYPE_VALUE_* indicates value or octets,
150	151	* OPTEE_MSG_ATTR_TYPE_TMEM_* indicates @tmem and
151	152	* OPTEE_MSG_ATTR_TYPE_RMEM_* indicates @rmem,
152	153	* OPTEE_MSG_ATTR_TYPE_NONE indicates that none of the members are used.
@@ -157,6 +158,7 @@ struct optee_msg_param { · 4 unmodified lines
157	158	struct optee_msg_param_tmem tmem;
158	159	struct optee_msg_param_rmem rmem;
159	160	struct optee_msg_param_value value;
	161	+ u8 octets[24];
160	162	} u;
161	163	};
162	164

drivers/thermal/intel/int340x_thermal/int340x_thermal_zone.c+4−0

@@ -237,6 +237,8 @@ struct int34x_thermal_zone int340x_thermal_zone_add(struct acpi_device adev, · 236 unmodified lines
237	237	if (ACPI_FAILURE(status))
238	238	trip_cnt = 0;
239	239	else {
	240	+ int i;
	241	+
240	242	int34x_thermal_zone->aux_trips =
241	243	kcalloc(trip_cnt,
242	244	sizeof(*int34x_thermal_zone->aux_trips),
@@ -247,6 +249,8 @@ struct int34x_thermal_zone int340x_thermal_zone_add(struct acpi_device adev, · 4 unmodified lines
247	249	}
248	250	trip_mask = BIT(trip_cnt) - 1;
249	251	int34x_thermal_zone->aux_trip_nr = trip_cnt;
	252	+ for (i = 0; i < trip_cnt; ++i)
	253	+ int34x_thermal_zone->aux_trips[i] = THERMAL_TEMP_INVALID;
250	254	}
251	255
252	256	trip_cnt = int340x_thermal_read_trips(int34x_thermal_zone);

drivers/thermal/intel/therm_throt.c+11−4

@@ -621,6 +621,17 @@ bool x86_thermal_enabled(void) · 620 unmodified lines
621	621	return atomic_read(&therm_throt_en);
622	622	}
623	623
	624	+void __init therm_lvt_init(void)
	625	+{
	626	+ /*
	627	+ * This function is only called on boot CPU. Save the init thermal
	628	+ * LVT value on BSP and use that value to restore APs' thermal LVT
	629	+ * entry BIOS programmed later
	630	+ */
	631	+ if (intel_thermal_supported(&boot_cpu_data))
	632	+ lvtthmr_init = apic_read(APIC_LVTTHMR);
	633	+}
	634	+
624	635	void intel_init_thermal(struct cpuinfo_x86 *c)
625	636	{
626	637	unsigned int cpu = smp_processor_id();
@@ -630,10 +641,6 @@ void intel_init_thermal(struct cpuinfo_x86 *c) · 3 unmodified lines
630	641	if (!intel_thermal_supported(c))
631	642	return;
632	643
633		− /* On the BSP? */
634		− if (c == &boot_cpu_data)
635		− lvtthmr_init = apic_read(APIC_LVTTHMR);
636		−
637	644	/*
638	645	* First check if its enabled already, in which case there might
639	646	* be some SMM goo which handles it, so we can't even put a handler

drivers/thermal/intel/x86_pkg_temp_thermal.c+1−1

@@ -166,7 +166,7 @@ static int sys_get_trip_temp(struct thermal_zone_device *tzd, · 165 unmodified lines
166	166	if (thres_reg_value)
167	167	temp = zonedev->tj_max - thres_reg_value 1000;
168	168	else
169		− *temp = 0;
	169	+ *temp = THERMAL_TEMP_INVALID;
170	170	pr_debug("sys_get_trip_temp %d\n", *temp);
171	171
172	172	return 0;

drivers/thermal/qcom/qcom-spmi-adc-tm5.c+1−1

@@ -441,7 +441,7 @@ static int adc_tm5_get_dt_channel_data(struct adc_tm5_chip *adc_tm, · 440 unmodified lines
441	441
442	442	if (args.args_count != 1 \|\| args.args[0] >= ADC5_MAX_CHANNEL) {
443	443	dev_err(dev, "%s: invalid ADC channel number %d\n", name, chan);
444		− return ret;
	444	+ return -EINVAL;
445	445	}
446	446	channel->adc_channel = args.args[0];
447	447

drivers/thermal/ti-soc-thermal/ti-bandgap.c+1−1

@@ -770,7 +770,7 @@ static int ti_bandgap_tshut_init(struct ti_bandgap *bgp, · 769 unmodified lines
770	770	}
771	771
772	772	/**
773		− * ti_bandgap_alert_init() - setup and initialize talert handling
	773	+ * ti_bandgap_talert_init() - setup and initialize talert handling
774	774	* @bgp: pointer to struct ti_bandgap
775	775	* @pdev: pointer to device struct platform_device
776	776	*

drivers/thunderbolt/dma_port.c+6−5

@@ -366,15 +366,15 @@ int dma_port_flash_read(struct tb_dma_port *dma, unsigned int address, · 365 unmodified lines
366	366	void *buf, size_t size)
367	367	{
368	368	unsigned int retries = DMA_PORT_RETRIES;
369		− unsigned int offset;
370		−
371		− offset = address & 3;
372		− address = address & ~3;
373	369
374	370	do {
375		− u32 nbytes = min_t(u32, size, MAIL_DATA_DWORDS * 4);
	371	+ unsigned int offset;
	372	+ size_t nbytes;
376	373	int ret;
377	374
	375	+ offset = address & 3;
	376	+ nbytes = min_t(size_t, size + offset, MAIL_DATA_DWORDS * 4);
	377	+
378	378	ret = dma_port_flash_read_block(dma, address, dma->buf,
379	379	ALIGN(nbytes, 4));
380	380	if (ret) {
@@ -386,6 +386,7 @@ int dma_port_flash_read(struct tb_dma_port *dma, unsigned int address, · 5 unmodified lines
386	386	return ret;
387	387	}
388	388
	389	+ nbytes -= offset;
389	390	memcpy(buf, dma->buf + offset, nbytes);
390	391
391	392	size -= nbytes;

drivers/thunderbolt/usb4.c+5−4

@@ -68,15 +68,15 @@ static int usb4_do_read_data(u16 address, void *buf, size_t size, · 67 unmodified lines
68	68	unsigned int retries = USB4_DATA_RETRIES;
69	69	unsigned int offset;
70	70
71		− offset = address & 3;
72		− address = address & ~3;
73		−
74	71	do {
75		− size_t nbytes = min_t(size_t, size, USB4_DATA_DWORDS * 4);
76	72	unsigned int dwaddress, dwords;
77	73	u8 data[USB4_DATA_DWORDS * 4];
	74	+ size_t nbytes;
78	75	int ret;
79	76
	77	+ offset = address & 3;
	78	+ nbytes = min_t(size_t, size + offset, USB4_DATA_DWORDS * 4);
	79	+
80	80	dwaddress = address / 4;
81	81	dwords = ALIGN(nbytes, 4) / 4;
82	82
@@ -87,6 +87,7 @@ static int usb4_do_read_data(u16 address, void *buf, size_t size, · 4 unmodified lines
87	87	return ret;
88	88	}
89	89
	90	+ nbytes -= offset;
90	91	memcpy(buf, data + offset, nbytes);
91	92
92	93	size -= nbytes;

drivers/tty/serial/8250/8250.h+17−15

@@ -7,6 +7,7 @@ · 6 unmodified lines
7	7	* Copyright (C) 2001 Russell King.
8	8	*/
9	9
	10	+#include <linux/bits.h>
10	11	#include <linux/serial_8250.h>
11	12	#include <linux/serial_reg.h>
12	13	#include <linux/dmaengine.h>
@@ -70,24 +71,25 @@ struct serial8250_config { · 57 unmodified lines
70	71	unsigned int flags;
71	72	};
72	73
73		−#define UART_CAP_FIFO (1 << 8) /* UART has FIFO */
74		−#define UART_CAP_EFR (1 << 9) /* UART has EFR */
75		−#define UART_CAP_SLEEP (1 << 10) /* UART has IER sleep */
76		−#define UART_CAP_AFE (1 << 11) /* MCR-based hw flow control */
77		−#define UART_CAP_UUE (1 << 12) /* UART needs IER bit 6 set (Xscale) */
78		−#define UART_CAP_RTOIE (1 << 13) /* UART needs IER bit 4 set (Xscale, Tegra) */
79		−#define UART_CAP_HFIFO (1 << 14) /* UART has a "hidden" FIFO */
80		−#define UART_CAP_RPM (1 << 15) /* Runtime PM is active while idle */
81		−#define UART_CAP_IRDA (1 << 16) /* UART supports IrDA line discipline */
82		−#define UART_CAP_MINI (1 << 17) /* Mini UART on BCM283X family lacks:
	74	+#define UART_CAP_FIFO BIT(8) /* UART has FIFO */
	75	+#define UART_CAP_EFR BIT(9) /* UART has EFR */
	76	+#define UART_CAP_SLEEP BIT(10) /* UART has IER sleep */
	77	+#define UART_CAP_AFE BIT(11) /* MCR-based hw flow control */
	78	+#define UART_CAP_UUE BIT(12) /* UART needs IER bit 6 set (Xscale) */
	79	+#define UART_CAP_RTOIE BIT(13) /* UART needs IER bit 4 set (Xscale, Tegra) */
	80	+#define UART_CAP_HFIFO BIT(14) /* UART has a "hidden" FIFO */
	81	+#define UART_CAP_RPM BIT(15) /* Runtime PM is active while idle */
	82	+#define UART_CAP_IRDA BIT(16) /* UART supports IrDA line discipline */
	83	+#define UART_CAP_MINI BIT(17) /* Mini UART on BCM283X family lacks:
83	84	* STOP PARITY EPAR SPAR WLEN5 WLEN6
84	85	*/
85	86
86		−#define UART_BUG_QUOT (1 << 0) /* UART has buggy quot LSB */
87		−#define UART_BUG_TXEN (1 << 1) /* UART has buggy TX IIR status */
88		−#define UART_BUG_NOMSR (1 << 2) /* UART has buggy MSR status bits (Au1x00) */
89		−#define UART_BUG_THRE (1 << 3) /* UART has buggy THRE reassertion */
90		−#define UART_BUG_PARITY (1 << 4) /* UART mishandles parity if FIFO enabled */
	87	+#define UART_BUG_QUOT BIT(0) /* UART has buggy quot LSB */
	88	+#define UART_BUG_TXEN BIT(1) /* UART has buggy TX IIR status */
	89	+#define UART_BUG_NOMSR BIT(2) /* UART has buggy MSR status bits (Au1x00) */
	90	+#define UART_BUG_THRE BIT(3) /* UART has buggy THRE reassertion */
	91	+#define UART_BUG_PARITY BIT(4) /* UART mishandles parity if FIFO enabled */
	92	+#define UART_BUG_TXRACE BIT(5) /* UART Tx fails to set remote DR */
91	93
92	94
93	95	#ifdef CONFIG_SERIAL_8250_SHARE_IRQ

drivers/tty/serial/8250/8250_aspeed_vuart.c+1−0

@@ -437,6 +437,7 @@ static int aspeed_vuart_probe(struct platform_device *pdev) · 436 unmodified lines
437	437	port.port.status = UPSTAT_SYNC_FIFO;
438	438	port.port.dev = &pdev->dev;
439	439	port.port.has_sysrq = IS_ENABLED(CONFIG_SERIAL_8250_CONSOLE);
	440	+ port.bugs \|= UART_BUG_TXRACE;
440	441
441	442	rc = sysfs_create_group(&vuart->dev->kobj, &aspeed_vuart_attr_group);
442	443	if (rc < 0)

drivers/tty/serial/8250/8250_dw.c+1−0

@@ -714,6 +714,7 @@ static const struct acpi_device_id dw8250_acpi_match[] = { · 713 unmodified lines
714	714	{ "APMC0D08", 0},
715	715	{ "AMD0020", 0 },
716	716	{ "AMDI0020", 0 },
	717	+ { "AMDI0022", 0 },
717	718	{ "BRCM2032", 0 },
718	719	{ "HISI0031", 0 },
719	720	{ },

drivers/tty/serial/8250/8250_exar.c+5−1

@@ -553,7 +553,11 @@ static void pci_xr17v35x_exit(struct pci_dev *pcidev) · 552 unmodified lines
553	553	{
554	554	struct exar8250 *priv = pci_get_drvdata(pcidev);
555	555	struct uart_8250_port *port = serial8250_get_port(priv->line[0]);
556		− struct platform_device *pdev = port->port.private_data;
	556	+ struct platform_device *pdev;
	557	+
	558	+ pdev = port->port.private_data;
	559	+ if (!pdev)
	560	+ return;
557	561
558	562	device_remove_software_node(&pdev->dev);
559	563	platform_device_unregister(pdev);

drivers/tty/serial/8250/8250_pci.c+35−12

@@ -56,6 +56,8 @@ struct serial_private { · 55 unmodified lines
56	56	int line[];
57	57	};
58	58
	59	+#define PCI_DEVICE_ID_HPE_PCI_SERIAL 0x37e
	60	+
59	61	static const struct pci_device_id pci_use_msi[] = {
60	62	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_NETMOS, PCI_DEVICE_ID_NETMOS_9900,
61	63	0xA000, 0x1000) },
@@ -63,6 +65,8 @@ static const struct pci_device_id pci_use_msi[] = { · 1 unmodified line
63	65	0xA000, 0x1000) },
64	66	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_NETMOS, PCI_DEVICE_ID_NETMOS_9922,
65	67	0xA000, 0x1000) },
	68	+ { PCI_DEVICE_SUB(PCI_VENDOR_ID_HP_3PAR, PCI_DEVICE_ID_HPE_PCI_SERIAL,
	69	+ PCI_ANY_ID, PCI_ANY_ID) },
66	70	{ }
67	71	};
68	72
@@ -1997,6 +2001,16 @@ static struct pci_serial_quirk pci_serial_quirks[] = { · 1928 unmodified lines
1997	2001	.init = pci_hp_diva_init,
1998	2002	.setup = pci_hp_diva_setup,
1999	2003	},
	2004	+ /*
	2005	+ * HPE PCI serial device
	2006	+ */
	2007	+ {
	2008	+ .vendor = PCI_VENDOR_ID_HP_3PAR,
	2009	+ .device = PCI_DEVICE_ID_HPE_PCI_SERIAL,
	2010	+ .subvendor = PCI_ANY_ID,
	2011	+ .subdevice = PCI_ANY_ID,
	2012	+ .setup = pci_hp_diva_setup,
	2013	+ },
2000	2014	/*
2001	2015	* Intel
2002	2016	*/
@@ -3944,21 +3958,26 @@ pciserial_init_ports(struct pci_dev dev, const struct pciserial_board board) · 1941 unmodified lines
3944	3958	uart.port.flags = UPF_SKIP_TEST \| UPF_BOOT_AUTOCONF \| UPF_SHARE_IRQ;
3945	3959	uart.port.uartclk = board->base_baud * 16;
3946	3960
3947		− if (pci_match_id(pci_use_msi, dev)) {
3948		− dev_dbg(&dev->dev, "Using MSI(-X) interrupts\n");
3949		− pci_set_master(dev);
3950		− rc = pci_alloc_irq_vectors(dev, 1, 1, PCI_IRQ_ALL_TYPES);
	3961	+ if (board->flags & FL_NOIRQ) {
	3962	+ uart.port.irq = 0;
3951	3963	} else {
3952		− dev_dbg(&dev->dev, "Using legacy interrupts\n");
3953		− rc = pci_alloc_irq_vectors(dev, 1, 1, PCI_IRQ_LEGACY);
3954		− }
3955		− if (rc < 0) {
3956		− kfree(priv);
3957		− priv = ERR_PTR(rc);
3958		− goto err_deinit;
	3964	+ if (pci_match_id(pci_use_msi, dev)) {
	3965	+ dev_dbg(&dev->dev, "Using MSI(-X) interrupts\n");
	3966	+ pci_set_master(dev);
	3967	+ rc = pci_alloc_irq_vectors(dev, 1, 1, PCI_IRQ_ALL_TYPES);
	3968	+ } else {
	3969	+ dev_dbg(&dev->dev, "Using legacy interrupts\n");
	3970	+ rc = pci_alloc_irq_vectors(dev, 1, 1, PCI_IRQ_LEGACY);
	3971	+ }
	3972	+ if (rc < 0) {
	3973	+ kfree(priv);
	3974	+ priv = ERR_PTR(rc);
	3975	+ goto err_deinit;
	3976	+ }
	3977	+
	3978	+ uart.port.irq = pci_irq_vector(dev, 0);
3959	3979	}
3960	3980
3961		− uart.port.irq = pci_irq_vector(dev, 0);
3962	3981	uart.port.dev = &dev->dev;
3963	3982
3964	3983	for (i = 0; i < nr_ports; i++) {
@@ -4973,6 +4992,10 @@ static const struct pci_device_id serial_pci_tbl[] = { · 1008 unmodified lines
4973	4992	{ PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_DIVA_AUX,
4974	4993	PCI_ANY_ID, PCI_ANY_ID, 0, 0,
4975	4994	pbn_b2_1_115200 },
	4995	+ /* HPE PCI serial device */
	4996	+ { PCI_VENDOR_ID_HP_3PAR, PCI_DEVICE_ID_HPE_PCI_SERIAL,
	4997	+ PCI_ANY_ID, PCI_ANY_ID, 0, 0,
	4998	+ pbn_b1_1_115200 },
4976	4999
4977	5000	{ PCI_VENDOR_ID_DCI, PCI_DEVICE_ID_DCI_PCCOM2,
4978	5001	PCI_ANY_ID, PCI_ANY_ID, 0, 0,

drivers/tty/serial/8250/8250_port.c+12−0

@@ -1809,6 +1809,18 @@ void serial8250_tx_chars(struct uart_8250_port *up) · 1808 unmodified lines
1809	1809	count = up->tx_loadsz;
1810	1810	do {
1811	1811	serial_out(up, UART_TX, xmit->buf[xmit->tail]);
	1812	+ if (up->bugs & UART_BUG_TXRACE) {
	1813	+ /*
	1814	+ * The Aspeed BMC virtual UARTs have a bug where data
	1815	+ * may get stuck in the BMC's Tx FIFO from bursts of
	1816	+ * writes on the APB interface.
	1817	+ *
	1818	+ * Delay back-to-back writes by a read cycle to avoid
	1819	+ * stalling the VUART. Read a register that won't have
	1820	+ * side-effects and discard the result.
	1821	+ */
	1822	+ serial_in(up, UART_SCR);
	1823	+ }
1812	1824	xmit->tail = (xmit->tail + 1) & (UART_XMIT_SIZE - 1);
1813	1825	port->icount.tx++;
1814	1826	if (uart_circ_empty(xmit))

