Merge tag 'v4.17-rc4' into patchwork

Linux 4.17-rc4 * tag 'v4.17-rc4': (920 commits) Linux 4.17-rc4 KVM: x86: remove APIC Timer periodic/oneshot spikes genksyms: fix typo in parse.tab.{c,h} generation rules kbuild: replace hardcoded bison in cmd_bison_h with $(YACC) gcc-plugins: fix build condition of SANCOV plugin MAINTAINERS: Update Kbuild entry with a few paths Revert "usb: host: ehci: Use dma_pool_zalloc()" platform/x86: Kconfig: Fix dell-laptop dependency chain. platform/x86: asus-wireless: Fix NULL pointer dereference arm64: vgic-v2: Fix proxying of cpuif access KVM: arm/arm64: vgic_init: Cleanup reference to process_maintenance KVM: arm64: Fix order of vcpu_write_sys_reg() arguments MAINTAINERS & files: Canonize the e-mails I use at files media: imx-media-csi: Fix inconsistent IS_ERR and PTR_ERR tools: power/acpi, revert to LD = gcc bdi: Fix oops in wb_workfn() RDMA/cma: Do not query GID during QP state transition to RTR IB/mlx4: Fix integer overflow when calculating optimal MTT size IB/hfi1: Fix memory leak in exception path in get_irq_affinity() IB/{hfi1, rdmavt}: Fix memory leak in hfi1_alloc_devdata() upon failure ...

Mauro Carvalho Chehab committed 8 years ago

Commit 71db1cd2 parents 7d95fb7 75bc37f999 files · +12047 −6119Browse files

Unified · Split

999 files

Documentation/bpf/bpf_devel_QA.txt +9 −1
Documentation/core-api/kernel-api.rst +13 −0
Documentation/devicetree/bindings/input/atmel,maxtouch.txt +7 −0
Documentation/devicetree/bindings/serial/amlogic,meson-uart.txt +1 −1
Documentation/devicetree/bindings/serial/mvebu-uart.txt +1 −1
Documentation/devicetree/bindings/serial/renesas,sci-serial.txt +2 −0
Documentation/devicetree/bindings/thermal/exynos-thermal.txt +5 −18
Documentation/devicetree/bindings/thermal/thermal.txt +1 −15
Documentation/devicetree/bindings/timer/nuvoton,npcm7xx-timer.txt +21 −0
Documentation/devicetree/bindings/timer/nxp,tpm-timer.txt +1 −1
Documentation/devicetree/bindings/usb/usb-xhci.txt +4 −1
Documentation/doc-guide/parse-headers.rst +2 −2
Documentation/driver-api/firmware/request_firmware.rst +8 −8
Documentation/driver-api/infrastructure.rst +1 −1
Documentation/driver-api/usb/typec.rst +1 −1
Documentation/i2c/dev-interface +14 −18
Documentation/ioctl/ioctl-number.txt +0 −2
Documentation/livepatch/shadow-vars.txt +29 −12
Documentation/media/uapi/rc/keytable.c.rst +1 −1
Documentation/media/uapi/v4l/v4l2grab.c.rst +1 −1
Documentation/networking/filter.txt +3 −3
Documentation/networking/ip-sysctl.txt +4 −19
Documentation/power/suspend-and-cpuhotplug.txt +1 −1
Documentation/process/magic-number.rst +0 −3
Documentation/sphinx/parse-headers.pl +2 −2
Documentation/trace/ftrace.rst +11 −3
Documentation/translations/zh_CN/video4linux/v4l2-framework.txt +2 −2
Documentation/virtual/kvm/api.txt +8 −1
Documentation/virtual/kvm/arm/psci.txt +30 −0
MAINTAINERS +49 −43
Makefile +2 −2
arch/arm/boot/dts/gemini-nas4220b.dts +14 −14
arch/arm/boot/dts/omap4.dtsi +4 −4
arch/arm/configs/gemini_defconfig +25 −2
arch/arm/configs/socfpga_defconfig +1 −0
arch/arm/include/asm/kvm_host.h +3 −0
arch/arm/include/uapi/asm/kvm.h +6 −0
arch/arm/kvm/guest.c +13 −0
arch/arm/mach-omap2/Makefile +1 −5
arch/arm/mach-omap2/pm-asm-offsets.c +3 −0
arch/arm/mach-omap2/sleep33xx.S +0 −1
arch/arm/mach-omap2/sleep43xx.S +0 −1
arch/arm/mach-s3c24xx/mach-jive.c +2 −2
arch/arm64/Makefile +4 −0
arch/arm64/boot/dts/amlogic/meson-gx-p23x-q20x.dtsi +4 −0
arch/arm64/boot/dts/amlogic/meson-gxl-s905x-libretech-cc.dts +12 −0
arch/arm64/boot/dts/amlogic/meson-gxl-s905x-nexbox-a95x.dts +4 −0
arch/arm64/boot/dts/amlogic/meson-gxl-s905x-p212.dtsi +4 −0
arch/arm64/boot/dts/amlogic/meson-gxl.dtsi +61 −0
arch/arm64/boot/dts/amlogic/meson-gxm-khadas-vim2.dts +4 −0
arch/arm64/boot/dts/amlogic/meson-gxm.dtsi +17 −0
arch/arm64/boot/dts/arm/juno-motherboard.dtsi +0 −2
arch/arm64/boot/dts/broadcom/stingray/stingray-sata.dtsi +40 −40
arch/arm64/include/asm/kvm_emulate.h +1 −1
arch/arm64/include/asm/kvm_host.h +3 −0
arch/arm64/include/asm/module.h +1 −1
arch/arm64/include/asm/pgtable.h +2 −2
arch/arm64/include/uapi/asm/kvm.h +6 −0
arch/arm64/kernel/cpufeature.c +1 −0
arch/arm64/kernel/module-plts.c +1 −1
arch/arm64/kernel/module.c +1 −1
arch/arm64/kernel/ptrace.c +10 −10
arch/arm64/kernel/traps.c +3 −2
arch/arm64/kvm/guest.c +13 −1
arch/arm64/kvm/hyp/vgic-v2-cpuif-proxy.c +19 −5
arch/arm64/kvm/sys_regs.c +2 −4
arch/arm64/lib/Makefile +4 −0
arch/arm64/mm/flush.c +1 −1
arch/arm64/mm/kasan_init.c +2 −2
arch/hexagon/include/asm/io.h +6 −0
arch/hexagon/lib/checksum.c +1 −0
arch/mips/boot/dts/img/boston.dts +6 −0
arch/mips/include/asm/io.h +3 −1
arch/mips/include/asm/uaccess.h +9 −2
arch/mips/lib/memset.S +8 −3
arch/parisc/Makefile +3 −0
arch/parisc/kernel/Makefile +1 −1
arch/parisc/kernel/drivers.c +4 −3
arch/parisc/kernel/pci.c +1 −1
arch/parisc/kernel/time.c +1 −1
arch/parisc/kernel/traps.c +11 −0
arch/parisc/mm/init.c +1 −1
arch/powerpc/include/asm/powernv.h +1 −1
arch/powerpc/kernel/eeh_pe.c +2 −1
arch/powerpc/kernel/idle_book3s.S +2 −2
arch/powerpc/kernel/mce_power.c +2 −5
arch/powerpc/kernel/setup_64.c +11 −0
arch/powerpc/kernel/smp.c +42 −7
arch/powerpc/kvm/booke.c +7 −0
arch/powerpc/lib/feature-fixups.c +1 −1
arch/powerpc/mm/mem.c +2 −0
arch/powerpc/platforms/cell/spufs/sched.c +1 −1
arch/powerpc/platforms/powernv/memtrace.c +0 −17
arch/powerpc/platforms/powernv/npu-dma.c +73 −15
arch/powerpc/platforms/powernv/opal-rtc.c +5 −3
arch/powerpc/sysdev/xive/native.c +4 −0
arch/riscv/Kconfig +1 −3
arch/riscv/include/asm/Kbuild +0 −1
arch/riscv/kernel/vdso/Makefile +1 −1
arch/s390/Kbuild +1 −0
arch/s390/Kconfig +24 −8
arch/s390/boot/Makefile +0 −6
arch/s390/boot/compressed/.gitignore +1 −0
arch/s390/configs/debug_defconfig +20 −10
arch/s390/configs/gcov_defconfig +0 −661
arch/s390/configs/performance_defconfig +15 −5
arch/s390/defconfig +11 −2
arch/s390/hypfs/inode.c +1 −1
arch/s390/include/asm/kexec.h +23 −0
arch/s390/include/asm/purgatory.h +17 −0
arch/s390/include/asm/setup.h +27 −13
arch/s390/include/asm/thread_info.h +3 −0
arch/s390/include/uapi/asm/signal.h +16 −7
arch/s390/kernel/Makefile +3 −0
arch/s390/kernel/asm-offsets.c +5 −0
arch/s390/kernel/compat_wrapper.c +1 −0
arch/s390/kernel/kexec_elf.c +147 −0
arch/s390/kernel/kexec_image.c +76 −0
arch/s390/kernel/machine_kexec_file.c +245 −0
arch/s390/kernel/module.c +2 −2
arch/s390/kernel/nospec-branch.c +1 −0
arch/s390/kernel/perf_cpum_cf_events.c +5 −4
arch/s390/kernel/process.c +10 −0
arch/s390/kernel/setup.c +1 −0
arch/s390/kernel/syscalls/syscall.tbl +1 −0
arch/s390/kernel/uprobes.c +9 −0
arch/s390/purgatory/.gitignore +2 −0
arch/s390/purgatory/Makefile +37 −0
arch/s390/purgatory/head.S +279 −0
arch/s390/purgatory/purgatory.c +42 −0
arch/sparc/include/uapi/asm/oradax.h +1 −1
arch/sparc/kernel/vio.c +1 −1
arch/x86/Kconfig +4 −0
arch/x86/entry/entry_64_compat.S +4 −4
arch/x86/events/intel/core.c +6 −3
arch/x86/events/intel/uncore_snbep.c +37 −0
arch/x86/include/asm/asm.h +0 −2
arch/x86/include/asm/cpufeatures.h +1 −0
arch/x86/include/asm/ftrace.h +17 −2
arch/x86/include/asm/irq_vectors.h +0 −7
arch/x86/include/asm/jailhouse_para.h +1 −1
arch/x86/include/asm/kvm_host.h +1 −0
arch/x86/include/asm/pgtable.h +5 −0
arch/x86/include/asm/pgtable_64_types.h +4 −4
arch/x86/include/asm/processor.h +0 −2
arch/x86/include/uapi/asm/msgbuf.h +31 −0
arch/x86/include/uapi/asm/shmbuf.h +42 −0
arch/x86/kernel/acpi/boot.c +4 −0
arch/x86/kernel/cpu/common.c +5 −1
arch/x86/kernel/cpu/intel.c +3 −0
arch/x86/kernel/cpu/microcode/core.c +2 −4
arch/x86/kernel/cpu/microcode/intel.c +0 −2
arch/x86/kernel/jailhouse.c +1 −1
arch/x86/kernel/kexec-bzimage64.c +2 −3
arch/x86/kernel/ldt.c +1 −1
arch/x86/kernel/pci-nommu.c +0 −90
arch/x86/kernel/setup.c +6 −0
arch/x86/kernel/smpboot.c +42 −5
arch/x86/kernel/tsc.c +12 −12
arch/x86/kvm/lapic.c +20 −17
arch/x86/kvm/svm.c +18 −13
arch/x86/kvm/vmx.c +59 −50
arch/x86/kvm/x86.c +12 −3
arch/x86/kvm/x86.h +0 −7
arch/x86/mm/dump_pagetables.c +6 −5
arch/x86/mm/pageattr.c +32 −12
arch/x86/mm/pti.c +23 −3
arch/x86/net/bpf_jit_comp.c +14 −4
arch/x86/power/hibernate_64.c +1 −1
arch/x86/xen/enlighten_pv.c +31 −55
block/bfq-iosched.c +9 −1
block/blk-cgroup.c +12 −16
block/blk-core.c +8 −7
block/blk-mq.c +66 −15
block/blk-mq.h +6 −1
block/genhd.c +12 −0
block/partition-generic.c +6 −4
crypto/api.c +8 −3
crypto/drbg.c +2 −0
drivers/acpi/acpi_video.c +25 −2
drivers/acpi/acpi_watchdog.c +49 −10
drivers/acpi/button.c +23 −1
drivers/acpi/scan.c +1 −1
drivers/acpi/sleep.c +13 −0
drivers/amba/bus.c +11 −6
drivers/android/binder.c +8 −0
drivers/atm/iphase.c +2 −2
drivers/base/dma-coherent.c +3 −2
drivers/base/dma-mapping.c +2 −4
drivers/base/firmware_loader/fallback.c +2 −2
drivers/base/firmware_loader/fallback.h +1 −1
drivers/block/loop.c +43 −21
drivers/block/loop.h +0 −1
drivers/block/rbd.c +69 −32
drivers/block/swim.c +22 −27
drivers/block/swim3.c +3 −3
drivers/bus/Kconfig +1 −0
drivers/cdrom/cdrom.c +1 −1
drivers/char/random.c +137 −35
drivers/char/virtio_console.c +71 −86
drivers/clk/clk-cs2000-cp.c +1 −1
drivers/clk/clk-mux.c +9 −1
drivers/clk/clk-stm32mp1.c +23 −31
drivers/clk/clk.c +4 −3
drivers/clk/meson/clk-regmap.c +10 −1
drivers/clk/meson/gxbb-aoclk.h +0 −2
drivers/clk/meson/meson8b.c +3 −2
drivers/clocksource/Kconfig +8 −0
drivers/clocksource/Makefile +1 −0
drivers/clocksource/timer-imx-tpm.c +34 −11
drivers/clocksource/timer-npcm7xx.c +215 −0
drivers/cpufreq/Kconfig.arm +0 −10
drivers/cpufreq/brcmstb-avs-cpufreq.c +1 −322
drivers/cpufreq/cppc_cpufreq.c +44 −2
drivers/cpufreq/powernv-cpufreq.c +11 −3
drivers/dax/device.c +2 −0
drivers/firmware/arm_scmi/clock.c +1 −1
drivers/fpga/altera-ps-spi.c +1 −1
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +5 −2
drivers/gpu/drm/amd/amdkfd/Kconfig +1 −0
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c +9 −8
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +9 −1
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_color.c +0 −7
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c +3 −2
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c +22 −32
drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c +10 −6
drivers/gpu/drm/amd/powerplay/inc/vega12/smu9_driver_if.h +3 −1
drivers/gpu/drm/bridge/dumb-vga-dac.c +3 −1
drivers/gpu/drm/drm_dp_dual_mode_helper.c +32 −7
drivers/gpu/drm/drm_edid.c +3 −8
drivers/gpu/drm/exynos/exynos_drm_fb.c +14 −59
drivers/gpu/drm/i915/gvt/cmd_parser.c +1 −0
drivers/gpu/drm/i915/gvt/display.c +10 −0
drivers/gpu/drm/i915/gvt/dmabuf.c +1 −0
drivers/gpu/drm/i915/gvt/fb_decoder.c +9 −18
drivers/gpu/drm/i915/gvt/gtt.c +45 −7
drivers/gpu/drm/i915/gvt/gtt.h +1 −1
drivers/gpu/drm/i915/gvt/handlers.c +1 −0
drivers/gpu/drm/i915/gvt/kvmgt.c +1 −1
drivers/gpu/drm/i915/i915_drv.c +15 −12
drivers/gpu/drm/i915/i915_gem_execbuffer.c +1 −1
drivers/gpu/drm/i915/i915_pmu.c +27 −10
drivers/gpu/drm/i915/intel_audio.c +1 −1
drivers/gpu/drm/i915/intel_bios.c +9 −4
drivers/gpu/drm/i915/intel_cdclk.c +14 −2
drivers/gpu/drm/i915/intel_csr.c +1 −0
drivers/gpu/drm/i915/intel_drv.h +2 −2
drivers/gpu/drm/i915/intel_fbdev.c +1 −1
drivers/gpu/drm/i915/intel_lrc.c +9 −0
drivers/gpu/drm/i915/intel_runtime_pm.c +5 −6
drivers/gpu/drm/msm/disp/mdp4/mdp4_crtc.c +1 −0
drivers/gpu/drm/msm/disp/mdp5/mdp5_crtc.c +1 −0
drivers/gpu/drm/msm/disp/mdp_format.c +2 −1
drivers/gpu/drm/msm/disp/mdp_kms.h +1 −1
drivers/gpu/drm/msm/dsi/dsi_host.c +12 −4
drivers/gpu/drm/msm/dsi/phy/dsi_phy.c +109 −0
drivers/gpu/drm/msm/dsi/phy/dsi_phy.h +2 −0
drivers/gpu/drm/msm/dsi/phy/dsi_phy_10nm.c +0 −28
drivers/gpu/drm/msm/msm_fb.c +2 −1
drivers/gpu/drm/msm/msm_fbdev.c +2 −9
drivers/gpu/drm/msm/msm_gem.c +11 −9
drivers/gpu/drm/msm/msm_kms.h +4 −1
drivers/gpu/drm/qxl/qxl_cmd.c +2 −4
drivers/gpu/drm/qxl/qxl_drv.h +1 −0
drivers/gpu/drm/qxl/qxl_ioctl.c +2 −2
drivers/gpu/drm/qxl/qxl_release.c +9 −9
drivers/gpu/drm/sun4i/sun4i_lvds.c +0 −55
drivers/gpu/drm/vc4/vc4_bo.c +2 −0
drivers/gpu/drm/vc4/vc4_crtc.c +45 −1
drivers/gpu/drm/vc4/vc4_validate_shaders.c +1 −0
drivers/gpu/drm/virtio/virtgpu_vq.c +2 −2
drivers/gpu/drm/vmwgfx/vmwgfx_fb.c +10 −21
drivers/gpu/drm/vmwgfx/vmwgfx_kms.c +11 −3
drivers/hid/hid-ids.h +3 −0
drivers/hid/hid-input.c +17 −7
drivers/hid/hidraw.c +5 −0
drivers/hid/i2c-hid/i2c-hid.c +13 −0
drivers/hid/wacom_wac.c +46 −30
drivers/hwmon/k10temp.c +14 −3
drivers/hwmon/nct6683.c +2 −2
drivers/hwmon/scmi-hwmon.c +4 −1
drivers/i2c/busses/Kconfig +0 −3
drivers/i2c/busses/i2c-sprd.c +18 −4
drivers/i2c/i2c-dev.c +1 −1
drivers/infiniband/Kconfig +4 −1
drivers/infiniband/core/cache.c +35 −20
drivers/infiniband/core/cma.c +43 −17
drivers/infiniband/core/iwpm_util.c +4 −1
drivers/infiniband/core/mad.c +2 −2
drivers/infiniband/core/roce_gid_mgmt.c +14 −12
drivers/infiniband/core/ucma.c +28 −16
drivers/infiniband/core/uverbs_cmd.c +6 −0
drivers/infiniband/core/uverbs_ioctl.c +9 −0
drivers/infiniband/core/uverbs_std_types_flow_action.c +6 −6
drivers/infiniband/core/verbs.c +1 −0
drivers/infiniband/hw/cxgb4/cq.c +10 −1
drivers/infiniband/hw/cxgb4/device.c +8 −1
drivers/infiniband/hw/cxgb4/iw_cxgb4.h +5 −1
drivers/infiniband/hw/cxgb4/qp.c +2 −2
drivers/infiniband/hw/cxgb4/resource.c +24 −2
drivers/infiniband/hw/hfi1/affinity.c +5 −6
drivers/infiniband/hw/hfi1/driver.c +15 −4
drivers/infiniband/hw/hfi1/hfi.h +4 −4
drivers/infiniband/hw/hfi1/init.c +31 −12
drivers/infiniband/hw/hfi1/pcie.c +0 −3
drivers/infiniband/hw/hfi1/platform.c +1 −0
drivers/infiniband/hw/hfi1/qsfp.c +2 −0
drivers/infiniband/hw/hfi1/ruc.c +40 −10
drivers/infiniband/hw/hfi1/ud.c +2 −2
drivers/infiniband/hw/hns/hns_roce_hem.c +6 −6
drivers/infiniband/hw/hns/hns_roce_hw_v2.c +29 −20
drivers/infiniband/hw/hns/hns_roce_qp.c +1 −1
drivers/infiniband/hw/mlx4/mr.c +1 −1
drivers/infiniband/hw/mlx4/qp.c +2 −1
drivers/infiniband/hw/mlx5/Kconfig +1 −0
drivers/infiniband/hw/mlx5/main.c +3 −6
drivers/infiniband/hw/mlx5/mr.c +23 −9
drivers/infiniband/hw/mlx5/qp.c +13 −9
drivers/infiniband/hw/nes/nes_nic.c +1 −1
drivers/infiniband/sw/rxe/rxe_opcode.c +1 −1
drivers/infiniband/sw/rxe/rxe_req.c +0 −1
drivers/infiniband/sw/rxe/rxe_resp.c +1 −5
drivers/infiniband/ulp/ipoib/ipoib_main.c +1 −1
drivers/infiniband/ulp/srp/Kconfig +1 −1
drivers/infiniband/ulp/srpt/Kconfig +1 −1
drivers/input/evdev.c +6 −1
drivers/input/input-leds.c +4 −4
drivers/input/mouse/alps.c +1 −1
drivers/input/rmi4/rmi_spi.c +5 −2
drivers/input/touchscreen/Kconfig +1 −1
drivers/input/touchscreen/atmel_mxt_ts.c +124 −76
drivers/iommu/amd_iommu.c +1 −1
drivers/iommu/dma-iommu.c +25 −29
drivers/iommu/dmar.c +1 −1
drivers/iommu/intel_irq_remapping.c +1 −1
drivers/iommu/rockchip-iommu.c +9 −2
drivers/irqchip/qcom-irq-combiner.c +2 −2
drivers/isdn/mISDN/dsp_hwec.c +5 −3
drivers/isdn/mISDN/l1oip_core.c +11 −3
drivers/md/bcache/alloc.c +4 −1
drivers/md/bcache/bcache.h +4 −0
drivers/md/bcache/debug.c +1 −2
drivers/md/bcache/io.c +3 −5
drivers/md/bcache/request.c +1 −4
drivers/md/bcache/super.c +52 −23
drivers/md/bcache/writeback.c +3 −1
drivers/md/md.c +4 −2
drivers/md/raid1.c +20 −5
drivers/media/i2c/saa7115.c +1 −1
drivers/media/i2c/saa711x_regs.h +1 −1
drivers/media/i2c/tda7432.c +1 −1
drivers/media/i2c/tvp5150.c +1 −1
drivers/media/i2c/tvp5150_reg.h +1 −1
drivers/media/i2c/tvp7002.c +1 −1
drivers/media/i2c/tvp7002_reg.h +1 −1
drivers/media/media-devnode.c +1 −1
drivers/media/pci/bt8xx/bttv-audio-hook.c +1 −1
drivers/media/pci/bt8xx/bttv-audio-hook.h +1 −1
drivers/media/pci/bt8xx/bttv-cards.c +2 −2
drivers/media/pci/bt8xx/bttv-driver.c +1 −1
drivers/media/pci/bt8xx/bttv-i2c.c +1 −1
drivers/media/pci/cx23885/cx23885-input.c +1 −1
drivers/media/pci/cx88/cx88-alsa.c +2 −2
drivers/media/pci/cx88/cx88-blackbird.c +1 −1
drivers/media/pci/cx88/cx88-core.c +1 −1
drivers/media/pci/cx88/cx88-i2c.c +1 −1
drivers/media/pci/cx88/cx88-video.c +1 −1
drivers/media/radio/radio-aimslab.c +1 −1
drivers/media/radio/radio-aztech.c +1 −1
drivers/media/radio/radio-gemtek.c +1 −1
drivers/media/radio/radio-maxiradio.c +1 −1
drivers/media/radio/radio-rtrack2.c +1 −1
drivers/media/radio/radio-sf16fmi.c +1 −1
drivers/media/radio/radio-terratec.c +1 −1
drivers/media/radio/radio-trust.c +1 −1
drivers/media/radio/radio-typhoon.c +1 −1
drivers/media/radio/radio-zoltrix.c +1 −1
drivers/media/rc/keymaps/rc-avermedia-m135a.c +1 −1
drivers/media/rc/keymaps/rc-encore-enltv-fm53.c +1 −1
drivers/media/rc/keymaps/rc-encore-enltv2.c +1 −1
drivers/media/rc/keymaps/rc-kaiomy.c +1 −1
drivers/media/rc/keymaps/rc-kworld-plus-tv-analog.c +1 −1
drivers/media/rc/keymaps/rc-pixelview-new.c +1 −1
drivers/media/tuners/tea5761.c +2 −2
drivers/media/tuners/tea5767.c +2 −2
drivers/media/tuners/tuner-xc2028-types.h +1 −1
drivers/media/tuners/tuner-xc2028.c +2 −2
drivers/media/tuners/tuner-xc2028.h +1 −1
drivers/media/usb/em28xx/em28xx-camera.c +1 −1
drivers/media/usb/em28xx/em28xx-cards.c +1 −1
drivers/media/usb/em28xx/em28xx-core.c +2 −2
drivers/media/usb/em28xx/em28xx-dvb.c +2 −2
drivers/media/usb/em28xx/em28xx-i2c.c +1 −1
drivers/media/usb/em28xx/em28xx-input.c +1 −1
drivers/media/usb/em28xx/em28xx-video.c +2 −2
drivers/media/usb/em28xx/em28xx.h +1 −1
drivers/media/usb/gspca/zc3xx-reg.h +1 −1
drivers/media/usb/tm6000/tm6000-cards.c +1 −1
drivers/media/usb/tm6000/tm6000-core.c +1 −1
drivers/media/usb/tm6000/tm6000-i2c.c +1 −1
drivers/media/usb/tm6000/tm6000-regs.h +1 −1
drivers/media/usb/tm6000/tm6000-usb-isoc.h +1 −1
drivers/media/usb/tm6000/tm6000-video.c +1 −1
drivers/media/usb/tm6000/tm6000.h +1 −1
drivers/media/v4l2-core/v4l2-dev.c +2 −2
drivers/media/v4l2-core/v4l2-ioctl.c +1 −1
drivers/media/v4l2-core/videobuf-core.c +3 −3
drivers/media/v4l2-core/videobuf-dma-contig.c +1 −1
drivers/media/v4l2-core/videobuf-dma-sg.c +3 −3
drivers/media/v4l2-core/videobuf-vmalloc.c +2 −2
drivers/memory/emif-asm-offsets.c +1 −71
drivers/message/fusion/mptsas.c +1 −0
drivers/mmc/host/renesas_sdhi_internal_dmac.c +33 −6
drivers/mmc/host/sdhci-pci-core.c +23 −2
drivers/mtd/chips/cfi_cmdset_0001.c +28 −5
drivers/mtd/chips/cfi_cmdset_0002.c +6 −3
drivers/mtd/nand/core.c +0 −3
drivers/mtd/nand/raw/marvell_nand.c +8 −17
drivers/mtd/nand/raw/tango_nand.c +1 −1
drivers/mtd/spi-nor/cadence-quadspi.c +17 −2
drivers/net/bonding/bond_main.c +1 −2
drivers/net/dsa/mv88e6xxx/hwtstamp.c +10 −2
drivers/net/ethernet/amd/xgbe/xgbe-common.h +8 −0
drivers/net/ethernet/amd/xgbe/xgbe-debugfs.c +16 −0
drivers/net/ethernet/amd/xgbe/xgbe-main.c +1 −0
drivers/net/ethernet/amd/xgbe/xgbe-mdio.c +19 −5
drivers/net/ethernet/amd/xgbe/xgbe-pci.c +2 −0
drivers/net/ethernet/amd/xgbe/xgbe-phy-v2.c +178 −18
drivers/net/ethernet/amd/xgbe/xgbe.h +9 −0
drivers/net/ethernet/broadcom/bcmsysport.c +13 −5
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c +27 −22
drivers/net/ethernet/broadcom/bnxt/bnxt_nvm_defs.h +0 −2
drivers/net/ethernet/freescale/ucc_geth_ethtool.c +1 −1
drivers/net/ethernet/hisilicon/hns/hnae.h +1 −1
drivers/net/ethernet/ibm/ibmvnic.c +56 −33
drivers/net/ethernet/ibm/ibmvnic.h +0 −1
drivers/net/ethernet/intel/ice/ice_adminq_cmd.h +1 −1
drivers/net/ethernet/intel/ice/ice_common.c +17 −5
drivers/net/ethernet/intel/ice/ice_hw_autogen.h +0 −2
drivers/net/ethernet/intel/ice/ice_main.c +0 −4
drivers/net/ethernet/intel/ice/ice_sched.c +2 −2
drivers/net/ethernet/intel/igb/igb_main.c +16 −1
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c +1 −1
drivers/net/ethernet/marvell/mvpp2.c +31 −13
drivers/net/ethernet/mellanox/mlx4/main.c +1 −1
drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c +5 −3
drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +3 −2
drivers/net/ethernet/mellanox/mlx5/core/en_selftest.c +1 −1
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +2 −1
drivers/net/ethernet/mellanox/mlx5/core/en_tx.c +10 −10
drivers/net/ethernet/mellanox/mlx5/core/fs_core.c +16 −10
drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c +5 −7
drivers/net/ethernet/netronome/nfp/flower/action.c +8 −2
drivers/net/ethernet/netronome/nfp/flower/cmsg.c +37 −7
drivers/net/ethernet/netronome/nfp/flower/cmsg.h +6 −1
drivers/net/ethernet/netronome/nfp/flower/main.c +5 −3
drivers/net/ethernet/netronome/nfp/flower/main.h +6 −2
drivers/net/ethernet/netronome/nfp/nfp_app_nic.c +1 −1
drivers/net/ethernet/netronome/nfp/nfp_main.h +3 −1
drivers/net/ethernet/netronome/nfp/nfp_net_main.c +18 −13
drivers/net/ethernet/netronome/nfp/nfpcore/nfp_mutex.c +4 −1
drivers/net/ethernet/netronome/nfp/nfpcore/nfp_nsp.c +1 −2
drivers/net/ethernet/qlogic/qed/qed_ll2.c +1 −1
drivers/net/ethernet/qlogic/qed/qed_roce.c +1 −1
drivers/net/ethernet/qualcomm/rmnet/rmnet_config.c +6 −5
drivers/net/ethernet/realtek/8139too.c +1 −1
drivers/net/ethernet/sfc/ef10.c +50 −38
drivers/net/ethernet/sfc/efx.c +143 −0
drivers/net/ethernet/sfc/efx.h +21 −0
drivers/net/ethernet/sfc/farch.c +34 −7
drivers/net/ethernet/sfc/net_driver.h +61 −0
drivers/net/ethernet/sfc/rx.c +90 −34
drivers/net/ethernet/stmicro/stmmac/dwmac4.h +1 −1
drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c +0 −7
drivers/net/ethernet/stmicro/stmmac/stmmac_main.c +6 −1
drivers/net/ethernet/ti/cpsw.c +3 −1
drivers/net/macsec.c +2 −3
drivers/net/phy/marvell.c +9 −0
drivers/net/phy/microchip.c +177 −1
drivers/net/phy/phy_device.c +10 −1
drivers/net/ppp/pppoe.c +4 −0
drivers/net/team/team.c +31 −7
drivers/net/tun.c +1 −6
drivers/net/usb/qmi_wwan.c +14 −0
drivers/net/virtio_net.c +48 −31
drivers/net/vmxnet3/vmxnet3_drv.c +13 −4
drivers/net/vmxnet3/vmxnet3_int.h +2 −2
drivers/net/wireless/broadcom/brcm80211/brcmfmac/firmware.c +20 −16
drivers/net/wireless/intel/iwlwifi/fw/api/scan.h +5 −8
drivers/net/wireless/intel/iwlwifi/iwl-nvm-parse.c +95 −16
drivers/net/wireless/intel/iwlwifi/iwl-nvm-parse.h +4 −2
drivers/net/wireless/intel/iwlwifi/mvm/mac80211.c +2 −1
drivers/net/wireless/realtek/rtlwifi/btcoexist/halbtcoutsrc.c +0 −15
drivers/net/wireless/realtek/rtlwifi/rtl8723be/hw.c +7 −4
drivers/net/wireless/realtek/rtlwifi/wifi.h +5 −0
drivers/nvdimm/Kconfig +1 −2
drivers/nvdimm/dimm_devs.c +12 −10
drivers/nvdimm/of_pmem.c +1 −1
drivers/nvme/host/Kconfig +1 −1
drivers/nvme/host/core.c +2 −25
drivers/nvme/host/fabrics.c +6 −0
drivers/nvme/host/multipath.c +23 −1
drivers/nvme/host/nvme.h +12 −0
drivers/nvme/target/Kconfig +1 −1
drivers/nvme/target/loop.c +6 −0
drivers/of/fdt.c +5 −2
drivers/parisc/ccio-dma.c +1 −1
drivers/pci/dwc/pcie-kirin.c +1 −1
drivers/pci/host/pci-aardvark.c +30 −23
drivers/pci/pci-driver.c +3 −2
drivers/pci/pci.c +2 −2
drivers/platform/x86/Kconfig +1 −1
drivers/platform/x86/asus-wireless.c +3 −1
drivers/rapidio/devices/rio_mport_cdev.c +9 −10
drivers/remoteproc/qcom_q6v5_pil.c +2 −0
drivers/remoteproc/remoteproc_core.c +2 −2
drivers/rpmsg/rpmsg_char.c +2 −0
drivers/rtc/rtc-opal.c +23 −14
drivers/s390/block/dasd_alias.c +11 −2
drivers/s390/block/dasd_diag.c +0 −1
drivers/s390/char/sclp_early_core.c +1 −1
drivers/s390/cio/chsc.c +11 −3
drivers/s390/cio/vfio_ccw_fsm.c +12 −7
drivers/s390/net/qeth_core.h +0 −2
drivers/s390/net/qeth_core_main.c +71 −87
drivers/s390/net/qeth_core_mpc.h +12 −0
drivers/s390/net/qeth_l2_main.c +33 −27
drivers/s390/net/smsgiucv.c +1 −1
drivers/sbus/char/oradax.c +1 −1
drivers/scsi/fnic/fnic_trace.c +1 −1
drivers/scsi/isci/port_config.c +1 −2
drivers/scsi/megaraid/megaraid_sas_fusion.c +3 −3
drivers/scsi/scsi_debug.c +24 −9
drivers/scsi/scsi_transport_iscsi.c +18 −11
drivers/scsi/sd.c +2 −0
drivers/scsi/sd_zbc.c +82 −58
drivers/scsi/storvsc_drv.c +5 −2
drivers/scsi/ufs/ufshcd.c +40 −0
drivers/slimbus/messaging.c +1 −1
drivers/soc/bcm/raspberrypi-power.c +1 −1
drivers/staging/media/imx/imx-media-csi.c +1 −1
drivers/staging/wilc1000/host_interface.c +1 −1
drivers/target/target_core_iblock.c +4 −4
drivers/target/target_core_pscsi.c +2 −0
drivers/tty/n_gsm.c +22 −1
drivers/tty/serial/earlycon.c +4 −2
drivers/tty/serial/imx.c +18 −1
drivers/tty/serial/mvebu-uart.c +0 −1
drivers/tty/serial/qcom_geni_serial.c +6 −4
drivers/tty/serial/xilinx_uartps.c +1 −1
drivers/tty/tty_io.c +4 −1
drivers/tty/tty_ldisc.c +13 −16
drivers/uio/uio_hv_generic.c +23 −49
drivers/usb/Kconfig +1 −0
drivers/usb/core/config.c +3 −1
drivers/usb/core/hcd.c +13 −6
drivers/usb/core/hub.c +9 −1
drivers/usb/core/phy.c +66 −27
drivers/usb/core/phy.h +21 −1
drivers/usb/core/quirks.c +3 −0
drivers/usb/dwc2/core.h +2 −0
drivers/usb/dwc2/gadget.c +21 −0
drivers/usb/dwc2/hcd.c +8 −5
drivers/usb/dwc2/pci.c +3 −1
drivers/usb/dwc3/gadget.c +2 −2
drivers/usb/gadget/function/f_phonet.c +1 −1
drivers/usb/host/ehci-mem.c +2 −1
drivers/usb/host/ehci-sched.c +4 −2
drivers/usb/host/xhci-dbgtty.c +5 −3
drivers/usb/host/xhci-pci.c +4 −1
drivers/usb/host/xhci-plat.c +23 −9
drivers/usb/host/xhci.c +1 −0
drivers/usb/host/xhci.h +2 −1
drivers/usb/musb/musb_dsps.c +0 −2
drivers/usb/musb/musb_gadget.c +2 −1
drivers/usb/musb/musb_host.c +4 −1
drivers/usb/serial/Kconfig +1 −0
drivers/usb/serial/cp210x.c +1 −0
drivers/usb/serial/ftdi_sio.c +2 −1
drivers/usb/serial/option.c +5 −0
drivers/usb/serial/usb-serial-simple.c +7 −0
drivers/usb/serial/visor.c +35 −34
drivers/usb/typec/tcpm.c +1 −0
drivers/usb/typec/tps6598x.c +39 −8
drivers/usb/typec/ucsi/Makefile +1 −1
drivers/usb/typec/ucsi/ucsi.c +1 −1
drivers/usb/usbip/stub_main.c +5 −0
drivers/usb/usbip/usbip_common.h +1 −1
drivers/usb/usbip/usbip_event.c +0 −4
drivers/usb/usbip/vhci_hcd.c +13 −0
drivers/virt/vboxguest/vboxguest_core.c +38 −32
drivers/virt/vboxguest/vboxguest_core.h +9 −0
drivers/virt/vboxguest/vboxguest_linux.c +16 −3
drivers/virt/vboxguest/vboxguest_utils.c +13 −4
drivers/watchdog/aspeed_wdt.c +7 −0
drivers/watchdog/renesas_wdt.c +4 −2
drivers/watchdog/sch311x_wdt.c +1 −1
drivers/watchdog/w83977f_wdt.c +1 −1
drivers/watchdog/wafer5823wdt.c +1 −1
drivers/xen/xen-pciback/conf_space_quirks.c +1 −1
drivers/xen/xen-pciback/pci_stub.c +4 −4
drivers/xen/xenbus/xenbus_dev_frontend.c +2 −1
fs/afs/server.c +8 −1
fs/autofs4/root.c +1 −1
fs/binfmt_elf.c +4 −4
fs/btrfs/ctree.h +25 −0
fs/btrfs/delayed-inode.c +16 −4
fs/btrfs/delayed-ref.c +14 −5
fs/btrfs/delayed-ref.h +1 −0
fs/btrfs/disk-io.c +1 −0
fs/btrfs/extent-tree.c +64 −16
fs/btrfs/file.c +1 −1
fs/btrfs/inode.c +12 −8
fs/btrfs/print-tree.c +15 −10
fs/btrfs/print-tree.h +1 −1
fs/btrfs/qgroup.c +41 −2
fs/btrfs/relocation.c +1 −1
fs/btrfs/send.c +4 −0
fs/btrfs/transaction.c +1 −0
fs/btrfs/transaction.h +14 −0
fs/ceph/inode.c +7 −3
fs/ceph/xattr.c +25 −3
fs/cifs/Kconfig +1 −1
fs/cifs/cifs_debug.h +1 −1
fs/cifs/cifssmb.c +3 −0
fs/cifs/connect.c +16 −16
fs/cifs/dir.c +5 −4
fs/cifs/file.c +1 −1
fs/cifs/smb2ops.c +16 −6
fs/cifs/smb2pdu.c +7 −6
fs/cifs/smb2pdu.h +1 −1
fs/cifs/smbdirect.c +18 −24
fs/cifs/transport.c +6 −3
fs/ecryptfs/crypto.c +28 −13
fs/ecryptfs/file.c +16 −5
fs/ecryptfs/inode.c +1 −2
fs/ecryptfs/keystore.c +1 −1
fs/ext2/file.c +2 −2
fs/ext4/balloc.c +5 −4
fs/ext4/extents.c +11 −5
fs/ext4/super.c +1 −0
fs/fs-writeback.c +5 −4
fs/isofs/compress.c +16 −3
fs/isofs/inode.c +3 −0
fs/jbd2/transaction.c +1 −0
fs/jffs2/super.c +1 −1
fs/namespace.c +3 −2
fs/notify/fanotify/fanotify.c +15 −19
fs/notify/fsnotify.c +11 −14
fs/orangefs/super.c +5 −0
fs/proc/base.c +6 −0
fs/proc/loadavg.c +1 −1
fs/proc/task_mmu.c +5 −1
fs/quota/dquot.c +1 −1
fs/super.c +4 −5
fs/udf/unicode.c +6 −0
fs/xfs/libxfs/xfs_attr.c +8 −1
fs/xfs/libxfs/xfs_bmap.c +4 −0
fs/xfs/libxfs/xfs_inode_buf.c +21 −0
fs/xfs/xfs_file.c +19 −5
include/asm-generic/vmlinux.lds.h +1 −1
include/drm/drm_hdcp.h +1 −1
include/dt-bindings/clock/stm32mp1-clks.h +2 −2
include/kvm/arm_psci.h +14 −2
include/kvm/arm_vgic.h +1 −0
include/linux/backing-dev-defs.h +5 −0
include/linux/backing-dev.h +16 −14
include/linux/blk-mq.h +3 −0
include/linux/blkdev.h +6 −0
include/linux/bpf.h +5 −3
include/linux/clk-provider.h +3 −0
include/linux/compiler-clang.h +3 −0
include/linux/coresight-pmu.h +1 −12
include/linux/device.h +4 −2
include/linux/ethtool.h +2 −0
include/linux/fsnotify_backend.h +2 −4
include/linux/genhd.h +3 −1
include/linux/hid.h +8 −1
include/linux/hrtimer.h +2 −0
include/linux/if_vlan.h +5 −2
include/linux/livepatch.h +13 −6
include/linux/microchipphy.h +8 −0
include/linux/mlx5/driver.h +3 −9
include/linux/mtd/flashchip.h +1 −0
include/linux/remoteproc.h +1 −1
include/linux/serial_core.h +14 −7
include/linux/shrinker.h +5 −2
include/linux/stringhash.h +2 −2
include/linux/textsearch.h +2 −2
include/linux/thread_info.h +1 −5
include/linux/ti-emif-sram.h +75 −0
include/linux/timekeeper_internal.h +0 −2
include/linux/timekeeping.h +25 −12
include/linux/timekeeping32.h +0 −3
include/linux/timer.h +0 −2
include/linux/tty.h +1 −1
include/linux/usb/composite.h +1 −1
include/linux/vbox_utils.h +0 −23
include/linux/virtio.h +3 −0
include/linux/wait_bit.h +17 −0
include/media/i2c/tvp7002.h +1 −1
include/media/videobuf-core.h +2 −2
include/media/videobuf-dma-sg.h +2 −2
include/media/videobuf-vmalloc.h +1 −1
include/net/ife.h +2 −1
include/net/llc_conn.h +1 −0
include/net/tls.h +1 −0
include/scsi/scsi_dbg.h +0 −2
include/soc/bcm2835/raspberrypi-firmware.h +2 −2
include/sound/control.h +5 −2
include/trace/events/initcall.h +11 −3
include/trace/events/ufs.h +27 −0
include/trace/events/workqueue.h +2 −0
include/uapi/linux/if_infiniband.h +1 −1
include/uapi/linux/kvm.h +7 −0
include/uapi/linux/perf_event.h +15 −3
include/uapi/linux/random.h +3 −0
include/uapi/linux/rds.h +1 −1
include/uapi/linux/sysctl.h +0 −18
include/uapi/linux/time.h +0 −1
include/uapi/linux/tls.h +1 −1
include/uapi/linux/virtio_balloon.h +15 −0
include/uapi/rdma/cxgb3-abi.h +1 −1
include/uapi/rdma/cxgb4-abi.h +1 −1
include/uapi/rdma/hns-abi.h +1 −1
include/uapi/rdma/ib_user_cm.h +1 −1
include/uapi/rdma/ib_user_ioctl_verbs.h +1 −1
include/uapi/rdma/ib_user_mad.h +1 −1
include/uapi/rdma/ib_user_sa.h +1 −1
include/uapi/rdma/ib_user_verbs.h +1 −1
include/uapi/rdma/mlx4-abi.h +1 −1
include/uapi/rdma/mlx5-abi.h +1 −1
include/uapi/rdma/mthca-abi.h +1 −1
include/uapi/rdma/nes-abi.h +1 −1
include/uapi/rdma/qedr-abi.h +1 −1
include/uapi/rdma/rdma_user_cm.h +1 −1
include/uapi/rdma/rdma_user_ioctl.h +1 −1
include/uapi/rdma/rdma_user_rxe.h +1 −1
include/xen/interface/io/sndif.h +306 −16
kernel/bpf/arraymap.c +2 −1
kernel/bpf/core.c +29 −16
kernel/bpf/sockmap.c +73 −29
kernel/bpf/syscall.c +2 −2
kernel/events/callchain.c +11 −14
kernel/events/core.c +6 −2
kernel/events/uprobes.c +3 −4
kernel/fork.c +1 −2
kernel/kprobes.c +1 −1
kernel/livepatch/shadow.c +71 −37
kernel/module.c +2 −1
kernel/sysctl_binary.c +1 −19
kernel/time/clocksource.c +44 −19
kernel/time/hrtimer.c +14 −2
kernel/time/posix-cpu-timers.c +3 −1
kernel/time/posix-stubs.c +0 −2
kernel/time/posix-timers.c +17 −9
kernel/time/tick-common.c +0 −15
kernel/time/tick-internal.h +0 −6
kernel/time/tick-oneshot.c +5 −6
kernel/time/tick-sched.c +5 −14
kernel/time/timekeeping.c +37 −48
kernel/time/timekeeping.h +1 −0
kernel/trace/bpf_trace.c +21 −4
kernel/trace/ftrace.c +2 −2
kernel/trace/trace.c +1 −1
kernel/trace/trace_entries.h +1 −1
kernel/trace/trace_events_filter.c +7 −7
kernel/trace/trace_events_hist.c +12 −0
kernel/trace/trace_kprobe.c +0 −2
kernel/trace/trace_stack.c +1 −1
kernel/trace/trace_uprobe.c +14 −21
kernel/tracepoint.c +2 −2
lib/dma-direct.c +2 −1
lib/errseq.c +9 −14
lib/kobject.c +5 −6
lib/swiotlb.c +1 −1
lib/textsearch.c +23 −17
mm/backing-dev.c +2 −1
mm/filemap.c +4 −5
mm/huge_memory.c +4 −1
mm/memcontrol.c +1 −1
mm/migrate.c +19 −3
mm/mmap.c +10 −1
mm/page-writeback.c +9 −9
mm/rmap.c +0 −3
mm/vmscan.c +20 −1
net/bridge/br_if.c +2 −2
net/bridge/netfilter/ebtables.c +6 −5
net/caif/chnl_net.c +1 −1
net/ceph/messenger.c +7 −0
net/ceph/mon_client.c +11 −3
net/compat.c +4 −2
net/core/dev.c +1 −1
net/core/dev_addr_lists.c +1 −1
net/core/ethtool.c +5 −0
net/core/filter.c +1 −0
net/core/neighbour.c +26 −14
net/dccp/ccids/ccid2.c +12 −2
net/dccp/timer.c +1 −1
net/dns_resolver/dns_key.c +5 −7
net/ife/ife.c +36 −2
net/ipv4/ip_output.c +5 −3
net/ipv4/route.c +53 −65
net/ipv4/tcp.c +9 −6
net/ipv4/tcp_bbr.c +3 −1
net/ipv4/tcp_input.c +2 −5
net/ipv6/netfilter/Kconfig +28 −27
net/ipv6/route.c +8 −1
net/ipv6/seg6_iptunnel.c +1 −1
net/l2tp/l2tp_core.c +20 −20
net/l2tp/l2tp_core.h +2 −1
net/l2tp/l2tp_debugfs.c +16 −2
net/l2tp/l2tp_netlink.c +8 −3
net/l2tp/l2tp_ppp.c +27 −7
net/llc/af_llc.c +12 −2
net/llc/llc_c_ac.c +1 −8
net/llc/llc_conn.c +21 −1
net/netfilter/Kconfig +1 −0
net/netfilter/ipvs/ip_vs_ctl.c +0 −8
net/netfilter/ipvs/ip_vs_sync.c +80 −75
net/netfilter/nf_conntrack_expect.c +4 −1
net/netfilter/nf_conntrack_extend.c +2 −0
net/netfilter/nf_conntrack_sip.c +12 −4
net/netfilter/nf_tables_api.c +38 −31
net/netfilter/xt_connmark.c +29 −20
net/packet/af_packet.c +58 −25
net/packet/internal.h +5 −5
net/qrtr/qrtr.c +1 −0
net/rds/ib_cm.c +2 −1
net/rds/recv.c +1 −0
net/sched/act_ife.c +7 −2
net/sched/sch_fq.c +25 −12
net/sctp/inqueue.c +1 −1
net/sctp/ipv6.c +32 −29
net/sctp/sm_statefuns.c +7 −1
net/sctp/stream.c +2 −0
net/smc/af_smc.c +33 −38
net/smc/smc_core.c +19 −3
net/smc/smc_core.h +2 −1
net/strparser/strparser.c +4 −5
net/sunrpc/rpc_pipe.c +1 −0
net/tipc/monitor.c +1 −1
net/tipc/name_table.c +21 −13
net/tipc/name_table.h +1 −1
net/tipc/net.c +2 −0
net/tipc/netlink.c +4 −1
net/tipc/node.c +5 −8
net/tipc/socket.c +3 −1
net/tipc/subscr.c +4 −1
net/tls/tls_main.c +7 −0
net/tls/tls_sw.c +9 −1
net/vmw_vsock/af_vsock.c +6 −0
samples/livepatch/livepatch-shadow-fix1.c +32 −11
samples/livepatch/livepatch-shadow-fix2.c +18 −15
samples/sockmap/Makefile +5 −2
scripts/Makefile.gcc-plugins +1 −1
scripts/Makefile.lib +1 −1
scripts/extract_xc3028.pl +1 −1
scripts/genksyms/Makefile +2 −2
scripts/mod/sumversion.c +1 −8
scripts/split-man.pl +1 −1
security/commoncap.c +2 −0
sound/core/control.c +1 −1
sound/core/pcm_compat.c +6 −3
sound/core/pcm_native.c +15 −15
sound/core/rawmidi_compat.c +12 −6
sound/core/seq/oss/seq_oss_event.c +9 −6
sound/core/seq/oss/seq_oss_midi.c +2 −0
sound/core/seq/oss/seq_oss_synth.c +49 −36
sound/core/seq/oss/seq_oss_synth.h +2 −1
sound/core/seq/seq_virmidi.c +2 −2
sound/drivers/aloop.c +15 −2
sound/drivers/opl3/opl3_synth.c +5 −2
sound/firewire/amdtp-stream.c +3 −2
sound/firewire/dice/dice-stream.c +1 −1
sound/firewire/dice/dice.c +1 −1
sound/pci/asihpi/hpimsginit.c +9 −4
sound/pci/asihpi/hpioctl.c +3 −1
sound/pci/hda/hda_hwdep.c +11 −1
sound/pci/hda/hda_intel.c +2 −1
sound/pci/hda/patch_hdmi.c +8 −1
sound/pci/hda/patch_realtek.c +9 −1
sound/pci/rme9652/hdspm.c +14 −10
sound/pci/rme9652/rme9652.c +4 −2
sound/soc/amd/acp-da7219-max98357a.c +1 −1
sound/soc/codecs/adau17x1.c +20 −6
sound/soc/codecs/adau17x1.h +2 −1
sound/soc/codecs/msm8916-wcd-analog.c +6 −3
sound/soc/codecs/rt5514.c +3 −0
sound/soc/fsl/fsl_esai.c +7 −0
sound/soc/fsl/fsl_ssi.c +11 −3
sound/soc/intel/Kconfig +13 −9
sound/soc/omap/omap-dmic.c +11 −3
sound/soc/sh/rcar/core.c +2 −2
sound/soc/soc-topology.c +9 −5
sound/usb/line6/midi.c +1 −1
sound/usb/mixer.c +4 −3
sound/usb/mixer_maps.c +3 −0
sound/usb/stream.c +1 −1
sound/usb/usx2y/us122l.c +1 −1
sound/usb/usx2y/usX2Yhwdep.c +1 −1
sound/usb/usx2y/usx2yhwdeppcm.c +1 −1
tools/arch/arm/include/uapi/asm/kvm.h +9 −0
tools/arch/x86/include/asm/required-features.h +1 −7
tools/arch/x86/include/uapi/asm/kvm.h +18 −1
tools/bpf/Makefile +2 −0
tools/bpf/bpf_dbg.c +5 −2
tools/include/linux/compiler.h +15 −5
tools/include/linux/coresight-pmu.h +1 −12
tools/include/uapi/asm-generic/mman-common.h +3 −0
tools/include/uapi/linux/bpf.h +1 −0
tools/include/uapi/linux/if_link.h +39 −0
tools/include/uapi/linux/kvm.h +20 −1
tools/include/uapi/linux/perf_event.h +15 −3
tools/include/uapi/sound/asound.h +1 −0
tools/lib/subcmd/parse-options.c +3 −3
tools/objtool/Makefile +2 −2
tools/perf/Documentation/perf-config.txt +5 −0
tools/perf/Documentation/perf-mem.txt +31 −11
tools/perf/Documentation/perf-sched.txt +2 −2
tools/perf/Documentation/perf-script.txt +9 −8
tools/perf/Documentation/perf-stat.txt +1 −1
tools/perf/Makefile.config +2 −2
tools/perf/arch/arm/include/arch-tests.h +12 −0
tools/perf/arch/arm/tests/Build +2 −0
tools/perf/arch/arm/tests/arch-tests.c +16 −0
tools/perf/arch/arm/util/auxtrace.c +1 −12
tools/perf/arch/arm/util/cs-etm.c +1 −12
tools/perf/arch/arm/util/cs-etm.h +1 −12
tools/perf/arch/arm/util/pmu.c +1 −12
tools/perf/arch/s390/util/auxtrace.c +1 −0
tools/perf/arch/s390/util/header.c +0 −18
tools/perf/arch/x86/Makefile +1 −1
tools/perf/arch/x86/annotate/instructions.c +66 −1
tools/perf/arch/x86/entry/syscalls/syscall_64.tbl +358 −354
tools/perf/builtin-help.c +1 −1
tools/perf/builtin-mem.c +2 −2
tools/perf/builtin-script.c +6 −3
tools/perf/builtin-stat.c +40 −14
tools/perf/builtin-version.c +3 −0
tools/perf/perf.c +2 −2
tools/perf/pmu-events/arch/s390/mapfile.csv +5 −5
tools/perf/tests/attr/test-record-group-sampling +3 −0
tools/perf/tests/bpf-script-example.c +1 −1
tools/perf/tests/bpf-script-test-kbuild.c +0 −1
tools/perf/tests/builtin-test.c +1 −0
tools/perf/tests/mmap-basic.c +1 −1
tools/perf/tests/shell/record+probe_libc_inet_pton.sh +2 −4
tools/perf/trace/beauty/mmap.c +3 −0
tools/perf/ui/browsers/annotate.c +5 −0
tools/perf/ui/browsers/hists.c +1 −1
tools/perf/util/annotate.c +21 −5
tools/perf/util/annotate.h +9 −0
tools/perf/util/cs-etm-decoder/cs-etm-decoder.c +1 −2
tools/perf/util/cs-etm.c +1 −2
tools/perf/util/cs-etm.h +1 −12
tools/perf/util/event.c +3 −1
tools/perf/util/evsel.c +16 −8
tools/perf/util/evsel.h +1 −0
tools/perf/util/generate-cmdlist.sh +1 −1
tools/perf/util/header.c +2 −1
tools/perf/util/machine.c +18 −12
tools/perf/util/parse-events.y +4 −4
tools/perf/util/pmu.c +14 −14
tools/perf/util/symbol.c +3 −5
tools/perf/util/syscalltbl.c +3 −3
tools/perf/util/trace-event-scripting.c +2 −2
tools/power/acpi/Makefile.config +1 −0
tools/testing/nvdimm/test/nfit.c +65 −19
tools/testing/selftests/bpf/.gitignore +3 −0
tools/testing/selftests/bpf/test_progs.c +2 −2
tools/testing/selftests/bpf/test_sock.c +1 −0
tools/testing/selftests/bpf/test_sock_addr.c +1 −0
tools/testing/selftests/bpf/test_sock_addr.sh +2 −2
tools/testing/selftests/filesystems/Makefile +3 −5
tools/testing/selftests/firmware/Makefile +1 −0
tools/testing/selftests/firmware/fw_lib.sh +6 −4
tools/testing/selftests/firmware/fw_run_tests.sh +1 −1
tools/testing/selftests/ftrace/test.d/trigger/inter-event/trigger-extended-error-support.tc +1 −1
tools/testing/selftests/ftrace/test.d/trigger/inter-event/trigger-multi-actions-accept.tc +44 −0
tools/testing/selftests/kvm/Makefile +3 −2
tools/testing/selftests/kvm/include/kvm_util.h +9 −6
tools/testing/selftests/kvm/include/vmx.h +494 −0
tools/testing/selftests/kvm/lib/kvm_util.c +13 −7
tools/testing/selftests/kvm/lib/sparsebit.c +2 −2
tools/testing/selftests/kvm/lib/vmx.c +243 −0
tools/testing/selftests/kvm/vmx_tsc_adjust_test.c +231 −0
tools/testing/selftests/lib.mk +4 −4
tools/testing/selftests/net/Makefile +1 −0
tools/testing/selftests/x86/test_syscall_vdso.c +21 −14
virt/kvm/arm/arm.c +10 −5
virt/kvm/arm/psci.c +60 −0
virt/kvm/arm/vgic/vgic-init.c +1 −1
virt/kvm/arm/vgic/vgic-mmio-v2.c +5 −0
virt/kvm/arm/vgic/vgic-mmio.c +8 −2
virt/kvm/arm/vgic/vgic-v2.c +22 −16
virt/kvm/arm/vgic/vgic-v3.c +29 −20
virt/kvm/arm/vgic/vgic.c +25 −27
virt/kvm/arm/vgic/vgic.h +14 −0

Documentation/bpf/bpf_devel_QA.txt+9−1

@@ -557,6 +557,14 @@ A: Although LLVM IR generation and optimization try to stay architecture · 556 unmodified lines
557	557	pulls in some header files containing file scope host assembly codes.
558	558	- You can add "-fno-jump-tables" to work around the switch table issue.
559	559
560		− Otherwise, you can use bpf target.
	560	+ Otherwise, you can use bpf target. Additionally, you _must_ use bpf target
	561	+ when:
	562	+
	563	+ - Your program uses data structures with pointer or long / unsigned long
	564	+ types that interface with BPF helpers or context data structures. Access
	565	+ into these structures is verified by the BPF verifier and may result
	566	+ in verification failures if the native architecture is not aligned with
	567	+ the BPF architecture, e.g. 64-bit. An example of this is
	568	+ BPF_PROG_TYPE_SK_MSG require '-target bpf'
561	569
562	570	Happy BPF hacking!

Documentation/core-api/kernel-api.rst+13−0

@@ -136,6 +136,19 @@ Sorting · 135 unmodified lines
136	136	.. kernel-doc:: lib/list_sort.c
137	137	:export:
138	138
	139	+Text Searching
	140	+--------------
	141	+
	142	+.. kernel-doc:: lib/textsearch.c
	143	+ :doc: ts_intro
	144	+
	145	+.. kernel-doc:: lib/textsearch.c
	146	+ :export:
	147	+
	148	+.. kernel-doc:: include/linux/textsearch.h
	149	+ :functions: textsearch_find textsearch_next \
	150	+ textsearch_get_pattern textsearch_get_pattern_len
	151	+
139	152	UUID/GUID
140	153	---------
141	154

Documentation/devicetree/bindings/input/atmel,maxtouch.txt+7−0

@@ -4,6 +4,13 @@ Required properties: · 3 unmodified lines
4	4	- compatible:
5	5	atmel,maxtouch
6	6
	7	+ The following compatibles have been used in various products but are
	8	+ deprecated:
	9	+ atmel,qt602240_ts
	10	+ atmel,atmel_mxt_ts
	11	+ atmel,atmel_mxt_tp
	12	+ atmel,mXT224
	13	+
7	14	- reg: The I2C address of the device
8	15
9	16	- interrupts: The sink for the touchpad's IRQ output

Documentation/devicetree/bindings/serial/amlogic,meson-uart.txt+1−1

@@ -21,7 +21,7 @@ Required properties: · 20 unmodified lines
21	21	- interrupts : identifier to the device interrupt
22	22	- clocks : a list of phandle + clock-specifier pairs, one for each
23	23	entry in clock names.
24		−- clocks-names :
	24	+- clock-names :
25	25	* "xtal" for external xtal clock identifier
26	26	* "pclk" for the bus core clock, either the clk81 clock or the gate clock
27	27	* "baud" for the source of the baudrate generator, can be either the xtal

Documentation/devicetree/bindings/serial/mvebu-uart.txt+1−1

@@ -24,7 +24,7 @@ Required properties: · 23 unmodified lines
24	24	- Must contain two elements for the extended variant of the IP
25	25	(marvell,armada-3700-uart-ext): "uart-tx" and "uart-rx",
26	26	respectively the UART TX interrupt and the UART RX interrupt. A
27		− corresponding interrupts-names property must be defined.
	27	+ corresponding interrupt-names property must be defined.
28	28	- For backward compatibility reasons, a single element interrupts
29	29	property is also supported for the standard variant of the IP,
30	30	containing only the UART sum interrupt. This form is deprecated

Documentation/devicetree/bindings/serial/renesas,sci-serial.txt+2−0

@@ -17,6 +17,8 @@ Required properties: · 16 unmodified lines
17	17	- "renesas,scifa-r8a7745" for R8A7745 (RZ/G1E) SCIFA compatible UART.
18	18	- "renesas,scifb-r8a7745" for R8A7745 (RZ/G1E) SCIFB compatible UART.
19	19	- "renesas,hscif-r8a7745" for R8A7745 (RZ/G1E) HSCIF compatible UART.
	20	+ - "renesas,scif-r8a77470" for R8A77470 (RZ/G1C) SCIF compatible UART.
	21	+ - "renesas,hscif-r8a77470" for R8A77470 (RZ/G1C) HSCIF compatible UART.
20	22	- "renesas,scif-r8a7778" for R8A7778 (R-Car M1) SCIF compatible UART.
21	23	- "renesas,scif-r8a7779" for R8A7779 (R-Car H1) SCIF compatible UART.
22	24	- "renesas,scif-r8a7790" for R8A7790 (R-Car H2) SCIF compatible UART.

Documentation/devicetree/bindings/thermal/exynos-thermal.txt+5−18

@@ -49,19 +49,6 @@ on the SoC (only first trip points defined in DT will be configured): · 48 unmodified lines
49	49	- samsung,exynos5433-tmu: 8
50	50	- samsung,exynos7-tmu: 8
51	51
52		−Following properties are mandatory (depending on SoC):
53		−- samsung,tmu_gain: Gain value for internal TMU operation.
54		−- samsung,tmu_reference_voltage: Value of TMU IP block's reference voltage
55		−- samsung,tmu_noise_cancel_mode: Mode for noise cancellation
56		−- samsung,tmu_efuse_value: Default level of temperature - it is needed when
57		− in factory fusing produced wrong value
58		−- samsung,tmu_min_efuse_value: Minimum temperature fused value
59		−- samsung,tmu_max_efuse_value: Maximum temperature fused value
60		−- samsung,tmu_first_point_trim: First point trimming value
61		−- samsung,tmu_second_point_trim: Second point trimming value
62		−- samsung,tmu_default_temp_offset: Default temperature offset
63		−- samsung,tmu_cal_type: Callibration type
64		−
65	52	** Optional properties:
66	53
67	54	- vtmu-supply: This entry is optional and provides the regulator node supplying
@@ -78,7 +65,7 @@ Example 1): · 10 unmodified lines
78	65	clocks = <&clock 383>;
79	66	clock-names = "tmu_apbif";
80	67	vtmu-supply = <&tmu_regulator_node>;
81		− #include "exynos4412-tmu-sensor-conf.dtsi"
	68	+ #thermal-sensor-cells = <0>;
82	69	};
83	70
84	71	Example 2):
@@ -89,7 +76,7 @@ Example 2): · 4 unmodified lines
89	76	interrupts = <0 58 0>;
90	77	clocks = <&clock 21>;
91	78	clock-names = "tmu_apbif";
92		− #include "exynos5440-tmu-sensor-conf.dtsi"
	79	+ #thermal-sensor-cells = <0>;
93	80	};
94	81
95	82	Example 3): (In case of Exynos5420 "with misplaced TRIMINFO register")
@@ -99,7 +86,7 @@ Example 3): (In case of Exynos5420 "with misplaced TRIMINFO register") · 3 unmodified lines
99	86	interrupts = <0 184 0>;
100	87	clocks = <&clock 318>, <&clock 318>;
101	88	clock-names = "tmu_apbif", "tmu_triminfo_apbif";
102		− #include "exynos4412-tmu-sensor-conf.dtsi"
	89	+ #thermal-sensor-cells = <0>;
103	90	};
104	91
105	92	tmu_cpu3: tmu@1006c000 {
@@ -108,7 +95,7 @@ Example 3): (In case of Exynos5420 "with misplaced TRIMINFO register") · 2 unmodified lines
108	95	interrupts = <0 185 0>;
109	96	clocks = <&clock 318>, <&clock 319>;
110	97	clock-names = "tmu_apbif", "tmu_triminfo_apbif";
111		− #include "exynos4412-tmu-sensor-conf.dtsi"
	98	+ #thermal-sensor-cells = <0>;
112	99	};
113	100
114	101	tmu_gpu: tmu@100a0000 {
@@ -117,7 +104,7 @@ Example 3): (In case of Exynos5420 "with misplaced TRIMINFO register") · 2 unmodified lines
117	104	interrupts = <0 215 0>;
118	105	clocks = <&clock 319>, <&clock 318>;
119	106	clock-names = "tmu_apbif", "tmu_triminfo_apbif";
120		− #include "exynos4412-tmu-sensor-conf.dtsi"
	107	+ #thermal-sensor-cells = <0>;
121	108	};
122	109
123	110	Note: For multi-instance tmu each instance should have an alias correctly

Documentation/devicetree/bindings/thermal/thermal.txt+1−15

@@ -55,8 +55,7 @@ of heat dissipation). For example a fan's cooling states correspond to · 54 unmodified lines
55	55	the different fan speeds possible. Cooling states are referred to by
56	56	single unsigned integers, where larger numbers mean greater heat
57	57	dissipation. The precise set of cooling states associated with a device
58		−(as referred to by the cooling-min-level and cooling-max-level
59		−properties) should be defined in a particular device's binding.
	58	+should be defined in a particular device's binding.
60	59	For more examples of cooling devices, refer to the example sections below.
61	60
62	61	Required properties:
@@ -69,15 +68,6 @@ Required properties: · 6 unmodified lines
69	68	See Cooling device maps section below for more details
70	69	on how consumers refer to cooling devices.
71	70
72		−Optional properties:
73		−- cooling-min-level: An integer indicating the smallest
74		− Type: unsigned cooling state accepted. Typically 0.
75		− Size: one cell
76		−
77		−- cooling-max-level: An integer indicating the largest
78		− Type: unsigned cooling state accepted.
79		− Size: one cell
80		−
81	71	* Trip points
82	72
83	73	The trip node is a node to describe a point in the temperature domain
@@ -226,8 +216,6 @@ cpus { · 142 unmodified lines
226	216	396000 950000
227	217	198000 850000
228	218	>;
229		− cooling-min-level = <0>;
230		− cooling-max-level = <3>;
231	219	#cooling-cells = <2>; /* min followed by max */
232	220	};
233	221	...
@@ -241,8 +229,6 @@ cpus { · 7 unmodified lines
241	229	*/
242	230	fan0: fan@48 {
243	231	...
244		− cooling-min-level = <0>;
245		− cooling-max-level = <9>;
246	232	#cooling-cells = <2>; /* min followed by max */
247	233	};
248	234	};

Documentation/devicetree/bindings/timer/nuvoton,npcm7xx-timer.txt+21−0

@@ -0,0 +1,21 @@
	1	+Nuvoton NPCM7xx timer
	2	+
	3	+Nuvoton NPCM7xx have three timer modules, each timer module provides five 24-bit
	4	+timer counters.
	5	+
	6	+Required properties:
	7	+- compatible : "nuvoton,npcm750-timer" for Poleg NPCM750.
	8	+- reg : Offset and length of the register set for the device.
	9	+- interrupts : Contain the timer interrupt with flags for
	10	+ falling edge.
	11	+- clocks : phandle of timer reference clock (usually a 25 MHz clock).
	12	+
	13	+Example:
	14	+
	15	+timer@f0008000 {
	16	+ compatible = "nuvoton,npcm750-timer";
	17	+ interrupts = <GIC_SPI 32 IRQ_TYPE_LEVEL_HIGH>;
	18	+ reg = <0xf0008000 0x50>;
	19	+ clocks = <&clk NPCM7XX_CLK_TIMER>;
	20	+};
	21	+

Documentation/devicetree/bindings/timer/nxp,tpm-timer.txt+1−1

@@ -15,7 +15,7 @@ Required properties: · 14 unmodified lines
15	15	- interrupts : Should be the clock event device interrupt.
16	16	- clocks : The clocks provided by the SoC to drive the timer, must contain
17	17	an entry for each entry in clock-names.
18		−- clock-names : Must include the following entries: "igp" and "per".
	18	+- clock-names : Must include the following entries: "ipg" and "per".
19	19
20	20	Example:
21	21	tpm5: tpm@40260000 {

Documentation/devicetree/bindings/usb/usb-xhci.txt+4−1

@@ -28,7 +28,10 @@ Required properties: · 27 unmodified lines
28	28	- interrupts: one XHCI interrupt should be described here.
29	29
30	30	Optional properties:
31		− - clocks: reference to a clock
	31	+ - clocks: reference to the clocks
	32	+ - clock-names: mandatory if there is a second clock, in this case
	33	+ the name must be "core" for the first clock and "reg" for the
	34	+ second one
32	35	- usb2-lpm-disable: indicate if we don't want to enable USB2 HW LPM
33	36	- usb3-lpm-capable: determines if platform is USB3 LPM capable
34	37	- quirk-broken-port-ped: set if the controller has broken port disable mechanism

Documentation/doc-guide/parse-headers.rst+2−2

@@ -177,14 +177,14 @@ BUGS · 176 unmodified lines
177	177	****
178	178
179	179
180		−Report bugs to Mauro Carvalho Chehab <mchehab@s-opensource.com>
	180	+Report bugs to Mauro Carvalho Chehab <mchehab@kernel.org>
181	181
182	182
183	183	COPYRIGHT
184	184	*********
185	185
186	186
187		−Copyright (c) 2016 by Mauro Carvalho Chehab <mchehab@s-opensource.com>.
	187	+Copyright (c) 2016 by Mauro Carvalho Chehab <mchehab+samsung@kernel.org>.
188	188
189	189	License GPLv2: GNU GPL version 2 <http://gnu.org/licenses/gpl.html>.
190	190

Documentation/driver-api/firmware/request_firmware.rst+8−8

@@ -17,17 +17,17 @@ an error is returned. · 16 unmodified lines
17	17
18	18	request_firmware
19	19	----------------
20		−.. kernel-doc:: drivers/base/firmware_class.c
	20	+.. kernel-doc:: drivers/base/firmware_loader/main.c
21	21	:functions: request_firmware
22	22
23	23	request_firmware_direct
24	24	-----------------------
25		−.. kernel-doc:: drivers/base/firmware_class.c
	25	+.. kernel-doc:: drivers/base/firmware_loader/main.c
26	26	:functions: request_firmware_direct
27	27
28	28	request_firmware_into_buf
29	29	-------------------------
30		−.. kernel-doc:: drivers/base/firmware_class.c
	30	+.. kernel-doc:: drivers/base/firmware_loader/main.c
31	31	:functions: request_firmware_into_buf
32	32
33	33	Asynchronous firmware requests
@@ -41,7 +41,7 @@ in atomic contexts. · 7 unmodified lines
41	41
42	42	request_firmware_nowait
43	43	-----------------------
44		−.. kernel-doc:: drivers/base/firmware_class.c
	44	+.. kernel-doc:: drivers/base/firmware_loader/main.c
45	45	:functions: request_firmware_nowait
46	46
47	47	Special optimizations on reboot
@@ -50,12 +50,12 @@ Special optimizations on reboot · 2 unmodified lines
50	50	Some devices have an optimization in place to enable the firmware to be
51	51	retained during system reboot. When such optimizations are used the driver
52	52	author must ensure the firmware is still available on resume from suspend,
53		−this can be done with firmware_request_cache() insted of requesting for the
54		−firmare to be loaded.
	53	+this can be done with firmware_request_cache() instead of requesting for the
	54	+firmware to be loaded.
55	55
56	56	firmware_request_cache()
57		−-----------------------
58		−.. kernel-doc:: drivers/base/firmware_class.c
	57	+------------------------
	58	+.. kernel-doc:: drivers/base/firmware_loader/main.c
59	59	:functions: firmware_request_cache
60	60
61	61	request firmware API expected driver use

Documentation/driver-api/infrastructure.rst+1−1

@@ -28,7 +28,7 @@ Device Drivers Base · 27 unmodified lines
28	28	.. kernel-doc:: drivers/base/node.c
29	29	:internal:
30	30
31		−.. kernel-doc:: drivers/base/firmware_class.c
	31	+.. kernel-doc:: drivers/base/firmware_loader/main.c
32	32	:export:
33	33
34	34	.. kernel-doc:: drivers/base/transport_class.c

Documentation/driver-api/usb/typec.rst+1−1

@@ -210,7 +210,7 @@ If the connector is dual-role capable, there may also be a switch for the data · 209 unmodified lines
210	210	role. USB Type-C Connector Class does not supply separate API for them. The
211	211	port drivers can use USB Role Class API with those.
212	212
213		−Illustration of the muxes behind a connector that supports an alternate mode:
	213	+Illustration of the muxes behind a connector that supports an alternate mode::
214	214
215	215	------------------------
216	216	\| Connector \|

Documentation/i2c/dev-interface+14−18

@@ -9,8 +9,8 @@ i2c adapters present on your system at a given time. i2cdetect is part of · 8 unmodified lines
9	9	the i2c-tools package.
10	10
11	11	I2C device files are character device files with major device number 89
12		−and a minor device number corresponding to the number assigned as
13		−explained above. They should be called "i2c-%d" (i2c-0, i2c-1, ...,
	12	+and a minor device number corresponding to the number assigned as
	13	+explained above. They should be called "i2c-%d" (i2c-0, i2c-1, ...,
14	14	i2c-10, ...). All 256 minor device numbers are reserved for i2c.
15	15
16	16
@@ -23,11 +23,6 @@ First, you need to include these two headers: · 6 unmodified lines
23	23	#include <linux/i2c-dev.h>
24	24	#include <i2c/smbus.h>
25	25
26		−(Please note that there are two files named "i2c-dev.h" out there. One is
27		−distributed with the Linux kernel and the other one is included in the
28		−source tree of i2c-tools. They used to be different in content but since 2012
29		−they're identical. You should use "linux/i2c-dev.h").
30		−
31	26	Now, you have to decide which adapter you want to access. You should
32	27	inspect /sys/class/i2c-dev/ or run "i2cdetect -l" to decide this.
33	28	Adapter numbers are assigned somewhat dynamically, so you can not
@@ -38,7 +33,7 @@ Next thing, open the device file, as follows: · 4 unmodified lines
38	33	int file;
39	34	int adapter_nr = 2; /* probably dynamically determined */
40	35	char filename[20];
41		−
	36	+
42	37	snprintf(filename, 19, "/dev/i2c-%d", adapter_nr);
43	38	file = open(filename, O_RDWR);
44	39	if (file < 0) {
@@ -72,8 +67,10 @@ the device supports them. Both are illustrated below. · 27 unmodified lines
72	67	/* res contains the read word */
73	68	}
74	69
75		− /* Using I2C Write, equivalent of
76		− i2c_smbus_write_word_data(file, reg, 0x6543) */
	70	+ /*
	71	+ * Using I2C Write, equivalent of
	72	+ * i2c_smbus_write_word_data(file, reg, 0x6543)
	73	+ */
77	74	buf[0] = reg;
78	75	buf[1] = 0x43;
79	76	buf[2] = 0x65;
@@ -140,14 +137,14 @@ ioctl(file, I2C_RDWR, struct i2c_rdwr_ioctl_data *msgset) · 60 unmodified lines
140	137	set in each message, overriding the values set with the above ioctl's.
141	138
142	139	ioctl(file, I2C_SMBUS, struct i2c_smbus_ioctl_data *args)
143		− Not meant to be called directly; instead, use the access functions
144		− below.
	140	+ If possible, use the provided i2c_smbus_* methods described below instead
	141	+ of issuing direct ioctls.
145	142
146	143	You can do plain i2c transactions by using read(2) and write(2) calls.
147	144	You do not need to pass the address byte; instead, set it through
148	145	ioctl I2C_SLAVE before you try to access the device.
149	146
150		−You can do SMBus level transactions (see documentation file smbus-protocol
	147	+You can do SMBus level transactions (see documentation file smbus-protocol
151	148	for details) through the following functions:
152	149	__s32 i2c_smbus_write_quick(int file, __u8 value);
153	150	__s32 i2c_smbus_read_byte(int file);
@@ -158,7 +155,7 @@ for details) through the following functions: · 4 unmodified lines
158	155	__s32 i2c_smbus_write_word_data(int file, __u8 command, __u16 value);
159	156	__s32 i2c_smbus_process_call(int file, __u8 command, __u16 value);
160	157	__s32 i2c_smbus_read_block_data(int file, __u8 command, __u8 *values);
161		− __s32 i2c_smbus_write_block_data(int file, __u8 command, __u8 length,
	158	+ __s32 i2c_smbus_write_block_data(int file, __u8 command, __u8 length,
162	159	__u8 *values);
163	160	All these transactions return -1 on failure; you can read errno to see
164	161	what happened. The 'write' transactions return 0 on success; the
@@ -166,10 +163,9 @@ what happened. The 'write' transactions return 0 on success; the · 1 unmodified line
166	163	returns the number of values read. The block buffers need not be longer
167	164	than 32 bytes.
168	165
169		−The above functions are all inline functions, that resolve to calls to
170		−the i2c_smbus_access function, that on its turn calls a specific ioctl
171		−with the data in a specific format. Read the source code if you
172		−want to know what happens behind the screens.
	166	+The above functions are made available by linking against the libi2c library,
	167	+which is provided by the i2c-tools project. See:
	168	+https://git.kernel.org/pub/scm/utils/i2c-tools/i2c-tools.git/.
173	169
174	170
175	171	Implementation details

Documentation/ioctl/ioctl-number.txt+0−2

@@ -217,7 +217,6 @@ Code Seq#(hex) Include File Comments · 216 unmodified lines
217	217	'd' 02-40 pcmcia/ds.h conflict!
218	218	'd' F0-FF linux/digi1.h
219	219	'e' all linux/digi1.h conflict!
220		−'e' 00-1F drivers/net/irda/irtty-sir.h conflict!
221	220	'f' 00-1F linux/ext2_fs.h conflict!
222	221	'f' 00-1F linux/ext3_fs.h conflict!
223	222	'f' 00-0F fs/jfs/jfs_dinode.h conflict!
@@ -247,7 +246,6 @@ Code Seq#(hex) Include File Comments · 23 unmodified lines
247	246	'm' all linux/synclink.h conflict!
248	247	'm' 00-19 drivers/message/fusion/mptctl.h conflict!
249	248	'm' 00 drivers/scsi/megaraid/megaraid_ioctl.h conflict!
250		−'m' 00-1F net/irda/irmod.h conflict!
251	249	'n' 00-7F linux/ncp_fs.h and fs/ncpfs/ioctl.c
252	250	'n' 80-8F uapi/linux/nilfs2_api.h NILFS2
253	251	'n' E0-FF linux/matroxfb.h matroxfb

Documentation/livepatch/shadow-vars.txt+29−12

@@ -34,9 +34,13 @@ meta-data and shadow-data: · 33 unmodified lines
34	34	- data[] - storage for shadow data
35	35
36	36	It is important to note that the klp_shadow_alloc() and
37		−klp_shadow_get_or_alloc() calls, described below, store a copy of the
38		−data that the functions are provided. Callers should provide whatever
39		−mutual exclusion is required of the shadow data.
	37	+klp_shadow_get_or_alloc() are zeroing the variable by default.
	38	+They also allow to call a custom constructor function when a non-zero
	39	+value is needed. Callers should provide whatever mutual exclusion
	40	+is required.
	41	+
	42	+Note that the constructor is called under klp_shadow_lock spinlock. It allows
	43	+to do actions that can be done only once when a new variable is allocated.
40	44
41	45	* klp_shadow_get() - retrieve a shadow variable data pointer
42	46	- search hashtable for <obj, id> pair
@@ -47,7 +51,7 @@ mutual exclusion is required of the shadow data. · 4 unmodified lines
47	51	- WARN and return NULL
48	52	- if <obj, id> doesn't already exist
49	53	- allocate a new shadow variable
50		− - copy data into the new shadow variable
	54	+ - initialize the variable using a custom constructor and data when provided
51	55	- add <obj, id> to the global hashtable
52	56
53	57	* klp_shadow_get_or_alloc() - get existing or alloc a new shadow variable
@@ -56,16 +60,20 @@ mutual exclusion is required of the shadow data. · 2 unmodified lines
56	60	- return existing shadow variable
57	61	- if <obj, id> doesn't already exist
58	62	- allocate a new shadow variable
59		− - copy data into the new shadow variable
	63	+ - initialize the variable using a custom constructor and data when provided
60	64	- add <obj, id> pair to the global hashtable
61	65
62	66	* klp_shadow_free() - detach and free a <obj, id> shadow variable
63	67	- find and remove a <obj, id> reference from global hashtable
64		− - if found, free shadow variable
	68	+ - if found
	69	+ - call destructor function if defined
	70	+ - free shadow variable
65	71
66	72	* klp_shadow_free_all() - detach and free all <*, id> shadow variables
67	73	- find and remove any <*, id> references from global hashtable
68		− - if found, free shadow variable
	74	+ - if found
	75	+ - call destructor function if defined
	76	+ - free shadow variable
69	77
70	78
71	79	2. Use cases
@@ -107,7 +115,8 @@ struct sta_info sta_info_alloc(struct ieee80211_sub_if_data sdata, · 35 unmodified lines
107	115	sta = kzalloc(sizeof(*sta) + hw->sta_data_size, gfp);
108	116
109	117	/* Attach a corresponding shadow variable, then initialize it */
110		− ps_lock = klp_shadow_alloc(sta, PS_LOCK, NULL, sizeof(*ps_lock), gfp);
	118	+ ps_lock = klp_shadow_alloc(sta, PS_LOCK, sizeof(*ps_lock), gfp,
	119	+ NULL, NULL);
111	120	if (!ps_lock)
112	121	goto shadow_fail;
113	122	spin_lock_init(ps_lock);
@@ -131,7 +140,7 @@ variable: · 17 unmodified lines
131	140
132	141	void sta_info_free(struct ieee80211_local local, struct sta_info sta)
133	142	{
134		− klp_shadow_free(sta, PS_LOCK);
	143	+ klp_shadow_free(sta, PS_LOCK, NULL);
135	144	kfree(sta);
136	145	...
137	146
@@ -148,16 +157,24 @@ shadow variables to parents already in-flight. · 10 unmodified lines
148	157	For commit 1d147bfa6429, a good spot to allocate a shadow spinlock is
149	158	inside ieee80211_sta_ps_deliver_wakeup():
150	159
	160	+int ps_lock_shadow_ctor(void obj, void shadow_data, void *ctor_data)
	161	+{
	162	+ spinlock_t *lock = shadow_data;
	163	+
	164	+ spin_lock_init(lock);
	165	+ return 0;
	166	+}
	167	+
151	168	#define PS_LOCK 1
152	169	void ieee80211_sta_ps_deliver_wakeup(struct sta_info *sta)
153	170	{
154		− DEFINE_SPINLOCK(ps_lock_fallback);
155	171	spinlock_t *ps_lock;
156	172
157	173	/* sync with ieee80211_tx_h_unicast_ps_buf */
158	174	ps_lock = klp_shadow_get_or_alloc(sta, PS_LOCK,
159		− &ps_lock_fallback, sizeof(ps_lock_fallback),
160		− GFP_ATOMIC);
	175	+ sizeof(*ps_lock), GFP_ATOMIC,
	176	+ ps_lock_shadow_ctor, NULL);
	177	+
161	178	if (ps_lock)
162	179	spin_lock(ps_lock);
163	180	...