drivers/tty/serial/rp2.c+17−35

@@ -195,7 +195,6 @@ struct rp2_card { · 194 unmodified lines
195	195	void __iomem *bar0;
196	196	void __iomem *bar1;
197	197	spinlock_t card_lock;
198		− struct completion fw_loaded;
199	198	};
200	199
201	200	#define RP_ID(prod) PCI_VDEVICE(RP, (prod))
@@ -662,17 +661,10 @@ static void rp2_remove_ports(struct rp2_card *card) · 460 unmodified lines
662	661	card->initialized_ports = 0;
663	662	}
664	663
665		−static void rp2_fw_cb(const struct firmware fw, void context)
	664	+static int rp2_load_firmware(struct rp2_card card, const struct firmware fw)
666	665	{
667		− struct rp2_card *card = context;
668	666	resource_size_t phys_base;
669		− int i, rc = -ENOENT;
670		−
671		− if (!fw) {
672		− dev_err(&card->pdev->dev, "cannot find '%s' firmware image\n",
673		− RP2_FW_NAME);
674		− goto no_fw;
675		− }
	667	+ int i, rc = 0;
676	668
677	669	phys_base = pci_resource_start(card->pdev, 1);
678	670
@@ -718,23 +710,13 @@ static void rp2_fw_cb(const struct firmware fw, void context) · 39 unmodified lines
718	710	card->initialized_ports++;
719	711	}
720	712
721		− release_firmware(fw);
722		−no_fw:
723		− /*
724		− * rp2_fw_cb() is called from a workqueue long after rp2_probe()
725		− * has already returned success. So if something failed here,
726		− * we'll just leave the now-dormant device in place until somebody
727		− * unbinds it.
728		− */
729		− if (rc)
730		− dev_warn(&card->pdev->dev, "driver initialization failed\n");
731		−
732		− complete(&card->fw_loaded);
	713	+ return rc;
733	714	}
734	715
735	716	static int rp2_probe(struct pci_dev *pdev,
736	717	const struct pci_device_id *id)
737	718	{
	719	+ const struct firmware *fw;
738	720	struct rp2_card *card;
739	721	struct rp2_uart_port *ports;
740	722	void __iomem * const *bars;
@@ -745,7 +727,6 @@ static int rp2_probe(struct pci_dev *pdev, · 4 unmodified lines
745	727	return -ENOMEM;
746	728	pci_set_drvdata(pdev, card);
747	729	spin_lock_init(&card->card_lock);
748		− init_completion(&card->fw_loaded);
749	730
750	731	rc = pcim_enable_device(pdev);
751	732	if (rc)
@@ -778,21 +759,23 @@ static int rp2_probe(struct pci_dev *pdev, · 26 unmodified lines
778	759	return -ENOMEM;
779	760	card->ports = ports;
780	761
781		− rc = devm_request_irq(&pdev->dev, pdev->irq, rp2_uart_interrupt,
782		− IRQF_SHARED, DRV_NAME, card);
783		− if (rc)
	762	+ rc = request_firmware(&fw, RP2_FW_NAME, &pdev->dev);
	763	+ if (rc < 0) {
	764	+ dev_err(&pdev->dev, "cannot find '%s' firmware image\n",
	765	+ RP2_FW_NAME);
784	766	return rc;
	767	+ }
785	768
786		− /*
787		− * Only catastrophic errors (e.g. ENOMEM) are reported here.
788		− * If the FW image is missing, we'll find out in rp2_fw_cb()
789		− * and print an error message.
790		− */
791		− rc = request_firmware_nowait(THIS_MODULE, 1, RP2_FW_NAME, &pdev->dev,
792		− GFP_KERNEL, card, rp2_fw_cb);
	769	+ rc = rp2_load_firmware(card, fw);
	770	+
	771	+ release_firmware(fw);
	772	+ if (rc < 0)
	773	+ return rc;
	774	+
	775	+ rc = devm_request_irq(&pdev->dev, pdev->irq, rp2_uart_interrupt,
	776	+ IRQF_SHARED, DRV_NAME, card);
793	777	if (rc)
794	778	return rc;
795		− dev_dbg(&pdev->dev, "waiting for firmware blob...\n");
796	779
797	780	return 0;
798	781	}
@@ -801,7 +784,6 @@ static void rp2_remove(struct pci_dev *pdev) · 2 unmodified lines
801	784	{
802	785	struct rp2_card *card = pci_get_drvdata(pdev);
803	786
804		− wait_for_completion(&card->fw_loaded);
805	787	rp2_remove_ports(card);
806	788	}
807	789

drivers/tty/serial/serial-tegra.c+1−1

@@ -338,7 +338,7 @@ static void tegra_uart_fifo_reset(struct tegra_uart_port *tup, u8 fcr_bits) · 337 unmodified lines
338	338
339	339	do {
340	340	lsr = tegra_uart_read(tup, UART_LSR);
341		− if ((lsr \| UART_LSR_TEMT) && !(lsr & UART_LSR_DR))
	341	+ if ((lsr & UART_LSR_TEMT) && !(lsr & UART_LSR_DR))
342	342	break;
343	343	udelay(1);
344	344	} while (--tmout);

drivers/tty/serial/serial_core.c+5−3

@@ -863,9 +863,11 @@ static int uart_set_info(struct tty_struct tty, struct tty_port port, · 862 unmodified lines
863	863	goto check_and_exit;
864	864	}
865	865
866		− retval = security_locked_down(LOCKDOWN_TIOCSSERIAL);
867		− if (retval && (change_irq \|\| change_port))
868		− goto exit;
	866	+ if (change_irq \|\| change_port) {
	867	+ retval = security_locked_down(LOCKDOWN_TIOCSSERIAL);
	868	+ if (retval)
	869	+ goto exit;
	870	+ }
869	871
870	872	/*
871	873	* Ask the low level driver to verify the settings.

drivers/tty/serial/sh-sci.c+2−2

@@ -1023,10 +1023,10 @@ static int scif_set_rtrg(struct uart_port *port, int rx_trig) · 1022 unmodified lines
1023	1023	{
1024	1024	unsigned int bits;
1025	1025
	1026	+ if (rx_trig >= port->fifosize)
	1027	+ rx_trig = port->fifosize - 1;
1026	1028	if (rx_trig < 1)
1027	1029	rx_trig = 1;
1028		− if (rx_trig >= port->fifosize)
1029		− rx_trig = port->fifosize;
1030	1030
1031	1031	/* HSCIF can be set to an arbitrary level. */
1032	1032	if (sci_getreg(port, HSRTRGR)->size) {

drivers/usb/cdns3/cdns3-gadget.c+6−6

@@ -2007,7 +2007,7 @@ static void cdns3_configure_dmult(struct cdns3_device *priv_dev, · 2006 unmodified lines
2007	2007	else
2008	2008	mask = BIT(priv_ep->num);
2009	2009
2010		− if (priv_ep->type != USB_ENDPOINT_XFER_ISOC) {
	2010	+ if (priv_ep->type != USB_ENDPOINT_XFER_ISOC && !priv_ep->dir) {
2011	2011	cdns3_set_register_bit(&regs->tdl_from_trb, mask);
2012	2012	cdns3_set_register_bit(&regs->tdl_beh, mask);
2013	2013	cdns3_set_register_bit(&regs->tdl_beh2, mask);
@@ -2046,15 +2046,13 @@ int cdns3_ep_config(struct cdns3_endpoint *priv_ep, bool enable) · 32 unmodified lines
2046	2046	case USB_ENDPOINT_XFER_INT:
2047	2047	ep_cfg = EP_CFG_EPTYPE(USB_ENDPOINT_XFER_INT);
2048	2048
2049		− if ((priv_dev->dev_ver == DEV_VER_V2 && !priv_ep->dir) \|\|
2050		− priv_dev->dev_ver > DEV_VER_V2)
	2049	+ if (priv_dev->dev_ver >= DEV_VER_V2 && !priv_ep->dir)
2051	2050	ep_cfg \|= EP_CFG_TDL_CHK;
2052	2051	break;
2053	2052	case USB_ENDPOINT_XFER_BULK:
2054	2053	ep_cfg = EP_CFG_EPTYPE(USB_ENDPOINT_XFER_BULK);
2055	2054
2056		− if ((priv_dev->dev_ver == DEV_VER_V2 && !priv_ep->dir) \|\|
2057		− priv_dev->dev_ver > DEV_VER_V2)
	2055	+ if (priv_dev->dev_ver >= DEV_VER_V2 && !priv_ep->dir)
2058	2056	ep_cfg \|= EP_CFG_TDL_CHK;
2059	2057	break;
2060	2058	default:
@@ -3268,8 +3266,10 @@ static int __cdns3_gadget_init(struct cdns *cdns) · 1207 unmodified lines
3268	3266	pm_runtime_get_sync(cdns->dev);
3269	3267
3270	3268	ret = cdns3_gadget_start(cdns);
3271		− if (ret)
	3269	+ if (ret) {
	3270	+ pm_runtime_put_sync(cdns->dev);
3272	3271	return ret;
	3272	+ }
3273	3273
3274	3274	/*
3275	3275	* Because interrupt line can be shared with other components in

drivers/usb/cdns3/cdnsp-gadget.c+7−7

@@ -422,17 +422,17 @@ unmap: · 421 unmodified lines
422	422	int cdnsp_ep_dequeue(struct cdnsp_ep pep, struct cdnsp_request preq)
423	423	{
424	424	struct cdnsp_device *pdev = pep->pdev;
425		− int ret;
	425	+ int ret_stop = 0;
	426	+ int ret_rem;
426	427
427	428	trace_cdnsp_request_dequeue(preq);
428	429
429		− if (GET_EP_CTX_STATE(pep->out_ctx) == EP_STATE_RUNNING) {
430		− ret = cdnsp_cmd_stop_ep(pdev, pep);
431		− if (ret)
432		− return ret;
433		− }
	430	+ if (GET_EP_CTX_STATE(pep->out_ctx) == EP_STATE_RUNNING)
	431	+ ret_stop = cdnsp_cmd_stop_ep(pdev, pep);
	432	+
	433	+ ret_rem = cdnsp_remove_request(pdev, preq, pep);
434	434
435		− return cdnsp_remove_request(pdev, preq, pep);
	435	+ return ret_rem ? ret_rem : ret_stop;
436	436	}
437	437
438	438	static void cdnsp_zero_in_ctx(struct cdnsp_device *pdev)

drivers/usb/cdns3/cdnsp-ring.c+4−3

@@ -1517,13 +1517,14 @@ irqreturn_t cdnsp_thread_irq_handler(int irq, void *data) · 1516 unmodified lines
1517	1517	{
1518	1518	struct cdnsp_device pdev = (struct cdnsp_device )data;
1519	1519	union cdnsp_trb *event_ring_deq;
	1520	+ unsigned long flags;
1520	1521	int counter = 0;
1521	1522
1522		− spin_lock(&pdev->lock);
	1523	+ spin_lock_irqsave(&pdev->lock, flags);
1523	1524
1524	1525	if (pdev->cdnsp_state & (CDNSP_STATE_HALTED \| CDNSP_STATE_DYING)) {
1525	1526	cdnsp_died(pdev);
1526		− spin_unlock(&pdev->lock);
	1527	+ spin_unlock_irqrestore(&pdev->lock, flags);
1527	1528	return IRQ_HANDLED;
1528	1529	}
1529	1530
@@ -1539,7 +1540,7 @@ irqreturn_t cdnsp_thread_irq_handler(int irq, void *data) · 9 unmodified lines
1539	1540
1540	1541	cdnsp_update_erst_dequeue(pdev, event_ring_deq, 1);
1541	1542
1542		− spin_unlock(&pdev->lock);
	1543	+ spin_unlock_irqrestore(&pdev->lock, flags);
1543	1544
1544	1545	return IRQ_HANDLED;
1545	1546	}

drivers/usb/chipidea/udc.c+1−0

@@ -2061,6 +2061,7 @@ static int udc_start(struct ci_hdrc *ci) · 2060 unmodified lines
2061	2061	ci->gadget.name = ci->platdata->name;
2062	2062	ci->gadget.otg_caps = otg_caps;
2063	2063	ci->gadget.sg_supported = 1;
	2064	+ ci->gadget.irq = ci->irq;
2064	2065
2065	2066	if (ci->platdata->flags & CI_HDRC_REQUIRES_ALIGNED_DMA)
2066	2067	ci->gadget.quirk_avoids_skb_reserve = 1;

drivers/usb/chipidea/usbmisc_imx.c+14−2

@@ -686,6 +686,16 @@ static int imx7d_charger_secondary_detection(struct imx_usbmisc_data *data) · 685 unmodified lines
686	686	int val;
687	687	unsigned long flags;
688	688
	689	+ /* Clear VDATSRCENB0 to disable VDP_SRC and IDM_SNK required by BC 1.2 spec */
	690	+ spin_lock_irqsave(&usbmisc->lock, flags);
	691	+ val = readl(usbmisc->base + MX7D_USB_OTG_PHY_CFG2);
	692	+ val &= ~MX7D_USB_OTG_PHY_CFG2_CHRG_VDATSRCENB0;
	693	+ writel(val, usbmisc->base + MX7D_USB_OTG_PHY_CFG2);
	694	+ spin_unlock_irqrestore(&usbmisc->lock, flags);
	695	+
	696	+ /* TVDMSRC_DIS */
	697	+ msleep(20);
	698	+
689	699	/* VDM_SRC is connected to D- and IDP_SINK is connected to D+ */
690	700	spin_lock_irqsave(&usbmisc->lock, flags);
691	701	val = readl(usbmisc->base + MX7D_USB_OTG_PHY_CFG2);
@@ -695,7 +705,8 @@ static int imx7d_charger_secondary_detection(struct imx_usbmisc_data *data) · 3 unmodified lines
695	705	usbmisc->base + MX7D_USB_OTG_PHY_CFG2);
696	706	spin_unlock_irqrestore(&usbmisc->lock, flags);
697	707
698		− usleep_range(1000, 2000);
	708	+ /* TVDMSRC_ON */
	709	+ msleep(40);
699	710
700	711	/*
701	712	* Per BC 1.2, check voltage of D+:
@@ -798,7 +809,8 @@ static int imx7d_charger_primary_detection(struct imx_usbmisc_data *data) · 96 unmodified lines
798	809	usbmisc->base + MX7D_USB_OTG_PHY_CFG2);
799	810	spin_unlock_irqrestore(&usbmisc->lock, flags);
800	811
801		− usleep_range(1000, 2000);
	812	+ /* TVDPSRC_ON */
	813	+ msleep(40);
802	814
803	815	/* Check if D- is less than VDAT_REF to determine an SDP per BC 1.2 */
804	816	val = readl(usbmisc->base + MX7D_USB_OTG_PHY_STATUS);

drivers/usb/core/devio.c+8−3

@@ -1218,7 +1218,12 @@ static int do_proc_bulk(struct usb_dev_state *ps, · 1217 unmodified lines
1218	1218	ret = usbfs_increase_memory_usage(len1 + sizeof(struct urb));
1219	1219	if (ret)
1220	1220	return ret;
1221		− tbuf = kmalloc(len1, GFP_KERNEL);
	1221	+
	1222	+ /*
	1223	+ * len1 can be almost arbitrarily large. Don't WARN if it's
	1224	+ * too big, just fail the request.
	1225	+ */
	1226	+ tbuf = kmalloc(len1, GFP_KERNEL \| __GFP_NOWARN);
1222	1227	if (!tbuf) {
1223	1228	ret = -ENOMEM;
1224	1229	goto done;
@@ -1696,7 +1701,7 @@ static int proc_do_submiturb(struct usb_dev_state ps, struct usbdevfs_urb uurb · 471 unmodified lines
1696	1701	if (num_sgs) {
1697	1702	as->urb->sg = kmalloc_array(num_sgs,
1698	1703	sizeof(struct scatterlist),
1699		− GFP_KERNEL);
	1704	+ GFP_KERNEL \| __GFP_NOWARN);
1700	1705	if (!as->urb->sg) {
1701	1706	ret = -ENOMEM;
1702	1707	goto error;
@@ -1731,7 +1736,7 @@ static int proc_do_submiturb(struct usb_dev_state ps, struct usbdevfs_urb uurb · 28 unmodified lines
1731	1736	(uurb_start - as->usbm->vm_start);
1732	1737	} else {
1733	1738	as->urb->transfer_buffer = kmalloc(uurb->buffer_length,
1734		− GFP_KERNEL);
	1739	+ GFP_KERNEL \| __GFP_NOWARN);
1735	1740	if (!as->urb->transfer_buffer) {
1736	1741	ret = -ENOMEM;
1737	1742	goto error;

drivers/usb/core/hub.c+7−0

@@ -41,6 +41,8 @@ · 40 unmodified lines
41	41	#define USB_VENDOR_GENESYS_LOGIC 0x05e3
42	42	#define USB_VENDOR_SMSC 0x0424
43	43	#define USB_PRODUCT_USB5534B 0x5534
	44	+#define USB_VENDOR_CYPRESS 0x04b4
	45	+#define USB_PRODUCT_CY7C65632 0x6570
44	46	#define HUB_QUIRK_CHECK_PORT_AUTOSUSPEND 0x01
45	47	#define HUB_QUIRK_DISABLE_AUTOSUSPEND 0x02
46	48
@@ -5697,6 +5699,11 @@ static const struct usb_device_id hub_id_table[] = { · 5650 unmodified lines
5697	5699	.idProduct = USB_PRODUCT_USB5534B,
5698	5700	.bInterfaceClass = USB_CLASS_HUB,
5699	5701	.driver_info = HUB_QUIRK_DISABLE_AUTOSUSPEND},
	5702	+ { .match_flags = USB_DEVICE_ID_MATCH_VENDOR
	5703	+ \| USB_DEVICE_ID_MATCH_PRODUCT,
	5704	+ .idVendor = USB_VENDOR_CYPRESS,
	5705	+ .idProduct = USB_PRODUCT_CY7C65632,
	5706	+ .driver_info = HUB_QUIRK_DISABLE_AUTOSUSPEND},
5700	5707	{ .match_flags = USB_DEVICE_ID_MATCH_VENDOR
5701	5708	\| USB_DEVICE_ID_MATCH_INT_CLASS,
5702	5709	.idVendor = USB_VENDOR_GENESYS_LOGIC,

drivers/usb/dwc3/core.c+1−7

@@ -1671,8 +1671,8 @@ static int dwc3_remove(struct platform_device *pdev) · 1670 unmodified lines
1671	1671
1672	1672	pm_runtime_get_sync(&pdev->dev);
1673	1673
1674		− dwc3_debugfs_exit(dwc);
1675	1674	dwc3_core_exit_mode(dwc);
	1675	+ dwc3_debugfs_exit(dwc);
1676	1676
1677	1677	dwc3_core_exit(dwc);
1678	1678	dwc3_ulpi_exit(dwc);
@@ -1690,11 +1690,6 @@ static int dwc3_remove(struct platform_device *pdev) · 11 unmodified lines
1690	1690	return 0;
1691	1691	}
1692	1692
1693		−static void dwc3_shutdown(struct platform_device *pdev)
1694		−{
1695		− dwc3_remove(pdev);
1696		−}
1697		−
1698	1693	#ifdef CONFIG_PM
1699	1694	static int dwc3_core_init_for_resume(struct dwc3 *dwc)
1700	1695	{
@@ -2012,7 +2007,6 @@ MODULE_DEVICE_TABLE(acpi, dwc3_acpi_match); · 311 unmodified lines
2012	2007	static struct platform_driver dwc3_driver = {
2013	2008	.probe = dwc3_probe,
2014	2009	.remove = dwc3_remove,
2015		− .shutdown = dwc3_shutdown,
2016	2010	.driver = {
2017	2011	.name = "dwc3",
2018	2012	.of_match_table = of_match_ptr(of_dwc3_match),

drivers/usb/dwc3/debug.h+3−0

@@ -413,9 +413,12 @@ static inline const char *dwc3_gadget_generic_cmd_status_string(int status) · 412 unmodified lines
413	413
414	414
415	415	#ifdef CONFIG_DEBUG_FS
	416	+extern void dwc3_debugfs_create_endpoint_dir(struct dwc3_ep *dep);
416	417	extern void dwc3_debugfs_init(struct dwc3 *d);
417	418	extern void dwc3_debugfs_exit(struct dwc3 *d);
418	419	#else
	420	+static inline void dwc3_debugfs_create_endpoint_dir(struct dwc3_ep *dep)
	421	+{ }
419	422	static inline void dwc3_debugfs_init(struct dwc3 *d)
420	423	{ }
421	424	static inline void dwc3_debugfs_exit(struct dwc3 *d)

drivers/usb/dwc3/debugfs.c+2−19

@@ -886,30 +886,14 @@ static void dwc3_debugfs_create_endpoint_files(struct dwc3_ep *dep, · 885 unmodified lines
886	886	}
887	887	}
888	888
889		−static void dwc3_debugfs_create_endpoint_dir(struct dwc3_ep *dep,
890		− struct dentry *parent)
	889	+void dwc3_debugfs_create_endpoint_dir(struct dwc3_ep *dep)
891	890	{
892	891	struct dentry *dir;
893	892
894		− dir = debugfs_create_dir(dep->name, parent);
	893	+ dir = debugfs_create_dir(dep->name, dep->dwc->root);
895	894	dwc3_debugfs_create_endpoint_files(dep, dir);
896	895	}
897	896
898		−static void dwc3_debugfs_create_endpoint_dirs(struct dwc3 *dwc,
899		− struct dentry *parent)
900		−{
901		− int i;
902		−
903		− for (i = 0; i < dwc->num_eps; i++) {
904		− struct dwc3_ep *dep = dwc->eps[i];
905		−
906		− if (!dep)
907		− continue;
908		−
909		− dwc3_debugfs_create_endpoint_dir(dep, parent);
910		− }
911		−}
912		−
913	897	void dwc3_debugfs_init(struct dwc3 *dwc)
914	898	{
915	899	struct dentry *root;
@@ -940,7 +924,6 @@ void dwc3_debugfs_init(struct dwc3 *dwc) · 24 unmodified lines
940	924	&dwc3_testmode_fops);
941	925	debugfs_create_file("link_state", 0644, root, dwc,
942	926	&dwc3_link_state_fops);
943		− dwc3_debugfs_create_endpoint_dirs(dwc, root);
944	927	}
945	928	}
946	929

drivers/usb/dwc3/dwc3-meson-g12a.c+10−3

@@ -651,7 +651,7 @@ static int dwc3_meson_g12a_setup_regmaps(struct dwc3_meson_g12a *priv, · 650 unmodified lines
651	651	return PTR_ERR(priv->usb_glue_regmap);
652	652
653	653	/* Create a regmap for each USB2 PHY control register set */
654		− for (i = 0; i < priv->usb2_ports; i++) {
	654	+ for (i = 0; i < priv->drvdata->num_phys; i++) {
655	655	struct regmap_config u2p_regmap_config = {
656	656	.reg_bits = 8,
657	657	.val_bits = 32,
@@ -659,6 +659,9 @@ static int dwc3_meson_g12a_setup_regmaps(struct dwc3_meson_g12a *priv, · 1 unmodified line
659	659	.max_register = U2P_R1,
660	660	};
661	661
	662	+ if (!strstr(priv->drvdata->phy_names[i], "usb2"))
	663	+ continue;
	664	+
662	665	u2p_regmap_config.name = devm_kasprintf(priv->dev, GFP_KERNEL,
663	666	"u2p-%d", i);
664	667	if (!u2p_regmap_config.name)
@@ -772,13 +775,13 @@ static int dwc3_meson_g12a_probe(struct platform_device *pdev) · 107 unmodified lines
772	775
773	776	ret = priv->drvdata->usb_init(priv);
774	777	if (ret)
775		− goto err_disable_clks;
	778	+ goto err_disable_regulator;
776	779
777	780	/* Init PHYs */
778	781	for (i = 0 ; i < PHY_COUNT ; ++i) {
779	782	ret = phy_init(priv->phys[i]);
780	783	if (ret)
781		− goto err_disable_clks;
	784	+ goto err_disable_regulator;
782	785	}
783	786
784	787	/* Set PHY Power */
@@ -816,6 +819,10 @@ err_phys_exit: · 31 unmodified lines
816	819	for (i = 0 ; i < PHY_COUNT ; ++i)
817	820	phy_exit(priv->phys[i]);
818	821
	822	+err_disable_regulator:
	823	+ if (priv->vbus)
	824	+ regulator_disable(priv->vbus);
	825	+
819	826	err_disable_clks:
820	827	clk_bulk_disable_unprepare(priv->drvdata->num_clks,
821	828	priv->drvdata->clks);

drivers/usb/dwc3/ep0.c+3−0

@@ -292,6 +292,9 @@ static struct dwc3_ep dwc3_wIndex_to_dep(struct dwc3 dwc, __le16 wIndex_le) · 291 unmodified lines
292	292	epnum \|= 1;
293	293
294	294	dep = dwc->eps[epnum];
	295	+ if (dep == NULL)
	296	+ return NULL;
	297	+
295	298	if (dep->flags & DWC3_EP_ENABLED)
296	299	return dep;
297	300

drivers/usb/dwc3/gadget.c+19−12

@@ -1244,6 +1244,7 @@ static int dwc3_prepare_trbs_sg(struct dwc3_ep *dep, · 1243 unmodified lines
1244	1244	req->start_sg = sg_next(s);
1245	1245
1246	1246	req->num_queued_sgs++;
	1247	+ req->num_pending_sgs--;
1247	1248
1248	1249	/*
1249	1250	* The number of pending SG entries may not correspond to the
@@ -1251,7 +1252,7 @@ static int dwc3_prepare_trbs_sg(struct dwc3_ep *dep, · 1 unmodified line
1251	1252	* don't include unused SG entries.
1252	1253	*/
1253	1254	if (length == 0) {
1254		− req->num_pending_sgs -= req->request.num_mapped_sgs - req->num_queued_sgs;
	1255	+ req->num_pending_sgs = 0;
1255	1256	break;
1256	1257	}
1257	1258
@@ -2260,13 +2261,10 @@ static int dwc3_gadget_pullup(struct usb_gadget *g, int is_on) · 1002 unmodified lines
2260	2261	}
2261	2262
2262	2263	/*
2263		− * Synchronize any pending event handling before executing the controller
2264		− * halt routine.
	2264	+ * Synchronize and disable any further event handling while controller
	2265	+ * is being enabled/disabled.
2265	2266	*/
2266		− if (!is_on) {
2267		− dwc3_gadget_disable_irq(dwc);
2268		− synchronize_irq(dwc->irq_gadget);
2269		− }
	2267	+ disable_irq(dwc->irq_gadget);
2270	2268
2271	2269	spin_lock_irqsave(&dwc->lock, flags);
2272	2270
@@ -2304,6 +2302,8 @@ static int dwc3_gadget_pullup(struct usb_gadget *g, int is_on) · 31 unmodified lines
2304	2302
2305	2303	ret = dwc3_gadget_run_stop(dwc, is_on, false);
2306	2304	spin_unlock_irqrestore(&dwc->lock, flags);
	2305	+ enable_irq(dwc->irq_gadget);
	2306	+
2307	2307	pm_runtime_put(dwc->dev);
2308	2308
2309	2309	return ret;
@@ -2753,6 +2753,8 @@ static int dwc3_gadget_init_endpoint(struct dwc3 *dwc, u8 epnum) · 443 unmodified lines
2753	2753	INIT_LIST_HEAD(&dep->started_list);
2754	2754	INIT_LIST_HEAD(&dep->cancelled_list);
2755	2755
	2756	+ dwc3_debugfs_create_endpoint_dir(dep);
	2757	+
2756	2758	return 0;
2757	2759	}
2758	2760
@@ -2796,6 +2798,7 @@ static void dwc3_gadget_free_endpoints(struct dwc3 *dwc) · 37 unmodified lines
2796	2798	list_del(&dep->endpoint.ep_list);
2797	2799	}
2798	2800
	2801	+ debugfs_remove_recursive(debugfs_lookup(dep->name, dwc->root));
2799	2802	kfree(dep);
2800	2803	}
2801	2804	}
@@ -2873,15 +2876,15 @@ static int dwc3_gadget_ep_reclaim_trb_sg(struct dwc3_ep *dep, · 71 unmodified lines
2873	2876	struct dwc3_trb *trb = &dep->trb_pool[dep->trb_dequeue];
2874	2877	struct scatterlist *sg = req->sg;
2875	2878	struct scatterlist *s;
2876		− unsigned int pending = req->num_pending_sgs;
	2879	+ unsigned int num_queued = req->num_queued_sgs;
2877	2880	unsigned int i;
2878	2881	int ret = 0;
2879	2882
2880		− for_each_sg(sg, s, pending, i) {
	2883	+ for_each_sg(sg, s, num_queued, i) {
2881	2884	trb = &dep->trb_pool[dep->trb_dequeue];
2882	2885
2883	2886	req->sg = sg_next(s);
2884		− req->num_pending_sgs--;
	2887	+ req->num_queued_sgs--;
2885	2888
2886	2889	ret = dwc3_gadget_ep_reclaim_completed_trb(dep, req,
2887	2890	trb, event, status, true);
@@ -2904,7 +2907,7 @@ static int dwc3_gadget_ep_reclaim_trb_linear(struct dwc3_ep *dep, · 16 unmodified lines
2904	2907
2905	2908	static bool dwc3_gadget_ep_request_completed(struct dwc3_request *req)
2906	2909	{
2907		− return req->num_pending_sgs == 0;
	2910	+ return req->num_pending_sgs == 0 && req->num_queued_sgs == 0;
2908	2911	}
2909	2912
2910	2913	static int dwc3_gadget_ep_cleanup_completed_request(struct dwc3_ep *dep,
@@ -2913,7 +2916,7 @@ static int dwc3_gadget_ep_cleanup_completed_request(struct dwc3_ep *dep, · 2 unmodified lines
2913	2916	{
2914	2917	int ret;
2915	2918
2916		− if (req->num_pending_sgs)
	2919	+ if (req->request.num_mapped_sgs)
2917	2920	ret = dwc3_gadget_ep_reclaim_trb_sg(dep, req, event,
2918	2921	status);
2919	2922	else
@@ -4045,6 +4048,7 @@ err5: · 1125 unmodified lines
4045	4048	dwc3_gadget_free_endpoints(dwc);
4046	4049	err4:
4047	4050	usb_put_gadget(dwc->gadget);
	4051	+ dwc->gadget = NULL;
4048	4052	err3:
4049	4053	dma_free_coherent(dwc->sysdev, DWC3_BOUNCE_SIZE, dwc->bounce,
4050	4054	dwc->bounce_addr);
@@ -4064,6 +4068,9 @@ err0: · 13 unmodified lines
4064	4068
4065	4069	void dwc3_gadget_exit(struct dwc3 *dwc)
4066	4070	{
	4071	+ if (!dwc->gadget)
	4072	+ return;
	4073	+
4067	4074	usb_del_gadget(dwc->gadget);
4068	4075	dwc3_gadget_free_endpoints(dwc);
4069	4076	usb_put_gadget(dwc->gadget);

drivers/usb/gadget/config.c+8−0

@@ -164,6 +164,14 @@ int usb_assign_descriptors(struct usb_function *f, · 163 unmodified lines
164	164	{
165	165	struct usb_gadget *g = f->config->cdev->gadget;
166	166
	167	+ /* super-speed-plus descriptor falls back to super-speed one,
	168	+ * if such a descriptor was provided, thus avoiding a NULL
	169	+ * pointer dereference if a 5gbps capable gadget is used with
	170	+ * a 10gbps capable config (device port + cable + host port)
	171	+ */
	172	+ if (!ssp)
	173	+ ssp = ss;
	174	+
167	175	if (fs) {
168	176	f->fs_descriptors = usb_copy_descriptors(fs);
169	177	if (!f->fs_descriptors)

drivers/usb/gadget/function/f_ecm.c+1−1

@@ -791,7 +791,7 @@ ecm_bind(struct usb_configuration c, struct usb_function f) · 790 unmodified lines
791	791	fs_ecm_notify_desc.bEndpointAddress;
792	792
793	793	status = usb_assign_descriptors(f, ecm_fs_function, ecm_hs_function,
794		− ecm_ss_function, NULL);
	794	+ ecm_ss_function, ecm_ss_function);
795	795	if (status)
796	796	goto fail;
797	797

drivers/usb/gadget/function/f_eem.c+3−3

@@ -302,7 +302,7 @@ static int eem_bind(struct usb_configuration c, struct usb_function f) · 301 unmodified lines
302	302	eem_ss_out_desc.bEndpointAddress = eem_fs_out_desc.bEndpointAddress;
303	303
304	304	status = usb_assign_descriptors(f, eem_fs_function, eem_hs_function,
305		− eem_ss_function, NULL);
	305	+ eem_ss_function, eem_ss_function);
306	306	if (status)
307	307	goto fail;
308	308
@@ -495,7 +495,7 @@ static int eem_unwrap(struct gether *port, · 186 unmodified lines
495	495	skb2 = skb_clone(skb, GFP_ATOMIC);
496	496	if (unlikely(!skb2)) {
497	497	DBG(cdev, "unable to unframe EEM packet\n");
498		− continue;
	498	+ goto next;
499	499	}
500	500	skb_trim(skb2, len - ETH_FCS_LEN);
501	501
@@ -505,7 +505,7 @@ static int eem_unwrap(struct gether *port, · 3 unmodified lines
505	505	GFP_ATOMIC);
506	506	if (unlikely(!skb3)) {
507	507	dev_kfree_skb_any(skb2);
508		− continue;
	508	+ goto next;
509	509	}
510	510	dev_kfree_skb_any(skb2);
511	511	skb_queue_tail(list, skb3);

drivers/usb/gadget/function/f_fs.c+3−0

@@ -3567,6 +3567,9 @@ static void ffs_func_unbind(struct usb_configuration *c, · 3566 unmodified lines
3567	3567	ffs->func = NULL;
3568	3568	}
3569	3569
	3570	+ /* Drain any pending AIO completions */
	3571	+ drain_workqueue(ffs->io_completion_wq);
	3572	+
3570	3573	if (!--opts->refcnt)
3571	3574	functionfs_unbind(ffs);
3572	3575

drivers/usb/gadget/function/f_hid.c+2−1

@@ -802,7 +802,8 @@ static int hidg_bind(struct usb_configuration c, struct usb_function f) · 801 unmodified lines
802	802	hidg_fs_out_ep_desc.bEndpointAddress;
803	803
804	804	status = usb_assign_descriptors(f, hidg_fs_descriptors,
805		− hidg_hs_descriptors, hidg_ss_descriptors, NULL);
	805	+ hidg_hs_descriptors, hidg_ss_descriptors,
	806	+ hidg_ss_descriptors);
806	807	if (status)
807	808	goto fail;
808	809

drivers/usb/gadget/function/f_loopback.c+1−1

@@ -207,7 +207,7 @@ autoconf_fail: · 206 unmodified lines
207	207	ss_loop_sink_desc.bEndpointAddress = fs_loop_sink_desc.bEndpointAddress;
208	208
209	209	ret = usb_assign_descriptors(f, fs_loopback_descs, hs_loopback_descs,
210		− ss_loopback_descs, NULL);
	210	+ ss_loopback_descs, ss_loopback_descs);
211	211	if (ret)
212	212	return ret;
213	213

drivers/usb/gadget/function/f_ncm.c+5−5

@@ -583,7 +583,7 @@ static void ncm_do_notify(struct f_ncm *ncm) · 582 unmodified lines
583	583	data[0] = cpu_to_le32(ncm_bitrate(cdev->gadget));
584	584	data[1] = data[0];
585	585
586		− DBG(cdev, "notify speed %d\n", ncm_bitrate(cdev->gadget));
	586	+ DBG(cdev, "notify speed %u\n", ncm_bitrate(cdev->gadget));
587	587	ncm->notify_state = NCM_NOTIFY_CONNECT;
588	588	break;
589	589	}
@@ -1101,11 +1101,11 @@ static struct sk_buff ncm_wrap_ntb(struct gether port, · 511 unmodified lines
1101	1101	ncm->ndp_dgram_count = 1;
1102	1102
1103	1103	/* Note: we skip opts->next_ndp_index */
1104		− }
1105	1104
1106		− /* Delay the timer. */
1107		− hrtimer_start(&ncm->task_timer, TX_TIMEOUT_NSECS,
1108		− HRTIMER_MODE_REL_SOFT);
	1105	+ /* Start the timer. */
	1106	+ hrtimer_start(&ncm->task_timer, TX_TIMEOUT_NSECS,
	1107	+ HRTIMER_MODE_REL_SOFT);
	1108	+ }
1109	1109
1110	1110	/* Add the datagram position entries */
1111	1111	ntb_ndp = skb_put_zero(ncm->skb_tx_ndp, dgram_idx_len);