Documentation/media/uapi/rc/keytable.c.rst+1−1

@@ -7,7 +7,7 @@ file: uapi/v4l/keytable.c · 6 unmodified lines
7	7
8	8	/* keytable.c - This program allows checking/replacing keys at IR
9	9
10		− Copyright (C) 2006-2009 Mauro Carvalho Chehab <mchehab@infradead.org>
	10	+ Copyright (C) 2006-2009 Mauro Carvalho Chehab <mchehab@kernel.org>
11	11
12	12	This program is free software; you can redistribute it and/or modify
13	13	it under the terms of the GNU General Public License as published by

Documentation/media/uapi/v4l/v4l2grab.c.rst+1−1

@@ -6,7 +6,7 @@ file: media/v4l/v4l2grab.c · 5 unmodified lines
6	6	.. code-block:: c
7	7
8	8	/* V4L2 video picture grabber
9		− Copyright (C) 2009 Mauro Carvalho Chehab <mchehab@infradead.org>
	9	+ Copyright (C) 2009 Mauro Carvalho Chehab <mchehab@kernel.org>
10	10
11	11	This program is free software; you can redistribute it and/or modify
12	12	it under the terms of the GNU General Public License as published by

Documentation/networking/filter.txt+3−3

@@ -169,7 +169,7 @@ access to BPF code as well. · 168 unmodified lines
169	169	BPF engine and instruction set
170	170	------------------------------
171	171
172		−Under tools/net/ there's a small helper tool called bpf_asm which can
	172	+Under tools/bpf/ there's a small helper tool called bpf_asm which can
173	173	be used to write low-level filters for example scenarios mentioned in the
174	174	previous section. Asm-like syntax mentioned here has been implemented in
175	175	bpf_asm and will be used for further explanations (instead of dealing with
@@ -359,7 +359,7 @@ $ ./bpf_asm -c foo · 183 unmodified lines
359	359	In particular, as usage with xt_bpf or cls_bpf can result in more complex BPF
360	360	filters that might not be obvious at first, it's good to test filters before
361	361	attaching to a live system. For that purpose, there's a small tool called
362		−bpf_dbg under tools/net/ in the kernel source directory. This debugger allows
	362	+bpf_dbg under tools/bpf/ in the kernel source directory. This debugger allows
363	363	for testing BPF filters against given pcap files, single stepping through the
364	364	BPF code on the pcap's packets and to do BPF machine register dumps.
365	365
@@ -483,7 +483,7 @@ Example output from dmesg: · 117 unmodified lines
483	483	[ 3389.935851] JIT code: 00000030: 00 e8 28 94 ff e0 83 f8 01 75 07 b8 ff ff 00 00
484	484	[ 3389.935852] JIT code: 00000040: eb 02 31 c0 c9 c3
485	485
486		−In the kernel source tree under tools/net/, there's bpf_jit_disasm for
	486	+In the kernel source tree under tools/bpf/, there's bpf_jit_disasm for
487	487	generating disassembly out of the kernel log's hexdump:
488	488
489	489	# ./bpf_jit_disasm

Documentation/networking/ip-sysctl.txt+4−19

@@ -1390,26 +1390,26 @@ mld_qrv - INTEGER · 1389 unmodified lines
1390	1390	Default: 2 (as specified by RFC3810 9.1)
1391	1391	Minimum: 1 (as specified by RFC6636 4.5)
1392	1392
1393		−max_dst_opts_cnt - INTEGER
	1393	+max_dst_opts_number - INTEGER
1394	1394	Maximum number of non-padding TLVs allowed in a Destination
1395	1395	options extension header. If this value is less than zero
1396	1396	then unknown options are disallowed and the number of known
1397	1397	TLVs allowed is the absolute value of this number.
1398	1398	Default: 8
1399	1399
1400		−max_hbh_opts_cnt - INTEGER
	1400	+max_hbh_opts_number - INTEGER
1401	1401	Maximum number of non-padding TLVs allowed in a Hop-by-Hop
1402	1402	options extension header. If this value is less than zero
1403	1403	then unknown options are disallowed and the number of known
1404	1404	TLVs allowed is the absolute value of this number.
1405	1405	Default: 8
1406	1406
1407		−max dst_opts_len - INTEGER
	1407	+max_dst_opts_length - INTEGER
1408	1408	Maximum length allowed for a Destination options extension
1409	1409	header.
1410	1410	Default: INT_MAX (unlimited)
1411	1411
1412		−max hbh_opts_len - INTEGER
	1412	+max_hbh_length - INTEGER
1413	1413	Maximum length allowed for a Hop-by-Hop options extension
1414	1414	header.
1415	1415	Default: INT_MAX (unlimited)
@@ -2126,18 +2126,3 @@ max_dgram_qlen - INTEGER · 710 unmodified lines
2126	2126
2127	2127	Default: 10
2128	2128
2129		−
2130		−UNDOCUMENTED:
2131		−
2132		−/proc/sys/net/irda/*
2133		− fast_poll_increase FIXME
2134		− warn_noreply_time FIXME
2135		− discovery_slots FIXME
2136		− slot_timeout FIXME
2137		− max_baud_rate FIXME
2138		− discovery_timeout FIXME
2139		− lap_keepalive_time FIXME
2140		− max_noreply_time FIXME
2141		− max_tx_data_size FIXME
2142		− max_tx_window FIXME
2143		− min_tx_turn_time FIXME

Documentation/power/suspend-and-cpuhotplug.txt+1−1

@@ -168,7 +168,7 @@ update on the CPUs, as discussed below: · 167 unmodified lines
168	168
169	169	[Please bear in mind that the kernel requests the microcode images from
170	170	userspace, using the request_firmware() function defined in
171		−drivers/base/firmware_class.c]
	171	+drivers/base/firmware_loader/main.c]
172	172
173	173
174	174	a. When all the CPUs are identical:

Documentation/process/magic-number.rst+0−3

@@ -157,8 +157,5 @@ memory management. See ``include/sound/sndmagic.h`` for complete list of them. M · 156 unmodified lines
157	157	OSS sound drivers have their magic numbers constructed from the soundcard PCI
158	158	ID - these are not listed here as well.
159	159
160		−IrDA subsystem also uses large number of own magic numbers, see
161		−``include/net/irda/irda.h`` for a complete list of them.
162		−
163	160	HFS is another larger user of magic numbers - you can find them in
164	161	``fs/hfs/hfs.h``.

Documentation/sphinx/parse-headers.pl+2−2

@@ -387,11 +387,11 @@ tree for more details. · 386 unmodified lines
387	387
388	388	=head1 BUGS
389	389
390		−Report bugs to Mauro Carvalho Chehab <mchehab@s-opensource.com>
	390	+Report bugs to Mauro Carvalho Chehab <mchehab@kernel.org>
391	391
392	392	=head1 COPYRIGHT
393	393
394		−Copyright (c) 2016 by Mauro Carvalho Chehab <mchehab@s-opensource.com>.
	394	+Copyright (c) 2016 by Mauro Carvalho Chehab <mchehab+samsung@kernel.org>.
395	395
396	396	License GPLv2: GNU GPL version 2 <http://gnu.org/licenses/gpl.html>.
397	397

Documentation/trace/ftrace.rst+11−3

@@ -461,9 +461,17 @@ of ftrace. Here is a list of some of the key files: · 460 unmodified lines
461	461	and ticks at the same rate as the hardware clocksource.
462	462
463	463	boot:
464		− Same as mono. Used to be a separate clock which accounted
465		− for the time spent in suspend while CLOCK_MONOTONIC did
466		− not.
	464	+ This is the boot clock (CLOCK_BOOTTIME) and is based on the
	465	+ fast monotonic clock, but also accounts for time spent in
	466	+ suspend. Since the clock access is designed for use in
	467	+ tracing in the suspend path, some side effects are possible
	468	+ if clock is accessed after the suspend time is accounted before
	469	+ the fast mono clock is updated. In this case, the clock update
	470	+ appears to happen slightly sooner than it normally would have.
	471	+ Also on 32-bit systems, it's possible that the 64-bit boot offset
	472	+ sees a partial update. These effects are rare and post
	473	+ processing should be able to handle them. See comments in the
	474	+ ktime_get_boot_fast_ns() function for more information.
467	475
468	476	To set a clock, simply echo the clock name into this file::
469	477

Documentation/translations/zh_CN/video4linux/v4l2-framework.txt+2−2

@@ -6,7 +6,7 @@ communicating in English you can also ask the Chinese maintainer for · 5 unmodified lines
6	6	help. Contact the Chinese maintainer if this translation is outdated
7	7	or if there is a problem with the translation.
8	8
9		−Maintainer: Mauro Carvalho Chehab <mchehab@infradead.org>
	9	+Maintainer: Mauro Carvalho Chehab <mchehab@kernel.org>
10	10	Chinese maintainer: Fu Wei <tekkamanninja@gmail.com>
11	11	---------------------------------------------------------------------
12	12	Documentation/video4linux/v4l2-framework.txt 的中文翻译
@@ -14,7 +14,7 @@ Documentation/video4linux/v4l2-framework.txt 的中文翻译 · 1 unmodified line
14	14	如果想评论或更新本文的内容，请直接联系原文档的维护者。如果你使用英文
15	15	交流有困难的话，也可以向中文版维护者求助。如果本翻译更新不及时或者翻
16	16	译存在问题，请联系中文版维护者。
17		−英文版维护者： Mauro Carvalho Chehab <mchehab@infradead.org>
	17	+英文版维护者： Mauro Carvalho Chehab <mchehab@kernel.org>
18	18	中文版维护者：傅炜 Fu Wei <tekkamanninja@gmail.com>
19	19	中文版翻译者：傅炜 Fu Wei <tekkamanninja@gmail.com>
20	20	中文版校译者：傅炜 Fu Wei <tekkamanninja@gmail.com>

Documentation/virtual/kvm/api.txt+8−1

@@ -1960,6 +1960,9 @@ ARM 32-bit VFP control registers have the following id bit patterns: · 1959 unmodified lines
1960	1960	ARM 64-bit FP registers have the following id bit patterns:
1961	1961	0x4030 0000 0012 0 <regno:12>
1962	1962
	1963	+ARM firmware pseudo-registers have the following bit pattern:
	1964	+ 0x4030 0000 0014 <regno:16>
	1965	+
1963	1966
1964	1967	arm64 registers are mapped using the lower 32 bits. The upper 16 of
1965	1968	that is the register group type, or coprocessor number:
@@ -1976,6 +1979,9 @@ arm64 CCSIDR registers are demultiplexed by CSSELR value: · 10 unmodified lines
1976	1979	arm64 system registers have the following id bit patterns:
1977	1980	0x6030 0000 0013 <op0:2> <op1:3> <crn:4> <crm:4> <op2:3>
1978	1981
	1982	+arm64 firmware pseudo-registers have the following bit pattern:
	1983	+ 0x6030 0000 0014 <regno:16>
	1984	+
1979	1985
1980	1986	MIPS registers are mapped using the lower 32 bits. The upper 16 of that is
1981	1987	the register group type:
@@ -2510,7 +2516,8 @@ Possible features: · 528 unmodified lines
2510	2516	and execute guest code when KVM_RUN is called.
2511	2517	- KVM_ARM_VCPU_EL1_32BIT: Starts the CPU in a 32bit mode.
2512	2518	Depends on KVM_CAP_ARM_EL1_32BIT (arm64 only).
2513		− - KVM_ARM_VCPU_PSCI_0_2: Emulate PSCI v0.2 for the CPU.
	2519	+ - KVM_ARM_VCPU_PSCI_0_2: Emulate PSCI v0.2 (or a future revision
	2520	+ backward compatible with v0.2) for the CPU.
2514	2521	Depends on KVM_CAP_ARM_PSCI_0_2.
2515	2522	- KVM_ARM_VCPU_PMU_V3: Emulate PMUv3 for the CPU.
2516	2523	Depends on KVM_CAP_ARM_PMU_V3.

Documentation/virtual/kvm/arm/psci.txt+30−0

@@ -0,0 +1,30 @@
	1	+KVM implements the PSCI (Power State Coordination Interface)
	2	+specification in order to provide services such as CPU on/off, reset
	3	+and power-off to the guest.
	4	+
	5	+The PSCI specification is regularly updated to provide new features,
	6	+and KVM implements these updates if they make sense from a virtualization
	7	+point of view.
	8	+
	9	+This means that a guest booted on two different versions of KVM can
	10	+observe two different "firmware" revisions. This could cause issues if
	11	+a given guest is tied to a particular PSCI revision (unlikely), or if
	12	+a migration causes a different PSCI version to be exposed out of the
	13	+blue to an unsuspecting guest.
	14	+
	15	+In order to remedy this situation, KVM exposes a set of "firmware
	16	+pseudo-registers" that can be manipulated using the GET/SET_ONE_REG
	17	+interface. These registers can be saved/restored by userspace, and set
	18	+to a convenient value if required.
	19	+
	20	+The following register is defined:
	21	+
	22	+* KVM_REG_ARM_PSCI_VERSION:
	23	+
	24	+ - Only valid if the vcpu has the KVM_ARM_VCPU_PSCI_0_2 feature set
	25	+ (and thus has already been initialized)
	26	+ - Returns the current PSCI version on GET_ONE_REG (defaulting to the
	27	+ highest PSCI version implemented by KVM and compatible with v0.2)
	28	+ - Allows any PSCI version implemented by KVM and compatible with
	29	+ v0.2 to be set with SET_ONE_REG
	30	+ - Affects the whole VM (even if the register view is per-vcpu)

MAINTAINERS+49−43

@@ -564,8 +564,9 @@ S: Maintained · 563 unmodified lines
564	564	F: drivers/media/dvb-frontends/af9033*
565	565
566	566	AFFS FILE SYSTEM
	567	+M: David Sterba <dsterba@suse.com>
567	568	L: linux-fsdevel@vger.kernel.org
568		−S: Orphan
	569	+S: Odd Fixes
569	570	F: Documentation/filesystems/affs.txt
570	571	F: fs/affs/
571	572
@@ -905,6 +906,8 @@ ANDROID ION DRIVER · 333 unmodified lines
905	906	M: Laura Abbott <labbott@redhat.com>
906	907	M: Sumit Semwal <sumit.semwal@linaro.org>
907	908	L: devel@driverdev.osuosl.org
	909	+L: dri-devel@lists.freedesktop.org
	910	+L: linaro-mm-sig@lists.linaro.org (moderated for non-subscribers)
908	911	S: Supported
909	912	F: drivers/staging/android/ion
910	913	F: drivers/staging/android/uapi/ion.h
@@ -1208,7 +1211,6 @@ F: drivers//alpine* · 297 unmodified lines
1208	1211	ARM/ARTPEC MACHINE SUPPORT
1209	1212	M: Jesper Nilsson <jesper.nilsson@axis.com>
1210	1213	M: Lars Persson <lars.persson@axis.com>
1211		−M: Niklas Cassel <niklas.cassel@axis.com>
1212	1214	S: Maintained
1213	1215	L: linux-arm-kernel@axis.com
1214	1216	F: arch/arm/mach-artpec
@@ -1373,7 +1375,8 @@ F: arch/arm/mach-ebsa110/ · 158 unmodified lines
1373	1375	F: drivers/net/ethernet/amd/am79c961a.*
1374	1376
1375	1377	ARM/ENERGY MICRO (SILICON LABS) EFM32 SUPPORT
1376		−M: Uwe Kleine-König <kernel@pengutronix.de>
	1378	+M: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
	1379	+R: Pengutronix Kernel Team <kernel@pengutronix.de>
1377	1380	L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers)
1378	1381	S: Maintained
1379	1382	N: efm32
@@ -1401,7 +1404,8 @@ F: arch/arm/mach-footbridge/ · 21 unmodified lines
1401	1404
1402	1405	ARM/FREESCALE IMX / MXC ARM ARCHITECTURE
1403	1406	M: Shawn Guo <shawnguo@kernel.org>
1404		−M: Sascha Hauer <kernel@pengutronix.de>
	1407	+M: Sascha Hauer <s.hauer@pengutronix.de>
	1408	+R: Pengutronix Kernel Team <kernel@pengutronix.de>
1405	1409	R: Fabio Estevam <fabio.estevam@nxp.com>
1406	1410	L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers)
1407	1411	S: Maintained
@@ -1416,7 +1420,8 @@ F: include/soc/imx/ · 8 unmodified lines
1416	1420
1417	1421	ARM/FREESCALE VYBRID ARM ARCHITECTURE
1418	1422	M: Shawn Guo <shawnguo@kernel.org>
1419		−M: Sascha Hauer <kernel@pengutronix.de>
	1423	+M: Sascha Hauer <s.hauer@pengutronix.de>
	1424	+R: Pengutronix Kernel Team <kernel@pengutronix.de>
1420	1425	R: Stefan Agner <stefan@agner.ch>
1421	1426	L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers)
1422	1427	S: Maintained
@@ -2549,7 +2554,6 @@ F: Documentation/devicetree/bindings/sound/axentia,* · 1126 unmodified lines
2549	2554	F: sound/soc/atmel/tse850-pcm5142.c
2550	2555
2551	2556	AZ6007 DVB DRIVER
2552		−M: Mauro Carvalho Chehab <mchehab@s-opensource.com>
2553	2557	M: Mauro Carvalho Chehab <mchehab@kernel.org>
2554	2558	L: linux-media@vger.kernel.org
2555	2559	W: https://linuxtv.org
@@ -2614,7 +2618,7 @@ S: Maintained · 58 unmodified lines
2614	2618	F: drivers/net/hamradio/baycom*
2615	2619
2616	2620	BCACHE (BLOCK LAYER CACHE)
2617		−M: Michael Lyle <mlyle@lyle.org>
	2621	+M: Coly Li <colyli@suse.de>
2618	2622	M: Kent Overstreet <kent.overstreet@gmail.com>
2619	2623	L: linux-bcache@vger.kernel.org
2620	2624	W: http://bcache.evilpiepirate.org
@@ -3078,7 +3082,6 @@ F: include/linux/btrfs* · 457 unmodified lines
3078	3082	F: include/uapi/linux/btrfs*
3079	3083
3080	3084	BTTV VIDEO4LINUX DRIVER
3081		−M: Mauro Carvalho Chehab <mchehab@s-opensource.com>
3082	3085	M: Mauro Carvalho Chehab <mchehab@kernel.org>
3083	3086	L: linux-media@vger.kernel.org
3084	3087	W: https://linuxtv.org
@@ -3807,7 +3810,6 @@ S: Maintained · 722 unmodified lines
3807	3810	F: drivers/media/dvb-frontends/cx24120*
3808	3811
3809	3812	CX88 VIDEO4LINUX DRIVER
3810		−M: Mauro Carvalho Chehab <mchehab@s-opensource.com>
3811	3813	M: Mauro Carvalho Chehab <mchehab@kernel.org>
3812	3814	L: linux-media@vger.kernel.org
3813	3815	W: https://linuxtv.org
@@ -4245,6 +4247,9 @@ F: include/trace/events/fs_dax.h · 431 unmodified lines
4245	4247
4246	4248	DEVICE DIRECT ACCESS (DAX)
4247	4249	M: Dan Williams <dan.j.williams@intel.com>
	4250	+M: Dave Jiang <dave.jiang@intel.com>
	4251	+M: Ross Zwisler <ross.zwisler@linux.intel.com>
	4252	+M: Vishal Verma <vishal.l.verma@intel.com>
4248	4253	L: linux-nvdimm@lists.01.org
4249	4254	S: Supported
4250	4255	F: drivers/dax/
@@ -5045,7 +5050,6 @@ F: drivers/edac/thunderx_edac* · 794 unmodified lines
5045	5050
5046	5051	EDAC-CORE
5047	5052	M: Borislav Petkov <bp@alien8.de>
5048		−M: Mauro Carvalho Chehab <mchehab@s-opensource.com>
5049	5053	M: Mauro Carvalho Chehab <mchehab@kernel.org>
5050	5054	L: linux-edac@vger.kernel.org
5051	5055	T: git git://git.kernel.org/pub/scm/linux/kernel/git/bp/bp.git for-next
@@ -5074,7 +5078,6 @@ S: Maintained · 22 unmodified lines
5074	5078	F: drivers/edac/fsl_ddr_edac.*
5075	5079
5076	5080	EDAC-GHES
5077		−M: Mauro Carvalho Chehab <mchehab@s-opensource.com>
5078	5081	M: Mauro Carvalho Chehab <mchehab@kernel.org>
5079	5082	L: linux-edac@vger.kernel.org
5080	5083	S: Maintained
@@ -5091,21 +5094,18 @@ S: Maintained · 10 unmodified lines
5091	5094	F: drivers/edac/i5000_edac.c
5092	5095
5093	5096	EDAC-I5400
5094		−M: Mauro Carvalho Chehab <mchehab@s-opensource.com>
5095	5097	M: Mauro Carvalho Chehab <mchehab@kernel.org>
5096	5098	L: linux-edac@vger.kernel.org
5097	5099	S: Maintained
5098	5100	F: drivers/edac/i5400_edac.c
5099	5101
5100	5102	EDAC-I7300
5101		−M: Mauro Carvalho Chehab <mchehab@s-opensource.com>
5102	5103	M: Mauro Carvalho Chehab <mchehab@kernel.org>
5103	5104	L: linux-edac@vger.kernel.org
5104	5105	S: Maintained
5105	5106	F: drivers/edac/i7300_edac.c
5106	5107
5107	5108	EDAC-I7CORE
5108		−M: Mauro Carvalho Chehab <mchehab@s-opensource.com>
5109	5109	M: Mauro Carvalho Chehab <mchehab@kernel.org>
5110	5110	L: linux-edac@vger.kernel.org
5111	5111	S: Maintained
@@ -5155,7 +5155,6 @@ S: Maintained · 43 unmodified lines
5155	5155	F: drivers/edac/r82600_edac.c
5156	5156
5157	5157	EDAC-SBRIDGE
5158		−M: Mauro Carvalho Chehab <mchehab@s-opensource.com>
5159	5158	M: Mauro Carvalho Chehab <mchehab@kernel.org>
5160	5159	L: linux-edac@vger.kernel.org
5161	5160	S: Maintained
@@ -5214,7 +5213,6 @@ S: Maintained · 52 unmodified lines
5214	5213	F: drivers/net/ethernet/ibm/ehea/
5215	5214
5216	5215	EM28XX VIDEO4LINUX DRIVER
5217		−M: Mauro Carvalho Chehab <mchehab@s-opensource.com>
5218	5216	M: Mauro Carvalho Chehab <mchehab@kernel.org>
5219	5217	L: linux-media@vger.kernel.org
5220	5218	W: https://linuxtv.org
@@ -5652,7 +5650,8 @@ F: drivers/net/ethernet/freescale/fec.h · 431 unmodified lines
5652	5650	F: Documentation/devicetree/bindings/net/fsl-fec.txt
5653	5651
5654	5652	FREESCALE IMX / MXC FRAMEBUFFER DRIVER
5655		−M: Sascha Hauer <kernel@pengutronix.de>
	5653	+M: Sascha Hauer <s.hauer@pengutronix.de>
	5654	+R: Pengutronix Kernel Team <kernel@pengutronix.de>
5656	5655	L: linux-fbdev@vger.kernel.org
5657	5656	L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers)
5658	5657	S: Maintained
@@ -5784,6 +5783,14 @@ F: fs/crypto/ · 125 unmodified lines
5784	5783	F: include/linux/fscrypt*.h
5785	5784	F: Documentation/filesystems/fscrypt.rst
5786	5785
	5786	+FSNOTIFY: FILESYSTEM NOTIFICATION INFRASTRUCTURE
	5787	+M: Jan Kara <jack@suse.cz>
	5788	+R: Amir Goldstein <amir73il@gmail.com>
	5789	+L: linux-fsdevel@vger.kernel.org
	5790	+S: Maintained
	5791	+F: fs/notify/
	5792	+F: include/linux/fsnotify*.h
	5793	+
5787	5794	FUJITSU LAPTOP EXTRAS
5788	5795	M: Jonathan Woithe <jwoithe@just42.net>
5789	5796	L: platform-driver-x86@vger.kernel.org
@@ -6256,7 +6263,7 @@ S: Odd Fixes · 466 unmodified lines
6256	6263	F: drivers/media/usb/hdpvr/
6257	6264
6258	6265	HEWLETT PACKARD ENTERPRISE ILO NMI WATCHDOG DRIVER
6259		−M: Jimmy Vance <jimmy.vance@hpe.com>
	6266	+M: Jerry Hoemann <jerry.hoemann@hpe.com>
6260	6267	S: Supported
6261	6268	F: Documentation/watchdog/hpwdt.txt
6262	6269	F: drivers/watchdog/hpwdt.c
@@ -7396,16 +7403,6 @@ S: Obsolete · 1133 unmodified lines
7396	7403	F: include/uapi/linux/ipx.h
7397	7404	F: drivers/staging/ipx/
7398	7405
7399		−IRDA SUBSYSTEM
7400		−M: Samuel Ortiz <samuel@sortiz.org>
7401		−L: irda-users@lists.sourceforge.net (subscribers-only)
7402		−L: netdev@vger.kernel.org
7403		−W: http://irda.sourceforge.net/
7404		−S: Obsolete
7405		−T: git git://git.kernel.org/pub/scm/linux/kernel/git/sameo/irda-2.6.git
7406		−F: Documentation/networking/irda.txt
7407		−F: drivers/staging/irda/
7408		−
7409	7406	IRQ DOMAINS (IRQ NUMBER MAPPING LIBRARY)
7410	7407	M: Marc Zyngier <marc.zyngier@arm.com>
7411	7408	S: Maintained
@@ -7670,9 +7667,11 @@ L: linux-kbuild@vger.kernel.org · 258 unmodified lines
7670	7667	S: Maintained
7671	7668	F: Documentation/kbuild/
7672	7669	F: Makefile
7673		−F: scripts/Makefile.*
	7670	+F: scripts/Kbuild*
	7671	+F: scripts/Makefile*
7674	7672	F: scripts/basic/
7675	7673	F: scripts/mk*
	7674	+F: scripts/mod/
7676	7675	F: scripts/package/
7677	7676
7678	7677	KERNEL JANITORS
@@ -7738,7 +7737,7 @@ F: arch/x86/include/asm/svm.h · 59 unmodified lines
7738	7737	F: arch/x86/kvm/svm.c
7739	7738
7740	7739	KERNEL VIRTUAL MACHINE FOR ARM (KVM/arm)
7741		−M: Christoffer Dall <christoffer.dall@linaro.org>
	7740	+M: Christoffer Dall <christoffer.dall@arm.com>
7742	7741	M: Marc Zyngier <marc.zyngier@arm.com>
7743	7742	L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers)
7744	7743	L: kvmarm@lists.cs.columbia.edu
@@ -7752,7 +7751,7 @@ F: virt/kvm/arm/ · 7 unmodified lines
7752	7751	F: include/kvm/arm_*
7753	7752
7754	7753	KERNEL VIRTUAL MACHINE FOR ARM64 (KVM/arm64)
7755		−M: Christoffer Dall <christoffer.dall@linaro.org>
	7754	+M: Christoffer Dall <christoffer.dall@arm.com>
7756	7755	M: Marc Zyngier <marc.zyngier@arm.com>
7757	7756	L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers)
7758	7757	L: kvmarm@lists.cs.columbia.edu
@@ -8048,6 +8047,9 @@ F: tools/lib/lockdep/ · 289 unmodified lines
8048	8047
8049	8048	LIBNVDIMM BLK: MMIO-APERTURE DRIVER
8050	8049	M: Ross Zwisler <ross.zwisler@linux.intel.com>
	8050	+M: Dan Williams <dan.j.williams@intel.com>
	8051	+M: Vishal Verma <vishal.l.verma@intel.com>
	8052	+M: Dave Jiang <dave.jiang@intel.com>
8051	8053	L: linux-nvdimm@lists.01.org
8052	8054	Q: https://patchwork.kernel.org/project/linux-nvdimm/list/
8053	8055	S: Supported
@@ -8056,6 +8058,9 @@ F: drivers/nvdimm/region_devs.c · 2 unmodified lines
8056	8058
8057	8059	LIBNVDIMM BTT: BLOCK TRANSLATION TABLE
8058	8060	M: Vishal Verma <vishal.l.verma@intel.com>
	8061	+M: Dan Williams <dan.j.williams@intel.com>
	8062	+M: Ross Zwisler <ross.zwisler@linux.intel.com>
	8063	+M: Dave Jiang <dave.jiang@intel.com>
8059	8064	L: linux-nvdimm@lists.01.org
8060	8065	Q: https://patchwork.kernel.org/project/linux-nvdimm/list/
8061	8066	S: Supported
@@ -8063,6 +8068,9 @@ F: drivers/nvdimm/btt* · 1 unmodified line
8063	8068
8064	8069	LIBNVDIMM PMEM: PERSISTENT MEMORY DRIVER
8065	8070	M: Ross Zwisler <ross.zwisler@linux.intel.com>
	8071	+M: Dan Williams <dan.j.williams@intel.com>
	8072	+M: Vishal Verma <vishal.l.verma@intel.com>
	8073	+M: Dave Jiang <dave.jiang@intel.com>
8066	8074	L: linux-nvdimm@lists.01.org
8067	8075	Q: https://patchwork.kernel.org/project/linux-nvdimm/list/
8068	8076	S: Supported
@@ -8078,6 +8086,9 @@ F: Documentation/devicetree/bindings/pmem/pmem-region.txt · 9 unmodified lines
8078	8086
8079	8087	LIBNVDIMM: NON-VOLATILE MEMORY DEVICE SUBSYSTEM
8080	8088	M: Dan Williams <dan.j.williams@intel.com>
	8089	+M: Ross Zwisler <ross.zwisler@linux.intel.com>
	8090	+M: Vishal Verma <vishal.l.verma@intel.com>
	8091	+M: Dave Jiang <dave.jiang@intel.com>
8081	8092	L: linux-nvdimm@lists.01.org
8082	8093	Q: https://patchwork.kernel.org/project/linux-nvdimm/list/
8083	8094	T: git git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm.git
@@ -8852,7 +8863,6 @@ F: Documentation/devicetree/bindings/media/nvidia,tegra-vde.txt · 768 unmodified lines
8852	8863	F: drivers/staging/media/tegra-vde/
8853	8864
8854	8865	MEDIA INPUT INFRASTRUCTURE (V4L/DVB)
8855		−M: Mauro Carvalho Chehab <mchehab@s-opensource.com>
8856	8866	M: Mauro Carvalho Chehab <mchehab@kernel.org>
8857	8867	P: LinuxTV.org Project
8858	8868	L: linux-media@vger.kernel.org
@@ -9706,6 +9716,7 @@ W: https://fedorahosted.org/dropwatch/ · 847 unmodified lines
9706	9716	F: net/core/drop_monitor.c
9707	9717
9708	9718	NETWORKING DRIVERS
	9719	+M: "David S. Miller" <davem@davemloft.net>
9709	9720	L: netdev@vger.kernel.org
9710	9721	W: http://www.linuxfoundation.org/en/Net
9711	9722	Q: http://patchwork.ozlabs.org/project/netdev/list/
@@ -9765,6 +9776,7 @@ F: include/uapi/linux/net_namespace.h · 53 unmodified lines
9765	9776	F: tools/testing/selftests/net/
9766	9777	F: lib/net_utils.c
9767	9778	F: lib/random32.c
	9779	+F: Documentation/networking/
9768	9780
9769	9781	NETWORKING [IPSEC]
9770	9782	M: Steffen Klassert <steffen.klassert@secunet.com>
@@ -10881,7 +10893,6 @@ F: drivers/pci/host/ · 1110 unmodified lines
10881	10893	F: drivers/pci/dwc/
10882	10894
10883	10895	PCIE DRIVER FOR AXIS ARTPEC
10884		−M: Niklas Cassel <niklas.cassel@axis.com>
10885	10896	M: Jesper Nilsson <jesper.nilsson@axis.com>
10886	10897	L: linux-arm-kernel@axis.com
10887	10898	L: linux-pci@vger.kernel.org
@@ -12240,7 +12251,6 @@ S: Odd Fixes · 1352 unmodified lines
12240	12251	F: drivers/media/i2c/saa6588*
12241	12252
12242	12253	SAA7134 VIDEO4LINUX DRIVER
12243		−M: Mauro Carvalho Chehab <mchehab@s-opensource.com>
12244	12254	M: Mauro Carvalho Chehab <mchehab@kernel.org>
12245	12255	L: linux-media@vger.kernel.org
12246	12256	W: https://linuxtv.org
@@ -12479,6 +12489,7 @@ F: drivers/scsi/st_*.h · 232 unmodified lines
12479	12489	SCTP PROTOCOL
12480	12490	M: Vlad Yasevich <vyasevich@gmail.com>
12481	12491	M: Neil Horman <nhorman@tuxdriver.com>
	12492	+M: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
12482	12493	L: linux-sctp@vger.kernel.org
12483	12494	W: http://lksctp.sourceforge.net
12484	12495	S: Maintained
@@ -12744,7 +12755,6 @@ S: Maintained · 259 unmodified lines
12744	12755	F: drivers/media/radio/si4713/radio-usb-si4713.c
12745	12756
12746	12757	SIANO DVB DRIVER
12747		−M: Mauro Carvalho Chehab <mchehab@s-opensource.com>
12748	12758	M: Mauro Carvalho Chehab <mchehab@kernel.org>
12749	12759	L: linux-media@vger.kernel.org
12750	12760	W: https://linuxtv.org
@@ -12816,7 +12826,8 @@ F: include/linux/siphash.h · 65 unmodified lines
12816	12826
12817	12827	SIOX
12818	12828	M: Gavin Schenk <g.schenk@eckelmann.de>
12819		−M: Uwe Kleine-König <kernel@pengutronix.de>
	12829	+M: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
	12830	+R: Pengutronix Kernel Team <kernel@pengutronix.de>
12820	12831	S: Supported
12821	12832	F: drivers/siox/*
12822	12833	F: include/trace/events/siox.h
@@ -13734,7 +13745,6 @@ S: Maintained · 911 unmodified lines
13734	13745	F: drivers/media/i2c/tda9840*
13735	13746
13736	13747	TEA5761 TUNER DRIVER
13737		−M: Mauro Carvalho Chehab <mchehab@s-opensource.com>
13738	13748	M: Mauro Carvalho Chehab <mchehab@kernel.org>
13739	13749	L: linux-media@vger.kernel.org
13740	13750	W: https://linuxtv.org
@@ -13743,7 +13753,6 @@ S: Odd fixes · 2 unmodified lines
13743	13753	F: drivers/media/tuners/tea5761.*
13744	13754
13745	13755	TEA5767 TUNER DRIVER
13746		−M: Mauro Carvalho Chehab <mchehab@s-opensource.com>
13747	13756	M: Mauro Carvalho Chehab <mchehab@kernel.org>
13748	13757	L: linux-media@vger.kernel.org
13749	13758	W: https://linuxtv.org
@@ -13833,7 +13842,6 @@ S: Supported · 83 unmodified lines
13833	13842	F: drivers/iommu/tegra*
13834	13843
13835	13844	TEGRA KBC DRIVER
13836		−M: Rakesh Iyer <riyer@nvidia.com>
13837	13845	M: Laxman Dewangan <ldewangan@nvidia.com>
13838	13846	S: Supported
13839	13847	F: drivers/input/keyboard/tegra-kbc.c
@@ -13936,7 +13944,7 @@ THUNDERBOLT DRIVER · 96 unmodified lines
13936	13944	M: Andreas Noever <andreas.noever@gmail.com>
13937	13945	M: Michael Jamet <michael.jamet@intel.com>
13938	13946	M: Mika Westerberg <mika.westerberg@linux.intel.com>
13939		−M: Yehezkel Bernat <yehezkel.bernat@intel.com>
	13947	+M: Yehezkel Bernat <YehezkelShB@gmail.com>
13940	13948	T: git git://git.kernel.org/pub/scm/linux/kernel/git/westeri/thunderbolt.git
13941	13949	S: Maintained
13942	13950	F: Documentation/admin-guide/thunderbolt.rst
@@ -13946,7 +13954,7 @@ F: include/linux/thunderbolt.h · 3 unmodified lines
13946	13954	THUNDERBOLT NETWORK DRIVER
13947	13955	M: Michael Jamet <michael.jamet@intel.com>
13948	13956	M: Mika Westerberg <mika.westerberg@linux.intel.com>
13949		−M: Yehezkel Bernat <yehezkel.bernat@intel.com>
	13957	+M: Yehezkel Bernat <YehezkelShB@gmail.com>
13950	13958	L: netdev@vger.kernel.org
13951	13959	S: Maintained
13952	13960	F: drivers/net/thunderbolt.c
@@ -14160,7 +14168,6 @@ F: Documentation/networking/tlan.txt · 207 unmodified lines
14160	14168	F: drivers/net/ethernet/ti/tlan.*
14161	14169
14162	14170	TM6000 VIDEO4LINUX DRIVER
14163		−M: Mauro Carvalho Chehab <mchehab@s-opensource.com>
14164	14171	M: Mauro Carvalho Chehab <mchehab@kernel.org>
14165	14172	L: linux-media@vger.kernel.org
14166	14173	W: https://linuxtv.org
@@ -15393,7 +15400,6 @@ S: Maintained · 1226 unmodified lines
15393	15400	F: arch/x86/entry/vdso/
15394	15401
15395	15402	XC2028/3028 TUNER DRIVER
15396		−M: Mauro Carvalho Chehab <mchehab@s-opensource.com>
15397	15403	M: Mauro Carvalho Chehab <mchehab@kernel.org>
15398	15404	L: linux-media@vger.kernel.org
15399	15405	W: https://linuxtv.org

Makefile+2−2

@@ -2,8 +2,8 @@ · 1 unmodified line
2	2	VERSION = 4
3	3	PATCHLEVEL = 17
4	4	SUBLEVEL = 0
5		−EXTRAVERSION = -rc1
6		−NAME = Fearless Coyote
	5	+EXTRAVERSION = -rc4
	6	+NAME = Merciless Moray
7	7
8	8	# DOCUMENTATION
9	9	# To see a list of typical targets execute "make help"

arch/arm/boot/dts/gemini-nas4220b.dts+14−14

@@ -134,37 +134,37 @@ · 133 unmodified lines
134	134	function = "gmii";
135	135	groups = "gmii_gmac0_grp";
136	136	};
137		− /* Settings come from OpenWRT */
	137	+ /* Settings come from OpenWRT, pins on SL3516 */
138	138	conf0 {
139		− pins = "R8 GMAC0 RXDV", "U11 GMAC1 RXDV";
	139	+ pins = "V8 GMAC0 RXDV", "T10 GMAC1 RXDV";
140	140	skew-delay = <0>;
141	141	};
142	142	conf1 {
143		− pins = "T8 GMAC0 RXC", "T11 GMAC1 RXC";
	143	+ pins = "Y7 GMAC0 RXC", "Y11 GMAC1 RXC";
144	144	skew-delay = <15>;
145	145	};
146	146	conf2 {
147		− pins = "P8 GMAC0 TXEN", "V11 GMAC1 TXEN";
	147	+ pins = "T8 GMAC0 TXEN", "W11 GMAC1 TXEN";
148	148	skew-delay = <7>;
149	149	};
150	150	conf3 {
151		− pins = "V7 GMAC0 TXC";
	151	+ pins = "U8 GMAC0 TXC";
152	152	skew-delay = <11>;
153	153	};
154	154	conf4 {
155		− pins = "P10 GMAC1 TXC";
	155	+ pins = "V11 GMAC1 TXC";
156	156	skew-delay = <10>;
157	157	};
158	158	conf5 {
159	159	/* The data lines all have default skew */
160		− pins = "U8 GMAC0 RXD0", "V8 GMAC0 RXD1",
161		− "P9 GMAC0 RXD2", "R9 GMAC0 RXD3",
162		− "U7 GMAC0 TXD0", "T7 GMAC0 TXD1",
163		− "R7 GMAC0 TXD2", "P7 GMAC0 TXD3",
164		− "R11 GMAC1 RXD0", "P11 GMAC1 RXD1",
165		− "V12 GMAC1 RXD2", "U12 GMAC1 RXD3",
166		− "R10 GMAC1 TXD0", "T10 GMAC1 TXD1",
167		− "U10 GMAC1 TXD2", "V10 GMAC1 TXD3";
	160	+ pins = "W8 GMAC0 RXD0", "V9 GMAC0 RXD1",
	161	+ "Y8 GMAC0 RXD2", "U9 GMAC0 RXD3",
	162	+ "T7 GMAC0 TXD0", "U6 GMAC0 TXD1",
	163	+ "V7 GMAC0 TXD2", "U7 GMAC0 TXD3",
	164	+ "Y12 GMAC1 RXD0", "V12 GMAC1 RXD1",
	165	+ "T11 GMAC1 RXD2", "W12 GMAC1 RXD3",
	166	+ "U10 GMAC1 TXD0", "Y10 GMAC1 TXD1",
	167	+ "W10 GMAC1 TXD2", "T9 GMAC1 TXD3";
168	168	skew-delay = <7>;
169	169	};
170	170	/* Set up drive strength on GMAC0 to 16 mA */

arch/arm/boot/dts/omap4.dtsi+4−4

@@ -163,10 +163,10 @@ · 162 unmodified lines
163	163
164	164	cm2: cm2@8000 {
165	165	compatible = "ti,omap4-cm2", "simple-bus";
166		− reg = <0x8000 0x3000>;
	166	+ reg = <0x8000 0x2000>;
167	167	#address-cells = <1>;
168	168	#size-cells = <1>;
169		− ranges = <0 0x8000 0x3000>;
	169	+ ranges = <0 0x8000 0x2000>;
170	170
171	171	cm2_clocks: clocks {
172	172	#address-cells = <1>;
@@ -250,11 +250,11 @@ · 77 unmodified lines
250	250
251	251	prm: prm@6000 {
252	252	compatible = "ti,omap4-prm";
253		− reg = <0x6000 0x3000>;
	253	+ reg = <0x6000 0x2000>;
254	254	interrupts = <GIC_SPI 11 IRQ_TYPE_LEVEL_HIGH>;
255	255	#address-cells = <1>;
256	256	#size-cells = <1>;
257		− ranges = <0 0x6000 0x3000>;
	257	+ ranges = <0 0x6000 0x2000>;
258	258
259	259	prm_clocks: clocks {
260	260	#address-cells = <1>;

arch/arm/configs/gemini_defconfig+25−2

@@ -1,6 +1,7 @@
1	1	# CONFIG_LOCALVERSION_AUTO is not set
2	2	CONFIG_SYSVIPC=y
3	3	CONFIG_NO_HZ_IDLE=y
	4	+CONFIG_HIGH_RES_TIMERS=y
4	5	CONFIG_BSD_PROCESS_ACCT=y
5	6	CONFIG_USER_NS=y
6	7	CONFIG_RELAY=y
@@ -12,15 +13,21 @@ CONFIG_ARCH_GEMINI=y · 5 unmodified lines
12	13	CONFIG_PCI=y
13	14	CONFIG_PREEMPT=y
14	15	CONFIG_AEABI=y
	16	+CONFIG_HIGHMEM=y
	17	+CONFIG_CMA=y
15	18	CONFIG_CMDLINE="console=ttyS0,115200n8"
16	19	CONFIG_KEXEC=y
17	20	CONFIG_BINFMT_MISC=y
18	21	CONFIG_PM=y
	22	+CONFIG_NET=y
	23	+CONFIG_UNIX=y
	24	+CONFIG_INET=y
19	25	CONFIG_UEVENT_HELPER_PATH="/sbin/hotplug"
20	26	CONFIG_DEVTMPFS=y
21	27	CONFIG_MTD=y
22	28	CONFIG_MTD_BLOCK=y
23	29	CONFIG_MTD_CFI=y
	30	+CONFIG_MTD_JEDECPROBE=y
24	31	CONFIG_MTD_CFI_INTELEXT=y
25	32	CONFIG_MTD_CFI_AMDSTD=y
26	33	CONFIG_MTD_CFI_STAA=y
@@ -33,6 +40,11 @@ CONFIG_BLK_DEV_SD=y · 6 unmodified lines
33	40	# CONFIG_SCSI_LOWLEVEL is not set
34	41	CONFIG_ATA=y
35	42	CONFIG_PATA_FTIDE010=y
	43	+CONFIG_NETDEVICES=y
	44	+CONFIG_GEMINI_ETHERNET=y
	45	+CONFIG_MDIO_BITBANG=y
	46	+CONFIG_MDIO_GPIO=y
	47	+CONFIG_REALTEK_PHY=y
36	48	CONFIG_INPUT_EVDEV=y
37	49	CONFIG_KEYBOARD_GPIO=y
38	50	# CONFIG_INPUT_MOUSE is not set
@@ -43,9 +55,19 @@ CONFIG_SERIAL_8250_NR_UARTS=1 · 4 unmodified lines
43	55	CONFIG_SERIAL_8250_RUNTIME_UARTS=1
44	56	CONFIG_SERIAL_OF_PLATFORM=y
45	57	# CONFIG_HW_RANDOM is not set
46		−# CONFIG_HWMON is not set
	58	+CONFIG_I2C_GPIO=y
	59	+CONFIG_SPI=y
	60	+CONFIG_SPI_GPIO=y
	61	+CONFIG_SENSORS_GPIO_FAN=y
	62	+CONFIG_SENSORS_LM75=y
	63	+CONFIG_THERMAL=y
47	64	CONFIG_WATCHDOG=y
48		−CONFIG_GEMINI_WATCHDOG=y
	65	+CONFIG_REGULATOR=y
	66	+CONFIG_REGULATOR_FIXED_VOLTAGE=y
	67	+CONFIG_DRM=y
	68	+CONFIG_DRM_PANEL_ILITEK_IL9322=y
	69	+CONFIG_DRM_TVE200=y
	70	+CONFIG_LOGO=y
49	71	CONFIG_USB=y
50	72	CONFIG_USB_MON=y
51	73	CONFIG_USB_FOTG210_HCD=y
@@ -54,6 +76,7 @@ CONFIG_NEW_LEDS=y · 2 unmodified lines
54	76	CONFIG_LEDS_CLASS=y
55	77	CONFIG_LEDS_GPIO=y
56	78	CONFIG_LEDS_TRIGGERS=y
	79	+CONFIG_LEDS_TRIGGER_DISK=y
57	80	CONFIG_LEDS_TRIGGER_HEARTBEAT=y
58	81	CONFIG_RTC_CLASS=y
59	82	CONFIG_DMADEVICES=y

arch/arm/configs/socfpga_defconfig+1−0

@@ -57,6 +57,7 @@ CONFIG_MTD_M25P80=y · 56 unmodified lines
57	57	CONFIG_MTD_NAND=y
58	58	CONFIG_MTD_NAND_DENALI_DT=y
59	59	CONFIG_MTD_SPI_NOR=y
	60	+# CONFIG_MTD_SPI_NOR_USE_4K_SECTORS is not set
60	61	CONFIG_SPI_CADENCE_QUADSPI=y
61	62	CONFIG_OF_OVERLAY=y
62	63	CONFIG_OF_CONFIGFS=y

arch/arm/include/asm/kvm_host.h+3−0

@@ -77,6 +77,9 @@ struct kvm_arch { · 76 unmodified lines
77	77	/* Interrupt controller */
78	78	struct vgic_dist vgic;
79	79	int max_vcpus;
	80	+
	81	+ /* Mandated version of PSCI */
	82	+ u32 psci_version;
80	83	};
81	84
82	85	#define KVM_NR_MEM_OBJS 40

arch/arm/include/uapi/asm/kvm.h+6−0

@@ -195,6 +195,12 @@ struct kvm_arch_memory_slot { · 194 unmodified lines
195	195	#define KVM_REG_ARM_VFP_FPINST 0x1009
196	196	#define KVM_REG_ARM_VFP_FPINST2 0x100A
197	197
	198	+/* KVM-as-firmware specific pseudo-registers */
	199	+#define KVM_REG_ARM_FW (0x0014 << KVM_REG_ARM_COPROC_SHIFT)
	200	+#define KVM_REG_ARM_FW_REG(r) (KVM_REG_ARM \| KVM_REG_SIZE_U64 \| \
	201	+ KVM_REG_ARM_FW \| ((r) & 0xffff))
	202	+#define KVM_REG_ARM_PSCI_VERSION KVM_REG_ARM_FW_REG(0)
	203	+
198	204	/* Device Control API: ARM VGIC */
199	205	#define KVM_DEV_ARM_VGIC_GRP_ADDR 0
200	206	#define KVM_DEV_ARM_VGIC_GRP_DIST_REGS 1

arch/arm/kvm/guest.c+13−0

@@ -22,6 +22,7 @@ · 21 unmodified lines
22	22	#include <linux/module.h>
23	23	#include <linux/vmalloc.h>
24	24	#include <linux/fs.h>
	25	+#include <kvm/arm_psci.h>
25	26	#include <asm/cputype.h>
26	27	#include <linux/uaccess.h>
27	28	#include <asm/kvm.h>
@@ -176,6 +177,7 @@ static unsigned long num_core_regs(void) · 148 unmodified lines
176	177	unsigned long kvm_arm_num_regs(struct kvm_vcpu *vcpu)
177	178	{
178	179	return num_core_regs() + kvm_arm_num_coproc_regs(vcpu)
	180	+ + kvm_arm_get_fw_num_regs(vcpu)
179	181	+ NUM_TIMER_REGS;
180	182	}
181	183
@@ -196,6 +198,11 @@ int kvm_arm_copy_reg_indices(struct kvm_vcpu vcpu, u64 __user uindices) · 14 unmodified lines
196	198	uindices++;
197	199	}
198	200
	201	+ ret = kvm_arm_copy_fw_reg_indices(vcpu, uindices);
	202	+ if (ret)
	203	+ return ret;
	204	+ uindices += kvm_arm_get_fw_num_regs(vcpu);
	205	+
199	206	ret = copy_timer_indices(vcpu, uindices);
200	207	if (ret)
201	208	return ret;
@@ -214,6 +221,9 @@ int kvm_arm_get_reg(struct kvm_vcpu vcpu, const struct kvm_one_reg reg) · 12 unmodified lines
214	221	if ((reg->id & KVM_REG_ARM_COPROC_MASK) == KVM_REG_ARM_CORE)
215	222	return get_core_reg(vcpu, reg);
216	223
	224	+ if ((reg->id & KVM_REG_ARM_COPROC_MASK) == KVM_REG_ARM_FW)
	225	+ return kvm_arm_get_fw_reg(vcpu, reg);
	226	+
217	227	if (is_timer_reg(reg->id))
218	228	return get_timer_reg(vcpu, reg);
219	229
@@ -230,6 +240,9 @@ int kvm_arm_set_reg(struct kvm_vcpu vcpu, const struct kvm_one_reg reg) · 10 unmodified lines
230	240	if ((reg->id & KVM_REG_ARM_COPROC_MASK) == KVM_REG_ARM_CORE)
231	241	return set_core_reg(vcpu, reg);
232	242
	243	+ if ((reg->id & KVM_REG_ARM_COPROC_MASK) == KVM_REG_ARM_FW)
	244	+ return kvm_arm_set_fw_reg(vcpu, reg);
	245	+
233	246	if (is_timer_reg(reg->id))
234	247	return set_timer_reg(vcpu, reg);
235	248

arch/arm/mach-omap2/Makefile+1−5

@@ -243,8 +243,4 @@ arch/arm/mach-omap2/pm-asm-offsets.s: arch/arm/mach-omap2/pm-asm-offsets.c · 242 unmodified lines
243	243	include/generated/ti-pm-asm-offsets.h: arch/arm/mach-omap2/pm-asm-offsets.s FORCE
244	244	$(call filechk,offsets,__TI_PM_ASM_OFFSETS_H__)
245	245
246		−# For rule to generate ti-emif-asm-offsets.h dependency
247		−include drivers/memory/Makefile.asm-offsets
248		−
249		−arch/arm/mach-omap2/sleep33xx.o: include/generated/ti-pm-asm-offsets.h include/generated/ti-emif-asm-offsets.h
250		−arch/arm/mach-omap2/sleep43xx.o: include/generated/ti-pm-asm-offsets.h include/generated/ti-emif-asm-offsets.h
	246	+$(obj)/sleep33xx.o $(obj)/sleep43xx.o: include/generated/ti-pm-asm-offsets.h

arch/arm/mach-omap2/pm-asm-offsets.c+3−0

@@ -7,9 +7,12 @@ · 6 unmodified lines
7	7
8	8	#include <linux/kbuild.h>
9	9	#include <linux/platform_data/pm33xx.h>
	10	+#include <linux/ti-emif-sram.h>
10	11
11	12	int main(void)
12	13	{
	14	+ ti_emif_asm_offsets();
	15	+
13	16	DEFINE(AMX3_PM_WFI_FLAGS_OFFSET,
14	17	offsetof(struct am33xx_pm_sram_data, wfi_flags));
15	18	DEFINE(AMX3_PM_L2_AUX_CTRL_VAL_OFFSET,

arch/arm/mach-omap2/sleep33xx.S+0−1

@@ -6,7 +6,6 @@ · 5 unmodified lines
6	6	* Dave Gerlach, Vaibhav Bedia
7	7	*/
8	8
9		−#include <generated/ti-emif-asm-offsets.h>
10	9	#include <generated/ti-pm-asm-offsets.h>
11	10	#include <linux/linkage.h>
12	11	#include <linux/ti-emif-sram.h>

arch/arm/mach-omap2/sleep43xx.S+0−1

@@ -6,7 +6,6 @@ · 5 unmodified lines
6	6	* Dave Gerlach, Vaibhav Bedia
7	7	*/
8	8
9		−#include <generated/ti-emif-asm-offsets.h>
10	9	#include <generated/ti-pm-asm-offsets.h>
11	10	#include <linux/linkage.h>
12	11	#include <linux/ti-emif-sram.h>

arch/arm/mach-s3c24xx/mach-jive.c+2−2

@@ -427,9 +427,9 @@ static struct gpiod_lookup_table jive_wm8750_gpiod_table = { · 426 unmodified lines
427	427	.dev_id = "spi_gpio",
428	428	.table = {
429	429	GPIO_LOOKUP("GPIOB", 4,
430		− "gpio-sck", GPIO_ACTIVE_HIGH),
	430	+ "sck", GPIO_ACTIVE_HIGH),
431	431	GPIO_LOOKUP("GPIOB", 9,
432		− "gpio-mosi", GPIO_ACTIVE_HIGH),
	432	+ "mosi", GPIO_ACTIVE_HIGH),
433	433	GPIO_LOOKUP("GPIOH", 10,
434	434	"cs", GPIO_ACTIVE_HIGH),
435	435	{ },

arch/arm64/Makefile+4−0

@@ -56,7 +56,11 @@ KBUILD_AFLAGS += $(lseinstr) $(brokengasinst) · 55 unmodified lines
56	56	KBUILD_CFLAGS += $(call cc-option,-mabi=lp64)
57	57	KBUILD_AFLAGS += $(call cc-option,-mabi=lp64)
58	58
	59	+ifeq ($(cc-name),clang)
	60	+KBUILD_CFLAGS += -DCONFIG_ARCH_SUPPORTS_INT128
	61	+else
59	62	KBUILD_CFLAGS += $(call cc-ifversion, -ge, 0500, -DCONFIG_ARCH_SUPPORTS_INT128)
	63	+endif
60	64
61	65	ifeq ($(CONFIG_CPU_BIG_ENDIAN), y)
62	66	KBUILD_CPPFLAGS += -mbig-endian

arch/arm64/boot/dts/amlogic/meson-gx-p23x-q20x.dtsi+4−0

@@ -212,3 +212,7 @@ · 211 unmodified lines
212	212	pinctrl-0 = <&uart_ao_a_pins>;
213	213	pinctrl-names = "default";
214	214	};
	215	+
	216	+&usb0 {
	217	+ status = "okay";
	218	+};

arch/arm64/boot/dts/amlogic/meson-gxl-s905x-libretech-cc.dts+12−0

@@ -271,3 +271,15 @@ · 270 unmodified lines
271	271	pinctrl-0 = <&uart_ao_a_pins>;
272	272	pinctrl-names = "default";
273	273	};
	274	+
	275	+&usb0 {
	276	+ status = "okay";
	277	+};
	278	+
	279	+&usb2_phy0 {
	280	+ /*
	281	+ * even though the schematics don't show it:
	282	+ * HDMI_5V is also used as supply for the USB VBUS.
	283	+ */
	284	+ phy-supply = <&hdmi_5v>;
	285	+};

arch/arm64/boot/dts/amlogic/meson-gxl-s905x-nexbox-a95x.dts+4−0

@@ -215,3 +215,7 @@ · 214 unmodified lines
215	215	pinctrl-0 = <&uart_ao_a_pins>;
216	216	pinctrl-names = "default";
217	217	};
	218	+
	219	+&usb0 {
	220	+ status = "okay";
	221	+};

arch/arm64/boot/dts/amlogic/meson-gxl-s905x-p212.dtsi+4−0

@@ -185,3 +185,7 @@ · 184 unmodified lines
185	185	pinctrl-0 = <&uart_ao_a_pins>;
186	186	pinctrl-names = "default";
187	187	};
	188	+
	189	+&usb0 {
	190	+ status = "okay";
	191	+};

arch/arm64/boot/dts/amlogic/meson-gxl.dtsi+61−0

@@ -20,6 +20,67 @@ · 19 unmodified lines
20	20	no-map;
21	21	};
22	22	};
	23	+
	24	+ soc {
	25	+ usb0: usb@c9000000 {
	26	+ status = "disabled";
	27	+ compatible = "amlogic,meson-gxl-dwc3";
	28	+ #address-cells = <2>;
	29	+ #size-cells = <2>;
	30	+ ranges;
	31	+
	32	+ clocks = <&clkc CLKID_USB>;
	33	+ clock-names = "usb_general";
	34	+ resets = <&reset RESET_USB_OTG>;
	35	+ reset-names = "usb_otg";
	36	+
	37	+ dwc3: dwc3@c9000000 {
	38	+ compatible = "snps,dwc3";
	39	+ reg = <0x0 0xc9000000 0x0 0x100000>;
	40	+ interrupts = <GIC_SPI 30 IRQ_TYPE_LEVEL_HIGH>;
	41	+ dr_mode = "host";
	42	+ maximum-speed = "high-speed";
	43	+ snps,dis_u2_susphy_quirk;
	44	+ phys = <&usb3_phy>, <&usb2_phy0>, <&usb2_phy1>;
	45	+ };
	46	+ };
	47	+ };
	48	+};
	49	+
	50	+&apb {
	51	+ usb2_phy0: phy@78000 {
	52	+ compatible = "amlogic,meson-gxl-usb2-phy";
	53	+ #phy-cells = <0>;
	54	+ reg = <0x0 0x78000 0x0 0x20>;
	55	+ clocks = <&clkc CLKID_USB>;
	56	+ clock-names = "phy";
	57	+ resets = <&reset RESET_USB_OTG>;
	58	+ reset-names = "phy";
	59	+ status = "okay";
	60	+ };
	61	+
	62	+ usb2_phy1: phy@78020 {
	63	+ compatible = "amlogic,meson-gxl-usb2-phy";
	64	+ #phy-cells = <0>;
	65	+ reg = <0x0 0x78020 0x0 0x20>;
	66	+ clocks = <&clkc CLKID_USB>;
	67	+ clock-names = "phy";
	68	+ resets = <&reset RESET_USB_OTG>;
	69	+ reset-names = "phy";
	70	+ status = "okay";
	71	+ };
	72	+
	73	+ usb3_phy: phy@78080 {
	74	+ compatible = "amlogic,meson-gxl-usb3-phy";
	75	+ #phy-cells = <0>;
	76	+ reg = <0x0 0x78080 0x0 0x20>;
	77	+ interrupts = <GIC_SPI 16 IRQ_TYPE_LEVEL_HIGH>;
	78	+ clocks = <&clkc CLKID_USB>, <&clkc_AO CLKID_AO_CEC_32K>;
	79	+ clock-names = "phy", "peripheral";
	80	+ resets = <&reset RESET_USB_OTG>, <&reset RESET_USB_OTG>;
	81	+ reset-names = "phy", "peripheral";
	82	+ status = "okay";
	83	+ };
23	84	};
24	85
25	86	&ethmac {

arch/arm64/boot/dts/amlogic/meson-gxm-khadas-vim2.dts+4−0

@@ -406,3 +406,7 @@ · 405 unmodified lines
406	406	status = "okay";
407	407	vref-supply = <&vddio_ao18>;
408	408	};
	409	+
	410	+&usb0 {
	411	+ status = "okay";
	412	+};

arch/arm64/boot/dts/amlogic/meson-gxm.dtsi+17−0

@@ -80,6 +80,19 @@ · 79 unmodified lines
80	80	};
81	81	};
82	82
	83	+&apb {
	84	+ usb2_phy2: phy@78040 {
	85	+ compatible = "amlogic,meson-gxl-usb2-phy";
	86	+ #phy-cells = <0>;
	87	+ reg = <0x0 0x78040 0x0 0x20>;
	88	+ clocks = <&clkc CLKID_USB>;
	89	+ clock-names = "phy";
	90	+ resets = <&reset RESET_USB_OTG>;
	91	+ reset-names = "phy";
	92	+ status = "okay";
	93	+ };
	94	+};
	95	+
83	96	&clkc_AO {
84	97	compatible = "amlogic,meson-gxm-aoclkc", "amlogic,meson-gx-aoclkc";
85	98	};
@@ -100,3 +113,7 @@ · 14 unmodified lines
100	113	&hdmi_tx {
101	114	compatible = "amlogic,meson-gxm-dw-hdmi", "amlogic,meson-gx-dw-hdmi";
102	115	};
	116	+
	117	+&dwc3 {
	118	+ phys = <&usb3_phy>, <&usb2_phy0>, <&usb2_phy1>, <&usb2_phy2>;
	119	+};

arch/arm64/boot/dts/arm/juno-motherboard.dtsi+0−2

@@ -56,8 +56,6 @@ · 55 unmodified lines
56	56
57	57	gpio_keys {
58	58	compatible = "gpio-keys";
59		− #address-cells = <1>;
60		− #size-cells = <0>;
61	59
62	60	power-button {
63	61	debounce_interval = <50>;

arch/arm64/boot/dts/broadcom/stingray/stingray-sata.dtsi+40−40

@@ -36,11 +36,11 @@ · 35 unmodified lines
36	36	#size-cells = <1>;
37	37	ranges = <0x0 0x0 0x67d00000 0x00800000>;
38	38
39		− sata0: ahci@210000 {
	39	+ sata0: ahci@0 {
40	40	compatible = "brcm,iproc-ahci", "generic-ahci";
41		− reg = <0x00210000 0x1000>;
	41	+ reg = <0x00000000 0x1000>;
42	42	reg-names = "ahci";
43		− interrupts = <GIC_SPI 339 IRQ_TYPE_LEVEL_HIGH>;
	43	+ interrupts = <GIC_SPI 321 IRQ_TYPE_LEVEL_HIGH>;
44	44	#address-cells = <1>;
45	45	#size-cells = <0>;
46	46	status = "disabled";
@@ -52,9 +52,9 @@ · 5 unmodified lines
52	52	};
53	53	};
54	54
55		− sata_phy0: sata_phy@212100 {
	55	+ sata_phy0: sata_phy@2100 {
56	56	compatible = "brcm,iproc-sr-sata-phy";
57		− reg = <0x00212100 0x1000>;
	57	+ reg = <0x00002100 0x1000>;
58	58	reg-names = "phy";
59	59	#address-cells = <1>;
60	60	#size-cells = <0>;
@@ -66,11 +66,11 @@ · 5 unmodified lines
66	66	};
67	67	};
68	68
69		− sata1: ahci@310000 {
	69	+ sata1: ahci@10000 {
70	70	compatible = "brcm,iproc-ahci", "generic-ahci";
71		− reg = <0x00310000 0x1000>;
	71	+ reg = <0x00010000 0x1000>;
72	72	reg-names = "ahci";
73		− interrupts = <GIC_SPI 347 IRQ_TYPE_LEVEL_HIGH>;
	73	+ interrupts = <GIC_SPI 323 IRQ_TYPE_LEVEL_HIGH>;
74	74	#address-cells = <1>;
75	75	#size-cells = <0>;
76	76	status = "disabled";
@@ -82,9 +82,9 @@ · 5 unmodified lines
82	82	};
83	83	};
84	84
85		− sata_phy1: sata_phy@312100 {
	85	+ sata_phy1: sata_phy@12100 {
86	86	compatible = "brcm,iproc-sr-sata-phy";
87		− reg = <0x00312100 0x1000>;
	87	+ reg = <0x00012100 0x1000>;
88	88	reg-names = "phy";
89	89	#address-cells = <1>;
90	90	#size-cells = <0>;
@@ -96,11 +96,11 @@ · 5 unmodified lines
96	96	};
97	97	};
98	98
99		− sata2: ahci@120000 {
	99	+ sata2: ahci@20000 {
100	100	compatible = "brcm,iproc-ahci", "generic-ahci";
101		− reg = <0x00120000 0x1000>;
	101	+ reg = <0x00020000 0x1000>;
102	102	reg-names = "ahci";
103		− interrupts = <GIC_SPI 333 IRQ_TYPE_LEVEL_HIGH>;
	103	+ interrupts = <GIC_SPI 325 IRQ_TYPE_LEVEL_HIGH>;
104	104	#address-cells = <1>;
105	105	#size-cells = <0>;
106	106	status = "disabled";
@@ -112,9 +112,9 @@ · 5 unmodified lines
112	112	};
113	113	};
114	114
115		− sata_phy2: sata_phy@122100 {
	115	+ sata_phy2: sata_phy@22100 {
116	116	compatible = "brcm,iproc-sr-sata-phy";
117		− reg = <0x00122100 0x1000>;
	117	+ reg = <0x00022100 0x1000>;
118	118	reg-names = "phy";
119	119	#address-cells = <1>;
120	120	#size-cells = <0>;
@@ -126,11 +126,11 @@ · 5 unmodified lines
126	126	};
127	127	};
128	128
129		− sata3: ahci@130000 {
	129	+ sata3: ahci@30000 {
130	130	compatible = "brcm,iproc-ahci", "generic-ahci";
131		− reg = <0x00130000 0x1000>;
	131	+ reg = <0x00030000 0x1000>;
132	132	reg-names = "ahci";
133		− interrupts = <GIC_SPI 335 IRQ_TYPE_LEVEL_HIGH>;
	133	+ interrupts = <GIC_SPI 327 IRQ_TYPE_LEVEL_HIGH>;
134	134	#address-cells = <1>;
135	135	#size-cells = <0>;
136	136	status = "disabled";
@@ -142,9 +142,9 @@ · 5 unmodified lines
142	142	};
143	143	};
144	144
145		− sata_phy3: sata_phy@132100 {
	145	+ sata_phy3: sata_phy@32100 {
146	146	compatible = "brcm,iproc-sr-sata-phy";
147		− reg = <0x00132100 0x1000>;
	147	+ reg = <0x00032100 0x1000>;
148	148	reg-names = "phy";
149	149	#address-cells = <1>;
150	150	#size-cells = <0>;
@@ -156,11 +156,11 @@ · 5 unmodified lines
156	156	};
157	157	};
158	158
159		− sata4: ahci@330000 {
	159	+ sata4: ahci@100000 {
160	160	compatible = "brcm,iproc-ahci", "generic-ahci";
161		− reg = <0x00330000 0x1000>;
	161	+ reg = <0x00100000 0x1000>;
162	162	reg-names = "ahci";
163		− interrupts = <GIC_SPI 351 IRQ_TYPE_LEVEL_HIGH>;
	163	+ interrupts = <GIC_SPI 329 IRQ_TYPE_LEVEL_HIGH>;
164	164	#address-cells = <1>;
165	165	#size-cells = <0>;
166	166	status = "disabled";
@@ -172,9 +172,9 @@ · 5 unmodified lines
172	172	};
173	173	};
174	174
175		− sata_phy4: sata_phy@332100 {
	175	+ sata_phy4: sata_phy@102100 {
176	176	compatible = "brcm,iproc-sr-sata-phy";
177		− reg = <0x00332100 0x1000>;
	177	+ reg = <0x00102100 0x1000>;
178	178	reg-names = "phy";
179	179	#address-cells = <1>;
180	180	#size-cells = <0>;
@@ -186,11 +186,11 @@ · 5 unmodified lines
186	186	};
187	187	};
188	188
189		− sata5: ahci@400000 {
	189	+ sata5: ahci@110000 {
190	190	compatible = "brcm,iproc-ahci", "generic-ahci";
191		− reg = <0x00400000 0x1000>;
	191	+ reg = <0x00110000 0x1000>;
192	192	reg-names = "ahci";
193		− interrupts = <GIC_SPI 353 IRQ_TYPE_LEVEL_HIGH>;
	193	+ interrupts = <GIC_SPI 331 IRQ_TYPE_LEVEL_HIGH>;
194	194	#address-cells = <1>;
195	195	#size-cells = <0>;
196	196	status = "disabled";
@@ -202,9 +202,9 @@ · 5 unmodified lines
202	202	};
203	203	};
204	204
205		− sata_phy5: sata_phy@402100 {
	205	+ sata_phy5: sata_phy@112100 {
206	206	compatible = "brcm,iproc-sr-sata-phy";
207		− reg = <0x00402100 0x1000>;
	207	+ reg = <0x00112100 0x1000>;
208	208	reg-names = "phy";
209	209	#address-cells = <1>;
210	210	#size-cells = <0>;
@@ -216,11 +216,11 @@ · 5 unmodified lines
216	216	};
217	217	};
218	218
219		− sata6: ahci@410000 {
	219	+ sata6: ahci@120000 {
220	220	compatible = "brcm,iproc-ahci", "generic-ahci";
221		− reg = <0x00410000 0x1000>;
	221	+ reg = <0x00120000 0x1000>;
222	222	reg-names = "ahci";
223		− interrupts = <GIC_SPI 355 IRQ_TYPE_LEVEL_HIGH>;
	223	+ interrupts = <GIC_SPI 333 IRQ_TYPE_LEVEL_HIGH>;
224	224	#address-cells = <1>;
225	225	#size-cells = <0>;
226	226	status = "disabled";
@@ -232,9 +232,9 @@ · 5 unmodified lines
232	232	};
233	233	};
234	234
235		− sata_phy6: sata_phy@412100 {
	235	+ sata_phy6: sata_phy@122100 {
236	236	compatible = "brcm,iproc-sr-sata-phy";
237		− reg = <0x00412100 0x1000>;
	237	+ reg = <0x00122100 0x1000>;
238	238	reg-names = "phy";
239	239	#address-cells = <1>;
240	240	#size-cells = <0>;
@@ -246,11 +246,11 @@ · 5 unmodified lines
246	246	};
247	247	};
248	248
249		− sata7: ahci@420000 {
	249	+ sata7: ahci@130000 {
250	250	compatible = "brcm,iproc-ahci", "generic-ahci";
251		− reg = <0x00420000 0x1000>;
	251	+ reg = <0x00130000 0x1000>;
252	252	reg-names = "ahci";
253		− interrupts = <GIC_SPI 357 IRQ_TYPE_LEVEL_HIGH>;
	253	+ interrupts = <GIC_SPI 335 IRQ_TYPE_LEVEL_HIGH>;
254	254	#address-cells = <1>;
255	255	#size-cells = <0>;
256	256	status = "disabled";
@@ -262,9 +262,9 @@ · 5 unmodified lines
262	262	};
263	263	};
264	264
265		− sata_phy7: sata_phy@422100 {
	265	+ sata_phy7: sata_phy@132100 {
266	266	compatible = "brcm,iproc-sr-sata-phy";
267		− reg = <0x00422100 0x1000>;
	267	+ reg = <0x00132100 0x1000>;
268	268	reg-names = "phy";
269	269	#address-cells = <1>;
270	270	#size-cells = <0>;

arch/arm64/include/asm/kvm_emulate.h+1−1

@@ -333,7 +333,7 @@ static inline void kvm_vcpu_set_be(struct kvm_vcpu *vcpu) · 332 unmodified lines
333	333	} else {
334	334	u64 sctlr = vcpu_read_sys_reg(vcpu, SCTLR_EL1);
335	335	sctlr \|= (1 << 25);
336		− vcpu_write_sys_reg(vcpu, SCTLR_EL1, sctlr);
	336	+ vcpu_write_sys_reg(vcpu, sctlr, SCTLR_EL1);
337	337	}
338	338	}
339	339

arch/arm64/include/asm/kvm_host.h+3−0

@@ -75,6 +75,9 @@ struct kvm_arch { · 74 unmodified lines
75	75
76	76	/* Interrupt controller */
77	77	struct vgic_dist vgic;
	78	+
	79	+ /* Mandated version of PSCI */
	80	+ u32 psci_version;
78	81	};
79	82
80	83	#define KVM_NR_MEM_OBJS 40

arch/arm64/include/asm/module.h+1−1

@@ -39,7 +39,7 @@ struct mod_arch_specific { · 38 unmodified lines
39	39	u64 module_emit_plt_entry(struct module mod, void loc, const Elf64_Rela *rela,
40	40	Elf64_Sym *sym);
41	41
42		−u64 module_emit_adrp_veneer(struct module mod, void loc, u64 val);
	42	+u64 module_emit_veneer_for_adrp(struct module mod, void loc, u64 val);
43	43
44	44	#ifdef CONFIG_RANDOMIZE_BASE
45	45	extern u64 module_alloc_base;

arch/arm64/include/asm/pgtable.h+2−2

@@ -230,7 +230,7 @@ static inline void set_pte(pte_t *ptep, pte_t pte) · 229 unmodified lines
230	230	}
231	231	}
232	232
233		−extern void __sync_icache_dcache(pte_t pteval, unsigned long addr);
	233	+extern void __sync_icache_dcache(pte_t pteval);
234	234
235	235	/*
236	236	* PTE bits configuration in the presence of hardware Dirty Bit Management
@@ -253,7 +253,7 @@ static inline void set_pte_at(struct mm_struct *mm, unsigned long addr, · 16 unmodified lines
253	253	pte_t old_pte;
254	254
255	255	if (pte_present(pte) && pte_user_exec(pte) && !pte_special(pte))
256		− __sync_icache_dcache(pte, addr);
	256	+ __sync_icache_dcache(pte);
257	257
258	258	/*
259	259	* If the existing pte is valid, check for potential race with

arch/arm64/include/uapi/asm/kvm.h+6−0

@@ -206,6 +206,12 @@ struct kvm_arch_memory_slot { · 205 unmodified lines
206	206	#define KVM_REG_ARM_TIMER_CNT ARM64_SYS_REG(3, 3, 14, 3, 2)
207	207	#define KVM_REG_ARM_TIMER_CVAL ARM64_SYS_REG(3, 3, 14, 0, 2)
208	208
	209	+/* KVM-as-firmware specific pseudo-registers */
	210	+#define KVM_REG_ARM_FW (0x0014 << KVM_REG_ARM_COPROC_SHIFT)
	211	+#define KVM_REG_ARM_FW_REG(r) (KVM_REG_ARM64 \| KVM_REG_SIZE_U64 \| \
	212	+ KVM_REG_ARM_FW \| ((r) & 0xffff))
	213	+#define KVM_REG_ARM_PSCI_VERSION KVM_REG_ARM_FW_REG(0)
	214	+
209	215	/* Device Control API: ARM VGIC */
210	216	#define KVM_DEV_ARM_VGIC_GRP_ADDR 0
211	217	#define KVM_DEV_ARM_VGIC_GRP_DIST_REGS 1

arch/arm64/kernel/cpufeature.c+1−0

@@ -868,6 +868,7 @@ static bool unmap_kernel_at_el0(const struct arm64_cpu_capabilities *entry, · 867 unmodified lines
868	868	static const struct midr_range kpti_safe_list[] = {
869	869	MIDR_ALL_VERSIONS(MIDR_CAVIUM_THUNDERX2),
870	870	MIDR_ALL_VERSIONS(MIDR_BRCM_VULCAN),
	871	+ { /* sentinel */ }
871	872	};
872	873	char const *str = "command line option";
873	874

arch/arm64/kernel/module-plts.c+1−1

@@ -43,7 +43,7 @@ u64 module_emit_plt_entry(struct module mod, void loc, const Elf64_Rela *rela, · 42 unmodified lines
43	43	}
44	44
45	45	#ifdef CONFIG_ARM64_ERRATUM_843419
46		−u64 module_emit_adrp_veneer(struct module mod, void loc, u64 val)
	46	+u64 module_emit_veneer_for_adrp(struct module mod, void loc, u64 val)
47	47	{
48	48	struct mod_plt_sec *pltsec = !in_init(mod, loc) ? &mod->arch.core :
49	49	&mod->arch.init;

arch/arm64/kernel/module.c+1−1

@@ -215,7 +215,7 @@ static int reloc_insn_adrp(struct module mod, __le32 place, u64 val) · 214 unmodified lines
215	215	insn &= ~BIT(31);
216	216	} else {
217	217	/* out of range for ADR -> emit a veneer */
218		− val = module_emit_adrp_veneer(mod, place, val & ~0xfff);
	218	+ val = module_emit_veneer_for_adrp(mod, place, val & ~0xfff);
219	219	if (!val)
220	220	return -ENOEXEC;
221	221	insn = aarch64_insn_gen_branch_imm((u64)place, val,

arch/arm64/kernel/ptrace.c+10−10

@@ -25,6 +25,7 @@ · 24 unmodified lines
25	25	#include <linux/sched/signal.h>
26	26	#include <linux/sched/task_stack.h>
27	27	#include <linux/mm.h>
	28	+#include <linux/nospec.h>
28	29	#include <linux/smp.h>
29	30	#include <linux/ptrace.h>
30	31	#include <linux/user.h>
@@ -249,15 +250,20 @@ static struct perf_event *ptrace_hbp_get_event(unsigned int note_type, · 218 unmodified lines
249	250
250	251	switch (note_type) {
251	252	case NT_ARM_HW_BREAK:
252		− if (idx < ARM_MAX_BRP)
253		− bp = tsk->thread.debug.hbp_break[idx];
	253	+ if (idx >= ARM_MAX_BRP)
	254	+ goto out;
	255	+ idx = array_index_nospec(idx, ARM_MAX_BRP);
	256	+ bp = tsk->thread.debug.hbp_break[idx];
254	257	break;
255	258	case NT_ARM_HW_WATCH:
256		− if (idx < ARM_MAX_WRP)
257		− bp = tsk->thread.debug.hbp_watch[idx];
	259	+ if (idx >= ARM_MAX_WRP)
	260	+ goto out;
	261	+ idx = array_index_nospec(idx, ARM_MAX_WRP);
	262	+ bp = tsk->thread.debug.hbp_watch[idx];
258	263	break;
259	264	}
260	265
	266	+out:
261	267	return bp;
262	268	}
263	269
@@ -1458,9 +1464,7 @@ static int compat_ptrace_gethbpregs(struct task_struct *tsk, compat_long_t num, · 1194 unmodified lines
1458	1464	{
1459	1465	int ret;
1460	1466	u32 kdata;
1461		− mm_segment_t old_fs = get_fs();
1462	1467
1463		− set_fs(KERNEL_DS);
1464	1468	/* Watchpoint */
1465	1469	if (num < 0) {
1466	1470	ret = compat_ptrace_hbp_get(NT_ARM_HW_WATCH, tsk, num, &kdata);
@@ -1471,7 +1475,6 @@ static int compat_ptrace_gethbpregs(struct task_struct *tsk, compat_long_t num, · 4 unmodified lines
1471	1475	} else {
1472	1476	ret = compat_ptrace_hbp_get(NT_ARM_HW_BREAK, tsk, num, &kdata);
1473	1477	}
1474		− set_fs(old_fs);
1475	1478
1476	1479	if (!ret)
1477	1480	ret = put_user(kdata, data);
@@ -1484,7 +1487,6 @@ static int compat_ptrace_sethbpregs(struct task_struct *tsk, compat_long_t num, · 6 unmodified lines
1484	1487	{
1485	1488	int ret;
1486	1489	u32 kdata = 0;
1487		− mm_segment_t old_fs = get_fs();
1488	1490
1489	1491	if (num == 0)
1490	1492	return 0;
@@ -1493,12 +1495,10 @@ static int compat_ptrace_sethbpregs(struct task_struct *tsk, compat_long_t num, · 2 unmodified lines
1493	1495	if (ret)
1494	1496	return ret;
1495	1497
1496		− set_fs(KERNEL_DS);
1497	1498	if (num < 0)
1498	1499	ret = compat_ptrace_hbp_set(NT_ARM_HW_WATCH, tsk, num, &kdata);
1499	1500	else
1500	1501	ret = compat_ptrace_hbp_set(NT_ARM_HW_BREAK, tsk, num, &kdata);
1501		− set_fs(old_fs);
1502	1502
1503	1503	return ret;
1504	1504	}

arch/arm64/kernel/traps.c+3−2

@@ -277,7 +277,8 @@ void arm64_skip_faulting_instruction(struct pt_regs *regs, unsigned long size) · 276 unmodified lines
277	277	* If we were single stepping, we want to get the step exception after
278	278	* we return from the trap.
279	279	*/
280		− user_fastforward_single_step(current);
	280	+ if (user_mode(regs))
	281	+ user_fastforward_single_step(current);
281	282	}
282	283
283	284	static LIST_HEAD(undef_hook);
@@ -366,7 +367,7 @@ void force_signal_inject(int signal, int code, unsigned long address) · 82 unmodified lines
366	367	}
367	368
368	369	/* Force signals we don't understand to SIGKILL */
369		− if (WARN_ON(signal != SIGKILL \|\|
	370	+ if (WARN_ON(signal != SIGKILL &&
370	371	siginfo_layout(signal, code) != SIL_FAULT)) {
371	372	signal = SIGKILL;
372	373	}

arch/arm64/kvm/guest.c+13−1

@@ -25,6 +25,7 @@ · 24 unmodified lines
25	25	#include <linux/module.h>
26	26	#include <linux/vmalloc.h>
27	27	#include <linux/fs.h>
	28	+#include <kvm/arm_psci.h>
28	29	#include <asm/cputype.h>
29	30	#include <linux/uaccess.h>
30	31	#include <asm/kvm.h>
@@ -205,7 +206,7 @@ static int get_timer_reg(struct kvm_vcpu vcpu, const struct kvm_one_reg reg) · 174 unmodified lines
205	206	unsigned long kvm_arm_num_regs(struct kvm_vcpu *vcpu)
206	207	{
207	208	return num_core_regs() + kvm_arm_num_sys_reg_descs(vcpu)
208		− + NUM_TIMER_REGS;
	209	+ + kvm_arm_get_fw_num_regs(vcpu) + NUM_TIMER_REGS;
209	210	}
210	211
211	212	/**
@@ -225,6 +226,11 @@ int kvm_arm_copy_reg_indices(struct kvm_vcpu vcpu, u64 __user uindices) · 13 unmodified lines
225	226	uindices++;
226	227	}
227	228
	229	+ ret = kvm_arm_copy_fw_reg_indices(vcpu, uindices);
	230	+ if (ret)
	231	+ return ret;
	232	+ uindices += kvm_arm_get_fw_num_regs(vcpu);
	233	+
228	234	ret = copy_timer_indices(vcpu, uindices);
229	235	if (ret)
230	236	return ret;
@@ -243,6 +249,9 @@ int kvm_arm_get_reg(struct kvm_vcpu vcpu, const struct kvm_one_reg reg) · 12 unmodified lines
243	249	if ((reg->id & KVM_REG_ARM_COPROC_MASK) == KVM_REG_ARM_CORE)
244	250	return get_core_reg(vcpu, reg);
245	251
	252	+ if ((reg->id & KVM_REG_ARM_COPROC_MASK) == KVM_REG_ARM_FW)
	253	+ return kvm_arm_get_fw_reg(vcpu, reg);
	254	+
246	255	if (is_timer_reg(reg->id))
247	256	return get_timer_reg(vcpu, reg);
248	257
@@ -259,6 +268,9 @@ int kvm_arm_set_reg(struct kvm_vcpu vcpu, const struct kvm_one_reg reg) · 10 unmodified lines
259	268	if ((reg->id & KVM_REG_ARM_COPROC_MASK) == KVM_REG_ARM_CORE)
260	269	return set_core_reg(vcpu, reg);
261	270
	271	+ if ((reg->id & KVM_REG_ARM_COPROC_MASK) == KVM_REG_ARM_FW)
	272	+ return kvm_arm_set_fw_reg(vcpu, reg);
	273	+
262	274	if (is_timer_reg(reg->id))
263	275	return set_timer_reg(vcpu, reg);
264	276

arch/arm64/kvm/hyp/vgic-v2-cpuif-proxy.c+19−5

@@ -18,11 +18,20 @@ · 17 unmodified lines
18	18	#include <linux/compiler.h>
19	19	#include <linux/irqchip/arm-gic.h>
20	20	#include <linux/kvm_host.h>
	21	+#include <linux/swab.h>
21	22
22	23	#include <asm/kvm_emulate.h>
23	24	#include <asm/kvm_hyp.h>
24	25	#include <asm/kvm_mmu.h>
25	26
	27	+static bool __hyp_text __is_be(struct kvm_vcpu *vcpu)
	28	+{
	29	+ if (vcpu_mode_is_32bit(vcpu))
	30	+ return !!(read_sysreg_el2(spsr) & COMPAT_PSR_E_BIT);
	31	+
	32	+ return !!(read_sysreg(SCTLR_EL1) & SCTLR_ELx_EE);
	33	+}
	34	+
26	35	/*
27	36	* __vgic_v2_perform_cpuif_access -- perform a GICV access on behalf of the
28	37	* guest.
@@ -64,14 +73,19 @@ int __hyp_text __vgic_v2_perform_cpuif_access(struct kvm_vcpu *vcpu) · 35 unmodified lines
64	73	addr += fault_ipa - vgic->vgic_cpu_base;
65	74
66	75	if (kvm_vcpu_dabt_iswrite(vcpu)) {
67		− u32 data = vcpu_data_guest_to_host(vcpu,
68		− vcpu_get_reg(vcpu, rd),
69		− sizeof(u32));
	76	+ u32 data = vcpu_get_reg(vcpu, rd);
	77	+ if (__is_be(vcpu)) {
	78	+ /* guest pre-swabbed data, undo this for writel() */
	79	+ data = swab32(data);
	80	+ }
70	81	writel_relaxed(data, addr);
71	82	} else {
72	83	u32 data = readl_relaxed(addr);
73		− vcpu_set_reg(vcpu, rd, vcpu_data_host_to_guest(vcpu, data,
74		− sizeof(u32)));
	84	+ if (__is_be(vcpu)) {
	85	+ /* guest expects swabbed data */
	86	+ data = swab32(data);
	87	+ }
	88	+ vcpu_set_reg(vcpu, rd, data);
75	89	}
76	90
77	91	return 1;

arch/arm64/kvm/sys_regs.c+2−4

@@ -996,14 +996,12 @@ static u64 read_id_reg(struct sys_reg_desc const *r, bool raz) · 995 unmodified lines
996	996
997	997	if (id == SYS_ID_AA64PFR0_EL1) {
998	998	if (val & (0xfUL << ID_AA64PFR0_SVE_SHIFT))
999		− pr_err_once("kvm [%i]: SVE unsupported for guests, suppressing\n",
1000		− task_pid_nr(current));
	999	+ kvm_debug("SVE unsupported for guests, suppressing\n");
1001	1000
1002	1001	val &= ~(0xfUL << ID_AA64PFR0_SVE_SHIFT);
1003	1002	} else if (id == SYS_ID_AA64MMFR1_EL1) {
1004	1003	if (val & (0xfUL << ID_AA64MMFR1_LOR_SHIFT))
1005		− pr_err_once("kvm [%i]: LORegions unsupported for guests, suppressing\n",
1006		− task_pid_nr(current));
	1004	+ kvm_debug("LORegions unsupported for guests, suppressing\n");
1007	1005
1008	1006	val &= ~(0xfUL << ID_AA64MMFR1_LOR_SHIFT);
1009	1007	}

arch/arm64/lib/Makefile+4−0

@@ -19,5 +19,9 @@ CFLAGS_atomic_ll_sc.o := -fcall-used-x0 -ffixed-x1 -ffixed-x2 \ · 18 unmodified lines
19	19	-fcall-saved-x13 -fcall-saved-x14 -fcall-saved-x15 \
20	20	-fcall-saved-x18 -fomit-frame-pointer
21	21	CFLAGS_REMOVE_atomic_ll_sc.o := -pg
	22	+GCOV_PROFILE_atomic_ll_sc.o := n
	23	+KASAN_SANITIZE_atomic_ll_sc.o := n
	24	+KCOV_INSTRUMENT_atomic_ll_sc.o := n
	25	+UBSAN_SANITIZE_atomic_ll_sc.o := n
22	26
23	27	lib-$(CONFIG_ARCH_HAS_UACCESS_FLUSHCACHE) += uaccess_flushcache.o

arch/arm64/mm/flush.c+1−1

@@ -58,7 +58,7 @@ void copy_to_user_page(struct vm_area_struct vma, struct page page, · 57 unmodified lines
58	58	flush_ptrace_access(vma, page, uaddr, dst, len);
59	59	}
60	60
61		−void __sync_icache_dcache(pte_t pte, unsigned long addr)
	61	+void __sync_icache_dcache(pte_t pte)
62	62	{
63	63	struct page *page = pte_page(pte);
64	64

arch/arm64/mm/kasan_init.c+2−2

@@ -204,7 +204,7 @@ void __init kasan_init(void) · 203 unmodified lines
204	204	clear_pgds(KASAN_SHADOW_START, KASAN_SHADOW_END);
205	205
206	206	kasan_map_populate(kimg_shadow_start, kimg_shadow_end,
207		− pfn_to_nid(virt_to_pfn(lm_alias(_text))));
	207	+ early_pfn_to_nid(virt_to_pfn(lm_alias(_text))));
208	208
209	209	kasan_populate_zero_shadow((void *)KASAN_SHADOW_START,
210	210	(void *)mod_shadow_start);
@@ -224,7 +224,7 @@ void __init kasan_init(void) · 13 unmodified lines
224	224
225	225	kasan_map_populate((unsigned long)kasan_mem_to_shadow(start),
226	226	(unsigned long)kasan_mem_to_shadow(end),
227		− pfn_to_nid(virt_to_pfn(start)));
	227	+ early_pfn_to_nid(virt_to_pfn(start)));
228	228	}
229	229
230	230	/*

arch/hexagon/include/asm/io.h+6−0

@@ -216,6 +216,12 @@ static inline void memcpy_toio(volatile void __iomem dst, const void src, · 215 unmodified lines
216	216	memcpy((void *) dst, src, count);
217	217	}
218	218
	219	+static inline void memset_io(volatile void __iomem *addr, int value,
	220	+ size_t size)
	221	+{
	222	+ memset((void __force *)addr, value, size);
	223	+}
	224	+
219	225	#define PCI_IO_ADDR (volatile void __iomem *)
220	226
221	227	/*

arch/hexagon/lib/checksum.c+1−0

@@ -199,3 +199,4 @@ csum_partial_copy_nocheck(const void src, void dst, int len, __wsum sum) · 198 unmodified lines
199	199	memcpy(dst, src, len);
200	200	return csum_partial(dst, len, sum);
201	201	}
	202	+EXPORT_SYMBOL(csum_partial_copy_nocheck);

arch/mips/boot/dts/img/boston.dts+6−0

@@ -51,6 +51,8 @@ · 50 unmodified lines
51	51	ranges = <0x02000000 0 0x40000000
52	52	0x40000000 0 0x40000000>;
53	53
	54	+ bus-range = <0x00 0xff>;
	55	+
54	56	interrupt-map-mask = <0 0 0 7>;
55	57	interrupt-map = <0 0 0 1 &pci0_intc 1>,
56	58	<0 0 0 2 &pci0_intc 2>,
@@ -79,6 +81,8 @@ · 22 unmodified lines
79	81	ranges = <0x02000000 0 0x20000000
80	82	0x20000000 0 0x20000000>;
81	83
	84	+ bus-range = <0x00 0xff>;
	85	+
82	86	interrupt-map-mask = <0 0 0 7>;
83	87	interrupt-map = <0 0 0 1 &pci1_intc 1>,
84	88	<0 0 0 2 &pci1_intc 2>,
@@ -107,6 +111,8 @@ · 22 unmodified lines
107	111	ranges = <0x02000000 0 0x16000000
108	112	0x16000000 0 0x100000>;
109	113
	114	+ bus-range = <0x00 0xff>;
	115	+
110	116	interrupt-map-mask = <0 0 0 7>;
111	117	interrupt-map = <0 0 0 1 &pci2_intc 1>,
112	118	<0 0 0 2 &pci2_intc 2>,

arch/mips/include/asm/io.h+3−1

@@ -307,7 +307,7 @@ static inline void iounmap(const volatile void __iomem *addr) · 306 unmodified lines
307	307	#if defined(CONFIG_CPU_CAVIUM_OCTEON) \|\| defined(CONFIG_LOONGSON3_ENHANCEMENT)
308	308	#define war_io_reorder_wmb() wmb()
309	309	#else
310		−#define war_io_reorder_wmb() do { } while (0)
	310	+#define war_io_reorder_wmb() barrier()
311	311	#endif
312	312
313	313	#define __BUILD_MEMORY_SINGLE(pfx, bwlq, type, irq) \
@@ -377,6 +377,8 @@ static inline type pfx##read##bwlq(const volatile void __iomem *mem) \ · 63 unmodified lines
377	377	BUG(); \
378	378	} \
379	379	\
	380	+ /* prevent prefetching of coherent DMA data prematurely */ \
	381	+ rmb(); \
380	382	return pfx##ioswab##bwlq(__mem, __val); \
381	383	}
382	384

arch/mips/include/asm/uaccess.h+9−2

@@ -654,6 +654,13 @@ __clear_user(void __user *addr, __kernel_size_t size) · 653 unmodified lines
654	654	{
655	655	__kernel_size_t res;
656	656
	657	+#ifdef CONFIG_CPU_MICROMIPS
	658	+/* micromips memset / bzero also clobbers t7 & t8 */
	659	+#define bzero_clobbers "$4", "$5", "$6", __UA_t0, __UA_t1, "$15", "$24", "$31"
	660	+#else
	661	+#define bzero_clobbers "$4", "$5", "$6", __UA_t0, __UA_t1, "$31"
	662	+#endif /* CONFIG_CPU_MICROMIPS */
	663	+
657	664	if (eva_kernel_access()) {
658	665	__asm__ __volatile__(
659	666	"move\t$4, %1\n\t"
@@ -663,7 +670,7 @@ __clear_user(void __user *addr, __kernel_size_t size) · 3 unmodified lines
663	670	"move\t%0, $6"
664	671	: "=r" (res)
665	672	: "r" (addr), "r" (size)
666		− : "$4", "$5", "$6", __UA_t0, __UA_t1, "$31");
	673	+ : bzero_clobbers);
667	674	} else {
668	675	might_fault();
669	676	__asm__ __volatile__(
@@ -674,7 +681,7 @@ __clear_user(void __user *addr, __kernel_size_t size) · 4 unmodified lines
674	681	"move\t%0, $6"
675	682	: "=r" (res)
676	683	: "r" (addr), "r" (size)
677		− : "$4", "$5", "$6", __UA_t0, __UA_t1, "$31");
	684	+ : bzero_clobbers);
678	685	}
679	686
680	687	return res;

arch/mips/lib/memset.S+8−3

@@ -219,7 +219,7 @@ · 218 unmodified lines
219	219	1: PTR_ADDIU a0, 1 /* fill bytewise */
220	220	R10KCBARRIER(0(ra))
221	221	bne t1, a0, 1b
222		− sb a1, -1(a0)
	222	+ EX(sb, a1, -1(a0), .Lsmall_fixup\@)
223	223
224	224	2: jr ra /* done */
225	225	move a2, zero
@@ -252,13 +252,18 @@ · 26 unmodified lines
252	252	PTR_L t0, TI_TASK($28)
253	253	andi a2, STORMASK
254	254	LONG_L t0, THREAD_BUADDR(t0)
255		− LONG_ADDU a2, t1
	255	+ LONG_ADDU a2, a0
256	256	jr ra
257	257	LONG_SUBU a2, t0
258	258
259	259	.Llast_fixup\@:
260	260	jr ra
261		− andi v1, a2, STORMASK
	261	+ nop
	262	+
	263	+.Lsmall_fixup\@:
	264	+ PTR_SUBU a2, t1, a0
	265	+ jr ra
	266	+ PTR_ADDIU a2, 1
262	267
263	268	.endm
264	269

arch/parisc/Makefile+3−0

@@ -123,6 +123,9 @@ INSTALL_TARGETS = zinstall install · 122 unmodified lines
123	123
124	124	PHONY += bzImage $(BOOT_TARGETS) $(INSTALL_TARGETS)
125	125
	126	+# Default kernel to build
	127	+all: bzImage
	128	+
126	129	zImage: vmlinuz
127	130	Image: vmlinux
128	131

arch/parisc/kernel/Makefile+1−1

@@ -23,7 +23,7 @@ obj-$(CONFIG_SMP) += smp.o · 22 unmodified lines
23	23	obj-$(CONFIG_PA11) += pci-dma.o
24	24	obj-$(CONFIG_PCI) += pci.o
25	25	obj-$(CONFIG_MODULES) += module.o
26		−obj-$(CONFIG_64BIT) += binfmt_elf32.o sys_parisc32.o signal32.o
	26	+obj-$(CONFIG_64BIT) += sys_parisc32.o signal32.o
27	27	obj-$(CONFIG_STACKTRACE)+= stacktrace.o
28	28	obj-$(CONFIG_AUDIT) += audit.o
29	29	obj64-$(CONFIG_AUDIT) += compat_audit.o

arch/parisc/kernel/drivers.c+4−3

@@ -448,7 +448,8 @@ static int match_by_id(struct device * dev, void * data) · 447 unmodified lines
448	448	* Checks all the children of @parent for a matching @id. If none
449	449	* found, it allocates a new device and returns it.
450	450	*/
451		−static struct parisc_device * alloc_tree_node(struct device *parent, char id)
	451	+static struct parisc_device * __init alloc_tree_node(
	452	+ struct device *parent, char id)
452	453	{
453	454	struct match_id_data d = {
454	455	.id = id,
@@ -825,8 +826,8 @@ static void walk_lower_bus(struct parisc_device *dev) · 370 unmodified lines
825	826	* devices which are not physically connected (such as extra serial &
826	827	* keyboard ports). This problem is not yet solved.
827	828	*/
828		−static void walk_native_bus(unsigned long io_io_low, unsigned long io_io_high,
829		− struct device *parent)
	829	+static void __init walk_native_bus(unsigned long io_io_low,
	830	+ unsigned long io_io_high, struct device *parent)
830	831	{
831	832	int i, devices_found = 0;
832	833	unsigned long hpa = io_io_low;

arch/parisc/kernel/pci.c+1−1

@@ -174,7 +174,7 @@ void pcibios_set_master(struct pci_dev *dev) · 173 unmodified lines
174	174	* pcibios_init_bridge() initializes cache line and default latency
175	175	* for pci controllers and pci-pci bridges
176	176	*/
177		−void __init pcibios_init_bridge(struct pci_dev *dev)
	177	+void __ref pcibios_init_bridge(struct pci_dev *dev)
178	178	{
179	179	unsigned short bridge_ctl, bridge_ctl_new;
180	180

arch/parisc/kernel/time.c+1−1

@@ -205,7 +205,7 @@ static int __init rtc_init(void) · 204 unmodified lines
205	205	device_initcall(rtc_init);
206	206	#endif
207	207
208		−void read_persistent_clock(struct timespec *ts)
	208	+void read_persistent_clock64(struct timespec64 *ts)
209	209	{
210	210	static struct pdc_tod tod_data;
211	211	if (pdc_tod_read(&tod_data) == 0) {

arch/parisc/kernel/traps.c+11−0

@@ -837,6 +837,17 @@ void __init initialize_ivt(const void *iva) · 836 unmodified lines
837	837	if (pdc_instr(&instr) == PDC_OK)
838	838	ivap[0] = instr;
839	839
	840	+ /*
	841	+ * Rules for the checksum of the HPMC handler:
	842	+ * 1. The IVA does not point to PDC/PDH space (ie: the OS has installed
	843	+ * its own IVA).
	844	+ * 2. The word at IVA + 32 is nonzero.
	845	+ * 3. If Length (IVA + 60) is not zero, then Length (IVA + 60) and
	846	+ * Address (IVA + 56) are word-aligned.
	847	+ * 4. The checksum of the 8 words starting at IVA + 32 plus the sum of
	848	+ * the Length/4 words starting at Address is zero.
	849	+ */
	850	+
840	851	/* Compute Checksum for HPMC handler */
841	852	length = os_hpmc_size;
842	853	ivap[7] = length;

arch/parisc/mm/init.c+1−1

@@ -516,7 +516,7 @@ static void __init map_pages(unsigned long start_vaddr, · 515 unmodified lines
516	516	}
517	517	}
518	518
519		−void free_initmem(void)
	519	+void __ref free_initmem(void)
520	520	{
521	521	unsigned long init_begin = (unsigned long)__init_begin;
522	522	unsigned long init_end = (unsigned long)__init_end;

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

@@ -15,7 +15,7 @@ · 14 unmodified lines
15	15	extern void powernv_set_nmmu_ptcr(unsigned long ptcr);
16	16	extern struct npu_context pnv_npu2_init_context(struct pci_dev gpdev,
17	17	unsigned long flags,
18		− struct npu_context (cb)(struct npu_context , void ),
	18	+ void (cb)(struct npu_context , void *),
19	19	void *priv);
20	20	extern void pnv_npu2_destroy_context(struct npu_context *context,
21	21	struct pci_dev *gpdev);

arch/powerpc/kernel/eeh_pe.c+2−1

@@ -807,7 +807,8 @@ static void eeh_restore_bridge_bars(struct eeh_dev *edev) · 806 unmodified lines
807	807	eeh_ops->write_config(pdn, 15*4, 4, edev->config_space[15]);
808	808
809	809	/* PCI Command: 0x4 */
810		− eeh_ops->write_config(pdn, PCI_COMMAND, 4, edev->config_space[1]);
	810	+ eeh_ops->write_config(pdn, PCI_COMMAND, 4, edev->config_space[1] \|
	811	+ PCI_COMMAND_MEMORY \| PCI_COMMAND_MASTER);
811	812
812	813	/* Check the PCIe link is ready */
813	814	eeh_bridge_check_link(edev);

arch/powerpc/kernel/idle_book3s.S+2−2

@@ -553,12 +553,12 @@ ALT_FTR_SECTION_END_IFSET(CPU_FTR_ARCH_300) · 552 unmodified lines
553	553	#ifdef CONFIG_KVM_BOOK3S_HV_POSSIBLE
554	554	lbz r0,HSTATE_HWTHREAD_STATE(r13)
555	555	cmpwi r0,KVM_HWTHREAD_IN_KERNEL
556		− beq 1f
	556	+ beq 0f
557	557	li r0,KVM_HWTHREAD_IN_KERNEL
558	558	stb r0,HSTATE_HWTHREAD_STATE(r13)
559	559	/* Order setting hwthread_state vs. testing hwthread_req */
560	560	sync
561		− lbz r0,HSTATE_HWTHREAD_REQ(r13)
	561	+0: lbz r0,HSTATE_HWTHREAD_REQ(r13)
562	562	cmpwi r0,0
563	563	beq 1f
564	564	b kvm_start_guest

arch/powerpc/kernel/mce_power.c+2−5

@@ -441,7 +441,6 @@ static int mce_handle_ierror(struct pt_regs *regs, · 440 unmodified lines
441	441	if (pfn != ULONG_MAX) {
442	442	*phys_addr =
443	443	(pfn << PAGE_SHIFT);
444		− handled = 1;
445	444	}
446	445	}
447	446	}
@@ -532,9 +531,7 @@ static int mce_handle_derror(struct pt_regs *regs, · 84 unmodified lines
532	531	* kernel/exception-64s.h
533	532	*/
534	533	if (get_paca()->in_mce < MAX_MCE_DEPTH)
535		− if (!mce_find_instr_ea_and_pfn(regs, addr,
536		− phys_addr))
537		− handled = 1;
	534	+ mce_find_instr_ea_and_pfn(regs, addr, phys_addr);
538	535	}
539	536	found = 1;
540	537	}
@@ -572,7 +569,7 @@ static long mce_handle_error(struct pt_regs *regs, · 31 unmodified lines
572	569	const struct mce_ierror_table itable[])
573	570	{
574	571	struct mce_error_info mce_err = { 0 };
575		− uint64_t addr, phys_addr;
	572	+ uint64_t addr, phys_addr = ULONG_MAX;
576	573	uint64_t srr1 = regs->msr;
577	574	long handled;
578	575

arch/powerpc/kernel/setup_64.c+11−0

@@ -890,6 +890,17 @@ static void __ref init_fallback_flush(void) · 889 unmodified lines
890	890	return;
891	891
892	892	l1d_size = ppc64_caches.l1d.size;
	893	+
	894	+ /*
	895	+ * If there is no d-cache-size property in the device tree, l1d_size
	896	+ * could be zero. That leads to the loop in the asm wrapping around to
	897	+ * 2^64-1, and then walking off the end of the fallback area and
	898	+ * eventually causing a page fault which is fatal. Just default to
	899	+ * something vaguely sane.
	900	+ */
	901	+ if (!l1d_size)
	902	+ l1d_size = (64 * 1024);
	903	+
893	904	limit = min(ppc64_bolted_size(), ppc64_rma_size);
894	905
895	906	/*

arch/powerpc/kernel/smp.c+42−7

@@ -566,10 +566,35 @@ void crash_send_ipi(void (crash_ipi_callback)(struct pt_regs )) · 565 unmodified lines
566	566	#endif
567	567
568	568	#ifdef CONFIG_NMI_IPI
569		−static void stop_this_cpu(struct pt_regs *regs)
570		−#else
	569	+static void nmi_stop_this_cpu(struct pt_regs *regs)
	570	+{
	571	+ /*
	572	+ * This is a special case because it never returns, so the NMI IPI
	573	+ * handling would never mark it as done, which makes any later
	574	+ * smp_send_nmi_ipi() call spin forever. Mark it done now.
	575	+ *
	576	+ * IRQs are already hard disabled by the smp_handle_nmi_ipi.
	577	+ */
	578	+ nmi_ipi_lock();
	579	+ nmi_ipi_busy_count--;
	580	+ nmi_ipi_unlock();
	581	+
	582	+ /* Remove this CPU */
	583	+ set_cpu_online(smp_processor_id(), false);
	584	+
	585	+ spin_begin();
	586	+ while (1)
	587	+ spin_cpu_relax();
	588	+}
	589	+
	590	+void smp_send_stop(void)
	591	+{
	592	+ smp_send_nmi_ipi(NMI_IPI_ALL_OTHERS, nmi_stop_this_cpu, 1000000);
	593	+}
	594	+
	595	+#else /* CONFIG_NMI_IPI */
	596	+
571	597	static void stop_this_cpu(void *dummy)
572		−#endif
573	598	{
574	599	/* Remove this CPU */
575	600	set_cpu_online(smp_processor_id(), false);
@@ -582,12 +607,22 @@ static void stop_this_cpu(void *dummy) · 6 unmodified lines
582	607
583	608	void smp_send_stop(void)
584	609	{
585		−#ifdef CONFIG_NMI_IPI
586		− smp_send_nmi_ipi(NMI_IPI_ALL_OTHERS, stop_this_cpu, 1000000);
587		−#else
	610	+ static bool stopped = false;
	611	+
	612	+ /*
	613	+ * Prevent waiting on csd lock from a previous smp_send_stop.
	614	+ * This is racy, but in general callers try to do the right
	615	+ * thing and only fire off one smp_send_stop (e.g., see
	616	+ * kernel/panic.c)
	617	+ */
	618	+ if (stopped)
	619	+ return;
	620	+
	621	+ stopped = true;
	622	+
588	623	smp_call_function(stop_this_cpu, NULL, 0);
589		−#endif
590	624	}
	625	+#endif /* CONFIG_NMI_IPI */
591	626
592	627	struct thread_info *current_set[NR_CPUS];
593	628

arch/powerpc/kvm/booke.c+7−0

@@ -305,6 +305,13 @@ void kvmppc_core_queue_fpunavail(struct kvm_vcpu *vcpu) · 304 unmodified lines
305	305	kvmppc_booke_queue_irqprio(vcpu, BOOKE_IRQPRIO_FP_UNAVAIL);
306	306	}
307	307
	308	+#ifdef CONFIG_ALTIVEC
	309	+void kvmppc_core_queue_vec_unavail(struct kvm_vcpu *vcpu)
	310	+{
	311	+ kvmppc_booke_queue_irqprio(vcpu, BOOKE_IRQPRIO_ALTIVEC_UNAVAIL);
	312	+}
	313	+#endif
	314	+
308	315	void kvmppc_core_queue_dec(struct kvm_vcpu *vcpu)
309	316	{
310	317	kvmppc_booke_queue_irqprio(vcpu, BOOKE_IRQPRIO_DECREMENTER);

arch/powerpc/lib/feature-fixups.c+1−1

@@ -55,7 +55,7 @@ static int patch_alt_instruction(unsigned int src, unsigned int dest, · 54 unmodified lines
55	55	unsigned int target = (unsigned int )branch_target(src);
56	56
57	57	/* Branch within the section doesn't need translating */
58		− if (target < alt_start \|\| target >= alt_end) {
	58	+ if (target < alt_start \|\| target > alt_end) {
59	59	instr = translate_branch(dest, src);
60	60	if (!instr)
61	61	return 1;

arch/powerpc/mm/mem.c+2−0

@@ -133,6 +133,7 @@ int __meminit arch_add_memory(int nid, u64 start, u64 size, struct vmem_altmap * · 132 unmodified lines
133	133	start, start + size, rc);
134	134	return -EFAULT;
135	135	}
	136	+ flush_inval_dcache_range(start, start + size);
136	137
137	138	return __add_pages(nid, start_pfn, nr_pages, altmap, want_memblock);
138	139	}
@@ -159,6 +160,7 @@ int __meminit arch_remove_memory(u64 start, u64 size, struct vmem_altmap *altmap · 20 unmodified lines
159	160
160	161	/* Remove htab bolted mappings for this section of memory */
161	162	start = (unsigned long)__va(start);
	163	+ flush_inval_dcache_range(start, start + size);
162	164	ret = remove_section_mapping(start, start + size);
163	165
164	166	/* Ensure all vmalloc mappings are flushed in case they also

arch/powerpc/platforms/cell/spufs/sched.c+1−1

@@ -1093,7 +1093,7 @@ static int show_spu_loadavg(struct seq_file s, void private) · 1092 unmodified lines
1093	1093	LOAD_INT(c), LOAD_FRAC(c),
1094	1094	count_active_contexts(),
1095	1095	atomic_read(&nr_spu_contexts),
1096		− idr_get_cursor(&task_active_pid_ns(current)->idr));
	1096	+ idr_get_cursor(&task_active_pid_ns(current)->idr) - 1);
1097	1097	return 0;
1098	1098	}
1099	1099

arch/powerpc/platforms/powernv/memtrace.c+0−17

@@ -82,19 +82,6 @@ static const struct file_operations memtrace_fops = { · 81 unmodified lines
82	82	.open = simple_open,
83	83	};
84	84
85		−static void flush_memory_region(u64 base, u64 size)
86		−{
87		− unsigned long line_size = ppc64_caches.l1d.size;
88		− u64 end = base + size;
89		− u64 addr;
90		−
91		− base = round_down(base, line_size);
92		− end = round_up(end, line_size);
93		−
94		− for (addr = base; addr < end; addr += line_size)
95		− asm volatile("dcbf 0,%0" : "=r" (addr) :: "memory");
96		−}
97		−
98	85	static int check_memblock_online(struct memory_block mem, void arg)
99	86	{
100	87	if (mem->state != MEM_ONLINE)
@@ -132,10 +119,6 @@ static bool memtrace_offline_pages(u32 nid, u64 start_pfn, u64 nr_pages) · 31 unmodified lines
132	119	walk_memory_range(start_pfn, end_pfn, (void *)MEM_OFFLINE,
133	120	change_memblock_state);
134	121
135		− /* RCU grace period? */
136		− flush_memory_region((u64)__va(start_pfn << PAGE_SHIFT),
137		− nr_pages << PAGE_SHIFT);
138		−
139	122	lock_device_hotplug();
140	123	remove_memory(nid, start_pfn << PAGE_SHIFT, nr_pages << PAGE_SHIFT);
141	124	unlock_device_hotplug();

arch/powerpc/platforms/powernv/npu-dma.c+73−15

@@ -33,6 +33,19 @@ · 32 unmodified lines
33	33
34	34	#define npu_to_phb(x) container_of(x, struct pnv_phb, npu)
35	35
	36	+/*
	37	+ * spinlock to protect initialisation of an npu_context for a particular
	38	+ * mm_struct.
	39	+ */
	40	+static DEFINE_SPINLOCK(npu_context_lock);
	41	+
	42	+/*
	43	+ * When an address shootdown range exceeds this threshold we invalidate the
	44	+ * entire TLB on the GPU for the given PID rather than each specific address in
	45	+ * the range.
	46	+ */
	47	+#define ATSD_THRESHOLD (210241024)
	48	+
36	49	/*
37	50	* Other types of TCE cache invalidation are not functional in the
38	51	* hardware.
@@ -401,7 +414,7 @@ struct npu_context { · 362 unmodified lines
401	414	bool nmmu_flush;
402	415
403	416	/* Callback to stop translation requests on a given GPU */
404		− struct npu_context (release_cb)(struct npu_context , void );
	417	+ void (release_cb)(struct npu_context context, void *priv);
405	418
406	419	/*
407	420	* Private pointer passed to the above callback for usage by
@@ -671,11 +684,19 @@ static void pnv_npu2_mn_invalidate_range(struct mmu_notifier *mn, · 263 unmodified lines
671	684	struct npu_context *npu_context = mn_to_npu_context(mn);
672	685	unsigned long address;
673	686
674		− for (address = start; address < end; address += PAGE_SIZE)
675		− mmio_invalidate(npu_context, 1, address, false);
	687	+ if (end - start > ATSD_THRESHOLD) {
	688	+ /*
	689	+ * Just invalidate the entire PID if the address range is too
	690	+ * large.
	691	+ */
	692	+ mmio_invalidate(npu_context, 0, 0, true);
	693	+ } else {
	694	+ for (address = start; address < end; address += PAGE_SIZE)
	695	+ mmio_invalidate(npu_context, 1, address, false);
676	696
677		− /* Do the flush only on the final addess == end */
678		− mmio_invalidate(npu_context, 1, address, true);
	697	+ /* Do the flush only on the final addess == end */
	698	+ mmio_invalidate(npu_context, 1, address, true);
	699	+ }
679	700	}
680	701
681	702	static const struct mmu_notifier_ops nv_nmmu_notifier_ops = {
@@ -696,11 +717,12 @@ static const struct mmu_notifier_ops nv_nmmu_notifier_ops = { · 14 unmodified lines
696	717	* Returns an error if there no contexts are currently available or a
697	718	* npu_context which should be passed to pnv_npu2_handle_fault().
698	719	*
699		− * mmap_sem must be held in write mode.
	720	+ * mmap_sem must be held in write mode and must not be called from interrupt
	721	+ * context.
700	722	*/
701	723	struct npu_context pnv_npu2_init_context(struct pci_dev gpdev,
702	724	unsigned long flags,
703		− struct npu_context (cb)(struct npu_context , void ),
	725	+ void (cb)(struct npu_context , void *),
704	726	void *priv)
705	727	{
706	728	int rc;
@@ -743,7 +765,9 @@ struct npu_context pnv_npu2_init_context(struct pci_dev gpdev, · 36 unmodified lines
743	765	/*
744	766	* Setup the NPU context table for a particular GPU. These need to be
745	767	* per-GPU as we need the tables to filter ATSDs when there are no
746		− * active contexts on a particular GPU.
	768	+ * active contexts on a particular GPU. It is safe for these to be
	769	+ * called concurrently with destroy as the OPAL call takes appropriate
	770	+ * locks and refcounts on init/destroy.
747	771	*/
748	772	rc = opal_npu_init_context(nphb->opal_id, mm->context.id, flags,
749	773	PCI_DEVID(gpdev->bus->number, gpdev->devfn));
@@ -754,8 +778,29 @@ struct npu_context pnv_npu2_init_context(struct pci_dev gpdev, · 4 unmodified lines
754	778	* We store the npu pci device so we can more easily get at the
755	779	* associated npus.
756	780	*/
	781	+ spin_lock(&npu_context_lock);
757	782	npu_context = mm->context.npu_context;
	783	+ if (npu_context) {
	784	+ if (npu_context->release_cb != cb \|\|
	785	+ npu_context->priv != priv) {
	786	+ spin_unlock(&npu_context_lock);
	787	+ opal_npu_destroy_context(nphb->opal_id, mm->context.id,
	788	+ PCI_DEVID(gpdev->bus->number,
	789	+ gpdev->devfn));
	790	+ return ERR_PTR(-EINVAL);
	791	+ }
	792	+
	793	+ WARN_ON(!kref_get_unless_zero(&npu_context->kref));
	794	+ }
	795	+ spin_unlock(&npu_context_lock);
	796	+
758	797	if (!npu_context) {
	798	+ /*
	799	+ * We can set up these fields without holding the
	800	+ * npu_context_lock as the npu_context hasn't been returned to
	801	+ * the caller meaning it can't be destroyed. Parallel allocation
	802	+ * is protected against by mmap_sem.
	803	+ */
759	804	rc = -ENOMEM;
760	805	npu_context = kzalloc(sizeof(struct npu_context), GFP_KERNEL);
761	806	if (npu_context) {
@@ -774,8 +819,6 @@ struct npu_context pnv_npu2_init_context(struct pci_dev gpdev, · 12 unmodified lines
774	819	}
775	820
776	821	mm->context.npu_context = npu_context;
777		− } else {
778		− WARN_ON(!kref_get_unless_zero(&npu_context->kref));
779	822	}
780	823
781	824	npu_context->release_cb = cb;
@@ -814,15 +857,16 @@ static void pnv_npu2_release_context(struct kref *kref) · 32 unmodified lines
814	857	mm_context_remove_copro(npu_context->mm);
815	858
816	859	npu_context->mm->context.npu_context = NULL;
817		− mmu_notifier_unregister(&npu_context->mn,
818		− npu_context->mm);
819		−
820		− kfree(npu_context);
821	860	}
822	861
	862	+/*
	863	+ * Destroy a context on the given GPU. May free the npu_context if it is no
	864	+ * longer active on any GPUs. Must not be called from interrupt context.
	865	+ */
823	866	void pnv_npu2_destroy_context(struct npu_context *npu_context,
824	867	struct pci_dev *gpdev)
825	868	{
	869	+ int removed;
826	870	struct pnv_phb *nphb;
827	871	struct npu *npu;
828	872	struct pci_dev *npdev = pnv_pci_get_npu_dev(gpdev, 0);
@@ -844,7 +888,21 @@ void pnv_npu2_destroy_context(struct npu_context *npu_context, · 15 unmodified lines
844	888	WRITE_ONCE(npu_context->npdev[npu->index][nvlink_index], NULL);
845	889	opal_npu_destroy_context(nphb->opal_id, npu_context->mm->context.id,
846	890	PCI_DEVID(gpdev->bus->number, gpdev->devfn));
847		− kref_put(&npu_context->kref, pnv_npu2_release_context);
	891	+ spin_lock(&npu_context_lock);
	892	+ removed = kref_put(&npu_context->kref, pnv_npu2_release_context);
	893	+ spin_unlock(&npu_context_lock);
	894	+
	895	+ /*
	896	+ * We need to do this outside of pnv_npu2_release_context so that it is
	897	+ * outside the spinlock as mmu_notifier_destroy uses SRCU.
	898	+ */
	899	+ if (removed) {
	900	+ mmu_notifier_unregister(&npu_context->mn,
	901	+ npu_context->mm);
	902	+
	903	+ kfree(npu_context);
	904	+ }
	905	+
848	906	}
849	907	EXPORT_SYMBOL(pnv_npu2_destroy_context);
850	908

arch/powerpc/platforms/powernv/opal-rtc.c+5−3

@@ -48,10 +48,12 @@ unsigned long __init opal_get_boot_time(void) · 47 unmodified lines
48	48
49	49	while (rc == OPAL_BUSY \|\| rc == OPAL_BUSY_EVENT) {
50	50	rc = opal_rtc_read(&__y_m_d, &__h_m_s_ms);
51		− if (rc == OPAL_BUSY_EVENT)
	51	+ if (rc == OPAL_BUSY_EVENT) {
	52	+ mdelay(OPAL_BUSY_DELAY_MS);
52	53	opal_poll_events(NULL);
53		− else if (rc == OPAL_BUSY)
54		− mdelay(10);
	54	+ } else if (rc == OPAL_BUSY) {
	55	+ mdelay(OPAL_BUSY_DELAY_MS);
	56	+ }
55	57	}
56	58	if (rc != OPAL_SUCCESS)
57	59	return 0;

arch/powerpc/sysdev/xive/native.c+4−0

@@ -389,6 +389,10 @@ static void xive_native_setup_cpu(unsigned int cpu, struct xive_cpu *xc) · 388 unmodified lines
389	389	if (xive_pool_vps == XIVE_INVALID_VP)
390	390	return;
391	391
	392	+ /* Check if pool VP already active, if it is, pull it */
	393	+ if (in_be32(xive_tima + TM_QW2_HV_POOL + TM_WORD2) & TM_QW2W2_VP)
	394	+ in_be64(xive_tima + TM_SPC_PULL_POOL_CTX);
	395	+
392	396	/* Enable the pool VP */
393	397	vp = xive_pool_vps + cpu;
394	398	pr_debug("CPU %d setting up pool VP 0x%x\n", cpu, vp);

arch/riscv/Kconfig+1−3

@@ -11,6 +11,7 @@ config RISCV · 10 unmodified lines
11	11	select ARCH_WANT_FRAME_POINTERS
12	12	select CLONE_BACKWARDS
13	13	select COMMON_CLK
	14	+ select DMA_DIRECT_OPS
14	15	select GENERIC_CLOCKEVENTS
15	16	select GENERIC_CPU_DEVICES
16	17	select GENERIC_IRQ_SHOW
@@ -89,9 +90,6 @@ config PGTABLE_LEVELS · 72 unmodified lines
89	90	config HAVE_KPROBES
90	91	def_bool n
91	92
92		−config DMA_DIRECT_OPS
93		− def_bool y
94		−
95	93	menu "Platform type"
96	94
97	95	choice

arch/riscv/include/asm/Kbuild+0−1

@@ -15,7 +15,6 @@ generic-y += fcntl.h · 14 unmodified lines
15	15	generic-y += futex.h
16	16	generic-y += hardirq.h
17	17	generic-y += hash.h
18		−generic-y += handle_irq.h
19	18	generic-y += hw_irq.h
20	19	generic-y += ioctl.h
21	20	generic-y += ioctls.h

arch/riscv/kernel/vdso/Makefile+1−1

@@ -52,7 +52,7 @@ $(obj)/%.so: $(obj)/%.so.dbg FORCE · 51 unmodified lines
52	52	# Add -lgcc so rv32 gets static muldi3 and lshrdi3 definitions.
53	53	# Make sure only to export the intended __vdso_xxx symbol offsets.
54	54	quiet_cmd_vdsold = VDSOLD $@
55		− cmd_vdsold = $(CC) $(KCFLAGS) -nostdlib $(SYSCFLAGS_$(@F)) \
	55	+ cmd_vdsold = $(CC) $(KCFLAGS) $(call cc-option, -no-pie) -nostdlib $(SYSCFLAGS_$(@F)) \
56	56	-Wl,-T,$(filter-out FORCE,$^) -o $@.tmp -lgcc && \
57	57	$(CROSS_COMPILE)objcopy \
58	58	$(patsubst %, -G __vdso_%, $(vdso-syms)) $@.tmp $@

arch/s390/Kbuild+1−0

@@ -8,3 +8,4 @@ obj-$(CONFIG_APPLDATA_BASE) += appldata/ · 7 unmodified lines
8	8	obj-y += net/
9	9	obj-$(CONFIG_PCI) += pci/
10	10	obj-$(CONFIG_NUMA) += numa/
	11	+obj-$(CONFIG_ARCH_HAS_KEXEC_PURGATORY) += purgatory/

arch/s390/Kconfig+24−8

@@ -47,10 +47,6 @@ config PGSTE · 46 unmodified lines
47	47	config ARCH_SUPPORTS_DEBUG_PAGEALLOC
48	48	def_bool y
49	49
50		−config KEXEC
51		− def_bool y
52		− select KEXEC_CORE
53		−
54	50	config AUDIT_ARCH
55	51	def_bool y
56	52
@@ -290,12 +286,12 @@ config MARCH_Z13 · 233 unmodified lines
290	286	older machines.
291	287
292	288	config MARCH_Z14
293		− bool "IBM z14"
	289	+ bool "IBM z14 ZR1 and z14"
294	290	select HAVE_MARCH_Z14_FEATURES
295	291	help
296		− Select this to enable optimizations for IBM z14 (3906 series).
297		− The kernel will be slightly faster but will not work on older
298		− machines.
	292	+ Select this to enable optimizations for IBM z14 ZR1 and z14 (3907
	293	+ and 3906 series). The kernel will be slightly faster but will not
	294	+ work on older machines.
299	295
300	296	endchoice
301	297
@@ -525,6 +521,26 @@ source kernel/Kconfig.preempt · 223 unmodified lines
525	521
526	522	source kernel/Kconfig.hz
527	523
	524	+config KEXEC
	525	+ def_bool y
	526	+ select KEXEC_CORE
	527	+
	528	+config KEXEC_FILE
	529	+ bool "kexec file based system call"
	530	+ select KEXEC_CORE
	531	+ select BUILD_BIN2C
	532	+ depends on CRYPTO
	533	+ depends on CRYPTO_SHA256
	534	+ depends on CRYPTO_SHA256_S390
	535	+ help
	536	+ Enable the kexec file based system call. In contrast to the normal
	537	+ kexec system call this system call takes file descriptors for the
	538	+ kernel and initramfs as arguments.
	539	+
	540	+config ARCH_HAS_KEXEC_PURGATORY
	541	+ def_bool y
	542	+ depends on KEXEC_FILE
	543	+
528	544	config ARCH_RANDOM
529	545	def_bool y
530	546	prompt "s390 architectural random number generation API"

arch/s390/boot/Makefile+0−6

@@ -3,12 +3,6 @@ · 2 unmodified lines
3	3	# Makefile for the linux s390-specific parts of the memory manager.
4	4	#
5	5
6		−COMPILE_VERSION := __linux_compile_version_id__`hostname \| \
7		− tr -c '[0-9A-Za-z]' '_'`__`date \| \
8		− tr -c '[0-9A-Za-z]' '_'`_t
9		−
10		−ccflags-y := -DCOMPILE_VERSION=$(COMPILE_VERSION) -gstabs -I.
11		−
12	6	targets := image
13	7	targets += bzImage
14	8	subdir- := compressed

arch/s390/boot/compressed/.gitignore+1−0

@@ -1,3 +1,4 @@
1	1	sizes.h
2	2	vmlinux
3	3	vmlinux.lds
	4	+vmlinux.bin.full

arch/s390/configs/debug_defconfigrenamed from arch/s390/configs/default_defconfig+20−10

@@ -24,13 +24,13 @@ CONFIG_CPUSETS=y · 23 unmodified lines
24	24	CONFIG_CGROUP_DEVICE=y
25	25	CONFIG_CGROUP_CPUACCT=y
26	26	CONFIG_CGROUP_PERF=y
27		−CONFIG_CHECKPOINT_RESTORE=y
28	27	CONFIG_NAMESPACES=y
29	28	CONFIG_USER_NS=y
30	29	CONFIG_SCHED_AUTOGROUP=y
31	30	CONFIG_BLK_DEV_INITRD=y
32	31	CONFIG_EXPERT=y
33	32	# CONFIG_SYSFS_SYSCALL is not set
	33	+CONFIG_CHECKPOINT_RESTORE=y
34	34	CONFIG_BPF_SYSCALL=y
35	35	CONFIG_USERFAULTFD=y
36	36	# CONFIG_COMPAT_BRK is not set
@@ -59,10 +59,11 @@ CONFIG_CFQ_GROUP_IOSCHED=y · 22 unmodified lines
59	59	CONFIG_DEFAULT_DEADLINE=y
60	60	CONFIG_LIVEPATCH=y
61	61	CONFIG_TUNE_ZEC12=y
62		−CONFIG_NR_CPUS=256
	62	+CONFIG_NR_CPUS=512
63	63	CONFIG_NUMA=y
64	64	CONFIG_PREEMPT=y
65	65	CONFIG_HZ_100=y
	66	+CONFIG_KEXEC_FILE=y
66	67	CONFIG_MEMORY_HOTPLUG=y
67	68	CONFIG_MEMORY_HOTREMOVE=y
68	69	CONFIG_KSM=y
@@ -305,7 +306,6 @@ CONFIG_IP6_NF_SECURITY=m · 236 unmodified lines
305	306	CONFIG_IP6_NF_NAT=m
306	307	CONFIG_IP6_NF_TARGET_MASQUERADE=m
307	308	CONFIG_NF_TABLES_BRIDGE=m
308		−CONFIG_NET_SCTPPROBE=m
309	309	CONFIG_RDS=m
310	310	CONFIG_RDS_RDMA=m
311	311	CONFIG_RDS_TCP=m
@@ -364,11 +364,11 @@ CONFIG_NET_ACT_SIMP=m · 52 unmodified lines
364	364	CONFIG_NET_ACT_SKBEDIT=m
365	365	CONFIG_NET_ACT_CSUM=m
366	366	CONFIG_DNS_RESOLVER=y
	367	+CONFIG_OPENVSWITCH=m
367	368	CONFIG_NETLINK_DIAG=m
368	369	CONFIG_CGROUP_NET_PRIO=y
369	370	CONFIG_BPF_JIT=y
370	371	CONFIG_NET_PKTGEN=m
371		−CONFIG_NET_TCPPROBE=m
372	372	CONFIG_DEVTMPFS=y
373	373	CONFIG_DMA_CMA=y
374	374	CONFIG_CMA_SIZE_MBYTES=0
@@ -380,9 +380,9 @@ CONFIG_BLK_DEV_DRBD=m · 5 unmodified lines
380	380	CONFIG_BLK_DEV_NBD=m
381	381	CONFIG_BLK_DEV_RAM=y
382	382	CONFIG_BLK_DEV_RAM_SIZE=32768
383		−CONFIG_BLK_DEV_RAM_DAX=y
384	383	CONFIG_VIRTIO_BLK=y
385	384	CONFIG_BLK_DEV_RBD=m
	385	+CONFIG_BLK_DEV_NVME=m
386	386	CONFIG_ENCLOSURE_SERVICES=m
387	387	CONFIG_GENWQE=m
388	388	CONFIG_RAID_ATTRS=m
@@ -461,6 +461,7 @@ CONFIG_PPTP=m · 72 unmodified lines
461	461	CONFIG_PPPOL2TP=m
462	462	CONFIG_PPP_ASYNC=m
463	463	CONFIG_PPP_SYNC_TTY=m
	464	+CONFIG_INPUT_EVDEV=y
464	465	# CONFIG_INPUT_KEYBOARD is not set
465	466	# CONFIG_INPUT_MOUSE is not set
466	467	# CONFIG_SERIO is not set
@@ -474,6 +475,9 @@ CONFIG_WATCHDOG=y · 7 unmodified lines
474	475	CONFIG_WATCHDOG_NOWAYOUT=y
475	476	CONFIG_SOFT_WATCHDOG=m
476	477	CONFIG_DIAG288_WATCHDOG=m
	478	+CONFIG_DRM=y
	479	+CONFIG_DRM_VIRTIO_GPU=y
	480	+CONFIG_FRAMEBUFFER_CONSOLE=y
477	481	# CONFIG_HID is not set
478	482	# CONFIG_USB_SUPPORT is not set
479	483	CONFIG_INFINIBAND=m
@@ -482,7 +486,9 @@ CONFIG_MLX4_INFINIBAND=m · 2 unmodified lines
482	486	CONFIG_MLX5_INFINIBAND=m
483	487	CONFIG_VFIO=m
484	488	CONFIG_VFIO_PCI=m
	489	+CONFIG_VIRTIO_PCI=m
485	490	CONFIG_VIRTIO_BALLOON=m
	491	+CONFIG_VIRTIO_INPUT=y
486	492	CONFIG_EXT4_FS=y
487	493	CONFIG_EXT4_FS_POSIX_ACL=y
488	494	CONFIG_EXT4_FS_SECURITY=y
@@ -641,6 +647,8 @@ CONFIG_ATOMIC64_SELFTEST=y · 152 unmodified lines
641	647	CONFIG_TEST_BPF=m
642	648	CONFIG_BUG_ON_DATA_CORRUPTION=y
643	649	CONFIG_S390_PTDUMP=y
	650	+CONFIG_PERSISTENT_KEYRINGS=y
	651	+CONFIG_BIG_KEYS=y
644	652	CONFIG_ENCRYPTED_KEYS=m
645	653	CONFIG_SECURITY=y
646	654	CONFIG_SECURITY_NETWORK=y
@@ -649,17 +657,20 @@ CONFIG_SECURITY_SELINUX=y · 2 unmodified lines
649	657	CONFIG_SECURITY_SELINUX_BOOTPARAM=y
650	658	CONFIG_SECURITY_SELINUX_BOOTPARAM_VALUE=0
651	659	CONFIG_SECURITY_SELINUX_DISABLE=y
	660	+CONFIG_INTEGRITY_SIGNATURE=y
	661	+CONFIG_INTEGRITY_ASYMMETRIC_KEYS=y
652	662	CONFIG_IMA=y
	663	+CONFIG_IMA_DEFAULT_HASH_SHA256=y
	664	+CONFIG_IMA_WRITE_POLICY=y
653	665	CONFIG_IMA_APPRAISE=y
654		−CONFIG_CRYPTO_RSA=m
655	666	CONFIG_CRYPTO_DH=m
656	667	CONFIG_CRYPTO_ECDH=m
657	668	CONFIG_CRYPTO_USER=m
	669	+# CONFIG_CRYPTO_MANAGER_DISABLE_TESTS is not set
658	670	CONFIG_CRYPTO_PCRYPT=m
659	671	CONFIG_CRYPTO_CRYPTD=m
660	672	CONFIG_CRYPTO_MCRYPTD=m
661	673	CONFIG_CRYPTO_TEST=m
662		−CONFIG_CRYPTO_GCM=m
663	674	CONFIG_CRYPTO_CHACHA20POLY1305=m
664	675	CONFIG_CRYPTO_LRW=m
665	676	CONFIG_CRYPTO_PCBC=m
@@ -707,9 +718,8 @@ CONFIG_CRYPTO_DES_S390=m · 41 unmodified lines
707	718	CONFIG_CRYPTO_AES_S390=m
708	719	CONFIG_CRYPTO_GHASH_S390=m
709	720	CONFIG_CRYPTO_CRC32_S390=y
710		−CONFIG_ASYMMETRIC_KEY_TYPE=y
711		−CONFIG_ASYMMETRIC_PUBLIC_KEY_SUBTYPE=m
712		−CONFIG_X509_CERTIFICATE_PARSER=m
	721	+CONFIG_PKCS7_MESSAGE_PARSER=y
	722	+CONFIG_SYSTEM_TRUSTED_KEYRING=y
713	723	CONFIG_CRC7=m
714	724	CONFIG_CRC8=m
715	725	CONFIG_RANDOM32_SELFTEST=y