drivers/usb/gadget/function/f_printer.c+2−1

@@ -1101,7 +1101,8 @@ autoconf_fail: · 1100 unmodified lines
1101	1101	ss_ep_out_desc.bEndpointAddress = fs_ep_out_desc.bEndpointAddress;
1102	1102
1103	1103	ret = usb_assign_descriptors(f, fs_printer_function,
1104		− hs_printer_function, ss_printer_function, NULL);
	1104	+ hs_printer_function, ss_printer_function,
	1105	+ ss_printer_function);
1105	1106	if (ret)
1106	1107	return ret;
1107	1108

drivers/usb/gadget/function/f_rndis.c+1−1

@@ -789,7 +789,7 @@ rndis_bind(struct usb_configuration c, struct usb_function f) · 788 unmodified lines
789	789	ss_notify_desc.bEndpointAddress = fs_notify_desc.bEndpointAddress;
790	790
791	791	status = usb_assign_descriptors(f, eth_fs_function, eth_hs_function,
792		− eth_ss_function, NULL);
	792	+ eth_ss_function, eth_ss_function);
793	793	if (status)
794	794	goto fail;
795	795

drivers/usb/gadget/function/f_serial.c+1−1

@@ -233,7 +233,7 @@ static int gser_bind(struct usb_configuration c, struct usb_function f) · 232 unmodified lines
233	233	gser_ss_out_desc.bEndpointAddress = gser_fs_out_desc.bEndpointAddress;
234	234
235	235	status = usb_assign_descriptors(f, gser_fs_function, gser_hs_function,
236		− gser_ss_function, NULL);
	236	+ gser_ss_function, gser_ss_function);
237	237	if (status)
238	238	goto fail;
239	239	dev_dbg(&cdev->gadget->dev, "generic ttyGS%d: %s speed IN/%s OUT/%s\n",

drivers/usb/gadget/function/f_sourcesink.c+2−1

@@ -431,7 +431,8 @@ no_iso: · 430 unmodified lines
431	431	ss_iso_sink_desc.bEndpointAddress = fs_iso_sink_desc.bEndpointAddress;
432	432
433	433	ret = usb_assign_descriptors(f, fs_source_sink_descs,
434		− hs_source_sink_descs, ss_source_sink_descs, NULL);
	434	+ hs_source_sink_descs, ss_source_sink_descs,
	435	+ ss_source_sink_descs);
435	436	if (ret)
436	437	return ret;
437	438

drivers/usb/gadget/function/f_subset.c+1−1

@@ -358,7 +358,7 @@ geth_bind(struct usb_configuration c, struct usb_function f) · 357 unmodified lines
358	358	fs_subset_out_desc.bEndpointAddress;
359	359
360	360	status = usb_assign_descriptors(f, fs_eth_function, hs_eth_function,
361		− ss_eth_function, NULL);
	361	+ ss_eth_function, ss_eth_function);
362	362	if (status)
363	363	goto fail;
364	364

drivers/usb/gadget/function/f_tcm.c+2−1

@@ -2057,7 +2057,8 @@ static int tcm_bind(struct usb_configuration c, struct usb_function f) · 2056 unmodified lines
2057	2057	uasp_fs_cmd_desc.bEndpointAddress = uasp_ss_cmd_desc.bEndpointAddress;
2058	2058
2059	2059	ret = usb_assign_descriptors(f, uasp_fs_function_desc,
2060		− uasp_hs_function_desc, uasp_ss_function_desc, NULL);
	2060	+ uasp_hs_function_desc, uasp_ss_function_desc,
	2061	+ uasp_ss_function_desc);
2061	2062	if (ret)
2062	2063	goto ep_fail;
2063	2064

drivers/usb/gadget/udc/renesas_usb3.c+3−2

@@ -1488,7 +1488,7 @@ static void usb3_start_pipen(struct renesas_usb3_ep *usb3_ep, · 1487 unmodified lines
1488	1488	struct renesas_usb3_request *usb3_req)
1489	1489	{
1490	1490	struct renesas_usb3 *usb3 = usb3_ep_to_usb3(usb3_ep);
1491		− struct renesas_usb3_request *usb3_req_first = usb3_get_request(usb3_ep);
	1491	+ struct renesas_usb3_request *usb3_req_first;
1492	1492	unsigned long flags;
1493	1493	int ret = -EAGAIN;
1494	1494	u32 enable_bits = 0;
@@ -1496,7 +1496,8 @@ static void usb3_start_pipen(struct renesas_usb3_ep *usb3_ep, · 1 unmodified line
1496	1496	spin_lock_irqsave(&usb3->lock, flags);
1497	1497	if (usb3_ep->halt \|\| usb3_ep->started)
1498	1498	goto out;
1499		− if (usb3_req != usb3_req_first)
	1499	+ usb3_req_first = __usb3_get_request(usb3_ep);
	1500	+ if (!usb3_req_first \|\| usb3_req != usb3_req_first)
1500	1501	goto out;
1501	1502
1502	1503	if (usb3_pn_change(usb3, usb3_ep->num) < 0)

drivers/usb/host/xhci-pci.c+6−1

@@ -59,6 +59,7 @@ · 58 unmodified lines
59	59	#define PCI_DEVICE_ID_INTEL_MAPLE_RIDGE_XHCI 0x1138
60	60	#define PCI_DEVICE_ID_INTEL_ALDER_LAKE_XHCI 0x461e
61	61
	62	+#define PCI_DEVICE_ID_AMD_RENOIR_XHCI 0x1639
62	63	#define PCI_DEVICE_ID_AMD_PROMONTORYA_4 0x43b9
63	64	#define PCI_DEVICE_ID_AMD_PROMONTORYA_3 0x43ba
64	65	#define PCI_DEVICE_ID_AMD_PROMONTORYA_2 0x43bb
@@ -182,6 +183,10 @@ static void xhci_pci_quirks(struct device dev, struct xhci_hcd xhci) · 117 unmodified lines
182	183	(pdev->device == PCI_DEVICE_ID_AMD_PROMONTORYA_1)))
183	184	xhci->quirks \|= XHCI_U2_DISABLE_WAKE;
184	185
	186	+ if (pdev->vendor == PCI_VENDOR_ID_AMD &&
	187	+ pdev->device == PCI_DEVICE_ID_AMD_RENOIR_XHCI)
	188	+ xhci->quirks \|= XHCI_BROKEN_D3COLD;
	189	+
185	190	if (pdev->vendor == PCI_VENDOR_ID_INTEL) {
186	191	xhci->quirks \|= XHCI_LPM_SUPPORT;
187	192	xhci->quirks \|= XHCI_INTEL_HOST;
@@ -539,7 +544,7 @@ static int xhci_pci_suspend(struct usb_hcd *hcd, bool do_wakeup) · 351 unmodified lines
539	544	* Systems with the TI redriver that loses port status change events
540	545	* need to have the registers polled during D3, so avoid D3cold.
541	546	*/
542		− if (xhci->quirks & XHCI_COMP_MODE_QUIRK)
	547	+ if (xhci->quirks & (XHCI_COMP_MODE_QUIRK \| XHCI_BROKEN_D3COLD))
543	548	pci_d3cold_disable(pdev);
544	549
545	550	if (xhci->quirks & XHCI_PME_STUCK_QUIRK)

drivers/usb/host/xhci-ring.c+7−7

@@ -828,14 +828,10 @@ static void xhci_giveback_invalidated_tds(struct xhci_virt_ep *ep) · 827 unmodified lines
828	828	list_for_each_entry_safe(td, tmp_td, &ep->cancelled_td_list,
829	829	cancelled_td_list) {
830	830
831		− /*
832		− * Doesn't matter what we pass for status, since the core will
833		− * just overwrite it (because the URB has been unlinked).
834		− */
835	831	ring = xhci_urb_to_transfer_ring(ep->xhci, td->urb);
836	832
837	833	if (td->cancel_status == TD_CLEARED)
838		− xhci_td_cleanup(ep->xhci, td, ring, 0);
	834	+ xhci_td_cleanup(ep->xhci, td, ring, td->status);
839	835
840	836	if (ep->xhci->xhc_state & XHCI_STATE_DYING)
841	837	return;
@@ -937,14 +933,18 @@ static int xhci_invalidate_cancelled_tds(struct xhci_virt_ep *ep) · 95 unmodified lines
937	933	continue;
938	934	}
939	935	/*
940		− * If ring stopped on the TD we need to cancel, then we have to
	936	+ * If a ring stopped on the TD we need to cancel then we have to
941	937	* move the xHC endpoint ring dequeue pointer past this TD.
	938	+ * Rings halted due to STALL may show hw_deq is past the stalled
	939	+ * TD, but still require a set TR Deq command to flush xHC cache.
942	940	*/
943	941	hw_deq = xhci_get_hw_deq(xhci, ep->vdev, ep->ep_index,
944	942	td->urb->stream_id);
945	943	hw_deq &= ~0xf;
946	944
947		− if (trb_in_td(xhci, td->start_seg, td->first_trb,
	945	+ if (td->cancel_status == TD_HALTED) {
	946	+ cached_td = td;
	947	+ } else if (trb_in_td(xhci, td->start_seg, td->first_trb,
948	948	td->last_trb, hw_deq, false)) {
949	949	switch (td->cancel_status) {
950	950	case TD_CLEARED: /* TD is already no-op */

drivers/usb/host/xhci.h+1−0

@@ -1892,6 +1892,7 @@ struct xhci_hcd { · 1891 unmodified lines
1892	1892	#define XHCI_DISABLE_SPARSE BIT_ULL(38)
1893	1893	#define XHCI_SG_TRB_CACHE_SIZE_QUIRK BIT_ULL(39)
1894	1894	#define XHCI_NO_SOFT_RETRY BIT_ULL(40)
	1895	+#define XHCI_BROKEN_D3COLD BIT_ULL(41)
1895	1896
1896	1897	unsigned int num_active_eps;
1897	1898	unsigned int limit_active_eps;

drivers/usb/misc/brcmstb-usb-pinmap.c+2−0

@@ -263,6 +263,8 @@ static int __init brcmstb_usb_pinmap_probe(struct platform_device *pdev) · 262 unmodified lines
263	263	return -EINVAL;
264	264
265	265	r = platform_get_resource(pdev, IORESOURCE_MEM, 0);
	266	+ if (!r)
	267	+ return -EINVAL;
266	268
267	269	pdata = devm_kzalloc(&pdev->dev,
268	270	sizeof(*pdata) +

drivers/usb/misc/trancevibrator.c+2−2

@@ -61,9 +61,9 @@ static ssize_t speed_store(struct device dev, struct device_attribute attr, · 60 unmodified lines
61	61	/* Set speed */
62	62	retval = usb_control_msg(tv->udev, usb_sndctrlpipe(tv->udev, 0),
63	63	0x01, /* vendor request: set speed */
64		− USB_DIR_IN \| USB_TYPE_VENDOR \| USB_RECIP_OTHER,
	64	+ USB_DIR_OUT \| USB_TYPE_VENDOR \| USB_RECIP_OTHER,
65	65	tv->speed, /* speed value */
66		− 0, NULL, 0, USB_CTRL_GET_TIMEOUT);
	66	+ 0, NULL, 0, USB_CTRL_SET_TIMEOUT);
67	67	if (retval) {
68	68	tv->speed = old;
69	69	dev_dbg(&tv->udev->dev, "retval = %d\n", retval);

drivers/usb/misc/uss720.c+1−0

@@ -736,6 +736,7 @@ static int uss720_probe(struct usb_interface *intf, · 735 unmodified lines
736	736	parport_announce_port(pp);
737	737
738	738	usb_set_intfdata(intf, pp);
	739	+ usb_put_dev(usbdev);
739	740	return 0;
740	741
741	742	probe_abort:

drivers/usb/musb/musb_core.c+1−2

@@ -2009,9 +2009,8 @@ static void musb_pm_runtime_check_session(struct musb *musb) · 2008 unmodified lines
2009	2009	schedule_delayed_work(&musb->irq_work,
2010	2010	msecs_to_jiffies(1000));
2011	2011	musb->quirk_retries--;
2012		− break;
2013	2012	}
2014		− fallthrough;
	2013	+ break;
2015	2014	case MUSB_QUIRK_B_INVALID_VBUS_91:
2016	2015	if (musb->quirk_retries && !musb->flush_irq_work) {
2017	2016	musb_dbg(musb,

drivers/usb/serial/cp210x.c+78−6

@@ -252,9 +252,11 @@ struct cp210x_serial_private { · 251 unmodified lines
252	252	u8 gpio_input;
253	253	#endif
254	254	u8 partnum;
	255	+ u32 fw_version;
255	256	speed_t min_speed;
256	257	speed_t max_speed;
257	258	bool use_actual_rate;
	259	+ bool no_flow_control;
258	260	};
259	261
260	262	enum cp210x_event_state {
@@ -398,6 +400,7 @@ struct cp210x_special_chars { · 137 unmodified lines
398	400
399	401	/* CP210X_VENDOR_SPECIFIC values */
400	402	#define CP210X_READ_2NCONFIG 0x000E
	403	+#define CP210X_GET_FW_VER_2N 0x0010
401	404	#define CP210X_READ_LATCH 0x00C2
402	405	#define CP210X_GET_PARTNUM 0x370B
403	406	#define CP210X_GET_PORTCONFIG 0x370C
@@ -537,6 +540,12 @@ struct cp210x_single_port_config { · 133 unmodified lines
537	540	#define CP210X_2NCONFIG_GPIO_RSTLATCH_IDX 587
538	541	#define CP210X_2NCONFIG_GPIO_CONTROL_IDX 600
539	542
	543	+/* CP2102N QFN20 port configuration values */
	544	+#define CP2102N_QFN20_GPIO2_TXLED_MODE BIT(2)
	545	+#define CP2102N_QFN20_GPIO3_RXLED_MODE BIT(3)
	546	+#define CP2102N_QFN20_GPIO1_RS485_MODE BIT(4)
	547	+#define CP2102N_QFN20_GPIO0_CLK_MODE BIT(6)
	548	+
540	549	/* CP210X_VENDOR_SPECIFIC, CP210X_WRITE_LATCH call writes these 0x2 bytes. */
541	550	struct cp210x_gpio_write {
542	551	u8 mask;
@@ -1122,6 +1131,7 @@ static bool cp210x_termios_change(const struct ktermios *a, const struct ktermio · 579 unmodified lines
1122	1131	static void cp210x_set_flow_control(struct tty_struct *tty,
1123	1132	struct usb_serial_port port, struct ktermios old_termios)
1124	1133	{
	1134	+ struct cp210x_serial_private *priv = usb_get_serial_data(port->serial);
1125	1135	struct cp210x_port_private *port_priv = usb_get_serial_port_data(port);
1126	1136	struct cp210x_special_chars chars;
1127	1137	struct cp210x_flow_ctl flow_ctl;
@@ -1129,6 +1139,15 @@ static void cp210x_set_flow_control(struct tty_struct *tty, · 1 unmodified line
1129	1139	u32 ctl_hs;
1130	1140	int ret;
1131	1141
	1142	+ /*
	1143	+ * Some CP2102N interpret ulXonLimit as ulFlowReplace (erratum
	1144	+ * CP2102N_E104). Report back that flow control is not supported.
	1145	+ */
	1146	+ if (priv->no_flow_control) {
	1147	+ tty->termios.c_cflag &= ~CRTSCTS;
	1148	+ tty->termios.c_iflag &= ~(IXON \| IXOFF);
	1149	+ }
	1150	+
1132	1151	if (old_termios &&
1133	1152	C_CRTSCTS(tty) == (old_termios->c_cflag & CRTSCTS) &&
1134	1153	I_IXON(tty) == (old_termios->c_iflag & IXON) &&
@@ -1185,19 +1204,20 @@ static void cp210x_set_flow_control(struct tty_struct *tty, · 50 unmodified lines
1185	1204	port_priv->crtscts = false;
1186	1205	}
1187	1206
1188		− if (I_IXOFF(tty))
	1207	+ if (I_IXOFF(tty)) {
1189	1208	flow_repl \|= CP210X_SERIAL_AUTO_RECEIVE;
1190		− else
	1209	+
	1210	+ flow_ctl.ulXonLimit = cpu_to_le32(128);
	1211	+ flow_ctl.ulXoffLimit = cpu_to_le32(128);
	1212	+ } else {
1191	1213	flow_repl &= ~CP210X_SERIAL_AUTO_RECEIVE;
	1214	+ }
1192	1215
1193	1216	if (I_IXON(tty))
1194	1217	flow_repl \|= CP210X_SERIAL_AUTO_TRANSMIT;
1195	1218	else
1196	1219	flow_repl &= ~CP210X_SERIAL_AUTO_TRANSMIT;
1197	1220
1198		− flow_ctl.ulXonLimit = cpu_to_le32(128);
1199		− flow_ctl.ulXoffLimit = cpu_to_le32(128);
1200		−
1201	1221	dev_dbg(&port->dev, "%s - ctrl = 0x%02x, flow = 0x%02x\n", __func__,
1202	1222	ctl_hs, flow_repl);
1203	1223
@@ -1733,7 +1753,19 @@ static int cp2102n_gpioconf_init(struct usb_serial *serial) · 529 unmodified lines
1733	1753	priv->gpio_pushpull = (gpio_pushpull >> 3) & 0x0f;
1734	1754
1735	1755	/* 0 indicates GPIO mode, 1 is alternate function */
1736		− priv->gpio_altfunc = (gpio_ctrl >> 2) & 0x0f;
	1756	+ if (priv->partnum == CP210X_PARTNUM_CP2102N_QFN20) {
	1757	+ /* QFN20 is special... */
	1758	+ if (gpio_ctrl & CP2102N_QFN20_GPIO0_CLK_MODE) /* GPIO 0 */
	1759	+ priv->gpio_altfunc \|= BIT(0);
	1760	+ if (gpio_ctrl & CP2102N_QFN20_GPIO1_RS485_MODE) /* GPIO 1 */
	1761	+ priv->gpio_altfunc \|= BIT(1);
	1762	+ if (gpio_ctrl & CP2102N_QFN20_GPIO2_TXLED_MODE) /* GPIO 2 */
	1763	+ priv->gpio_altfunc \|= BIT(2);
	1764	+ if (gpio_ctrl & CP2102N_QFN20_GPIO3_RXLED_MODE) /* GPIO 3 */
	1765	+ priv->gpio_altfunc \|= BIT(3);
	1766	+ } else {
	1767	+ priv->gpio_altfunc = (gpio_ctrl >> 2) & 0x0f;
	1768	+ }
1737	1769
1738	1770	if (priv->partnum == CP210X_PARTNUM_CP2102N_QFN28) {
1739	1771	/*
@@ -1908,6 +1940,45 @@ static void cp210x_init_max_speed(struct usb_serial *serial) · 168 unmodified lines
1908	1940	priv->use_actual_rate = use_actual_rate;
1909	1941	}
1910	1942
	1943	+static int cp210x_get_fw_version(struct usb_serial *serial, u16 value)
	1944	+{
	1945	+ struct cp210x_serial_private *priv = usb_get_serial_data(serial);
	1946	+ u8 ver[3];
	1947	+ int ret;
	1948	+
	1949	+ ret = cp210x_read_vendor_block(serial, REQTYPE_DEVICE_TO_HOST, value,
	1950	+ ver, sizeof(ver));
	1951	+ if (ret)
	1952	+ return ret;
	1953	+
	1954	+ dev_dbg(&serial->interface->dev, "%s - %d.%d.%d\n", __func__,
	1955	+ ver[0], ver[1], ver[2]);
	1956	+
	1957	+ priv->fw_version = ver[0] << 16 \| ver[1] << 8 \| ver[2];
	1958	+
	1959	+ return 0;
	1960	+}
	1961	+
	1962	+static void cp210x_determine_quirks(struct usb_serial *serial)
	1963	+{
	1964	+ struct cp210x_serial_private *priv = usb_get_serial_data(serial);
	1965	+ int ret;
	1966	+
	1967	+ switch (priv->partnum) {
	1968	+ case CP210X_PARTNUM_CP2102N_QFN28:
	1969	+ case CP210X_PARTNUM_CP2102N_QFN24:
	1970	+ case CP210X_PARTNUM_CP2102N_QFN20:
	1971	+ ret = cp210x_get_fw_version(serial, CP210X_GET_FW_VER_2N);
	1972	+ if (ret)
	1973	+ break;
	1974	+ if (priv->fw_version <= 0x10004)
	1975	+ priv->no_flow_control = true;
	1976	+ break;
	1977	+ default:
	1978	+ break;
	1979	+ }
	1980	+}
	1981	+
1911	1982	static int cp210x_attach(struct usb_serial *serial)
1912	1983	{
1913	1984	int result;
@@ -1928,6 +1999,7 @@ static int cp210x_attach(struct usb_serial *serial) · 14 unmodified lines
1928	1999
1929	2000	usb_set_serial_data(serial, priv);
1930	2001
	2002	+ cp210x_determine_quirks(serial);
1931	2003	cp210x_init_max_speed(serial);
1932	2004
1933	2005	result = cp210x_gpio_init(serial);