arch/s390/configs/gcov_defconfig+0−661

@@ -1,661 +0,0 @@
1		−CONFIG_SYSVIPC=y
2		−CONFIG_POSIX_MQUEUE=y
3		−CONFIG_AUDIT=y
4		−CONFIG_NO_HZ_IDLE=y
5		−CONFIG_HIGH_RES_TIMERS=y
6		−CONFIG_BSD_PROCESS_ACCT=y
7		−CONFIG_BSD_PROCESS_ACCT_V3=y
8		−CONFIG_TASKSTATS=y
9		−CONFIG_TASK_DELAY_ACCT=y
10		−CONFIG_TASK_XACCT=y
11		−CONFIG_TASK_IO_ACCOUNTING=y
12		−CONFIG_IKCONFIG=y
13		−CONFIG_IKCONFIG_PROC=y
14		−CONFIG_NUMA_BALANCING=y
15		−# CONFIG_NUMA_BALANCING_DEFAULT_ENABLED is not set
16		−CONFIG_MEMCG=y
17		−CONFIG_MEMCG_SWAP=y
18		−CONFIG_BLK_CGROUP=y
19		−CONFIG_CFS_BANDWIDTH=y
20		−CONFIG_RT_GROUP_SCHED=y
21		−CONFIG_CGROUP_PIDS=y
22		−CONFIG_CGROUP_FREEZER=y
23		−CONFIG_CGROUP_HUGETLB=y
24		−CONFIG_CPUSETS=y
25		−CONFIG_CGROUP_DEVICE=y
26		−CONFIG_CGROUP_CPUACCT=y
27		−CONFIG_CGROUP_PERF=y
28		−CONFIG_CHECKPOINT_RESTORE=y
29		−CONFIG_NAMESPACES=y
30		−CONFIG_USER_NS=y
31		−CONFIG_SCHED_AUTOGROUP=y
32		−CONFIG_BLK_DEV_INITRD=y
33		−CONFIG_EXPERT=y
34		−# CONFIG_SYSFS_SYSCALL is not set
35		−CONFIG_BPF_SYSCALL=y
36		−CONFIG_USERFAULTFD=y
37		−# CONFIG_COMPAT_BRK is not set
38		−CONFIG_PROFILING=y
39		−CONFIG_OPROFILE=m
40		−CONFIG_KPROBES=y
41		−CONFIG_JUMP_LABEL=y
42		−CONFIG_GCOV_KERNEL=y
43		−CONFIG_GCOV_PROFILE_ALL=y
44		−CONFIG_MODULES=y
45		−CONFIG_MODULE_FORCE_LOAD=y
46		−CONFIG_MODULE_UNLOAD=y
47		−CONFIG_MODULE_FORCE_UNLOAD=y
48		−CONFIG_MODVERSIONS=y
49		−CONFIG_MODULE_SRCVERSION_ALL=y
50		−CONFIG_BLK_DEV_INTEGRITY=y
51		−CONFIG_BLK_DEV_THROTTLING=y
52		−CONFIG_BLK_WBT=y
53		−CONFIG_BLK_WBT_SQ=y
54		−CONFIG_PARTITION_ADVANCED=y
55		−CONFIG_IBM_PARTITION=y
56		−CONFIG_BSD_DISKLABEL=y
57		−CONFIG_MINIX_SUBPARTITION=y
58		−CONFIG_SOLARIS_X86_PARTITION=y
59		−CONFIG_UNIXWARE_DISKLABEL=y
60		−CONFIG_CFQ_GROUP_IOSCHED=y
61		−CONFIG_DEFAULT_DEADLINE=y
62		−CONFIG_LIVEPATCH=y
63		−CONFIG_TUNE_ZEC12=y
64		−CONFIG_NR_CPUS=512
65		−CONFIG_NUMA=y
66		−CONFIG_HZ_100=y
67		−CONFIG_MEMORY_HOTPLUG=y
68		−CONFIG_MEMORY_HOTREMOVE=y
69		−CONFIG_KSM=y
70		−CONFIG_TRANSPARENT_HUGEPAGE=y
71		−CONFIG_CLEANCACHE=y
72		−CONFIG_FRONTSWAP=y
73		−CONFIG_MEM_SOFT_DIRTY=y
74		−CONFIG_ZSWAP=y
75		−CONFIG_ZBUD=m
76		−CONFIG_ZSMALLOC=m
77		−CONFIG_ZSMALLOC_STAT=y
78		−CONFIG_DEFERRED_STRUCT_PAGE_INIT=y
79		−CONFIG_IDLE_PAGE_TRACKING=y
80		−CONFIG_PCI=y
81		−CONFIG_HOTPLUG_PCI=y
82		−CONFIG_HOTPLUG_PCI_S390=y
83		−CONFIG_CHSC_SCH=y
84		−CONFIG_CRASH_DUMP=y
85		−CONFIG_BINFMT_MISC=m
86		−CONFIG_HIBERNATION=y
87		−CONFIG_NET=y
88		−CONFIG_PACKET=y
89		−CONFIG_PACKET_DIAG=m
90		−CONFIG_UNIX=y
91		−CONFIG_UNIX_DIAG=m
92		−CONFIG_XFRM_USER=m
93		−CONFIG_NET_KEY=m
94		−CONFIG_SMC=m
95		−CONFIG_SMC_DIAG=m
96		−CONFIG_INET=y
97		−CONFIG_IP_MULTICAST=y
98		−CONFIG_IP_ADVANCED_ROUTER=y
99		−CONFIG_IP_MULTIPLE_TABLES=y
100		−CONFIG_IP_ROUTE_MULTIPATH=y
101		−CONFIG_IP_ROUTE_VERBOSE=y
102		−CONFIG_NET_IPIP=m
103		−CONFIG_NET_IPGRE_DEMUX=m
104		−CONFIG_NET_IPGRE=m
105		−CONFIG_NET_IPGRE_BROADCAST=y
106		−CONFIG_IP_MROUTE=y
107		−CONFIG_IP_MROUTE_MULTIPLE_TABLES=y
108		−CONFIG_IP_PIMSM_V1=y
109		−CONFIG_IP_PIMSM_V2=y
110		−CONFIG_SYN_COOKIES=y
111		−CONFIG_NET_IPVTI=m
112		−CONFIG_INET_AH=m
113		−CONFIG_INET_ESP=m
114		−CONFIG_INET_IPCOMP=m
115		−CONFIG_INET_XFRM_MODE_TRANSPORT=m
116		−CONFIG_INET_XFRM_MODE_TUNNEL=m
117		−CONFIG_INET_XFRM_MODE_BEET=m
118		−CONFIG_INET_DIAG=m
119		−CONFIG_INET_UDP_DIAG=m
120		−CONFIG_TCP_CONG_ADVANCED=y
121		−CONFIG_TCP_CONG_HSTCP=m
122		−CONFIG_TCP_CONG_HYBLA=m
123		−CONFIG_TCP_CONG_SCALABLE=m
124		−CONFIG_TCP_CONG_LP=m
125		−CONFIG_TCP_CONG_VENO=m
126		−CONFIG_TCP_CONG_YEAH=m
127		−CONFIG_TCP_CONG_ILLINOIS=m
128		−CONFIG_IPV6_ROUTER_PREF=y
129		−CONFIG_INET6_AH=m
130		−CONFIG_INET6_ESP=m
131		−CONFIG_INET6_IPCOMP=m
132		−CONFIG_IPV6_MIP6=m
133		−CONFIG_INET6_XFRM_MODE_TRANSPORT=m
134		−CONFIG_INET6_XFRM_MODE_TUNNEL=m
135		−CONFIG_INET6_XFRM_MODE_BEET=m
136		−CONFIG_INET6_XFRM_MODE_ROUTEOPTIMIZATION=m
137		−CONFIG_IPV6_VTI=m
138		−CONFIG_IPV6_SIT=m
139		−CONFIG_IPV6_GRE=m
140		−CONFIG_IPV6_MULTIPLE_TABLES=y
141		−CONFIG_IPV6_SUBTREES=y
142		−CONFIG_NETFILTER=y
143		−CONFIG_NF_CONNTRACK=m
144		−CONFIG_NF_CONNTRACK_SECMARK=y
145		−CONFIG_NF_CONNTRACK_EVENTS=y
146		−CONFIG_NF_CONNTRACK_TIMEOUT=y
147		−CONFIG_NF_CONNTRACK_TIMESTAMP=y
148		−CONFIG_NF_CONNTRACK_AMANDA=m
149		−CONFIG_NF_CONNTRACK_FTP=m
150		−CONFIG_NF_CONNTRACK_H323=m
151		−CONFIG_NF_CONNTRACK_IRC=m
152		−CONFIG_NF_CONNTRACK_NETBIOS_NS=m
153		−CONFIG_NF_CONNTRACK_SNMP=m
154		−CONFIG_NF_CONNTRACK_PPTP=m
155		−CONFIG_NF_CONNTRACK_SANE=m
156		−CONFIG_NF_CONNTRACK_SIP=m
157		−CONFIG_NF_CONNTRACK_TFTP=m
158		−CONFIG_NF_CT_NETLINK=m
159		−CONFIG_NF_CT_NETLINK_TIMEOUT=m
160		−CONFIG_NF_TABLES=m
161		−CONFIG_NFT_EXTHDR=m
162		−CONFIG_NFT_META=m
163		−CONFIG_NFT_CT=m
164		−CONFIG_NFT_COUNTER=m
165		−CONFIG_NFT_LOG=m
166		−CONFIG_NFT_LIMIT=m
167		−CONFIG_NFT_NAT=m
168		−CONFIG_NFT_COMPAT=m
169		−CONFIG_NFT_HASH=m
170		−CONFIG_NETFILTER_XT_SET=m
171		−CONFIG_NETFILTER_XT_TARGET_AUDIT=m
172		−CONFIG_NETFILTER_XT_TARGET_CHECKSUM=m
173		−CONFIG_NETFILTER_XT_TARGET_CLASSIFY=m
174		−CONFIG_NETFILTER_XT_TARGET_CONNMARK=m
175		−CONFIG_NETFILTER_XT_TARGET_CONNSECMARK=m
176		−CONFIG_NETFILTER_XT_TARGET_CT=m
177		−CONFIG_NETFILTER_XT_TARGET_DSCP=m
178		−CONFIG_NETFILTER_XT_TARGET_HMARK=m
179		−CONFIG_NETFILTER_XT_TARGET_IDLETIMER=m
180		−CONFIG_NETFILTER_XT_TARGET_LOG=m
181		−CONFIG_NETFILTER_XT_TARGET_MARK=m
182		−CONFIG_NETFILTER_XT_TARGET_NFLOG=m
183		−CONFIG_NETFILTER_XT_TARGET_NFQUEUE=m
184		−CONFIG_NETFILTER_XT_TARGET_TEE=m
185		−CONFIG_NETFILTER_XT_TARGET_TPROXY=m
186		−CONFIG_NETFILTER_XT_TARGET_TRACE=m
187		−CONFIG_NETFILTER_XT_TARGET_SECMARK=m
188		−CONFIG_NETFILTER_XT_TARGET_TCPMSS=m
189		−CONFIG_NETFILTER_XT_TARGET_TCPOPTSTRIP=m
190		−CONFIG_NETFILTER_XT_MATCH_ADDRTYPE=m
191		−CONFIG_NETFILTER_XT_MATCH_BPF=m
192		−CONFIG_NETFILTER_XT_MATCH_CLUSTER=m
193		−CONFIG_NETFILTER_XT_MATCH_COMMENT=m
194		−CONFIG_NETFILTER_XT_MATCH_CONNBYTES=m
195		−CONFIG_NETFILTER_XT_MATCH_CONNLABEL=m
196		−CONFIG_NETFILTER_XT_MATCH_CONNLIMIT=m
197		−CONFIG_NETFILTER_XT_MATCH_CONNMARK=m
198		−CONFIG_NETFILTER_XT_MATCH_CONNTRACK=m
199		−CONFIG_NETFILTER_XT_MATCH_CPU=m
200		−CONFIG_NETFILTER_XT_MATCH_DCCP=m
201		−CONFIG_NETFILTER_XT_MATCH_DEVGROUP=m
202		−CONFIG_NETFILTER_XT_MATCH_DSCP=m
203		−CONFIG_NETFILTER_XT_MATCH_ESP=m
204		−CONFIG_NETFILTER_XT_MATCH_HASHLIMIT=m
205		−CONFIG_NETFILTER_XT_MATCH_HELPER=m
206		−CONFIG_NETFILTER_XT_MATCH_IPRANGE=m
207		−CONFIG_NETFILTER_XT_MATCH_IPVS=m
208		−CONFIG_NETFILTER_XT_MATCH_LENGTH=m
209		−CONFIG_NETFILTER_XT_MATCH_LIMIT=m
210		−CONFIG_NETFILTER_XT_MATCH_MAC=m
211		−CONFIG_NETFILTER_XT_MATCH_MARK=m
212		−CONFIG_NETFILTER_XT_MATCH_MULTIPORT=m
213		−CONFIG_NETFILTER_XT_MATCH_NFACCT=m
214		−CONFIG_NETFILTER_XT_MATCH_OSF=m
215		−CONFIG_NETFILTER_XT_MATCH_OWNER=m
216		−CONFIG_NETFILTER_XT_MATCH_POLICY=m
217		−CONFIG_NETFILTER_XT_MATCH_PHYSDEV=m
218		−CONFIG_NETFILTER_XT_MATCH_PKTTYPE=m
219		−CONFIG_NETFILTER_XT_MATCH_QUOTA=m
220		−CONFIG_NETFILTER_XT_MATCH_RATEEST=m
221		−CONFIG_NETFILTER_XT_MATCH_REALM=m
222		−CONFIG_NETFILTER_XT_MATCH_RECENT=m
223		−CONFIG_NETFILTER_XT_MATCH_STATE=m
224		−CONFIG_NETFILTER_XT_MATCH_STATISTIC=m
225		−CONFIG_NETFILTER_XT_MATCH_STRING=m
226		−CONFIG_NETFILTER_XT_MATCH_TCPMSS=m
227		−CONFIG_NETFILTER_XT_MATCH_TIME=m
228		−CONFIG_NETFILTER_XT_MATCH_U32=m
229		−CONFIG_IP_SET=m
230		−CONFIG_IP_SET_BITMAP_IP=m
231		−CONFIG_IP_SET_BITMAP_IPMAC=m
232		−CONFIG_IP_SET_BITMAP_PORT=m
233		−CONFIG_IP_SET_HASH_IP=m
234		−CONFIG_IP_SET_HASH_IPPORT=m
235		−CONFIG_IP_SET_HASH_IPPORTIP=m
236		−CONFIG_IP_SET_HASH_IPPORTNET=m
237		−CONFIG_IP_SET_HASH_NETPORTNET=m
238		−CONFIG_IP_SET_HASH_NET=m
239		−CONFIG_IP_SET_HASH_NETNET=m
240		−CONFIG_IP_SET_HASH_NETPORT=m
241		−CONFIG_IP_SET_HASH_NETIFACE=m
242		−CONFIG_IP_SET_LIST_SET=m
243		−CONFIG_IP_VS=m
244		−CONFIG_IP_VS_PROTO_TCP=y
245		−CONFIG_IP_VS_PROTO_UDP=y
246		−CONFIG_IP_VS_PROTO_ESP=y
247		−CONFIG_IP_VS_PROTO_AH=y
248		−CONFIG_IP_VS_RR=m
249		−CONFIG_IP_VS_WRR=m
250		−CONFIG_IP_VS_LC=m
251		−CONFIG_IP_VS_WLC=m
252		−CONFIG_IP_VS_LBLC=m
253		−CONFIG_IP_VS_LBLCR=m
254		−CONFIG_IP_VS_DH=m
255		−CONFIG_IP_VS_SH=m
256		−CONFIG_IP_VS_SED=m
257		−CONFIG_IP_VS_NQ=m
258		−CONFIG_IP_VS_FTP=m
259		−CONFIG_IP_VS_PE_SIP=m
260		−CONFIG_NF_CONNTRACK_IPV4=m
261		−CONFIG_NF_TABLES_IPV4=m
262		−CONFIG_NFT_CHAIN_ROUTE_IPV4=m
263		−CONFIG_NF_TABLES_ARP=m
264		−CONFIG_NFT_CHAIN_NAT_IPV4=m
265		−CONFIG_IP_NF_IPTABLES=m
266		−CONFIG_IP_NF_MATCH_AH=m
267		−CONFIG_IP_NF_MATCH_ECN=m
268		−CONFIG_IP_NF_MATCH_RPFILTER=m
269		−CONFIG_IP_NF_MATCH_TTL=m
270		−CONFIG_IP_NF_FILTER=m
271		−CONFIG_IP_NF_TARGET_REJECT=m
272		−CONFIG_IP_NF_NAT=m
273		−CONFIG_IP_NF_TARGET_MASQUERADE=m
274		−CONFIG_IP_NF_MANGLE=m
275		−CONFIG_IP_NF_TARGET_CLUSTERIP=m
276		−CONFIG_IP_NF_TARGET_ECN=m
277		−CONFIG_IP_NF_TARGET_TTL=m
278		−CONFIG_IP_NF_RAW=m
279		−CONFIG_IP_NF_SECURITY=m
280		−CONFIG_IP_NF_ARPTABLES=m
281		−CONFIG_IP_NF_ARPFILTER=m
282		−CONFIG_IP_NF_ARP_MANGLE=m
283		−CONFIG_NF_CONNTRACK_IPV6=m
284		−CONFIG_NF_TABLES_IPV6=m
285		−CONFIG_NFT_CHAIN_ROUTE_IPV6=m
286		−CONFIG_NFT_CHAIN_NAT_IPV6=m
287		−CONFIG_IP6_NF_IPTABLES=m
288		−CONFIG_IP6_NF_MATCH_AH=m
289		−CONFIG_IP6_NF_MATCH_EUI64=m
290		−CONFIG_IP6_NF_MATCH_FRAG=m
291		−CONFIG_IP6_NF_MATCH_OPTS=m
292		−CONFIG_IP6_NF_MATCH_HL=m
293		−CONFIG_IP6_NF_MATCH_IPV6HEADER=m
294		−CONFIG_IP6_NF_MATCH_MH=m
295		−CONFIG_IP6_NF_MATCH_RPFILTER=m
296		−CONFIG_IP6_NF_MATCH_RT=m
297		−CONFIG_IP6_NF_TARGET_HL=m
298		−CONFIG_IP6_NF_FILTER=m
299		−CONFIG_IP6_NF_TARGET_REJECT=m
300		−CONFIG_IP6_NF_MANGLE=m
301		−CONFIG_IP6_NF_RAW=m
302		−CONFIG_IP6_NF_SECURITY=m
303		−CONFIG_IP6_NF_NAT=m
304		−CONFIG_IP6_NF_TARGET_MASQUERADE=m
305		−CONFIG_NF_TABLES_BRIDGE=m
306		−CONFIG_NET_SCTPPROBE=m
307		−CONFIG_RDS=m
308		−CONFIG_RDS_RDMA=m
309		−CONFIG_RDS_TCP=m
310		−CONFIG_L2TP=m
311		−CONFIG_L2TP_DEBUGFS=m
312		−CONFIG_L2TP_V3=y
313		−CONFIG_L2TP_IP=m
314		−CONFIG_L2TP_ETH=m
315		−CONFIG_BRIDGE=m
316		−CONFIG_VLAN_8021Q=m
317		−CONFIG_VLAN_8021Q_GVRP=y
318		−CONFIG_NET_SCHED=y
319		−CONFIG_NET_SCH_CBQ=m
320		−CONFIG_NET_SCH_HTB=m
321		−CONFIG_NET_SCH_HFSC=m
322		−CONFIG_NET_SCH_PRIO=m
323		−CONFIG_NET_SCH_MULTIQ=m
324		−CONFIG_NET_SCH_RED=m
325		−CONFIG_NET_SCH_SFB=m
326		−CONFIG_NET_SCH_SFQ=m
327		−CONFIG_NET_SCH_TEQL=m
328		−CONFIG_NET_SCH_TBF=m
329		−CONFIG_NET_SCH_GRED=m
330		−CONFIG_NET_SCH_DSMARK=m
331		−CONFIG_NET_SCH_NETEM=m
332		−CONFIG_NET_SCH_DRR=m
333		−CONFIG_NET_SCH_MQPRIO=m
334		−CONFIG_NET_SCH_CHOKE=m
335		−CONFIG_NET_SCH_QFQ=m
336		−CONFIG_NET_SCH_CODEL=m
337		−CONFIG_NET_SCH_FQ_CODEL=m
338		−CONFIG_NET_SCH_INGRESS=m
339		−CONFIG_NET_SCH_PLUG=m
340		−CONFIG_NET_CLS_BASIC=m
341		−CONFIG_NET_CLS_TCINDEX=m
342		−CONFIG_NET_CLS_ROUTE4=m
343		−CONFIG_NET_CLS_FW=m
344		−CONFIG_NET_CLS_U32=m
345		−CONFIG_CLS_U32_PERF=y
346		−CONFIG_CLS_U32_MARK=y
347		−CONFIG_NET_CLS_RSVP=m
348		−CONFIG_NET_CLS_RSVP6=m
349		−CONFIG_NET_CLS_FLOW=m
350		−CONFIG_NET_CLS_CGROUP=y
351		−CONFIG_NET_CLS_BPF=m
352		−CONFIG_NET_CLS_ACT=y
353		−CONFIG_NET_ACT_POLICE=m
354		−CONFIG_NET_ACT_GACT=m
355		−CONFIG_GACT_PROB=y
356		−CONFIG_NET_ACT_MIRRED=m
357		−CONFIG_NET_ACT_IPT=m
358		−CONFIG_NET_ACT_NAT=m
359		−CONFIG_NET_ACT_PEDIT=m
360		−CONFIG_NET_ACT_SIMP=m
361		−CONFIG_NET_ACT_SKBEDIT=m
362		−CONFIG_NET_ACT_CSUM=m
363		−CONFIG_DNS_RESOLVER=y
364		−CONFIG_NETLINK_DIAG=m
365		−CONFIG_CGROUP_NET_PRIO=y
366		−CONFIG_BPF_JIT=y
367		−CONFIG_NET_PKTGEN=m
368		−CONFIG_NET_TCPPROBE=m
369		−CONFIG_DEVTMPFS=y
370		−CONFIG_DMA_CMA=y
371		−CONFIG_CMA_SIZE_MBYTES=0
372		−CONFIG_CONNECTOR=y
373		−CONFIG_ZRAM=m
374		−CONFIG_BLK_DEV_LOOP=m
375		−CONFIG_BLK_DEV_CRYPTOLOOP=m
376		−CONFIG_BLK_DEV_DRBD=m
377		−CONFIG_BLK_DEV_NBD=m
378		−CONFIG_BLK_DEV_RAM=y
379		−CONFIG_BLK_DEV_RAM_SIZE=32768
380		−CONFIG_BLK_DEV_RAM_DAX=y
381		−CONFIG_VIRTIO_BLK=y
382		−CONFIG_ENCLOSURE_SERVICES=m
383		−CONFIG_GENWQE=m
384		−CONFIG_RAID_ATTRS=m
385		−CONFIG_SCSI=y
386		−CONFIG_BLK_DEV_SD=y
387		−CONFIG_CHR_DEV_ST=m
388		−CONFIG_CHR_DEV_OSST=m
389		−CONFIG_BLK_DEV_SR=m
390		−CONFIG_CHR_DEV_SG=y
391		−CONFIG_CHR_DEV_SCH=m
392		−CONFIG_SCSI_ENCLOSURE=m
393		−CONFIG_SCSI_CONSTANTS=y
394		−CONFIG_SCSI_LOGGING=y
395		−CONFIG_SCSI_SPI_ATTRS=m
396		−CONFIG_SCSI_FC_ATTRS=y
397		−CONFIG_SCSI_SAS_LIBSAS=m
398		−CONFIG_SCSI_SRP_ATTRS=m
399		−CONFIG_ISCSI_TCP=m
400		−CONFIG_SCSI_DEBUG=m
401		−CONFIG_ZFCP=y
402		−CONFIG_SCSI_VIRTIO=m
403		−CONFIG_SCSI_DH=y
404		−CONFIG_SCSI_DH_RDAC=m
405		−CONFIG_SCSI_DH_HP_SW=m
406		−CONFIG_SCSI_DH_EMC=m
407		−CONFIG_SCSI_DH_ALUA=m
408		−CONFIG_SCSI_OSD_INITIATOR=m
409		−CONFIG_SCSI_OSD_ULD=m
410		−CONFIG_MD=y
411		−CONFIG_BLK_DEV_MD=y
412		−CONFIG_MD_LINEAR=m
413		−CONFIG_MD_MULTIPATH=m
414		−CONFIG_MD_FAULTY=m
415		−CONFIG_BLK_DEV_DM=m
416		−CONFIG_DM_CRYPT=m
417		−CONFIG_DM_SNAPSHOT=m
418		−CONFIG_DM_THIN_PROVISIONING=m
419		−CONFIG_DM_MIRROR=m
420		−CONFIG_DM_LOG_USERSPACE=m
421		−CONFIG_DM_RAID=m
422		−CONFIG_DM_ZERO=m
423		−CONFIG_DM_MULTIPATH=m
424		−CONFIG_DM_MULTIPATH_QL=m
425		−CONFIG_DM_MULTIPATH_ST=m
426		−CONFIG_DM_DELAY=m
427		−CONFIG_DM_UEVENT=y
428		−CONFIG_DM_FLAKEY=m
429		−CONFIG_DM_VERITY=m
430		−CONFIG_DM_SWITCH=m
431		−CONFIG_NETDEVICES=y
432		−CONFIG_BONDING=m
433		−CONFIG_DUMMY=m
434		−CONFIG_EQUALIZER=m
435		−CONFIG_IFB=m
436		−CONFIG_MACVLAN=m
437		−CONFIG_MACVTAP=m
438		−CONFIG_VXLAN=m
439		−CONFIG_TUN=m
440		−CONFIG_VETH=m
441		−CONFIG_VIRTIO_NET=m
442		−CONFIG_NLMON=m
443		−# CONFIG_NET_VENDOR_ARC is not set
444		−# CONFIG_NET_VENDOR_CHELSIO is not set
445		−# CONFIG_NET_VENDOR_INTEL is not set
446		−# CONFIG_NET_VENDOR_MARVELL is not set
447		−CONFIG_MLX4_EN=m
448		−CONFIG_MLX5_CORE=m
449		−CONFIG_MLX5_CORE_EN=y
450		−# CONFIG_NET_VENDOR_NATSEMI is not set
451		−CONFIG_PPP=m
452		−CONFIG_PPP_BSDCOMP=m
453		−CONFIG_PPP_DEFLATE=m
454		−CONFIG_PPP_MPPE=m
455		−CONFIG_PPPOE=m
456		−CONFIG_PPTP=m
457		−CONFIG_PPPOL2TP=m
458		−CONFIG_PPP_ASYNC=m
459		−CONFIG_PPP_SYNC_TTY=m
460		−# CONFIG_INPUT_KEYBOARD is not set
461		−# CONFIG_INPUT_MOUSE is not set
462		−# CONFIG_SERIO is not set
463		−CONFIG_LEGACY_PTY_COUNT=0
464		−CONFIG_HW_RANDOM_VIRTIO=m
465		−CONFIG_RAW_DRIVER=m
466		−CONFIG_HANGCHECK_TIMER=m
467		−CONFIG_TN3270_FS=y
468		−# CONFIG_HWMON is not set
469		−CONFIG_WATCHDOG=y
470		−CONFIG_WATCHDOG_NOWAYOUT=y
471		−CONFIG_SOFT_WATCHDOG=m
472		−CONFIG_DIAG288_WATCHDOG=m
473		−# CONFIG_HID is not set
474		−# CONFIG_USB_SUPPORT is not set
475		−CONFIG_INFINIBAND=m
476		−CONFIG_INFINIBAND_USER_ACCESS=m
477		−CONFIG_MLX4_INFINIBAND=m
478		−CONFIG_MLX5_INFINIBAND=m
479		−CONFIG_VFIO=m
480		−CONFIG_VFIO_PCI=m
481		−CONFIG_VIRTIO_BALLOON=m
482		−CONFIG_EXT4_FS=y
483		−CONFIG_EXT4_FS_POSIX_ACL=y
484		−CONFIG_EXT4_FS_SECURITY=y
485		−CONFIG_EXT4_ENCRYPTION=y
486		−CONFIG_JBD2_DEBUG=y
487		−CONFIG_JFS_FS=m
488		−CONFIG_JFS_POSIX_ACL=y
489		−CONFIG_JFS_SECURITY=y
490		−CONFIG_JFS_STATISTICS=y
491		−CONFIG_XFS_FS=y
492		−CONFIG_XFS_QUOTA=y
493		−CONFIG_XFS_POSIX_ACL=y
494		−CONFIG_XFS_RT=y
495		−CONFIG_GFS2_FS=m
496		−CONFIG_GFS2_FS_LOCKING_DLM=y
497		−CONFIG_OCFS2_FS=m
498		−CONFIG_BTRFS_FS=y
499		−CONFIG_BTRFS_FS_POSIX_ACL=y
500		−CONFIG_NILFS2_FS=m
501		−CONFIG_FS_DAX=y
502		−CONFIG_EXPORTFS_BLOCK_OPS=y
503		−CONFIG_FANOTIFY=y
504		−CONFIG_FANOTIFY_ACCESS_PERMISSIONS=y
505		−CONFIG_QUOTA_NETLINK_INTERFACE=y
506		−CONFIG_QFMT_V1=m
507		−CONFIG_QFMT_V2=m
508		−CONFIG_AUTOFS4_FS=m
509		−CONFIG_FUSE_FS=y
510		−CONFIG_CUSE=m
511		−CONFIG_OVERLAY_FS=m
512		−CONFIG_OVERLAY_FS_REDIRECT_DIR=y
513		−CONFIG_FSCACHE=m
514		−CONFIG_CACHEFILES=m
515		−CONFIG_ISO9660_FS=y
516		−CONFIG_JOLIET=y
517		−CONFIG_ZISOFS=y
518		−CONFIG_UDF_FS=m
519		−CONFIG_MSDOS_FS=m
520		−CONFIG_VFAT_FS=m
521		−CONFIG_NTFS_FS=m
522		−CONFIG_NTFS_RW=y
523		−CONFIG_PROC_KCORE=y
524		−CONFIG_TMPFS=y
525		−CONFIG_TMPFS_POSIX_ACL=y
526		−CONFIG_HUGETLBFS=y
527		−CONFIG_CONFIGFS_FS=m
528		−CONFIG_ECRYPT_FS=m
529		−CONFIG_CRAMFS=m
530		−CONFIG_SQUASHFS=m
531		−CONFIG_SQUASHFS_XATTR=y
532		−CONFIG_SQUASHFS_LZO=y
533		−CONFIG_SQUASHFS_XZ=y
534		−CONFIG_ROMFS_FS=m
535		−CONFIG_NFS_FS=m
536		−CONFIG_NFS_V3_ACL=y
537		−CONFIG_NFS_V4=m
538		−CONFIG_NFS_SWAP=y
539		−CONFIG_NFSD=m
540		−CONFIG_NFSD_V3_ACL=y
541		−CONFIG_NFSD_V4=y
542		−CONFIG_NFSD_V4_SECURITY_LABEL=y
543		−CONFIG_CIFS=m
544		−CONFIG_CIFS_STATS=y
545		−CONFIG_CIFS_STATS2=y
546		−CONFIG_CIFS_WEAK_PW_HASH=y
547		−CONFIG_CIFS_UPCALL=y
548		−CONFIG_CIFS_XATTR=y
549		−CONFIG_CIFS_POSIX=y
550		−# CONFIG_CIFS_DEBUG is not set
551		−CONFIG_CIFS_DFS_UPCALL=y
552		−CONFIG_NLS_DEFAULT="utf8"
553		−CONFIG_NLS_CODEPAGE_437=m
554		−CONFIG_NLS_CODEPAGE_850=m
555		−CONFIG_NLS_ASCII=m
556		−CONFIG_NLS_ISO8859_1=m
557		−CONFIG_NLS_ISO8859_15=m
558		−CONFIG_NLS_UTF8=m
559		−CONFIG_DLM=m
560		−CONFIG_PRINTK_TIME=y
561		−CONFIG_DEBUG_INFO=y
562		−CONFIG_DEBUG_INFO_DWARF4=y
563		−CONFIG_GDB_SCRIPTS=y
564		−# CONFIG_ENABLE_MUST_CHECK is not set
565		−CONFIG_FRAME_WARN=1024
566		−CONFIG_UNUSED_SYMBOLS=y
567		−CONFIG_MAGIC_SYSRQ=y
568		−CONFIG_DEBUG_MEMORY_INIT=y
569		−CONFIG_PANIC_ON_OOPS=y
570		−CONFIG_RCU_TORTURE_TEST=m
571		−CONFIG_RCU_CPU_STALL_TIMEOUT=60
572		−CONFIG_LATENCYTOP=y
573		−CONFIG_SCHED_TRACER=y
574		−CONFIG_FTRACE_SYSCALLS=y
575		−CONFIG_STACK_TRACER=y
576		−CONFIG_BLK_DEV_IO_TRACE=y
577		−CONFIG_FUNCTION_PROFILER=y
578		−CONFIG_HIST_TRIGGERS=y
579		−CONFIG_LKDTM=m
580		−CONFIG_PERCPU_TEST=m
581		−CONFIG_ATOMIC64_SELFTEST=y
582		−CONFIG_TEST_BPF=m
583		−CONFIG_BUG_ON_DATA_CORRUPTION=y
584		−CONFIG_S390_PTDUMP=y
585		−CONFIG_PERSISTENT_KEYRINGS=y
586		−CONFIG_BIG_KEYS=y
587		−CONFIG_ENCRYPTED_KEYS=m
588		−CONFIG_SECURITY=y
589		−CONFIG_SECURITY_NETWORK=y
590		−CONFIG_SECURITY_SELINUX=y
591		−CONFIG_SECURITY_SELINUX_BOOTPARAM=y
592		−CONFIG_SECURITY_SELINUX_BOOTPARAM_VALUE=0
593		−CONFIG_SECURITY_SELINUX_DISABLE=y
594		−CONFIG_INTEGRITY_SIGNATURE=y
595		−CONFIG_INTEGRITY_ASYMMETRIC_KEYS=y
596		−CONFIG_IMA=y
597		−CONFIG_IMA_WRITE_POLICY=y
598		−CONFIG_IMA_APPRAISE=y
599		−CONFIG_CRYPTO_DH=m
600		−CONFIG_CRYPTO_ECDH=m
601		−CONFIG_CRYPTO_USER=m
602		−# CONFIG_CRYPTO_MANAGER_DISABLE_TESTS is not set
603		−CONFIG_CRYPTO_PCRYPT=m
604		−CONFIG_CRYPTO_CRYPTD=m
605		−CONFIG_CRYPTO_MCRYPTD=m
606		−CONFIG_CRYPTO_TEST=m
607		−CONFIG_CRYPTO_CHACHA20POLY1305=m
608		−CONFIG_CRYPTO_LRW=m
609		−CONFIG_CRYPTO_PCBC=m
610		−CONFIG_CRYPTO_KEYWRAP=m
611		−CONFIG_CRYPTO_XCBC=m
612		−CONFIG_CRYPTO_VMAC=m
613		−CONFIG_CRYPTO_CRC32=m
614		−CONFIG_CRYPTO_MICHAEL_MIC=m
615		−CONFIG_CRYPTO_RMD128=m
616		−CONFIG_CRYPTO_RMD160=m
617		−CONFIG_CRYPTO_RMD256=m
618		−CONFIG_CRYPTO_RMD320=m
619		−CONFIG_CRYPTO_SHA512=m
620		−CONFIG_CRYPTO_SHA3=m
621		−CONFIG_CRYPTO_TGR192=m
622		−CONFIG_CRYPTO_WP512=m
623		−CONFIG_CRYPTO_AES_TI=m
624		−CONFIG_CRYPTO_ANUBIS=m
625		−CONFIG_CRYPTO_BLOWFISH=m
626		−CONFIG_CRYPTO_CAMELLIA=m
627		−CONFIG_CRYPTO_CAST5=m
628		−CONFIG_CRYPTO_CAST6=m
629		−CONFIG_CRYPTO_FCRYPT=m
630		−CONFIG_CRYPTO_KHAZAD=m
631		−CONFIG_CRYPTO_SALSA20=m
632		−CONFIG_CRYPTO_SEED=m
633		−CONFIG_CRYPTO_SERPENT=m
634		−CONFIG_CRYPTO_TEA=m
635		−CONFIG_CRYPTO_TWOFISH=m
636		−CONFIG_CRYPTO_842=m
637		−CONFIG_CRYPTO_LZ4=m
638		−CONFIG_CRYPTO_LZ4HC=m
639		−CONFIG_CRYPTO_ANSI_CPRNG=m
640		−CONFIG_CRYPTO_USER_API_HASH=m
641		−CONFIG_CRYPTO_USER_API_SKCIPHER=m
642		−CONFIG_CRYPTO_USER_API_RNG=m
643		−CONFIG_CRYPTO_USER_API_AEAD=m
644		−CONFIG_ZCRYPT=m
645		−CONFIG_PKEY=m
646		−CONFIG_CRYPTO_PAES_S390=m
647		−CONFIG_CRYPTO_SHA1_S390=m
648		−CONFIG_CRYPTO_SHA256_S390=m
649		−CONFIG_CRYPTO_SHA512_S390=m
650		−CONFIG_CRYPTO_DES_S390=m
651		−CONFIG_CRYPTO_AES_S390=m
652		−CONFIG_CRYPTO_GHASH_S390=m
653		−CONFIG_CRYPTO_CRC32_S390=y
654		−CONFIG_CRC7=m
655		−CONFIG_CRC8=m
656		−CONFIG_CORDIC=m
657		−CONFIG_CMM=m
658		−CONFIG_APPLDATA_BASE=y
659		−CONFIG_KVM=m
660		−CONFIG_KVM_S390_UCONTROL=y
661		−CONFIG_VHOST_NET=m

arch/s390/configs/performance_defconfig+15−5

@@ -25,13 +25,13 @@ CONFIG_CPUSETS=y · 24 unmodified lines
25	25	CONFIG_CGROUP_DEVICE=y
26	26	CONFIG_CGROUP_CPUACCT=y
27	27	CONFIG_CGROUP_PERF=y
28		−CONFIG_CHECKPOINT_RESTORE=y
29	28	CONFIG_NAMESPACES=y
30	29	CONFIG_USER_NS=y
31	30	CONFIG_SCHED_AUTOGROUP=y
32	31	CONFIG_BLK_DEV_INITRD=y
33	32	CONFIG_EXPERT=y
34	33	# CONFIG_SYSFS_SYSCALL is not set
	34	+CONFIG_CHECKPOINT_RESTORE=y
35	35	CONFIG_BPF_SYSCALL=y
36	36	CONFIG_USERFAULTFD=y
37	37	# CONFIG_COMPAT_BRK is not set
@@ -45,6 +45,8 @@ CONFIG_MODULE_UNLOAD=y · 7 unmodified lines
45	45	CONFIG_MODULE_FORCE_UNLOAD=y
46	46	CONFIG_MODVERSIONS=y
47	47	CONFIG_MODULE_SRCVERSION_ALL=y
	48	+CONFIG_MODULE_SIG=y
	49	+CONFIG_MODULE_SIG_SHA256=y
48	50	CONFIG_BLK_DEV_INTEGRITY=y
49	51	CONFIG_BLK_DEV_THROTTLING=y
50	52	CONFIG_BLK_WBT=y
@@ -62,6 +64,7 @@ CONFIG_TUNE_ZEC12=y · 11 unmodified lines
62	64	CONFIG_NR_CPUS=512
63	65	CONFIG_NUMA=y
64	66	CONFIG_HZ_100=y
	67	+CONFIG_KEXEC_FILE=y
65	68	CONFIG_MEMORY_HOTPLUG=y
66	69	CONFIG_MEMORY_HOTREMOVE=y
67	70	CONFIG_KSM=y
@@ -301,7 +304,6 @@ CONFIG_IP6_NF_SECURITY=m · 233 unmodified lines
301	304	CONFIG_IP6_NF_NAT=m
302	305	CONFIG_IP6_NF_TARGET_MASQUERADE=m
303	306	CONFIG_NF_TABLES_BRIDGE=m
304		−CONFIG_NET_SCTPPROBE=m
305	307	CONFIG_RDS=m
306	308	CONFIG_RDS_RDMA=m
307	309	CONFIG_RDS_TCP=m
@@ -359,11 +361,11 @@ CONFIG_NET_ACT_SIMP=m · 51 unmodified lines
359	361	CONFIG_NET_ACT_SKBEDIT=m
360	362	CONFIG_NET_ACT_CSUM=m
361	363	CONFIG_DNS_RESOLVER=y
	364	+CONFIG_OPENVSWITCH=m
362	365	CONFIG_NETLINK_DIAG=m
363	366	CONFIG_CGROUP_NET_PRIO=y
364	367	CONFIG_BPF_JIT=y
365	368	CONFIG_NET_PKTGEN=m
366		−CONFIG_NET_TCPPROBE=m
367	369	CONFIG_DEVTMPFS=y
368	370	CONFIG_DMA_CMA=y
369	371	CONFIG_CMA_SIZE_MBYTES=0
@@ -375,8 +377,9 @@ CONFIG_BLK_DEV_DRBD=m · 5 unmodified lines
375	377	CONFIG_BLK_DEV_NBD=m
376	378	CONFIG_BLK_DEV_RAM=y
377	379	CONFIG_BLK_DEV_RAM_SIZE=32768
378		−CONFIG_BLK_DEV_RAM_DAX=y
379	380	CONFIG_VIRTIO_BLK=y
	381	+CONFIG_BLK_DEV_RBD=m
	382	+CONFIG_BLK_DEV_NVME=m
380	383	CONFIG_ENCLOSURE_SERVICES=m
381	384	CONFIG_GENWQE=m
382	385	CONFIG_RAID_ATTRS=m
@@ -455,6 +458,7 @@ CONFIG_PPTP=m · 72 unmodified lines
455	458	CONFIG_PPPOL2TP=m
456	459	CONFIG_PPP_ASYNC=m
457	460	CONFIG_PPP_SYNC_TTY=m
	461	+CONFIG_INPUT_EVDEV=y
458	462	# CONFIG_INPUT_KEYBOARD is not set
459	463	# CONFIG_INPUT_MOUSE is not set
460	464	# CONFIG_SERIO is not set
@@ -468,6 +472,9 @@ CONFIG_WATCHDOG=y · 7 unmodified lines
468	472	CONFIG_WATCHDOG_NOWAYOUT=y
469	473	CONFIG_SOFT_WATCHDOG=m
470	474	CONFIG_DIAG288_WATCHDOG=m
	475	+CONFIG_DRM=y
	476	+CONFIG_DRM_VIRTIO_GPU=y
	477	+CONFIG_FRAMEBUFFER_CONSOLE=y
471	478	# CONFIG_HID is not set
472	479	# CONFIG_USB_SUPPORT is not set
473	480	CONFIG_INFINIBAND=m
@@ -476,7 +483,9 @@ CONFIG_MLX4_INFINIBAND=m · 2 unmodified lines
476	483	CONFIG_MLX5_INFINIBAND=m
477	484	CONFIG_VFIO=m
478	485	CONFIG_VFIO_PCI=m
	486	+CONFIG_VIRTIO_PCI=m
479	487	CONFIG_VIRTIO_BALLOON=m
	488	+CONFIG_VIRTIO_INPUT=y
480	489	CONFIG_EXT4_FS=y
481	490	CONFIG_EXT4_FS_POSIX_ACL=y
482	491	CONFIG_EXT4_FS_SECURITY=y
@@ -507,7 +516,6 @@ CONFIG_AUTOFS4_FS=m · 24 unmodified lines
507	516	CONFIG_FUSE_FS=y
508	517	CONFIG_CUSE=m
509	518	CONFIG_OVERLAY_FS=m
510		−CONFIG_OVERLAY_FS_REDIRECT_DIR=y
511	519	CONFIG_FSCACHE=m
512	520	CONFIG_CACHEFILES=m
513	521	CONFIG_ISO9660_FS=y
@@ -592,8 +600,10 @@ CONFIG_SECURITY_SELINUX_DISABLE=y · 78 unmodified lines
592	600	CONFIG_INTEGRITY_SIGNATURE=y
593	601	CONFIG_INTEGRITY_ASYMMETRIC_KEYS=y
594	602	CONFIG_IMA=y
	603	+CONFIG_IMA_DEFAULT_HASH_SHA256=y
595	604	CONFIG_IMA_WRITE_POLICY=y
596	605	CONFIG_IMA_APPRAISE=y
	606	+CONFIG_CRYPTO_FIPS=y
597	607	CONFIG_CRYPTO_DH=m
598	608	CONFIG_CRYPTO_ECDH=m
599	609	CONFIG_CRYPTO_USER=m

arch/s390/defconfig+11−2

@@ -8,6 +8,7 @@ CONFIG_TASKSTATS=y · 7 unmodified lines
8	8	CONFIG_TASK_DELAY_ACCT=y
9	9	CONFIG_TASK_XACCT=y
10	10	CONFIG_TASK_IO_ACCOUNTING=y
	11	+# CONFIG_CPU_ISOLATION is not set
11	12	CONFIG_IKCONFIG=y
12	13	CONFIG_IKCONFIG_PROC=y
13	14	CONFIG_CGROUPS=y
@@ -23,12 +24,12 @@ CONFIG_CPUSETS=y · 9 unmodified lines
23	24	CONFIG_CGROUP_DEVICE=y
24	25	CONFIG_CGROUP_CPUACCT=y
25	26	CONFIG_CGROUP_PERF=y
26		−CONFIG_CHECKPOINT_RESTORE=y
27	27	CONFIG_NAMESPACES=y
28	28	CONFIG_USER_NS=y
29	29	CONFIG_BLK_DEV_INITRD=y
30	30	CONFIG_EXPERT=y
31	31	# CONFIG_SYSFS_SYSCALL is not set
	32	+CONFIG_CHECKPOINT_RESTORE=y
32	33	CONFIG_BPF_SYSCALL=y
33	34	CONFIG_USERFAULTFD=y
34	35	# CONFIG_COMPAT_BRK is not set
@@ -47,6 +48,7 @@ CONFIG_LIVEPATCH=y · 12 unmodified lines
47	48	CONFIG_NR_CPUS=256
48	49	CONFIG_NUMA=y
49	50	CONFIG_HZ_100=y
	51	+CONFIG_KEXEC_FILE=y
50	52	CONFIG_MEMORY_HOTPLUG=y
51	53	CONFIG_MEMORY_HOTREMOVE=y
52	54	CONFIG_KSM=y
@@ -129,10 +131,13 @@ CONFIG_EQUALIZER=m · 76 unmodified lines
129	131	CONFIG_TUN=m
130	132	CONFIG_VIRTIO_NET=y
131	133	# CONFIG_NET_VENDOR_ALACRITECH is not set
	134	+# CONFIG_NET_VENDOR_CORTINA is not set
132	135	# CONFIG_NET_VENDOR_SOLARFLARE is not set
	136	+# CONFIG_NET_VENDOR_SOCIONEXT is not set
133	137	# CONFIG_NET_VENDOR_SYNOPSYS is not set
134	138	# CONFIG_INPUT is not set
135	139	# CONFIG_SERIO is not set
	140	+# CONFIG_VT is not set
136	141	CONFIG_DEVKMEM=y
137	142	CONFIG_RAW_DRIVER=m
138	143	CONFIG_VIRTIO_BALLOON=y
@@ -177,13 +182,15 @@ CONFIG_TRACER_SNAPSHOT_PER_CPU_SWAP=y · 38 unmodified lines
177	182	CONFIG_STACK_TRACER=y
178	183	CONFIG_BLK_DEV_IO_TRACE=y
179	184	CONFIG_FUNCTION_PROFILER=y
180		−CONFIG_KPROBES_SANITY_TEST=y
	185	+# CONFIG_RUNTIME_TESTING_MENU is not set
181	186	CONFIG_S390_PTDUMP=y
182	187	CONFIG_CRYPTO_CRYPTD=m
	188	+CONFIG_CRYPTO_AUTHENC=m
183	189	CONFIG_CRYPTO_TEST=m
184	190	CONFIG_CRYPTO_CCM=m
185	191	CONFIG_CRYPTO_GCM=m
186	192	CONFIG_CRYPTO_CBC=y
	193	+CONFIG_CRYPTO_CFB=m
187	194	CONFIG_CRYPTO_CTS=m
188	195	CONFIG_CRYPTO_LRW=m
189	196	CONFIG_CRYPTO_PCBC=m
@@ -213,6 +220,8 @@ CONFIG_CRYPTO_KHAZAD=m · 23 unmodified lines
213	220	CONFIG_CRYPTO_SALSA20=m
214	221	CONFIG_CRYPTO_SEED=m
215	222	CONFIG_CRYPTO_SERPENT=m
	223	+CONFIG_CRYPTO_SM4=m
	224	+CONFIG_CRYPTO_SPECK=m
216	225	CONFIG_CRYPTO_TEA=m
217	226	CONFIG_CRYPTO_TWOFISH=m
218	227	CONFIG_CRYPTO_DEFLATE=m

arch/s390/hypfs/inode.c+1−1

@@ -320,7 +320,7 @@ static void hypfs_kill_super(struct super_block *sb) · 319 unmodified lines
320	320
321	321	if (sb->s_root)
322	322	hypfs_delete_tree(sb->s_root);
323		− if (sb_info->update_file)
	323	+ if (sb_info && sb_info->update_file)
324	324	hypfs_remove(sb_info->update_file);
325	325	kfree(sb->s_fs_info);
326	326	sb->s_fs_info = NULL;

arch/s390/include/asm/kexec.h+23−0

@@ -46,4 +46,27 @@ · 45 unmodified lines
46	46	static inline void crash_setup_regs(struct pt_regs *newregs,
47	47	struct pt_regs *oldregs) { }
48	48
	49	+struct kimage;
	50	+struct s390_load_data {
	51	+ /* Pointer to the kernel buffer. Used to register cmdline etc.. */
	52	+ void *kernel_buf;
	53	+
	54	+ /* Total size of loaded segments in memory. Used as an offset. */
	55	+ size_t memsz;
	56	+
	57	+ /* Load address of initrd. Used to register INITRD_START in kernel. */
	58	+ unsigned long initrd_load_addr;
	59	+};
	60	+
	61	+int kexec_file_add_purgatory(struct kimage *image,
	62	+ struct s390_load_data *data);
	63	+int kexec_file_add_initrd(struct kimage *image,
	64	+ struct s390_load_data *data,
	65	+ char *initrd, unsigned long initrd_len);
	66	+int kexec_file_update_kernel(struct kimage iamge,
	67	+ struct s390_load_data *data);
	68	+
	69	+extern const struct kexec_file_ops s390_kexec_image_ops;
	70	+extern const struct kexec_file_ops s390_kexec_elf_ops;
	71	+
49	72	#endif /_S390_KEXEC_H /

arch/s390/include/asm/purgatory.h+17−0

@@ -0,0 +1,17 @@
	1	+/* SPDX-License-Identifier: GPL-2.0 */
	2	+/*
	3	+ * Copyright IBM Corp. 2018
	4	+ *
	5	+ * Author(s): Philipp Rudo <prudo@linux.vnet.ibm.com>
	6	+ */
	7	+
	8	+#ifndef _S390_PURGATORY_H_
	9	+#define _S390_PURGATORY_H_
	10	+#ifndef __ASSEMBLY__
	11	+
	12	+#include <linux/purgatory.h>
	13	+
	14	+int verify_sha256_digest(void);
	15	+
	16	+#endif /* __ASSEMBLY__ */
	17	+#endif /* _S390_PURGATORY_H_ */

arch/s390/include/asm/setup.h+27−13

@@ -1,7 +1,7 @@
1	1	/* SPDX-License-Identifier: GPL-2.0 */
2	2	/*
3	3	* S390 version
4		− * Copyright IBM Corp. 1999, 2010
	4	+ * Copyright IBM Corp. 1999, 2017
5	5	*/
6	6	#ifndef _ASM_S390_SETUP_H
7	7	#define _ASM_S390_SETUP_H
@@ -37,17 +37,31 @@ · 29 unmodified lines
37	37	#define LPP_MAGIC _BITUL(31)
38	38	#define LPP_PID_MASK _AC(0xffffffff, UL)
39	39
	40	+/* Offsets to entry points in kernel/head.S */
	41	+
	42	+#define STARTUP_NORMAL_OFFSET 0x10000
	43	+#define STARTUP_KDUMP_OFFSET 0x10010
	44	+
	45	+/* Offsets to parameters in kernel/head.S */
	46	+
	47	+#define IPL_DEVICE_OFFSET 0x10400
	48	+#define INITRD_START_OFFSET 0x10408
	49	+#define INITRD_SIZE_OFFSET 0x10410
	50	+#define OLDMEM_BASE_OFFSET 0x10418
	51	+#define OLDMEM_SIZE_OFFSET 0x10420
	52	+#define COMMAND_LINE_OFFSET 0x10480
	53	+
40	54	#ifndef __ASSEMBLY__
41	55
42	56	#include <asm/lowcore.h>
43	57	#include <asm/types.h>
44	58
45		−#define IPL_DEVICE ((unsigned long ) (0x10400))
46		−#define INITRD_START ((unsigned long ) (0x10408))
47		−#define INITRD_SIZE ((unsigned long ) (0x10410))
48		−#define OLDMEM_BASE ((unsigned long ) (0x10418))
49		−#define OLDMEM_SIZE ((unsigned long ) (0x10420))
50		−#define COMMAND_LINE ((char *) (0x10480))
	59	+#define IPL_DEVICE ((unsigned long ) (IPL_DEVICE_OFFSET))
	60	+#define INITRD_START ((unsigned long ) (INITRD_START_OFFSET))
	61	+#define INITRD_SIZE ((unsigned long ) (INITRD_SIZE_OFFSET))
	62	+#define OLDMEM_BASE ((unsigned long ) (OLDMEM_BASE_OFFSET))
	63	+#define OLDMEM_SIZE ((unsigned long ) (OLDMEM_SIZE_OFFSET))
	64	+#define COMMAND_LINE ((char *) (COMMAND_LINE_OFFSET))
51	65
52	66	extern int memory_end_set;
53	67	extern unsigned long memory_end;
@@ -121,12 +135,12 @@ extern void (*_machine_power_off)(void); · 67 unmodified lines
121	135
122	136	#else /* __ASSEMBLY__ */
123	137
124		−#define IPL_DEVICE 0x10400
125		−#define INITRD_START 0x10408
126		−#define INITRD_SIZE 0x10410
127		−#define OLDMEM_BASE 0x10418
128		−#define OLDMEM_SIZE 0x10420
129		−#define COMMAND_LINE 0x10480
	138	+#define IPL_DEVICE (IPL_DEVICE_OFFSET)
	139	+#define INITRD_START (INITRD_START_OFFSET)
	140	+#define INITRD_SIZE (INITRD_SIZE_OFFSET)
	141	+#define OLDMEM_BASE (OLDMEM_BASE_OFFSET)
	142	+#define OLDMEM_SIZE (OLDMEM_SIZE_OFFSET)
	143	+#define COMMAND_LINE (COMMAND_LINE_OFFSET)
130	144
131	145	#endif /* __ASSEMBLY__ */
132	146	#endif /* _ASM_S390_SETUP_H */

arch/s390/include/asm/thread_info.h+3−0

@@ -45,6 +45,9 @@ struct thread_info { · 44 unmodified lines
45	45	void arch_release_task_struct(struct task_struct *tsk);
46	46	int arch_dup_task_struct(struct task_struct dst, struct task_struct src);
47	47
	48	+void arch_setup_new_exec(void);
	49	+#define arch_setup_new_exec arch_setup_new_exec
	50	+
48	51	#endif
49	52
50	53	/*

arch/s390/include/uapi/asm/signal.h+16−7

@@ -97,22 +97,31 @@ typedef unsigned long sigset_t; · 96 unmodified lines
97	97	#include <asm-generic/signal-defs.h>
98	98
99	99	#ifndef __KERNEL__
100		−/* Here we must cater to libcs that poke about in kernel headers. */
101	100
	101	+/*
	102	+ * There are two system calls in regard to sigaction, sys_rt_sigaction
	103	+ * and sys_sigaction. Internally the kernel uses the struct old_sigaction
	104	+ * for the older sys_sigaction system call, and the kernel version of the
	105	+ * struct sigaction for the newer sys_rt_sigaction.
	106	+ *
	107	+ * The uapi definition for struct sigaction has made a strange distinction
	108	+ * between 31-bit and 64-bit in the past. For 64-bit the uapi structure
	109	+ * looks like the kernel struct sigaction, but for 31-bit it used to
	110	+ * look like the kernel struct old_sigaction. That practically made the
	111	+ * structure unusable for either system call. To get around this problem
	112	+ * the glibc always had its own definitions for the sigaction structures.
	113	+ *
	114	+ * The current struct sigaction uapi definition below is suitable for the
	115	+ * sys_rt_sigaction system call only.
	116	+ */
102	117	struct sigaction {
103	118	union {
104	119	__sighandler_t _sa_handler;
105	120	void (_sa_sigaction)(int, struct siginfo , void *);
106	121	} _u;
107		−#ifndef __s390x__ /* lovely */
108		− sigset_t sa_mask;
109		− unsigned long sa_flags;
110		− void (*sa_restorer)(void);
111		−#else /* __s390x__ */
112	122	unsigned long sa_flags;
113	123	void (*sa_restorer)(void);
114	124	sigset_t sa_mask;
115		−#endif /* __s390x__ */
116	125	};
117	126
118	127	#define sa_handler _u._sa_handler

arch/s390/kernel/Makefile+3−0

@@ -82,6 +82,9 @@ obj-$(CONFIG_FUNCTION_TRACER) += mcount.o ftrace.o · 81 unmodified lines
82	82	obj-$(CONFIG_CRASH_DUMP) += crash_dump.o
83	83	obj-$(CONFIG_UPROBES) += uprobes.o
84	84
	85	+obj-$(CONFIG_KEXEC_FILE) += machine_kexec_file.o kexec_image.o
	86	+obj-$(CONFIG_KEXEC_FILE) += kexec_elf.o
	87	+
85	88	obj-$(CONFIG_PERF_EVENTS) += perf_event.o perf_cpum_cf.o perf_cpum_sf.o
86	89	obj-$(CONFIG_PERF_EVENTS) += perf_cpum_cf_events.o perf_regs.o
87	90

arch/s390/kernel/asm-offsets.c+5−0

@@ -10,6 +10,7 @@ · 9 unmodified lines
10	10	#include <linux/kbuild.h>
11	11	#include <linux/kvm_host.h>
12	12	#include <linux/sched.h>
	13	+#include <linux/purgatory.h>
13	14	#include <asm/idle.h>
14	15	#include <asm/vdso.h>
15	16	#include <asm/pgtable.h>
@@ -204,5 +205,9 @@ int main(void) · 188 unmodified lines
204	205	OFFSET(__GMAP_ASCE, gmap, asce);
205	206	OFFSET(__SIE_PROG0C, kvm_s390_sie_block, prog0c);
206	207	OFFSET(__SIE_PROG20, kvm_s390_sie_block, prog20);
	208	+ /* kexec_sha_region */
	209	+ OFFSET(__KEXEC_SHA_REGION_START, kexec_sha_region, start);
	210	+ OFFSET(__KEXEC_SHA_REGION_LEN, kexec_sha_region, len);
	211	+ DEFINE(__KEXEC_SHA_REGION_SIZE, sizeof(struct kexec_sha_region));
207	212	return 0;
208	213	}

arch/s390/kernel/compat_wrapper.c+1−0

@@ -182,3 +182,4 @@ COMPAT_SYSCALL_WRAP6(copy_file_range, int, fd_in, loff_t __user *, off_in, int, · 181 unmodified lines
182	182	COMPAT_SYSCALL_WRAP2(s390_guarded_storage, int, command, struct gs_cb *, gs_cb);
183	183	COMPAT_SYSCALL_WRAP5(statx, int, dfd, const char __user , path, unsigned, flags, unsigned, mask, struct statx __user , buffer);
184	184	COMPAT_SYSCALL_WRAP4(s390_sthyi, unsigned long, code, void __user , info, u64 __user , rc, unsigned long, flags);
	185	+COMPAT_SYSCALL_WRAP5(kexec_file_load, int, kernel_fd, int, initrd_fd, unsigned long, cmdline_len, const char __user *, cmdline_ptr, unsigned long, flags)

arch/s390/kernel/kexec_elf.c+147−0

@@ -0,0 +1,147 @@
	1	+// SPDX-License-Identifier: GPL-2.0
	2	+/*
	3	+ * ELF loader for kexec_file_load system call.
	4	+ *
	5	+ * Copyright IBM Corp. 2018
	6	+ *
	7	+ * Author(s): Philipp Rudo <prudo@linux.vnet.ibm.com>
	8	+ */
	9	+
	10	+#include <linux/errno.h>
	11	+#include <linux/kernel.h>
	12	+#include <linux/kexec.h>
	13	+#include <asm/setup.h>
	14	+
	15	+static int kexec_file_add_elf_kernel(struct kimage *image,
	16	+ struct s390_load_data *data,
	17	+ char *kernel, unsigned long kernel_len)
	18	+{
	19	+ struct kexec_buf buf;
	20	+ const Elf_Ehdr *ehdr;
	21	+ const Elf_Phdr *phdr;
	22	+ int i, ret;
	23	+
	24	+ ehdr = (Elf_Ehdr *)kernel;
	25	+ buf.image = image;
	26	+
	27	+ phdr = (void *)ehdr + ehdr->e_phoff;
	28	+ for (i = 0; i < ehdr->e_phnum; i++, phdr++) {
	29	+ if (phdr->p_type != PT_LOAD)
	30	+ continue;
	31	+
	32	+ buf.buffer = kernel + phdr->p_offset;
	33	+ buf.bufsz = phdr->p_filesz;
	34	+
	35	+ buf.mem = ALIGN(phdr->p_paddr, phdr->p_align);
	36	+ buf.memsz = phdr->p_memsz;
	37	+
	38	+ if (phdr->p_paddr == 0) {
	39	+ data->kernel_buf = buf.buffer;
	40	+ data->memsz += STARTUP_NORMAL_OFFSET;
	41	+
	42	+ buf.buffer += STARTUP_NORMAL_OFFSET;
	43	+ buf.bufsz -= STARTUP_NORMAL_OFFSET;
	44	+
	45	+ buf.mem += STARTUP_NORMAL_OFFSET;
	46	+ buf.memsz -= STARTUP_NORMAL_OFFSET;
	47	+ }
	48	+
	49	+ if (image->type == KEXEC_TYPE_CRASH)
	50	+ buf.mem += crashk_res.start;
	51	+
	52	+ ret = kexec_add_buffer(&buf);
	53	+ if (ret)
	54	+ return ret;
	55	+
	56	+ data->memsz += buf.memsz;
	57	+ }
	58	+
	59	+ return 0;
	60	+}
	61	+
	62	+static void s390_elf_load(struct kimage image,
	63	+ char *kernel, unsigned long kernel_len,
	64	+ char *initrd, unsigned long initrd_len,
	65	+ char *cmdline, unsigned long cmdline_len)
	66	+{
	67	+ struct s390_load_data data = {0};
	68	+ const Elf_Ehdr *ehdr;
	69	+ const Elf_Phdr *phdr;
	70	+ size_t size;
	71	+ int i, ret;
	72	+
	73	+ /* image->fobs->probe already checked for valid ELF magic number. */
	74	+ ehdr = (Elf_Ehdr *)kernel;
	75	+
	76	+ if (ehdr->e_type != ET_EXEC \|\|
	77	+ ehdr->e_ident[EI_CLASS] != ELFCLASS64 \|\|
	78	+ !elf_check_arch(ehdr))
	79	+ return ERR_PTR(-EINVAL);
	80	+
	81	+ if (!ehdr->e_phnum \|\| ehdr->e_phentsize != sizeof(Elf_Phdr))
	82	+ return ERR_PTR(-EINVAL);
	83	+
	84	+ size = ehdr->e_ehsize + ehdr->e_phoff;
	85	+ size += ehdr->e_phentsize * ehdr->e_phnum;
	86	+ if (size > kernel_len)
	87	+ return ERR_PTR(-EINVAL);
	88	+
	89	+ phdr = (void *)ehdr + ehdr->e_phoff;
	90	+ size = ALIGN(size, phdr->p_align);
	91	+ for (i = 0; i < ehdr->e_phnum; i++, phdr++) {
	92	+ if (phdr->p_type == PT_INTERP)
	93	+ return ERR_PTR(-EINVAL);
	94	+
	95	+ if (phdr->p_offset > kernel_len)
	96	+ return ERR_PTR(-EINVAL);
	97	+
	98	+ size += ALIGN(phdr->p_filesz, phdr->p_align);
	99	+ }
	100	+
	101	+ if (size > kernel_len)
	102	+ return ERR_PTR(-EINVAL);
	103	+
	104	+ ret = kexec_file_add_elf_kernel(image, &data, kernel, kernel_len);
	105	+ if (ret)
	106	+ return ERR_PTR(ret);
	107	+
	108	+ if (!data.memsz)
	109	+ return ERR_PTR(-EINVAL);
	110	+
	111	+ if (initrd) {
	112	+ ret = kexec_file_add_initrd(image, &data, initrd, initrd_len);
	113	+ if (ret)
	114	+ return ERR_PTR(ret);
	115	+ }
	116	+
	117	+ ret = kexec_file_add_purgatory(image, &data);
	118	+ if (ret)
	119	+ return ERR_PTR(ret);
	120	+
	121	+ return kexec_file_update_kernel(image, &data);
	122	+}
	123	+
	124	+static int s390_elf_probe(const char *buf, unsigned long len)
	125	+{
	126	+ const Elf_Ehdr *ehdr;
	127	+
	128	+ if (len < sizeof(Elf_Ehdr))
	129	+ return -ENOEXEC;
	130	+
	131	+ ehdr = (Elf_Ehdr *)buf;
	132	+
	133	+ /* Only check the ELF magic number here and do proper validity check
	134	+ * in the loader. Any check here that fails would send the erroneous
	135	+ * ELF file to the image loader that does not care what it gets.
	136	+ * (Most likely) causing behavior not intended by the user.
	137	+ */
	138	+ if (memcmp(ehdr->e_ident, ELFMAG, SELFMAG) != 0)
	139	+ return -ENOEXEC;
	140	+
	141	+ return 0;
	142	+}
	143	+
	144	+const struct kexec_file_ops s390_kexec_elf_ops = {
	145	+ .probe = s390_elf_probe,
	146	+ .load = s390_elf_load,
	147	+};

arch/s390/kernel/kexec_image.c+76−0

@@ -0,0 +1,76 @@
	1	+// SPDX-License-Identifier: GPL-2.0
	2	+/*
	3	+ * Image loader for kexec_file_load system call.
	4	+ *
	5	+ * Copyright IBM Corp. 2018
	6	+ *
	7	+ * Author(s): Philipp Rudo <prudo@linux.vnet.ibm.com>
	8	+ */
	9	+
	10	+#include <linux/errno.h>
	11	+#include <linux/kernel.h>
	12	+#include <linux/kexec.h>
	13	+#include <asm/setup.h>
	14	+
	15	+static int kexec_file_add_image_kernel(struct kimage *image,
	16	+ struct s390_load_data *data,
	17	+ char *kernel, unsigned long kernel_len)
	18	+{
	19	+ struct kexec_buf buf;
	20	+ int ret;
	21	+
	22	+ buf.image = image;
	23	+
	24	+ buf.buffer = kernel + STARTUP_NORMAL_OFFSET;
	25	+ buf.bufsz = kernel_len - STARTUP_NORMAL_OFFSET;
	26	+
	27	+ buf.mem = STARTUP_NORMAL_OFFSET;
	28	+ if (image->type == KEXEC_TYPE_CRASH)
	29	+ buf.mem += crashk_res.start;
	30	+ buf.memsz = buf.bufsz;
	31	+
	32	+ ret = kexec_add_buffer(&buf);
	33	+
	34	+ data->kernel_buf = kernel;
	35	+ data->memsz += buf.memsz + STARTUP_NORMAL_OFFSET;
	36	+
	37	+ return ret;
	38	+}
	39	+
	40	+static void s390_image_load(struct kimage image,
	41	+ char *kernel, unsigned long kernel_len,
	42	+ char *initrd, unsigned long initrd_len,
	43	+ char *cmdline, unsigned long cmdline_len)
	44	+{
	45	+ struct s390_load_data data = {0};
	46	+ int ret;
	47	+
	48	+ ret = kexec_file_add_image_kernel(image, &data, kernel, kernel_len);
	49	+ if (ret)
	50	+ return ERR_PTR(ret);
	51	+
	52	+ if (initrd) {
	53	+ ret = kexec_file_add_initrd(image, &data, initrd, initrd_len);
	54	+ if (ret)
	55	+ return ERR_PTR(ret);
	56	+ }
	57	+
	58	+ ret = kexec_file_add_purgatory(image, &data);
	59	+ if (ret)
	60	+ return ERR_PTR(ret);
	61	+
	62	+ return kexec_file_update_kernel(image, &data);
	63	+}
	64	+
	65	+static int s390_image_probe(const char *buf, unsigned long len)
	66	+{
	67	+ /* Can't reliably tell if an image is valid. Therefore give the
	68	+ * user whatever he wants.
	69	+ */
	70	+ return 0;
	71	+}
	72	+
	73	+const struct kexec_file_ops s390_kexec_image_ops = {
	74	+ .probe = s390_image_probe,
	75	+ .load = s390_image_load,
	76	+};

arch/s390/kernel/machine_kexec_file.c+245−0

@@ -0,0 +1,245 @@
	1	+// SPDX-License-Identifier: GPL-2.0
	2	+/*
	3	+ * s390 code for kexec_file_load system call
	4	+ *
	5	+ * Copyright IBM Corp. 2018
	6	+ *
	7	+ * Author(s): Philipp Rudo <prudo@linux.vnet.ibm.com>
	8	+ */
	9	+
	10	+#include <linux/elf.h>
	11	+#include <linux/kexec.h>
	12	+#include <asm/setup.h>
	13	+
	14	+const struct kexec_file_ops * const kexec_file_loaders[] = {
	15	+ &s390_kexec_elf_ops,
	16	+ &s390_kexec_image_ops,
	17	+ NULL,
	18	+};
	19	+
	20	+int kexec_file_update_kernel(struct kimage image,
	21	+ struct s390_load_data *data)
	22	+{
	23	+ unsigned long *loc;
	24	+
	25	+ if (image->cmdline_buf_len >= ARCH_COMMAND_LINE_SIZE)
	26	+ return ERR_PTR(-EINVAL);
	27	+
	28	+ if (image->cmdline_buf_len)
	29	+ memcpy(data->kernel_buf + COMMAND_LINE_OFFSET,
	30	+ image->cmdline_buf, image->cmdline_buf_len);
	31	+
	32	+ if (image->type == KEXEC_TYPE_CRASH) {
	33	+ loc = (unsigned long *)(data->kernel_buf + OLDMEM_BASE_OFFSET);
	34	+ *loc = crashk_res.start;
	35	+
	36	+ loc = (unsigned long *)(data->kernel_buf + OLDMEM_SIZE_OFFSET);
	37	+ *loc = crashk_res.end - crashk_res.start + 1;
	38	+ }
	39	+
	40	+ if (image->initrd_buf) {
	41	+ loc = (unsigned long *)(data->kernel_buf + INITRD_START_OFFSET);
	42	+ *loc = data->initrd_load_addr;
	43	+
	44	+ loc = (unsigned long *)(data->kernel_buf + INITRD_SIZE_OFFSET);
	45	+ *loc = image->initrd_buf_len;
	46	+ }
	47	+
	48	+ return NULL;
	49	+}
	50	+
	51	+static int kexec_file_update_purgatory(struct kimage *image)
	52	+{
	53	+ u64 entry, type;
	54	+ int ret;
	55	+
	56	+ if (image->type == KEXEC_TYPE_CRASH) {
	57	+ entry = STARTUP_KDUMP_OFFSET;
	58	+ type = KEXEC_TYPE_CRASH;
	59	+ } else {
	60	+ entry = STARTUP_NORMAL_OFFSET;
	61	+ type = KEXEC_TYPE_DEFAULT;
	62	+ }
	63	+
	64	+ ret = kexec_purgatory_get_set_symbol(image, "kernel_entry", &entry,
	65	+ sizeof(entry), false);
	66	+ if (ret)
	67	+ return ret;
	68	+
	69	+ ret = kexec_purgatory_get_set_symbol(image, "kernel_type", &type,
	70	+ sizeof(type), false);
	71	+ if (ret)
	72	+ return ret;
	73	+
	74	+ if (image->type == KEXEC_TYPE_CRASH) {
	75	+ u64 crash_size;
	76	+
	77	+ ret = kexec_purgatory_get_set_symbol(image, "crash_start",
	78	+ &crashk_res.start,
	79	+ sizeof(crashk_res.start),
	80	+ false);
	81	+ if (ret)
	82	+ return ret;
	83	+
	84	+ crash_size = crashk_res.end - crashk_res.start + 1;
	85	+ ret = kexec_purgatory_get_set_symbol(image, "crash_size",
	86	+ &crash_size,
	87	+ sizeof(crash_size),
	88	+ false);
	89	+ }
	90	+ return ret;
	91	+}
	92	+
	93	+int kexec_file_add_purgatory(struct kimage image, struct s390_load_data data)
	94	+{
	95	+ struct kexec_buf buf;
	96	+ int ret;
	97	+
	98	+ buf.image = image;
	99	+
	100	+ data->memsz = ALIGN(data->memsz, PAGE_SIZE);
	101	+ buf.mem = data->memsz;
	102	+ if (image->type == KEXEC_TYPE_CRASH)
	103	+ buf.mem += crashk_res.start;
	104	+
	105	+ ret = kexec_load_purgatory(image, &buf);
	106	+ if (ret)
	107	+ return ret;
	108	+
	109	+ ret = kexec_file_update_purgatory(image);
	110	+ return ret;
	111	+}
	112	+
	113	+int kexec_file_add_initrd(struct kimage image, struct s390_load_data data,
	114	+ char *initrd, unsigned long initrd_len)
	115	+{
	116	+ struct kexec_buf buf;
	117	+ int ret;
	118	+
	119	+ buf.image = image;
	120	+
	121	+ buf.buffer = initrd;
	122	+ buf.bufsz = initrd_len;
	123	+
	124	+ data->memsz = ALIGN(data->memsz, PAGE_SIZE);
	125	+ buf.mem = data->memsz;
	126	+ if (image->type == KEXEC_TYPE_CRASH)
	127	+ buf.mem += crashk_res.start;
	128	+ buf.memsz = buf.bufsz;
	129	+
	130	+ data->initrd_load_addr = buf.mem;
	131	+ data->memsz += buf.memsz;
	132	+
	133	+ ret = kexec_add_buffer(&buf);
	134	+ return ret;
	135	+}
	136	+
	137	+/*
	138	+ * The kernel is loaded to a fixed location. Turn off kexec_locate_mem_hole
	139	+ * and provide kbuf->mem by hand.
	140	+ */
	141	+int arch_kexec_walk_mem(struct kexec_buf *kbuf,
	142	+ int (func)(struct resource , void *))
	143	+{
	144	+ return 1;
	145	+}
	146	+
	147	+int arch_kexec_apply_relocations_add(struct purgatory_info *pi,
	148	+ Elf_Shdr *section,
	149	+ const Elf_Shdr *relsec,
	150	+ const Elf_Shdr *symtab)
	151	+{
	152	+ Elf_Rela *relas;
	153	+ int i;
	154	+
	155	+ relas = (void *)pi->ehdr + relsec->sh_offset;
	156	+
	157	+ for (i = 0; i < relsec->sh_size / sizeof(*relas); i++) {
	158	+ const Elf_Sym sym; / symbol to relocate */
	159	+ unsigned long addr; /* final location after relocation */
	160	+ unsigned long val; /* relocated symbol value */
	161	+ void loc; / tmp location to modify */
	162	+
	163	+ sym = (void *)pi->ehdr + symtab->sh_offset;
	164	+ sym += ELF64_R_SYM(relas[i].r_info);
	165	+
	166	+ if (sym->st_shndx == SHN_UNDEF)
	167	+ return -ENOEXEC;
	168	+
	169	+ if (sym->st_shndx == SHN_COMMON)
	170	+ return -ENOEXEC;
	171	+
	172	+ if (sym->st_shndx >= pi->ehdr->e_shnum &&
	173	+ sym->st_shndx != SHN_ABS)
	174	+ return -ENOEXEC;
	175	+
	176	+ loc = pi->purgatory_buf;
	177	+ loc += section->sh_offset;
	178	+ loc += relas[i].r_offset;
	179	+
	180	+ val = sym->st_value;
	181	+ if (sym->st_shndx != SHN_ABS)
	182	+ val += pi->sechdrs[sym->st_shndx].sh_addr;
	183	+ val += relas[i].r_addend;
	184	+
	185	+ addr = section->sh_addr + relas[i].r_offset;
	186	+
	187	+ switch (ELF64_R_TYPE(relas[i].r_info)) {
	188	+ case R_390_8: /* Direct 8 bit. */
	189	+ (u8 )loc = val;
	190	+ break;
	191	+ case R_390_12: /* Direct 12 bit. */
	192	+ (u16 )loc &= 0xf000;
	193	+ (u16 )loc \|= val & 0xfff;
	194	+ break;
	195	+ case R_390_16: /* Direct 16 bit. */
	196	+ (u16 )loc = val;
	197	+ break;
	198	+ case R_390_20: /* Direct 20 bit. */
	199	+ (u32 )loc &= 0xf00000ff;
	200	+ (u32 )loc \|= (val & 0xfff) << 16; /* DL */
	201	+ (u32 )loc \|= (val & 0xff000) >> 4; /* DH */
	202	+ break;
	203	+ case R_390_32: /* Direct 32 bit. */
	204	+ (u32 )loc = val;
	205	+ break;
	206	+ case R_390_64: /* Direct 64 bit. */
	207	+ (u64 )loc = val;
	208	+ break;
	209	+ case R_390_PC16: /* PC relative 16 bit. */
	210	+ (u16 )loc = (val - addr);
	211	+ break;
	212	+ case R_390_PC16DBL: /* PC relative 16 bit shifted by 1. */
	213	+ (u16 )loc = (val - addr) >> 1;
	214	+ break;
	215	+ case R_390_PC32DBL: /* PC relative 32 bit shifted by 1. */
	216	+ (u32 )loc = (val - addr) >> 1;
	217	+ break;
	218	+ case R_390_PC32: /* PC relative 32 bit. */
	219	+ (u32 )loc = (val - addr);
	220	+ break;
	221	+ case R_390_PC64: /* PC relative 64 bit. */
	222	+ (u64 )loc = (val - addr);
	223	+ break;
	224	+ default:
	225	+ break;
	226	+ }
	227	+ }
	228	+ return 0;
	229	+}
	230	+
	231	+int arch_kexec_kernel_image_probe(struct kimage image, void buf,
	232	+ unsigned long buf_len)
	233	+{
	234	+ /* A kernel must be at least large enough to contain head.S. During
	235	+ * load memory in head.S will be accessed, e.g. to register the next
	236	+ * command line. If the next kernel were smaller the current kernel
	237	+ * will panic at load.
	238	+ *
	239	+ * 0x11000 = sizeof(head.S)
	240	+ */
	241	+ if (buf_len < 0x11000)
	242	+ return -ENOEXEC;
	243	+
	244	+ return kexec_image_probe_default(image, buf, buf_len);
	245	+}

arch/s390/kernel/module.c+2−2

@@ -465,11 +465,11 @@ int module_finalize(const Elf_Ehdr *hdr, · 464 unmodified lines
465	465	apply_alternatives(aseg, aseg + s->sh_size);
466	466
467	467	if (IS_ENABLED(CONFIG_EXPOLINE) &&
468		− (!strcmp(".nospec_call_table", secname)))
	468	+ (!strncmp(".s390_indirect", secname, 14)))
469	469	nospec_revert(aseg, aseg + s->sh_size);
470	470
471	471	if (IS_ENABLED(CONFIG_EXPOLINE) &&
472		− (!strcmp(".nospec_return_table", secname)))
	472	+ (!strncmp(".s390_return", secname, 12)))
473	473	nospec_revert(aseg, aseg + s->sh_size);
474	474	}
475	475

arch/s390/kernel/nospec-branch.c+1−0

@@ -1,6 +1,7 @@
1	1	// SPDX-License-Identifier: GPL-2.0
2	2	#include <linux/module.h>
3	3	#include <linux/device.h>
	4	+#include <linux/cpu.h>
4	5	#include <asm/nospec-branch.h>
5	6
6	7	static int __init nobp_setup_early(char *str)

arch/s390/kernel/perf_cpum_cf_events.c+5−4

@@ -123,7 +123,7 @@ CPUMF_EVENT_ATTR(cf_zec12, L1I_OFFBOOK_L3_SOURCED_WRITES_IV, 0x00a1); · 122 unmodified lines
123	123	CPUMF_EVENT_ATTR(cf_zec12, TX_NC_TABORT, 0x00b1);
124	124	CPUMF_EVENT_ATTR(cf_zec12, TX_C_TABORT_NO_SPECIAL, 0x00b2);
125	125	CPUMF_EVENT_ATTR(cf_zec12, TX_C_TABORT_SPECIAL, 0x00b3);
126		−CPUMF_EVENT_ATTR(cf_z13, L1D_WRITES_RO_EXCL, 0x0080);
	126	+CPUMF_EVENT_ATTR(cf_z13, L1D_RO_EXCL_WRITES, 0x0080);
127	127	CPUMF_EVENT_ATTR(cf_z13, DTLB1_WRITES, 0x0081);
128	128	CPUMF_EVENT_ATTR(cf_z13, DTLB1_MISSES, 0x0082);
129	129	CPUMF_EVENT_ATTR(cf_z13, DTLB1_HPAGE_WRITES, 0x0083);
@@ -179,7 +179,7 @@ CPUMF_EVENT_ATTR(cf_z13, TX_C_TABORT_NO_SPECIAL, 0x00db); · 49 unmodified lines
179	179	CPUMF_EVENT_ATTR(cf_z13, TX_C_TABORT_SPECIAL, 0x00dc);
180	180	CPUMF_EVENT_ATTR(cf_z13, MT_DIAG_CYCLES_ONE_THR_ACTIVE, 0x01c0);
181	181	CPUMF_EVENT_ATTR(cf_z13, MT_DIAG_CYCLES_TWO_THR_ACTIVE, 0x01c1);
182		−CPUMF_EVENT_ATTR(cf_z14, L1D_WRITES_RO_EXCL, 0x0080);
	182	+CPUMF_EVENT_ATTR(cf_z14, L1D_RO_EXCL_WRITES, 0x0080);
183	183	CPUMF_EVENT_ATTR(cf_z14, DTLB2_WRITES, 0x0081);
184	184	CPUMF_EVENT_ATTR(cf_z14, DTLB2_MISSES, 0x0082);
185	185	CPUMF_EVENT_ATTR(cf_z14, DTLB2_HPAGE_WRITES, 0x0083);
@@ -371,7 +371,7 @@ static struct attribute *cpumcf_zec12_pmu_event_attr[] __initdata = { · 185 unmodified lines
371	371	};
372	372
373	373	static struct attribute *cpumcf_z13_pmu_event_attr[] __initdata = {
374		− CPUMF_EVENT_PTR(cf_z13, L1D_WRITES_RO_EXCL),
	374	+ CPUMF_EVENT_PTR(cf_z13, L1D_RO_EXCL_WRITES),
375	375	CPUMF_EVENT_PTR(cf_z13, DTLB1_WRITES),
376	376	CPUMF_EVENT_PTR(cf_z13, DTLB1_MISSES),
377	377	CPUMF_EVENT_PTR(cf_z13, DTLB1_HPAGE_WRITES),
@@ -431,7 +431,7 @@ static struct attribute *cpumcf_z13_pmu_event_attr[] __initdata = { · 53 unmodified lines
431	431	};
432	432
433	433	static struct attribute *cpumcf_z14_pmu_event_attr[] __initdata = {
434		− CPUMF_EVENT_PTR(cf_z14, L1D_WRITES_RO_EXCL),
	434	+ CPUMF_EVENT_PTR(cf_z14, L1D_RO_EXCL_WRITES),
435	435	CPUMF_EVENT_PTR(cf_z14, DTLB2_WRITES),
436	436	CPUMF_EVENT_PTR(cf_z14, DTLB2_MISSES),
437	437	CPUMF_EVENT_PTR(cf_z14, DTLB2_HPAGE_WRITES),
@@ -583,6 +583,7 @@ __init const struct attribute_group **cpumf_cf_event_group(void) · 145 unmodified lines
583	583	model = cpumcf_z13_pmu_event_attr;
584	584	break;
585	585	case 0x3906:
	586	+ case 0x3907:
586	587	model = cpumcf_z14_pmu_event_attr;
587	588	break;
588	589	default:

arch/s390/kernel/process.c+10−0

@@ -29,6 +29,7 @@ · 28 unmodified lines
29	29	#include <linux/random.h>
30	30	#include <linux/export.h>
31	31	#include <linux/init_task.h>
	32	+#include <asm/cpu_mf.h>
32	33	#include <asm/io.h>
33	34	#include <asm/processor.h>
34	35	#include <asm/vtimer.h>
@@ -48,6 +49,15 @@ void flush_thread(void) · 13 unmodified lines
48	49	{
49	50	}
50	51
	52	+void arch_setup_new_exec(void)
	53	+{
	54	+ if (S390_lowcore.current_pid != current->pid) {
	55	+ S390_lowcore.current_pid = current->pid;
	56	+ if (test_facility(40))
	57	+ lpp(&S390_lowcore.lpp);
	58	+ }
	59	+}
	60	+
51	61	void arch_release_task_struct(struct task_struct *tsk)
52	62	{
53	63	runtime_instr_release(tsk);

arch/s390/kernel/setup.c+1−0

@@ -821,6 +821,7 @@ static int __init setup_hwcaps(void) · 820 unmodified lines
821	821	strcpy(elf_platform, "z13");
822	822	break;
823	823	case 0x3906:
	824	+ case 0x3907:
824	825	strcpy(elf_platform, "z14");
825	826	break;
826	827	}

arch/s390/kernel/syscalls/syscall.tbl+1−0

@@ -388,3 +388,4 @@ · 387 unmodified lines
388	388	378 common s390_guarded_storage sys_s390_guarded_storage compat_sys_s390_guarded_storage
389	389	379 common statx sys_statx compat_sys_statx
390	390	380 common s390_sthyi sys_s390_sthyi compat_sys_s390_sthyi
	391	+381 common kexec_file_load sys_kexec_file_load compat_sys_kexec_file_load

arch/s390/kernel/uprobes.c+9−0

@@ -150,6 +150,15 @@ unsigned long arch_uretprobe_hijack_return_addr(unsigned long trampoline, · 149 unmodified lines
150	150	return orig;
151	151	}
152	152
	153	+bool arch_uretprobe_is_alive(struct return_instance *ret, enum rp_check ctx,
	154	+ struct pt_regs *regs)
	155	+{
	156	+ if (ctx == RP_CHECK_CHAIN_CALL)
	157	+ return user_stack_pointer(regs) <= ret->stack;
	158	+ else
	159	+ return user_stack_pointer(regs) < ret->stack;
	160	+}
	161	+
153	162	/* Instruction Emulation */
154	163
155	164	static void adjust_psw_addr(psw_t *psw, unsigned long len)

arch/s390/purgatory/.gitignore+2−0

@@ -0,0 +1,2 @@
	1	+kexec-purgatory.c
	2	+purgatory.ro

arch/s390/purgatory/Makefile+37−0

@@ -0,0 +1,37 @@
	1	+# SPDX-License-Identifier: GPL-2.0
	2	+
	3	+OBJECT_FILES_NON_STANDARD := y
	4	+
	5	+purgatory-y := head.o purgatory.o string.o sha256.o mem.o
	6	+
	7	+targets += $(purgatory-y) purgatory.ro kexec-purgatory.c
	8	+PURGATORY_OBJS = $(addprefix $(obj)/,$(purgatory-y))
	9	+
	10	+$(obj)/sha256.o: $(srctree)/lib/sha256.c
	11	+ $(call if_changed_rule,cc_o_c)
	12	+
	13	+$(obj)/mem.o: $(srctree)/arch/s390/lib/mem.S
	14	+ $(call if_changed_rule,as_o_S)
	15	+
	16	+$(obj)/string.o: $(srctree)/arch/s390/lib/string.c
	17	+ $(call if_changed_rule,cc_o_c)
	18	+
	19	+LDFLAGS_purgatory.ro := -e purgatory_start -r --no-undefined -nostdlib
	20	+LDFLAGS_purgatory.ro += -z nodefaultlib
	21	+KBUILD_CFLAGS := -fno-strict-aliasing -Wall -Wstrict-prototypes
	22	+KBUILD_CFLAGS += -Wno-pointer-sign -Wno-sign-compare
	23	+KBUILD_CFLAGS += -fno-zero-initialized-in-bss -fno-builtin -ffreestanding
	24	+KBUILD_CFLAGS += -c -MD -Os -m64
	25	+KBUILD_CFLAGS += $(call cc-option,-fno-PIE)
	26	+
	27	+$(obj)/purgatory.ro: $(PURGATORY_OBJS) FORCE
	28	+ $(call if_changed,ld)
	29	+
	30	+CMD_BIN2C = $(objtree)/scripts/basic/bin2c
	31	+quiet_cmd_bin2c = BIN2C $@
	32	+ cmd_bin2c = $(CMD_BIN2C) kexec_purgatory < $< > $@
	33	+
	34	+$(obj)/kexec-purgatory.c: $(obj)/purgatory.ro FORCE
	35	+ $(call if_changed,bin2c)
	36	+
	37	+obj-$(CONFIG_ARCH_HAS_KEXEC_PURGATORY) += kexec-purgatory.o

arch/s390/purgatory/head.S+279−0

@@ -0,0 +1,279 @@
	1	+/* SPDX-License-Identifier: GPL-2.0 */
	2	+/*
	3	+ * Purgatory setup code
	4	+ *
	5	+ * Copyright IBM Corp. 2018
	6	+ *
	7	+ * Author(s): Philipp Rudo <prudo@linux.vnet.ibm.com>
	8	+ */
	9	+
	10	+#include <linux/linkage.h>
	11	+#include <asm/asm-offsets.h>
	12	+#include <asm/page.h>
	13	+#include <asm/sigp.h>
	14	+
	15	+/* The purgatory is the code running between two kernels. It's main purpose
	16	+ * is to verify that the next kernel was not corrupted after load and to
	17	+ * start it.
	18	+ *
	19	+ * If the next kernel is a crash kernel there are some peculiarities to
	20	+ * consider:
	21	+ *
	22	+ * First the purgatory is called twice. Once only to verify the
	23	+ * sha digest. So if the crash kernel got corrupted the old kernel can try
	24	+ * to trigger a stand-alone dumper. And once to actually load the crash kernel.
	25	+ *
	26	+ * Second the purgatory also has to swap the crash memory region with its
	27	+ * destination at address 0. As the purgatory is part of crash memory this
	28	+ * requires some finesse. The tactic here is that the purgatory first copies
	29	+ * itself to the end of the destination and then swaps the rest of the
	30	+ * memory running from there.
	31	+ */
	32	+
	33	+#define bufsz purgatory_end-stack
	34	+
	35	+.macro MEMCPY dst,src,len
	36	+ lgr %r0,\dst
	37	+ lgr %r1,\len
	38	+ lgr %r2,\src
	39	+ lgr %r3,\len
	40	+
	41	+20: mvcle %r0,%r2,0
	42	+ jo 20b
	43	+.endm
	44	+
	45	+.macro MEMSWAP dst,src,buf,len
	46	+10: cghi \len,bufsz
	47	+ jh 11f
	48	+ lgr %r4,\len
	49	+ j 12f
	50	+11: lghi %r4,bufsz
	51	+
	52	+12: MEMCPY \buf,\dst,%r4
	53	+ MEMCPY \dst,\src,%r4
	54	+ MEMCPY \src,\buf,%r4
	55	+
	56	+ agr \dst,%r4
	57	+ agr \src,%r4
	58	+ sgr \len,%r4
	59	+
	60	+ cghi \len,0
	61	+ jh 10b
	62	+.endm
	63	+
	64	+.macro START_NEXT_KERNEL base
	65	+ lg %r4,kernel_entry-\base(%r13)
	66	+ lg %r5,load_psw_mask-\base(%r13)
	67	+ ogr %r4,%r5
	68	+ stg %r4,0(%r0)
	69	+
	70	+ xgr %r0,%r0
	71	+ diag %r0,%r0,0x308
	72	+.endm
	73	+
	74	+.text
	75	+.align PAGE_SIZE
	76	+ENTRY(purgatory_start)
	77	+ /* The purgatory might be called after a diag308 so better set
	78	+ * architecture and addressing mode.
	79	+ */
	80	+ lhi %r1,1
	81	+ sigp %r1,%r0,SIGP_SET_ARCHITECTURE
	82	+ sam64
	83	+
	84	+ larl %r5,gprregs
	85	+ stmg %r6,%r15,0(%r5)
	86	+
	87	+ basr %r13,0
	88	+.base_crash:
	89	+
	90	+ /* Setup stack */
	91	+ larl %r15,purgatory_end
	92	+ aghi %r15,-160
	93	+
	94	+ /* If the next kernel is KEXEC_TYPE_CRASH the purgatory is called
	95	+ * directly with a flag passed in %r2 whether the purgatory shall do
	96	+ * checksum verification only (%r2 = 0 -> verification only).
	97	+ *
	98	+ * Check now and preserve over C function call by storing in
	99	+ * %r10 whith
	100	+ * 1 -> checksum verification only
	101	+ * 0 -> load new kernel
	102	+ */
	103	+ lghi %r10,0
	104	+ lg %r11,kernel_type-.base_crash(%r13)
	105	+ cghi %r11,1 /* KEXEC_TYPE_CRASH */
	106	+ jne .do_checksum_verification
	107	+ cghi %r2,0 /* checksum verification only */
	108	+ jne .do_checksum_verification
	109	+ lghi %r10,1
	110	+
	111	+.do_checksum_verification:
	112	+ brasl %r14,verify_sha256_digest
	113	+
	114	+ cghi %r10,1 /* checksum verification only */
	115	+ je .return_old_kernel
	116	+ cghi %r2,0 /* checksum match */
	117	+ jne .disabled_wait
	118	+
	119	+ /* If the next kernel is a crash kernel the purgatory has to swap
	120	+ * the mem regions first.
	121	+ */
	122	+ cghi %r11,1 /* KEXEC_TYPE_CRASH */
	123	+ je .start_crash_kernel
	124	+
	125	+ /* start normal kernel */
	126	+ START_NEXT_KERNEL .base_crash
	127	+
	128	+.return_old_kernel:
	129	+ lmg %r6,%r15,gprregs-.base_crash(%r13)
	130	+ br %r14
	131	+
	132	+.disabled_wait:
	133	+ lpswe disabled_wait_psw-.base_crash(%r13)
	134	+
	135	+.start_crash_kernel:
	136	+ /* Location of purgatory_start in crash memory */
	137	+ lgr %r8,%r13
	138	+ aghi %r8,-(.base_crash-purgatory_start)
	139	+
	140	+ /* Destination for this code i.e. end of memory to be swapped. */
	141	+ lg %r9,crash_size-.base_crash(%r13)
	142	+ aghi %r9,-(purgatory_end-purgatory_start)
	143	+
	144	+ /* Destination in crash memory, i.e. same as r9 but in crash memory. */
	145	+ lg %r10,crash_start-.base_crash(%r13)
	146	+ agr %r10,%r9
	147	+
	148	+ /* Buffer location (in crash memory) and size. As the purgatory is
	149	+ * behind the point of no return it can re-use the stack as buffer.
	150	+ */
	151	+ lghi %r11,bufsz
	152	+ larl %r12,stack
	153	+
	154	+ MEMCPY %r12,%r9,%r11 /* dst -> (crash) buf */
	155	+ MEMCPY %r9,%r8,%r11 /* self -> dst */
	156	+
	157	+ /* Jump to new location. */
	158	+ lgr %r7,%r9
	159	+ aghi %r7,.jump_to_dst-purgatory_start
	160	+ br %r7
	161	+
	162	+.jump_to_dst:
	163	+ basr %r13,0
	164	+.base_dst:
	165	+
	166	+ /* clear buffer */
	167	+ MEMCPY %r12,%r10,%r11 /* (crash) buf -> (crash) dst */
	168	+
	169	+ /* Load new buffer location after jump */
	170	+ larl %r7,stack
	171	+ aghi %r10,stack-purgatory_start
	172	+ MEMCPY %r10,%r7,%r11 /* (new) buf -> (crash) buf */
	173	+
	174	+ /* Now the code is set up to run from its designated location. Start
	175	+ * swapping the rest of crash memory now.
	176	+ *
	177	+ * The registers will be used as follow:
	178	+ *
	179	+ * %r0-%r4 reserved for macros defined above
	180	+ * %r5-%r6 tmp registers
	181	+ * %r7 pointer to current struct sha region
	182	+ * %r8 index to iterate over all sha regions
	183	+ * %r9 pointer in crash memory
	184	+ * %r10 pointer in old kernel
	185	+ * %r11 total size (still) to be moved
	186	+ * %r12 pointer to buffer
	187	+ */
	188	+ lgr %r12,%r7
	189	+ lgr %r11,%r9
	190	+ lghi %r10,0
	191	+ lg %r9,crash_start-.base_dst(%r13)
	192	+ lghi %r8,16 /* KEXEC_SEGMENTS_MAX */
	193	+ larl %r7,purgatory_sha_regions
	194	+
	195	+ j .loop_first
	196	+
	197	+ /* Loop over all purgatory_sha_regions. */
	198	+.loop_next:
	199	+ aghi %r8,-1
	200	+ cghi %r8,0
	201	+ je .loop_out
	202	+
	203	+ aghi %r7,__KEXEC_SHA_REGION_SIZE
	204	+
	205	+.loop_first:
	206	+ lg %r5,__KEXEC_SHA_REGION_START(%r7)
	207	+ cghi %r5,0
	208	+ je .loop_next
	209	+
	210	+ /* Copy [end last sha region, start current sha region) */
	211	+ /* Note: kexec_sha_region->start points in crash memory */
	212	+ sgr %r5,%r9
	213	+ MEMCPY %r9,%r10,%r5
	214	+
	215	+ agr %r9,%r5
	216	+ agr %r10,%r5
	217	+ sgr %r11,%r5
	218	+
	219	+ /* Swap sha region */
	220	+ lg %r6,__KEXEC_SHA_REGION_LEN(%r7)
	221	+ MEMSWAP %r9,%r10,%r12,%r6
	222	+ sg %r11,__KEXEC_SHA_REGION_LEN(%r7)
	223	+ j .loop_next
	224	+
	225	+.loop_out:
	226	+ /* Copy rest of crash memory */
	227	+ MEMCPY %r9,%r10,%r11
	228	+
	229	+ /* start crash kernel */
	230	+ START_NEXT_KERNEL .base_dst
	231	+
	232	+
	233	+load_psw_mask:
	234	+ .long 0x00080000,0x80000000
	235	+
	236	+ .align 8
	237	+disabled_wait_psw:
	238	+ .quad 0x0002000180000000
	239	+ .quad 0x0000000000000000 + .do_checksum_verification
	240	+
	241	+gprregs:
	242	+ .rept 10
	243	+ .quad 0
	244	+ .endr
	245	+
	246	+purgatory_sha256_digest:
	247	+ .global purgatory_sha256_digest
	248	+ .rept 32 /* SHA256_DIGEST_SIZE */
	249	+ .byte 0
	250	+ .endr
	251	+
	252	+purgatory_sha_regions:
	253	+ .global purgatory_sha_regions
	254	+ .rept 16 * __KEXEC_SHA_REGION_SIZE /* KEXEC_SEGMENTS_MAX */
	255	+ .byte 0
	256	+ .endr
	257	+
	258	+kernel_entry:
	259	+ .global kernel_entry
	260	+ .quad 0
	261	+
	262	+kernel_type:
	263	+ .global kernel_type
	264	+ .quad 0
	265	+
	266	+crash_start:
	267	+ .global crash_start
	268	+ .quad 0
	269	+
	270	+crash_size:
	271	+ .global crash_size
	272	+ .quad 0
	273	+
	274	+ .align PAGE_SIZE
	275	+stack:
	276	+ /* The buffer to move this code must be as big as the code. */
	277	+ .skip stack-purgatory_start
	278	+ .align PAGE_SIZE
	279	+purgatory_end:

arch/s390/purgatory/purgatory.c+42−0

@@ -0,0 +1,42 @@
	1	+// SPDX-License-Identifier: GPL-2.0
	2	+/*
	3	+ * Purgatory code running between two kernels.
	4	+ *
	5	+ * Copyright IBM Corp. 2018
	6	+ *
	7	+ * Author(s): Philipp Rudo <prudo@linux.vnet.ibm.com>
	8	+ */
	9	+
	10	+#include <linux/kexec.h>
	11	+#include <linux/sha256.h>
	12	+#include <linux/string.h>
	13	+#include <asm/purgatory.h>
	14	+
	15	+struct kexec_sha_region purgatory_sha_regions[KEXEC_SEGMENT_MAX];
	16	+u8 purgatory_sha256_digest[SHA256_DIGEST_SIZE];
	17	+
	18	+u64 kernel_entry;
	19	+u64 kernel_type;
	20	+
	21	+u64 crash_start;
	22	+u64 crash_size;
	23	+
	24	+int verify_sha256_digest(void)
	25	+{
	26	+ struct kexec_sha_region ptr, end;
	27	+ u8 digest[SHA256_DIGEST_SIZE];
	28	+ struct sha256_state sctx;
	29	+
	30	+ sha256_init(&sctx);
	31	+ end = purgatory_sha_regions + ARRAY_SIZE(purgatory_sha_regions);
	32	+
	33	+ for (ptr = purgatory_sha_regions; ptr < end; ptr++)
	34	+ sha256_update(&sctx, (uint8_t *)(ptr->start), ptr->len);
	35	+
	36	+ sha256_final(&sctx, digest);
	37	+
	38	+ if (memcmp(digest, purgatory_sha256_digest, sizeof(digest)))
	39	+ return 1;
	40	+
	41	+ return 0;
	42	+}

arch/sparc/include/uapi/asm/oradax.h+1−1

@@ -3,7 +3,7 @@ · 2 unmodified lines
3	3	*
4	4	* This program is free software: you can redistribute it and/or modify
5	5	* it under the terms of the GNU General Public License as published by
6		− * the Free Software Foundation, either version 3 of the License, or
	6	+ * the Free Software Foundation, either version 2 of the License, or
7	7	* (at your option) any later version.
8	8	*
9	9	* This program is distributed in the hope that it will be useful,

arch/sparc/kernel/vio.c+1−1

@@ -403,7 +403,7 @@ static struct vio_dev vio_create_one(struct mdesc_handle hp, u64 mp, · 402 unmodified lines
403	403	if (err) {
404	404	printk(KERN_ERR "VIO: Could not register device %s, err=%d\n",
405	405	dev_name(&vdev->dev), err);
406		− kfree(vdev);
	406	+ put_device(&vdev->dev);
407	407	return NULL;
408	408	}
409	409	if (vdev->dp)

arch/x86/Kconfig+4−0

@@ -52,6 +52,7 @@ config X86 · 51 unmodified lines
52	52	select ARCH_HAS_DEVMEM_IS_ALLOWED
53	53	select ARCH_HAS_ELF_RANDOMIZE
54	54	select ARCH_HAS_FAST_MULTIPLIER
	55	+ select ARCH_HAS_FILTER_PGPROT
55	56	select ARCH_HAS_FORTIFY_SOURCE
56	57	select ARCH_HAS_GCOV_PROFILE_ALL
57	58	select ARCH_HAS_KCOV if X86_64
@@ -273,6 +274,9 @@ config ARCH_HAS_CPU_RELAX · 215 unmodified lines
273	274	config ARCH_HAS_CACHE_LINE_SIZE
274	275	def_bool y
275	276
	277	+config ARCH_HAS_FILTER_PGPROT
	278	+ def_bool y
	279	+
276	280	config HAVE_SETUP_PER_CPU_AREA
277	281	def_bool y
278	282

arch/x86/entry/entry_64_compat.S+4−4

@@ -84,13 +84,13 @@ ENTRY(entry_SYSENTER_compat) · 83 unmodified lines
84	84	pushq %rdx /* pt_regs->dx */
85	85	pushq %rcx /* pt_regs->cx */
86	86	pushq $-ENOSYS /* pt_regs->ax */
87		− pushq $0 /* pt_regs->r8 = 0 */
	87	+ pushq %r8 /* pt_regs->r8 */
88	88	xorl %r8d, %r8d /* nospec r8 */
89		− pushq $0 /* pt_regs->r9 = 0 */
	89	+ pushq %r9 /* pt_regs->r9 */
90	90	xorl %r9d, %r9d /* nospec r9 */
91		− pushq $0 /* pt_regs->r10 = 0 */
	91	+ pushq %r10 /* pt_regs->r10 */
92	92	xorl %r10d, %r10d /* nospec r10 */
93		− pushq $0 /* pt_regs->r11 = 0 */
	93	+ pushq %r11 /* pt_regs->r11 */
94	94	xorl %r11d, %r11d /* nospec r11 */
95	95	pushq %rbx /* pt_regs->rbx */
96	96	xorl %ebx, %ebx /* nospec rbx */

arch/x86/events/intel/core.c+6−3

@@ -3339,7 +3339,8 @@ static void intel_pmu_cpu_starting(int cpu) · 3338 unmodified lines
3339	3339
3340	3340	cpuc->lbr_sel = NULL;
3341	3341
3342		− flip_smm_bit(&x86_pmu.attr_freeze_on_smi);
	3342	+ if (x86_pmu.version > 1)
	3343	+ flip_smm_bit(&x86_pmu.attr_freeze_on_smi);
3343	3344
3344	3345	if (!cpuc->shared_regs)
3345	3346	return;
@@ -3502,6 +3503,8 @@ static __initconst const struct x86_pmu core_pmu = { · 156 unmodified lines
3502	3503	.cpu_dying = intel_pmu_cpu_dying,
3503	3504	};
3504	3505
	3506	+static struct attribute *intel_pmu_attrs[];
	3507	+
3505	3508	static __initconst const struct x86_pmu intel_pmu = {
3506	3509	.name = "Intel",
3507	3510	.handle_irq = intel_pmu_handle_irq,
@@ -3533,6 +3536,8 @@ static __initconst const struct x86_pmu intel_pmu = { · 25 unmodified lines
3533	3536	.format_attrs = intel_arch3_formats_attr,
3534	3537	.events_sysfs_show = intel_event_sysfs_show,
3535	3538
	3539	+ .attrs = intel_pmu_attrs,
	3540	+
3536	3541	.cpu_prepare = intel_pmu_cpu_prepare,
3537	3542	.cpu_starting = intel_pmu_cpu_starting,
3538	3543	.cpu_dying = intel_pmu_cpu_dying,
@@ -3911,8 +3916,6 @@ __init int intel_pmu_init(void) · 372 unmodified lines
3911	3916
3912	3917	x86_pmu.max_pebs_events = min_t(unsigned, MAX_PEBS_EVENTS, x86_pmu.num_counters);
3913	3918
3914		−
3915		− x86_pmu.attrs = intel_pmu_attrs;
3916	3919	/*
3917	3920	* Quirk: v2 perfmon does not report fixed-purpose events, so
3918	3921	* assume at least 3 events, when not running in a hypervisor:

arch/x86/events/intel/uncore_snbep.c+37−0

@@ -3028,10 +3028,27 @@ static struct intel_uncore_type bdx_uncore_cbox = { · 3027 unmodified lines
3028	3028	.format_group = &hswep_uncore_cbox_format_group,
3029	3029	};
3030	3030
	3031	+static struct intel_uncore_type bdx_uncore_sbox = {
	3032	+ .name = "sbox",
	3033	+ .num_counters = 4,
	3034	+ .num_boxes = 4,
	3035	+ .perf_ctr_bits = 48,
	3036	+ .event_ctl = HSWEP_S0_MSR_PMON_CTL0,
	3037	+ .perf_ctr = HSWEP_S0_MSR_PMON_CTR0,
	3038	+ .event_mask = HSWEP_S_MSR_PMON_RAW_EVENT_MASK,
	3039	+ .box_ctl = HSWEP_S0_MSR_PMON_BOX_CTL,
	3040	+ .msr_offset = HSWEP_SBOX_MSR_OFFSET,
	3041	+ .ops = &hswep_uncore_sbox_msr_ops,
	3042	+ .format_group = &hswep_uncore_sbox_format_group,
	3043	+};
	3044	+
	3045	+#define BDX_MSR_UNCORE_SBOX 3
	3046	+
3031	3047	static struct intel_uncore_type *bdx_msr_uncores[] = {
3032	3048	&bdx_uncore_ubox,
3033	3049	&bdx_uncore_cbox,
3034	3050	&hswep_uncore_pcu,
	3051	+ &bdx_uncore_sbox,
3035	3052	NULL,
3036	3053	};
3037	3054
@@ -3043,10 +3060,25 @@ static struct event_constraint bdx_uncore_pcu_constraints[] = { · 5 unmodified lines
3043	3060
3044	3061	void bdx_uncore_cpu_init(void)
3045	3062	{
	3063	+ int pkg = topology_phys_to_logical_pkg(0);
	3064	+
3046	3065	if (bdx_uncore_cbox.num_boxes > boot_cpu_data.x86_max_cores)
3047	3066	bdx_uncore_cbox.num_boxes = boot_cpu_data.x86_max_cores;
3048	3067	uncore_msr_uncores = bdx_msr_uncores;
3049	3068
	3069	+ /* BDX-DE doesn't have SBOX */
	3070	+ if (boot_cpu_data.x86_model == 86) {
	3071	+ uncore_msr_uncores[BDX_MSR_UNCORE_SBOX] = NULL;
	3072	+ /* Detect systems with no SBOXes */
	3073	+ } else if (uncore_extra_pci_dev[pkg].dev[HSWEP_PCI_PCU_3]) {
	3074	+ struct pci_dev *pdev;
	3075	+ u32 capid4;
	3076	+
	3077	+ pdev = uncore_extra_pci_dev[pkg].dev[HSWEP_PCI_PCU_3];
	3078	+ pci_read_config_dword(pdev, 0x94, &capid4);
	3079	+ if (((capid4 >> 6) & 0x3) == 0)
	3080	+ bdx_msr_uncores[BDX_MSR_UNCORE_SBOX] = NULL;
	3081	+ }
3050	3082	hswep_uncore_pcu.constraints = bdx_uncore_pcu_constraints;
3051	3083	}
3052	3084
@@ -3264,6 +3296,11 @@ static const struct pci_device_id bdx_uncore_pci_ids[] = { · 211 unmodified lines
3264	3296	PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x6f46),
3265	3297	.driver_data = UNCORE_PCI_DEV_DATA(UNCORE_EXTRA_PCI_DEV, 2),
3266	3298	},
	3299	+ { /* PCU.3 (for Capability registers) */
	3300	+ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x6fc0),
	3301	+ .driver_data = UNCORE_PCI_DEV_DATA(UNCORE_EXTRA_PCI_DEV,
	3302	+ HSWEP_PCI_PCU_3),
	3303	+ },
3267	3304	{ /* end: all zeroes */ }
3268	3305	};
3269	3306

arch/x86/include/asm/asm.h+0−2

@@ -136,7 +136,6 @@ · 135 unmodified lines
136	136	#endif
137	137
138	138	#ifndef __ASSEMBLY__
139		−#ifndef __BPF__
140	139	/*
141	140	* This output constraint should be used for any inline asm which has a "call"
142	141	* instruction. Otherwise the asm may be inserted before the frame pointer
@@ -146,6 +145,5 @@ · 3 unmodified lines
146	145	register unsigned long current_stack_pointer asm(_ASM_SP);
147	146	#define ASM_CALL_CONSTRAINT "+r" (current_stack_pointer)
148	147	#endif
149		−#endif
150	148
151	149	#endif /* _ASM_X86_ASM_H */

arch/x86/include/asm/cpufeatures.h+1−0

@@ -320,6 +320,7 @@ · 319 unmodified lines
320	320	#define X86_FEATURE_AVX512_VPOPCNTDQ (1632+14) / POPCNT for vectors of DW/QW */
321	321	#define X86_FEATURE_LA57 (1632+16) / 5-level page tables */
322	322	#define X86_FEATURE_RDPID (1632+22) / RDPID instruction */
	323	+#define X86_FEATURE_CLDEMOTE (1632+25) / CLDEMOTE instruction */
323	324
324	325	/* AMD-defined CPU features, CPUID level 0x80000007 (EBX), word 17 */
325	326	#define X86_FEATURE_OVERFLOW_RECOV (1732+ 0) / MCA overflow recovery support */

arch/x86/include/asm/ftrace.h+17−2

@@ -46,7 +46,21 @@ int ftrace_int3_handler(struct pt_regs *regs); · 45 unmodified lines
46	46	#endif /* CONFIG_FUNCTION_TRACER */
47	47
48	48
49		−#if !defined(__ASSEMBLY__) && !defined(COMPILE_OFFSETS)
	49	+#ifndef __ASSEMBLY__
	50	+
	51	+#define ARCH_HAS_SYSCALL_MATCH_SYM_NAME
	52	+static inline bool arch_syscall_match_sym_name(const char sym, const char name)
	53	+{
	54	+ /*
	55	+ * Compare the symbol name with the system call name. Skip the
	56	+ * "__x64_sys", "__ia32_sys" or simple "sys" prefix.
	57	+ */
	58	+ return !strcmp(sym + 3, name + 3) \|\|
	59	+ (!strncmp(sym, "__x64_", 6) && !strcmp(sym + 9, name + 3)) \|\|
	60	+ (!strncmp(sym, "__ia32_", 7) && !strcmp(sym + 10, name + 3));
	61	+}
	62	+
	63	+#ifndef COMPILE_OFFSETS
50	64
51	65	#if defined(CONFIG_FTRACE_SYSCALLS) && defined(CONFIG_IA32_EMULATION)
52	66	#include <asm/compat.h>
@@ -67,6 +81,7 @@ static inline bool arch_trace_is_compat_syscall(struct pt_regs *regs) · 14 unmodified lines
67	81	return false;
68	82	}
69	83	#endif /* CONFIG_FTRACE_SYSCALLS && CONFIG_IA32_EMULATION */
70		−#endif /* !__ASSEMBLY__ && !COMPILE_OFFSETS */
	84	+#endif /* !COMPILE_OFFSETS */
	85	+#endif /* !__ASSEMBLY__ */
71	86
72	87	#endif /* _ASM_X86_FTRACE_H */

arch/x86/include/asm/irq_vectors.h+0−7

@@ -34,11 +34,6 @@ · 33 unmodified lines
34	34	* (0x80 is the syscall vector, 0x30-0x3f are for ISA)
35	35	*/
36	36	#define FIRST_EXTERNAL_VECTOR 0x20
37		−/*
38		− * We start allocating at 0x21 to spread out vectors evenly between
39		− * priority levels. (0x80 is the syscall vector)
40		− */
41		−#define VECTOR_OFFSET_START 1
42	37
43	38	/*
44	39	* Reserve the lowest usable vector (and hence lowest priority) 0x20 for
@@ -119,8 +114,6 @@ · 74 unmodified lines
119	114	#define FIRST_SYSTEM_VECTOR NR_VECTORS
120	115	#endif
121	116
122		−#define FPU_IRQ 13
123		−
124	117	/*
125	118	* Size the maximum number of interrupts.
126	119	*

arch/x86/include/asm/jailhouse_para.h+1−1

@@ -1,4 +1,4 @@
1		−/* SPDX-License-Identifier: GPL2.0 */
	1	+/* SPDX-License-Identifier: GPL-2.0 */
2	2
3	3	/*
4	4	* Jailhouse paravirt detection

arch/x86/include/asm/kvm_host.h+1−0

@@ -1013,6 +1013,7 @@ struct kvm_x86_ops { · 1012 unmodified lines
1013	1013
1014	1014	bool (*has_wbinvd_exit)(void);
1015	1015
	1016	+ u64 (read_l1_tsc_offset)(struct kvm_vcpu vcpu);
1016	1017	void (write_tsc_offset)(struct kvm_vcpu vcpu, u64 offset);
1017	1018
1018	1019	void (get_exit_info)(struct kvm_vcpu vcpu, u64 info1, u64 info2);

arch/x86/include/asm/pgtable.h+5−0

@@ -601,6 +601,11 @@ static inline pgprot_t pgprot_modify(pgprot_t oldprot, pgprot_t newprot) · 600 unmodified lines
601	601
602	602	#define canon_pgprot(p) __pgprot(massage_pgprot(p))
603	603
	604	+static inline pgprot_t arch_filter_pgprot(pgprot_t prot)
	605	+{
	606	+ return canon_pgprot(prot);
	607	+}
	608	+
604	609	static inline int is_new_memtype_allowed(u64 paddr, unsigned long size,
605	610	enum page_cache_mode pcm,
606	611	enum page_cache_mode new_pcm)

arch/x86/include/asm/pgtable_64_types.h+4−4

@@ -105,14 +105,14 @@ extern unsigned int ptrs_per_p4d; · 104 unmodified lines
105	105	#define LDT_PGD_ENTRY (pgtable_l5_enabled ? LDT_PGD_ENTRY_L5 : LDT_PGD_ENTRY_L4)
106	106	#define LDT_BASE_ADDR (LDT_PGD_ENTRY << PGDIR_SHIFT)
107	107
108		−#define __VMALLOC_BASE_L4 0xffffc90000000000
109		−#define __VMALLOC_BASE_L5 0xffa0000000000000
	108	+#define __VMALLOC_BASE_L4 0xffffc90000000000UL
	109	+#define __VMALLOC_BASE_L5 0xffa0000000000000UL
110	110
111	111	#define VMALLOC_SIZE_TB_L4 32UL
112	112	#define VMALLOC_SIZE_TB_L5 12800UL
113	113
114		−#define __VMEMMAP_BASE_L4 0xffffea0000000000
115		−#define __VMEMMAP_BASE_L5 0xffd4000000000000
	114	+#define __VMEMMAP_BASE_L4 0xffffea0000000000UL
	115	+#define __VMEMMAP_BASE_L5 0xffd4000000000000UL
116	116
117	117	#ifdef CONFIG_DYNAMIC_MEMORY_LAYOUT
118	118	# define VMALLOC_START vmalloc_base

arch/x86/include/asm/processor.h+0−2

@@ -749,13 +749,11 @@ enum idle_boot_override {IDLE_NO_OVERRIDE=0, IDLE_HALT, IDLE_NOMWAIT, · 748 unmodified lines
749	749	extern void enable_sep_cpu(void);
750	750	extern int sysenter_setup(void);
751	751
752		−extern void early_trap_init(void);
753	752	void early_trap_pf_init(void);
754	753
755	754	/* Defined in head.S */
756	755	extern struct desc_ptr early_gdt_descr;
757	756
758		−extern void cpu_set_gdt(int);
759	757	extern void switch_to_new_gdt(int);
760	758	extern void load_direct_gdt(int);
761	759	extern void load_fixmap_gdt(int);

arch/x86/include/uapi/asm/msgbuf.h+31−0

@@ -1 +1,32 @@
	1	+/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
	2	+#ifndef __ASM_X64_MSGBUF_H
	3	+#define __ASM_X64_MSGBUF_H
	4	+
	5	+#if !defined(__x86_64__) \|\| !defined(__ILP32__)
1	6	#include <asm-generic/msgbuf.h>
	7	+#else
	8	+/*
	9	+ * The msqid64_ds structure for x86 architecture with x32 ABI.
	10	+ *
	11	+ * On x86-32 and x86-64 we can just use the generic definition, but
	12	+ * x32 uses the same binary layout as x86_64, which is differnet
	13	+ * from other 32-bit architectures.
	14	+ */
	15	+
	16	+struct msqid64_ds {
	17	+ struct ipc64_perm msg_perm;
	18	+ __kernel_time_t msg_stime; /* last msgsnd time */
	19	+ __kernel_time_t msg_rtime; /* last msgrcv time */
	20	+ __kernel_time_t msg_ctime; /* last change time */
	21	+ __kernel_ulong_t msg_cbytes; /* current number of bytes on queue */
	22	+ __kernel_ulong_t msg_qnum; /* number of messages in queue */
	23	+ __kernel_ulong_t msg_qbytes; /* max number of bytes on queue */
	24	+ __kernel_pid_t msg_lspid; /* pid of last msgsnd */
	25	+ __kernel_pid_t msg_lrpid; /* last receive pid */
	26	+ __kernel_ulong_t __unused4;
	27	+ __kernel_ulong_t __unused5;
	28	+};
	29	+
	30	+#endif
	31	+
	32	+#endif /* __ASM_GENERIC_MSGBUF_H */

arch/x86/include/uapi/asm/shmbuf.h+42−0

@@ -1 +1,43 @@
	1	+/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
	2	+#ifndef __ASM_X86_SHMBUF_H
	3	+#define __ASM_X86_SHMBUF_H
	4	+
	5	+#if !defined(__x86_64__) \|\| !defined(__ILP32__)
1	6	#include <asm-generic/shmbuf.h>
	7	+#else
	8	+/*
	9	+ * The shmid64_ds structure for x86 architecture with x32 ABI.
	10	+ *
	11	+ * On x86-32 and x86-64 we can just use the generic definition, but
	12	+ * x32 uses the same binary layout as x86_64, which is differnet
	13	+ * from other 32-bit architectures.
	14	+ */
	15	+
	16	+struct shmid64_ds {
	17	+ struct ipc64_perm shm_perm; /* operation perms */
	18	+ size_t shm_segsz; /* size of segment (bytes) */
	19	+ __kernel_time_t shm_atime; /* last attach time */
	20	+ __kernel_time_t shm_dtime; /* last detach time */
	21	+ __kernel_time_t shm_ctime; /* last change time */
	22	+ __kernel_pid_t shm_cpid; /* pid of creator */
	23	+ __kernel_pid_t shm_lpid; /* pid of last operator */
	24	+ __kernel_ulong_t shm_nattch; /* no. of current attaches */
	25	+ __kernel_ulong_t __unused4;
	26	+ __kernel_ulong_t __unused5;
	27	+};
	28	+
	29	+struct shminfo64 {
	30	+ __kernel_ulong_t shmmax;
	31	+ __kernel_ulong_t shmmin;
	32	+ __kernel_ulong_t shmmni;
	33	+ __kernel_ulong_t shmseg;
	34	+ __kernel_ulong_t shmall;
	35	+ __kernel_ulong_t __unused1;
	36	+ __kernel_ulong_t __unused2;
	37	+ __kernel_ulong_t __unused3;
	38	+ __kernel_ulong_t __unused4;
	39	+};
	40	+
	41	+#endif
	42	+
	43	+#endif /* __ASM_X86_SHMBUF_H */

arch/x86/kernel/acpi/boot.c+4−0

@@ -215,6 +215,10 @@ acpi_parse_x2apic(struct acpi_subtable_header *header, const unsigned long end) · 214 unmodified lines
215	215	apic_id = processor->local_apic_id;
216	216	enabled = processor->lapic_flags & ACPI_MADT_ENABLED;
217	217
	218	+ /* Ignore invalid ID */
	219	+ if (apic_id == 0xffffffff)
	220	+ return 0;
	221	+
218	222	/*
219	223	* We need to register disabled CPU as well to permit
220	224	* counting disabled CPUs. This allows us to size

arch/x86/kernel/cpu/common.c+5−1

@@ -848,6 +848,11 @@ void get_cpu_cap(struct cpuinfo_x86 *c) · 847 unmodified lines
848	848	c->x86_power = edx;
849	849	}
850	850
	851	+ if (c->extended_cpuid_level >= 0x80000008) {
	852	+ cpuid(0x80000008, &eax, &ebx, &ecx, &edx);
	853	+ c->x86_capability[CPUID_8000_0008_EBX] = ebx;
	854	+ }
	855	+
851	856	if (c->extended_cpuid_level >= 0x8000000a)
852	857	c->x86_capability[CPUID_8000_000A_EDX] = cpuid_edx(0x8000000a);
853	858
@@ -871,7 +876,6 @@ static void get_cpu_address_sizes(struct cpuinfo_x86 *c) · 17 unmodified lines
871	876
872	877	c->x86_virt_bits = (eax >> 8) & 0xff;
873	878	c->x86_phys_bits = eax & 0xff;
874		− c->x86_capability[CPUID_8000_0008_EBX] = ebx;
875	879	}
876	880	#ifdef CONFIG_X86_32
877	881	else if (cpu_has(c, X86_FEATURE_PAE) \|\| cpu_has(c, X86_FEATURE_PSE36))

arch/x86/kernel/cpu/intel.c+3−0

@@ -835,6 +835,9 @@ static const struct _tlb_table intel_tlb_table[] = { · 834 unmodified lines
835	835	{ 0x5d, TLB_DATA_4K_4M, 256, " TLB_DATA 4 KByte and 4 MByte pages" },
836	836	{ 0x61, TLB_INST_4K, 48, " TLB_INST 4 KByte pages, full associative" },
837	837	{ 0x63, TLB_DATA_1G, 4, " TLB_DATA 1 GByte pages, 4-way set associative" },
	838	+ { 0x6b, TLB_DATA_4K, 256, " TLB_DATA 4 KByte pages, 8-way associative" },
	839	+ { 0x6c, TLB_DATA_2M_4M, 128, " TLB_DATA 2 MByte or 4 MByte pages, 8-way associative" },
	840	+ { 0x6d, TLB_DATA_1G, 16, " TLB_DATA 1 GByte pages, fully associative" },
838	841	{ 0x76, TLB_INST_2M_4M, 8, " TLB_INST 2-MByte or 4-MByte pages, fully associative" },
839	842	{ 0xb0, TLB_INST_4K, 128, " TLB_INST 4 KByte pages, 4-way set associative" },
840	843	{ 0xb1, TLB_INST_2M_4M, 4, " TLB_INST 2M pages, 4-way, 8 entries or 4M pages, 4-way entries" },

arch/x86/kernel/cpu/microcode/core.c+2−4

@@ -564,14 +564,12 @@ static int __reload_late(void *info) · 563 unmodified lines
564	564	apply_microcode_local(&err);
565	565	spin_unlock(&update_lock);
566	566
	567	+ /* siblings return UCODE_OK because their engine got updated already */
567	568	if (err > UCODE_NFOUND) {
568	569	pr_warn("Error reloading microcode on CPU %d\n", cpu);
569		− return -1;
570		− /* siblings return UCODE_OK because their engine got updated already */
	570	+ ret = -1;
571	571	} else if (err == UCODE_UPDATED \|\| err == UCODE_OK) {
572	572	ret = 1;
573		− } else {
574		− return ret;
575	573	}
576	574
577	575	/*

arch/x86/kernel/cpu/microcode/intel.c+0−2

@@ -485,7 +485,6 @@ static void show_saved_mc(void) · 484 unmodified lines
485	485	*/
486	486	static void save_mc_for_early(u8 *mc, unsigned int size)
487	487	{
488		−#ifdef CONFIG_HOTPLUG_CPU
489	488	/* Synchronization during CPU hotplug. */
490	489	static DEFINE_MUTEX(x86_cpu_microcode_mutex);
491	490
@@ -495,7 +494,6 @@ static void save_mc_for_early(u8 *mc, unsigned int size) · 3 unmodified lines
495	494	show_saved_mc();
496	495
497	496	mutex_unlock(&x86_cpu_microcode_mutex);
498		−#endif
499	497	}
500	498
501	499	static bool load_builtin_intel_microcode(struct cpio_data *cp)

arch/x86/kernel/jailhouse.c+1−1

@@ -1,4 +1,4 @@
1		−// SPDX-License-Identifier: GPL2.0
	1	+// SPDX-License-Identifier: GPL-2.0
2	2	/*
3	3	* Jailhouse paravirt_ops implementation
4	4	*

arch/x86/kernel/kexec-bzimage64.c+2−3

@@ -398,11 +398,10 @@ static void bzImage64_load(struct kimage image, char *kernel, · 397 unmodified lines
398	398	* little bit simple
399	399	*/
400	400	efi_map_sz = efi_get_runtime_map_size();
401		− efi_map_sz = ALIGN(efi_map_sz, 16);
402	401	params_cmdline_sz = sizeof(struct boot_params) + cmdline_len +
403	402	MAX_ELFCOREHDR_STR_LEN;
404	403	params_cmdline_sz = ALIGN(params_cmdline_sz, 16);
405		− kbuf.bufsz = params_cmdline_sz + efi_map_sz +
	404	+ kbuf.bufsz = params_cmdline_sz + ALIGN(efi_map_sz, 16) +
406	405	sizeof(struct setup_data) +
407	406	sizeof(struct efi_setup_data);
408	407
@@ -410,7 +409,7 @@ static void bzImage64_load(struct kimage image, char *kernel, · 1 unmodified line
410	409	if (!params)
411	410	return ERR_PTR(-ENOMEM);
412	411	efi_map_offset = params_cmdline_sz;
413		− efi_setup_data_offset = efi_map_offset + efi_map_sz;
	412	+ efi_setup_data_offset = efi_map_offset + ALIGN(efi_map_sz, 16);
414	413
415	414	/* Copy setup header onto bootparams. Documentation/x86/boot.txt */
416	415	setup_header_size = 0x0202 + kernel[0x0201] - setup_hdr_offset;

arch/x86/kernel/ldt.c+1−1

@@ -166,7 +166,7 @@ map_ldt_struct(struct mm_struct mm, struct ldt_struct ldt, int slot) · 165 unmodified lines
166	166	*/
167	167	pte_prot = __pgprot(__PAGE_KERNEL_RO & ~_PAGE_GLOBAL);
168	168	/* Filter out unsuppored __PAGE_KERNEL* bits: */
169		− pgprot_val(pte_prot) \|= __supported_pte_mask;
	169	+ pgprot_val(pte_prot) &= __supported_pte_mask;
170	170	pte = pfn_pte(pfn, pte_prot);
171	171	set_pte_at(mm, va, ptep, pte);
172	172	pte_unmap_unlock(ptep, ptl);

arch/x86/kernel/pci-nommu.c+0−90

@@ -1,90 +0,0 @@
1		−// SPDX-License-Identifier: GPL-2.0
2		−/* Fallback functions when the main IOMMU code is not compiled in. This
3		− code is roughly equivalent to i386. */
4		−#include <linux/dma-direct.h>
5		−#include <linux/scatterlist.h>
6		−#include <linux/string.h>
7		−#include <linux/gfp.h>
8		−#include <linux/pci.h>
9		−#include <linux/mm.h>
10		−
11		−#include <asm/processor.h>
12		−#include <asm/iommu.h>
13		−#include <asm/dma.h>
14		−
15		−#define NOMMU_MAPPING_ERROR 0
16		−
17		−static int
18		−check_addr(char name, struct device hwdev, dma_addr_t bus, size_t size)
19		−{
20		− if (hwdev && !dma_capable(hwdev, bus, size)) {
21		− if (*hwdev->dma_mask >= DMA_BIT_MASK(32))
22		− printk(KERN_ERR
23		− "nommu_%s: overflow %Lx+%zu of device mask %Lx\n",
24		− name, (long long)bus, size,
25		− (long long)*hwdev->dma_mask);
26		− return 0;
27		− }
28		− return 1;
29		−}
30		−
31		−static dma_addr_t nommu_map_page(struct device dev, struct page page,
32		− unsigned long offset, size_t size,
33		− enum dma_data_direction dir,
34		− unsigned long attrs)
35		−{
36		− dma_addr_t bus = phys_to_dma(dev, page_to_phys(page)) + offset;
37		− WARN_ON(size == 0);
38		− if (!check_addr("map_single", dev, bus, size))
39		− return NOMMU_MAPPING_ERROR;
40		− return bus;
41		−}
42		−
43		−/* Map a set of buffers described by scatterlist in streaming
44		− * mode for DMA. This is the scatter-gather version of the
45		− * above pci_map_single interface. Here the scatter gather list
46		− * elements are each tagged with the appropriate dma address
47		− * and length. They are obtained via sg_dma_{address,length}(SG).
48		− *
49		− * NOTE: An implementation may be able to use a smaller number of
50		− * DMA address/length pairs than there are SG table elements.
51		− * (for example via virtual mapping capabilities)
52		− * The routine returns the number of addr/length pairs actually
53		− * used, at most nents.
54		− *
55		− * Device ownership issues as mentioned above for pci_map_single are
56		− * the same here.
57		− */
58		−static int nommu_map_sg(struct device hwdev, struct scatterlist sg,
59		− int nents, enum dma_data_direction dir,
60		− unsigned long attrs)
61		−{
62		− struct scatterlist *s;
63		− int i;
64		−
65		− WARN_ON(nents == 0 \|\| sg[0].length == 0);
66		−
67		− for_each_sg(sg, s, nents, i) {
68		− BUG_ON(!sg_page(s));
69		− s->dma_address = sg_phys(s);
70		− if (!check_addr("map_sg", hwdev, s->dma_address, s->length))
71		− return 0;
72		− s->dma_length = s->length;
73		− }
74		− return nents;
75		−}
76		−
77		−static int nommu_mapping_error(struct device *dev, dma_addr_t dma_addr)
78		−{
79		− return dma_addr == NOMMU_MAPPING_ERROR;
80		−}
81		−
82		−const struct dma_map_ops nommu_dma_ops = {
83		− .alloc = dma_generic_alloc_coherent,
84		− .free = dma_generic_free_coherent,
85		− .map_sg = nommu_map_sg,
86		− .map_page = nommu_map_page,
87		− .is_phys = 1,
88		− .mapping_error = nommu_mapping_error,
89		− .dma_supported = x86_dma_supported,
90		−};

arch/x86/kernel/setup.c+6−0

@@ -50,6 +50,7 @@ · 49 unmodified lines
50	50	#include <linux/init_ohci1394_dma.h>
51	51	#include <linux/kvm_para.h>
52	52	#include <linux/dma-contiguous.h>
	53	+#include <xen/xen.h>
53	54
54	55	#include <linux/errno.h>
55	56	#include <linux/kernel.h>
@@ -534,6 +535,11 @@ static void __init reserve_crashkernel(void) · 478 unmodified lines
534	535	high = true;
535	536	}
536	537
	538	+ if (xen_pv_domain()) {
	539	+ pr_info("Ignoring crashkernel for a Xen PV domain\n");
	540	+ return;
	541	+ }
	542	+
537	543	/* 0 means: find the address automatically */
538	544	if (crash_base <= 0) {
539	545	/*

arch/x86/kernel/smpboot.c+42−5

@@ -77,6 +77,8 @@ · 76 unmodified lines
77	77	#include <asm/i8259.h>
78	78	#include <asm/misc.h>
79	79	#include <asm/qspinlock.h>
	80	+#include <asm/intel-family.h>
	81	+#include <asm/cpu_device_id.h>
80	82
81	83	/* Number of siblings per CPU package */
82	84	int smp_num_siblings = 1;
@@ -390,15 +392,47 @@ static bool match_smt(struct cpuinfo_x86 c, struct cpuinfo_x86 o) · 307 unmodified lines
390	392	return false;
391	393	}
392	394
	395	+/*
	396	+ * Define snc_cpu[] for SNC (Sub-NUMA Cluster) CPUs.
	397	+ *
	398	+ * These are Intel CPUs that enumerate an LLC that is shared by
	399	+ * multiple NUMA nodes. The LLC on these systems is shared for
	400	+ * off-package data access but private to the NUMA node (half
	401	+ * of the package) for on-package access.
	402	+ *
	403	+ * CPUID (the source of the information about the LLC) can only
	404	+ * enumerate the cache as being shared or unshared, but not
	405	+ * this particular configuration. The CPU in this case enumerates
	406	+ * the cache to be shared across the entire package (spanning both
	407	+ * NUMA nodes).
	408	+ */
	409	+
	410	+static const struct x86_cpu_id snc_cpu[] = {
	411	+ { X86_VENDOR_INTEL, 6, INTEL_FAM6_SKYLAKE_X },
	412	+ {}
	413	+};
	414	+
393	415	static bool match_llc(struct cpuinfo_x86 c, struct cpuinfo_x86 o)
394	416	{
395	417	int cpu1 = c->cpu_index, cpu2 = o->cpu_index;
396	418
397		− if (per_cpu(cpu_llc_id, cpu1) != BAD_APICID &&
398		− per_cpu(cpu_llc_id, cpu1) == per_cpu(cpu_llc_id, cpu2))
399		− return topology_sane(c, o, "llc");
	419	+ /* Do not match if we do not have a valid APICID for cpu: */
	420	+ if (per_cpu(cpu_llc_id, cpu1) == BAD_APICID)
	421	+ return false;
400	422
401		− return false;
	423	+ /* Do not match if LLC id does not match: */
	424	+ if (per_cpu(cpu_llc_id, cpu1) != per_cpu(cpu_llc_id, cpu2))
	425	+ return false;
	426	+
	427	+ /*
	428	+ * Allow the SNC topology without warning. Return of false
	429	+ * means 'c' does not share the LLC of 'o'. This will be
	430	+ * reflected to userspace.
	431	+ */
	432	+ if (!topology_same_node(c, o) && x86_match_cpu(snc_cpu))
	433	+ return false;
	434	+
	435	+ return topology_sane(c, o, "llc");
402	436	}
403	437
404	438	/*
@@ -456,7 +490,8 @@ static struct sched_domain_topology_level x86_topology[] = { · 51 unmodified lines
456	490
457	491	/*
458	492	* Set if a package/die has multiple NUMA nodes inside.
459		− * AMD Magny-Cours and Intel Cluster-on-Die have this.
	493	+ * AMD Magny-Cours, Intel Cluster-on-Die, and Intel
	494	+ * Sub-NUMA Clustering have this.
460	495	*/
461	496	static bool x86_has_numa_in_package;
462	497
@@ -1536,6 +1571,8 @@ static inline void mwait_play_dead(void) · 1073 unmodified lines
1536	1571	void *mwait_ptr;
1537	1572	int i;
1538	1573
	1574	+ if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD)
	1575	+ return;
1539	1576	if (!this_cpu_has(X86_FEATURE_MWAIT))
1540	1577	return;
1541	1578	if (!this_cpu_has(X86_FEATURE_CLFLUSH))

arch/x86/kernel/tsc.c+12−12

@@ -317,7 +317,7 @@ static unsigned long calc_hpet_ref(u64 deltatsc, u64 hpet1, u64 hpet2) · 316 unmodified lines
317	317	hpet2 -= hpet1;
318	318	tmp = ((u64)hpet2 * hpet_readl(HPET_PERIOD));
319	319	do_div(tmp, 1000000);
320		− do_div(deltatsc, tmp);
	320	+ deltatsc = div64_u64(deltatsc, tmp);
321	321
322	322	return (unsigned long) deltatsc;
323	323	}
@@ -1067,6 +1067,7 @@ static struct clocksource clocksource_tsc_early = { · 743 unmodified lines
1067	1067	.resume = tsc_resume,
1068	1068	.mark_unstable = tsc_cs_mark_unstable,
1069	1069	.tick_stable = tsc_cs_tick_stable,
	1070	+ .list = LIST_HEAD_INIT(clocksource_tsc_early.list),
1070	1071	};
1071	1072
1072	1073	/*
@@ -1086,6 +1087,7 @@ static struct clocksource clocksource_tsc = { · 13 unmodified lines
1086	1087	.resume = tsc_resume,
1087	1088	.mark_unstable = tsc_cs_mark_unstable,
1088	1089	.tick_stable = tsc_cs_tick_stable,
	1090	+ .list = LIST_HEAD_INIT(clocksource_tsc.list),
1089	1091	};
1090	1092
1091	1093	void mark_tsc_unstable(char *reason)
@@ -1098,13 +1100,9 @@ void mark_tsc_unstable(char *reason) · 6 unmodified lines
1098	1100	clear_sched_clock_stable();
1099	1101	disable_sched_clock_irqtime();
1100	1102	pr_info("Marking TSC unstable due to %s\n", reason);
1101		− /* Change only the rating, when not registered */
1102		− if (clocksource_tsc.mult) {
1103		− clocksource_mark_unstable(&clocksource_tsc);
1104		− } else {
1105		− clocksource_tsc.flags \|= CLOCK_SOURCE_UNSTABLE;
1106		− clocksource_tsc.rating = 0;
1107		− }
	1103	+
	1104	+ clocksource_mark_unstable(&clocksource_tsc_early);
	1105	+ clocksource_mark_unstable(&clocksource_tsc);
1108	1106	}
1109	1107
1110	1108	EXPORT_SYMBOL_GPL(mark_tsc_unstable);
@@ -1244,7 +1242,7 @@ static void tsc_refine_calibration_work(struct work_struct *work) · 133 unmodified lines
1244	1242
1245	1243	/* Don't bother refining TSC on unstable systems */
1246	1244	if (tsc_unstable)
1247		− return;
	1245	+ goto unreg;
1248	1246
1249	1247	/*
1250	1248	* Since the work is started early in boot, we may be
@@ -1297,11 +1295,12 @@ static void tsc_refine_calibration_work(struct work_struct *work) · 46 unmodified lines
1297	1295
1298	1296	out:
1299	1297	if (tsc_unstable)
1300		− return;
	1298	+ goto unreg;
1301	1299
1302	1300	if (boot_cpu_has(X86_FEATURE_ART))
1303	1301	art_related_clocksource = &clocksource_tsc;
1304	1302	clocksource_register_khz(&clocksource_tsc, tsc_khz);
	1303	+unreg:
1305	1304	clocksource_unregister(&clocksource_tsc_early);
1306	1305	}
1307	1306
@@ -1311,8 +1310,8 @@ static int __init init_tsc_clocksource(void) · 3 unmodified lines
1311	1310	if (!boot_cpu_has(X86_FEATURE_TSC) \|\| tsc_disabled > 0 \|\| !tsc_khz)
1312	1311	return 0;
1313	1312
1314		− if (check_tsc_unstable())
1315		− return 0;
	1313	+ if (tsc_unstable)
	1314	+ goto unreg;
1316	1315
1317	1316	if (tsc_clocksource_reliable)
1318	1317	clocksource_tsc.flags &= ~CLOCK_SOURCE_MUST_VERIFY;
@@ -1328,6 +1327,7 @@ static int __init init_tsc_clocksource(void) · 9 unmodified lines
1328	1327	if (boot_cpu_has(X86_FEATURE_ART))
1329	1328	art_related_clocksource = &clocksource_tsc;
1330	1329	clocksource_register_khz(&clocksource_tsc, tsc_khz);
	1330	+unreg:
1331	1331	clocksource_unregister(&clocksource_tsc_early);
1332	1332	return 0;
1333	1333	}

arch/x86/kvm/lapic.c+20−17

@@ -1463,23 +1463,6 @@ static void start_sw_tscdeadline(struct kvm_lapic *apic) · 1462 unmodified lines
1463	1463	local_irq_restore(flags);
1464	1464	}
1465	1465
1466		−static void start_sw_period(struct kvm_lapic *apic)
1467		−{
1468		− if (!apic->lapic_timer.period)
1469		− return;
1470		−
1471		− if (apic_lvtt_oneshot(apic) &&
1472		− ktime_after(ktime_get(),
1473		− apic->lapic_timer.target_expiration)) {
1474		− apic_timer_expired(apic);
1475		− return;
1476		− }
1477		−
1478		− hrtimer_start(&apic->lapic_timer.timer,
1479		− apic->lapic_timer.target_expiration,
1480		− HRTIMER_MODE_ABS_PINNED);
1481		−}
1482		−
1483	1466	static void update_target_expiration(struct kvm_lapic *apic, uint32_t old_divisor)
1484	1467	{
1485	1468	ktime_t now, remaining;
@@ -1546,6 +1529,26 @@ static void advance_periodic_target_expiration(struct kvm_lapic *apic) · 60 unmodified lines
1546	1529	apic->lapic_timer.period);
1547	1530	}
1548	1531
	1532	+static void start_sw_period(struct kvm_lapic *apic)
	1533	+{
	1534	+ if (!apic->lapic_timer.period)
	1535	+ return;
	1536	+
	1537	+ if (ktime_after(ktime_get(),
	1538	+ apic->lapic_timer.target_expiration)) {
	1539	+ apic_timer_expired(apic);
	1540	+
	1541	+ if (apic_lvtt_oneshot(apic))
	1542	+ return;
	1543	+
	1544	+ advance_periodic_target_expiration(apic);
	1545	+ }
	1546	+
	1547	+ hrtimer_start(&apic->lapic_timer.timer,
	1548	+ apic->lapic_timer.target_expiration,
	1549	+ HRTIMER_MODE_ABS_PINNED);
	1550	+}
	1551	+
1549	1552	bool kvm_lapic_hv_timer_in_use(struct kvm_vcpu *vcpu)
1550	1553	{
1551	1554	if (!lapic_in_kernel(vcpu))

arch/x86/kvm/svm.c+18−13

@@ -1423,12 +1423,23 @@ static void init_sys_seg(struct vmcb_seg *seg, uint32_t type) · 1422 unmodified lines
1423	1423	seg->base = 0;
1424	1424	}
1425	1425
	1426	+static u64 svm_read_l1_tsc_offset(struct kvm_vcpu *vcpu)
	1427	+{
	1428	+ struct vcpu_svm *svm = to_svm(vcpu);
	1429	+
	1430	+ if (is_guest_mode(vcpu))
	1431	+ return svm->nested.hsave->control.tsc_offset;
	1432	+
	1433	+ return vcpu->arch.tsc_offset;
	1434	+}
	1435	+
1426	1436	static void svm_write_tsc_offset(struct kvm_vcpu *vcpu, u64 offset)
1427	1437	{
1428	1438	struct vcpu_svm *svm = to_svm(vcpu);
1429	1439	u64 g_tsc_offset = 0;
1430	1440
1431	1441	if (is_guest_mode(vcpu)) {
	1442	+ /* Write L1's TSC offset. */
1432	1443	g_tsc_offset = svm->vmcb->control.tsc_offset -
1433	1444	svm->nested.hsave->control.tsc_offset;
1434	1445	svm->nested.hsave->control.tsc_offset = offset;
@@ -3322,6 +3333,7 @@ static int nested_svm_vmexit(struct vcpu_svm *svm) · 1887 unmodified lines
3322	3333	/* Restore the original control entries */
3323	3334	copy_vmcb_control_area(vmcb, hsave);
3324	3335
	3336	+ svm->vcpu.arch.tsc_offset = svm->vmcb->control.tsc_offset;
3325	3337	kvm_clear_exception_queue(&svm->vcpu);
3326	3338	kvm_clear_interrupt_queue(&svm->vcpu);
3327	3339
@@ -3482,10 +3494,12 @@ static void enter_svm_guest_mode(struct vcpu_svm *svm, u64 vmcb_gpa, · 154 unmodified lines
3482	3494	/* We don't want to see VMMCALLs from a nested guest */
3483	3495	clr_intercept(svm, INTERCEPT_VMMCALL);
3484	3496
	3497	+ svm->vcpu.arch.tsc_offset += nested_vmcb->control.tsc_offset;
	3498	+ svm->vmcb->control.tsc_offset = svm->vcpu.arch.tsc_offset;
	3499	+
3485	3500	svm->vmcb->control.virt_ext = nested_vmcb->control.virt_ext;
3486	3501	svm->vmcb->control.int_vector = nested_vmcb->control.int_vector;
3487	3502	svm->vmcb->control.int_state = nested_vmcb->control.int_state;
3488		− svm->vmcb->control.tsc_offset += nested_vmcb->control.tsc_offset;
3489	3503	svm->vmcb->control.event_inj = nested_vmcb->control.event_inj;
3490	3504	svm->vmcb->control.event_inj_err = nested_vmcb->control.event_inj_err;
3491	3505
@@ -4035,12 +4049,6 @@ static int svm_get_msr(struct kvm_vcpu vcpu, struct msr_data msr_info) · 543 unmodified lines
4035	4049	struct vcpu_svm *svm = to_svm(vcpu);
4036	4050
4037	4051	switch (msr_info->index) {
4038		− case MSR_IA32_TSC: {
4039		− msr_info->data = svm->vmcb->control.tsc_offset +
4040		− kvm_scale_tsc(vcpu, rdtsc());
4041		−
4042		− break;
4043		− }
4044	4052	case MSR_STAR:
4045	4053	msr_info->data = svm->vmcb->save.star;
4046	4054	break;
@@ -4193,9 +4201,6 @@ static int svm_set_msr(struct kvm_vcpu vcpu, struct msr_data msr) · 146 unmodified lines
4193	4201	svm->vmcb->save.g_pat = data;
4194	4202	mark_dirty(svm->vmcb, VMCB_NPT);
4195	4203	break;
4196		− case MSR_IA32_TSC:
4197		− kvm_write_tsc(vcpu, msr);
4198		− break;
4199	4204	case MSR_IA32_SPEC_CTRL:
4200	4205	if (!msr->host_initiated &&
4201	4206	!guest_cpuid_has(vcpu, X86_FEATURE_IBRS))
@@ -5265,9 +5270,8 @@ static int svm_update_pi_irte(struct kvm *kvm, unsigned int host_irq, · 1063 unmodified lines
5265	5270	}
5266	5271
5267	5272	if (!ret && svm) {
5268		− trace_kvm_pi_irte_update(svm->vcpu.vcpu_id,
5269		− host_irq, e->gsi,
5270		− vcpu_info.vector,
	5273	+ trace_kvm_pi_irte_update(host_irq, svm->vcpu.vcpu_id,
	5274	+ e->gsi, vcpu_info.vector,
5271	5275	vcpu_info.pi_desc_addr, set);
5272	5276	}
5273	5277
@@ -7102,6 +7106,7 @@ static struct kvm_x86_ops svm_x86_ops __ro_after_init = { · 1828 unmodified lines
7102	7106
7103	7107	.has_wbinvd_exit = svm_has_wbinvd_exit,
7104	7108
	7109	+ .read_l1_tsc_offset = svm_read_l1_tsc_offset,
7105	7110	.write_tsc_offset = svm_write_tsc_offset,
7106	7111
7107	7112	.set_tdp_cr3 = set_tdp_cr3,

arch/x86/kvm/vmx.c+59−50

@@ -2880,18 +2880,15 @@ static void setup_msrs(struct vcpu_vmx *vmx) · 2879 unmodified lines
2880	2880	vmx_update_msr_bitmap(&vmx->vcpu);
2881	2881	}
2882	2882
2883		−/*
2884		− * reads and returns guest's timestamp counter "register"
2885		− * guest_tsc = (host_tsc * tsc multiplier) >> 48 + tsc_offset
2886		− * -- Intel TSC Scaling for Virtualization White Paper, sec 1.3
2887		− */
2888		−static u64 guest_read_tsc(struct kvm_vcpu *vcpu)
	2883	+static u64 vmx_read_l1_tsc_offset(struct kvm_vcpu *vcpu)
2889	2884	{
2890		− u64 host_tsc, tsc_offset;
	2885	+ struct vmcs12 *vmcs12 = get_vmcs12(vcpu);
2891	2886
2892		− host_tsc = rdtsc();
2893		− tsc_offset = vmcs_read64(TSC_OFFSET);
2894		− return kvm_scale_tsc(vcpu, host_tsc) + tsc_offset;
	2887	+ if (is_guest_mode(vcpu) &&
	2888	+ (vmcs12->cpu_based_vm_exec_control & CPU_BASED_USE_TSC_OFFSETING))
	2889	+ return vcpu->arch.tsc_offset - vmcs12->tsc_offset;
	2890	+
	2891	+ return vcpu->arch.tsc_offset;
2895	2892	}
2896	2893
2897	2894	/*
@@ -3524,9 +3521,6 @@ static int vmx_get_msr(struct kvm_vcpu vcpu, struct msr_data msr_info) · 626 unmodified lines
3524	3521	#endif
3525	3522	case MSR_EFER:
3526	3523	return kvm_get_msr_common(vcpu, msr_info);
3527		− case MSR_IA32_TSC:
3528		− msr_info->data = guest_read_tsc(vcpu);
3529		− break;
3530	3524	case MSR_IA32_SPEC_CTRL:
3531	3525	if (!msr_info->host_initiated &&
3532	3526	!guest_cpuid_has(vcpu, X86_FEATURE_IBRS) &&
@@ -3646,9 +3640,6 @@ static int vmx_set_msr(struct kvm_vcpu vcpu, struct msr_data msr_info) · 113 unmodified lines
3646	3640	return 1;
3647	3641	vmcs_write64(GUEST_BNDCFGS, data);
3648	3642	break;
3649		− case MSR_IA32_TSC:
3650		− kvm_write_tsc(vcpu, msr_info);
3651		− break;
3652	3643	case MSR_IA32_SPEC_CTRL:
3653	3644	if (!msr_info->host_initiated &&
3654	3645	!guest_cpuid_has(vcpu, X86_FEATURE_IBRS) &&
@@ -4553,12 +4544,6 @@ static void vmx_flush_tlb(struct kvm_vcpu *vcpu, bool invalidate_gpa) · 898 unmodified lines
4553	4544	__vmx_flush_tlb(vcpu, to_vmx(vcpu)->vpid, invalidate_gpa);
4554	4545	}
4555	4546
4556		−static void vmx_flush_tlb_ept_only(struct kvm_vcpu *vcpu)
4557		−{
4558		− if (enable_ept)
4559		− vmx_flush_tlb(vcpu, true);
4560		−}
4561		−
4562	4547	static void vmx_decache_cr0_guest_bits(struct kvm_vcpu *vcpu)
4563	4548	{
4564	4549	ulong cr0_guest_owned_bits = vcpu->arch.cr0_guest_owned_bits;
@@ -9287,7 +9272,7 @@ static void vmx_set_virtual_x2apic_mode(struct kvm_vcpu *vcpu, bool set) · 4722 unmodified lines
9287	9272	} else {
9288	9273	sec_exec_control &= ~SECONDARY_EXEC_VIRTUALIZE_X2APIC_MODE;
9289	9274	sec_exec_control \|= SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES;
9290		− vmx_flush_tlb_ept_only(vcpu);
	9275	+ vmx_flush_tlb(vcpu, true);
9291	9276	}
9292	9277	vmcs_write32(SECONDARY_VM_EXEC_CONTROL, sec_exec_control);
9293	9278
@@ -9315,7 +9300,7 @@ static void vmx_set_apic_access_page_addr(struct kvm_vcpu *vcpu, hpa_t hpa) · 21 unmodified lines
9315	9300	!nested_cpu_has2(get_vmcs12(&vmx->vcpu),
9316	9301	SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES)) {
9317	9302	vmcs_write64(APIC_ACCESS_ADDR, hpa);
9318		− vmx_flush_tlb_ept_only(vcpu);
	9303	+ vmx_flush_tlb(vcpu, true);
9319	9304	}
9320	9305	}
9321	9306
@@ -10608,6 +10593,16 @@ static inline bool nested_vmx_prepare_msr_bitmap(struct kvm_vcpu *vcpu, · 1286 unmodified lines
10608	10593	return true;
10609	10594	}
10610	10595
	10596	+static int nested_vmx_check_apic_access_controls(struct kvm_vcpu *vcpu,
	10597	+ struct vmcs12 *vmcs12)
	10598	+{
	10599	+ if (nested_cpu_has2(vmcs12, SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES) &&
	10600	+ !page_address_valid(vcpu, vmcs12->apic_access_addr))
	10601	+ return -EINVAL;
	10602	+ else
	10603	+ return 0;
	10604	+}
	10605	+
10611	10606	static int nested_vmx_check_apicv_controls(struct kvm_vcpu *vcpu,
10612	10607	struct vmcs12 *vmcs12)
10613	10608	{
@@ -11176,11 +11171,8 @@ static int prepare_vmcs02(struct kvm_vcpu vcpu, struct vmcs12 vmcs12, · 562 unmodified lines
11176	11171	vmcs_write64(GUEST_IA32_PAT, vmx->vcpu.arch.pat);
11177	11172	}
11178	11173
11179		− if (vmcs12->cpu_based_vm_exec_control & CPU_BASED_USE_TSC_OFFSETING)
11180		− vmcs_write64(TSC_OFFSET,
11181		− vcpu->arch.tsc_offset + vmcs12->tsc_offset);
11182		− else
11183		− vmcs_write64(TSC_OFFSET, vcpu->arch.tsc_offset);
	11174	+ vmcs_write64(TSC_OFFSET, vcpu->arch.tsc_offset);
	11175	+
11184	11176	if (kvm_has_tsc_control)
11185	11177	decache_tsc_multiplier(vmx);
11186	11178
@@ -11222,7 +11214,7 @@ static int prepare_vmcs02(struct kvm_vcpu vcpu, struct vmcs12 vmcs12, · 35 unmodified lines
11222	11214	}
11223	11215	} else if (nested_cpu_has2(vmcs12,
11224	11216	SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES)) {
11225		− vmx_flush_tlb_ept_only(vcpu);
	11217	+ vmx_flush_tlb(vcpu, true);
11226	11218	}
11227	11219
11228	11220	/*
@@ -11299,6 +11291,9 @@ static int check_vmentry_prereqs(struct kvm_vcpu vcpu, struct vmcs12 vmcs12) · 70 unmodified lines
11299	11291	if (nested_vmx_check_msr_bitmap_controls(vcpu, vmcs12))
11300	11292	return VMXERR_ENTRY_INVALID_CONTROL_FIELD;
11301	11293
	11294	+ if (nested_vmx_check_apic_access_controls(vcpu, vmcs12))
	11295	+ return VMXERR_ENTRY_INVALID_CONTROL_FIELD;
	11296	+
11302	11297	if (nested_vmx_check_tpr_shadow_controls(vcpu, vmcs12))
11303	11298	return VMXERR_ENTRY_INVALID_CONTROL_FIELD;
11304	11299
@@ -11420,6 +11415,7 @@ static int enter_vmx_non_root_mode(struct kvm_vcpu *vcpu, bool from_vmentry) · 115 unmodified lines
11420	11415	struct vmcs12 *vmcs12 = get_vmcs12(vcpu);
11421	11416	u32 msr_entry_idx;
11422	11417	u32 exit_qual;
	11418	+ int r;
11423	11419
11424	11420	enter_guest_mode(vcpu);
11425	11421
@@ -11429,26 +11425,21 @@ static int enter_vmx_non_root_mode(struct kvm_vcpu *vcpu, bool from_vmentry) · 3 unmodified lines
11429	11425	vmx_switch_vmcs(vcpu, &vmx->nested.vmcs02);
11430	11426	vmx_segment_cache_clear(vmx);
11431	11427
11432		− if (prepare_vmcs02(vcpu, vmcs12, from_vmentry, &exit_qual)) {
11433		− leave_guest_mode(vcpu);
11434		− vmx_switch_vmcs(vcpu, &vmx->vmcs01);
11435		− nested_vmx_entry_failure(vcpu, vmcs12,
11436		− EXIT_REASON_INVALID_STATE, exit_qual);
11437		− return 1;
11438		− }
	11428	+ if (vmcs12->cpu_based_vm_exec_control & CPU_BASED_USE_TSC_OFFSETING)
	11429	+ vcpu->arch.tsc_offset += vmcs12->tsc_offset;
	11430	+
	11431	+ r = EXIT_REASON_INVALID_STATE;
	11432	+ if (prepare_vmcs02(vcpu, vmcs12, from_vmentry, &exit_qual))
	11433	+ goto fail;
11439	11434
11440	11435	nested_get_vmcs12_pages(vcpu, vmcs12);
11441	11436
	11437	+ r = EXIT_REASON_MSR_LOAD_FAIL;
11442	11438	msr_entry_idx = nested_vmx_load_msr(vcpu,
11443	11439	vmcs12->vm_entry_msr_load_addr,
11444	11440	vmcs12->vm_entry_msr_load_count);
11445		− if (msr_entry_idx) {
11446		− leave_guest_mode(vcpu);
11447		− vmx_switch_vmcs(vcpu, &vmx->vmcs01);
11448		− nested_vmx_entry_failure(vcpu, vmcs12,
11449		− EXIT_REASON_MSR_LOAD_FAIL, msr_entry_idx);
11450		− return 1;
11451		− }
	11441	+ if (msr_entry_idx)
	11442	+ goto fail;
11452	11443
11453	11444	/*
11454	11445	* Note no nested_vmx_succeed or nested_vmx_fail here. At this point
@@ -11457,6 +11448,14 @@ static int enter_vmx_non_root_mode(struct kvm_vcpu *vcpu, bool from_vmentry) · 2 unmodified lines
11457	11448	* the success flag) when L2 exits (see nested_vmx_vmexit()).
11458	11449	*/
11459	11450	return 0;
	11451	+
	11452	+fail:
	11453	+ if (vmcs12->cpu_based_vm_exec_control & CPU_BASED_USE_TSC_OFFSETING)
	11454	+ vcpu->arch.tsc_offset -= vmcs12->tsc_offset;
	11455	+ leave_guest_mode(vcpu);
	11456	+ vmx_switch_vmcs(vcpu, &vmx->vmcs01);
	11457	+ nested_vmx_entry_failure(vcpu, vmcs12, r, exit_qual);
	11458	+ return 1;
11460	11459	}
11461	11460
11462	11461	/*
@@ -12028,6 +12027,9 @@ static void nested_vmx_vmexit(struct kvm_vcpu *vcpu, u32 exit_reason, · 565 unmodified lines
12028	12027
12029	12028	leave_guest_mode(vcpu);
12030	12029
	12030	+ if (vmcs12->cpu_based_vm_exec_control & CPU_BASED_USE_TSC_OFFSETING)
	12031	+ vcpu->arch.tsc_offset -= vmcs12->tsc_offset;
	12032	+
12031	12033	if (likely(!vmx->fail)) {
12032	12034	if (exit_reason == -1)
12033	12035	sync_vmcs12(vcpu, vmcs12);
@@ -12065,7 +12067,7 @@ static void nested_vmx_vmexit(struct kvm_vcpu *vcpu, u32 exit_reason, · 31 unmodified lines
12065	12067	} else if (!nested_cpu_has_ept(vmcs12) &&
12066	12068	nested_cpu_has2(vmcs12,
12067	12069	SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES)) {
12068		− vmx_flush_tlb_ept_only(vcpu);
	12070	+ vmx_flush_tlb(vcpu, true);
12069	12071	}
12070	12072
12071	12073	/* This is needed for same reason as it was needed in prepare_vmcs02 */
@@ -12224,10 +12226,16 @@ static inline int u64_shl_div_u64(u64 a, unsigned int shift, · 152 unmodified lines
12224	12226
12225	12227	static int vmx_set_hv_timer(struct kvm_vcpu *vcpu, u64 guest_deadline_tsc)
12226	12228	{
12227		− struct vcpu_vmx *vmx = to_vmx(vcpu);
12228		− u64 tscl = rdtsc();
12229		− u64 guest_tscl = kvm_read_l1_tsc(vcpu, tscl);
12230		− u64 delta_tsc = max(guest_deadline_tsc, guest_tscl) - guest_tscl;
	12229	+ struct vcpu_vmx *vmx;
	12230	+ u64 tscl, guest_tscl, delta_tsc;
	12231	+
	12232	+ if (kvm_mwait_in_guest(vcpu->kvm))
	12233	+ return -EOPNOTSUPP;
	12234	+
	12235	+ vmx = to_vmx(vcpu);
	12236	+ tscl = rdtsc();
	12237	+ guest_tscl = kvm_read_l1_tsc(vcpu, tscl);
	12238	+ delta_tsc = max(guest_deadline_tsc, guest_tscl) - guest_tscl;
12231	12239
12232	12240	/* Convert to host delta tsc if tsc scaling is enabled */
12233	12241	if (vcpu->arch.tsc_scaling_ratio != kvm_default_tsc_scaling_ratio &&
@@ -12533,7 +12541,7 @@ static int vmx_update_pi_irte(struct kvm *kvm, unsigned int host_irq, · 299 unmodified lines
12533	12541	vcpu_info.pi_desc_addr = __pa(vcpu_to_pi_desc(vcpu));
12534	12542	vcpu_info.vector = irq.vector;
12535	12543
12536		− trace_kvm_pi_irte_update(vcpu->vcpu_id, host_irq, e->gsi,
	12544	+ trace_kvm_pi_irte_update(host_irq, vcpu->vcpu_id, e->gsi,
12537	12545	vcpu_info.vector, vcpu_info.pi_desc_addr, set);
12538	12546
12539	12547	if (set)
@@ -12712,6 +12720,7 @@ static struct kvm_x86_ops vmx_x86_ops __ro_after_init = { · 172 unmodified lines
12712	12720
12713	12721	.has_wbinvd_exit = cpu_has_vmx_wbinvd_exit,
12714	12722
	12723	+ .read_l1_tsc_offset = vmx_read_l1_tsc_offset,
12715	12724	.write_tsc_offset = vmx_write_tsc_offset,
12716	12725
12717	12726	.set_tdp_cr3 = vmx_set_cr3,

arch/x86/kvm/x86.c+12−3

@@ -1490,7 +1490,7 @@ static void kvm_track_tsc_matching(struct kvm_vcpu *vcpu) · 1489 unmodified lines
1490	1490
1491	1491	static void update_ia32_tsc_adjust_msr(struct kvm_vcpu *vcpu, s64 offset)
1492	1492	{
1493		− u64 curr_offset = vcpu->arch.tsc_offset;
	1493	+ u64 curr_offset = kvm_x86_ops->read_l1_tsc_offset(vcpu);
1494	1494	vcpu->arch.ia32_tsc_adjust_msr += offset - curr_offset;
1495	1495	}
1496	1496
@@ -1532,7 +1532,9 @@ static u64 kvm_compute_tsc_offset(struct kvm_vcpu *vcpu, u64 target_tsc) · 35 unmodified lines
1532	1532
1533	1533	u64 kvm_read_l1_tsc(struct kvm_vcpu *vcpu, u64 host_tsc)
1534	1534	{
1535		− return vcpu->arch.tsc_offset + kvm_scale_tsc(vcpu, host_tsc);
	1535	+ u64 tsc_offset = kvm_x86_ops->read_l1_tsc_offset(vcpu);
	1536	+
	1537	+ return tsc_offset + kvm_scale_tsc(vcpu, host_tsc);
1536	1538	}
1537	1539	EXPORT_SYMBOL_GPL(kvm_read_l1_tsc);
1538	1540
@@ -2362,6 +2364,9 @@ int kvm_set_msr_common(struct kvm_vcpu vcpu, struct msr_data msr_info) · 823 unmodified lines
2362	2364	return 1;
2363	2365	vcpu->arch.smbase = data;
2364	2366	break;
	2367	+ case MSR_IA32_TSC:
	2368	+ kvm_write_tsc(vcpu, msr_info);
	2369	+ break;
2365	2370	case MSR_SMI_COUNT:
2366	2371	if (!msr_info->host_initiated)
2367	2372	return 1;
@@ -2605,6 +2610,9 @@ int kvm_get_msr_common(struct kvm_vcpu vcpu, struct msr_data msr_info) · 237 unmodified lines
2605	2610	case MSR_IA32_UCODE_REV:
2606	2611	msr_info->data = vcpu->arch.microcode_version;
2607	2612	break;
	2613	+ case MSR_IA32_TSC:
	2614	+ msr_info->data = kvm_scale_tsc(vcpu, rdtsc()) + vcpu->arch.tsc_offset;
	2615	+ break;
2608	2616	case MSR_MTRRcap:
2609	2617	case 0x200 ... 0x2ff:
2610	2618	return kvm_mtrr_get_msr(vcpu, msr_info->index, &msr_info->data);
@@ -2819,7 +2827,8 @@ out: · 208 unmodified lines
2819	2827	static inline bool kvm_can_mwait_in_guest(void)
2820	2828	{
2821	2829	return boot_cpu_has(X86_FEATURE_MWAIT) &&
2822		− !boot_cpu_has_bug(X86_BUG_MONITOR);
	2830	+ !boot_cpu_has_bug(X86_BUG_MONITOR) &&
	2831	+ boot_cpu_has(X86_FEATURE_ARAT);
2823	2832	}
2824	2833
2825	2834	int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)

arch/x86/kvm/x86.h+0−7

@@ -302,13 +302,6 @@ static inline u64 nsec_to_cycles(struct kvm_vcpu *vcpu, u64 nsec) · 301 unmodified lines
302	302	__rem; \
303	303	})
304	304
305		−#define KVM_X86_DISABLE_EXITS_MWAIT (1 << 0)
306		−#define KVM_X86_DISABLE_EXITS_HTL (1 << 1)
307		−#define KVM_X86_DISABLE_EXITS_PAUSE (1 << 2)
308		−#define KVM_X86_DISABLE_VALID_EXITS (KVM_X86_DISABLE_EXITS_MWAIT \| \
309		− KVM_X86_DISABLE_EXITS_HTL \| \
310		− KVM_X86_DISABLE_EXITS_PAUSE)
311		−
312	305	static inline bool kvm_mwait_in_guest(struct kvm *kvm)
313	306	{
314	307	return kvm->arch.mwait_in_guest;

arch/x86/mm/dump_pagetables.c+6−5

@@ -18,6 +18,7 @@ · 17 unmodified lines
18	18	#include <linux/init.h>
19	19	#include <linux/sched.h>
20	20	#include <linux/seq_file.h>
	21	+#include <linux/highmem.h>
21	22
22	23	#include <asm/pgtable.h>
23	24
@@ -334,16 +335,16 @@ static void walk_pte_level(struct seq_file m, struct pg_state st, pmd_t addr, · 310 unmodified lines
334	335	pgprotval_t eff_in, unsigned long P)
335	336	{
336	337	int i;
337		− pte_t *start;
	338	+ pte_t *pte;
338	339	pgprotval_t prot, eff;
339	340
340		− start = (pte_t *)pmd_page_vaddr(addr);
341	341	for (i = 0; i < PTRS_PER_PTE; i++) {
342		− prot = pte_flags(*start);
343		− eff = effective_prot(eff_in, prot);
344	342	st->current_address = normalize_addr(P + i * PTE_LEVEL_MULT);
	343	+ pte = pte_offset_map(&addr, st->current_address);
	344	+ prot = pte_flags(*pte);
	345	+ eff = effective_prot(eff_in, prot);
345	346	note_page(m, st, __pgprot(prot), eff, 5);
346		− start++;
	347	+ pte_unmap(pte);
347	348	}
348	349	}
349	350	#ifdef CONFIG_KASAN

arch/x86/mm/pageattr.c+32−12

@@ -93,6 +93,18 @@ void arch_report_meminfo(struct seq_file *m) · 92 unmodified lines
93	93	static inline void split_page_count(int level) { }
94	94	#endif
95	95
	96	+static inline int
	97	+within(unsigned long addr, unsigned long start, unsigned long end)
	98	+{
	99	+ return addr >= start && addr < end;
	100	+}
	101	+
	102	+static inline int
	103	+within_inclusive(unsigned long addr, unsigned long start, unsigned long end)
	104	+{
	105	+ return addr >= start && addr <= end;
	106	+}
	107	+
96	108	#ifdef CONFIG_X86_64
97	109
98	110	static inline unsigned long highmap_start_pfn(void)
@@ -106,20 +118,25 @@ static inline unsigned long highmap_end_pfn(void) · 7 unmodified lines
106	118	return __pa_symbol(roundup(_brk_end, PMD_SIZE) - 1) >> PAGE_SHIFT;
107	119	}
108	120
109		−#endif
110		−
111		−static inline int
112		−within(unsigned long addr, unsigned long start, unsigned long end)
	121	+static bool __cpa_pfn_in_highmap(unsigned long pfn)
113	122	{
114		− return addr >= start && addr < end;
	123	+ /*
	124	+ * Kernel text has an alias mapping at a high address, known
	125	+ * here as "highmap".
	126	+ */
	127	+ return within_inclusive(pfn, highmap_start_pfn(), highmap_end_pfn());
115	128	}
116	129
117		−static inline int
118		−within_inclusive(unsigned long addr, unsigned long start, unsigned long end)
	130	+#else
	131	+
	132	+static bool __cpa_pfn_in_highmap(unsigned long pfn)
119	133	{
120		− return addr >= start && addr <= end;
	134	+ /* There is no highmap on 32-bit */
	135	+ return false;
121	136	}
122	137
	138	+#endif
	139	+
123	140	/*
124	141	* Flushing functions
125	142	*/
@@ -172,7 +189,7 @@ static void __cpa_flush_all(void *arg) · 46 unmodified lines
172	189
173	190	static void cpa_flush_all(unsigned long cache)
174	191	{
175		− BUG_ON(irqs_disabled());
	192	+ BUG_ON(irqs_disabled() && !early_boot_irqs_disabled);
176	193
177	194	on_each_cpu(__cpa_flush_all, (void *) cache, 1);
178	195	}
@@ -236,7 +253,7 @@ static void cpa_flush_array(unsigned long *start, int numpages, int cache, · 57 unmodified lines
236	253	unsigned long do_wbinvd = cache && numpages >= 1024; /* 4M threshold */
237	254	#endif
238	255
239		− BUG_ON(irqs_disabled());
	256	+ BUG_ON(irqs_disabled() && !early_boot_irqs_disabled);
240	257
241	258	on_each_cpu(__cpa_flush_all, (void *) do_wbinvd, 1);
242	259
@@ -1183,6 +1200,10 @@ static int __cpa_process_fault(struct cpa_data *cpa, unsigned long vaddr, · 940 unmodified lines
1183	1200	cpa->numpages = 1;
1184	1201	cpa->pfn = __pa(vaddr) >> PAGE_SHIFT;
1185	1202	return 0;
	1203	+
	1204	+ } else if (__cpa_pfn_in_highmap(cpa->pfn)) {
	1205	+ /* Faults in the highmap are OK, so do not warn: */
	1206	+ return -EFAULT;
1186	1207	} else {
1187	1208	WARN(1, KERN_WARNING "CPA: called for zero pte. "
1188	1209	"vaddr = %lx cpa->vaddr = %lx\n", vaddr,
@@ -1335,8 +1356,7 @@ static int cpa_process_alias(struct cpa_data *cpa) · 146 unmodified lines
1335	1356	* to touch the high mapped kernel as well:
1336	1357	*/
1337	1358	if (!within(vaddr, (unsigned long)_text, _brk_end) &&
1338		− within_inclusive(cpa->pfn, highmap_start_pfn(),
1339		− highmap_end_pfn())) {
	1359	+ __cpa_pfn_in_highmap(cpa->pfn)) {
1340	1360	unsigned long temp_cpa_vaddr = (cpa->pfn << PAGE_SHIFT) +
1341	1361	__START_KERNEL_map - phys_base;
1342	1362	alias_cpa = *cpa;