drivers/usb/serial/ftdi_sio.c+4−0

@@ -611,6 +611,7 @@ static const struct usb_device_id id_table_combined[] = { · 610 unmodified lines
611	611	.driver_info = (kernel_ulong_t)&ftdi_jtag_quirk },
612	612	{ USB_DEVICE(FTDI_VID, FTDI_NT_ORIONLX_PLUS_PID) },
613	613	{ USB_DEVICE(FTDI_VID, FTDI_NT_ORION_IO_PID) },
	614	+ { USB_DEVICE(FTDI_VID, FTDI_NT_ORIONMX_PID) },
614	615	{ USB_DEVICE(FTDI_VID, FTDI_SYNAPSE_SS200_PID) },
615	616	{ USB_DEVICE(FTDI_VID, FTDI_CUSTOMWARE_MINIPLEX_PID) },
616	617	{ USB_DEVICE(FTDI_VID, FTDI_CUSTOMWARE_MINIPLEX2_PID) },
@@ -1034,6 +1035,9 @@ static const struct usb_device_id id_table_combined[] = { · 417 unmodified lines
1034	1035	/* Sienna devices */
1035	1036	{ USB_DEVICE(FTDI_VID, FTDI_SIENNA_PID) },
1036	1037	{ USB_DEVICE(ECHELON_VID, ECHELON_U20_PID) },
	1038	+ /* IDS GmbH devices */
	1039	+ { USB_DEVICE(IDS_VID, IDS_SI31A_PID) },
	1040	+ { USB_DEVICE(IDS_VID, IDS_CM31A_PID) },
1037	1041	/* U-Blox devices */
1038	1042	{ USB_DEVICE(UBLOX_VID, UBLOX_C099F9P_ZED_PID) },
1039	1043	{ USB_DEVICE(UBLOX_VID, UBLOX_C099F9P_ODIN_PID) },

drivers/usb/serial/ftdi_sio_ids.h+8−0

@@ -581,6 +581,7 @@ · 580 unmodified lines
581	581	#define FTDI_NT_ORIONLXM_PID 0x7c90 /* OrionLXm Substation Automation Platform */
582	582	#define FTDI_NT_ORIONLX_PLUS_PID 0x7c91 /* OrionLX+ Substation Automation Platform */
583	583	#define FTDI_NT_ORION_IO_PID 0x7c92 /* Orion I/O */
	584	+#define FTDI_NT_ORIONMX_PID 0x7c93 /* OrionMX */
584	585
585	586	/*
586	587	* Synapse Wireless product ids (FTDI_VID)
@@ -1567,6 +1568,13 @@ · 980 unmodified lines
1567	1568	#define UNJO_VID 0x22B7
1568	1569	#define UNJO_ISODEBUG_V1_PID 0x150D
1569	1570
	1571	+/*
	1572	+ * IDS GmbH
	1573	+ */
	1574	+#define IDS_VID 0x2CAF
	1575	+#define IDS_SI31A_PID 0x13A2
	1576	+#define IDS_CM31A_PID 0x13A3
	1577	+
1570	1578	/*
1571	1579	* U-Blox products (http://www.u-blox.com).
1572	1580	*/

drivers/usb/serial/omninet.c+5−3

@@ -1,6 +1,6 @@
1	1	// SPDX-License-Identifier: GPL-2.0
2	2	/*
3		− * USB ZyXEL omni.net LCD PLUS driver
	3	+ * USB ZyXEL omni.net driver
4	4	*
5	5	* Copyright (C) 2013,2017 Johan Hovold <johan@kernel.org>
6	6	*
@@ -22,10 +22,11 @@ · 15 unmodified lines
22	22	#include <linux/usb/serial.h>
23	23
24	24	#define DRIVER_AUTHOR "Alessandro Zummo"
25		−#define DRIVER_DESC "USB ZyXEL omni.net LCD PLUS Driver"
	25	+#define DRIVER_DESC "USB ZyXEL omni.net Driver"
26	26
27	27	#define ZYXEL_VENDOR_ID 0x0586
28	28	#define ZYXEL_OMNINET_ID 0x1000
	29	+#define ZYXEL_OMNI_56K_PLUS_ID 0x1500
29	30	/* This one seems to be a re-branded ZyXEL device */
30	31	#define BT_IGNITIONPRO_ID 0x2000
31	32
@@ -40,6 +41,7 @@ static void omninet_port_remove(struct usb_serial_port *port); · 8 unmodified lines
40	41
41	42	static const struct usb_device_id id_table[] = {
42	43	{ USB_DEVICE(ZYXEL_VENDOR_ID, ZYXEL_OMNINET_ID) },
	44	+ { USB_DEVICE(ZYXEL_VENDOR_ID, ZYXEL_OMNI_56K_PLUS_ID) },
43	45	{ USB_DEVICE(ZYXEL_VENDOR_ID, BT_IGNITIONPRO_ID) },
44	46	{ } /* Terminating entry */
45	47	};
@@ -50,7 +52,7 @@ static struct usb_serial_driver zyxel_omninet_device = { · 4 unmodified lines
50	52	.owner = THIS_MODULE,
51	53	.name = "omninet",
52	54	},
53		− .description = "ZyXEL - omni.net lcd plus usb",
	55	+ .description = "ZyXEL - omni.net usb",
54	56	.id_table = id_table,
55	57	.num_bulk_out = 2,
56	58	.calc_num_ports = omninet_calc_num_ports,

drivers/usb/serial/option.c+4−0

@@ -1240,6 +1240,10 @@ static const struct usb_device_id option_ids[] = { · 1239 unmodified lines
1240	1240	.driver_info = NCTRL(0) \| RSVD(1) },
1241	1241	{ USB_DEVICE_INTERFACE_CLASS(TELIT_VENDOR_ID, 0x1901, 0xff), /* Telit LN940 (MBIM) */
1242	1242	.driver_info = NCTRL(0) },
	1243	+ { USB_DEVICE_INTERFACE_CLASS(TELIT_VENDOR_ID, 0x7010, 0xff), /* Telit LE910-S1 (RNDIS) */
	1244	+ .driver_info = NCTRL(2) },
	1245	+ { USB_DEVICE_INTERFACE_CLASS(TELIT_VENDOR_ID, 0x7011, 0xff), /* Telit LE910-S1 (ECM) */
	1246	+ .driver_info = NCTRL(2) },
1243	1247	{ USB_DEVICE(TELIT_VENDOR_ID, 0x9010), /* Telit SBL FN980 flashing device */
1244	1248	.driver_info = NCTRL(0) \| ZLP },
1245	1249	{ USB_DEVICE_AND_INTERFACE_INFO(ZTE_VENDOR_ID, ZTE_PRODUCT_MF622, 0xff, 0xff, 0xff) }, /* ZTE WCDMA products */

drivers/usb/serial/pl2303.c+1−0

@@ -113,6 +113,7 @@ static const struct usb_device_id id_table[] = { · 112 unmodified lines
113	113	{ USB_DEVICE(SONY_VENDOR_ID, SONY_QN3USB_PRODUCT_ID) },
114	114	{ USB_DEVICE(SANWA_VENDOR_ID, SANWA_PRODUCT_ID) },
115	115	{ USB_DEVICE(ADLINK_VENDOR_ID, ADLINK_ND6530_PRODUCT_ID) },
	116	+ { USB_DEVICE(ADLINK_VENDOR_ID, ADLINK_ND6530GC_PRODUCT_ID) },
116	117	{ USB_DEVICE(SMART_VENDOR_ID, SMART_PRODUCT_ID) },
117	118	{ USB_DEVICE(AT_VENDOR_ID, AT_VTKIT3_PRODUCT_ID) },
118	119	{ } /* Terminating entry */

drivers/usb/serial/pl2303.h+1−0

@@ -158,6 +158,7 @@ · 157 unmodified lines
158	158	/* ADLINK ND-6530 RS232,RS485 and RS422 adapter */
159	159	#define ADLINK_VENDOR_ID 0x0b63
160	160	#define ADLINK_ND6530_PRODUCT_ID 0x6530
	161	+#define ADLINK_ND6530GC_PRODUCT_ID 0x653a
161	162
162	163	/* SMART USB Serial Adapter */
163	164	#define SMART_VENDOR_ID 0x0b8c

drivers/usb/serial/quatech2.c+3−3

@@ -416,7 +416,7 @@ static void qt2_close(struct usb_serial_port *port) · 415 unmodified lines
416	416
417	417	/* flush the port transmit buffer */
418	418	i = usb_control_msg(serial->dev,
419		− usb_rcvctrlpipe(serial->dev, 0),
	419	+ usb_sndctrlpipe(serial->dev, 0),
420	420	QT2_FLUSH_DEVICE, 0x40, 1,
421	421	port_priv->device_port, NULL, 0, QT2_USB_TIMEOUT);
422	422
@@ -426,7 +426,7 @@ static void qt2_close(struct usb_serial_port *port) · 3 unmodified lines
426	426
427	427	/* flush the port receive buffer */
428	428	i = usb_control_msg(serial->dev,
429		− usb_rcvctrlpipe(serial->dev, 0),
	429	+ usb_sndctrlpipe(serial->dev, 0),
430	430	QT2_FLUSH_DEVICE, 0x40, 0,
431	431	port_priv->device_port, NULL, 0, QT2_USB_TIMEOUT);
432	432
@@ -639,7 +639,7 @@ static int qt2_attach(struct usb_serial *serial) · 206 unmodified lines
639	639	int status;
640	640
641	641	/* power on unit */
642		− status = usb_control_msg(serial->dev, usb_rcvctrlpipe(serial->dev, 0),
	642	+ status = usb_control_msg(serial->dev, usb_sndctrlpipe(serial->dev, 0),
643	643	0xc2, 0x40, 0x8000, 0, NULL, 0,
644	644	QT2_USB_TIMEOUT);
645	645	if (status < 0) {

drivers/usb/serial/ti_usb_3410_5052.c+3−0

@@ -37,6 +37,7 @@ · 36 unmodified lines
37	37	/* Vendor and product ids */
38	38	#define TI_VENDOR_ID 0x0451
39	39	#define IBM_VENDOR_ID 0x04b3
	40	+#define STARTECH_VENDOR_ID 0x14b0
40	41	#define TI_3410_PRODUCT_ID 0x3410
41	42	#define IBM_4543_PRODUCT_ID 0x4543
42	43	#define IBM_454B_PRODUCT_ID 0x454b
@@ -370,6 +371,7 @@ static const struct usb_device_id ti_id_table_3410[] = { · 327 unmodified lines
370	371	{ USB_DEVICE(MXU1_VENDOR_ID, MXU1_1131_PRODUCT_ID) },
371	372	{ USB_DEVICE(MXU1_VENDOR_ID, MXU1_1150_PRODUCT_ID) },
372	373	{ USB_DEVICE(MXU1_VENDOR_ID, MXU1_1151_PRODUCT_ID) },
	374	+ { USB_DEVICE(STARTECH_VENDOR_ID, TI_3410_PRODUCT_ID) },
373	375	{ } /* terminator */
374	376	};
375	377
@@ -408,6 +410,7 @@ static const struct usb_device_id ti_id_table_combined[] = { · 32 unmodified lines
408	410	{ USB_DEVICE(MXU1_VENDOR_ID, MXU1_1131_PRODUCT_ID) },
409	411	{ USB_DEVICE(MXU1_VENDOR_ID, MXU1_1150_PRODUCT_ID) },
410	412	{ USB_DEVICE(MXU1_VENDOR_ID, MXU1_1151_PRODUCT_ID) },
	413	+ { USB_DEVICE(STARTECH_VENDOR_ID, TI_3410_PRODUCT_ID) },
411	414	{ } /* terminator */
412	415	};
413	416

drivers/usb/typec/mux.c+5−4

@@ -191,6 +191,7 @@ static void typec_mux_match(struct fwnode_handle fwnode, const char *id, · 190 unmodified lines
191	191	bool match;
192	192	int nval;
193	193	u16 *val;
	194	+ int ret;
194	195	int i;
195	196
196	197	/*
@@ -218,10 +219,10 @@ static void typec_mux_match(struct fwnode_handle fwnode, const char *id, · 21 unmodified lines
218	219	if (!val)
219	220	return ERR_PTR(-ENOMEM);
220	221
221		− nval = fwnode_property_read_u16_array(fwnode, "svid", val, nval);
222		− if (nval < 0) {
	222	+ ret = fwnode_property_read_u16_array(fwnode, "svid", val, nval);
	223	+ if (ret < 0) {
223	224	kfree(val);
224		− return ERR_PTR(nval);
	225	+ return ERR_PTR(ret);
225	226	}
226	227
227	228	for (i = 0; i < nval; i++) {
@@ -238,7 +239,7 @@ find_mux: · 10 unmodified lines
238	239	dev = class_find_device(&typec_mux_class, NULL, fwnode,
239	240	mux_fwnode_match);
240	241
241		− return dev ? to_typec_switch(dev) : ERR_PTR(-EPROBE_DEFER);
	242	+ return dev ? to_typec_mux(dev) : ERR_PTR(-EPROBE_DEFER);
242	243	}
243	244
244	245	/**

drivers/usb/typec/mux/intel_pmc_mux.c+11−4

@@ -582,10 +582,15 @@ static int pmc_usb_probe_iom(struct pmc_usb *pmc) · 581 unmodified lines
582	582	acpi_dev_free_resource_list(&resource_list);
583	583
584	584	if (!pmc->iom_base) {
585		− put_device(&adev->dev);
	585	+ acpi_dev_put(adev);
586	586	return -ENOMEM;
587	587	}
588	588
	589	+ if (IS_ERR(pmc->iom_base)) {
	590	+ acpi_dev_put(adev);
	591	+ return PTR_ERR(pmc->iom_base);
	592	+ }
	593	+
589	594	pmc->iom_adev = adev;
590	595
591	596	return 0;
@@ -636,8 +641,10 @@ static int pmc_usb_probe(struct platform_device *pdev) · 44 unmodified lines
636	641	break;
637	642
638	643	ret = pmc_usb_register_port(pmc, i, fwnode);
639		− if (ret)
	644	+ if (ret) {
	645	+ fwnode_handle_put(fwnode);
640	646	goto err_remove_ports;
	647	+ }
641	648	}
642	649
643	650	platform_set_drvdata(pdev, pmc);
@@ -651,7 +658,7 @@ err_remove_ports: · 7 unmodified lines
651	658	usb_role_switch_unregister(pmc->port[i].usb_sw);
652	659	}
653	660
654		− put_device(&pmc->iom_adev->dev);
	661	+ acpi_dev_put(pmc->iom_adev);
655	662
656	663	return ret;
657	664	}
@@ -667,7 +674,7 @@ static int pmc_usb_remove(struct platform_device *pdev) · 9 unmodified lines
667	674	usb_role_switch_unregister(pmc->port[i].usb_sw);
668	675	}
669	676
670		− put_device(&pmc->iom_adev->dev);
	677	+ acpi_dev_put(pmc->iom_adev);
671	678
672	679	return 0;
673	680	}