arch/x86/mm/pti.c+23−3

@@ -421,6 +421,16 @@ static inline bool pti_kernel_image_global_ok(void) · 420 unmodified lines
421	421	if (boot_cpu_has(X86_FEATURE_K8))
422	422	return false;
423	423
	424	+ /*
	425	+ * RANDSTRUCT derives its hardening benefits from the
	426	+ * attacker's lack of knowledge about the layout of kernel
	427	+ * data structures. Keep the kernel image non-global in
	428	+ * cases where RANDSTRUCT is in use to help keep the layout a
	429	+ * secret.
	430	+ */
	431	+ if (IS_ENABLED(CONFIG_GCC_PLUGIN_RANDSTRUCT))
	432	+ return false;
	433	+
424	434	return true;
425	435	}
426	436
@@ -430,12 +440,24 @@ static inline bool pti_kernel_image_global_ok(void) · 3 unmodified lines
430	440	*/
431	441	void pti_clone_kernel_text(void)
432	442	{
	443	+ /*
	444	+ * rodata is part of the kernel image and is normally
	445	+ * readable on the filesystem or on the web. But, do not
	446	+ * clone the areas past rodata, they might contain secrets.
	447	+ */
433	448	unsigned long start = PFN_ALIGN(_text);
434		− unsigned long end = ALIGN((unsigned long)_end, PMD_PAGE_SIZE);
	449	+ unsigned long end = (unsigned long)__end_rodata_hpage_align;
435	450
436	451	if (!pti_kernel_image_global_ok())
437	452	return;
438	453
	454	+ pr_debug("mapping partial kernel image into user address space\n");
	455	+
	456	+ /*
	457	+ * Note that this will undo _some_ of the work that
	458	+ * pti_set_kernel_image_nonglobal() did to clear the
	459	+ * global bit.
	460	+ */
439	461	pti_clone_pmds(start, end, _PAGE_RW);
440	462	}
441	463
@@ -458,8 +480,6 @@ void pti_set_kernel_image_nonglobal(void) · 16 unmodified lines
458	480	if (pti_kernel_image_global_ok())
459	481	return;
460	482
461		− pr_debug("set kernel image non-global\n");
462		−
463	483	set_memory_nonglobal(start, (end - start) >> PAGE_SHIFT);
464	484	}
465	485

arch/x86/net/bpf_jit_comp.c+14−4

@@ -1027,7 +1027,17 @@ emit_cond_jmp: /* convert BPF opcode to x86 */ · 1026 unmodified lines
1027	1027	break;
1028	1028
1029	1029	case BPF_JMP \| BPF_JA:
1030		− jmp_offset = addrs[i + insn->off] - addrs[i];
	1030	+ if (insn->off == -1)
	1031	+ /* -1 jmp instructions will always jump
	1032	+ * backwards two bytes. Explicitly handling
	1033	+ * this case avoids wasting too many passes
	1034	+ * when there are long sequences of replaced
	1035	+ * dead code.
	1036	+ */
	1037	+ jmp_offset = -2;
	1038	+ else
	1039	+ jmp_offset = addrs[i + insn->off] - addrs[i];
	1040	+
1031	1041	if (!jmp_offset)
1032	1042	/* optimize out nop jumps */
1033	1043	break;
@@ -1226,6 +1236,7 @@ skip_init_addrs: · 192 unmodified lines
1226	1236	for (pass = 0; pass < 20 \|\| image; pass++) {
1227	1237	proglen = do_jit(prog, addrs, image, oldproglen, &ctx);
1228	1238	if (proglen <= 0) {
	1239	+out_image:
1229	1240	image = NULL;
1230	1241	if (header)
1231	1242	bpf_jit_binary_free(header);
@@ -1236,8 +1247,7 @@ skip_init_addrs: · 4 unmodified lines
1236	1247	if (proglen != oldproglen) {
1237	1248	pr_err("bpf_jit: proglen=%d != oldproglen=%d\n",
1238	1249	proglen, oldproglen);
1239		− prog = orig_prog;
1240		− goto out_addrs;
	1250	+ goto out_image;
1241	1251	}
1242	1252	break;
1243	1253	}
@@ -1273,7 +1283,7 @@ skip_init_addrs: · 29 unmodified lines
1273	1283	prog = orig_prog;
1274	1284	}
1275	1285
1276		− if (!prog->is_func \|\| extra_pass) {
	1286	+ if (!image \|\| !prog->is_func \|\| extra_pass) {
1277	1287	out_addrs:
1278	1288	kfree(addrs);
1279	1289	kfree(jit_data);

arch/x86/power/hibernate_64.c+1−1

@@ -98,7 +98,7 @@ static int set_up_temporary_text_mapping(pgd_t *pgd) · 97 unmodified lines
98	98	set_pgd(pgd + pgd_index(restore_jump_address), new_pgd);
99	99	} else {
100	100	/* No p4d for 4-level paging: point the pgd to the pud page table */
101		− pgd_t new_pgd = __pgd(__pa(p4d) \| pgprot_val(pgtable_prot));
	101	+ pgd_t new_pgd = __pgd(__pa(pud) \| pgprot_val(pgtable_prot));
102	102	set_pgd(pgd + pgd_index(restore_jump_address), new_pgd);
103	103	}
104	104

arch/x86/xen/enlighten_pv.c+31−55

@@ -421,45 +421,33 @@ static void xen_load_gdt(const struct desc_ptr *dtr) · 420 unmodified lines
421	421	{
422	422	unsigned long va = dtr->address;
423	423	unsigned int size = dtr->size + 1;
424		− unsigned pages = DIV_ROUND_UP(size, PAGE_SIZE);
425		− unsigned long frames[pages];
426		− int f;
427		−
428		− /*
429		− * A GDT can be up to 64k in size, which corresponds to 8192
430		− * 8-byte entries, or 16 4k pages..
431		− */
	424	+ unsigned long pfn, mfn;
	425	+ int level;
	426	+ pte_t *ptep;
	427	+ void *virt;
432	428
433		− BUG_ON(size > 65536);
	429	+ /* @size should be at most GDT_SIZE which is smaller than PAGE_SIZE. */
	430	+ BUG_ON(size > PAGE_SIZE);
434	431	BUG_ON(va & ~PAGE_MASK);
435	432
436		− for (f = 0; va < dtr->address + size; va += PAGE_SIZE, f++) {
437		− int level;
438		− pte_t *ptep;
439		− unsigned long pfn, mfn;
440		− void *virt;
441		−
442		− /*
443		− * The GDT is per-cpu and is in the percpu data area.
444		− * That can be virtually mapped, so we need to do a
445		− * page-walk to get the underlying MFN for the
446		− * hypercall. The page can also be in the kernel's
447		− * linear range, so we need to RO that mapping too.
448		− */
449		− ptep = lookup_address(va, &level);
450		− BUG_ON(ptep == NULL);
451		−
452		− pfn = pte_pfn(*ptep);
453		− mfn = pfn_to_mfn(pfn);
454		− virt = __va(PFN_PHYS(pfn));
	433	+ /*
	434	+ * The GDT is per-cpu and is in the percpu data area.
	435	+ * That can be virtually mapped, so we need to do a
	436	+ * page-walk to get the underlying MFN for the
	437	+ * hypercall. The page can also be in the kernel's
	438	+ * linear range, so we need to RO that mapping too.
	439	+ */
	440	+ ptep = lookup_address(va, &level);
	441	+ BUG_ON(ptep == NULL);
455	442
456		− frames[f] = mfn;
	443	+ pfn = pte_pfn(*ptep);
	444	+ mfn = pfn_to_mfn(pfn);
	445	+ virt = __va(PFN_PHYS(pfn));
457	446
458		− make_lowmem_page_readonly((void *)va);
459		− make_lowmem_page_readonly(virt);
460		− }
	447	+ make_lowmem_page_readonly((void *)va);
	448	+ make_lowmem_page_readonly(virt);
461	449
462		− if (HYPERVISOR_set_gdt(frames, size / sizeof(struct desc_struct)))
	450	+ if (HYPERVISOR_set_gdt(&mfn, size / sizeof(struct desc_struct)))
463	451	BUG();
464	452	}
465	453
@@ -470,34 +458,22 @@ static void __init xen_load_gdt_boot(const struct desc_ptr *dtr) · 4 unmodified lines
470	458	{
471	459	unsigned long va = dtr->address;
472	460	unsigned int size = dtr->size + 1;
473		− unsigned pages = DIV_ROUND_UP(size, PAGE_SIZE);
474		− unsigned long frames[pages];
475		− int f;
476		−
477		− /*
478		− * A GDT can be up to 64k in size, which corresponds to 8192
479		− * 8-byte entries, or 16 4k pages..
480		− */
	461	+ unsigned long pfn, mfn;
	462	+ pte_t pte;
481	463
482		− BUG_ON(size > 65536);
	464	+ /* @size should be at most GDT_SIZE which is smaller than PAGE_SIZE. */
	465	+ BUG_ON(size > PAGE_SIZE);
483	466	BUG_ON(va & ~PAGE_MASK);
484	467
485		− for (f = 0; va < dtr->address + size; va += PAGE_SIZE, f++) {
486		− pte_t pte;
487		− unsigned long pfn, mfn;
	468	+ pfn = virt_to_pfn(va);
	469	+ mfn = pfn_to_mfn(pfn);
488	470
489		− pfn = virt_to_pfn(va);
490		− mfn = pfn_to_mfn(pfn);
	471	+ pte = pfn_pte(pfn, PAGE_KERNEL_RO);
491	472
492		− pte = pfn_pte(pfn, PAGE_KERNEL_RO);
493		−
494		− if (HYPERVISOR_update_va_mapping((unsigned long)va, pte, 0))
495		− BUG();
496		−
497		− frames[f] = mfn;
498		− }
	473	+ if (HYPERVISOR_update_va_mapping((unsigned long)va, pte, 0))
	474	+ BUG();
499	475
500		− if (HYPERVISOR_set_gdt(frames, size / sizeof(struct desc_struct)))
	476	+ if (HYPERVISOR_set_gdt(&mfn, size / sizeof(struct desc_struct)))
501	477	BUG();
502	478	}
503	479

block/bfq-iosched.c+9−1

@@ -4934,8 +4934,16 @@ static void bfq_prepare_request(struct request rq, struct bio bio) · 4933 unmodified lines
4934	4934	bool new_queue = false;
4935	4935	bool bfqq_already_existing = false, split = false;
4936	4936
4937		− if (!rq->elv.icq)
	4937	+ /*
	4938	+ * Even if we don't have an icq attached, we should still clear
	4939	+ * the scheduler pointers, as they might point to previously
	4940	+ * allocated bic/bfqq structs.
	4941	+ */
	4942	+ if (!rq->elv.icq) {
	4943	+ rq->elv.priv[0] = rq->elv.priv[1] = NULL;
4938	4944	return;
	4945	+ }
	4946	+
4939	4947	bic = icq_to_bic(rq->elv.icq);
4940	4948
4941	4949	spin_lock_irq(&bfqd->lock);

block/blk-cgroup.c+12−16

@@ -1177,26 +1177,20 @@ int blkcg_init_queue(struct request_queue *q) · 1176 unmodified lines
1177	1177
1178	1178	preloaded = !radix_tree_preload(GFP_KERNEL);
1179	1179
1180		− /*
1181		− * Make sure the root blkg exists and count the existing blkgs. As
1182		− * @q is bypassing at this point, blkg_lookup_create() can't be
1183		− * used. Open code insertion.
1184		− */
	1180	+ /* Make sure the root blkg exists. */
1185	1181	rcu_read_lock();
1186	1182	spin_lock_irq(q->queue_lock);
1187	1183	blkg = blkg_create(&blkcg_root, q, new_blkg);
	1184	+ if (IS_ERR(blkg))
	1185	+ goto err_unlock;
	1186	+ q->root_blkg = blkg;
	1187	+ q->root_rl.blkg = blkg;
1188	1188	spin_unlock_irq(q->queue_lock);
1189	1189	rcu_read_unlock();
1190	1190
1191	1191	if (preloaded)
1192	1192	radix_tree_preload_end();
1193	1193
1194		− if (IS_ERR(blkg))
1195		− return PTR_ERR(blkg);
1196		−
1197		− q->root_blkg = blkg;
1198		− q->root_rl.blkg = blkg;
1199		−
1200	1194	ret = blk_throtl_init(q);
1201	1195	if (ret) {
1202	1196	spin_lock_irq(q->queue_lock);
@@ -1204,6 +1198,13 @@ int blkcg_init_queue(struct request_queue *q) · 1 unmodified line
1204	1198	spin_unlock_irq(q->queue_lock);
1205	1199	}
1206	1200	return ret;
	1201	+
	1202	+err_unlock:
	1203	+ spin_unlock_irq(q->queue_lock);
	1204	+ rcu_read_unlock();
	1205	+ if (preloaded)
	1206	+ radix_tree_preload_end();
	1207	+ return PTR_ERR(blkg);
1207	1208	}
1208	1209
1209	1210	/**
@@ -1410,9 +1411,6 @@ void blkcg_deactivate_policy(struct request_queue *q, · 200 unmodified lines
1410	1411	__clear_bit(pol->plid, q->blkcg_pols);
1411	1412
1412	1413	list_for_each_entry(blkg, &q->blkg_list, q_node) {
1413		− /* grab blkcg lock too while removing @pd from @blkg */
1414		− spin_lock(&blkg->blkcg->lock);
1415		−
1416	1414	if (blkg->pd[pol->plid]) {
1417	1415	if (!blkg->pd[pol->plid]->offline &&
1418	1416	pol->pd_offline_fn) {
@@ -1422,8 +1420,6 @@ void blkcg_deactivate_policy(struct request_queue *q, · 3 unmodified lines
1422	1420	pol->pd_free_fn(blkg->pd[pol->plid]);
1423	1421	blkg->pd[pol->plid] = NULL;
1424	1422	}
1425		−
1426		− spin_unlock(&blkg->blkcg->lock);
1427	1423	}
1428	1424
1429	1425	spin_unlock_irq(q->queue_lock);

block/blk-core.c+8−7

@@ -201,6 +201,10 @@ void blk_rq_init(struct request_queue q, struct request rq) · 200 unmodified lines
201	201	rq->part = NULL;
202	202	seqcount_init(&rq->gstate_seq);
203	203	u64_stats_init(&rq->aborted_gstate_sync);
	204	+ /*
	205	+ * See comment of blk_mq_init_request
	206	+ */
	207	+ WRITE_ONCE(rq->gstate, MQ_RQ_GEN_INC);
204	208	}
205	209	EXPORT_SYMBOL(blk_rq_init);
206	210
@@ -915,7 +919,6 @@ int blk_queue_enter(struct request_queue *q, blk_mq_req_flags_t flags) · 708 unmodified lines
915	919
916	920	while (true) {
917	921	bool success = false;
918		− int ret;
919	922
920	923	rcu_read_lock();
921	924	if (percpu_ref_tryget_live(&q->q_usage_counter)) {
@@ -947,14 +950,12 @@ int blk_queue_enter(struct request_queue *q, blk_mq_req_flags_t flags) · 25 unmodified lines
947	950	*/
948	951	smp_rmb();
949	952
950		− ret = wait_event_interruptible(q->mq_freeze_wq,
951		− (atomic_read(&q->mq_freeze_depth) == 0 &&
952		− (preempt \|\| !blk_queue_preempt_only(q))) \|\|
953		− blk_queue_dying(q));
	953	+ wait_event(q->mq_freeze_wq,
	954	+ (atomic_read(&q->mq_freeze_depth) == 0 &&
	955	+ (preempt \|\| !blk_queue_preempt_only(q))) \|\|
	956	+ blk_queue_dying(q));
954	957	if (blk_queue_dying(q))
955	958	return -ENODEV;
956		− if (ret)
957		− return ret;
958	959	}
959	960	}
960	961

block/blk-mq.c+66−15

@@ -95,18 +95,15 @@ static void blk_mq_check_inflight(struct blk_mq_hw_ctx *hctx, · 94 unmodified lines
95	95	{
96	96	struct mq_inflight *mi = priv;
97	97
98		− if (blk_mq_rq_state(rq) == MQ_RQ_IN_FLIGHT) {
99		− /*
100		− * index[0] counts the specific partition that was asked
101		− * for. index[1] counts the ones that are active on the
102		− * whole device, so increment that if mi->part is indeed
103		− * a partition, and not a whole device.
104		− */
105		− if (rq->part == mi->part)
106		− mi->inflight[0]++;
107		− if (mi->part->partno)
108		− mi->inflight[1]++;
109		− }
	98	+ /*
	99	+ * index[0] counts the specific partition that was asked for. index[1]
	100	+ * counts the ones that are active on the whole device, so increment
	101	+ * that if mi->part is indeed a partition, and not a whole device.
	102	+ */
	103	+ if (rq->part == mi->part)
	104	+ mi->inflight[0]++;
	105	+ if (mi->part->partno)
	106	+ mi->inflight[1]++;
110	107	}
111	108
112	109	void blk_mq_in_flight(struct request_queue q, struct hd_struct part,
@@ -118,6 +115,25 @@ void blk_mq_in_flight(struct request_queue q, struct hd_struct part, · 5 unmodified lines
118	115	blk_mq_queue_tag_busy_iter(q, blk_mq_check_inflight, &mi);
119	116	}
120	117
	118	+static void blk_mq_check_inflight_rw(struct blk_mq_hw_ctx *hctx,
	119	+ struct request rq, void priv,
	120	+ bool reserved)
	121	+{
	122	+ struct mq_inflight *mi = priv;
	123	+
	124	+ if (rq->part == mi->part)
	125	+ mi->inflight[rq_data_dir(rq)]++;
	126	+}
	127	+
	128	+void blk_mq_in_flight_rw(struct request_queue q, struct hd_struct part,
	129	+ unsigned int inflight[2])
	130	+{
	131	+ struct mq_inflight mi = { .part = part, .inflight = inflight, };
	132	+
	133	+ inflight[0] = inflight[1] = 0;
	134	+ blk_mq_queue_tag_busy_iter(q, blk_mq_check_inflight_rw, &mi);
	135	+}
	136	+
121	137	void blk_freeze_queue_start(struct request_queue *q)
122	138	{
123	139	int freeze_depth;
@@ -2042,6 +2058,13 @@ static int blk_mq_init_request(struct blk_mq_tag_set set, struct request rq, · 1918 unmodified lines
2042	2058
2043	2059	seqcount_init(&rq->gstate_seq);
2044	2060	u64_stats_init(&rq->aborted_gstate_sync);
	2061	+ /*
	2062	+ * start gstate with gen 1 instead of 0, otherwise it will be equal
	2063	+ * to aborted_gstate, and be identified timed out by
	2064	+ * blk_mq_terminate_expired.
	2065	+ */
	2066	+ WRITE_ONCE(rq->gstate, MQ_RQ_GEN_INC);
	2067	+
2045	2068	return 0;
2046	2069	}
2047	2070
@@ -2329,7 +2352,7 @@ static void blk_mq_free_map_and_requests(struct blk_mq_tag_set *set, · 281 unmodified lines
2329	2352
2330	2353	static void blk_mq_map_swqueue(struct request_queue *q)
2331	2354	{
2332		− unsigned int i;
	2355	+ unsigned int i, hctx_idx;
2333	2356	struct blk_mq_hw_ctx *hctx;
2334	2357	struct blk_mq_ctx *ctx;
2335	2358	struct blk_mq_tag_set *set = q->tag_set;
@@ -2346,8 +2369,23 @@ static void blk_mq_map_swqueue(struct request_queue *q) · 10 unmodified lines
2346	2369
2347	2370	/*
2348	2371	* Map software to hardware queues.
	2372	+ *
	2373	+ * If the cpu isn't present, the cpu is mapped to first hctx.
2349	2374	*/
2350	2375	for_each_possible_cpu(i) {
	2376	+ hctx_idx = q->mq_map[i];
	2377	+ /* unmapped hw queue can be remapped after CPU topo changed */
	2378	+ if (!set->tags[hctx_idx] &&
	2379	+ !__blk_mq_alloc_rq_map(set, hctx_idx)) {
	2380	+ /*
	2381	+ * If tags initialization fail for some hctx,
	2382	+ * that hctx won't be brought online. In this
	2383	+ * case, remap the current ctx to hctx[0] which
	2384	+ * is guaranteed to always have tags allocated
	2385	+ */
	2386	+ q->mq_map[i] = 0;
	2387	+ }
	2388	+
2351	2389	ctx = per_cpu_ptr(q->queue_ctx, i);
2352	2390	hctx = blk_mq_map_queue(q, i);
2353	2391
@@ -2359,8 +2397,21 @@ static void blk_mq_map_swqueue(struct request_queue *q) · 5 unmodified lines
2359	2397	mutex_unlock(&q->sysfs_lock);
2360	2398
2361	2399	queue_for_each_hw_ctx(q, hctx, i) {
2362		− /* every hctx should get mapped by at least one CPU */
2363		− WARN_ON(!hctx->nr_ctx);
	2400	+ /*
	2401	+ * If no software queues are mapped to this hardware queue,
	2402	+ * disable it and free the request entries.
	2403	+ */
	2404	+ if (!hctx->nr_ctx) {
	2405	+ /* Never unmap queue 0. We need it as a
	2406	+ * fallback in case of a new remap fails
	2407	+ * allocation
	2408	+ */
	2409	+ if (i && set->tags[i])
	2410	+ blk_mq_free_map_and_requests(set, i);
	2411	+
	2412	+ hctx->tags = NULL;
	2413	+ continue;
	2414	+ }
2364	2415
2365	2416	hctx->tags = set->tags[i];
2366	2417	WARN_ON(!hctx->tags);

block/blk-mq.h+6−1

@@ -7,6 +7,9 @@ · 6 unmodified lines
7	7
8	8	struct blk_mq_tag_set;
9	9
	10	+/**
	11	+ * struct blk_mq_ctx - State for a software queue facing the submitting CPUs
	12	+ */
10	13	struct blk_mq_ctx {
11	14	struct {
12	15	spinlock_t lock;
@@ -185,7 +188,9 @@ static inline bool blk_mq_hw_queue_mapped(struct blk_mq_hw_ctx *hctx) · 172 unmodified lines
185	188	}
186	189
187	190	void blk_mq_in_flight(struct request_queue q, struct hd_struct part,
188		− unsigned int inflight[2]);
	191	+ unsigned int inflight[2]);
	192	+void blk_mq_in_flight_rw(struct request_queue q, struct hd_struct part,
	193	+ unsigned int inflight[2]);
189	194
190	195	static inline void blk_mq_put_dispatch_budget(struct blk_mq_hw_ctx *hctx)
191	196	{

block/genhd.c+12−0

@@ -82,6 +82,18 @@ void part_in_flight(struct request_queue q, struct hd_struct part, · 81 unmodified lines
82	82	}
83	83	}
84	84
	85	+void part_in_flight_rw(struct request_queue q, struct hd_struct part,
	86	+ unsigned int inflight[2])
	87	+{
	88	+ if (q->mq_ops) {
	89	+ blk_mq_in_flight_rw(q, part, inflight);
	90	+ return;
	91	+ }
	92	+
	93	+ inflight[0] = atomic_read(&part->in_flight[0]);
	94	+ inflight[1] = atomic_read(&part->in_flight[1]);
	95	+}
	96	+
85	97	struct hd_struct __disk_get_part(struct gendisk disk, int partno)
86	98	{
87	99	struct disk_part_tbl *ptbl = rcu_dereference(disk->part_tbl);

block/partition-generic.c+6−4

@@ -145,13 +145,15 @@ ssize_t part_stat_show(struct device *dev, · 144 unmodified lines
145	145	jiffies_to_msecs(part_stat_read(p, time_in_queue)));
146	146	}
147	147
148		−ssize_t part_inflight_show(struct device *dev,
149		− struct device_attribute attr, char buf)
	148	+ssize_t part_inflight_show(struct device dev, struct device_attribute attr,
	149	+ char *buf)
150	150	{
151	151	struct hd_struct *p = dev_to_part(dev);
	152	+ struct request_queue *q = part_to_disk(p)->queue;
	153	+ unsigned int inflight[2];
152	154
153		− return sprintf(buf, "%8u %8u\n", atomic_read(&p->in_flight[0]),
154		− atomic_read(&p->in_flight[1]));
	155	+ part_in_flight_rw(q, p, inflight);
	156	+ return sprintf(buf, "%8u %8u\n", inflight[0], inflight[1]);
155	157	}
156	158
157	159	#ifdef CONFIG_FAIL_MAKE_REQUEST

crypto/api.c+8−3

@@ -204,9 +204,14 @@ static struct crypto_alg crypto_alg_lookup(const char name, u32 type, · 203 unmodified lines
204	204
205	205	down_read(&crypto_alg_sem);
206	206	alg = __crypto_alg_lookup(name, type \| test, mask \| test);
207		− if (!alg && test)
208		− alg = __crypto_alg_lookup(name, type, mask) ?
209		− ERR_PTR(-ELIBBAD) : NULL;
	207	+ if (!alg && test) {
	208	+ alg = __crypto_alg_lookup(name, type, mask);
	209	+ if (alg && !crypto_is_larval(alg)) {
	210	+ /* Test failed */
	211	+ crypto_mod_put(alg);
	212	+ alg = ERR_PTR(-ELIBBAD);
	213	+ }
	214	+ }
210	215	up_read(&crypto_alg_sem);
211	216
212	217	return alg;

crypto/drbg.c+2−0

@@ -1134,8 +1134,10 @@ static inline void drbg_dealloc_state(struct drbg_state *drbg) · 1133 unmodified lines
1134	1134	if (!drbg)
1135	1135	return;
1136	1136	kzfree(drbg->Vbuf);
	1137	+ drbg->Vbuf = NULL;
1137	1138	drbg->V = NULL;
1138	1139	kzfree(drbg->Cbuf);
	1140	+ drbg->Cbuf = NULL;
1139	1141	drbg->C = NULL;
1140	1142	kzfree(drbg->scratchpadbuf);
1141	1143	drbg->scratchpadbuf = NULL;

drivers/acpi/acpi_video.c+25−2

@@ -2123,6 +2123,25 @@ static int __init intel_opregion_present(void) · 2122 unmodified lines
2123	2123	return opregion;
2124	2124	}
2125	2125
	2126	+static bool dmi_is_desktop(void)
	2127	+{
	2128	+ const char *chassis_type;
	2129	+
	2130	+ chassis_type = dmi_get_system_info(DMI_CHASSIS_TYPE);
	2131	+ if (!chassis_type)
	2132	+ return false;
	2133	+
	2134	+ if (!strcmp(chassis_type, "3") \|\| /* 3: Desktop */
	2135	+ !strcmp(chassis_type, "4") \|\| /* 4: Low Profile Desktop */
	2136	+ !strcmp(chassis_type, "5") \|\| /* 5: Pizza Box */
	2137	+ !strcmp(chassis_type, "6") \|\| /* 6: Mini Tower */
	2138	+ !strcmp(chassis_type, "7") \|\| /* 7: Tower */
	2139	+ !strcmp(chassis_type, "11")) /* 11: Main Server Chassis */
	2140	+ return true;
	2141	+
	2142	+ return false;
	2143	+}
	2144	+
2126	2145	int acpi_video_register(void)
2127	2146	{
2128	2147	int ret = 0;
@@ -2143,8 +2162,12 @@ int acpi_video_register(void) · 14 unmodified lines
2143	2162	* win8 ready (where we also prefer the native backlight driver, so
2144	2163	* normally the acpi_video code should not register there anyways).
2145	2164	*/
2146		− if (only_lcd == -1)
2147		− only_lcd = acpi_osi_is_win8();
	2165	+ if (only_lcd == -1) {
	2166	+ if (dmi_is_desktop() && acpi_osi_is_win8())
	2167	+ only_lcd = true;
	2168	+ else
	2169	+ only_lcd = false;
	2170	+ }
2148	2171
2149	2172	dmi_check_system(video_dmi_table);
2150	2173

drivers/acpi/acpi_watchdog.c+49−10

@@ -12,23 +12,64 @@ · 11 unmodified lines
12	12	#define pr_fmt(fmt) "ACPI: watchdog: " fmt
13	13
14	14	#include <linux/acpi.h>
	15	+#include <linux/dmi.h>
15	16	#include <linux/ioport.h>
16	17	#include <linux/platform_device.h>
17	18
18	19	#include "internal.h"
19	20
	21	+static const struct dmi_system_id acpi_watchdog_skip[] = {
	22	+ {
	23	+ /*
	24	+ * On Lenovo Z50-70 there are two issues with the WDAT
	25	+ * table. First some of the instructions use RTC SRAM
	26	+ * to store persistent information. This does not work well
	27	+ * with Linux RTC driver. Second, more important thing is
	28	+ * that the instructions do not actually reset the system.
	29	+ *
	30	+ * On this particular system iTCO_wdt seems to work just
	31	+ * fine so we prefer that over WDAT for now.
	32	+ *
	33	+ * See also https://bugzilla.kernel.org/show_bug.cgi?id=199033.
	34	+ */
	35	+ .ident = "Lenovo Z50-70",
	36	+ .matches = {
	37	+ DMI_MATCH(DMI_SYS_VENDOR, "LENOVO"),
	38	+ DMI_MATCH(DMI_PRODUCT_NAME, "20354"),
	39	+ DMI_MATCH(DMI_PRODUCT_VERSION, "Lenovo Z50-70"),
	40	+ },
	41	+ },
	42	+ {}
	43	+};
	44	+
	45	+static const struct acpi_table_wdat *acpi_watchdog_get_wdat(void)
	46	+{
	47	+ const struct acpi_table_wdat *wdat = NULL;
	48	+ acpi_status status;
	49	+
	50	+ if (acpi_disabled)
	51	+ return NULL;
	52	+
	53	+ if (dmi_check_system(acpi_watchdog_skip))
	54	+ return NULL;
	55	+
	56	+ status = acpi_get_table(ACPI_SIG_WDAT, 0,
	57	+ (struct acpi_table_header **)&wdat);
	58	+ if (ACPI_FAILURE(status)) {
	59	+ /* It is fine if there is no WDAT */
	60	+ return NULL;
	61	+ }
	62	+
	63	+ return wdat;
	64	+}
	65	+
20	66	/**
21	67	* Returns true if this system should prefer ACPI based watchdog instead of
22	68	* the native one (which are typically the same hardware).
23	69	*/
24	70	bool acpi_has_watchdog(void)
25	71	{
26		− struct acpi_table_header hdr;
27		−
28		− if (acpi_disabled)
29		− return false;
30		−
31		− return ACPI_SUCCESS(acpi_get_table_header(ACPI_SIG_WDAT, 0, &hdr));
	72	+ return !!acpi_watchdog_get_wdat();
32	73	}
33	74	EXPORT_SYMBOL_GPL(acpi_has_watchdog);
34	75
@@ -41,12 +82,10 @@ void __init acpi_watchdog_init(void) · 6 unmodified lines
41	82	struct platform_device *pdev;
42	83	struct resource *resources;
43	84	size_t nresources = 0;
44		− acpi_status status;
45	85	int i;
46	86
47		− status = acpi_get_table(ACPI_SIG_WDAT, 0,
48		− (struct acpi_table_header **)&wdat);
49		− if (ACPI_FAILURE(status)) {
	87	+ wdat = acpi_watchdog_get_wdat();
	88	+ if (!wdat) {
50	89	/* It is fine if there is no WDAT */
51	90	return;
52	91	}

drivers/acpi/button.c+23−1

@@ -635,4 +635,26 @@ module_param_call(lid_init_state, · 634 unmodified lines
635	635	NULL, 0644);
636	636	MODULE_PARM_DESC(lid_init_state, "Behavior for reporting LID initial state");
637	637
638		−module_acpi_driver(acpi_button_driver);
	638	+static int acpi_button_register_driver(struct acpi_driver *driver)
	639	+{
	640	+ /*
	641	+ * Modules such as nouveau.ko and i915.ko have a link time dependency
	642	+ * on acpi_lid_open(), and would therefore not be loadable on ACPI
	643	+ * capable kernels booted in non-ACPI mode if the return value of
	644	+ * acpi_bus_register_driver() is returned from here with ACPI disabled
	645	+ * when this driver is built as a module.
	646	+ */
	647	+ if (acpi_disabled)
	648	+ return 0;
	649	+
	650	+ return acpi_bus_register_driver(driver);
	651	+}
	652	+
	653	+static void acpi_button_unregister_driver(struct acpi_driver *driver)
	654	+{
	655	+ if (!acpi_disabled)
	656	+ acpi_bus_unregister_driver(driver);
	657	+}
	658	+
	659	+module_driver(acpi_button_driver, acpi_button_register_driver,
	660	+ acpi_button_unregister_driver);

drivers/acpi/scan.c+1−1

@@ -2166,10 +2166,10 @@ int __init acpi_scan_init(void) · 2165 unmodified lines
2166	2166	acpi_cmos_rtc_init();
2167	2167	acpi_container_init();
2168	2168	acpi_memory_hotplug_init();
	2169	+ acpi_watchdog_init();
2169	2170	acpi_pnp_init();
2170	2171	acpi_int340x_thermal_init();
2171	2172	acpi_amba_init();
2172		− acpi_watchdog_init();
2173	2173	acpi_init_lpit();
2174	2174
2175	2175	acpi_scan_add_handler(&generic_device_handler);

drivers/acpi/sleep.c+13−0

@@ -364,6 +364,19 @@ static const struct dmi_system_id acpisleep_dmi_table[] __initconst = { · 363 unmodified lines
364	364	DMI_MATCH(DMI_PRODUCT_NAME, "XPS 13 9360"),
365	365	},
366	366	},
	367	+ /*
	368	+ * ThinkPad X1 Tablet(2016) cannot do suspend-to-idle using
	369	+ * the Low Power S0 Idle firmware interface (see
	370	+ * https://bugzilla.kernel.org/show_bug.cgi?id=199057).
	371	+ */
	372	+ {
	373	+ .callback = init_no_lps0,
	374	+ .ident = "ThinkPad X1 Tablet(2016)",
	375	+ .matches = {
	376	+ DMI_MATCH(DMI_SYS_VENDOR, "LENOVO"),
	377	+ DMI_MATCH(DMI_PRODUCT_NAME, "20GGA00L00"),
	378	+ },
	379	+ },
367	380	{},
368	381	};
369	382

drivers/amba/bus.c+11−6

@@ -69,11 +69,12 @@ static ssize_t driver_override_show(struct device *_dev, · 68 unmodified lines
69	69	struct device_attribute attr, char buf)
70	70	{
71	71	struct amba_device *dev = to_amba_device(_dev);
	72	+ ssize_t len;
72	73
73		− if (!dev->driver_override)
74		− return 0;
75		−
76		− return sprintf(buf, "%s\n", dev->driver_override);
	74	+ device_lock(_dev);
	75	+ len = sprintf(buf, "%s\n", dev->driver_override);
	76	+ device_unlock(_dev);
	77	+ return len;
77	78	}
78	79
79	80	static ssize_t driver_override_store(struct device *_dev,
@@ -81,9 +82,10 @@ static ssize_t driver_override_store(struct device *_dev, · 1 unmodified line
81	82	const char *buf, size_t count)
82	83	{
83	84	struct amba_device *dev = to_amba_device(_dev);
84		− char driver_override, old = dev->driver_override, *cp;
	85	+ char driver_override, old, *cp;
85	86
86		− if (count > PATH_MAX)
	87	+ /* We need to keep extra room for a newline */
	88	+ if (count >= (PAGE_SIZE - 1))
87	89	return -EINVAL;
88	90
89	91	driver_override = kstrndup(buf, count, GFP_KERNEL);
@@ -94,12 +96,15 @@ static ssize_t driver_override_store(struct device *_dev, · 4 unmodified lines
94	96	if (cp)
95	97	*cp = '\0';
96	98
	99	+ device_lock(_dev);
	100	+ old = dev->driver_override;
97	101	if (strlen(driver_override)) {
98	102	dev->driver_override = driver_override;
99	103	} else {
100	104	kfree(driver_override);
101	105	dev->driver_override = NULL;
102	106	}
	107	+ device_unlock(_dev);
103	108
104	109	kfree(old);
105	110

drivers/android/binder.c+8−0

@@ -2839,6 +2839,14 @@ static void binder_transaction(struct binder_proc *proc, · 2838 unmodified lines
2839	2839	else
2840	2840	return_error = BR_DEAD_REPLY;
2841	2841	mutex_unlock(&context->context_mgr_node_lock);
	2842	+ if (target_node && target_proc == proc) {
	2843	+ binder_user_error("%d:%d got transaction to context manager from process owning it\n",
	2844	+ proc->pid, thread->pid);
	2845	+ return_error = BR_FAILED_REPLY;
	2846	+ return_error_param = -EINVAL;
	2847	+ return_error_line = __LINE__;
	2848	+ goto err_invalid_target_handle;
	2849	+ }
2842	2850	}
2843	2851	if (!target_node) {
2844	2852	/*

drivers/atm/iphase.c+2−2

@@ -671,7 +671,7 @@ static void ia_tx_poll (IADEV *iadev) { · 670 unmodified lines
671	671	if ((vcc->pop) && (skb1->len != 0))
672	672	{
673	673	vcc->pop(vcc, skb1);
674		− IF_EVENT(printk("Tansmit Done - skb 0x%lx return\n",
	674	+ IF_EVENT(printk("Transmit Done - skb 0x%lx return\n",
675	675	(long)skb1);)
676	676	}
677	677	else
@@ -1665,7 +1665,7 @@ static void tx_intr(struct atm_dev *dev) · 987 unmodified lines
1665	1665	status = readl(iadev->seg_reg+SEG_INTR_STATUS_REG);
1666	1666	if (status & TRANSMIT_DONE){
1667	1667
1668		− IF_EVENT(printk("Tansmit Done Intr logic run\n");)
	1668	+ IF_EVENT(printk("Transmit Done Intr logic run\n");)
1669	1669	spin_lock_irqsave(&iadev->tx_lock, flags);
1670	1670	ia_tx_poll(iadev);
1671	1671	spin_unlock_irqrestore(&iadev->tx_lock, flags);

drivers/base/dma-coherent.c+3−2

@@ -312,8 +312,9 @@ static int __dma_mmap_from_coherent(struct dma_coherent_mem *mem, · 311 unmodified lines
312	312	* This checks whether the memory was allocated from the per-device
313	313	* coherent memory pool and if so, maps that memory to the provided vma.
314	314	*
315		− * Returns 1 if we correctly mapped the memory, or 0 if the caller should
316		− * proceed with mapping memory from generic pools.
	315	+ * Returns 1 if @vaddr belongs to the device coherent pool and the caller
	316	+ * should return @ret, or 0 if they should proceed with mapping memory from
	317	+ * generic areas.
317	318	*/
318	319	int dma_mmap_from_dev_coherent(struct device dev, struct vm_area_struct vma,
319	320	void vaddr, size_t size, int ret)

drivers/base/dma-mapping.c+2−4

@@ -226,7 +226,6 @@ int dma_common_mmap(struct device dev, struct vm_area_struct vma, · 225 unmodified lines
226	226	#ifndef CONFIG_ARCH_NO_COHERENT_DMA_MMAP
227	227	unsigned long user_count = vma_pages(vma);
228	228	unsigned long count = PAGE_ALIGN(size) >> PAGE_SHIFT;
229		− unsigned long pfn = page_to_pfn(virt_to_page(cpu_addr));
230	229	unsigned long off = vma->vm_pgoff;
231	230
232	231	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
@@ -234,12 +233,11 @@ int dma_common_mmap(struct device dev, struct vm_area_struct vma, · 1 unmodified line
234	233	if (dma_mmap_from_dev_coherent(dev, vma, cpu_addr, size, &ret))
235	234	return ret;
236	235
237		− if (off < count && user_count <= (count - off)) {
	236	+ if (off < count && user_count <= (count - off))
238	237	ret = remap_pfn_range(vma, vma->vm_start,
239		− pfn + off,
	238	+ page_to_pfn(virt_to_page(cpu_addr)) + off,
240	239	user_count << PAGE_SHIFT,
241	240	vma->vm_page_prot);
242		− }
243	241	#endif /* !CONFIG_ARCH_NO_COHERENT_DMA_MMAP */
244	242
245	243	return ret;

drivers/base/firmware_loader/fallback.c+2−2

@@ -537,8 +537,8 @@ exit: · 536 unmodified lines
537	537	}
538	538
539	539	/**
540		− * fw_load_sysfs_fallback - load a firmware via the syfs fallback mechanism
541		− * @fw_sysfs: firmware syfs information for the firmware to load
	540	+ * fw_load_sysfs_fallback - load a firmware via the sysfs fallback mechanism
	541	+ * @fw_sysfs: firmware sysfs information for the firmware to load
542	542	* @opt_flags: flags of options, FW_OPT_*
543	543	* @timeout: timeout to wait for the load
544	544	*

drivers/base/firmware_loader/fallback.h+1−1

@@ -6,7 +6,7 @@ · 5 unmodified lines
6	6	#include <linux/device.h>
7	7
8	8	/**
9		− * struct firmware_fallback_config - firmware fallback configuratioon settings
	9	+ * struct firmware_fallback_config - firmware fallback configuration settings
10	10	*
11	11	* Helps describe and fine tune the fallback mechanism.
12	12	*

drivers/block/loop.c+43−21

@@ -451,25 +451,47 @@ static int lo_req_flush(struct loop_device lo, struct request rq) · 450 unmodified lines
451	451	static void lo_complete_rq(struct request *rq)
452	452	{
453	453	struct loop_cmd *cmd = blk_mq_rq_to_pdu(rq);
	454	+ blk_status_t ret = BLK_STS_OK;
454	455
455		− if (unlikely(req_op(cmd->rq) == REQ_OP_READ && cmd->use_aio &&
456		− cmd->ret >= 0 && cmd->ret < blk_rq_bytes(cmd->rq))) {
457		− struct bio *bio = cmd->rq->bio;
458		−
459		− bio_advance(bio, cmd->ret);
460		− zero_fill_bio(bio);
	456	+ if (!cmd->use_aio \|\| cmd->ret < 0 \|\| cmd->ret == blk_rq_bytes(rq) \|\|
	457	+ req_op(rq) != REQ_OP_READ) {
	458	+ if (cmd->ret < 0)
	459	+ ret = BLK_STS_IOERR;
	460	+ goto end_io;
461	461	}
462	462
463		− blk_mq_end_request(rq, cmd->ret < 0 ? BLK_STS_IOERR : BLK_STS_OK);
	463	+ /*
	464	+ * Short READ - if we got some data, advance our request and
	465	+ * retry it. If we got no data, end the rest with EIO.
	466	+ */
	467	+ if (cmd->ret) {
	468	+ blk_update_request(rq, BLK_STS_OK, cmd->ret);
	469	+ cmd->ret = 0;
	470	+ blk_mq_requeue_request(rq, true);
	471	+ } else {
	472	+ if (cmd->use_aio) {
	473	+ struct bio *bio = rq->bio;
	474	+
	475	+ while (bio) {
	476	+ zero_fill_bio(bio);
	477	+ bio = bio->bi_next;
	478	+ }
	479	+ }
	480	+ ret = BLK_STS_IOERR;
	481	+end_io:
	482	+ blk_mq_end_request(rq, ret);
	483	+ }
464	484	}
465	485
466	486	static void lo_rw_aio_do_completion(struct loop_cmd *cmd)
467	487	{
	488	+ struct request *rq = blk_mq_rq_from_pdu(cmd);
	489	+
468	490	if (!atomic_dec_and_test(&cmd->ref))
469	491	return;
470	492	kfree(cmd->bvec);
471	493	cmd->bvec = NULL;
472		− blk_mq_complete_request(cmd->rq);
	494	+ blk_mq_complete_request(rq);
473	495	}
474	496
475	497	static void lo_rw_aio_complete(struct kiocb *iocb, long ret, long ret2)
@@ -487,7 +509,7 @@ static int lo_rw_aio(struct loop_device lo, struct loop_cmd cmd, · 11 unmodified lines
487	509	{
488	510	struct iov_iter iter;
489	511	struct bio_vec *bvec;
490		− struct request *rq = cmd->rq;
	512	+ struct request *rq = blk_mq_rq_from_pdu(cmd);
491	513	struct bio *bio = rq->bio;
492	514	struct file *file = lo->lo_backing_file;
493	515	unsigned int offset;
@@ -1702,15 +1724,16 @@ EXPORT_SYMBOL(loop_unregister_transfer); · 1208 unmodified lines
1702	1724	static blk_status_t loop_queue_rq(struct blk_mq_hw_ctx *hctx,
1703	1725	const struct blk_mq_queue_data *bd)
1704	1726	{
1705		− struct loop_cmd *cmd = blk_mq_rq_to_pdu(bd->rq);
1706		− struct loop_device *lo = cmd->rq->q->queuedata;
	1727	+ struct request *rq = bd->rq;
	1728	+ struct loop_cmd *cmd = blk_mq_rq_to_pdu(rq);
	1729	+ struct loop_device *lo = rq->q->queuedata;
1707	1730
1708		− blk_mq_start_request(bd->rq);
	1731	+ blk_mq_start_request(rq);
1709	1732
1710	1733	if (lo->lo_state != Lo_bound)
1711	1734	return BLK_STS_IOERR;
1712	1735
1713		− switch (req_op(cmd->rq)) {
	1736	+ switch (req_op(rq)) {
1714	1737	case REQ_OP_FLUSH:
1715	1738	case REQ_OP_DISCARD:
1716	1739	case REQ_OP_WRITE_ZEROES:
@@ -1723,8 +1746,8 @@ static blk_status_t loop_queue_rq(struct blk_mq_hw_ctx *hctx, · 6 unmodified lines
1723	1746
1724	1747	/* always use the first bio's css */
1725	1748	#ifdef CONFIG_BLK_CGROUP
1726		− if (cmd->use_aio && cmd->rq->bio && cmd->rq->bio->bi_css) {
1727		− cmd->css = cmd->rq->bio->bi_css;
	1749	+ if (cmd->use_aio && rq->bio && rq->bio->bi_css) {
	1750	+ cmd->css = rq->bio->bi_css;
1728	1751	css_get(cmd->css);
1729	1752	} else
1730	1753	#endif
@@ -1736,8 +1759,9 @@ static blk_status_t loop_queue_rq(struct blk_mq_hw_ctx *hctx, · 5 unmodified lines
1736	1759
1737	1760	static void loop_handle_cmd(struct loop_cmd *cmd)
1738	1761	{
1739		− const bool write = op_is_write(req_op(cmd->rq));
1740		− struct loop_device *lo = cmd->rq->q->queuedata;
	1762	+ struct request *rq = blk_mq_rq_from_pdu(cmd);
	1763	+ const bool write = op_is_write(req_op(rq));
	1764	+ struct loop_device *lo = rq->q->queuedata;
1741	1765	int ret = 0;
1742	1766
1743	1767	if (write && (lo->lo_flags & LO_FLAGS_READ_ONLY)) {
@@ -1745,12 +1769,12 @@ static void loop_handle_cmd(struct loop_cmd *cmd) · 1 unmodified line
1745	1769	goto failed;
1746	1770	}
1747	1771
1748		− ret = do_req_filebacked(lo, cmd->rq);
	1772	+ ret = do_req_filebacked(lo, rq);
1749	1773	failed:
1750	1774	/* complete non-aio request */
1751	1775	if (!cmd->use_aio \|\| ret) {
1752	1776	cmd->ret = ret ? -EIO : 0;
1753		− blk_mq_complete_request(cmd->rq);
	1777	+ blk_mq_complete_request(rq);
1754	1778	}
1755	1779	}
1756	1780
@@ -1767,9 +1791,7 @@ static int loop_init_request(struct blk_mq_tag_set set, struct request rq, · 10 unmodified lines
1767	1791	{
1768	1792	struct loop_cmd *cmd = blk_mq_rq_to_pdu(rq);
1769	1793
1770		− cmd->rq = rq;
1771	1794	kthread_init_work(&cmd->work, loop_queue_work);
1772		−
1773	1795	return 0;
1774	1796	}
1775	1797

drivers/block/loop.h+0−1

@@ -66,7 +66,6 @@ struct loop_device { · 65 unmodified lines
66	66
67	67	struct loop_cmd {
68	68	struct kthread_work work;
69		− struct request *rq;
70	69	bool use_aio; /* use AIO interface to handle I/O */
71	70	atomic_t ref; /* only for aio */
72	71	long ret;

drivers/block/rbd.c+69−32

@@ -732,6 +732,7 @@ static struct rbd_client rbd_client_find(struct ceph_options ceph_opts) · 731 unmodified lines
732	732	*/
733	733	enum {
734	734	Opt_queue_depth,
	735	+ Opt_lock_timeout,
735	736	Opt_last_int,
736	737	/* int args above */
737	738	Opt_last_string,
@@ -740,11 +741,13 @@ enum { · 2 unmodified lines
740	741	Opt_read_write,
741	742	Opt_lock_on_read,
742	743	Opt_exclusive,
	744	+ Opt_notrim,
743	745	Opt_err
744	746	};
745	747
746	748	static match_table_t rbd_opts_tokens = {
747	749	{Opt_queue_depth, "queue_depth=%d"},
	750	+ {Opt_lock_timeout, "lock_timeout=%d"},
748	751	/* int args above */
749	752	/* string args above */
750	753	{Opt_read_only, "read_only"},
@@ -753,20 +756,25 @@ static match_table_t rbd_opts_tokens = { · 2 unmodified lines
753	756	{Opt_read_write, "rw"}, /* Alternate spelling */
754	757	{Opt_lock_on_read, "lock_on_read"},
755	758	{Opt_exclusive, "exclusive"},
	759	+ {Opt_notrim, "notrim"},
756	760	{Opt_err, NULL}
757	761	};
758	762
759	763	struct rbd_options {
760	764	int queue_depth;
	765	+ unsigned long lock_timeout;
761	766	bool read_only;
762	767	bool lock_on_read;
763	768	bool exclusive;
	769	+ bool trim;
764	770	};
765	771
766	772	#define RBD_QUEUE_DEPTH_DEFAULT BLKDEV_MAX_RQ
	773	+#define RBD_LOCK_TIMEOUT_DEFAULT 0 /* no timeout */
767	774	#define RBD_READ_ONLY_DEFAULT false
768	775	#define RBD_LOCK_ON_READ_DEFAULT false
769	776	#define RBD_EXCLUSIVE_DEFAULT false
	777	+#define RBD_TRIM_DEFAULT true
770	778
771	779	static int parse_rbd_opts_token(char c, void private)
772	780	{
@@ -796,6 +804,14 @@ static int parse_rbd_opts_token(char c, void private) · 23 unmodified lines
796	804	}
797	805	rbd_opts->queue_depth = intval;
798	806	break;
	807	+ case Opt_lock_timeout:
	808	+ /* 0 is "wait forever" (i.e. infinite timeout) */
	809	+ if (intval < 0 \|\| intval > INT_MAX / 1000) {
	810	+ pr_err("lock_timeout out of range\n");
	811	+ return -EINVAL;
	812	+ }
	813	+ rbd_opts->lock_timeout = msecs_to_jiffies(intval * 1000);
	814	+ break;
799	815	case Opt_read_only:
800	816	rbd_opts->read_only = true;
801	817	break;
@@ -808,6 +824,9 @@ static int parse_rbd_opts_token(char c, void private) · 6 unmodified lines
808	824	case Opt_exclusive:
809	825	rbd_opts->exclusive = true;
810	826	break;
	827	+ case Opt_notrim:
	828	+ rbd_opts->trim = false;
	829	+ break;
811	830	default:
812	831	/* libceph prints "bad option" msg */
813	832	return -EINVAL;
@@ -1392,7 +1411,7 @@ static bool rbd_img_is_write(struct rbd_img_request *img_req) · 578 unmodified lines
1392	1411	case OBJ_OP_DISCARD:
1393	1412	return true;
1394	1413	default:
1395		− rbd_assert(0);
	1414	+ BUG();
1396	1415	}
1397	1416	}
1398	1417
@@ -2466,7 +2485,7 @@ again: · 1067 unmodified lines
2466	2485	}
2467	2486	return false;
2468	2487	default:
2469		− rbd_assert(0);
	2488	+ BUG();
2470	2489	}
2471	2490	}
2472	2491
@@ -2494,7 +2513,7 @@ static bool __rbd_obj_handle_request(struct rbd_obj_request *obj_req) · 21 unmodified lines
2494	2513	}
2495	2514	return false;
2496	2515	default:
2497		− rbd_assert(0);
	2516	+ BUG();
2498	2517	}
2499	2518	}
2500	2519
@@ -3533,9 +3552,22 @@ static int rbd_obj_method_sync(struct rbd_device *rbd_dev, · 1032 unmodified lines
3533	3552	/*
3534	3553	* lock_rwsem must be held for read
3535	3554	*/
3536		−static void rbd_wait_state_locked(struct rbd_device *rbd_dev)
	3555	+static int rbd_wait_state_locked(struct rbd_device *rbd_dev, bool may_acquire)
3537	3556	{
3538	3557	DEFINE_WAIT(wait);
	3558	+ unsigned long timeout;
	3559	+ int ret = 0;
	3560	+
	3561	+ if (test_bit(RBD_DEV_FLAG_BLACKLISTED, &rbd_dev->flags))
	3562	+ return -EBLACKLISTED;
	3563	+
	3564	+ if (rbd_dev->lock_state == RBD_LOCK_STATE_LOCKED)
	3565	+ return 0;
	3566	+
	3567	+ if (!may_acquire) {
	3568	+ rbd_warn(rbd_dev, "exclusive lock required");
	3569	+ return -EROFS;
	3570	+ }
3539	3571
3540	3572	do {
3541	3573	/*
@@ -3547,12 +3579,22 @@ static void rbd_wait_state_locked(struct rbd_device *rbd_dev) · 5 unmodified lines
3547	3579	prepare_to_wait_exclusive(&rbd_dev->lock_waitq, &wait,
3548	3580	TASK_UNINTERRUPTIBLE);
3549	3581	up_read(&rbd_dev->lock_rwsem);
3550		− schedule();
	3582	+ timeout = schedule_timeout(ceph_timeout_jiffies(
	3583	+ rbd_dev->opts->lock_timeout));
3551	3584	down_read(&rbd_dev->lock_rwsem);
3552		− } while (rbd_dev->lock_state != RBD_LOCK_STATE_LOCKED &&
3553		− !test_bit(RBD_DEV_FLAG_BLACKLISTED, &rbd_dev->flags));
	3585	+ if (test_bit(RBD_DEV_FLAG_BLACKLISTED, &rbd_dev->flags)) {
	3586	+ ret = -EBLACKLISTED;
	3587	+ break;
	3588	+ }
	3589	+ if (!timeout) {
	3590	+ rbd_warn(rbd_dev, "timed out waiting for lock");
	3591	+ ret = -ETIMEDOUT;
	3592	+ break;
	3593	+ }
	3594	+ } while (rbd_dev->lock_state != RBD_LOCK_STATE_LOCKED);
3554	3595
3555	3596	finish_wait(&rbd_dev->lock_waitq, &wait);
	3597	+ return ret;
3556	3598	}
3557	3599
3558	3600	static void rbd_queue_workfn(struct work_struct *work)
@@ -3638,19 +3680,10 @@ static void rbd_queue_workfn(struct work_struct *work) · 79 unmodified lines
3638	3680	(op_type != OBJ_OP_READ \|\| rbd_dev->opts->lock_on_read);
3639	3681	if (must_be_locked) {
3640	3682	down_read(&rbd_dev->lock_rwsem);
3641		− if (rbd_dev->lock_state != RBD_LOCK_STATE_LOCKED &&
3642		− !test_bit(RBD_DEV_FLAG_BLACKLISTED, &rbd_dev->flags)) {
3643		− if (rbd_dev->opts->exclusive) {
3644		− rbd_warn(rbd_dev, "exclusive lock required");
3645		− result = -EROFS;
3646		− goto err_unlock;
3647		− }
3648		− rbd_wait_state_locked(rbd_dev);
3649		− }
3650		− if (test_bit(RBD_DEV_FLAG_BLACKLISTED, &rbd_dev->flags)) {
3651		− result = -EBLACKLISTED;
	3683	+ result = rbd_wait_state_locked(rbd_dev,
	3684	+ !rbd_dev->opts->exclusive);
	3685	+ if (result)
3652	3686	goto err_unlock;
3653		− }
3654	3687	}
3655	3688
3656	3689	img_request = rbd_img_request_create(rbd_dev, op_type, snapc);
@@ -3902,7 +3935,8 @@ static int rbd_init_disk(struct rbd_device *rbd_dev) · 245 unmodified lines
3902	3935	{
3903	3936	struct gendisk *disk;
3904	3937	struct request_queue *q;
3905		− u64 segment_size;
	3938	+ unsigned int objset_bytes =
	3939	+ rbd_dev->layout.object_size * rbd_dev->layout.stripe_count;
3906	3940	int err;
3907	3941
3908	3942	/* create gendisk info */
@@ -3942,20 +3976,19 @@ static int rbd_init_disk(struct rbd_device *rbd_dev) · 33 unmodified lines
3942	3976	blk_queue_flag_set(QUEUE_FLAG_NONROT, q);
3943	3977	/* QUEUE_FLAG_ADD_RANDOM is off by default for blk-mq */
3944	3978
3945		− /* set io sizes to object size */
3946		− segment_size = rbd_obj_bytes(&rbd_dev->header);
3947		− blk_queue_max_hw_sectors(q, segment_size / SECTOR_SIZE);
	3979	+ blk_queue_max_hw_sectors(q, objset_bytes >> SECTOR_SHIFT);
3948	3980	q->limits.max_sectors = queue_max_hw_sectors(q);
3949	3981	blk_queue_max_segments(q, USHRT_MAX);
3950	3982	blk_queue_max_segment_size(q, UINT_MAX);
3951		− blk_queue_io_min(q, segment_size);
3952		− blk_queue_io_opt(q, segment_size);
	3983	+ blk_queue_io_min(q, objset_bytes);
	3984	+ blk_queue_io_opt(q, objset_bytes);
3953	3985
3954		− /* enable the discard support */
3955		− blk_queue_flag_set(QUEUE_FLAG_DISCARD, q);
3956		− q->limits.discard_granularity = segment_size;
3957		− blk_queue_max_discard_sectors(q, segment_size / SECTOR_SIZE);
3958		− blk_queue_max_write_zeroes_sectors(q, segment_size / SECTOR_SIZE);
	3986	+ if (rbd_dev->opts->trim) {
	3987	+ blk_queue_flag_set(QUEUE_FLAG_DISCARD, q);
	3988	+ q->limits.discard_granularity = objset_bytes;
	3989	+ blk_queue_max_discard_sectors(q, objset_bytes >> SECTOR_SHIFT);
	3990	+ blk_queue_max_write_zeroes_sectors(q, objset_bytes >> SECTOR_SHIFT);
	3991	+ }
3959	3992
3960	3993	if (!ceph_test_opt(rbd_dev->rbd_client->client, NOCRC))
3961	3994	q->backing_dev_info->capabilities \|= BDI_CAP_STABLE_WRITES;
@@ -5179,8 +5212,10 @@ static int rbd_add_parse_args(const char *buf, · 1217 unmodified lines
5179	5212
5180	5213	rbd_opts->read_only = RBD_READ_ONLY_DEFAULT;
5181	5214	rbd_opts->queue_depth = RBD_QUEUE_DEPTH_DEFAULT;
	5215	+ rbd_opts->lock_timeout = RBD_LOCK_TIMEOUT_DEFAULT;
5182	5216	rbd_opts->lock_on_read = RBD_LOCK_ON_READ_DEFAULT;
5183	5217	rbd_opts->exclusive = RBD_EXCLUSIVE_DEFAULT;
	5218	+ rbd_opts->trim = RBD_TRIM_DEFAULT;
5184	5219
5185	5220	copts = ceph_parse_options(options, mon_addrs,
5186	5221	mon_addrs + mon_addrs_size - 1,
@@ -5216,6 +5251,8 @@ static void rbd_dev_image_unlock(struct rbd_device *rbd_dev) · 29 unmodified lines
5216	5251
5217	5252	static int rbd_add_acquire_lock(struct rbd_device *rbd_dev)
5218	5253	{
	5254	+ int ret;
	5255	+
5219	5256	if (!(rbd_dev->header.features & RBD_FEATURE_EXCLUSIVE_LOCK)) {
5220	5257	rbd_warn(rbd_dev, "exclusive-lock feature is not enabled");
5221	5258	return -EINVAL;
@@ -5223,9 +5260,9 @@ static int rbd_add_acquire_lock(struct rbd_device *rbd_dev) · 1 unmodified line
5223	5260
5224	5261	/* FIXME: "rbd map --exclusive" should be in interruptible */
5225	5262	down_read(&rbd_dev->lock_rwsem);
5226		− rbd_wait_state_locked(rbd_dev);
	5263	+ ret = rbd_wait_state_locked(rbd_dev, true);
5227	5264	up_read(&rbd_dev->lock_rwsem);
5228		− if (test_bit(RBD_DEV_FLAG_BLACKLISTED, &rbd_dev->flags)) {
	5265	+ if (ret) {
5229	5266	rbd_warn(rbd_dev, "failed to acquire exclusive lock");
5230	5267	return -EROFS;
5231	5268	}

drivers/block/swim.c+22−27

@@ -110,7 +110,7 @@ struct iwm { · 109 unmodified lines
110	110	/* Select values for swim_select and swim_readbit */
111	111
112	112	#define READ_DATA_0 0x074
113		−#define TWOMEG_DRIVE 0x075
	113	+#define ONEMEG_DRIVE 0x075
114	114	#define SINGLE_SIDED 0x076
115	115	#define DRIVE_PRESENT 0x077
116	116	#define DISK_IN 0x170
@@ -118,9 +118,9 @@ struct iwm { · 1 unmodified line
118	118	#define TRACK_ZERO 0x172
119	119	#define TACHO 0x173
120	120	#define READ_DATA_1 0x174
121		−#define MFM_MODE 0x175
	121	+#define GCR_MODE 0x175
122	122	#define SEEK_COMPLETE 0x176
123		−#define ONEMEG_MEDIA 0x177
	123	+#define TWOMEG_MEDIA 0x177
124	124
125	125	/* Bits in handshake register */
126	126
@@ -612,7 +612,6 @@ static void setup_medium(struct floppy_state *fs) · 485 unmodified lines
612	612	struct floppy_struct *g;
613	613	fs->disk_in = 1;
614	614	fs->write_protected = swim_readbit(base, WRITE_PROT);
615		− fs->type = swim_readbit(base, ONEMEG_MEDIA);
616	615
617	616	if (swim_track00(base))
618	617	printk(KERN_ERR
@@ -620,6 +619,9 @@ static void setup_medium(struct floppy_state *fs) · 1 unmodified line
620	619
621	620	swim_track00(base);
622	621
	622	+ fs->type = swim_readbit(base, TWOMEG_MEDIA) ?
	623	+ HD_MEDIA : DD_MEDIA;
	624	+ fs->head_number = swim_readbit(base, SINGLE_SIDED) ? 1 : 2;
623	625	get_floppy_geometry(fs, 0, &g);
624	626	fs->total_secs = g->size;
625	627	fs->secpercyl = g->head * g->sect;
@@ -646,7 +648,7 @@ static int floppy_open(struct block_device *bdev, fmode_t mode) · 20 unmodified lines
646	648
647	649	swim_write(base, setup, S_IBM_DRIVE \| S_FCLK_DIV2);
648	650	udelay(10);
649		− swim_drive(base, INTERNAL_DRIVE);
	651	+ swim_drive(base, fs->location);
650	652	swim_motor(base, ON);
651	653	swim_action(base, SETMFM);
652	654	if (fs->ejected)
@@ -656,6 +658,8 @@ static int floppy_open(struct block_device *bdev, fmode_t mode) · 3 unmodified lines
656	658	goto out;
657	659	}
658	660
	661	+ set_capacity(fs->disk, fs->total_secs);
	662	+
659	663	if (mode & FMODE_NDELAY)
660	664	return 0;
661	665
@@ -727,14 +731,9 @@ static int floppy_ioctl(struct block_device *bdev, fmode_t mode, · 65 unmodified lines
727	731	if (copy_to_user((void __user ) param, (void ) &floppy_type,
728	732	sizeof(struct floppy_struct)))
729	733	return -EFAULT;
730		− break;
731		−
732		− default:
733		− printk(KERN_DEBUG "SWIM floppy_ioctl: unknown cmd %d\n",
734		− cmd);
735		− return -ENOSYS;
	734	+ return 0;
736	735	}
737		− return 0;
	736	+ return -ENOTTY;
738	737	}
739	738
740	739	static int floppy_getgeo(struct block_device bdev, struct hd_geometry geo)
@@ -795,7 +794,7 @@ static struct kobject floppy_find(dev_t dev, int part, void *data) · 54 unmodified lines
795	794	struct swim_priv *swd = data;
796	795	int drive = (*part & 3);
797	796
798		− if (drive > swd->floppy_count)
	797	+ if (drive >= swd->floppy_count)
799	798	return NULL;
800	799
801	800	*part = 0;
@@ -813,10 +812,9 @@ static int swim_add_floppy(struct swim_priv *swd, enum drive_location location) · 11 unmodified lines
813	812
814	813	swim_motor(base, OFF);
815	814
816		− if (swim_readbit(base, SINGLE_SIDED))
817		− fs->head_number = 1;
818		− else
819		− fs->head_number = 2;
	815	+ fs->type = HD_MEDIA;
	816	+ fs->head_number = 2;
	817	+
820	818	fs->ref_count = 0;
821	819	fs->ejected = 1;
822	820
@@ -834,10 +832,12 @@ static int swim_floppy_init(struct swim_priv *swd) · 11 unmodified lines
834	832	/* scan floppy drives */
835	833
836	834	swim_drive(base, INTERNAL_DRIVE);
837		− if (swim_readbit(base, DRIVE_PRESENT))
	835	+ if (swim_readbit(base, DRIVE_PRESENT) &&
	836	+ !swim_readbit(base, ONEMEG_DRIVE))
838	837	swim_add_floppy(swd, INTERNAL_DRIVE);
839	838	swim_drive(base, EXTERNAL_DRIVE);
840		− if (swim_readbit(base, DRIVE_PRESENT))
	839	+ if (swim_readbit(base, DRIVE_PRESENT) &&
	840	+ !swim_readbit(base, ONEMEG_DRIVE))
841	841	swim_add_floppy(swd, EXTERNAL_DRIVE);
842	842
843	843	/* register floppy drives */
@@ -861,7 +861,6 @@ static int swim_floppy_init(struct swim_priv *swd) · 17 unmodified lines
861	861	&swd->lock);
862	862	if (!swd->unit[drive].disk->queue) {
863	863	err = -ENOMEM;
864		− put_disk(swd->unit[drive].disk);
865	864	goto exit_put_disks;
866	865	}
867	866	blk_queue_bounce_limit(swd->unit[drive].disk->queue,
@@ -911,7 +910,7 @@ static int swim_probe(struct platform_device *dev) · 43 unmodified lines
911	910	goto out;
912	911	}
913	912
914		− swim_base = ioremap(res->start, resource_size(res));
	913	+ swim_base = (struct swim __iomem *)res->start;
915	914	if (!swim_base) {
916	915	ret = -ENOMEM;
917	916	goto out_release_io;
@@ -923,7 +922,7 @@ static int swim_probe(struct platform_device *dev) · 5 unmodified lines
923	922	if (!get_swim_mode(swim_base)) {
924	923	printk(KERN_INFO "SWIM device not found !\n");
925	924	ret = -ENODEV;
926		− goto out_iounmap;
	925	+ goto out_release_io;
927	926	}
928	927
929	928	/* set platform driver data */
@@ -931,7 +930,7 @@ static int swim_probe(struct platform_device *dev) · 1 unmodified line
931	930	swd = kzalloc(sizeof(struct swim_priv), GFP_KERNEL);
932	931	if (!swd) {
933	932	ret = -ENOMEM;
934		− goto out_iounmap;
	933	+ goto out_release_io;
935	934	}
936	935	platform_set_drvdata(dev, swd);
937	936
@@ -945,8 +944,6 @@ static int swim_probe(struct platform_device *dev) · 7 unmodified lines
945	944
946	945	out_kfree:
947	946	kfree(swd);
948		−out_iounmap:
949		− iounmap(swim_base);
950	947	out_release_io:
951	948	release_mem_region(res->start, resource_size(res));
952	949	out:
@@ -974,8 +971,6 @@ static int swim_remove(struct platform_device *dev) · 21 unmodified lines
974	971	for (drive = 0; drive < swd->floppy_count; drive++)
975	972	floppy_eject(&swd->unit[drive]);
976	973
977		− iounmap(swd->base);
978		−
979	974	res = platform_get_resource(dev, IORESOURCE_MEM, 0);
980	975	if (res)
981	976	release_mem_region(res->start, resource_size(res));

drivers/block/swim3.c+3−3

@@ -148,7 +148,7 @@ struct swim3 { · 147 unmodified lines
148	148	#define MOTOR_ON 2
149	149	#define RELAX 3 /* also eject in progress */
150	150	#define READ_DATA_0 4
151		−#define TWOMEG_DRIVE 5
	151	+#define ONEMEG_DRIVE 5
152	152	#define SINGLE_SIDED 6 /* drive or diskette is 4MB type? */
153	153	#define DRIVE_PRESENT 7
154	154	#define DISK_IN 8
@@ -156,9 +156,9 @@ struct swim3 { · 1 unmodified line
156	156	#define TRACK_ZERO 10
157	157	#define TACHO 11
158	158	#define READ_DATA_1 12
159		−#define MFM_MODE 13
	159	+#define GCR_MODE 13
160	160	#define SEEK_COMPLETE 14
161		−#define ONEMEG_MEDIA 15
	161	+#define TWOMEG_MEDIA 15
162	162
163	163	/* Definitions of values used in writing and formatting */
164	164	#define DATA_ESCAPE 0x99

drivers/bus/Kconfig+1−0

@@ -33,6 +33,7 @@ config HISILICON_LPC · 32 unmodified lines
33	33	bool "Support for ISA I/O space on HiSilicon Hip06/7"
34	34	depends on ARM64 && (ARCH_HISI \|\| COMPILE_TEST)
35	35	select INDIRECT_PIO
	36	+ select MFD_CORE if ACPI
36	37	help
37	38	Driver to enable I/O access to devices attached to the Low Pin
38	39	Count bus on the HiSilicon Hip06/7 SoC.