drivers/usb/typec/tcpm/tcpm.c+107−43

@@ -401,6 +401,8 @@ struct tcpm_port { · 400 unmodified lines
401	401	unsigned int nr_src_pdo;
402	402	u32 snk_pdo[PDO_MAX_OBJECTS];
403	403	unsigned int nr_snk_pdo;
	404	+ u32 snk_vdo_v1[VDO_MAX_OBJECTS];
	405	+ unsigned int nr_snk_vdo_v1;
404	406	u32 snk_vdo[VDO_MAX_OBJECTS];
405	407	unsigned int nr_snk_vdo;
406	408
@@ -1547,33 +1549,43 @@ static int tcpm_pd_svdm(struct tcpm_port port, struct typec_altmode adev, · 1140 unmodified lines
1547	1549	if (PD_VDO_VID(p[0]) != USB_SID_PD)
1548	1550	break;
1549	1551
1550		− if (PD_VDO_SVDM_VER(p[0]) < svdm_version)
	1552	+ if (PD_VDO_SVDM_VER(p[0]) < svdm_version) {
1551	1553	typec_partner_set_svdm_version(port->partner,
1552	1554	PD_VDO_SVDM_VER(p[0]));
1553		− /* 6.4.4.3.1: Only respond as UFP (device) */
1554		− if (port->data_role == TYPEC_DEVICE &&
	1555	+ svdm_version = PD_VDO_SVDM_VER(p[0]);
	1556	+ }
	1557	+
	1558	+ port->ams = DISCOVER_IDENTITY;
	1559	+ /*
	1560	+ * PD2.0 Spec 6.10.3: respond with NAK as DFP (data host)
	1561	+ * PD3.1 Spec 6.4.4.2.5.1: respond with NAK if "invalid field" or
	1562	+ * "wrong configuation" or "Unrecognized"
	1563	+ */
	1564	+ if ((port->data_role == TYPEC_DEVICE \|\| svdm_version >= SVDM_VER_2_0) &&
1555	1565	port->nr_snk_vdo) {
1556		− /*
1557		− * Product Type DFP and Connector Type are not defined in SVDM
1558		− * version 1.0 and shall be set to zero.
1559		− */
1560		− if (typec_get_negotiated_svdm_version(typec) < SVDM_VER_2_0)
1561		− response[1] = port->snk_vdo[0] & ~IDH_DFP_MASK
1562		− & ~IDH_CONN_MASK;
1563		− else
1564		− response[1] = port->snk_vdo[0];
1565		− for (i = 1; i < port->nr_snk_vdo; i++)
1566		− response[i + 1] = port->snk_vdo[i];
1567		− rlen = port->nr_snk_vdo + 1;
	1566	+ if (svdm_version < SVDM_VER_2_0) {
	1567	+ for (i = 0; i < port->nr_snk_vdo_v1; i++)
	1568	+ response[i + 1] = port->snk_vdo_v1[i];
	1569	+ rlen = port->nr_snk_vdo_v1 + 1;
	1570	+
	1571	+ } else {
	1572	+ for (i = 0; i < port->nr_snk_vdo; i++)
	1573	+ response[i + 1] = port->snk_vdo[i];
	1574	+ rlen = port->nr_snk_vdo + 1;
	1575	+ }
1568	1576	}
1569	1577	break;
1570	1578	case CMD_DISCOVER_SVID:
	1579	+ port->ams = DISCOVER_SVIDS;
1571	1580	break;
1572	1581	case CMD_DISCOVER_MODES:
	1582	+ port->ams = DISCOVER_MODES;
1573	1583	break;
1574	1584	case CMD_ENTER_MODE:
	1585	+ port->ams = DFP_TO_UFP_ENTER_MODE;
1575	1586	break;
1576	1587	case CMD_EXIT_MODE:
	1588	+ port->ams = DFP_TO_UFP_EXIT_MODE;
1577	1589	break;
1578	1590	case CMD_ATTENTION:
1579	1591	/* Attention command does not have response */
@@ -1930,6 +1942,9 @@ static void vdm_run_state_machine(struct tcpm_port *port) · 350 unmodified lines
1930	1942	tcpm_log(port, "VDM Tx error, retry");
1931	1943	port->vdm_retries++;
1932	1944	port->vdm_state = VDM_STATE_READY;
	1945	+ if (PD_VDO_SVDM(vdo_hdr) && PD_VDO_CMDT(vdo_hdr) == CMDT_INIT)
	1946	+ tcpm_ams_finish(port);
	1947	+ } else {
1933	1948	tcpm_ams_finish(port);
1934	1949	}
1935	1950	break;
@@ -2176,20 +2191,25 @@ static void tcpm_handle_alert(struct tcpm_port port, const __le32 payload, · 240 unmodified lines
2176	2191
2177	2192	if (!type) {
2178	2193	tcpm_log(port, "Alert message received with no type");
	2194	+ tcpm_queue_message(port, PD_MSG_CTRL_NOT_SUPP);
2179	2195	return;
2180	2196	}
2181	2197
2182	2198	/* Just handling non-battery alerts for now */
2183	2199	if (!(type & USB_PD_ADO_TYPE_BATT_STATUS_CHANGE)) {
2184		− switch (port->state) {
2185		− case SRC_READY:
2186		− case SNK_READY:
	2200	+ if (port->pwr_role == TYPEC_SOURCE) {
	2201	+ port->upcoming_state = GET_STATUS_SEND;
	2202	+ tcpm_ams_start(port, GETTING_SOURCE_SINK_STATUS);
	2203	+ } else {
	2204	+ /*
	2205	+ * Do not check SinkTxOk here in case the Source doesn't set its Rp to
	2206	+ * SinkTxOk in time.
	2207	+ */
	2208	+ port->ams = GETTING_SOURCE_SINK_STATUS;
2187	2209	tcpm_set_state(port, GET_STATUS_SEND, 0);
2188		− break;
2189		− default:
2190		− tcpm_queue_message(port, PD_MSG_CTRL_WAIT);
2191		− break;
2192	2210	}
	2211	+ } else {
	2212	+ tcpm_queue_message(port, PD_MSG_CTRL_NOT_SUPP);
2193	2213	}
2194	2214	}
2195	2215
@@ -2287,6 +2307,12 @@ static void tcpm_pd_data_request(struct tcpm_port *port, · 91 unmodified lines
2287	2307	bool frs_enable;
2288	2308	int ret;
2289	2309
	2310	+ if (tcpm_vdm_ams(port) && type != PD_DATA_VENDOR_DEF) {
	2311	+ port->vdm_state = VDM_STATE_ERR_BUSY;
	2312	+ tcpm_ams_finish(port);
	2313	+ mod_vdm_delayed_work(port, 0);
	2314	+ }
	2315	+
2290	2316	switch (type) {
2291	2317	case PD_DATA_SOURCE_CAP:
2292	2318	for (i = 0; i < cnt; i++)
@@ -2417,14 +2443,22 @@ static void tcpm_pd_data_request(struct tcpm_port *port, · 124 unmodified lines
2417	2443	NONE_AMS);
2418	2444	break;
2419	2445	case PD_DATA_VENDOR_DEF:
2420		− tcpm_handle_vdm_request(port, msg->payload, cnt);
	2446	+ if (tcpm_vdm_ams(port) \|\| port->nr_snk_vdo)
	2447	+ tcpm_handle_vdm_request(port, msg->payload, cnt);
	2448	+ else if (port->negotiated_rev > PD_REV20)
	2449	+ tcpm_pd_handle_msg(port, PD_MSG_CTRL_NOT_SUPP, NONE_AMS);
2421	2450	break;
2422	2451	case PD_DATA_BIST:
2423	2452	port->bist_request = le32_to_cpu(msg->payload[0]);
2424	2453	tcpm_pd_handle_state(port, BIST_RX, BIST, 0);
2425	2454	break;
2426	2455	case PD_DATA_ALERT:
2427		− tcpm_handle_alert(port, msg->payload, cnt);
	2456	+ if (port->state != SRC_READY && port->state != SNK_READY)
	2457	+ tcpm_pd_handle_state(port, port->pwr_role == TYPEC_SOURCE ?
	2458	+ SRC_SOFT_RESET_WAIT_SNK_TX : SNK_SOFT_RESET,
	2459	+ NONE_AMS, 0);
	2460	+ else
	2461	+ tcpm_handle_alert(port, msg->payload, cnt);
2428	2462	break;
2429	2463	case PD_DATA_BATT_STATUS:
2430	2464	case PD_DATA_GET_COUNTRY_INFO:

111 more lines aren’t shown here.

drivers/usb/typec/tcpm/wcove.c+1−1

This diff is large: its 9 lines aren’t shown here.

drivers/usb/typec/ucsi/ucsi.c+2−1

This diff is large: its 18 lines aren’t shown here.

drivers/vdpa/mlx5/net/mlx5_vnet.c+18−1

This diff is large: its 62 lines aren’t shown here.

drivers/vfio/pci/Kconfig+1−0

This diff is large: its 8 lines aren’t shown here.

drivers/vfio/pci/vfio_pci_config.c+1−1

This diff is large: its 9 lines aren’t shown here.

drivers/vfio/platform/vfio_platform_common.c+1−1

This diff is large: its 9 lines aren’t shown here.

drivers/vfio/vfio_iommu_type1.c+1−1

This diff is large: its 9 lines aren’t shown here.

drivers/video/fbdev/core/fb_defio.c+35−0

This diff is large: its 62 lines aren’t shown here.

drivers/video/fbdev/core/fbmem.c+4−0

This diff is large: its 11 lines aren’t shown here.

drivers/video/fbdev/hgafb.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/afs/cmservice.c+5−0

This diff is large: its 33 lines aren’t shown here.

fs/afs/dir.c+3−1

This diff is large: its 11 lines aren’t shown here.

fs/afs/fsclient.c+4−0

This diff is large: its 32 lines aren’t shown here.

fs/afs/main.c+2−2

This diff is large: its 11 lines aren’t shown here.

fs/afs/vlclient.c+1−0

This diff is large: its 8 lines aren’t shown here.

fs/afs/write.c+9−6

This diff is large: its 56 lines aren’t shown here.

fs/btrfs/block-group.c+4−4

This diff is large: its 21 lines aren’t shown here.

fs/btrfs/compression.c+12−5

This diff is large: its 31 lines aren’t shown here.

fs/btrfs/disk-io.c+18−8

This diff is large: its 40 lines aren’t shown here.

fs/btrfs/extent-tree.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/btrfs/file-item.c+81−27

This diff is large: its 170 lines aren’t shown here.

fs/btrfs/file.c+2−2

This diff is large: its 18 lines aren’t shown here.

fs/btrfs/inode.c+18−1

This diff is large: its 48 lines aren’t shown here.

fs/btrfs/reflink.c+22−16

This diff is large: its 78 lines aren’t shown here.

fs/btrfs/tree-log.c+29−8

This diff is large: its 86 lines aren’t shown here.

fs/btrfs/zoned.c+18−5

This diff is large: its 58 lines aren’t shown here.

fs/cifs/cifs_ioctl.h+19−6

This diff is large: its 35 lines aren’t shown here.

fs/cifs/cifspdu.h+2−1

This diff is large: its 10 lines aren’t shown here.

fs/cifs/ioctl.c+105−38

This diff is large: its 189 lines aren’t shown here.

fs/cifs/smb2pdu.c+7−0

This diff is large: its 14 lines aren’t shown here.

fs/cifs/trace.h+17−12

This diff is large: its 97 lines aren’t shown here.

fs/coredump.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/debugfs/file.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/debugfs/inode.c+6−3

This diff is large: its 17 lines aren’t shown here.

fs/ext4/extents.c+23−20

This diff is large: its 64 lines aren’t shown here.

fs/ext4/fast_commit.c+90−80

This diff is large: its 378 lines aren’t shown here.

fs/ext4/fast_commit.h+0−19

This diff is large: its 31 lines aren’t shown here.

fs/ext4/ialloc.c+4−2

This diff is large: its 19 lines aren’t shown here.

fs/ext4/mballoc.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/ext4/namei.c+4−2

This diff is large: its 20 lines aren’t shown here.

fs/ext4/super.c+9−2

This diff is large: its 33 lines aren’t shown here.

fs/ext4/sysfs.c+4−0

This diff is large: its 20 lines aren’t shown here.

fs/gfs2/file.c+4−1

This diff is large: its 13 lines aren’t shown here.

fs/gfs2/glock.c+24−4

This diff is large: its 74 lines aren’t shown here.

fs/gfs2/glops.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/gfs2/log.c+3−3

This diff is large: its 29 lines aren’t shown here.

fs/gfs2/log.h+1−0

This diff is large: its 7 lines aren’t shown here.

fs/gfs2/lops.c+6−1

This diff is large: its 21 lines aren’t shown here.

fs/gfs2/lops.h+1−0

This diff is large: its 8 lines aren’t shown here.

fs/gfs2/util.c+1−0

This diff is large: its 8 lines aren’t shown here.

fs/hugetlbfs/inode.c+1−0

This diff is large: its 8 lines aren’t shown here.

fs/io-wq.c+15−14

This diff is large: its 69 lines aren’t shown here.

fs/io-wq.h+1−1

This diff is large: its 9 lines aren’t shown here.

fs/io_uring.c+43−15

This diff is large: its 145 lines aren’t shown here.

fs/netfs/Kconfig+1−1

This diff is large: its 9 lines aren’t shown here.

fs/netfs/read_helper.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/nfs/client.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/nfs/filelayout/filelayout.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/nfs/namespace.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/nfs/nfs4_fs.h+1−0

This diff is large: its 8 lines aren’t shown here.

fs/nfs/nfs4client.c+1−1

This diff is large: its 10 lines aren’t shown here.

fs/nfs/nfs4file.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/nfs/nfs4proc.c+32−9

This diff is large: its 125 lines aren’t shown here.

fs/nfs/nfstrace.h+0−4

This diff is large: its 11 lines aren’t shown here.

fs/nfs/pagelist.c+6−14

This diff is large: its 52 lines aren’t shown here.

fs/nfs/pnfs.c+8−9

This diff is large: its 40 lines aren’t shown here.

fs/nfs/super.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/notify/fanotify/fanotify_user.c+26−8

This diff is large: its 95 lines aren’t shown here.

fs/notify/fdinfo.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/ocfs2/file.c+50−5

This diff is large: its 97 lines aren’t shown here.

fs/proc/base.c+13−0

This diff is large: its 37 lines aren’t shown here.

fs/xfs/libxfs/xfs_ag_resv.c+15−3

This diff is large: its 26 lines aren’t shown here.

fs/xfs/libxfs/xfs_bmap.c+0−12

This diff is large: its 33 lines aren’t shown here.

fs/xfs/libxfs/xfs_inode_buf.c+42−4

This diff is large: its 67 lines aren’t shown here.

fs/xfs/libxfs/xfs_trans_inode.c+17−0

This diff is large: its 24 lines aren’t shown here.

fs/xfs/xfs_inode.c+29−0

This diff is large: its 56 lines aren’t shown here.

fs/xfs/xfs_ioctl.c+34−67

This diff is large: its 142 lines aren’t shown here.

fs/xfs/xfs_message.h+2−0

This diff is large: its 9 lines aren’t shown here.

include/asm-generic/vmlinux.lds.h+1−0

This diff is large: its 8 lines aren’t shown here.

include/dt-bindings/usb/pd.h+76−13

This diff is large: its 158 lines aren’t shown here.

include/linux/arch_topology.h+0−1

This diff is large: its 8 lines aren’t shown here.

include/linux/avf/virtchnl.h+1−0

This diff is large: its 8 lines aren’t shown here.

include/linux/cgroup-defs.h+3−3

This diff is large: its 27 lines aren’t shown here.

include/linux/cgroup.h+1−1

This diff is large: its 9 lines aren’t shown here.

include/linux/compiler_attributes.h+1−0

This diff is large: its 8 lines aren’t shown here.

include/linux/device.h+2−4

This diff is large: its 20 lines aren’t shown here.

include/linux/entry-kvm.h+2−1

This diff is large: its 17 lines aren’t shown here.

include/linux/fanotify.h+4−0

This diff is large: its 11 lines aren’t shown here.

include/linux/fb.h+3−0

This diff is large: its 10 lines aren’t shown here.

include/linux/hid.h+1−2

This diff is large: its 10 lines aren’t shown here.

include/linux/host1x.h+24−6

This diff is large: its 37 lines aren’t shown here.

include/linux/huge_mm.h+7−1

This diff is large: its 29 lines aren’t shown here.

include/linux/hugetlb.h+8−0

This diff is large: its 29 lines aren’t shown here.

include/linux/init.h+2−1

This diff is large: its 10 lines aren’t shown here.

include/linux/kvm_host.h+16−2

This diff is large: its 46 lines aren’t shown here.

include/linux/mfd/rohm-bd70528.h+1−3

This diff is large: its 11 lines aren’t shown here.

include/linux/mfd/rohm-bd71828.h+5−5

This diff is large: its 17 lines aren’t shown here.

include/linux/mlx4/device.h+1−0

This diff is large: its 8 lines aren’t shown here.

include/linux/mlx5/driver.h+26−22

This diff is large: its 83 lines aren’t shown here.

include/linux/mlx5/mlx5_ifc.h+2−0

This diff is large: its 9 lines aren’t shown here.

include/linux/mlx5/mpfs.h+18−0

This diff is large: its 19 lines aren’t shown here.

include/linux/mlx5/transobj.h+1−0

This diff is large: its 6 lines aren’t shown here.

include/linux/mm.h+3−0

This diff is large: its 24 lines aren’t shown here.

include/linux/mm_types.h+20−7

This diff is large: its 49 lines aren’t shown here.

include/linux/pci.h+2−0

This diff is large: its 16 lines aren’t shown here.

include/linux/pgtable.h+8−0

This diff is large: its 15 lines aren’t shown here.

include/linux/phy.h+5−0

This diff is large: its 12 lines aren’t shown here.

include/linux/platform_data/ti-sysc.h+1−0

This diff is large: its 8 lines aren’t shown here.

include/linux/ptp_clock_kernel.h+1−1

This diff is large: its 9 lines aren’t shown here.

include/linux/rmap.h+1−0

This diff is large: its 8 lines aren’t shown here.

include/linux/rtsx_pci.h+2−0

This diff is large: its 16 lines aren’t shown here.

include/linux/sched.h+8−0

This diff is large: its 20 lines aren’t shown here.

include/linux/socket.h+0−2

This diff is large: its 7 lines aren’t shown here.

include/linux/spi/spi.h+2−2

This diff is large: its 11 lines aren’t shown here.

include/linux/sunrpc/xprt.h+2−0

This diff is large: its 9 lines aren’t shown here.

include/linux/swapops.h+11−4

This diff is large: its 29 lines aren’t shown here.