drivers/cdrom/cdrom.c+1−1

@@ -2371,7 +2371,7 @@ static int cdrom_ioctl_media_changed(struct cdrom_device_info *cdi, · 2370 unmodified lines
2371	2371	if (!CDROM_CAN(CDC_SELECT_DISC) \|\| arg == CDSL_CURRENT)
2372	2372	return media_changed(cdi, 1);
2373	2373
2374		− if ((unsigned int)arg >= cdi->capacity)
	2374	+ if (arg >= cdi->capacity)
2375	2375	return -EINVAL;
2376	2376
2377	2377	info = kmalloc(sizeof(*info), GFP_KERNEL);

drivers/char/random.c+137−35

@@ -261,6 +261,7 @@ · 260 unmodified lines
261	261	#include <linux/ptrace.h>
262	262	#include <linux/workqueue.h>
263	263	#include <linux/irq.h>
	264	+#include <linux/ratelimit.h>
264	265	#include <linux/syscalls.h>
265	266	#include <linux/completion.h>
266	267	#include <linux/uuid.h>
@@ -427,8 +428,9 @@ struct crng_state primary_crng = { · 160 unmodified lines
427	428	* its value (from 0->1->2).
428	429	*/
429	430	static int crng_init = 0;
430		−#define crng_ready() (likely(crng_init > 0))
	431	+#define crng_ready() (likely(crng_init > 1))
431	432	static int crng_init_cnt = 0;
	433	+static unsigned long crng_global_init_time = 0;
432	434	#define CRNG_INIT_CNT_THRESH (2*CHACHA20_KEY_SIZE)
433	435	static void _extract_crng(struct crng_state *crng,
434	436	__u32 out[CHACHA20_BLOCK_WORDS]);
@@ -437,6 +439,16 @@ static void _crng_backtrack_protect(struct crng_state *crng, · 2 unmodified lines
437	439	static void process_random_ready_list(void);
438	440	static void _get_random_bytes(void *buf, int nbytes);
439	441
	442	+static struct ratelimit_state unseeded_warning =
	443	+ RATELIMIT_STATE_INIT("warn_unseeded_randomness", HZ, 3);
	444	+static struct ratelimit_state urandom_warning =
	445	+ RATELIMIT_STATE_INIT("warn_urandom_randomness", HZ, 3);
	446	+
	447	+static int ratelimit_disable __read_mostly;
	448	+
	449	+module_param_named(ratelimit_disable, ratelimit_disable, int, 0644);
	450	+MODULE_PARM_DESC(ratelimit_disable, "Disable random ratelimit suppression");
	451	+
440	452	/**********************************************************************
441	453	*
442	454	* OS independent entropy store. Here are the functions which handle
@@ -787,6 +799,43 @@ static void crng_initialize(struct crng_state *crng) · 344 unmodified lines
787	799	crng->init_time = jiffies - CRNG_RESEED_INTERVAL - 1;
788	800	}
789	801
	802	+#ifdef CONFIG_NUMA
	803	+static void do_numa_crng_init(struct work_struct *work)
	804	+{
	805	+ int i;
	806	+ struct crng_state *crng;
	807	+ struct crng_state **pool;
	808	+
	809	+ pool = kcalloc(nr_node_ids, sizeof(*pool), GFP_KERNEL\|__GFP_NOFAIL);
	810	+ for_each_online_node(i) {
	811	+ crng = kmalloc_node(sizeof(struct crng_state),
	812	+ GFP_KERNEL \| __GFP_NOFAIL, i);
	813	+ spin_lock_init(&crng->lock);
	814	+ crng_initialize(crng);
	815	+ pool[i] = crng;
	816	+ }
	817	+ mb();
	818	+ if (cmpxchg(&crng_node_pool, NULL, pool)) {
	819	+ for_each_node(i)
	820	+ kfree(pool[i]);
	821	+ kfree(pool);
	822	+ }
	823	+}
	824	+
	825	+static DECLARE_WORK(numa_crng_init_work, do_numa_crng_init);
	826	+
	827	+static void numa_crng_init(void)
	828	+{
	829	+ schedule_work(&numa_crng_init_work);
	830	+}
	831	+#else
	832	+static void numa_crng_init(void) {}
	833	+#endif
	834	+
	835	+/*
	836	+ * crng_fast_load() can be called by code in the interrupt service
	837	+ * path. So we can't afford to dilly-dally.
	838	+ */
790	839	static int crng_fast_load(const char *cp, size_t len)
791	840	{
792	841	unsigned long flags;
@@ -794,7 +843,7 @@ static int crng_fast_load(const char *cp, size_t len) · 1 unmodified line
794	843
795	844	if (!spin_trylock_irqsave(&primary_crng.lock, flags))
796	845	return 0;
797		− if (crng_ready()) {
	846	+ if (crng_init != 0) {
798	847	spin_unlock_irqrestore(&primary_crng.lock, flags);
799	848	return 0;
800	849	}
@@ -813,6 +862,51 @@ static int crng_fast_load(const char *cp, size_t len) · 12 unmodified lines
813	862	return 1;
814	863	}
815	864
	865	+/*
	866	+ * crng_slow_load() is called by add_device_randomness, which has two
	867	+ * attributes. (1) We can't trust the buffer passed to it is
	868	+ * guaranteed to be unpredictable (so it might not have any entropy at
	869	+ * all), and (2) it doesn't have the performance constraints of
	870	+ * crng_fast_load().
	871	+ *
	872	+ * So we do something more comprehensive which is guaranteed to touch
	873	+ * all of the primary_crng's state, and which uses a LFSR with a
	874	+ * period of 255 as part of the mixing algorithm. Finally, we do
	875	+ * not advance crng_init_cnt since buffer we may get may be something
	876	+ * like a fixed DMI table (for example), which might very well be
	877	+ * unique to the machine, but is otherwise unvarying.
	878	+ */
	879	+static int crng_slow_load(const char *cp, size_t len)
	880	+{
	881	+ unsigned long flags;
	882	+ static unsigned char lfsr = 1;
	883	+ unsigned char tmp;
	884	+ unsigned i, max = CHACHA20_KEY_SIZE;
	885	+ const char * src_buf = cp;
	886	+ char * dest_buf = (char *) &primary_crng.state[4];
	887	+
	888	+ if (!spin_trylock_irqsave(&primary_crng.lock, flags))
	889	+ return 0;
	890	+ if (crng_init != 0) {
	891	+ spin_unlock_irqrestore(&primary_crng.lock, flags);
	892	+ return 0;
	893	+ }
	894	+ if (len > max)
	895	+ max = len;
	896	+
	897	+ for (i = 0; i < max ; i++) {
	898	+ tmp = lfsr;
	899	+ lfsr >>= 1;
	900	+ if (tmp & 1)
	901	+ lfsr ^= 0xE1;
	902	+ tmp = dest_buf[i % CHACHA20_KEY_SIZE];
	903	+ dest_buf[i % CHACHA20_KEY_SIZE] ^= src_buf[i % len] ^ lfsr;
	904	+ lfsr += (tmp << 3) \| (tmp >> 5);
	905	+ }
	906	+ spin_unlock_irqrestore(&primary_crng.lock, flags);
	907	+ return 1;
	908	+}
	909	+
816	910	static void crng_reseed(struct crng_state crng, struct entropy_store r)
817	911	{
818	912	unsigned long flags;
@@ -831,7 +925,7 @@ static void crng_reseed(struct crng_state crng, struct entropy_store r) · 12 unmodified lines
831	925	_crng_backtrack_protect(&primary_crng, buf.block,
832	926	CHACHA20_KEY_SIZE);
833	927	}
834		− spin_lock_irqsave(&primary_crng.lock, flags);
	928	+ spin_lock_irqsave(&crng->lock, flags);
835	929	for (i = 0; i < 8; i++) {
836	930	unsigned long rv;
837	931	if (!arch_get_random_seed_long(&rv) &&
@@ -841,13 +935,26 @@ static void crng_reseed(struct crng_state crng, struct entropy_store r) · 3 unmodified lines
841	935	}
842	936	memzero_explicit(&buf, sizeof(buf));
843	937	crng->init_time = jiffies;
844		− spin_unlock_irqrestore(&primary_crng.lock, flags);
	938	+ spin_unlock_irqrestore(&crng->lock, flags);
845	939	if (crng == &primary_crng && crng_init < 2) {
846	940	invalidate_batched_entropy();
	941	+ numa_crng_init();
847	942	crng_init = 2;
848	943	process_random_ready_list();
849	944	wake_up_interruptible(&crng_init_wait);
850	945	pr_notice("random: crng init done\n");
	946	+ if (unseeded_warning.missed) {
	947	+ pr_notice("random: %d get_random_xx warning(s) missed "
	948	+ "due to ratelimiting\n",
	949	+ unseeded_warning.missed);
	950	+ unseeded_warning.missed = 0;
	951	+ }
	952	+ if (urandom_warning.missed) {
	953	+ pr_notice("random: %d urandom warning(s) missed "
	954	+ "due to ratelimiting\n",
	955	+ urandom_warning.missed);
	956	+ urandom_warning.missed = 0;
	957	+ }
851	958	}
852	959	}
853	960
@@ -856,8 +963,9 @@ static void _extract_crng(struct crng_state *crng, · 2 unmodified lines
856	963	{
857	964	unsigned long v, flags;
858	965
859		− if (crng_init > 1 &&
860		− time_after(jiffies, crng->init_time + CRNG_RESEED_INTERVAL))
	966	+ if (crng_ready() &&
	967	+ (time_after(crng_global_init_time, crng->init_time) \|\|
	968	+ time_after(jiffies, crng->init_time + CRNG_RESEED_INTERVAL)))
861	969	crng_reseed(crng, crng == &primary_crng ? &input_pool : NULL);
862	970	spin_lock_irqsave(&crng->lock, flags);
863	971	if (arch_get_random_long(&v))
@@ -981,10 +1089,8 @@ void add_device_randomness(const void *buf, unsigned int size) · 117 unmodified lines
981	1089	unsigned long time = random_get_entropy() ^ jiffies;
982	1090	unsigned long flags;
983	1091
984		− if (!crng_ready()) {
985		− crng_fast_load(buf, size);
986		− return;
987		− }
	1092	+ if (!crng_ready() && size)
	1093	+ crng_slow_load(buf, size);
988	1094
989	1095	trace_add_device_randomness(size, _RET_IP_);
990	1096	spin_lock_irqsave(&input_pool.lock, flags);
@@ -1139,7 +1245,7 @@ void add_interrupt_randomness(int irq, int irq_flags) · 148 unmodified lines
1139	1245	fast_mix(fast_pool);
1140	1246	add_interrupt_bench(cycles);
1141	1247
1142		− if (!crng_ready()) {
	1248	+ if (unlikely(crng_init == 0)) {
1143	1249	if ((fast_pool->count >= 64) &&
1144	1250	crng_fast_load((char *) fast_pool->pool,
1145	1251	sizeof(fast_pool->pool))) {
@@ -1489,8 +1595,9 @@ static void _warn_unseeded_randomness(const char func_name, void caller, · 343 unmodified lines
1489	1595	#ifndef CONFIG_WARN_ALL_UNSEEDED_RANDOM
1490	1596	print_once = true;
1491	1597	#endif
1492		− pr_notice("random: %s called from %pS with crng_init=%d\n",
1493		− func_name, caller, crng_init);
	1598	+ if (__ratelimit(&unseeded_warning))
	1599	+ pr_notice("random: %s called from %pS with crng_init=%d\n",
	1600	+ func_name, caller, crng_init);
1494	1601	}
1495	1602
1496	1603	/*
@@ -1680,28 +1787,14 @@ static void init_std_data(struct entropy_store *r) · 183 unmodified lines
1680	1787	*/
1681	1788	static int rand_initialize(void)
1682	1789	{
1683		−#ifdef CONFIG_NUMA
1684		− int i;
1685		− struct crng_state *crng;
1686		− struct crng_state **pool;
1687		−#endif
1688		−
1689	1790	init_std_data(&input_pool);
1690	1791	init_std_data(&blocking_pool);
1691	1792	crng_initialize(&primary_crng);
1692		−
1693		−#ifdef CONFIG_NUMA
1694		− pool = kcalloc(nr_node_ids, sizeof(*pool), GFP_KERNEL\|__GFP_NOFAIL);
1695		− for_each_online_node(i) {
1696		− crng = kmalloc_node(sizeof(struct crng_state),
1697		− GFP_KERNEL \| __GFP_NOFAIL, i);
1698		− spin_lock_init(&crng->lock);
1699		− crng_initialize(crng);
1700		− pool[i] = crng;
	1793	+ crng_global_init_time = jiffies;
	1794	+ if (ratelimit_disable) {
	1795	+ urandom_warning.interval = 0;
	1796	+ unseeded_warning.interval = 0;
1701	1797	}
1702		− mb();
1703		− crng_node_pool = pool;
1704		−#endif
1705	1798	return 0;
1706	1799	}
1707	1800	early_initcall(rand_initialize);
@@ -1769,9 +1862,10 @@ urandom_read(struct file file, char __user buf, size_t nbytes, loff_t *ppos) · 61 unmodified lines
1769	1862
1770	1863	if (!crng_ready() && maxwarn > 0) {
1771	1864	maxwarn--;
1772		− printk(KERN_NOTICE "random: %s: uninitialized urandom read "
1773		− "(%zd bytes read)\n",
1774		− current->comm, nbytes);
	1865	+ if (__ratelimit(&urandom_warning))
	1866	+ printk(KERN_NOTICE "random: %s: uninitialized "
	1867	+ "urandom read (%zd bytes read)\n",
	1868	+ current->comm, nbytes);
1775	1869	spin_lock_irqsave(&primary_crng.lock, flags);
1776	1870	crng_init_cnt = 0;
1777	1871	spin_unlock_irqrestore(&primary_crng.lock, flags);
@@ -1875,6 +1969,14 @@ static long random_ioctl(struct file *f, unsigned int cmd, unsigned long arg) · 97 unmodified lines
1875	1969	input_pool.entropy_count = 0;
1876	1970	blocking_pool.entropy_count = 0;
1877	1971	return 0;
	1972	+ case RNDRESEEDCRNG:
	1973	+ if (!capable(CAP_SYS_ADMIN))
	1974	+ return -EPERM;
	1975	+ if (crng_init < 2)
	1976	+ return -ENODATA;
	1977	+ crng_reseed(&primary_crng, NULL);
	1978	+ crng_global_init_time = jiffies - 1;
	1979	+ return 0;
1878	1980	default:
1879	1981	return -EINVAL;
1880	1982	}
@@ -2212,7 +2314,7 @@ void add_hwgenerator_randomness(const char *buffer, size_t count, · 331 unmodified lines
2212	2314	{
2213	2315	struct entropy_store *poolp = &input_pool;
2214	2316
2215		− if (!crng_ready()) {
	2317	+ if (unlikely(crng_init == 0)) {
2216	2318	crng_fast_load(buffer, count);
2217	2319	return;
2218	2320	}

drivers/char/virtio_console.c+71−86

@@ -422,7 +422,7 @@ static void reclaim_dma_bufs(void) · 421 unmodified lines
422	422	}
423	423	}
424	424
425		−static struct port_buffer alloc_buf(struct virtqueue vq, size_t buf_size,
	425	+static struct port_buffer alloc_buf(struct virtio_device vdev, size_t buf_size,
426	426	int pages)
427	427	{
428	428	struct port_buffer *buf;
@@ -445,16 +445,16 @@ static struct port_buffer alloc_buf(struct virtqueue vq, size_t buf_size, · 16 unmodified lines
445	445	return buf;
446	446	}
447	447
448		− if (is_rproc_serial(vq->vdev)) {
	448	+ if (is_rproc_serial(vdev)) {
449	449	/*
450	450	* Allocate DMA memory from ancestor. When a virtio
451	451	* device is created by remoteproc, the DMA memory is
452	452	* associated with the grandparent device:
453	453	* vdev => rproc => platform-dev.
454	454	*/
455		− if (!vq->vdev->dev.parent \|\| !vq->vdev->dev.parent->parent)
	455	+ if (!vdev->dev.parent \|\| !vdev->dev.parent->parent)
456	456	goto free_buf;
457		− buf->dev = vq->vdev->dev.parent->parent;
	457	+ buf->dev = vdev->dev.parent->parent;
458	458
459	459	/* Increase device refcnt to avoid freeing it */
460	460	get_device(buf->dev);
@@ -838,7 +838,7 @@ static ssize_t port_fops_write(struct file filp, const char __user ubuf, · 377 unmodified lines
838	838
839	839	count = min((size_t)(32 * 1024), count);
840	840
841		− buf = alloc_buf(port->out_vq, count, 0);
	841	+ buf = alloc_buf(port->portdev->vdev, count, 0);
842	842	if (!buf)
843	843	return -ENOMEM;
844	844
@@ -957,7 +957,7 @@ static ssize_t port_fops_splice_write(struct pipe_inode_info *pipe, · 112 unmodified lines
957	957	if (ret < 0)
958	958	goto error_out;
959	959
960		− buf = alloc_buf(port->out_vq, 0, pipe->nrbufs);
	960	+ buf = alloc_buf(port->portdev->vdev, 0, pipe->nrbufs);
961	961	if (!buf) {
962	962	ret = -ENOMEM;
963	963	goto error_out;
@@ -1374,7 +1374,7 @@ static unsigned int fill_queue(struct virtqueue vq, spinlock_t lock) · 410 unmodified lines
1374	1374
1375	1375	nr_added_bufs = 0;
1376	1376	do {
1377		− buf = alloc_buf(vq, PAGE_SIZE, 0);
	1377	+ buf = alloc_buf(vq->vdev, PAGE_SIZE, 0);
1378	1378	if (!buf)
1379	1379	break;
1380	1380
@@ -1402,7 +1402,6 @@ static int add_port(struct ports_device *portdev, u32 id) · 21 unmodified lines
1402	1402	{
1403	1403	char debugfs_name[16];
1404	1404	struct port *port;
1405		− struct port_buffer *buf;
1406	1405	dev_t devt;
1407	1406	unsigned int nr_added_bufs;
1408	1407	int err;
@@ -1513,8 +1512,6 @@ static int add_port(struct ports_device *portdev, u32 id) · 104 unmodified lines
1513	1512	return 0;
1514	1513
1515	1514	free_inbufs:
1516		− while ((buf = virtqueue_detach_unused_buf(port->in_vq)))
1517		− free_buf(buf, true);
1518	1515	free_device:
1519	1516	device_destroy(pdrvdata.class, port->dev->devt);
1520	1517	free_cdev:
@@ -1539,34 +1536,14 @@ static void remove_port(struct kref *kref) · 18 unmodified lines
1539	1536
1540	1537	static void remove_port_data(struct port *port)
1541	1538	{
1542		− struct port_buffer *buf;
1543		−
1544	1539	spin_lock_irq(&port->inbuf_lock);
1545	1540	/* Remove unused data this port might have received. */
1546	1541	discard_port_data(port);
1547	1542	spin_unlock_irq(&port->inbuf_lock);
1548	1543
1549		− /* Remove buffers we queued up for the Host to send us data in. */
1550		− do {
1551		− spin_lock_irq(&port->inbuf_lock);
1552		− buf = virtqueue_detach_unused_buf(port->in_vq);
1553		− spin_unlock_irq(&port->inbuf_lock);
1554		− if (buf)
1555		− free_buf(buf, true);
1556		− } while (buf);
1557		−
1558	1544	spin_lock_irq(&port->outvq_lock);
1559	1545	reclaim_consumed_buffers(port);
1560	1546	spin_unlock_irq(&port->outvq_lock);
1561		−
1562		− /* Free pending buffers from the out-queue. */
1563		− do {
1564		− spin_lock_irq(&port->outvq_lock);
1565		− buf = virtqueue_detach_unused_buf(port->out_vq);
1566		− spin_unlock_irq(&port->outvq_lock);
1567		− if (buf)
1568		− free_buf(buf, true);
1569		− } while (buf);
1570	1547	}
1571	1548
1572	1549	/*
@@ -1791,13 +1768,24 @@ static void control_work_handler(struct work_struct *work) · 218 unmodified lines
1791	1768	spin_unlock(&portdev->c_ivq_lock);
1792	1769	}
1793	1770
	1771	+static void flush_bufs(struct virtqueue *vq, bool can_sleep)
	1772	+{
	1773	+ struct port_buffer *buf;
	1774	+ unsigned int len;
	1775	+
	1776	+ while ((buf = virtqueue_get_buf(vq, &len)))
	1777	+ free_buf(buf, can_sleep);
	1778	+}
	1779	+
1794	1780	static void out_intr(struct virtqueue *vq)
1795	1781	{
1796	1782	struct port *port;
1797	1783
1798	1784	port = find_port_by_vq(vq->vdev->priv, vq);
1799		− if (!port)
	1785	+ if (!port) {
	1786	+ flush_bufs(vq, false);
1800	1787	return;
	1788	+ }
1801	1789
1802	1790	wake_up_interruptible(&port->waitqueue);
1803	1791	}
@@ -1808,8 +1796,10 @@ static void in_intr(struct virtqueue *vq) · 4 unmodified lines
1808	1796	unsigned long flags;
1809	1797
1810	1798	port = find_port_by_vq(vq->vdev->priv, vq);
1811		− if (!port)
	1799	+ if (!port) {
	1800	+ flush_bufs(vq, false);
1812	1801	return;
	1802	+ }
1813	1803
1814	1804	spin_lock_irqsave(&port->inbuf_lock, flags);
1815	1805	port->inbuf = get_inbuf(port);
@@ -1984,24 +1974,54 @@ static const struct file_operations portdev_fops = { · 168 unmodified lines
1984	1974
1985	1975	static void remove_vqs(struct ports_device *portdev)
1986	1976	{
	1977	+ struct virtqueue *vq;
	1978	+
	1979	+ virtio_device_for_each_vq(portdev->vdev, vq) {
	1980	+ struct port_buffer *buf;
	1981	+
	1982	+ flush_bufs(vq, true);
	1983	+ while ((buf = virtqueue_detach_unused_buf(vq)))
	1984	+ free_buf(buf, true);
	1985	+ }
1987	1986	portdev->vdev->config->del_vqs(portdev->vdev);
1988	1987	kfree(portdev->in_vqs);
1989	1988	kfree(portdev->out_vqs);
1990	1989	}
1991	1990
1992		−static void remove_controlq_data(struct ports_device *portdev)
	1991	+static void virtcons_remove(struct virtio_device *vdev)
1993	1992	{
1994		− struct port_buffer *buf;
1995		− unsigned int len;
	1993	+ struct ports_device *portdev;
	1994	+ struct port port, port2;
1996	1995
1997		− if (!use_multiport(portdev))
1998		− return;
	1996	+ portdev = vdev->priv;
1999	1997
2000		− while ((buf = virtqueue_get_buf(portdev->c_ivq, &len)))
2001		− free_buf(buf, true);
	1998	+ spin_lock_irq(&pdrvdata_lock);
	1999	+ list_del(&portdev->list);
	2000	+ spin_unlock_irq(&pdrvdata_lock);
2002	2001
2003		− while ((buf = virtqueue_detach_unused_buf(portdev->c_ivq)))
2004		− free_buf(buf, true);
	2002	+ /* Disable interrupts for vqs */
	2003	+ vdev->config->reset(vdev);
	2004	+ /* Finish up work that's lined up */
	2005	+ if (use_multiport(portdev))
	2006	+ cancel_work_sync(&portdev->control_work);
	2007	+ else
	2008	+ cancel_work_sync(&portdev->config_work);
	2009	+
	2010	+ list_for_each_entry_safe(port, port2, &portdev->ports, list)
	2011	+ unplug_port(port);
	2012	+
	2013	+ unregister_chrdev(portdev->chr_major, "virtio-portsdev");
	2014	+
	2015	+ /*
	2016	+ * When yanking out a device, we immediately lose the
	2017	+ * (device-side) queues. So there's no point in keeping the
	2018	+ * guest side around till we drop our final reference. This
	2019	+ * also means that any ports which are in an open state will
	2020	+ * have to just stop using the port, as the vqs are going
	2021	+ * away.
	2022	+ */
	2023	+ remove_vqs(portdev);
	2024	+ kfree(portdev);
2005	2025	}
2006	2026
2007	2027	/*
@@ -2070,6 +2090,7 @@ static int virtcons_probe(struct virtio_device *vdev) · 62 unmodified lines
2070	2090
2071	2091	spin_lock_init(&portdev->ports_lock);
2072	2092	INIT_LIST_HEAD(&portdev->ports);
	2093	+ INIT_LIST_HEAD(&portdev->list);
2073	2094
2074	2095	virtio_device_ready(portdev->vdev);
2075	2096
@@ -2087,8 +2108,15 @@ static int virtcons_probe(struct virtio_device *vdev) · 11 unmodified lines
2087	2108	if (!nr_added_bufs) {
2088	2109	dev_err(&vdev->dev,
2089	2110	"Error allocating buffers for control queue\n");
2090		− err = -ENOMEM;
2091		− goto free_vqs;
	2111	+ /*
	2112	+ * The host might want to notify mgmt sw about device
	2113	+ * add failure.
	2114	+ */
	2115	+ __send_control_msg(portdev, VIRTIO_CONSOLE_BAD_ID,
	2116	+ VIRTIO_CONSOLE_DEVICE_READY, 0);
	2117	+ /* Device was functional: we need full cleanup. */
	2118	+ virtcons_remove(vdev);
	2119	+ return -ENOMEM;
2092	2120	}
2093	2121	} else {
2094	2122	/*
@@ -2119,11 +2147,6 @@ static int virtcons_probe(struct virtio_device *vdev) · 24 unmodified lines
2119	2147
2120	2148	return 0;
2121	2149
2122		−free_vqs:
2123		− /* The host might want to notify mgmt sw about device add failure */
2124		− __send_control_msg(portdev, VIRTIO_CONSOLE_BAD_ID,
2125		− VIRTIO_CONSOLE_DEVICE_READY, 0);
2126		− remove_vqs(portdev);
2127	2150	free_chrdev:
2128	2151	unregister_chrdev(portdev->chr_major, "virtio-portsdev");
2129	2152	free:
@@ -2132,43 +2155,6 @@ fail: · 2 unmodified lines
2132	2155	return err;
2133	2156	}
2134	2157
2135		−static void virtcons_remove(struct virtio_device *vdev)
2136		−{
2137		− struct ports_device *portdev;
2138		− struct port port, port2;
2139		−
2140		− portdev = vdev->priv;
2141		−
2142		− spin_lock_irq(&pdrvdata_lock);
2143		− list_del(&portdev->list);
2144		− spin_unlock_irq(&pdrvdata_lock);
2145		−
2146		− /* Disable interrupts for vqs */
2147		− vdev->config->reset(vdev);
2148		− /* Finish up work that's lined up */
2149		− if (use_multiport(portdev))
2150		− cancel_work_sync(&portdev->control_work);
2151		− else
2152		− cancel_work_sync(&portdev->config_work);
2153		−
2154		− list_for_each_entry_safe(port, port2, &portdev->ports, list)
2155		− unplug_port(port);
2156		−
2157		− unregister_chrdev(portdev->chr_major, "virtio-portsdev");
2158		−
2159		− /*
2160		− * When yanking out a device, we immediately lose the
2161		− * (device-side) queues. So there's no point in keeping the
2162		− * guest side around till we drop our final reference. This
2163		− * also means that any ports which are in an open state will
2164		− * have to just stop using the port, as the vqs are going
2165		− * away.
2166		− */
2167		− remove_controlq_data(portdev);
2168		− remove_vqs(portdev);
2169		− kfree(portdev);
2170		−}
2171		−
2172	2158	static struct virtio_device_id id_table[] = {
2173	2159	{ VIRTIO_ID_CONSOLE, VIRTIO_DEV_ANY_ID },
2174	2160	{ 0 },
@@ -2209,7 +2195,6 @@ static int virtcons_freeze(struct virtio_device *vdev) · 34 unmodified lines
2209	2195	*/
2210	2196	if (use_multiport(portdev))
2211	2197	virtqueue_disable_cb(portdev->c_ivq);
2212		− remove_controlq_data(portdev);
2213	2198
2214	2199	list_for_each_entry(port, &portdev->ports, list) {
2215	2200	virtqueue_disable_cb(port->in_vq);

drivers/clk/clk-cs2000-cp.c+1−1

@@ -541,7 +541,7 @@ probe_err: · 540 unmodified lines
541	541	return ret;
542	542	}
543	543
544		−static int cs2000_resume(struct device *dev)
	544	+static int __maybe_unused cs2000_resume(struct device *dev)
545	545	{
546	546	struct cs2000_priv *priv = dev_get_drvdata(dev);
547	547

drivers/clk/clk-mux.c+9−1

@@ -112,10 +112,18 @@ static int clk_mux_set_parent(struct clk_hw *hw, u8 index) · 111 unmodified lines
112	112	return 0;
113	113	}
114	114
	115	+static int clk_mux_determine_rate(struct clk_hw *hw,
	116	+ struct clk_rate_request *req)
	117	+{
	118	+ struct clk_mux *mux = to_clk_mux(hw);
	119	+
	120	+ return clk_mux_determine_rate_flags(hw, req, mux->flags);
	121	+}
	122	+
115	123	const struct clk_ops clk_mux_ops = {
116	124	.get_parent = clk_mux_get_parent,
117	125	.set_parent = clk_mux_set_parent,
118		− .determine_rate = __clk_mux_determine_rate,
	126	+ .determine_rate = clk_mux_determine_rate,
119	127	};
120	128	EXPORT_SYMBOL_GPL(clk_mux_ops);
121	129

drivers/clk/clk-stm32mp1.c+23−31

@@ -216,7 +216,7 @@ static const char * const usart1_src[] = { · 215 unmodified lines
216	216	"pclk5", "pll3_q", "ck_hsi", "ck_csi", "pll4_q", "ck_hse"
217	217	};
218	218
219		−const char * const usart234578_src[] = {
	219	+static const char * const usart234578_src[] = {
220	220	"pclk1", "pll4_q", "ck_hsi", "ck_csi", "ck_hse"
221	221	};
222	222
@@ -224,10 +224,6 @@ static const char * const usart6_src[] = { · 1 unmodified line
224	224	"pclk2", "pll4_q", "ck_hsi", "ck_csi", "ck_hse"
225	225	};
226	226
227		−static const char * const dfsdm_src[] = {
228		− "pclk2", "ck_mcu"
229		−};
230		−
231	227	static const char * const fdcan_src[] = {
232	228	"ck_hse", "pll3_q", "pll4_q"
233	229	};
@@ -316,10 +312,8 @@ struct stm32_clk_mgate { · 82 unmodified lines
316	312	struct clock_config {
317	313	u32 id;
318	314	const char *name;
319		− union {
320		− const char *parent_name;
321		− const char * const *parent_names;
322		− };
	315	+ const char *parent_name;
	316	+ const char * const *parent_names;
323	317	int num_parents;
324	318	unsigned long flags;
325	319	void *cfg;
@@ -469,7 +463,7 @@ static void mp1_gate_clk_disable(struct clk_hw *hw) · 143 unmodified lines
469	463	}
470	464	}
471	465
472		−const struct clk_ops mp1_gate_clk_ops = {
	466	+static const struct clk_ops mp1_gate_clk_ops = {
473	467	.enable = mp1_gate_clk_enable,
474	468	.disable = mp1_gate_clk_disable,
475	469	.is_enabled = clk_gate_is_enabled,
@@ -698,7 +692,7 @@ static void mp1_mgate_clk_disable(struct clk_hw *hw) · 222 unmodified lines
698	692	mp1_gate_clk_disable(hw);
699	693	}
700	694
701		−const struct clk_ops mp1_mgate_clk_ops = {
	695	+static const struct clk_ops mp1_mgate_clk_ops = {
702	696	.enable = mp1_mgate_clk_enable,
703	697	.disable = mp1_mgate_clk_disable,
704	698	.is_enabled = clk_gate_is_enabled,
@@ -732,7 +726,7 @@ static int clk_mmux_set_parent(struct clk_hw *hw, u8 index) · 27 unmodified lines
732	726	return 0;
733	727	}
734	728
735		−const struct clk_ops clk_mmux_ops = {
	729	+static const struct clk_ops clk_mmux_ops = {
736	730	.get_parent = clk_mmux_get_parent,
737	731	.set_parent = clk_mmux_set_parent,
738	732	.determine_rate = __clk_mux_determine_rate,
@@ -1048,10 +1042,10 @@ struct stm32_pll_cfg { · 309 unmodified lines
1048	1042	u32 offset;
1049	1043	};
1050	1044
1051		−struct clk_hw _clk_register_pll(struct device dev,
1052		− struct clk_hw_onecell_data *clk_data,
1053		− void __iomem base, spinlock_t lock,
1054		− const struct clock_config *cfg)
	1045	+static struct clk_hw _clk_register_pll(struct device dev,
	1046	+ struct clk_hw_onecell_data *clk_data,
	1047	+ void __iomem base, spinlock_t lock,
	1048	+ const struct clock_config *cfg)
1055	1049	{
1056	1050	struct stm32_pll_cfg *stm_pll_cfg = cfg->cfg;
1057	1051
@@ -1405,7 +1399,8 @@ enum { · 347 unmodified lines
1405	1399	G_USBH,
1406	1400	G_ETHSTP,
1407	1401	G_RTCAPB,
1408		− G_TZC,
	1402	+ G_TZC1,
	1403	+ G_TZC2,
1409	1404	G_TZPC,
1410	1405	G_IWDG1,
1411	1406	G_BSEC,
@@ -1417,7 +1412,7 @@ enum { · 5 unmodified lines
1417	1412	G_LAST
1418	1413	};
1419	1414
1420		−struct stm32_mgate mp1_mgate[G_LAST];
	1415	+static struct stm32_mgate mp1_mgate[G_LAST];
1421	1416
1422	1417	#define _K_GATE(_id, _gate_offset, _gate_bit_idx, _gate_flags,\
1423	1418	_mgate, _ops)\
@@ -1440,7 +1435,7 @@ struct stm32_mgate mp1_mgate[G_LAST]; · 16 unmodified lines
1440	1435	&mp1_mgate[_id], &mp1_mgate_clk_ops)
1441	1436
1442	1437	/* Peripheral gates */
1443		−struct stm32_gate_cfg per_gate_cfg[G_LAST] = {
	1438	+static struct stm32_gate_cfg per_gate_cfg[G_LAST] = {
1444	1439	/* Multi gates */
1445	1440	K_GATE(G_MDIO, RCC_APB1ENSETR, 31, 0),
1446	1441	K_MGATE(G_DAC12, RCC_APB1ENSETR, 29, 0),
@@ -1506,7 +1501,8 @@ struct stm32_gate_cfg per_gate_cfg[G_LAST] = { · 59 unmodified lines
1506	1501	K_GATE(G_BSEC, RCC_APB5ENSETR, 16, 0),
1507	1502	K_GATE(G_IWDG1, RCC_APB5ENSETR, 15, 0),
1508	1503	K_GATE(G_TZPC, RCC_APB5ENSETR, 13, 0),
1509		− K_GATE(G_TZC, RCC_APB5ENSETR, 12, 0),
	1504	+ K_GATE(G_TZC2, RCC_APB5ENSETR, 12, 0),
	1505	+ K_GATE(G_TZC1, RCC_APB5ENSETR, 11, 0),
1510	1506	K_GATE(G_RTCAPB, RCC_APB5ENSETR, 8, 0),
1511	1507	K_MGATE(G_USART1, RCC_APB5ENSETR, 4, 0),
1512	1508	K_MGATE(G_I2C6, RCC_APB5ENSETR, 3, 0),
@@ -1600,7 +1596,7 @@ enum { · 87 unmodified lines
1600	1596	M_LAST
1601	1597	};
1602	1598
1603		−struct stm32_mmux ker_mux[M_LAST];
	1599	+static struct stm32_mmux ker_mux[M_LAST];
1604	1600
1605	1601	#define _K_MUX(_id, _offset, _shift, _width, _mux_flags, _mmux, _ops)\
1606	1602	[_id] = {\
@@ -1623,7 +1619,7 @@ struct stm32_mmux ker_mux[M_LAST]; · 16 unmodified lines
1623	1619	_K_MUX(_id, _offset, _shift, _width, _mux_flags,\
1624	1620	&ker_mux[_id], &clk_mmux_ops)
1625	1621
1626		−const struct stm32_mux_cfg ker_mux_cfg[M_LAST] = {
	1622	+static const struct stm32_mux_cfg ker_mux_cfg[M_LAST] = {
1627	1623	/* Kernel multi mux */
1628	1624	K_MMUX(M_SDMMC12, RCC_SDMMC12CKSELR, 0, 3, 0),
1629	1625	K_MMUX(M_SPI23, RCC_SPI2S23CKSELR, 0, 3, 0),
@@ -1860,7 +1856,8 @@ static const struct clock_config stm32mp1_clock_cfg[] = { · 230 unmodified lines
1860	1856	PCLK(USART1, "usart1", "pclk5", 0, G_USART1),
1861	1857	PCLK(RTCAPB, "rtcapb", "pclk5", CLK_IGNORE_UNUSED \|
1862	1858	CLK_IS_CRITICAL, G_RTCAPB),
1863		− PCLK(TZC, "tzc", "pclk5", CLK_IGNORE_UNUSED, G_TZC),
	1859	+ PCLK(TZC1, "tzc1", "ck_axi", CLK_IGNORE_UNUSED, G_TZC1),
	1860	+ PCLK(TZC2, "tzc2", "ck_axi", CLK_IGNORE_UNUSED, G_TZC2),
1864	1861	PCLK(TZPC, "tzpc", "pclk5", CLK_IGNORE_UNUSED, G_TZPC),
1865	1862	PCLK(IWDG1, "iwdg1", "pclk5", 0, G_IWDG1),
1866	1863	PCLK(BSEC, "bsec", "pclk5", CLK_IGNORE_UNUSED, G_BSEC),
@@ -1916,8 +1913,7 @@ static const struct clock_config stm32mp1_clock_cfg[] = { · 49 unmodified lines
1916	1913	KCLK(RNG1_K, "rng1_k", rng_src, 0, G_RNG1, M_RNG1),
1917	1914	KCLK(RNG2_K, "rng2_k", rng_src, 0, G_RNG2, M_RNG2),
1918	1915	KCLK(USBPHY_K, "usbphy_k", usbphy_src, 0, G_USBPHY, M_USBPHY),
1919		− KCLK(STGEN_K, "stgen_k", stgen_src, CLK_IGNORE_UNUSED,
1920		− G_STGEN, M_STGEN),
	1916	+ KCLK(STGEN_K, "stgen_k", stgen_src, CLK_IS_CRITICAL, G_STGEN, M_STGEN),
1921	1917	KCLK(SPDIF_K, "spdif_k", spdif_src, 0, G_SPDIF, M_SPDIF),
1922	1918	KCLK(SPI1_K, "spi1_k", spi123_src, 0, G_SPI1, M_SPI1),
1923	1919	KCLK(SPI2_K, "spi2_k", spi123_src, 0, G_SPI2, M_SPI23),
@@ -1948,8 +1944,8 @@ static const struct clock_config stm32mp1_clock_cfg[] = { · 24 unmodified lines
1948	1944	KCLK(FDCAN_K, "fdcan_k", fdcan_src, 0, G_FDCAN, M_FDCAN),
1949	1945	KCLK(SAI1_K, "sai1_k", sai_src, 0, G_SAI1, M_SAI1),
1950	1946	KCLK(SAI2_K, "sai2_k", sai2_src, 0, G_SAI2, M_SAI2),
1951		− KCLK(SAI3_K, "sai3_k", sai_src, 0, G_SAI2, M_SAI3),
1952		− KCLK(SAI4_K, "sai4_k", sai_src, 0, G_SAI2, M_SAI4),
	1947	+ KCLK(SAI3_K, "sai3_k", sai_src, 0, G_SAI3, M_SAI3),
	1948	+ KCLK(SAI4_K, "sai4_k", sai_src, 0, G_SAI4, M_SAI4),
1953	1949	KCLK(ADC12_K, "adc12_k", adc12_src, 0, G_ADC12, M_ADC12),
1954	1950	KCLK(DSI_K, "dsi_k", dsi_src, 0, G_DSI, M_DSI),
1955	1951	KCLK(ADFSDM_K, "adfsdm_k", sai_src, 0, G_ADFSDM, M_SAI1),
@@ -1992,10 +1988,6 @@ static const struct clock_config stm32mp1_clock_cfg[] = { · 36 unmodified lines
1992	1988	_DIV(RCC_MCO2CFGR, 4, 4, 0, NULL)),
1993	1989
1994	1990	/* Debug clocks */
1995		− FIXED_FACTOR(NO_ID, "ck_axi_div2", "ck_axi", 0, 1, 2),
1996		−
1997		− GATE(DBG, "ck_apb_dbg", "ck_axi_div2", 0, RCC_DBGCFGR, 8, 0),
1998		−
1999	1991	GATE(CK_DBG, "ck_sys_dbg", "ck_axi", 0, RCC_DBGCFGR, 8, 0),
2000	1992
2001	1993	COMPOSITE(CK_TRACE, "ck_trace", ck_trace_src, CLK_OPS_PARENT_ENABLE,

drivers/clk/clk.c+4−3

@@ -426,9 +426,9 @@ static bool mux_is_better_rate(unsigned long rate, unsigned long now, · 425 unmodified lines
426	426	return now <= rate && now > best;
427	427	}
428	428
429		−static int
430		−clk_mux_determine_rate_flags(struct clk_hw hw, struct clk_rate_request req,
431		− unsigned long flags)
	429	+int clk_mux_determine_rate_flags(struct clk_hw *hw,
	430	+ struct clk_rate_request *req,
	431	+ unsigned long flags)
432	432	{
433	433	struct clk_core core = hw->core, parent, *best_parent = NULL;
434	434	int i, num_parents, ret;
@@ -488,6 +488,7 @@ out: · 53 unmodified lines
488	488
489	489	return 0;
490	490	}
	491	+EXPORT_SYMBOL_GPL(clk_mux_determine_rate_flags);
491	492
492	493	struct clk __clk_lookup(const char name)
493	494	{

drivers/clk/meson/clk-regmap.c+10−1

@@ -153,10 +153,19 @@ static int clk_regmap_mux_set_parent(struct clk_hw *hw, u8 index) · 152 unmodified lines
153	153	val << mux->shift);
154	154	}
155	155
	156	+static int clk_regmap_mux_determine_rate(struct clk_hw *hw,
	157	+ struct clk_rate_request *req)
	158	+{
	159	+ struct clk_regmap *clk = to_clk_regmap(hw);
	160	+ struct clk_regmap_mux_data *mux = clk_get_regmap_mux_data(clk);
	161	+
	162	+ return clk_mux_determine_rate_flags(hw, req, mux->flags);
	163	+}
	164	+
156	165	const struct clk_ops clk_regmap_mux_ops = {
157	166	.get_parent = clk_regmap_mux_get_parent,
158	167	.set_parent = clk_regmap_mux_set_parent,
159		− .determine_rate = __clk_mux_determine_rate,
	168	+ .determine_rate = clk_regmap_mux_determine_rate,
160	169	};
161	170	EXPORT_SYMBOL_GPL(clk_regmap_mux_ops);
162	171

drivers/clk/meson/gxbb-aoclk.h+0−2

@@ -17,8 +17,6 @@ · 16 unmodified lines
17	17	#define AO_RTC_ALT_CLK_CNTL0 0x94
18	18	#define AO_RTC_ALT_CLK_CNTL1 0x98
19	19
20		−extern const struct clk_ops meson_aoclk_gate_regmap_ops;
21		−
22	20	struct aoclk_cec_32k {
23	21	struct clk_hw hw;
24	22	struct regmap *regmap;

drivers/clk/meson/meson8b.c+3−2

@@ -253,7 +253,7 @@ static struct clk_fixed_factor meson8b_fclk_div3_div = { · 252 unmodified lines
253	253	.mult = 1,
254	254	.div = 3,
255	255	.hw.init = &(struct clk_init_data){
256		− .name = "fclk_div_div3",
	256	+ .name = "fclk_div3_div",
257	257	.ops = &clk_fixed_factor_ops,
258	258	.parent_names = (const char *[]){ "fixed_pll" },
259	259	.num_parents = 1,
@@ -632,7 +632,8 @@ static struct clk_regmap meson8b_cpu_clk = { · 372 unmodified lines
632	632	.hw.init = &(struct clk_init_data){
633	633	.name = "cpu_clk",
634	634	.ops = &clk_regmap_mux_ro_ops,
635		− .parent_names = (const char *[]){ "xtal", "cpu_out_sel" },
	635	+ .parent_names = (const char *[]){ "xtal",
	636	+ "cpu_scale_out_sel" },
636	637	.num_parents = 2,
637	638	.flags = (CLK_SET_RATE_PARENT \|
638	639	CLK_SET_RATE_NO_REPARENT),

drivers/clocksource/Kconfig+8−0

@@ -133,6 +133,14 @@ config VT8500_TIMER · 132 unmodified lines
133	133	help
134	134	Enables support for the VT8500 driver.
135	135
	136	+config NPCM7XX_TIMER
	137	+ bool "NPCM7xx timer driver" if COMPILE_TEST
	138	+ depends on HAS_IOMEM
	139	+ select CLKSRC_MMIO
	140	+ help
	141	+ Enable 24-bit TIMER0 and TIMER1 counters in the NPCM7xx architecture,
	142	+ While TIMER0 serves as clockevent and TIMER1 serves as clocksource.
	143	+
136	144	config CADENCE_TTC_TIMER
137	145	bool "Cadence TTC timer driver" if COMPILE_TEST
138	146	depends on COMMON_CLK

drivers/clocksource/Makefile+1−0

@@ -56,6 +56,7 @@ obj-$(CONFIG_CLKSRC_NPS) += timer-nps.o · 55 unmodified lines
56	56	obj-$(CONFIG_OXNAS_RPS_TIMER) += timer-oxnas-rps.o
57	57	obj-$(CONFIG_OWL_TIMER) += owl-timer.o
58	58	obj-$(CONFIG_SPRD_TIMER) += timer-sprd.o
	59	+obj-$(CONFIG_NPCM7XX_TIMER) += timer-npcm7xx.o
59	60
60	61	obj-$(CONFIG_ARC_TIMERS) += arc_timer.o
61	62	obj-$(CONFIG_ARM_ARCH_TIMER) += arm_arch_timer.o

drivers/clocksource/timer-imx-tpm.c+34−11

@@ -17,9 +17,14 @@ · 16 unmodified lines
17	17	#include <linux/of_irq.h>
18	18	#include <linux/sched_clock.h>
19	19
	20	+#define TPM_PARAM 0x4
	21	+#define TPM_PARAM_WIDTH_SHIFT 16
	22	+#define TPM_PARAM_WIDTH_MASK (0xff << 16)
20	23	#define TPM_SC 0x10
21	24	#define TPM_SC_CMOD_INC_PER_CNT (0x1 << 3)
22	25	#define TPM_SC_CMOD_DIV_DEFAULT 0x3
	26	+#define TPM_SC_CMOD_DIV_MAX 0x7
	27	+#define TPM_SC_TOF_MASK (0x1 << 7)
23	28	#define TPM_CNT 0x14
24	29	#define TPM_MOD 0x18
25	30	#define TPM_STATUS 0x1c
@@ -29,8 +34,11 @@ · 3 unmodified lines
29	34	#define TPM_C0SC_MODE_SHIFT 2
30	35	#define TPM_C0SC_MODE_MASK 0x3c
31	36	#define TPM_C0SC_MODE_SW_COMPARE 0x4
	37	+#define TPM_C0SC_CHF_MASK (0x1 << 7)
32	38	#define TPM_C0V 0x24
33	39
	40	+static int counter_width;
	41	+static int rating;
34	42	static void __iomem *timer_base;
35	43	static struct clock_event_device clockevent_tpm;
36	44
@@ -83,10 +91,11 @@ static int __init tpm_clocksource_init(unsigned long rate) · 46 unmodified lines
83	91	tpm_delay_timer.freq = rate;
84	92	register_current_timer_delay(&tpm_delay_timer);
85	93
86		− sched_clock_register(tpm_read_sched_clock, 32, rate);
	94	+ sched_clock_register(tpm_read_sched_clock, counter_width, rate);
87	95
88	96	return clocksource_mmio_init(timer_base + TPM_CNT, "imx-tpm",
89		− rate, 200, 32, clocksource_mmio_readl_up);
	97	+ rate, rating, counter_width,
	98	+ clocksource_mmio_readl_up);
90	99	}
91	100
92	101	static int tpm_set_next_event(unsigned long delta,
@@ -105,7 +114,7 @@ static int tpm_set_next_event(unsigned long delta, · 12 unmodified lines
105	114	* of writing CNT registers which may cause the min_delta event got
106	115	* missed, so we need add a ETIME check here in case it happened.
107	116	*/
108		− return (int)((next - now) <= 0) ? -ETIME : 0;
	117	+ return (int)(next - now) <= 0 ? -ETIME : 0;
109	118	}
110	119
111	120	static int tpm_set_state_oneshot(struct clock_event_device *evt)
@@ -139,7 +148,6 @@ static struct clock_event_device clockevent_tpm = { · 27 unmodified lines
139	148	.set_state_oneshot = tpm_set_state_oneshot,
140	149	.set_next_event = tpm_set_next_event,
141	150	.set_state_shutdown = tpm_set_state_shutdown,
142		− .rating = 200,
143	151	};
144	152
145	153	static int __init tpm_clockevent_init(unsigned long rate, int irq)
@@ -149,10 +157,11 @@ static int __init tpm_clockevent_init(unsigned long rate, int irq) · 3 unmodified lines
149	157	ret = request_irq(irq, tpm_timer_interrupt, IRQF_TIMER \| IRQF_IRQPOLL,
150	158	"i.MX7ULP TPM Timer", &clockevent_tpm);
151	159
	160	+ clockevent_tpm.rating = rating;
152	161	clockevent_tpm.cpumask = cpumask_of(0);
153	162	clockevent_tpm.irq = irq;
154		− clockevents_config_and_register(&clockevent_tpm,
155		− rate, 300, 0xfffffffe);
	163	+ clockevents_config_and_register(&clockevent_tpm, rate, 300,
	164	+ GENMASK(counter_width - 1, 1));
156	165
157	166	return ret;
158	167	}
@@ -179,7 +188,7 @@ static int __init tpm_timer_init(struct device_node *np) · 20 unmodified lines
179	188	ipg = of_clk_get_by_name(np, "ipg");
180	189	per = of_clk_get_by_name(np, "per");
181	190	if (IS_ERR(ipg) \|\| IS_ERR(per)) {
182		− pr_err("tpm: failed to get igp or per clk\n");
	191	+ pr_err("tpm: failed to get ipg or per clk\n");
183	192	ret = -ENODEV;
184	193	goto err_clk_get;
185	194	}
@@ -197,6 +206,11 @@ static int __init tpm_timer_init(struct device_node *np) · 11 unmodified lines
197	206	goto err_per_clk_enable;
198	207	}
199	208
	209	+ counter_width = (readl(timer_base + TPM_PARAM) & TPM_PARAM_WIDTH_MASK)
	210	+ >> TPM_PARAM_WIDTH_SHIFT;
	211	+ /* use rating 200 for 32-bit counter and 150 for 16-bit counter */
	212	+ rating = counter_width == 0x20 ? 200 : 150;
	213	+
200	214	/*
201	215	* Initialize tpm module to a known state
202	216	* 1) Counter disabled
@@ -205,16 +219,25 @@ static int __init tpm_timer_init(struct device_node *np) · 2 unmodified lines
205	219	* 4) Channel0 disabled
206	220	* 5) DMA transfers disabled
207	221	*/
	222	+ /* make sure counter is disabled */
208	223	writel(0, timer_base + TPM_SC);
	224	+ /* TOF is W1C */
	225	+ writel(TPM_SC_TOF_MASK, timer_base + TPM_SC);
209	226	writel(0, timer_base + TPM_CNT);
210		− writel(0, timer_base + TPM_C0SC);
	227	+ /* CHF is W1C */
	228	+ writel(TPM_C0SC_CHF_MASK, timer_base + TPM_C0SC);
211	229
212		− /* increase per cnt, div 8 by default */
213		− writel(TPM_SC_CMOD_INC_PER_CNT \| TPM_SC_CMOD_DIV_DEFAULT,
	230	+ /*
	231	+ * increase per cnt,
	232	+ * div 8 for 32-bit counter and div 128 for 16-bit counter
	233	+ */
	234	+ writel(TPM_SC_CMOD_INC_PER_CNT \|
	235	+ (counter_width == 0x20 ?
	236	+ TPM_SC_CMOD_DIV_DEFAULT : TPM_SC_CMOD_DIV_MAX),
214	237	timer_base + TPM_SC);
215	238
216	239	/* set MOD register to maximum for free running mode */
217		− writel(0xffffffff, timer_base + TPM_MOD);
	240	+ writel(GENMASK(counter_width - 1, 0), timer_base + TPM_MOD);
218	241
219	242	rate = clk_get_rate(per) >> 3;
220	243	ret = tpm_clocksource_init(rate);

drivers/clocksource/timer-npcm7xx.c+215−0

@@ -0,0 +1,215 @@
	1	+// SPDX-License-Identifier: GPL-2.0
	2	+/*
	3	+ * Copyright (C) 2014-2018 Nuvoton Technologies tomer.maimon@nuvoton.com
	4	+ * All rights reserved.
	5	+ *
	6	+ * Copyright 2017 Google, Inc.
	7	+ */
	8	+
	9	+#include <linux/kernel.h>
	10	+#include <linux/sched.h>
	11	+#include <linux/init.h>
	12	+#include <linux/interrupt.h>
	13	+#include <linux/err.h>
	14	+#include <linux/clk.h>
	15	+#include <linux/io.h>
	16	+#include <linux/clockchips.h>
	17	+#include <linux/of_irq.h>
	18	+#include <linux/of_address.h>
	19	+#include "timer-of.h"
	20	+
	21	+/* Timers registers */
	22	+#define NPCM7XX_REG_TCSR0 0x0 /* Timer 0 Control and Status Register */
	23	+#define NPCM7XX_REG_TICR0 0x8 /* Timer 0 Initial Count Register */
	24	+#define NPCM7XX_REG_TCSR1 0x4 /* Timer 1 Control and Status Register */
	25	+#define NPCM7XX_REG_TICR1 0xc /* Timer 1 Initial Count Register */
	26	+#define NPCM7XX_REG_TDR1 0x14 /* Timer 1 Data Register */
	27	+#define NPCM7XX_REG_TISR 0x18 /* Timer Interrupt Status Register */
	28	+
	29	+/* Timers control */
	30	+#define NPCM7XX_Tx_RESETINT 0x1f
	31	+#define NPCM7XX_Tx_PERIOD BIT(27)
	32	+#define NPCM7XX_Tx_INTEN BIT(29)
	33	+#define NPCM7XX_Tx_COUNTEN BIT(30)
	34	+#define NPCM7XX_Tx_ONESHOT 0x0
	35	+#define NPCM7XX_Tx_OPER GENMASK(3, 27)
	36	+#define NPCM7XX_Tx_MIN_PRESCALE 0x1
	37	+#define NPCM7XX_Tx_TDR_MASK_BITS 24
	38	+#define NPCM7XX_Tx_MAX_CNT 0xFFFFFF
	39	+#define NPCM7XX_T0_CLR_INT 0x1
	40	+#define NPCM7XX_Tx_CLR_CSR 0x0
	41	+
	42	+/* Timers operating mode */
	43	+#define NPCM7XX_START_PERIODIC_Tx (NPCM7XX_Tx_PERIOD \| NPCM7XX_Tx_COUNTEN \| \
	44	+ NPCM7XX_Tx_INTEN \| \
	45	+ NPCM7XX_Tx_MIN_PRESCALE)
	46	+
	47	+#define NPCM7XX_START_ONESHOT_Tx (NPCM7XX_Tx_ONESHOT \| NPCM7XX_Tx_COUNTEN \| \
	48	+ NPCM7XX_Tx_INTEN \| \
	49	+ NPCM7XX_Tx_MIN_PRESCALE)
	50	+
	51	+#define NPCM7XX_START_Tx (NPCM7XX_Tx_COUNTEN \| NPCM7XX_Tx_PERIOD \| \
	52	+ NPCM7XX_Tx_MIN_PRESCALE)
	53	+
	54	+#define NPCM7XX_DEFAULT_CSR (NPCM7XX_Tx_CLR_CSR \| NPCM7XX_Tx_MIN_PRESCALE)
	55	+
	56	+static int npcm7xx_timer_resume(struct clock_event_device *evt)
	57	+{
	58	+ struct timer_of *to = to_timer_of(evt);
	59	+ u32 val;
	60	+
	61	+ val = readl(timer_of_base(to) + NPCM7XX_REG_TCSR0);
	62	+ val \|= NPCM7XX_Tx_COUNTEN;
	63	+ writel(val, timer_of_base(to) + NPCM7XX_REG_TCSR0);
	64	+
	65	+ return 0;
	66	+}
	67	+
	68	+static int npcm7xx_timer_shutdown(struct clock_event_device *evt)
	69	+{
	70	+ struct timer_of *to = to_timer_of(evt);
	71	+ u32 val;
	72	+
	73	+ val = readl(timer_of_base(to) + NPCM7XX_REG_TCSR0);
	74	+ val &= ~NPCM7XX_Tx_COUNTEN;
	75	+ writel(val, timer_of_base(to) + NPCM7XX_REG_TCSR0);
	76	+
	77	+ return 0;
	78	+}
	79	+
	80	+static int npcm7xx_timer_oneshot(struct clock_event_device *evt)
	81	+{
	82	+ struct timer_of *to = to_timer_of(evt);
	83	+ u32 val;
	84	+
	85	+ val = readl(timer_of_base(to) + NPCM7XX_REG_TCSR0);
	86	+ val &= ~NPCM7XX_Tx_OPER;
	87	+
	88	+ val = readl(timer_of_base(to) + NPCM7XX_REG_TCSR0);
	89	+ val \|= NPCM7XX_START_ONESHOT_Tx;
	90	+ writel(val, timer_of_base(to) + NPCM7XX_REG_TCSR0);
	91	+
	92	+ return 0;
	93	+}
	94	+
	95	+static int npcm7xx_timer_periodic(struct clock_event_device *evt)
	96	+{
	97	+ struct timer_of *to = to_timer_of(evt);
	98	+ u32 val;
	99	+
	100	+ val = readl(timer_of_base(to) + NPCM7XX_REG_TCSR0);
	101	+ val &= ~NPCM7XX_Tx_OPER;
	102	+
	103	+ writel(timer_of_period(to), timer_of_base(to) + NPCM7XX_REG_TICR0);
	104	+ val \|= NPCM7XX_START_PERIODIC_Tx;
	105	+
	106	+ writel(val, timer_of_base(to) + NPCM7XX_REG_TCSR0);
	107	+
	108	+ return 0;
	109	+}
	110	+
	111	+static int npcm7xx_clockevent_set_next_event(unsigned long evt,
	112	+ struct clock_event_device *clk)
	113	+{
	114	+ struct timer_of *to = to_timer_of(clk);
	115	+ u32 val;
	116	+
	117	+ writel(evt, timer_of_base(to) + NPCM7XX_REG_TICR0);
	118	+ val = readl(timer_of_base(to) + NPCM7XX_REG_TCSR0);
	119	+ val \|= NPCM7XX_START_Tx;
	120	+ writel(val, timer_of_base(to) + NPCM7XX_REG_TCSR0);
	121	+
	122	+ return 0;
	123	+}
	124	+
	125	+static irqreturn_t npcm7xx_timer0_interrupt(int irq, void *dev_id)
	126	+{
	127	+ struct clock_event_device evt = (struct clock_event_device )dev_id;
	128	+ struct timer_of *to = to_timer_of(evt);
	129	+
	130	+ writel(NPCM7XX_T0_CLR_INT, timer_of_base(to) + NPCM7XX_REG_TISR);
	131	+
	132	+ evt->event_handler(evt);
	133	+
	134	+ return IRQ_HANDLED;
	135	+}
	136	+
	137	+static struct timer_of npcm7xx_to = {
	138	+ .flags = TIMER_OF_IRQ \| TIMER_OF_BASE \| TIMER_OF_CLOCK,
	139	+
	140	+ .clkevt = {
	141	+ .name = "npcm7xx-timer0",
	142	+ .features = CLOCK_EVT_FEAT_PERIODIC \|
	143	+ CLOCK_EVT_FEAT_ONESHOT,
	144	+ .set_next_event = npcm7xx_clockevent_set_next_event,
	145	+ .set_state_shutdown = npcm7xx_timer_shutdown,
	146	+ .set_state_periodic = npcm7xx_timer_periodic,
	147	+ .set_state_oneshot = npcm7xx_timer_oneshot,
	148	+ .tick_resume = npcm7xx_timer_resume,
	149	+ .rating = 300,
	150	+ },
	151	+
	152	+ .of_irq = {
	153	+ .handler = npcm7xx_timer0_interrupt,
	154	+ .flags = IRQF_TIMER \| IRQF_IRQPOLL,
	155	+ },
	156	+};
	157	+
	158	+static void __init npcm7xx_clockevents_init(void)
	159	+{
	160	+ writel(NPCM7XX_DEFAULT_CSR,
	161	+ timer_of_base(&npcm7xx_to) + NPCM7XX_REG_TCSR0);
	162	+
	163	+ writel(NPCM7XX_Tx_RESETINT,
	164	+ timer_of_base(&npcm7xx_to) + NPCM7XX_REG_TISR);
	165	+
	166	+ npcm7xx_to.clkevt.cpumask = cpumask_of(0);
	167	+ clockevents_config_and_register(&npcm7xx_to.clkevt,
	168	+ timer_of_rate(&npcm7xx_to),
	169	+ 0x1, NPCM7XX_Tx_MAX_CNT);
	170	+}
	171	+
	172	+static void __init npcm7xx_clocksource_init(void)
	173	+{
	174	+ u32 val;
	175	+
	176	+ writel(NPCM7XX_DEFAULT_CSR,
	177	+ timer_of_base(&npcm7xx_to) + NPCM7XX_REG_TCSR1);
	178	+ writel(NPCM7XX_Tx_MAX_CNT,
	179	+ timer_of_base(&npcm7xx_to) + NPCM7XX_REG_TICR1);
	180	+
	181	+ val = readl(timer_of_base(&npcm7xx_to) + NPCM7XX_REG_TCSR1);
	182	+ val \|= NPCM7XX_START_Tx;
	183	+ writel(val, timer_of_base(&npcm7xx_to) + NPCM7XX_REG_TCSR1);
	184	+
	185	+ clocksource_mmio_init(timer_of_base(&npcm7xx_to) +
	186	+ NPCM7XX_REG_TDR1,
	187	+ "npcm7xx-timer1", timer_of_rate(&npcm7xx_to),
	188	+ 200, (unsigned int)NPCM7XX_Tx_TDR_MASK_BITS,
	189	+ clocksource_mmio_readl_down);
	190	+}
	191	+
	192	+static int __init npcm7xx_timer_init(struct device_node *np)
	193	+{
	194	+ int ret;
	195	+
	196	+ ret = timer_of_init(np, &npcm7xx_to);
	197	+ if (ret)
	198	+ return ret;
	199	+
	200	+ /* Clock input is divided by PRESCALE + 1 before it is fed */
	201	+ /* to the counter */
	202	+ npcm7xx_to.of_clk.rate = npcm7xx_to.of_clk.rate /
	203	+ (NPCM7XX_Tx_MIN_PRESCALE + 1);
	204	+
	205	+ npcm7xx_clocksource_init();
	206	+ npcm7xx_clockevents_init();
	207	+
	208	+ pr_info("Enabling NPCM7xx clocksource timer base: %px, IRQ: %d ",
	209	+ timer_of_base(&npcm7xx_to), timer_of_irq(&npcm7xx_to));
	210	+
	211	+ return 0;
	212	+}
	213	+
	214	+TIMER_OF_DECLARE(npcm7xx, "nuvoton,npcm750-timer", npcm7xx_timer_init);
	215	+

drivers/cpufreq/Kconfig.arm+0−10

@@ -71,16 +71,6 @@ config ARM_BRCMSTB_AVS_CPUFREQ · 70 unmodified lines
71	71
72	72	Say Y, if you have a Broadcom SoC with AVS support for DFS or DVFS.
73	73
74		−config ARM_BRCMSTB_AVS_CPUFREQ_DEBUG
75		− bool "Broadcom STB AVS CPUfreq driver sysfs debug capability"
76		− depends on ARM_BRCMSTB_AVS_CPUFREQ
77		− help
78		− Enabling this option turns on debug support via sysfs under
79		− /sys/kernel/debug/brcmstb-avs-cpufreq. It is possible to read all and
80		− write some AVS mailbox registers through sysfs entries.
81		−
82		− If in doubt, say N.
83		−
84	74	config ARM_EXYNOS5440_CPUFREQ
85	75	tristate "SAMSUNG EXYNOS5440"
86	76	depends on SOC_EXYNOS5440

drivers/cpufreq/brcmstb-avs-cpufreq.c+1−322

@@ -49,13 +49,6 @@ · 48 unmodified lines
49	49	#include <linux/platform_device.h>
50	50	#include <linux/semaphore.h>
51	51
52		−#ifdef CONFIG_ARM_BRCMSTB_AVS_CPUFREQ_DEBUG
53		−#include <linux/ctype.h>
54		−#include <linux/debugfs.h>
55		−#include <linux/slab.h>
56		−#include <linux/uaccess.h>
57		−#endif
58		−
59	52	/* Max number of arguments AVS calls take */
60	53	#define AVS_MAX_CMD_ARGS 4
61	54	/*
@@ -182,88 +175,11 @@ struct private_data { · 120 unmodified lines
182	175	void __iomem *base;
183	176	void __iomem *avs_intr_base;
184	177	struct device *dev;
185		−#ifdef CONFIG_ARM_BRCMSTB_AVS_CPUFREQ_DEBUG
186		− struct dentry *debugfs;
187		−#endif
188	178	struct completion done;
189	179	struct semaphore sem;
190	180	struct pmap pmap;
191	181	};
192	182
193		−#ifdef CONFIG_ARM_BRCMSTB_AVS_CPUFREQ_DEBUG
194		−
195		−enum debugfs_format {
196		− DEBUGFS_NORMAL,
197		− DEBUGFS_FLOAT,
198		− DEBUGFS_REV,
199		−};
200		−
201		−struct debugfs_data {
202		− struct debugfs_entry *entry;
203		− struct private_data *priv;
204		−};
205		−
206		−struct debugfs_entry {
207		− char *name;
208		− u32 offset;
209		− fmode_t mode;
210		− enum debugfs_format format;
211		−};
212		−
213		−#define DEBUGFS_ENTRY(name, mode, format) { \
214		− #name, AVS_MBOX_##name, mode, format \
215		−}
216		−
217		−/*
218		− * These are used for debugfs only. Otherwise we use AVS_MBOX_PARAM() directly.
219		− */
220		−#define AVS_MBOX_PARAM1 AVS_MBOX_PARAM(0)
221		−#define AVS_MBOX_PARAM2 AVS_MBOX_PARAM(1)
222		−#define AVS_MBOX_PARAM3 AVS_MBOX_PARAM(2)
223		−#define AVS_MBOX_PARAM4 AVS_MBOX_PARAM(3)
224		−
225		−/*
226		− * This table stores the name, access permissions and offset for each hardware
227		− * register and is used to generate debugfs entries.
228		− */
229		−static struct debugfs_entry debugfs_entries[] = {
230		− DEBUGFS_ENTRY(COMMAND, S_IWUSR, DEBUGFS_NORMAL),
231		− DEBUGFS_ENTRY(STATUS, S_IWUSR, DEBUGFS_NORMAL),
232		− DEBUGFS_ENTRY(VOLTAGE0, 0, DEBUGFS_FLOAT),
233		− DEBUGFS_ENTRY(TEMP0, 0, DEBUGFS_FLOAT),
234		− DEBUGFS_ENTRY(PV0, 0, DEBUGFS_FLOAT),
235		− DEBUGFS_ENTRY(MV0, 0, DEBUGFS_FLOAT),
236		− DEBUGFS_ENTRY(PARAM1, S_IWUSR, DEBUGFS_NORMAL),
237		− DEBUGFS_ENTRY(PARAM2, S_IWUSR, DEBUGFS_NORMAL),
238		− DEBUGFS_ENTRY(PARAM3, S_IWUSR, DEBUGFS_NORMAL),
239		− DEBUGFS_ENTRY(PARAM4, S_IWUSR, DEBUGFS_NORMAL),
240		− DEBUGFS_ENTRY(REVISION, 0, DEBUGFS_REV),
241		− DEBUGFS_ENTRY(PSTATE, 0, DEBUGFS_NORMAL),
242		− DEBUGFS_ENTRY(HEARTBEAT, 0, DEBUGFS_NORMAL),
243		− DEBUGFS_ENTRY(MAGIC, S_IWUSR, DEBUGFS_NORMAL),
244		− DEBUGFS_ENTRY(SIGMA_HVT, 0, DEBUGFS_NORMAL),
245		− DEBUGFS_ENTRY(SIGMA_SVT, 0, DEBUGFS_NORMAL),
246		− DEBUGFS_ENTRY(VOLTAGE1, 0, DEBUGFS_FLOAT),
247		− DEBUGFS_ENTRY(TEMP1, 0, DEBUGFS_FLOAT),
248		− DEBUGFS_ENTRY(PV1, 0, DEBUGFS_FLOAT),
249		− DEBUGFS_ENTRY(MV1, 0, DEBUGFS_FLOAT),
250		− DEBUGFS_ENTRY(FREQUENCY, 0, DEBUGFS_NORMAL),
251		−};
252		−
253		−static int brcm_avs_target_index(struct cpufreq_policy *, unsigned int);
254		−
255		−static char __strtolower(char s)
256		−{
257		− char *p;
258		−
259		− for (p = s; *p; p++)
260		− p = tolower(p);
261		−
262		− return s;
263		−}
264		−
265		−#endif /* CONFIG_ARM_BRCMSTB_AVS_CPUFREQ_DEBUG */
266		−
267	183	static void __iomem __map_region(const char name)
268	184	{
269	185	struct device_node *np;
@@ -516,238 +432,6 @@ brcm_avs_get_freq_table(struct device dev, struct private_data priv) · 246 unmodified lines
516	432	return table;
517	433	}
518	434
519		−#ifdef CONFIG_ARM_BRCMSTB_AVS_CPUFREQ_DEBUG
520		−
521		−#define MANT(x) (unsigned int)(abs((x)) / 1000)
522		−#define FRAC(x) (unsigned int)(abs((x)) - abs((x)) / 1000 * 1000)
523		−
524		−static int brcm_avs_debug_show(struct seq_file s, void data)
525		−{
526		− struct debugfs_data *dbgfs = s->private;
527		− void __iomem *base;
528		− u32 val, offset;
529		−
530		− if (!dbgfs) {
531		− seq_puts(s, "No device pointer\n");
532		− return 0;
533		− }
534		−
535		− base = dbgfs->priv->base;
536		− offset = dbgfs->entry->offset;
537		− val = readl(base + offset);
538		− switch (dbgfs->entry->format) {
539		− case DEBUGFS_NORMAL:
540		− seq_printf(s, "%u\n", val);
541		− break;
542		− case DEBUGFS_FLOAT:
543		− seq_printf(s, "%d.%03d\n", MANT(val), FRAC(val));
544		− break;
545		− case DEBUGFS_REV:
546		− seq_printf(s, "%c.%c.%c.%c\n", (val >> 24 & 0xff),
547		− (val >> 16 & 0xff), (val >> 8 & 0xff),
548		− val & 0xff);
549		− break;
550		− }
551		− seq_printf(s, "0x%08x\n", val);
552		−
553		− return 0;
554		−}
555		−
556		−#undef MANT
557		−#undef FRAC
558		−
559		−static ssize_t brcm_avs_seq_write(struct file file, const char __user buf,
560		− size_t size, loff_t *ppos)
561		−{
562		− struct seq_file *s = file->private_data;
563		− struct debugfs_data *dbgfs = s->private;
564		− struct private_data *priv = dbgfs->priv;
565		− void __iomem base, avs_intr_base;
566		− bool use_issue_command = false;
567		− unsigned long val, offset;
568		− char str[128];
569		− int ret;
570		− char *str_ptr = str;
571		−
572		− if (size >= sizeof(str))
573		− return -E2BIG;
574		−
575		− memset(str, 0, sizeof(str));
576		− ret = copy_from_user(str, buf, size);
577		− if (ret)
578		− return ret;
579		−
580		− base = priv->base;
581		− avs_intr_base = priv->avs_intr_base;
582		− offset = dbgfs->entry->offset;
583		− /*
584		− * Special case writing to "command" entry only: if the string starts
585		− * with a 'c', we use the driver's __issue_avs_command() function.
586		− * Otherwise, we perform a raw write. This should allow testing of raw
587		− * access as well as using the higher level function. (Raw access
588		− * doesn't clear the firmware return status after issuing the command.)
589		− */
590		− if (str_ptr[0] == 'c' && offset == AVS_MBOX_COMMAND) {
591		− use_issue_command = true;
592		− str_ptr++;
593		− }
594		− if (kstrtoul(str_ptr, 0, &val) != 0)
595		− return -EINVAL;
596		−
597		− /*
598		− * Setting the P-state is a special case. We need to update the CPU
599		− * frequency we report.
600		− */
601		− if (val == AVS_CMD_SET_PSTATE) {
602		− struct cpufreq_policy *policy;
603		− unsigned int pstate;
604		−
605		− policy = cpufreq_cpu_get(smp_processor_id());
606		− /* Read back the P-state we are about to set */
607		− pstate = readl(base + AVS_MBOX_PARAM(0));
608		− if (use_issue_command) {
609		− ret = brcm_avs_target_index(policy, pstate);
610		− return ret ? ret : size;
611		− }
612		− policy->cur = policy->freq_table[pstate].frequency;
613		− }
614		−
615		− if (use_issue_command) {
616		− ret = __issue_avs_command(priv, val, false, NULL);
617		− } else {
618		− /* Locking here is not perfect, but is only for debug. */
619		− ret = down_interruptible(&priv->sem);
620		− if (ret)
621		− return ret;
622		−
623		− writel(val, base + offset);
624		− /* We have to wake up the firmware to process a command. */
625		− if (offset == AVS_MBOX_COMMAND)
626		− writel(AVS_CPU_L2_INT_MASK,
627		− avs_intr_base + AVS_CPU_L2_SET0);
628		− up(&priv->sem);
629		− }
630		−
631		− return ret ? ret : size;
632		−}
633		−
634		−static struct debugfs_entry __find_debugfs_entry(const char name)
635		−{
636		− int i;
637		−
638		− for (i = 0; i < ARRAY_SIZE(debugfs_entries); i++)
639		− if (strcasecmp(debugfs_entries[i].name, name) == 0)
640		− return &debugfs_entries[i];
641		−
642		− return NULL;
643		−}
644		−
645		−static int brcm_avs_debug_open(struct inode inode, struct file file)
646		−{
647		− struct debugfs_data *data;
648		− fmode_t fmode;
649		− int ret;
650		−
651		− /*
652		− * seq_open(), which is called by single_open(), clears "write" access.
653		− * We need write access to some files, so we preserve our access mode
654		− * and restore it.
655		− */
656		− fmode = file->f_mode;
657		− /*
658		− * Check access permissions even for root. We don't want to be writing
659		− * to read-only registers. Access for regular users has already been
660		− * checked by the VFS layer.
661		− */
662		− if ((fmode & FMODE_WRITER) && !(inode->i_mode & S_IWUSR))
663		− return -EACCES;
664		−
665		− data = kmalloc(sizeof(*data), GFP_KERNEL);
666		− if (!data)
667		− return -ENOMEM;
668		− /*
669		− * We use the same file system operations for all our debug files. To
670		− * produce specific output, we look up the file name upon opening a
671		− * debugfs entry and map it to a memory offset. This offset is then used
672		− * in the generic "show" function to read a specific register.
673		− */
674		− data->entry = __find_debugfs_entry(file->f_path.dentry->d_iname);
675		− data->priv = inode->i_private;
676		−
677		− ret = single_open(file, brcm_avs_debug_show, data);
678		− if (ret)
679		− kfree(data);
680		− file->f_mode = fmode;
681		−
682		− return ret;
683		−}
684		−
685		−static int brcm_avs_debug_release(struct inode inode, struct file file)
686		−{
687		− struct seq_file *seq_priv = file->private_data;
688		− struct debugfs_data *data = seq_priv->private;
689		−
690		− kfree(data);
691		− return single_release(inode, file);
692		−}
693		−
694		−static const struct file_operations brcm_avs_debug_ops = {
695		− .open = brcm_avs_debug_open,
696		− .read = seq_read,
697		− .write = brcm_avs_seq_write,
698		− .llseek = seq_lseek,
699		− .release = brcm_avs_debug_release,
700		−};
701		−
702		−static void brcm_avs_cpufreq_debug_init(struct platform_device *pdev)
703		−{
704		− struct private_data *priv = platform_get_drvdata(pdev);
705		− struct dentry *dir;
706		− int i;
707		−
708		− if (!priv)
709		− return;
710		−
711		− dir = debugfs_create_dir(BRCM_AVS_CPUFREQ_NAME, NULL);
712		− if (IS_ERR_OR_NULL(dir))
713		− return;
714		− priv->debugfs = dir;
715		−
716		− for (i = 0; i < ARRAY_SIZE(debugfs_entries); i++) {
717		− /*
718		− * The DEBUGFS_ENTRY macro generates uppercase strings. We
719		− * convert them to lowercase before creating the debugfs
720		− * entries.
721		− */
722		− char *entry = __strtolower(debugfs_entries[i].name);
723		− fmode_t mode = debugfs_entries[i].mode;
724		−
725		− if (!debugfs_create_file(entry, S_IFREG \| S_IRUGO \| mode,
726		− dir, priv, &brcm_avs_debug_ops)) {
727		− priv->debugfs = NULL;
728		− debugfs_remove_recursive(dir);
729		− break;
730		− }
731		− }
732		−}
733		−
734		−static void brcm_avs_cpufreq_debug_exit(struct platform_device *pdev)
735		−{
736		− struct private_data *priv = platform_get_drvdata(pdev);
737		−
738		− if (priv && priv->debugfs) {
739		− debugfs_remove_recursive(priv->debugfs);
740		− priv->debugfs = NULL;
741		− }
742		−}
743		−
744		−#else
745		−
746		−static void brcm_avs_cpufreq_debug_init(struct platform_device *pdev) {}
747		−static void brcm_avs_cpufreq_debug_exit(struct platform_device *pdev) {}
748		−
749		−#endif /* CONFIG_ARM_BRCMSTB_AVS_CPUFREQ_DEBUG */
750		−
751	435	/*
752	436	* To ensure the right firmware is running we need to
753	437	* - check the MAGIC matches what we expect
@@ -1016,11 +700,8 @@ static int brcm_avs_cpufreq_probe(struct platform_device *pdev) · 262 unmodified lines
1016	700	return ret;
1017	701
1018	702	brcm_avs_driver.driver_data = pdev;
1019		− ret = cpufreq_register_driver(&brcm_avs_driver);
1020		− if (!ret)
1021		− brcm_avs_cpufreq_debug_init(pdev);
1022	703
1023		− return ret;
	704	+ return cpufreq_register_driver(&brcm_avs_driver);
1024	705	}
1025	706
1026	707	static int brcm_avs_cpufreq_remove(struct platform_device *pdev)
@@ -1032,8 +713,6 @@ static int brcm_avs_cpufreq_remove(struct platform_device *pdev) · 5 unmodified lines
1032	713	if (ret)
1033	714	return ret;
1034	715
1035		− brcm_avs_cpufreq_debug_exit(pdev);
1036		−
1037	716	priv = platform_get_drvdata(pdev);
1038	717	iounmap(priv->base);
1039	718	iounmap(priv->avs_intr_base);

drivers/cpufreq/cppc_cpufreq.c+44−2

@@ -126,6 +126,49 @@ static void cppc_cpufreq_stop_cpu(struct cpufreq_policy *policy) · 125 unmodified lines
126	126	cpu->perf_caps.lowest_perf, cpu_num, ret);
127	127	}
128	128
	129	+/*
	130	+ * The PCC subspace describes the rate at which platform can accept commands
	131	+ * on the shared PCC channel (including READs which do not count towards freq
	132	+ * trasition requests), so ideally we need to use the PCC values as a fallback
	133	+ * if we don't have a platform specific transition_delay_us
	134	+ */
	135	+#ifdef CONFIG_ARM64
	136	+#include <asm/cputype.h>
	137	+
	138	+static unsigned int cppc_cpufreq_get_transition_delay_us(int cpu)
	139	+{
	140	+ unsigned long implementor = read_cpuid_implementor();
	141	+ unsigned long part_num = read_cpuid_part_number();
	142	+ unsigned int delay_us = 0;
	143	+
	144	+ switch (implementor) {
	145	+ case ARM_CPU_IMP_QCOM:
	146	+ switch (part_num) {
	147	+ case QCOM_CPU_PART_FALKOR_V1:
	148	+ case QCOM_CPU_PART_FALKOR:
	149	+ delay_us = 10000;
	150	+ break;
	151	+ default:
	152	+ delay_us = cppc_get_transition_latency(cpu) / NSEC_PER_USEC;
	153	+ break;
	154	+ }
	155	+ break;
	156	+ default:
	157	+ delay_us = cppc_get_transition_latency(cpu) / NSEC_PER_USEC;
	158	+ break;
	159	+ }
	160	+
	161	+ return delay_us;
	162	+}
	163	+
	164	+#else
	165	+
	166	+static unsigned int cppc_cpufreq_get_transition_delay_us(int cpu)
	167	+{
	168	+ return cppc_get_transition_latency(cpu) / NSEC_PER_USEC;
	169	+}
	170	+#endif
	171	+
129	172	static int cppc_cpufreq_cpu_init(struct cpufreq_policy *policy)
130	173	{
131	174	struct cppc_cpudata *cpu;
@@ -162,8 +205,7 @@ static int cppc_cpufreq_cpu_init(struct cpufreq_policy *policy) · 30 unmodified lines
162	205	cpu->perf_caps.highest_perf;
163	206	policy->cpuinfo.max_freq = cppc_dmi_max_khz;
164	207
165		− policy->transition_delay_us = cppc_get_transition_latency(cpu_num) /
166		− NSEC_PER_USEC;
	208	+ policy->transition_delay_us = cppc_cpufreq_get_transition_delay_us(cpu_num);
167	209	policy->shared_type = cpu->shared_type;
168	210
169	211	if (policy->shared_type == CPUFREQ_SHARED_TYPE_ANY) {

drivers/cpufreq/powernv-cpufreq.c+11−3

@@ -679,6 +679,16 @@ void gpstate_timer_handler(struct timer_list *t) · 678 unmodified lines
679	679
680	680	if (!spin_trylock(&gpstates->gpstate_lock))
681	681	return;
	682	+ /*
	683	+ * If the timer has migrated to the different cpu then bring
	684	+ * it back to one of the policy->cpus
	685	+ */
	686	+ if (!cpumask_test_cpu(raw_smp_processor_id(), policy->cpus)) {
	687	+ gpstates->timer.expires = jiffies + msecs_to_jiffies(1);
	688	+ add_timer_on(&gpstates->timer, cpumask_first(policy->cpus));
	689	+ spin_unlock(&gpstates->gpstate_lock);
	690	+ return;
	691	+ }
682	692
683	693	/*
684	694	* If PMCR was last updated was using fast_swtich then
@@ -718,10 +728,8 @@ void gpstate_timer_handler(struct timer_list *t) · 33 unmodified lines
718	728	if (gpstate_idx != gpstates->last_lpstate_idx)
719	729	queue_gpstate_timer(gpstates);
720	730
	731	+ set_pstate(&freq_data);
721	732	spin_unlock(&gpstates->gpstate_lock);
722		−
723		− /* Timer may get migrated to a different cpu on cpu hot unplug */
724		− smp_call_function_any(policy->cpus, set_pstate, &freq_data, 1);
725	733	}
726	734
727	735	/*

drivers/dax/device.c+2−0

@@ -19,6 +19,7 @@ · 18 unmodified lines
19	19	#include <linux/dax.h>
20	20	#include <linux/fs.h>
21	21	#include <linux/mm.h>
	22	+#include <linux/mman.h>
22	23	#include "dax-private.h"
23	24	#include "dax.h"
24	25
@@ -540,6 +541,7 @@ static const struct file_operations dax_fops = { · 515 unmodified lines
540	541	.release = dax_release,
541	542	.get_unmapped_area = dax_get_unmapped_area,
542	543	.mmap = dax_mmap,
	544	+ .mmap_supported_flags = MAP_SYNC,
543	545	};
544	546
545	547	static void dev_dax_release(struct device *dev)

drivers/firmware/arm_scmi/clock.c+1−1

@@ -284,7 +284,7 @@ scmi_clock_info_get(const struct scmi_handle *handle, u32 clk_id) · 283 unmodified lines
284	284	struct clock_info *ci = handle->clk_priv;
285	285	struct scmi_clock_info *clk = ci->clk + clk_id;
286	286
287		− if (!clk->name \|\| !clk->name[0])
	287	+ if (!clk->name[0])
288	288	return NULL;
289	289
290	290	return clk;

drivers/fpga/altera-ps-spi.c+1−1

@@ -249,7 +249,7 @@ static int altera_ps_probe(struct spi_device *spi) · 248 unmodified lines
249	249
250	250	conf->data = of_id->data;
251	251	conf->spi = spi;
252		− conf->config = devm_gpiod_get(&spi->dev, "nconfig", GPIOD_OUT_HIGH);
	252	+ conf->config = devm_gpiod_get(&spi->dev, "nconfig", GPIOD_OUT_LOW);
253	253	if (IS_ERR(conf->config)) {
254	254	dev_err(&spi->dev, "Failed to get config gpio: %ld\n",
255	255	PTR_ERR(conf->config));

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c+5−2

@@ -1459,10 +1459,11 @@ static const u32 sgpr_init_compute_shader[] = · 1458 unmodified lines
1459	1459	static const u32 vgpr_init_regs[] =
1460	1460	{
1461	1461	mmCOMPUTE_STATIC_THREAD_MGMT_SE0, 0xffffffff,
1462		− mmCOMPUTE_RESOURCE_LIMITS, 0,
	1462	+ mmCOMPUTE_RESOURCE_LIMITS, 0x1000000, /* CU_GROUP_COUNT=1 */
1463	1463	mmCOMPUTE_NUM_THREAD_X, 256*4,
1464	1464	mmCOMPUTE_NUM_THREAD_Y, 1,
1465	1465	mmCOMPUTE_NUM_THREAD_Z, 1,
	1466	+ mmCOMPUTE_PGM_RSRC1, 0x100004f, /* VGPRS=15 (64 logical VGPRs), SGPRS=1 (16 SGPRs), BULKY=1 */
1466	1467	mmCOMPUTE_PGM_RSRC2, 20,
1467	1468	mmCOMPUTE_USER_DATA_0, 0xedcedc00,
1468	1469	mmCOMPUTE_USER_DATA_1, 0xedcedc01,
@@ -1479,10 +1480,11 @@ static const u32 vgpr_init_regs[] = · 10 unmodified lines
1479	1480	static const u32 sgpr1_init_regs[] =
1480	1481	{
1481	1482	mmCOMPUTE_STATIC_THREAD_MGMT_SE0, 0x0f,
1482		− mmCOMPUTE_RESOURCE_LIMITS, 0x1000000,
	1483	+ mmCOMPUTE_RESOURCE_LIMITS, 0x1000000, /* CU_GROUP_COUNT=1 */
1483	1484	mmCOMPUTE_NUM_THREAD_X, 256*5,
1484	1485	mmCOMPUTE_NUM_THREAD_Y, 1,
1485	1486	mmCOMPUTE_NUM_THREAD_Z, 1,
	1487	+ mmCOMPUTE_PGM_RSRC1, 0x240, /* SGPRS=9 (80 GPRS) */
1486	1488	mmCOMPUTE_PGM_RSRC2, 20,
1487	1489	mmCOMPUTE_USER_DATA_0, 0xedcedc00,
1488	1490	mmCOMPUTE_USER_DATA_1, 0xedcedc01,
@@ -1503,6 +1505,7 @@ static const u32 sgpr2_init_regs[] = · 14 unmodified lines
1503	1505	mmCOMPUTE_NUM_THREAD_X, 256*5,
1504	1506	mmCOMPUTE_NUM_THREAD_Y, 1,
1505	1507	mmCOMPUTE_NUM_THREAD_Z, 1,
	1508	+ mmCOMPUTE_PGM_RSRC1, 0x240, /* SGPRS=9 (80 GPRS) */
1506	1509	mmCOMPUTE_PGM_RSRC2, 20,
1507	1510	mmCOMPUTE_USER_DATA_0, 0xedcedc00,
1508	1511	mmCOMPUTE_USER_DATA_1, 0xedcedc01,

drivers/gpu/drm/amd/amdkfd/Kconfig+1−0

@@ -6,5 +6,6 @@ config HSA_AMD · 5 unmodified lines
6	6	tristate "HSA kernel driver for AMD GPU devices"
7	7	depends on DRM_AMDGPU && X86_64
8	8	imply AMD_IOMMU_V2
	9	+ select MMU_NOTIFIER
9	10	help
10	11	Enable this if you want to use HSA features on AMD GPU devices.