include/linux/tick.h+7−0

This diff is large: its 19 lines aren’t shown here.

include/linux/usb/pd.h+1−1

This diff is large: its 9 lines aren’t shown here.

include/linux/usb/pd_ext_sdb.h+0−4

This diff is large: its 9 lines aren’t shown here.

include/net/caif/caif_dev.h+1−1

This diff is large: its 9 lines aren’t shown here.

include/net/caif/cfcnfg.h+1−1

This diff is large: its 9 lines aren’t shown here.

include/net/caif/cfserl.h+1−0

This diff is large: its 6 lines aren’t shown here.

include/net/cfg80211.h+2−2

This diff is large: its 18 lines aren’t shown here.

include/net/mac80211.h+7−2

This diff is large: its 23 lines aren’t shown here.

include/net/net_namespace.h+13−1

This diff is large: its 34 lines aren’t shown here.

include/net/netfilter/nf_flow_table.h+0−1

This diff is large: its 8 lines aren’t shown here.

include/net/netfilter/nf_tables.h+0−6

This diff is large: its 17 lines aren’t shown here.

include/net/nfc/nci_core.h+1−0

This diff is large: its 8 lines aren’t shown here.

include/net/pkt_cls.h+11−0

This diff is large: its 18 lines aren’t shown here.

include/net/pkt_sched.h+1−6

This diff is large: its 14 lines aren’t shown here.

include/net/sch_generic.h+34−1

This diff is large: its 59 lines aren’t shown here.

include/net/sock.h+16−5

This diff is large: its 57 lines aren’t shown here.

include/net/tls.h+9−1

This diff is large: its 31 lines aren’t shown here.

include/sound/soc-dai.h+1−1

This diff is large: its 9 lines aren’t shown here.

include/uapi/asm-generic/unistd.h+1−2

This diff is large: its 10 lines aren’t shown here.

include/uapi/linux/in.h+3−0

This diff is large: its 10 lines aren’t shown here.

include/uapi/linux/input-event-codes.h+1−0

This diff is large: its 8 lines aren’t shown here.

include/uapi/linux/io_uring.h+10−9

This diff is large: its 50 lines aren’t shown here.

include/uapi/linux/kvm.h+3−2

This diff is large: its 19 lines aren’t shown here.

include/uapi/linux/virtio_ids.h+1−1

This diff is large: its 9 lines aren’t shown here.

init/Kconfig+1−40

This diff is large: its 55 lines aren’t shown here.

init/main.c+1−1

This diff is large: its 9 lines aren’t shown here.

kernel/bpf/Kconfig+89−0

This diff is large: its 90 lines aren’t shown here.

kernel/bpf/bpf_lsm.c+2−0

This diff is large: its 13 lines aren’t shown here.

kernel/bpf/btf.c+12−0

This diff is large: its 26 lines aren’t shown here.

kernel/bpf/helpers.c+24−18

This diff is large: its 90 lines aren’t shown here.

kernel/bpf/ringbuf.c+11−13

This diff is large: its 44 lines aren’t shown here.

kernel/bpf/syscall.c+2−1

This diff is large: its 10 lines aren’t shown here.

kernel/bpf/verifier.c+118−44

This diff is large: its 347 lines aren’t shown here.

kernel/cgroup/cgroup-v1.c+5−1

This diff is large: its 20 lines aren’t shown here.

kernel/cgroup/cgroup.c+9−12

This diff is large: its 71 lines aren’t shown here.

kernel/cgroup/cpuset.c+1−1

This diff is large: its 9 lines aren’t shown here.

kernel/cgroup/rdma.c+1−1

This diff is large: its 9 lines aren’t shown here.

kernel/cgroup/rstat.c+1−1

This diff is large: its 9 lines aren’t shown here.

kernel/crash_core.c+1−0

This diff is large: its 8 lines aren’t shown here.

kernel/entry/common.c+3−2

This diff is large: its 26 lines aren’t shown here.

kernel/events/core.c+2−0

This diff is large: its 10 lines aren’t shown here.

kernel/irq_work.c+0−3

This diff is large: its 10 lines aren’t shown here.

kernel/printk/printk_safe.c+1−1

This diff is large: its 13 lines aren’t shown here.

kernel/sched/core.c+0−1

This diff is large: its 8 lines aren’t shown here.

kernel/sched/debug.c+2−1

This diff is large: its 17 lines aren’t shown here.

kernel/sched/fair.c+42−30

This diff is large: its 157 lines aren’t shown here.

kernel/sched/pelt.h+1−10

This diff is large: its 25 lines aren’t shown here.

kernel/seccomp.c+16−14

This diff is large: its 45 lines aren’t shown here.

kernel/sysctl.c+24−5

This diff is large: its 43 lines aren’t shown here.

kernel/time/tick-sched.c+1−0

This diff is large: its 8 lines aren’t shown here.

kernel/trace/bpf_trace.c+12−20

This diff is large: its 69 lines aren’t shown here.

kernel/trace/ftrace.c+7−1

This diff is large: its 20 lines aren’t shown here.

kernel/trace/trace.c+1−12

This diff is large: its 48 lines aren’t shown here.

kernel/trace/trace_clock.c+3−3

This diff is large: its 21 lines aren’t shown here.

kernel/workqueue.c+10−2

This diff is large: its 44 lines aren’t shown here.

lib/crc64.c+1−1

This diff is large: its 9 lines aren’t shown here.

lib/percpu-refcount.c+3−3

This diff is large: its 27 lines aren’t shown here.

mm/debug_vm_pgtable.c+2−2

This diff is large: its 18 lines aren’t shown here.

mm/huge_memory.c+29−27

This diff is large: its 136 lines aren’t shown here.

mm/hugetlb.c+125−26

This diff is large: its 233 lines aren’t shown here.

mm/internal.h+39−14

This diff is large: its 67 lines aren’t shown here.

mm/kasan/init.c+2−2

This diff is large: its 11 lines aren’t shown here.

mm/kfence/core.c+3−3

This diff is large: its 14 lines aren’t shown here.

mm/memory-failure.c+33−3

This diff is large: its 65 lines aren’t shown here.

mm/memory.c+45−0

This diff is large: its 81 lines aren’t shown here.

mm/migrate.c+1−0

This diff is large: its 8 lines aren’t shown here.

mm/page_alloc.c+2−0

This diff is large: its 9 lines aren’t shown here.

mm/page_vma_mapped.c+17−10

This diff is large: its 56 lines aren’t shown here.

mm/pgtable-generic.c+2−3

This diff is large: its 12 lines aren’t shown here.

mm/rmap.c+27−12

This diff is large: its 98 lines aren’t shown here.

mm/slab_common.c+1−2

This diff is large: its 10 lines aren’t shown here.

mm/slub.c+16−21

This diff is large: its 123 lines aren’t shown here.

mm/sparse.c+11−2

This diff is large: its 34 lines aren’t shown here.

mm/swapfile.c+1−1

This diff is large: its 9 lines aren’t shown here.

mm/truncate.c+19−24

This diff is large: its 80 lines aren’t shown here.

net/Kconfig+0−27

This diff is large: its 38 lines aren’t shown here.

net/appletalk/aarp.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/batman-adv/bat_iv_ogm.c+3−1

This diff is large: its 12 lines aren’t shown here.

net/bluetooth/hci_core.c+6−1

This diff is large: its 15 lines aren’t shown here.

net/bluetooth/hci_sock.c+2−2

This diff is large: its 18 lines aren’t shown here.

net/bluetooth/smp.c+3−3

This diff is large: its 27 lines aren’t shown here.

net/bridge/br_private.h+2−2

This diff is large: its 11 lines aren’t shown here.

net/bridge/br_vlan_tunnel.c+24−14

This diff is large: its 91 lines aren’t shown here.

net/caif/caif_dev.c+9−4

This diff is large: its 60 lines aren’t shown here.

net/caif/caif_usb.c+13−1

This diff is large: its 43 lines aren’t shown here.

net/caif/cfcnfg.c+11−5

This diff is large: its 61 lines aren’t shown here.

net/caif/cfserl.c+5−0

This diff is large: its 12 lines aren’t shown here.

net/can/bcm.c+49−13

This diff is large: its 152 lines aren’t shown here.

net/can/isotp.c+81−29

This diff is large: its 212 lines aren’t shown here.

net/can/j1939/transport.c+40−14

This diff is large: its 171 lines aren’t shown here.

net/can/raw.c+48−14

This diff is large: its 132 lines aren’t shown here.

net/compat.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/core/dev.c+24−5

This diff is large: its 58 lines aren’t shown here.

net/core/devlink.c+2−2

This diff is large: its 25 lines aren’t shown here.

net/core/fib_rules.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/core/filter.c+1−0

This diff is large: its 8 lines aren’t shown here.

net/core/neighbour.c+1−0

This diff is large: its 8 lines aren’t shown here.

net/core/net_namespace.c+13−7

This diff is large: its 35 lines aren’t shown here.

net/core/rtnetlink.c+4−0

This diff is large: its 11 lines aren’t shown here.

net/core/skbuff.c+3−1

This diff is large: its 25 lines aren’t shown here.

net/core/sock.c+16−8

This diff is large: its 49 lines aren’t shown here.

net/dsa/master.c+3−2

This diff is large: its 19 lines aren’t shown here.

net/dsa/slave.c+7−5

This diff is large: its 21 lines aren’t shown here.

net/dsa/tag_8021q.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/ethtool/eeprom.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/ethtool/ioctl.c+5−5

This diff is large: its 45 lines aren’t shown here.

net/ethtool/stats.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/ethtool/strset.c+2−0

This diff is large: its 9 lines aren’t shown here.

net/hsr/hsr_device.c+2−0

This diff is large: its 16 lines aren’t shown here.

net/hsr/hsr_forward.c+21−9

This diff is large: its 79 lines aren’t shown here.

net/hsr/hsr_forward.h+4−4

This diff is large: its 13 lines aren’t shown here.

net/hsr/hsr_main.h+2−2

This diff is large: its 11 lines aren’t shown here.

net/hsr/hsr_slave.c+5−6

This diff is large: its 18 lines aren’t shown here.

net/ieee802154/nl-mac.c+6−4

This diff is large: its 39 lines aren’t shown here.

net/ieee802154/nl-phy.c+3−1

This diff is large: its 12 lines aren’t shown here.

net/ieee802154/nl802154.c+5−4

This diff is large: its 25 lines aren’t shown here.

net/ipv4/af_inet.c+2−2

This diff is large: its 18 lines aren’t shown here.

net/ipv4/bpf_tcp_ca.c+2−0

This diff is large: its 16 lines aren’t shown here.

net/ipv4/cipso_ipv4.c+1−0

This diff is large: its 8 lines aren’t shown here.

net/ipv4/devinet.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/ipv4/icmp.c+7−0

This diff is large: its 14 lines aren’t shown here.

net/ipv4/igmp.c+1−0

This diff is large: its 8 lines aren’t shown here.

net/ipv4/ipconfig.c+8−5

This diff is large: its 29 lines aren’t shown here.

net/ipv4/ping.c+7−5

This diff is large: its 29 lines aren’t shown here.

net/ipv4/route.c+14−1

This diff is large: its 29 lines aren’t shown here.

net/ipv4/udp.c+10−0

This diff is large: its 28 lines aren’t shown here.

net/ipv6/addrconf.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/ipv6/mcast.c+0−3

This diff is large: its 11 lines aren’t shown here.

net/ipv6/netfilter/nft_fib_ipv6.c+18−4

This diff is large: its 36 lines aren’t shown here.

net/ipv6/reassembly.c+3−1

This diff is large: its 18 lines aren’t shown here.

net/ipv6/route.c+6−2

This diff is large: its 25 lines aren’t shown here.

net/ipv6/sit.c+3−0

This diff is large: its 10 lines aren’t shown here.

net/ipv6/udp.c+3−0

This diff is large: its 10 lines aren’t shown here.

net/mac80211/debugfs.c+9−2

This diff is large: its 28 lines aren’t shown here.

net/mac80211/ieee80211_i.h+14−24

This diff is large: its 78 lines aren’t shown here.

net/mac80211/iface.c+15−15

This diff is large: its 78 lines aren’t shown here.

net/mac80211/key.c+7−0

This diff is large: its 21 lines aren’t shown here.

net/mac80211/key.h+2−0

This diff is large: its 9 lines aren’t shown here.

net/mac80211/main.c+6−1

This diff is large: its 22 lines aren’t shown here.

net/mac80211/mlme.c+8−0

This diff is large: its 32 lines aren’t shown here.

net/mac80211/rc80211_minstrel_ht.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/mac80211/rx.c+119−38

This diff is large: its 316 lines aren’t shown here.

net/mac80211/scan.c+16−5

This diff is large: its 30 lines aren’t shown here.

net/mac80211/sta_info.c+5−1

This diff is large: its 27 lines aren’t shown here.

net/mac80211/sta_info.h+32−1

This diff is large: its 61 lines aren’t shown here.

net/mac80211/tx.c+36−16

This diff is large: its 101 lines aren’t shown here.

net/mac80211/util.c+11−13

This diff is large: its 107 lines aren’t shown here.

net/mac80211/wpa.c+9−4

This diff is large: its 48 lines aren’t shown here.

net/mptcp/options.c+3−2

This diff is large: its 26 lines aren’t shown here.

net/mptcp/pm_netlink.c+4−4

This diff is large: its 31 lines aren’t shown here.

net/mptcp/protocol.c+59−29

This diff is large: its 218 lines aren’t shown here.

net/mptcp/protocol.h+2−2

This diff is large: its 25 lines aren’t shown here.

net/mptcp/sockopt.c+2−2

This diff is large: its 18 lines aren’t shown here.

net/mptcp/subflow.c+90−92

This diff is large: its 286 lines aren’t shown here.

net/netfilter/ipvs/ip_vs_ctl.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/netfilter/nf_conntrack_proto.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/netfilter/nf_flow_table_core.c+1−2

This diff is large: its 10 lines aren’t shown here.

net/netfilter/nf_flow_table_offload.c+4−3

This diff is large: its 15 lines aren’t shown here.

net/netfilter/nf_synproxy_core.c+5−0

This diff is large: its 19 lines aren’t shown here.

net/netfilter/nf_tables_api.c+100−69

This diff is large: its 329 lines aren’t shown here.

net/netfilter/nfnetlink_cthelper.c+6−2

This diff is large: its 17 lines aren’t shown here.

net/netfilter/nft_ct.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/netfilter/nft_set_pipapo.c+2−2

This diff is large: its 11 lines aren’t shown here.

net/netfilter/nft_set_pipapo.h+2−0

This diff is large: its 9 lines aren’t shown here.

net/netfilter/nft_set_pipapo_avx2.c+3−0

This diff is large: its 10 lines aren’t shown here.

net/netlink/af_netlink.c+4−2

This diff is large: its 16 lines aren’t shown here.

net/nfc/llcp_sock.c+2−0

This diff is large: its 16 lines aren’t shown here.

net/nfc/nci/core.c+1−0

This diff is large: its 8 lines aren’t shown here.

net/nfc/nci/hci.c+5−0

This diff is large: its 9 lines aren’t shown here.

net/nfc/rawsock.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/openvswitch/meter.c+8−0

This diff is large: its 15 lines aren’t shown here.

net/packet/af_packet.c+31−20

This diff is large: its 158 lines aren’t shown here.

net/qrtr/qrtr.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/rds/connection.c+17−6

This diff is large: its 30 lines aren’t shown here.

net/rds/recv.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/rds/tcp.c+2−2

This diff is large: its 11 lines aren’t shown here.

net/rds/tcp.h+2−1

This diff is large: its 10 lines aren’t shown here.

net/rds/tcp_listen.c+6−0

This diff is large: its 13 lines aren’t shown here.

net/sched/act_ct.c+17−14

This diff is large: its 61 lines aren’t shown here.

net/sched/cls_api.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/sched/sch_cake.c+11−7

This diff is large: its 81 lines aren’t shown here.

net/sched/sch_dsmark.c+2−1

This diff is large: its 10 lines aren’t shown here.

net/sched/sch_fq_pie.c+13−6

This diff is large: its 56 lines aren’t shown here.

net/sched/sch_generic.c+48−2

This diff is large: its 110 lines aren’t shown here.

net/sched/sch_htb.c+4−4

This diff is large: its 23 lines aren’t shown here.

net/sctp/socket.c+1−0

This diff is large: its 8 lines aren’t shown here.

net/sctp/sysctl.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/smc/smc_ism.c+10−1

This diff is large: its 25 lines aren’t shown here.

net/socket.c+0−13

This diff is large: its 20 lines aren’t shown here.

net/sunrpc/clnt.c+0−7

This diff is large: its 14 lines aren’t shown here.

net/sunrpc/xprt.c+32−8

This diff is large: its 88 lines aren’t shown here.

net/sunrpc/xprtrdma/rpc_rdma.c+18−9

This diff is large: its 47 lines aren’t shown here.

net/sunrpc/xprtrdma/transport.c+6−6

This diff is large: its 27 lines aren’t shown here.

net/sunrpc/xprtrdma/verbs.c+15−3

This diff is large: its 32 lines aren’t shown here.

net/sunrpc/xprtrdma/xprt_rdma.h+1−0

This diff is large: its 8 lines aren’t shown here.

net/sunrpc/xprtsock.c+2−0

This diff is large: its 9 lines aren’t shown here.

net/tipc/core.c+4−2

This diff is large: its 27 lines aren’t shown here.

net/tipc/core.h+3−7

This diff is large: its 24 lines aren’t shown here.

net/tipc/discover.c+2−2

This diff is large: its 18 lines aren’t shown here.

net/tipc/link.c+5−0

This diff is large: its 12 lines aren’t shown here.

net/tipc/link.h+1−0

This diff is large: its 6 lines aren’t shown here.

net/tipc/msg.c+2−7

This diff is large: its 21 lines aren’t shown here.

net/tipc/net.c+3−12

This diff is large: its 30 lines aren’t shown here.

net/tipc/node.c+6−6

This diff is large: its 25 lines aren’t shown here.

net/tipc/socket.c+4−1

This diff is large: its 12 lines aren’t shown here.

net/tipc/udp_media.c+2−0

This diff is large: its 16 lines aren’t shown here.

net/tls/tls_device.c+49−11

This diff is large: its 109 lines aren’t shown here.

net/tls/tls_device_fallback.c+7−0

This diff is large: its 14 lines aren’t shown here.

net/tls/tls_main.c+1−0

This diff is large: its 8 lines aren’t shown here.

net/tls/tls_sw.c+6−5

This diff is large: its 47 lines aren’t shown here.

net/unix/af_unix.c+4−3

This diff is large: its 25 lines aren’t shown here.

net/wireless/Makefile+1−1

This diff is large: its 9 lines aren’t shown here.

net/wireless/core.c+5−8

This diff is large: its 27 lines aren’t shown here.

net/wireless/pmsr.c+14−2

This diff is large: its 31 lines aren’t shown here.

net/wireless/sysfs.c+4−0

This diff is large: its 11 lines aren’t shown here.

net/wireless/util.c+8−2

This diff is large: its 38 lines aren’t shown here.

net/x25/af_x25.c+1−1

This diff is large: its 9 lines aren’t shown here.

samples/bpf/xdpsock_user.c+1−1

This diff is large: its 9 lines aren’t shown here.

samples/vfio-mdev/mdpy-fb.c+9−4

This diff is large: its 32 lines aren’t shown here.

scripts/Makefile.modfinal+1−1

This diff is large: its 9 lines aren’t shown here.

scripts/link-vmlinux.sh+1−1

This diff is large: its 9 lines aren’t shown here.

scripts/recordmcount.h+10−5

This diff is large: its 26 lines aren’t shown here.

sound/core/control_led.c+26−7

This diff is large: its 97 lines aren’t shown here.

sound/core/seq/seq_timer.c+9−1

This diff is large: its 18 lines aren’t shown here.

sound/core/timer.c+2−1

This diff is large: its 12 lines aren’t shown here.

sound/firewire/amdtp-stream.c+1−1

This diff is large: its 9 lines aren’t shown here.

sound/hda/intel-dsp-config.c+4−0

This diff is large: its 11 lines aren’t shown here.

sound/pci/hda/hda_codec.c+5−0

This diff is large: its 19 lines aren’t shown here.

sound/pci/hda/hda_generic.c+1−0

This diff is large: its 8 lines aren’t shown here.

sound/pci/hda/hda_intel.c+3−0

This diff is large: its 10 lines aren’t shown here.

sound/pci/hda/patch_cirrus.c+3−4

This diff is large: its 14 lines aren’t shown here.

sound/pci/hda/patch_realtek.c+62−1

This diff is large: its 140 lines aren’t shown here.

sound/soc/amd/raven/acp3x-pcm-dma.c+0−10

This diff is large: its 24 lines aren’t shown here.

sound/soc/amd/raven/acp3x.h+1−0

This diff is large: its 8 lines aren’t shown here.

sound/soc/amd/raven/pci-acp3x.c+15−0

This diff is large: its 36 lines aren’t shown here.

sound/soc/codecs/ak5558.c+1−1

This diff is large: its 9 lines aren’t shown here.

sound/soc/codecs/cs35l32.c+3−0

This diff is large: its 10 lines aren’t shown here.

sound/soc/codecs/cs35l33.c+1−0

This diff is large: its 8 lines aren’t shown here.

sound/soc/codecs/cs35l34.c+3−0

This diff is large: its 10 lines aren’t shown here.

sound/soc/codecs/cs42l42.c+3−0

This diff is large: its 10 lines aren’t shown here.

sound/soc/codecs/cs42l56.c+6−1

This diff is large: its 21 lines aren’t shown here.

sound/soc/codecs/cs42l73.c+3−0

This diff is large: its 10 lines aren’t shown here.

sound/soc/codecs/cs53l30.c+3−0

This diff is large: its 10 lines aren’t shown here.

sound/soc/codecs/da7219.c+1−4

This diff is large: its 12 lines aren’t shown here.

sound/soc/codecs/lpass-rx-macro.c+1−0

This diff is large: its 8 lines aren’t shown here.

sound/soc/codecs/lpass-tx-macro.c+1−0

This diff is large: its 8 lines aren’t shown here.

sound/soc/codecs/max98088.c+11−2

This diff is large: its 55 lines aren’t shown here.

sound/soc/codecs/rt5659.c+21−5

This diff is large: its 55 lines aren’t shown here.

sound/soc/codecs/rt5682-sdw.c+2−1

This diff is large: its 10 lines aren’t shown here.

sound/soc/codecs/rt711-sdca.c+2−2

This diff is large: its 16 lines aren’t shown here.

sound/soc/codecs/sti-sas.c+1−0

This diff is large: its 8 lines aren’t shown here.

sound/soc/codecs/tas2562.h+7−7

This diff is large: its 21 lines aren’t shown here.

sound/soc/fsl/Kconfig+1−0

This diff is large: its 8 lines aren’t shown here.

sound/soc/fsl/fsl-asoc-card.c+1−0

This diff is large: its 8 lines aren’t shown here.

sound/soc/generic/audio-graph-card.c+29−28

This diff is large: its 164 lines aren’t shown here.

sound/soc/generic/simple-card.c+83−85

This diff is large: its 290 lines aren’t shown here.

sound/soc/intel/boards/bytcr_rt5640.c+25−0

This diff is large: its 39 lines aren’t shown here.

sound/soc/qcom/lpass-cpu.c+80−11

This diff is large: its 135 lines aren’t shown here.

sound/soc/qcom/lpass.h+4−0

This diff is large: its 11 lines aren’t shown here.

sound/soc/soc-core.c+2−0

This diff is large: its 9 lines aren’t shown here.

sound/soc/soc-topology.c+3−3

This diff is large: its 27 lines aren’t shown here.

sound/soc/sof/intel/hda-dai.c+5−0

This diff is large: its 17 lines aren’t shown here.

sound/soc/sof/pm.c+1−0

This diff is large: its 8 lines aren’t shown here.

sound/soc/stm/stm32_sai_sub.c+1−4

This diff is large: its 12 lines aren’t shown here.

sound/usb/format.c+1−1

This diff is large: its 9 lines aren’t shown here.

sound/usb/mixer_quirks.c+1−1

This diff is large: its 9 lines aren’t shown here.

sound/usb/mixer_scarlett_gen2.c+51−30

This diff is large: its 102 lines aren’t shown here.

sound/usb/mixer_scarlett_gen2.h+1−1

This diff is large: its 8 lines aren’t shown here.

tools/arch/mips/include/uapi/asm/perf_regs.h+40−0

This diff is large: its 41 lines aren’t shown here.

tools/arch/x86/include/asm/disabled-features.h+2−5

This diff is large: its 14 lines aren’t shown here.

tools/arch/x86/include/uapi/asm/kvm.h+2−0

This diff is large: its 9 lines aren’t shown here.

tools/bootconfig/include/linux/bootconfig.h+4−0

This diff is large: its 9 lines aren’t shown here.

tools/bootconfig/main.c+1−0

This diff is large: its 8 lines aren’t shown here.

tools/bpf/bpftool/Documentation/bpftool-cgroup.rst+3−1

This diff is large: its 18 lines aren’t shown here.

tools/bpf/bpftool/Documentation/bpftool-prog.rst+1−1

This diff is large: its 9 lines aren’t shown here.

tools/bpf/bpftool/bash-completion/bpftool+3−3

This diff is large: its 27 lines aren’t shown here.

tools/bpf/bpftool/cgroup.c+2−1

This diff is large: its 10 lines aren’t shown here.

tools/bpf/bpftool/prog.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/include/uapi/asm-generic/unistd.h+1−2

This diff is large: its 10 lines aren’t shown here.

tools/include/uapi/linux/fs.h+1−1

This diff is large: its 9 lines aren’t shown here.

tools/include/uapi/linux/in.h+3−0

This diff is large: its 10 lines aren’t shown here.

tools/include/uapi/linux/kvm.h+3−2

This diff is large: its 19 lines aren’t shown here.

tools/include/uapi/linux/perf_event.h+1−1

This diff is large: its 9 lines aren’t shown here.

tools/lib/bpf/libbpf.c+3−0

This diff is large: its 10 lines aren’t shown here.

tools/lib/bpf/libbpf_internal.h+5−0

This diff is large: its 12 lines aren’t shown here.

tools/lib/bpf/xsk.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/objtool/arch/x86/decode.c+4−0

This diff is large: its 11 lines aren’t shown here.

tools/objtool/elf.c+24−1

This diff is large: its 39 lines aren’t shown here.

tools/perf/Documentation/perf-intel-pt.txt+3−3

This diff is large: its 13 lines aren’t shown here.

tools/perf/Documentation/perf-script.txt+4−3

This diff is large: its 19 lines aren’t shown here.

tools/perf/Makefile.config+0−1

This diff is large: its 8 lines aren’t shown here.

tools/perf/arch/mips/entry/syscalls/syscall_n64.tbl+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/arch/powerpc/entry/syscalls/syscall.tbl+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/arch/s390/entry/syscalls/syscall.tbl+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/arch/x86/entry/syscalls/syscall_64.tbl+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/builtin-buildid-list.c+3−0

This diff is large: its 10 lines aren’t shown here.

tools/perf/builtin-record.c+6−0

This diff is large: its 13 lines aren’t shown here.

tools/perf/builtin-stat.c+10−3

This diff is large: its 32 lines aren’t shown here.

tools/perf/check-headers.sh+1−0

This diff is large: its 8 lines aren’t shown here.

tools/perf/perf.c+2−2

This diff is large: its 18 lines aren’t shown here.

tools/perf/pmu-events/arch/powerpc/power10/cache.json+20−10

This diff is large: its 67 lines aren’t shown here.

tools/perf/pmu-events/arch/powerpc/power10/floating_point.json+1−1

This diff is large: its 8 lines aren’t shown here.

tools/perf/pmu-events/arch/powerpc/power10/frontend.json+77−47

This diff is large: its 294 lines aren’t shown here.

tools/perf/pmu-events/arch/powerpc/power10/locks.json+2−2

This diff is large: its 14 lines aren’t shown here.

tools/perf/pmu-events/arch/powerpc/power10/marked.json+28−33

This diff is large: its 175 lines aren’t shown here.

tools/perf/pmu-events/arch/powerpc/power10/memory.json+37−42

This diff is large: its 229 lines aren’t shown here.

tools/perf/pmu-events/arch/powerpc/power10/others.json+54−79

This diff is large: its 351 lines aren’t shown here.

tools/perf/pmu-events/arch/powerpc/power10/pipeline.json+65−70

This diff is large: its 362 lines aren’t shown here.

tools/perf/pmu-events/arch/powerpc/power10/pmc.json+4−4

This diff is large: its 26 lines aren’t shown here.

tools/perf/pmu-events/arch/powerpc/power10/translation.json+11−11

This diff is large: its 68 lines aren’t shown here.

tools/perf/pmu-events/jevents.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/scripts/python/exported-sql-viewer.py+9−3

This diff is large: its 33 lines aren’t shown here.

tools/perf/tests/attr/base-record+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/tests/pfm.c+2−2

This diff is large: its 11 lines aren’t shown here.

tools/perf/tests/shell/stat_bpf_counters.sh+2−2

This diff is large: its 12 lines aren’t shown here.

tools/perf/trace/beauty/include/linux/socket.h+0−2

This diff is large: its 7 lines aren’t shown here.

tools/perf/util/bpf_counter.c+6−4

This diff is large: its 34 lines aren’t shown here.

tools/perf/util/dwarf-aux.c+6−2

This diff is large: its 16 lines aren’t shown here.

tools/perf/util/env.c+1−0

This diff is large: its 8 lines aren’t shown here.

tools/perf/util/event.h+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/util/evlist.c+0−3

This diff is large: its 10 lines aren’t shown here.

tools/perf/util/evsel.c+1−0

This diff is large: its 8 lines aren’t shown here.

tools/perf/util/evsel.h+2−2

This diff is large: its 22 lines aren’t shown here.

tools/perf/util/intel-pt-decoder/intel-pt-decoder.c+5−1

This diff is large: its 21 lines aren’t shown here.

tools/perf/util/intel-pt.c+4−2

This diff is large: its 28 lines aren’t shown here.

tools/perf/util/machine.c+2−1

This diff is large: its 20 lines aren’t shown here.

tools/perf/util/metricgroup.c+8−6

This diff is large: its 44 lines aren’t shown here.

tools/perf/util/parse-events.c+11−2

This diff is large: its 28 lines aren’t shown here.

tools/perf/util/parse-events.l+1−0

This diff is large: its 8 lines aren’t shown here.

tools/perf/util/perf_api_probe.c+10−0

This diff is large: its 21 lines aren’t shown here.

tools/perf/util/perf_api_probe.h+1−0

This diff is large: its 7 lines aren’t shown here.

tools/perf/util/pfm.c+10−1

This diff is large: its 26 lines aren’t shown here.

tools/perf/util/probe-finder.c+3−0

This diff is large: its 10 lines aren’t shown here.

tools/perf/util/session.c+1−0

This diff is large: its 8 lines aren’t shown here.

tools/perf/util/stat-display.c+3−5

This diff is large: its 22 lines aren’t shown here.

tools/perf/util/symbol-elf.c+1−0

This diff is large: its 8 lines aren’t shown here.

tools/testing/selftests/bpf/network_helpers.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/testing/selftests/bpf/network_helpers.h+1−0

This diff is large: its 8 lines aren’t shown here.

tools/testing/selftests/bpf/prog_tests/ringbuf.c+48−1

This diff is large: its 64 lines aren’t shown here.

tools/testing/selftests/bpf/prog_tests/tc_redirect.c+785−0

This diff is large: its 786 lines aren’t shown here.

tools/testing/selftests/bpf/progs/test_tc_neigh.c+10−23

This diff is large: its 70 lines aren’t shown here.

tools/testing/selftests/bpf/progs/test_tc_neigh_fib.c+6−3

This diff is large: its 27 lines aren’t shown here.

tools/testing/selftests/bpf/progs/test_tc_peer.c+37−19

This diff is large: its 79 lines aren’t shown here.

tools/testing/selftests/bpf/test_tc_redirect.sh+0−216

This diff is large: its 217 lines aren’t shown here.

tools/testing/selftests/bpf/test_verifier.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/testing/selftests/bpf/verifier/and.c+2−0

This diff is large: its 9 lines aren’t shown here.

tools/testing/selftests/bpf/verifier/bounds.c+14−0

This diff is large: its 63 lines aren’t shown here.

tools/testing/selftests/bpf/verifier/dead_code.c+2−0

This diff is large: its 9 lines aren’t shown here.

tools/testing/selftests/bpf/verifier/jmp32.c+22−0

This diff is large: its 99 lines aren’t shown here.

tools/testing/selftests/bpf/verifier/jset.c+6−4

This diff is large: its 30 lines aren’t shown here.

tools/testing/selftests/bpf/verifier/stack_ptr.c+0−2

This diff is large: its 9 lines aren’t shown here.

tools/testing/selftests/bpf/verifier/unpriv.c+2−0

This diff is large: its 9 lines aren’t shown here.

tools/testing/selftests/bpf/verifier/value_ptr_arith.c+4−11

This diff is large: its 64 lines aren’t shown here.

tools/testing/selftests/kvm/.gitignore+1−0

This diff is large: its 7 lines aren’t shown here.

tools/testing/selftests/kvm/Makefile+2−1

This diff is large: its 17 lines aren’t shown here.

tools/testing/selftests/kvm/demand_paging_test.c+109−65

This diff is large: its 362 lines aren’t shown here.

tools/testing/selftests/kvm/hardware_disable_test.c+31−1

This diff is large: its 46 lines aren’t shown here.

tools/testing/selftests/kvm/include/kvm_util.h+9−5

This diff is large: its 57 lines aren’t shown here.

tools/testing/selftests/kvm/include/test_util.h+12−0

This diff is large: its 31 lines aren’t shown here.

tools/testing/selftests/kvm/kvm_page_table_test.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/testing/selftests/kvm/lib/kvm_util.c+263−67

This diff is large: its 573 lines aren’t shown here.

tools/testing/selftests/kvm/lib/kvm_util_internal.h+15−2

This diff is large: its 47 lines aren’t shown here.

tools/testing/selftests/kvm/lib/perf_test_util.c+4−2

This diff is large: its 27 lines aren’t shown here.

tools/testing/selftests/kvm/lib/rbtree.c+1−0

This diff is large: its 2 lines aren’t shown here.

tools/testing/selftests/kvm/lib/test_util.c+35−16

This diff is large: its 119 lines aren’t shown here.

tools/testing/selftests/kvm/lib/x86_64/processor.c+4−12

This diff is large: its 44 lines aren’t shown here.

tools/testing/selftests/kvm/memslot_modification_stress_test.c+11−7

This diff is large: its 38 lines aren’t shown here.

tools/testing/selftests/kvm/memslot_perf_test.c+1037−0

This diff is large: its 1038 lines aren’t shown here.

tools/testing/selftests/kvm/x86_64/get_cpuid_test.c+5−0

This diff is large: its 13 lines aren’t shown here.

tools/testing/selftests/kvm/x86_64/get_msr_index_features.c+2−6

This diff is large: its 22 lines aren’t shown here.

tools/testing/selftests/nci/.gitignore+1−0

This diff is large: its 2 lines aren’t shown here.

tools/testing/selftests/net/fib_tests.sh+25−0

This diff is large: its 38 lines aren’t shown here.

tools/testing/selftests/net/icmp.sh+74−0

This diff is large: its 75 lines aren’t shown here.

tools/testing/selftests/net/mptcp/mptcp_connect.sh+17−7

This diff is large: its 52 lines aren’t shown here.

tools/testing/selftests/net/udpgro_fwd.sh+1−1

This diff is large: its 6 lines aren’t shown here.

tools/testing/selftests/net/veth.sh+3−2

This diff is large: its 19 lines aren’t shown here.

tools/testing/selftests/netfilter/Makefile+1−1

This diff is large: its 9 lines aren’t shown here.

tools/testing/selftests/netfilter/nft_fib.sh+221−0

This diff is large: its 222 lines aren’t shown here.

tools/testing/selftests/proc/.gitignore+1−0

This diff is large: its 8 lines aren’t shown here.

tools/testing/selftests/tc-testing/tc-tests/qdiscs/fq_pie.json+4−4

This diff is large: its 16 lines aren’t shown here.

tools/testing/selftests/wireguard/netns.sh+1−0

This diff is large: its 8 lines aren’t shown here.

tools/testing/selftests/wireguard/qemu/kernel.config+0−1

This diff is large: its 8 lines aren’t shown here.

virt/kvm/kvm_main.c+4−2

This diff is large: its 27 lines aren’t shown here.

virt/lib/irqbypass.c+6−10

This diff is large: its 28 lines aren’t shown here.