drivers/gpu/drm/amd/amdkfd/kfd_chardev.c+9−8

@@ -749,12 +749,13 @@ static int kfd_ioctl_get_clock_counters(struct file *filep, · 748 unmodified lines
749	749	struct timespec64 time;
750	750
751	751	dev = kfd_device_by_id(args->gpu_id);
752		− if (dev == NULL)
753		− return -EINVAL;
754		−
755		− /* Reading GPU clock counter from KGD */
756		− args->gpu_clock_counter =
757		− dev->kfd2kgd->get_gpu_clock_counter(dev->kgd);
	752	+ if (dev)
	753	+ /* Reading GPU clock counter from KGD */
	754	+ args->gpu_clock_counter =
	755	+ dev->kfd2kgd->get_gpu_clock_counter(dev->kgd);
	756	+ else
	757	+ /* Node without GPU resource */
	758	+ args->gpu_clock_counter = 0;
758	759
759	760	/* No access to rdtsc. Using raw monotonic time */
760	761	getrawmonotonic64(&time);
@@ -1147,7 +1148,7 @@ err_unlock: · 386 unmodified lines
1147	1148	return ret;
1148	1149	}
1149	1150
1150		−bool kfd_dev_is_large_bar(struct kfd_dev *dev)
	1151	+static bool kfd_dev_is_large_bar(struct kfd_dev *dev)
1151	1152	{
1152	1153	struct kfd_local_mem_info mem_info;
1153	1154
@@ -1421,7 +1422,7 @@ static int kfd_ioctl_unmap_memory_from_gpu(struct file *filep, · 267 unmodified lines
1421	1422
1422	1423	pdd = kfd_get_process_device_data(dev, p);
1423	1424	if (!pdd) {
1424		− err = PTR_ERR(pdd);
	1425	+ err = -EINVAL;
1425	1426	goto bind_process_to_device_failed;
1426	1427	}
1427	1428

drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c+9−1

@@ -4557,6 +4557,7 @@ static int dm_update_crtcs_state(struct dc *dc, · 4556 unmodified lines
4557	4557	struct amdgpu_dm_connector *aconnector = NULL;
4558	4558	struct drm_connector_state *new_con_state = NULL;
4559	4559	struct dm_connector_state *dm_conn_state = NULL;
	4560	+ struct drm_plane_state *new_plane_state = NULL;
4560	4561
4561	4562	new_stream = NULL;
4562	4563
@@ -4564,6 +4565,13 @@ static int dm_update_crtcs_state(struct dc *dc, · 1 unmodified line
4564	4565	dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
4565	4566	acrtc = to_amdgpu_crtc(crtc);
4566	4567
	4568	+ new_plane_state = drm_atomic_get_new_plane_state(state, new_crtc_state->crtc->primary);
	4569	+
	4570	+ if (new_crtc_state->enable && new_plane_state && !new_plane_state->fb) {
	4571	+ ret = -EINVAL;
	4572	+ goto fail;
	4573	+ }
	4574	+
4567	4575	aconnector = amdgpu_dm_find_first_crtc_matching_connector(state, crtc);
4568	4576
4569	4577	/* TODO This hack should go away */
@@ -4760,7 +4768,7 @@ static int dm_update_planes_state(struct dc *dc, · 190 unmodified lines
4760	4768	if (!dm_old_crtc_state->stream)
4761	4769	continue;
4762	4770
4763		− DRM_DEBUG_DRIVER("Disabling DRM plane: %d on DRM crtc %d\n",
	4771	+ DRM_DEBUG_ATOMIC("Disabling DRM plane: %d on DRM crtc %d\n",
4764	4772	plane->base.id, old_plane_crtc->base.id);
4765	4773
4766	4774	if (!dc_remove_plane_from_context(

drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_color.c+0−7

@@ -138,13 +138,6 @@ int amdgpu_dm_set_regamma_lut(struct dm_crtc_state *crtc) · 137 unmodified lines
138	138	lut = (struct drm_color_lut *)blob->data;
139	139	lut_size = blob->length / sizeof(struct drm_color_lut);
140	140
141		− if (__is_lut_linear(lut, lut_size)) {
142		− /* Set to bypass if lut is set to linear */
143		− stream->out_transfer_func->type = TF_TYPE_BYPASS;
144		− stream->out_transfer_func->tf = TRANSFER_FUNCTION_LINEAR;
145		− return 0;
146		− }
147		−
148	141	gamma = dc_create_gamma();
149	142	if (!gamma)
150	143	return -ENOMEM;

drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c+3−2

@@ -329,14 +329,15 @@ void amdgpu_dm_irq_fini(struct amdgpu_device *adev) · 328 unmodified lines
329	329	{
330	330	int src;
331	331	struct irq_list_head *lh;
	332	+ unsigned long irq_table_flags;
332	333	DRM_DEBUG_KMS("DM_IRQ: releasing resources.\n");
333		−
334	334	for (src = 0; src < DAL_IRQ_SOURCES_NUMBER; src++) {
335		−
	335	+ DM_IRQ_TABLE_LOCK(adev, irq_table_flags);
336	336	/* The handler was removed from the table,
337	337	* it means it is safe to flush all the 'work'
338	338	* (because no code can schedule a new one). */
339	339	lh = &adev->dm.irq_handler_list_low_tab[src];
	340	+ DM_IRQ_TABLE_UNLOCK(adev, irq_table_flags);
340	341	flush_work(&lh->work);
341	342	}
342	343	}

drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c+22−32

@@ -161,6 +161,11 @@ dm_dp_mst_connector_destroy(struct drm_connector *connector) · 160 unmodified lines
161	161	struct amdgpu_dm_connector *amdgpu_dm_connector = to_amdgpu_dm_connector(connector);
162	162	struct amdgpu_encoder *amdgpu_encoder = amdgpu_dm_connector->mst_encoder;
163	163
	164	+ if (amdgpu_dm_connector->edid) {
	165	+ kfree(amdgpu_dm_connector->edid);
	166	+ amdgpu_dm_connector->edid = NULL;
	167	+ }
	168	+
164	169	drm_encoder_cleanup(&amdgpu_encoder->base);
165	170	kfree(amdgpu_encoder);
166	171	drm_connector_cleanup(connector);
@@ -181,28 +186,22 @@ static const struct drm_connector_funcs dm_dp_mst_connector_funcs = { · 14 unmodified lines
181	186	void dm_dp_mst_dc_sink_create(struct drm_connector *connector)
182	187	{
183	188	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
184		− struct edid *edid;
185	189	struct dc_sink *dc_sink;
186	190	struct dc_sink_init_data init_params = {
187	191	.link = aconnector->dc_link,
188	192	.sink_signal = SIGNAL_TYPE_DISPLAY_PORT_MST };
189	193
	194	+ /* FIXME none of this is safe. we shouldn't touch aconnector here in
	195	+ * atomic_check
	196	+ */
	197	+
190	198	/*
191	199	* TODO: Need to further figure out why ddc.algo is NULL while MST port exists
192	200	*/
193	201	if (!aconnector->port \|\| !aconnector->port->aux.ddc.algo)
194	202	return;
195	203
196		− edid = drm_dp_mst_get_edid(connector, &aconnector->mst_port->mst_mgr, aconnector->port);
197		−
198		− if (!edid) {
199		− drm_mode_connector_update_edid_property(
200		− &aconnector->base,
201		− NULL);
202		− return;
203		− }
204		−
205		− aconnector->edid = edid;
	204	+ ASSERT(aconnector->edid);
206	205
207	206	dc_sink = dc_link_add_remote_sink(
208	207	aconnector->dc_link,
@@ -215,9 +214,6 @@ void dm_dp_mst_dc_sink_create(struct drm_connector *connector) · 6 unmodified lines
215	214
216	215	amdgpu_dm_add_sink_to_freesync_module(
217	216	connector, aconnector->edid);
218		−
219		− drm_mode_connector_update_edid_property(
220		− &aconnector->base, aconnector->edid);
221	217	}
222	218
223	219	static int dm_dp_mst_get_modes(struct drm_connector *connector)
@@ -230,10 +226,6 @@ static int dm_dp_mst_get_modes(struct drm_connector *connector) · 6 unmodified lines
230	226
231	227	if (!aconnector->edid) {
232	228	struct edid *edid;
233		− struct dc_sink *dc_sink;
234		− struct dc_sink_init_data init_params = {
235		− .link = aconnector->dc_link,
236		− .sink_signal = SIGNAL_TYPE_DISPLAY_PORT_MST };
237	229	edid = drm_dp_mst_get_edid(connector, &aconnector->mst_port->mst_mgr, aconnector->port);
238	230
239	231	if (!edid) {
@@ -244,11 +236,17 @@ static int dm_dp_mst_get_modes(struct drm_connector *connector) · 4 unmodified lines
244	236	}
245	237
246	238	aconnector->edid = edid;
	239	+ }
247	240
	241	+ if (!aconnector->dc_sink) {
	242	+ struct dc_sink *dc_sink;
	243	+ struct dc_sink_init_data init_params = {
	244	+ .link = aconnector->dc_link,
	245	+ .sink_signal = SIGNAL_TYPE_DISPLAY_PORT_MST };
248	246	dc_sink = dc_link_add_remote_sink(
249	247	aconnector->dc_link,
250		− (uint8_t *)edid,
251		− (edid->extensions + 1) * EDID_LENGTH,
	248	+ (uint8_t *)aconnector->edid,
	249	+ (aconnector->edid->extensions + 1) * EDID_LENGTH,
252	250	&init_params);
253	251
254	252	dc_sink->priv = aconnector;
@@ -256,12 +254,12 @@ static int dm_dp_mst_get_modes(struct drm_connector *connector) · 1 unmodified line
256	254
257	255	if (aconnector->dc_sink)
258	256	amdgpu_dm_add_sink_to_freesync_module(
259		− connector, edid);
260		−
261		− drm_mode_connector_update_edid_property(
262		− &aconnector->base, edid);
	257	+ connector, aconnector->edid);
263	258	}
264	259
	260	+ drm_mode_connector_update_edid_property(
	261	+ &aconnector->base, aconnector->edid);
	262	+
265	263	ret = drm_add_edid_modes(connector, aconnector->edid);
266	264
267	265	return ret;
@@ -424,14 +422,6 @@ static void dm_dp_destroy_mst_connector(struct drm_dp_mst_topology_mgr *mgr, · 156 unmodified lines
424	422	dc_sink_release(aconnector->dc_sink);
425	423	aconnector->dc_sink = NULL;
426	424	}
427		− if (aconnector->edid) {
428		− kfree(aconnector->edid);
429		− aconnector->edid = NULL;
430		− }
431		−
432		− drm_mode_connector_update_edid_property(
433		− &aconnector->base,
434		− NULL);
435	425
436	426	aconnector->mst_connected = false;
437	427	}

drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c+10−6

@@ -4743,23 +4743,27 @@ static void smu7_check_dpm_table_updated(struct pp_hwmgr *hwmgr) · 4742 unmodified lines
4743	4743
4744	4744	for (i=0; i < dep_table->count; i++) {
4745	4745	if (dep_table->entries[i].vddc != odn_dep_table->entries[i].vddc) {
4746		− data->need_update_smu7_dpm_table \|= DPMTABLE_OD_UPDATE_VDDC;
4747		− break;
	4746	+ data->need_update_smu7_dpm_table \|= DPMTABLE_OD_UPDATE_VDDC \| DPMTABLE_OD_UPDATE_MCLK;
	4747	+ return;
4748	4748	}
4749	4749	}
4750		− if (i == dep_table->count)
	4750	+ if (i == dep_table->count && data->need_update_smu7_dpm_table & DPMTABLE_OD_UPDATE_VDDC) {
4751	4751	data->need_update_smu7_dpm_table &= ~DPMTABLE_OD_UPDATE_VDDC;
	4752	+ data->need_update_smu7_dpm_table \|= DPMTABLE_OD_UPDATE_MCLK;
	4753	+ }
4752	4754
4753	4755	dep_table = table_info->vdd_dep_on_sclk;
4754	4756	odn_dep_table = (struct phm_ppt_v1_clock_voltage_dependency_table *)&(odn_table->vdd_dependency_on_sclk);
4755	4757	for (i=0; i < dep_table->count; i++) {
4756	4758	if (dep_table->entries[i].vddc != odn_dep_table->entries[i].vddc) {
4757		− data->need_update_smu7_dpm_table \|= DPMTABLE_OD_UPDATE_VDDC;
4758		− break;
	4759	+ data->need_update_smu7_dpm_table \|= DPMTABLE_OD_UPDATE_VDDC \| DPMTABLE_OD_UPDATE_SCLK;
	4760	+ return;
4759	4761	}
4760	4762	}
4761		− if (i == dep_table->count)
	4763	+ if (i == dep_table->count && data->need_update_smu7_dpm_table & DPMTABLE_OD_UPDATE_VDDC) {
4762	4764	data->need_update_smu7_dpm_table &= ~DPMTABLE_OD_UPDATE_VDDC;
	4765	+ data->need_update_smu7_dpm_table \|= DPMTABLE_OD_UPDATE_SCLK;
	4766	+ }
4763	4767	}
4764	4768
4765	4769	static int smu7_odn_edit_dpm_table(struct pp_hwmgr *hwmgr,

drivers/gpu/drm/amd/powerplay/inc/vega12/smu9_driver_if.h+3−1

@@ -412,8 +412,10 @@ typedef struct { · 411 unmodified lines
412	412	QuadraticInt_t ReservedEquation2;
413	413	QuadraticInt_t ReservedEquation3;
414	414
	415	+ uint16_t MinVoltageUlvGfx;
	416	+ uint16_t MinVoltageUlvSoc;
415	417
416		− uint32_t Reserved[15];
	418	+ uint32_t Reserved[14];
417	419
418	420
419	421

drivers/gpu/drm/bridge/dumb-vga-dac.c+3−1

@@ -56,7 +56,9 @@ static int dumb_vga_get_modes(struct drm_connector *connector) · 55 unmodified lines
56	56	}
57	57
58	58	drm_mode_connector_update_edid_property(connector, edid);
59		− return drm_add_edid_modes(connector, edid);
	59	+ ret = drm_add_edid_modes(connector, edid);
	60	+ kfree(edid);
	61	+ return ret;
60	62
61	63	fallback:
62	64	/*

drivers/gpu/drm/drm_dp_dual_mode_helper.c+32−7

@@ -350,19 +350,44 @@ int drm_dp_dual_mode_set_tmds_output(enum drm_dp_dual_mode_type type, · 349 unmodified lines
350	350	{
351	351	uint8_t tmds_oen = enable ? 0 : DP_DUAL_MODE_TMDS_DISABLE;
352	352	ssize_t ret;
	353	+ int retry;
353	354
354	355	if (type < DRM_DP_DUAL_MODE_TYPE2_DVI)
355	356	return 0;
356	357
357		− ret = drm_dp_dual_mode_write(adapter, DP_DUAL_MODE_TMDS_OEN,
358		− &tmds_oen, sizeof(tmds_oen));
359		− if (ret) {
360		− DRM_DEBUG_KMS("Failed to %s TMDS output buffers\n",
361		− enable ? "enable" : "disable");
362		− return ret;
	358	+ /*
	359	+ * LSPCON adapters in low-power state may ignore the first write, so
	360	+ * read back and verify the written value a few times.
	361	+ */
	362	+ for (retry = 0; retry < 3; retry++) {
	363	+ uint8_t tmp;
	364	+
	365	+ ret = drm_dp_dual_mode_write(adapter, DP_DUAL_MODE_TMDS_OEN,
	366	+ &tmds_oen, sizeof(tmds_oen));
	367	+ if (ret) {
	368	+ DRM_DEBUG_KMS("Failed to %s TMDS output buffers (%d attempts)\n",
	369	+ enable ? "enable" : "disable",
	370	+ retry + 1);
	371	+ return ret;
	372	+ }
	373	+
	374	+ ret = drm_dp_dual_mode_read(adapter, DP_DUAL_MODE_TMDS_OEN,
	375	+ &tmp, sizeof(tmp));
	376	+ if (ret) {
	377	+ DRM_DEBUG_KMS("I2C read failed during TMDS output buffer %s (%d attempts)\n",
	378	+ enable ? "enabling" : "disabling",
	379	+ retry + 1);
	380	+ return ret;
	381	+ }
	382	+
	383	+ if (tmp == tmds_oen)
	384	+ return 0;
363	385	}
364	386
365		− return 0;
	387	+ DRM_DEBUG_KMS("I2C write value mismatch during TMDS output buffer %s\n",
	388	+ enable ? "enabling" : "disabling");
	389	+
	390	+ return -EIO;
366	391	}
367	392	EXPORT_SYMBOL(drm_dp_dual_mode_set_tmds_output);
368	393

drivers/gpu/drm/drm_edid.c+3−8

@@ -4451,6 +4451,7 @@ drm_reset_display_info(struct drm_connector *connector) · 4450 unmodified lines
4451	4451	info->max_tmds_clock = 0;
4452	4452	info->dvi_dual = false;
4453	4453	info->has_hdmi_infoframe = false;
	4454	+ memset(&info->hdmi, 0, sizeof(info->hdmi));
4454	4455
4455	4456	info->non_desktop = 0;
4456	4457	}
@@ -4462,17 +4463,11 @@ u32 drm_add_display_info(struct drm_connector connector, const struct edid edi · 5 unmodified lines
4462	4463
4463	4464	u32 quirks = edid_get_quirks(edid);
4464	4465
	4466	+ drm_reset_display_info(connector);
	4467	+
4465	4468	info->width_mm = edid->width_cm * 10;
4466	4469	info->height_mm = edid->height_cm * 10;
4467	4470
4468		− /* driver figures it out in this case */
4469		− info->bpc = 0;
4470		− info->color_formats = 0;
4471		− info->cea_rev = 0;
4472		− info->max_tmds_clock = 0;
4473		− info->dvi_dual = false;
4474		− info->has_hdmi_infoframe = false;
4475		−
4476	4471	info->non_desktop = !!(quirks & EDID_QUIRK_NON_DESKTOP);
4477	4472
4478	4473	DRM_DEBUG_KMS("non_desktop set to %d\n", info->non_desktop);

drivers/gpu/drm/exynos/exynos_drm_fb.c+14−59

@@ -18,6 +18,7 @@ · 17 unmodified lines
18	18	#include <drm/drm_fb_helper.h>
19	19	#include <drm/drm_atomic.h>
20	20	#include <drm/drm_atomic_helper.h>
	21	+#include <drm/drm_gem_framebuffer_helper.h>
21	22	#include <uapi/drm/exynos_drm.h>
22	23
23	24	#include "exynos_drm_drv.h"
@@ -26,20 +27,6 @@ · 2 unmodified lines
26	27	#include "exynos_drm_iommu.h"
27	28	#include "exynos_drm_crtc.h"
28	29
29		−#define to_exynos_fb(x) container_of(x, struct exynos_drm_fb, fb)
30		−
31		−/*
32		− * exynos specific framebuffer structure.
33		− *
34		− * @fb: drm framebuffer obejct.
35		− * @exynos_gem: array of exynos specific gem object containing a gem object.
36		− */
37		−struct exynos_drm_fb {
38		− struct drm_framebuffer fb;
39		− struct exynos_drm_gem *exynos_gem[MAX_FB_BUFFER];
40		− dma_addr_t dma_addr[MAX_FB_BUFFER];
41		−};
42		−
43	30	static int check_fb_gem_memory_type(struct drm_device *drm_dev,
44	31	struct exynos_drm_gem *exynos_gem)
45	32	{
@@ -66,40 +53,9 @@ static int check_fb_gem_memory_type(struct drm_device *drm_dev, · 20 unmodified lines
66	53	return 0;
67	54	}
68	55
69		−static void exynos_drm_fb_destroy(struct drm_framebuffer *fb)
70		−{
71		− struct exynos_drm_fb *exynos_fb = to_exynos_fb(fb);
72		− unsigned int i;
73		−
74		− drm_framebuffer_cleanup(fb);
75		−
76		− for (i = 0; i < ARRAY_SIZE(exynos_fb->exynos_gem); i++) {
77		− struct drm_gem_object *obj;
78		−
79		− if (exynos_fb->exynos_gem[i] == NULL)
80		− continue;
81		−
82		− obj = &exynos_fb->exynos_gem[i]->base;
83		− drm_gem_object_unreference_unlocked(obj);
84		− }
85		−
86		− kfree(exynos_fb);
87		− exynos_fb = NULL;
88		−}
89		−
90		−static int exynos_drm_fb_create_handle(struct drm_framebuffer *fb,
91		− struct drm_file *file_priv,
92		− unsigned int *handle)
93		−{
94		− struct exynos_drm_fb *exynos_fb = to_exynos_fb(fb);
95		−
96		− return drm_gem_handle_create(file_priv,
97		− &exynos_fb->exynos_gem[0]->base, handle);
98		−}
99		−
100	56	static const struct drm_framebuffer_funcs exynos_drm_fb_funcs = {
101		− .destroy = exynos_drm_fb_destroy,
102		− .create_handle = exynos_drm_fb_create_handle,
	57	+ .destroy = drm_gem_fb_destroy,
	58	+ .create_handle = drm_gem_fb_create_handle,
103	59	};
104	60
105	61	struct drm_framebuffer *
@@ -108,12 +64,12 @@ exynos_drm_framebuffer_init(struct drm_device *dev, · 2 unmodified lines
108	64	struct exynos_drm_gem **exynos_gem,
109	65	int count)
110	66	{
111		− struct exynos_drm_fb *exynos_fb;
	67	+ struct drm_framebuffer *fb;
112	68	int i;
113	69	int ret;
114	70
115		− exynos_fb = kzalloc(sizeof(*exynos_fb), GFP_KERNEL);
116		− if (!exynos_fb)
	71	+ fb = kzalloc(sizeof(*fb), GFP_KERNEL);
	72	+ if (!fb)
117	73	return ERR_PTR(-ENOMEM);
118	74
119	75	for (i = 0; i < count; i++) {
@@ -121,23 +77,21 @@ exynos_drm_framebuffer_init(struct drm_device *dev, · 1 unmodified line
121	77	if (ret < 0)
122	78	goto err;
123	79
124		− exynos_fb->exynos_gem[i] = exynos_gem[i];
125		− exynos_fb->dma_addr[i] = exynos_gem[i]->dma_addr
126		− + mode_cmd->offsets[i];
	80	+ fb->obj[i] = &exynos_gem[i]->base;
127	81	}
128	82
129		− drm_helper_mode_fill_fb_struct(dev, &exynos_fb->fb, mode_cmd);
	83	+ drm_helper_mode_fill_fb_struct(dev, fb, mode_cmd);
130	84
131		− ret = drm_framebuffer_init(dev, &exynos_fb->fb, &exynos_drm_fb_funcs);
	85	+ ret = drm_framebuffer_init(dev, fb, &exynos_drm_fb_funcs);
132	86	if (ret < 0) {
133	87	DRM_ERROR("failed to initialize framebuffer\n");
134	88	goto err;
135	89	}
136	90
137		− return &exynos_fb->fb;
	91	+ return fb;
138	92
139	93	err:
140		− kfree(exynos_fb);
	94	+ kfree(fb);
141	95	return ERR_PTR(ret);
142	96	}
143	97
@@ -191,12 +145,13 @@ err: · 47 unmodified lines
191	145
192	146	dma_addr_t exynos_drm_fb_dma_addr(struct drm_framebuffer *fb, int index)
193	147	{
194		− struct exynos_drm_fb *exynos_fb = to_exynos_fb(fb);
	148	+ struct exynos_drm_gem *exynos_gem;
195	149
196	150	if (WARN_ON_ONCE(index >= MAX_FB_BUFFER))
197	151	return 0;
198	152
199		− return exynos_fb->dma_addr[index];
	153	+ exynos_gem = to_exynos_gem(fb->obj[index]);
	154	+ return exynos_gem->dma_addr + fb->offsets[index];
200	155	}
201	156
202	157	static struct drm_mode_config_helper_funcs exynos_drm_mode_config_helpers = {

drivers/gpu/drm/i915/gvt/cmd_parser.c+1−0

@@ -1080,6 +1080,7 @@ static int cmd_handler_mi_user_interrupt(struct parser_exec_state *s) · 1079 unmodified lines
1080	1080	{
1081	1081	set_bit(cmd_interrupt_events[s->ring_id].mi_user_interrupt,
1082	1082	s->workload->pending_events);
	1083	+ patch_value(s, cmd_ptr(s, 0), MI_NOOP);
1083	1084	return 0;
1084	1085	}
1085	1086

drivers/gpu/drm/i915/gvt/display.c+10−0

@@ -169,6 +169,8 @@ static u8 dpcd_fix_data[DPCD_HEADER_SIZE] = { · 168 unmodified lines
169	169	static void emulate_monitor_status_change(struct intel_vgpu *vgpu)
170	170	{
171	171	struct drm_i915_private *dev_priv = vgpu->gvt->dev_priv;
	172	+ int pipe;
	173	+
172	174	vgpu_vreg_t(vgpu, SDEISR) &= ~(SDE_PORTB_HOTPLUG_CPT \|
173	175	SDE_PORTC_HOTPLUG_CPT \|
174	176	SDE_PORTD_HOTPLUG_CPT);
@@ -267,6 +269,14 @@ static void emulate_monitor_status_change(struct intel_vgpu *vgpu) · 92 unmodified lines
267	269	if (IS_BROADWELL(dev_priv))
268	270	vgpu_vreg_t(vgpu, PCH_ADPA) &= ~ADPA_CRT_HOTPLUG_MONITOR_MASK;
269	271
	272	+ /* Disable Primary/Sprite/Cursor plane */
	273	+ for_each_pipe(dev_priv, pipe) {
	274	+ vgpu_vreg_t(vgpu, DSPCNTR(pipe)) &= ~DISPLAY_PLANE_ENABLE;
	275	+ vgpu_vreg_t(vgpu, SPRCTL(pipe)) &= ~SPRITE_ENABLE;
	276	+ vgpu_vreg_t(vgpu, CURCNTR(pipe)) &= ~CURSOR_MODE;
	277	+ vgpu_vreg_t(vgpu, CURCNTR(pipe)) \|= CURSOR_MODE_DISABLE;
	278	+ }
	279	+
270	280	vgpu_vreg_t(vgpu, PIPECONF(PIPE_A)) \|= PIPECONF_ENABLE;
271	281	}
272	282

drivers/gpu/drm/i915/gvt/dmabuf.c+1−0

@@ -323,6 +323,7 @@ static void update_fb_info(struct vfio_device_gfx_plane_info *gvt_dmabuf, · 322 unmodified lines
323	323	struct intel_vgpu_fb_info *fb_info)
324	324	{
325	325	gvt_dmabuf->drm_format = fb_info->drm_format;
	326	+ gvt_dmabuf->drm_format_mod = fb_info->drm_format_mod;
326	327	gvt_dmabuf->width = fb_info->width;
327	328	gvt_dmabuf->height = fb_info->height;
328	329	gvt_dmabuf->stride = fb_info->stride;

drivers/gpu/drm/i915/gvt/fb_decoder.c+9−18

@@ -245,16 +245,13 @@ int intel_vgpu_decode_primary_plane(struct intel_vgpu *vgpu, · 244 unmodified lines
245	245	plane->hw_format = fmt;
246	246
247	247	plane->base = vgpu_vreg_t(vgpu, DSPSURF(pipe)) & I915_GTT_PAGE_MASK;
248		− if (!intel_gvt_ggtt_validate_range(vgpu, plane->base, 0)) {
249		− gvt_vgpu_err("invalid gma address: %lx\n",
250		− (unsigned long)plane->base);
	248	+ if (!intel_gvt_ggtt_validate_range(vgpu, plane->base, 0))
251	249	return -EINVAL;
252		− }
253	250
254	251	plane->base_gpa = intel_vgpu_gma_to_gpa(vgpu->gtt.ggtt_mm, plane->base);
255	252	if (plane->base_gpa == INTEL_GVT_INVALID_ADDR) {
256		− gvt_vgpu_err("invalid gma address: %lx\n",
257		− (unsigned long)plane->base);
	253	+ gvt_vgpu_err("Translate primary plane gma 0x%x to gpa fail\n",
	254	+ plane->base);
258	255	return -EINVAL;
259	256	}
260	257
@@ -371,16 +368,13 @@ int intel_vgpu_decode_cursor_plane(struct intel_vgpu *vgpu, · 110 unmodified lines
371	368	alpha_plane, alpha_force);
372	369
373	370	plane->base = vgpu_vreg_t(vgpu, CURBASE(pipe)) & I915_GTT_PAGE_MASK;
374		− if (!intel_gvt_ggtt_validate_range(vgpu, plane->base, 0)) {
375		− gvt_vgpu_err("invalid gma address: %lx\n",
376		− (unsigned long)plane->base);
	371	+ if (!intel_gvt_ggtt_validate_range(vgpu, plane->base, 0))
377	372	return -EINVAL;
378		− }
379	373
380	374	plane->base_gpa = intel_vgpu_gma_to_gpa(vgpu->gtt.ggtt_mm, plane->base);
381	375	if (plane->base_gpa == INTEL_GVT_INVALID_ADDR) {
382		− gvt_vgpu_err("invalid gma address: %lx\n",
383		− (unsigned long)plane->base);
	376	+ gvt_vgpu_err("Translate cursor plane gma 0x%x to gpa fail\n",
	377	+ plane->base);
384	378	return -EINVAL;
385	379	}
386	380
@@ -476,16 +470,13 @@ int intel_vgpu_decode_sprite_plane(struct intel_vgpu *vgpu, · 89 unmodified lines
476	470	plane->drm_format = drm_format;
477	471
478	472	plane->base = vgpu_vreg_t(vgpu, SPRSURF(pipe)) & I915_GTT_PAGE_MASK;
479		− if (!intel_gvt_ggtt_validate_range(vgpu, plane->base, 0)) {
480		− gvt_vgpu_err("invalid gma address: %lx\n",
481		− (unsigned long)plane->base);
	473	+ if (!intel_gvt_ggtt_validate_range(vgpu, plane->base, 0))
482	474	return -EINVAL;
483		− }
484	475
485	476	plane->base_gpa = intel_vgpu_gma_to_gpa(vgpu->gtt.ggtt_mm, plane->base);
486	477	if (plane->base_gpa == INTEL_GVT_INVALID_ADDR) {
487		− gvt_vgpu_err("invalid gma address: %lx\n",
488		− (unsigned long)plane->base);
	478	+ gvt_vgpu_err("Translate sprite plane gma 0x%x to gpa fail\n",
	479	+ plane->base);
489	480	return -EINVAL;
490	481	}
491	482

drivers/gpu/drm/i915/gvt/gtt.c+45−7

@@ -530,6 +530,16 @@ static void ggtt_set_guest_entry(struct intel_vgpu_mm *mm, · 529 unmodified lines
530	530	false, 0, mm->vgpu);
531	531	}
532	532
	533	+static void ggtt_get_host_entry(struct intel_vgpu_mm *mm,
	534	+ struct intel_gvt_gtt_entry *entry, unsigned long index)
	535	+{
	536	+ struct intel_gvt_gtt_pte_ops *pte_ops = mm->vgpu->gvt->gtt.pte_ops;
	537	+
	538	+ GEM_BUG_ON(mm->type != INTEL_GVT_MM_GGTT);
	539	+
	540	+ pte_ops->get_entry(NULL, entry, index, false, 0, mm->vgpu);
	541	+}
	542	+
533	543	static void ggtt_set_host_entry(struct intel_vgpu_mm *mm,
534	544	struct intel_gvt_gtt_entry *entry, unsigned long index)
535	545	{
@@ -1818,6 +1828,18 @@ int intel_vgpu_emulate_ggtt_mmio_read(struct intel_vgpu *vgpu, unsigned int off, · 1282 unmodified lines
1818	1828	return ret;
1819	1829	}
1820	1830
	1831	+static void ggtt_invalidate_pte(struct intel_vgpu *vgpu,
	1832	+ struct intel_gvt_gtt_entry *entry)
	1833	+{
	1834	+ struct intel_gvt_gtt_pte_ops *pte_ops = vgpu->gvt->gtt.pte_ops;
	1835	+ unsigned long pfn;
	1836	+
	1837	+ pfn = pte_ops->get_pfn(entry);
	1838	+ if (pfn != vgpu->gvt->gtt.scratch_mfn)
	1839	+ intel_gvt_hypervisor_dma_unmap_guest_page(vgpu,
	1840	+ pfn << PAGE_SHIFT);
	1841	+}
	1842	+
1821	1843	static int emulate_ggtt_mmio_write(struct intel_vgpu *vgpu, unsigned int off,
1822	1844	void *p_data, unsigned int bytes)
1823	1845	{
@@ -1844,10 +1866,10 @@ static int emulate_ggtt_mmio_write(struct intel_vgpu *vgpu, unsigned int off, · 20 unmodified lines
1844	1866
1845	1867	memcpy((void *)&e.val64 + (off & (info->gtt_entry_size - 1)), p_data,
1846	1868	bytes);
1847		− m = e;
1848	1869
1849	1870	if (ops->test_present(&e)) {
1850	1871	gfn = ops->get_pfn(&e);
	1872	+ m = e;
1851	1873
1852	1874	/* one PTE update may be issued in multiple writes and the
1853	1875	* first write may not construct a valid gfn
@@ -1868,8 +1890,12 @@ static int emulate_ggtt_mmio_write(struct intel_vgpu *vgpu, unsigned int off, · 14 unmodified lines
1868	1890	ops->set_pfn(&m, gvt->gtt.scratch_mfn);
1869	1891	} else
1870	1892	ops->set_pfn(&m, dma_addr >> PAGE_SHIFT);
1871		− } else
	1893	+ } else {
	1894	+ ggtt_get_host_entry(ggtt_mm, &m, g_gtt_index);
	1895	+ ggtt_invalidate_pte(vgpu, &m);
1872	1896	ops->set_pfn(&m, gvt->gtt.scratch_mfn);
	1897	+ ops->clear_present(&m);
	1898	+ }
1873	1899
1874	1900	out:
1875	1901	ggtt_set_host_entry(ggtt_mm, &m, g_gtt_index);
@@ -2030,7 +2056,7 @@ int intel_vgpu_init_gtt(struct intel_vgpu *vgpu) · 154 unmodified lines
2030	2056	return PTR_ERR(gtt->ggtt_mm);
2031	2057	}
2032	2058
2033		− intel_vgpu_reset_ggtt(vgpu);
	2059	+ intel_vgpu_reset_ggtt(vgpu, false);
2034	2060
2035	2061	return create_scratch_page_tree(vgpu);
2036	2062	}
@@ -2315,17 +2341,19 @@ void intel_vgpu_invalidate_ppgtt(struct intel_vgpu *vgpu) · 278 unmodified lines
2315	2341	/**
2316	2342	* intel_vgpu_reset_ggtt - reset the GGTT entry
2317	2343	* @vgpu: a vGPU
	2344	+ * @invalidate_old: invalidate old entries
2318	2345	*
2319	2346	* This function is called at the vGPU create stage
2320	2347	* to reset all the GGTT entries.
2321	2348	*
2322	2349	*/
2323		−void intel_vgpu_reset_ggtt(struct intel_vgpu *vgpu)
	2350	+void intel_vgpu_reset_ggtt(struct intel_vgpu *vgpu, bool invalidate_old)
2324	2351	{
2325	2352	struct intel_gvt *gvt = vgpu->gvt;
2326	2353	struct drm_i915_private *dev_priv = gvt->dev_priv;
2327	2354	struct intel_gvt_gtt_pte_ops *pte_ops = vgpu->gvt->gtt.pte_ops;
2328	2355	struct intel_gvt_gtt_entry entry = {.type = GTT_TYPE_GGTT_PTE};
	2356	+ struct intel_gvt_gtt_entry old_entry;
2329	2357	u32 index;
2330	2358	u32 num_entries;
2331	2359
@@ -2334,13 +2362,23 @@ void intel_vgpu_reset_ggtt(struct intel_vgpu *vgpu) · 2 unmodified lines
2334	2362
2335	2363	index = vgpu_aperture_gmadr_base(vgpu) >> PAGE_SHIFT;
2336	2364	num_entries = vgpu_aperture_sz(vgpu) >> PAGE_SHIFT;
2337		− while (num_entries--)
	2365	+ while (num_entries--) {
	2366	+ if (invalidate_old) {
	2367	+ ggtt_get_host_entry(vgpu->gtt.ggtt_mm, &old_entry, index);
	2368	+ ggtt_invalidate_pte(vgpu, &old_entry);
	2369	+ }
2338	2370	ggtt_set_host_entry(vgpu->gtt.ggtt_mm, &entry, index++);
	2371	+ }
2339	2372
2340	2373	index = vgpu_hidden_gmadr_base(vgpu) >> PAGE_SHIFT;
2341	2374	num_entries = vgpu_hidden_sz(vgpu) >> PAGE_SHIFT;
2342		− while (num_entries--)
	2375	+ while (num_entries--) {
	2376	+ if (invalidate_old) {
	2377	+ ggtt_get_host_entry(vgpu->gtt.ggtt_mm, &old_entry, index);
	2378	+ ggtt_invalidate_pte(vgpu, &old_entry);
	2379	+ }
2343	2380	ggtt_set_host_entry(vgpu->gtt.ggtt_mm, &entry, index++);
	2381	+ }
2344	2382
2345	2383	ggtt_invalidate(dev_priv);
2346	2384	}
@@ -2360,5 +2398,5 @@ void intel_vgpu_reset_gtt(struct intel_vgpu *vgpu) · 13 unmodified lines
2360	2398	* removing the shadow pages.
2361	2399	*/
2362	2400	intel_vgpu_destroy_all_ppgtt_mm(vgpu);
2363		− intel_vgpu_reset_ggtt(vgpu);
	2401	+ intel_vgpu_reset_ggtt(vgpu, true);
2364	2402	}

drivers/gpu/drm/i915/gvt/gtt.h+1−1

@@ -193,7 +193,7 @@ struct intel_vgpu_gtt { · 192 unmodified lines
193	193
194	194	extern int intel_vgpu_init_gtt(struct intel_vgpu *vgpu);
195	195	extern void intel_vgpu_clean_gtt(struct intel_vgpu *vgpu);
196		−void intel_vgpu_reset_ggtt(struct intel_vgpu *vgpu);
	196	+void intel_vgpu_reset_ggtt(struct intel_vgpu *vgpu, bool invalidate_old);
197	197	void intel_vgpu_invalidate_ppgtt(struct intel_vgpu *vgpu);
198	198
199	199	extern int intel_gvt_init_gtt(struct intel_gvt *gvt);

drivers/gpu/drm/i915/gvt/handlers.c+1−0

@@ -1150,6 +1150,7 @@ static int handle_g2v_notification(struct intel_vgpu *vgpu, int notification) · 1149 unmodified lines
1150	1150	switch (notification) {
1151	1151	case VGT_G2V_PPGTT_L3_PAGE_TABLE_CREATE:
1152	1152	root_entry_type = GTT_TYPE_PPGTT_ROOT_L3_ENTRY;
	1153	+ /* fall through */
1153	1154	case VGT_G2V_PPGTT_L4_PAGE_TABLE_CREATE:
1154	1155	mm = intel_vgpu_get_ppgtt_mm(vgpu, root_entry_type, pdps);
1155	1156	return PTR_ERR_OR_ZERO(mm);

drivers/gpu/drm/i915/gvt/kvmgt.c+1−1

@@ -1301,7 +1301,7 @@ static long intel_vgpu_ioctl(struct mdev_device *mdev, unsigned int cmd, · 1300 unmodified lines
1301	1301
1302	1302	}
1303	1303
1304		− return 0;
	1304	+ return -ENOTTY;
1305	1305	}
1306	1306
1307	1307	static ssize_t

drivers/gpu/drm/i915/i915_drv.c+15−12

@@ -1105,30 +1105,32 @@ static int i915_driver_init_hw(struct drm_i915_private *dev_priv) · 1104 unmodified lines
1105	1105
1106	1106	ret = i915_ggtt_probe_hw(dev_priv);
1107	1107	if (ret)
1108		− return ret;
	1108	+ goto err_perf;
1109	1109
1110		− /* WARNING: Apparently we must kick fbdev drivers before vgacon,
1111		− * otherwise the vga fbdev driver falls over. */
	1110	+ /*
	1111	+ * WARNING: Apparently we must kick fbdev drivers before vgacon,
	1112	+ * otherwise the vga fbdev driver falls over.
	1113	+ */
1112	1114	ret = i915_kick_out_firmware_fb(dev_priv);
1113	1115	if (ret) {
1114	1116	DRM_ERROR("failed to remove conflicting framebuffer drivers\n");
1115		− goto out_ggtt;
	1117	+ goto err_ggtt;
1116	1118	}
1117	1119
1118	1120	ret = i915_kick_out_vgacon(dev_priv);
1119	1121	if (ret) {
1120	1122	DRM_ERROR("failed to remove conflicting VGA console\n");
1121		− goto out_ggtt;
	1123	+ goto err_ggtt;
1122	1124	}
1123	1125
1124	1126	ret = i915_ggtt_init_hw(dev_priv);
1125	1127	if (ret)
1126		− return ret;
	1128	+ goto err_ggtt;
1127	1129
1128	1130	ret = i915_ggtt_enable_hw(dev_priv);
1129	1131	if (ret) {
1130	1132	DRM_ERROR("failed to enable GGTT\n");
1131		− goto out_ggtt;
	1133	+ goto err_ggtt;
1132	1134	}
1133	1135
1134	1136	pci_set_master(pdev);
@@ -1139,7 +1141,7 @@ static int i915_driver_init_hw(struct drm_i915_private *dev_priv) · 4 unmodified lines
1139	1141	if (ret) {
1140	1142	DRM_ERROR("failed to set DMA mask\n");
1141	1143
1142		− goto out_ggtt;
	1144	+ goto err_ggtt;
1143	1145	}
1144	1146	}
1145	1147
@@ -1157,7 +1159,7 @@ static int i915_driver_init_hw(struct drm_i915_private *dev_priv) · 11 unmodified lines
1157	1159	if (ret) {
1158	1160	DRM_ERROR("failed to set DMA mask\n");
1159	1161
1160		− goto out_ggtt;
	1162	+ goto err_ggtt;
1161	1163	}
1162	1164	}
1163	1165
@@ -1190,13 +1192,14 @@ static int i915_driver_init_hw(struct drm_i915_private *dev_priv) · 26 unmodified lines
1190	1192
1191	1193	ret = intel_gvt_init(dev_priv);
1192	1194	if (ret)
1193		− goto out_ggtt;
	1195	+ goto err_ggtt;
1194	1196
1195	1197	return 0;
1196	1198
1197		−out_ggtt:
	1199	+err_ggtt:
1198	1200	i915_ggtt_cleanup_hw(dev_priv);
1199		−
	1201	+err_perf:
	1202	+ i915_perf_fini(dev_priv);
1200	1203	return ret;
1201	1204	}
1202	1205

drivers/gpu/drm/i915/i915_gem_execbuffer.c+1−1

@@ -728,7 +728,7 @@ static int eb_lookup_vmas(struct i915_execbuffer *eb) · 727 unmodified lines
728	728
729	729	err = radix_tree_insert(handles_vma, handle, vma);
730	730	if (unlikely(err)) {
731		− kfree(lut);
	731	+ kmem_cache_free(eb->i915->luts, lut);
732	732	goto err_obj;
733	733	}
734	734

drivers/gpu/drm/i915/i915_pmu.c+27−10

@@ -473,20 +473,37 @@ static u64 get_rc6(struct drm_i915_private *i915) · 472 unmodified lines
473	473	spin_lock_irqsave(&i915->pmu.lock, flags);
474	474	spin_lock(&kdev->power.lock);
475	475
476		− if (!i915->pmu.sample[__I915_SAMPLE_RC6_ESTIMATED].cur)
477		− i915->pmu.suspended_jiffies_last =
478		− kdev->power.suspended_jiffies;
	476	+ /*
	477	+ * After the above branch intel_runtime_pm_get_if_in_use failed
	478	+ * to get the runtime PM reference we cannot assume we are in
	479	+ * runtime suspend since we can either: a) race with coming out
	480	+ * of it before we took the power.lock, or b) there are other
	481	+ * states than suspended which can bring us here.
	482	+ *
	483	+ * We need to double-check that we are indeed currently runtime
	484	+ * suspended and if not we cannot do better than report the last
	485	+ * known RC6 value.
	486	+ */
	487	+ if (kdev->power.runtime_status == RPM_SUSPENDED) {
	488	+ if (!i915->pmu.sample[__I915_SAMPLE_RC6_ESTIMATED].cur)
	489	+ i915->pmu.suspended_jiffies_last =
	490	+ kdev->power.suspended_jiffies;
479	491
480		− val = kdev->power.suspended_jiffies -
481		− i915->pmu.suspended_jiffies_last;
482		− val += jiffies - kdev->power.accounting_timestamp;
	492	+ val = kdev->power.suspended_jiffies -
	493	+ i915->pmu.suspended_jiffies_last;
	494	+ val += jiffies - kdev->power.accounting_timestamp;
483	495
484		− spin_unlock(&kdev->power.lock);
	496	+ val = jiffies_to_nsecs(val);
	497	+ val += i915->pmu.sample[__I915_SAMPLE_RC6].cur;
485	498
486		− val = jiffies_to_nsecs(val);
487		− val += i915->pmu.sample[__I915_SAMPLE_RC6].cur;
488		− i915->pmu.sample[__I915_SAMPLE_RC6_ESTIMATED].cur = val;
	499	+ i915->pmu.sample[__I915_SAMPLE_RC6_ESTIMATED].cur = val;
	500	+ } else if (i915->pmu.sample[__I915_SAMPLE_RC6_ESTIMATED].cur) {
	501	+ val = i915->pmu.sample[__I915_SAMPLE_RC6_ESTIMATED].cur;
	502	+ } else {
	503	+ val = i915->pmu.sample[__I915_SAMPLE_RC6].cur;
	504	+ }
489	505
	506	+ spin_unlock(&kdev->power.lock);
490	507	spin_unlock_irqrestore(&i915->pmu.lock, flags);
491	508	}
492	509

drivers/gpu/drm/i915/intel_audio.c+1−1

@@ -729,7 +729,7 @@ static void i915_audio_component_codec_wake_override(struct device *kdev, · 728 unmodified lines
729	729	struct drm_i915_private *dev_priv = kdev_to_i915(kdev);
730	730	u32 tmp;
731	731
732		− if (!IS_GEN9_BC(dev_priv))
	732	+ if (!IS_GEN9(dev_priv))
733	733	return;
734	734
735	735	i915_audio_component_get_power(kdev);

drivers/gpu/drm/i915/intel_bios.c+9−4

@@ -1256,7 +1256,6 @@ static void parse_ddi_port(struct drm_i915_private *dev_priv, enum port port, · 1255 unmodified lines
1256	1256	return;
1257	1257
1258	1258	aux_channel = child->aux_channel;
1259		− ddc_pin = child->ddc_pin;
1260	1259
1261	1260	is_dvi = child->device_type & DEVICE_TYPE_TMDS_DVI_SIGNALING;
1262	1261	is_dp = child->device_type & DEVICE_TYPE_DISPLAYPORT_OUTPUT;
@@ -1303,9 +1302,15 @@ static void parse_ddi_port(struct drm_i915_private *dev_priv, enum port port, · 40 unmodified lines
1303	1302	DRM_DEBUG_KMS("Port %c is internal DP\n", port_name(port));
1304	1303
1305	1304	if (is_dvi) {
1306		− info->alternate_ddc_pin = map_ddc_pin(dev_priv, ddc_pin);
1307		−
1308		− sanitize_ddc_pin(dev_priv, port);
	1305	+ ddc_pin = map_ddc_pin(dev_priv, child->ddc_pin);
	1306	+ if (intel_gmbus_is_valid_pin(dev_priv, ddc_pin)) {
	1307	+ info->alternate_ddc_pin = ddc_pin;
	1308	+ sanitize_ddc_pin(dev_priv, port);
	1309	+ } else {
	1310	+ DRM_DEBUG_KMS("Port %c has invalid DDC pin %d, "
	1311	+ "sticking to defaults\n",
	1312	+ port_name(port), ddc_pin);
	1313	+ }
1309	1314	}
1310	1315
1311	1316	if (is_dp) {

drivers/gpu/drm/i915/intel_cdclk.c+14−2

@@ -2140,10 +2140,22 @@ int intel_crtc_compute_min_cdclk(const struct intel_crtc_state *crtc_state) · 2139 unmodified lines
2140	2140	}
2141	2141	}
2142	2142
2143		− /* According to BSpec, "The CD clock frequency must be at least twice
	2143	+ /*
	2144	+ * According to BSpec, "The CD clock frequency must be at least twice
2144	2145	* the frequency of the Azalia BCLK." and BCLK is 96 MHz by default.
	2146	+ *
	2147	+ * FIXME: Check the actual, not default, BCLK being used.
	2148	+ *
	2149	+ * FIXME: This does not depend on ->has_audio because the higher CDCLK
	2150	+ * is required for audio probe, also when there are no audio capable
	2151	+ * displays connected at probe time. This leads to unnecessarily high
	2152	+ * CDCLK when audio is not required.
	2153	+ *
	2154	+ * FIXME: This limit is only applied when there are displays connected
	2155	+ * at probe time. If we probe without displays, we'll still end up using
	2156	+ * the platform minimum CDCLK, failing audio probe.
2145	2157	*/
2146		− if (crtc_state->has_audio && INTEL_GEN(dev_priv) >= 9)
	2158	+ if (INTEL_GEN(dev_priv) >= 9)
2147	2159	min_cdclk = max(2 * 96000, min_cdclk);
2148	2160
2149	2161	/*

drivers/gpu/drm/i915/intel_csr.c+1−0

@@ -35,6 +35,7 @@ · 34 unmodified lines
35	35	*/
36	36
37	37	#define I915_CSR_GLK "i915/glk_dmc_ver1_04.bin"
	38	+MODULE_FIRMWARE(I915_CSR_GLK);
38	39	#define GLK_CSR_VERSION_REQUIRED CSR_VERSION(1, 4)
39	40
40	41	#define I915_CSR_CNL "i915/cnl_dmc_ver1_07.bin"

drivers/gpu/drm/i915/intel_drv.h+2−2

@@ -49,12 +49,12 @@ · 48 unmodified lines
49	49	* check the condition before the timeout.
50	50	*/
51	51	#define __wait_for(OP, COND, US, Wmin, Wmax) ({ \
52		− unsigned long timeout__ = jiffies + usecs_to_jiffies(US) + 1; \
	52	+ const ktime_t end__ = ktime_add_ns(ktime_get_raw(), 1000ll * (US)); \
53	53	long wait__ = (Wmin); /* recommended min for usleep is 10 us */ \
54	54	int ret__; \
55	55	might_sleep(); \
56	56	for (;;) { \
57		− bool expired__ = time_after(jiffies, timeout__); \
	57	+ const bool expired__ = ktime_after(ktime_get_raw(), end__); \
58	58	OP; \
59	59	if (COND) { \
60	60	ret__ = 0; \

drivers/gpu/drm/i915/intel_fbdev.c+1−1

@@ -806,7 +806,7 @@ void intel_fbdev_output_poll_changed(struct drm_device *dev) · 805 unmodified lines
806	806	return;
807	807
808	808	intel_fbdev_sync(ifbdev);
809		− if (ifbdev->vma)
	809	+ if (ifbdev->vma \|\| ifbdev->helper.deferred_setup)
810	810	drm_fb_helper_hotplug_event(&ifbdev->helper);
811	811	}
812	812

drivers/gpu/drm/i915/intel_lrc.c+9−0

@@ -577,6 +577,8 @@ static void execlists_dequeue(struct intel_engine_cs *engine) · 576 unmodified lines
577	577	* know the next preemption status we see corresponds
578	578	* to this ELSP update.
579	579	*/
	580	+ GEM_BUG_ON(!execlists_is_active(execlists,
	581	+ EXECLISTS_ACTIVE_USER));
580	582	GEM_BUG_ON(!port_count(&port[0]));
581	583	if (port_count(&port[0]) > 1)
582	584	goto unlock;
@@ -738,6 +740,8 @@ execlists_cancel_port_requests(struct intel_engine_execlists * const execlists) · 155 unmodified lines
738	740	memset(port, 0, sizeof(*port));
739	741	port++;
740	742	}
	743	+
	744	+ execlists_clear_active(execlists, EXECLISTS_ACTIVE_USER);
741	745	}
742	746
743	747	static void execlists_cancel_requests(struct intel_engine_cs *engine)
@@ -1001,6 +1005,11 @@ static void execlists_submission_tasklet(unsigned long data) · 257 unmodified lines
1001	1005
1002	1006	if (fw)
1003	1007	intel_uncore_forcewake_put(dev_priv, execlists->fw_domains);
	1008	+
	1009	+ /* If the engine is now idle, so should be the flag; and vice versa. */
	1010	+ GEM_BUG_ON(execlists_is_active(&engine->execlists,
	1011	+ EXECLISTS_ACTIVE_USER) ==
	1012	+ !port_isset(engine->execlists.port));
1004	1013	}
1005	1014
1006	1015	static void queue_request(struct intel_engine_cs *engine,

drivers/gpu/drm/i915/intel_runtime_pm.c+5−6

@@ -641,19 +641,18 @@ void skl_enable_dc6(struct drm_i915_private *dev_priv) · 640 unmodified lines
641	641
642	642	DRM_DEBUG_KMS("Enabling DC6\n");
643	643
644		− gen9_set_dc_state(dev_priv, DC_STATE_EN_UPTO_DC6);
	644	+ /* Wa Display #1183: skl,kbl,cfl */
	645	+ if (IS_GEN9_BC(dev_priv))
	646	+ I915_WRITE(GEN8_CHICKEN_DCPR_1, I915_READ(GEN8_CHICKEN_DCPR_1) \|
	647	+ SKL_SELECT_ALTERNATE_DC_EXIT);
645	648
	649	+ gen9_set_dc_state(dev_priv, DC_STATE_EN_UPTO_DC6);
646	650	}
647	651
648	652	void skl_disable_dc6(struct drm_i915_private *dev_priv)
649	653	{
650	654	DRM_DEBUG_KMS("Disabling DC6\n");
651	655
652		− /* Wa Display #1183: skl,kbl,cfl */
653		− if (IS_GEN9_BC(dev_priv))
654		− I915_WRITE(GEN8_CHICKEN_DCPR_1, I915_READ(GEN8_CHICKEN_DCPR_1) \|
655		− SKL_SELECT_ALTERNATE_DC_EXIT);
656		−
657	656	gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
658	657	}
659	658

drivers/gpu/drm/msm/disp/mdp4/mdp4_crtc.c+1−0

@@ -351,6 +351,7 @@ static void mdp4_crtc_atomic_flush(struct drm_crtc *crtc, · 350 unmodified lines
351	351
352	352	spin_lock_irqsave(&dev->event_lock, flags);
353	353	mdp4_crtc->event = crtc->state->event;
	354	+ crtc->state->event = NULL;
354	355	spin_unlock_irqrestore(&dev->event_lock, flags);
355	356
356	357	blend_setup(crtc);

drivers/gpu/drm/msm/disp/mdp5/mdp5_crtc.c+1−0

@@ -708,6 +708,7 @@ static void mdp5_crtc_atomic_flush(struct drm_crtc *crtc, · 707 unmodified lines
708	708
709	709	spin_lock_irqsave(&dev->event_lock, flags);
710	710	mdp5_crtc->event = crtc->state->event;
	711	+ crtc->state->event = NULL;
711	712	spin_unlock_irqrestore(&dev->event_lock, flags);
712	713
713	714	/*

drivers/gpu/drm/msm/disp/mdp_format.c+2−1

@@ -171,7 +171,8 @@ uint32_t mdp_get_formats(uint32_t *pixel_formats, uint32_t max_formats, · 170 unmodified lines
171	171	return i;
172	172	}
173	173
174		−const struct msm_format mdp_get_format(struct msm_kms kms, uint32_t format)
	174	+const struct msm_format mdp_get_format(struct msm_kms kms, uint32_t format,
	175	+ uint64_t modifier)
175	176	{
176	177	int i;
177	178	for (i = 0; i < ARRAY_SIZE(formats); i++) {

drivers/gpu/drm/msm/disp/mdp_kms.h+1−1

@@ -98,7 +98,7 @@ struct mdp_format { · 97 unmodified lines
98	98	#define MDP_FORMAT_IS_YUV(mdp_format) ((mdp_format)->is_yuv)
99	99
100	100	uint32_t mdp_get_formats(uint32_t *formats, uint32_t max_formats, bool rgb_only);
101		−const struct msm_format mdp_get_format(struct msm_kms kms, uint32_t format);
	101	+const struct msm_format mdp_get_format(struct msm_kms kms, uint32_t format, uint64_t modifier);
102	102
103	103	/* MDP capabilities */
104	104	#define MDP_CAP_SMP BIT(0) /* Shared Memory Pool */

drivers/gpu/drm/msm/dsi/dsi_host.c+12−4

@@ -173,6 +173,7 @@ struct msm_dsi_host { · 172 unmodified lines
173	173
174	174	bool registered;
175	175	bool power_on;
	176	+ bool enabled;
176	177	int irq;
177	178	};
178	179
@@ -775,7 +776,7 @@ static inline enum dsi_cmd_dst_format dsi_get_cmd_fmt( · 596 unmodified lines
775	776	switch (mipi_fmt) {
776	777	case MIPI_DSI_FMT_RGB888: return CMD_DST_FORMAT_RGB888;
777	778	case MIPI_DSI_FMT_RGB666_PACKED:
778		− case MIPI_DSI_FMT_RGB666: return VID_DST_FORMAT_RGB666;
	779	+ case MIPI_DSI_FMT_RGB666: return CMD_DST_FORMAT_RGB666;
779	780	case MIPI_DSI_FMT_RGB565: return CMD_DST_FORMAT_RGB565;
780	781	default: return CMD_DST_FORMAT_RGB888;
781	782	}
@@ -986,13 +987,19 @@ static void dsi_set_tx_power_mode(int mode, struct msm_dsi_host *msm_host) · 204 unmodified lines
986	987
987	988	static void dsi_wait4video_done(struct msm_dsi_host *msm_host)
988	989	{
	990	+ u32 ret = 0;
	991	+ struct device *dev = &msm_host->pdev->dev;
	992	+
989	993	dsi_intr_ctrl(msm_host, DSI_IRQ_MASK_VIDEO_DONE, 1);
990	994
991	995	reinit_completion(&msm_host->video_comp);
992	996
993		− wait_for_completion_timeout(&msm_host->video_comp,
	997	+ ret = wait_for_completion_timeout(&msm_host->video_comp,
994	998	msecs_to_jiffies(70));
995	999
	1000	+ if (ret <= 0)
	1001	+ dev_err(dev, "wait for video done timed out\n");
	1002	+
996	1003	dsi_intr_ctrl(msm_host, DSI_IRQ_MASK_VIDEO_DONE, 0);
997	1004	}
998	1005
@@ -1001,7 +1008,7 @@ static void dsi_wait4video_eng_busy(struct msm_dsi_host *msm_host) · 2 unmodified lines
1001	1008	if (!(msm_host->mode_flags & MIPI_DSI_MODE_VIDEO))
1002	1009	return;
1003	1010
1004		− if (msm_host->power_on) {
	1011	+ if (msm_host->power_on && msm_host->enabled) {
1005	1012	dsi_wait4video_done(msm_host);
1006	1013	/* delay 4 ms to skip BLLP */
1007	1014	usleep_range(2000, 4000);
@@ -2203,7 +2210,7 @@ int msm_dsi_host_enable(struct mipi_dsi_host *host) · 1195 unmodified lines
2203	2210	* pm_runtime_put_autosuspend(&msm_host->pdev->dev);
2204	2211	* }
2205	2212	*/
2206		−
	2213	+ msm_host->enabled = true;
2207	2214	return 0;
2208	2215	}
2209	2216
@@ -2211,6 +2218,7 @@ int msm_dsi_host_disable(struct mipi_dsi_host *host) · 1 unmodified line
2211	2218	{
2212	2219	struct msm_dsi_host *msm_host = to_msm_dsi_host(host);
2213	2220
	2221	+ msm_host->enabled = false;
2214	2222	dsi_op_mode_config(msm_host,
2215	2223	!!(msm_host->mode_flags & MIPI_DSI_MODE_VIDEO), false);
2216	2224

drivers/gpu/drm/msm/dsi/phy/dsi_phy.c+109−0

@@ -265,6 +265,115 @@ int msm_dsi_dphy_timing_calc_v2(struct msm_dsi_dphy_timing *timing, · 264 unmodified lines
265	265	return 0;
266	266	}
267	267
	268	+int msm_dsi_dphy_timing_calc_v3(struct msm_dsi_dphy_timing *timing,
	269	+ struct msm_dsi_phy_clk_request *clk_req)
	270	+{
	271	+ const unsigned long bit_rate = clk_req->bitclk_rate;
	272	+ const unsigned long esc_rate = clk_req->escclk_rate;
	273	+ s32 ui, ui_x8, lpx;
	274	+ s32 tmax, tmin;
	275	+ s32 pcnt0 = 50;
	276	+ s32 pcnt1 = 50;
	277	+ s32 pcnt2 = 10;
	278	+ s32 pcnt3 = 30;
	279	+ s32 pcnt4 = 10;
	280	+ s32 pcnt5 = 2;
	281	+ s32 coeff = 1000; /* Precision, should avoid overflow */
	282	+ s32 hb_en, hb_en_ckln;
	283	+ s32 temp;
	284	+
	285	+ if (!bit_rate \|\| !esc_rate)
	286	+ return -EINVAL;
	287	+
	288	+ timing->hs_halfbyte_en = 0;
	289	+ hb_en = 0;
	290	+ timing->hs_halfbyte_en_ckln = 0;
	291	+ hb_en_ckln = 0;
	292	+
	293	+ ui = mult_frac(NSEC_PER_MSEC, coeff, bit_rate / 1000);
	294	+ ui_x8 = ui << 3;
	295	+ lpx = mult_frac(NSEC_PER_MSEC, coeff, esc_rate / 1000);
	296	+
	297	+ temp = S_DIV_ROUND_UP(38 * coeff, ui_x8);
	298	+ tmin = max_t(s32, temp, 0);
	299	+ temp = (95 * coeff) / ui_x8;
	300	+ tmax = max_t(s32, temp, 0);
	301	+ timing->clk_prepare = linear_inter(tmax, tmin, pcnt0, 0, false);
	302	+
	303	+ temp = 300 * coeff - (timing->clk_prepare << 3) * ui;
	304	+ tmin = S_DIV_ROUND_UP(temp, ui_x8) - 1;
	305	+ tmax = (tmin > 255) ? 511 : 255;
	306	+ timing->clk_zero = linear_inter(tmax, tmin, pcnt5, 0, false);
	307	+
	308	+ tmin = DIV_ROUND_UP(60 * coeff + 3 * ui, ui_x8);
	309	+ temp = 105 * coeff + 12 * ui - 20 * coeff;
	310	+ tmax = (temp + 3 * ui) / ui_x8;
	311	+ timing->clk_trail = linear_inter(tmax, tmin, pcnt3, 0, false);
	312	+
	313	+ temp = S_DIV_ROUND_UP(40 * coeff + 4 * ui, ui_x8);
	314	+ tmin = max_t(s32, temp, 0);
	315	+ temp = (85 * coeff + 6 * ui) / ui_x8;
	316	+ tmax = max_t(s32, temp, 0);
	317	+ timing->hs_prepare = linear_inter(tmax, tmin, pcnt1, 0, false);
	318	+
	319	+ temp = 145 * coeff + 10 * ui - (timing->hs_prepare << 3) * ui;
	320	+ tmin = S_DIV_ROUND_UP(temp, ui_x8) - 1;
	321	+ tmax = 255;
	322	+ timing->hs_zero = linear_inter(tmax, tmin, pcnt4, 0, false);
	323	+
	324	+ tmin = DIV_ROUND_UP(60 * coeff + 4 * ui, ui_x8) - 1;
	325	+ temp = 105 * coeff + 12 * ui - 20 * coeff;
	326	+ tmax = (temp / ui_x8) - 1;
	327	+ timing->hs_trail = linear_inter(tmax, tmin, pcnt3, 0, false);
	328	+
	329	+ temp = 50 * coeff + ((hb_en << 2) - 8) * ui;
	330	+ timing->hs_rqst = S_DIV_ROUND_UP(temp, ui_x8);
	331	+
	332	+ tmin = DIV_ROUND_UP(100 * coeff, ui_x8) - 1;
	333	+ tmax = 255;
	334	+ timing->hs_exit = linear_inter(tmax, tmin, pcnt2, 0, false);
	335	+
	336	+ temp = 50 * coeff + ((hb_en_ckln << 2) - 8) * ui;
	337	+ timing->hs_rqst_ckln = S_DIV_ROUND_UP(temp, ui_x8);
	338	+
	339	+ temp = 60 * coeff + 52 * ui - 43 * ui;
	340	+ tmin = DIV_ROUND_UP(temp, ui_x8) - 1;
	341	+ tmax = 63;
	342	+ timing->shared_timings.clk_post =
	343	+ linear_inter(tmax, tmin, pcnt2, 0, false);
	344	+
	345	+ temp = 8 * ui + (timing->clk_prepare << 3) * ui;
	346	+ temp += (((timing->clk_zero + 3) << 3) + 11) * ui;
	347	+ temp += hb_en_ckln ? (((timing->hs_rqst_ckln << 3) + 4) * ui) :
	348	+ (((timing->hs_rqst_ckln << 3) + 8) * ui);
	349	+ tmin = S_DIV_ROUND_UP(temp, ui_x8) - 1;
	350	+ tmax = 63;
	351	+ if (tmin > tmax) {
	352	+ temp = linear_inter(tmax << 1, tmin, pcnt2, 0, false);
	353	+ timing->shared_timings.clk_pre = temp >> 1;
	354	+ timing->shared_timings.clk_pre_inc_by_2 = 1;
	355	+ } else {
	356	+ timing->shared_timings.clk_pre =
	357	+ linear_inter(tmax, tmin, pcnt2, 0, false);
	358	+ timing->shared_timings.clk_pre_inc_by_2 = 0;
	359	+ }
	360	+
	361	+ timing->ta_go = 3;
	362	+ timing->ta_sure = 0;
	363	+ timing->ta_get = 4;
	364	+
	365	+ DBG("%d, %d, %d, %d, %d, %d, %d, %d, %d, %d, %d, %d, %d, %d, %d, %d",
	366	+ timing->shared_timings.clk_pre, timing->shared_timings.clk_post,
	367	+ timing->shared_timings.clk_pre_inc_by_2, timing->clk_zero,
	368	+ timing->clk_trail, timing->clk_prepare, timing->hs_exit,
	369	+ timing->hs_zero, timing->hs_prepare, timing->hs_trail,
	370	+ timing->hs_rqst, timing->hs_rqst_ckln, timing->hs_halfbyte_en,
	371	+ timing->hs_halfbyte_en_ckln, timing->hs_prep_dly,
	372	+ timing->hs_prep_dly_ckln);
	373	+
	374	+ return 0;
	375	+}
	376	+
268	377	void msm_dsi_phy_set_src_pll(struct msm_dsi_phy *phy, int pll_id, u32 reg,
269	378	u32 bit_mask)
270	379	{

drivers/gpu/drm/msm/dsi/phy/dsi_phy.h+2−0

@@ -101,6 +101,8 @@ int msm_dsi_dphy_timing_calc(struct msm_dsi_dphy_timing *timing, · 100 unmodified lines
101	101	struct msm_dsi_phy_clk_request *clk_req);
102	102	int msm_dsi_dphy_timing_calc_v2(struct msm_dsi_dphy_timing *timing,
103	103	struct msm_dsi_phy_clk_request *clk_req);
	104	+int msm_dsi_dphy_timing_calc_v3(struct msm_dsi_dphy_timing *timing,
	105	+ struct msm_dsi_phy_clk_request *clk_req);
104	106	void msm_dsi_phy_set_src_pll(struct msm_dsi_phy *phy, int pll_id, u32 reg,
105	107	u32 bit_mask);
106	108	int msm_dsi_phy_init_common(struct msm_dsi_phy *phy);

drivers/gpu/drm/msm/dsi/phy/dsi_phy_10nm.c+0−28

@@ -79,34 +79,6 @@ static void dsi_phy_hw_v3_0_lane_settings(struct msm_dsi_phy *phy) · 78 unmodified lines
79	79	dsi_phy_write(lane_base + REG_DSI_10nm_PHY_LN_TX_DCTRL(3), 0x04);
80	80	}
81	81
82		−static int msm_dsi_dphy_timing_calc_v3(struct msm_dsi_dphy_timing *timing,
83		− struct msm_dsi_phy_clk_request *clk_req)
84		−{
85		− /*
86		− * TODO: These params need to be computed, they're currently hardcoded
87		− * for a 1440x2560@60Hz panel with a byteclk of 100.618 Mhz, and a
88		− * default escape clock of 19.2 Mhz.
89		− */
90		−
91		− timing->hs_halfbyte_en = 0;
92		− timing->clk_zero = 0x1c;
93		− timing->clk_prepare = 0x07;
94		− timing->clk_trail = 0x07;
95		− timing->hs_exit = 0x23;
96		− timing->hs_zero = 0x21;
97		− timing->hs_prepare = 0x07;
98		− timing->hs_trail = 0x07;
99		− timing->hs_rqst = 0x05;
100		− timing->ta_sure = 0x00;
101		− timing->ta_go = 0x03;
102		− timing->ta_get = 0x04;
103		−
104		− timing->shared_timings.clk_pre = 0x2d;
105		− timing->shared_timings.clk_post = 0x0d;
106		−
107		− return 0;
108		−}
109		−
110	82	static int dsi_10nm_phy_enable(struct msm_dsi_phy *phy, int src_pll_id,
111	83	struct msm_dsi_phy_clk_request *clk_req)
112	84	{

drivers/gpu/drm/msm/msm_fb.c+2−1

@@ -183,7 +183,8 @@ static struct drm_framebuffer msm_framebuffer_init(struct drm_device dev, · 182 unmodified lines
183	183	hsub = drm_format_horz_chroma_subsampling(mode_cmd->pixel_format);
184	184	vsub = drm_format_vert_chroma_subsampling(mode_cmd->pixel_format);
185	185
186		− format = kms->funcs->get_format(kms, mode_cmd->pixel_format);
	186	+ format = kms->funcs->get_format(kms, mode_cmd->pixel_format,
	187	+ mode_cmd->modifier[0]);
187	188	if (!format) {
188	189	dev_err(dev->dev, "unsupported pixel format: %4.4s\n",
189	190	(char *)&mode_cmd->pixel_format);

drivers/gpu/drm/msm/msm_fbdev.c+2−9

@@ -92,8 +92,7 @@ static int msm_fbdev_create(struct drm_fb_helper *helper, · 91 unmodified lines
92	92
93	93	if (IS_ERR(fb)) {
94	94	dev_err(dev->dev, "failed to allocate fb\n");
95		− ret = PTR_ERR(fb);
96		− goto fail;
	95	+ return PTR_ERR(fb);
97	96	}
98	97
99	98	bo = msm_framebuffer_bo(fb, 0);
@@ -151,13 +150,7 @@ static int msm_fbdev_create(struct drm_fb_helper *helper, · 51 unmodified lines
151	150
152	151	fail_unlock:
153	152	mutex_unlock(&dev->struct_mutex);
154		−fail:
155		−
156		− if (ret) {
157		− if (fb)
158		− drm_framebuffer_remove(fb);
159		− }
160		−
	153	+ drm_framebuffer_remove(fb);
161	154	return ret;
162	155	}
163	156

drivers/gpu/drm/msm/msm_gem.c+11−9

@@ -132,17 +132,19 @@ static void put_pages(struct drm_gem_object *obj) · 131 unmodified lines
132	132	struct msm_gem_object *msm_obj = to_msm_bo(obj);
133	133
134	134	if (msm_obj->pages) {
135		− /* For non-cached buffers, ensure the new pages are clean
136		− * because display controller, GPU, etc. are not coherent:
137		− */
138		− if (msm_obj->flags & (MSM_BO_WC\|MSM_BO_UNCACHED))
139		− dma_unmap_sg(obj->dev->dev, msm_obj->sgt->sgl,
140		− msm_obj->sgt->nents, DMA_BIDIRECTIONAL);
	135	+ if (msm_obj->sgt) {
	136	+ /* For non-cached buffers, ensure the new
	137	+ * pages are clean because display controller,
	138	+ * GPU, etc. are not coherent:
	139	+ */
	140	+ if (msm_obj->flags & (MSM_BO_WC\|MSM_BO_UNCACHED))
	141	+ dma_unmap_sg(obj->dev->dev, msm_obj->sgt->sgl,
	142	+ msm_obj->sgt->nents,
	143	+ DMA_BIDIRECTIONAL);
141	144
142		− if (msm_obj->sgt)
143	145	sg_free_table(msm_obj->sgt);
144		−
145		− kfree(msm_obj->sgt);
	146	+ kfree(msm_obj->sgt);
	147	+ }
146	148
147	149	if (use_pages(obj))
148	150	drm_gem_put_pages(obj, msm_obj->pages, true, false);

drivers/gpu/drm/msm/msm_kms.h+4−1

@@ -48,8 +48,11 @@ struct msm_kms_funcs { · 47 unmodified lines
48	48	/* functions to wait for atomic commit completed on each CRTC */
49	49	void (wait_for_crtc_commit_done)(struct msm_kms kms,
50	50	struct drm_crtc *crtc);
	51	+ /* get msm_format w/ optional format modifiers from drm_mode_fb_cmd2 */
	52	+ const struct msm_format (get_format)(struct msm_kms *kms,
	53	+ const uint32_t format,
	54	+ const uint64_t modifiers);
51	55	/* misc: */
52		− const struct msm_format (get_format)(struct msm_kms *kms, uint32_t format);
53	56	long (round_pixclk)(struct msm_kms kms, unsigned long rate,
54	57	struct drm_encoder *encoder);
55	58	int (set_split_display)(struct msm_kms kms,

drivers/gpu/drm/qxl/qxl_cmd.c+2−4

@@ -179,10 +179,9 @@ qxl_push_command_ring_release(struct qxl_device qdev, struct qxl_release relea · 178 unmodified lines
179	179	uint32_t type, bool interruptible)
180	180	{
181	181	struct qxl_command cmd;
182		− struct qxl_bo_list *entry = list_first_entry(&release->bos, struct qxl_bo_list, tv.head);
183	182
184	183	cmd.type = type;
185		− cmd.data = qxl_bo_physical_address(qdev, to_qxl_bo(entry->tv.bo), release->release_offset);
	184	+ cmd.data = qxl_bo_physical_address(qdev, release->release_bo, release->release_offset);
186	185
187	186	return qxl_ring_push(qdev->command_ring, &cmd, interruptible);
188	187	}
@@ -192,10 +191,9 @@ qxl_push_cursor_ring_release(struct qxl_device qdev, struct qxl_release releas · 3 unmodified lines
192	191	uint32_t type, bool interruptible)
193	192	{
194	193	struct qxl_command cmd;
195		− struct qxl_bo_list *entry = list_first_entry(&release->bos, struct qxl_bo_list, tv.head);
196	194
197	195	cmd.type = type;
198		− cmd.data = qxl_bo_physical_address(qdev, to_qxl_bo(entry->tv.bo), release->release_offset);
	196	+ cmd.data = qxl_bo_physical_address(qdev, release->release_bo, release->release_offset);
199	197
200	198	return qxl_ring_push(qdev->cursor_ring, &cmd, interruptible);
201	199	}

drivers/gpu/drm/qxl/qxl_drv.h+1−0

@@ -167,6 +167,7 @@ struct qxl_release { · 166 unmodified lines
167	167
168	168	int id;
169	169	int type;
	170	+ struct qxl_bo *release_bo;
170	171	uint32_t release_offset;
171	172	uint32_t surface_release_id;
172	173	struct ww_acquire_ctx ticket;

drivers/gpu/drm/qxl/qxl_ioctl.c+2−2

@@ -182,9 +182,9 @@ static int qxl_process_single_command(struct qxl_device *qdev, · 181 unmodified lines
182	182	goto out_free_reloc;
183	183
184	184	/* TODO copy slow path code from i915 */
185		− fb_cmd = qxl_bo_kmap_atomic_page(qdev, cmd_bo, (release->release_offset & PAGE_SIZE));
	185	+ fb_cmd = qxl_bo_kmap_atomic_page(qdev, cmd_bo, (release->release_offset & PAGE_MASK));
186	186	unwritten = __copy_from_user_inatomic_nocache
187		− (fb_cmd + sizeof(union qxl_release_info) + (release->release_offset & ~PAGE_SIZE),
	187	+ (fb_cmd + sizeof(union qxl_release_info) + (release->release_offset & ~PAGE_MASK),
188	188	u64_to_user_ptr(cmd->command), cmd->command_size);
189	189
190	190	{

drivers/gpu/drm/qxl/qxl_release.c+9−9

@@ -173,6 +173,7 @@ qxl_release_free_list(struct qxl_release *release) · 172 unmodified lines
173	173	list_del(&entry->tv.head);
174	174	kfree(entry);
175	175	}
	176	+ release->release_bo = NULL;
176	177	}
177	178
178	179	void
@@ -296,7 +297,6 @@ int qxl_alloc_surface_release_reserved(struct qxl_device *qdev, · 117 unmodified lines
296	297	{
297	298	if (surface_cmd_type == QXL_SURFACE_CMD_DESTROY && create_rel) {
298	299	int idr_ret;
299		− struct qxl_bo_list *entry = list_first_entry(&create_rel->bos, struct qxl_bo_list, tv.head);
300	300	struct qxl_bo *bo;
301	301	union qxl_release_info *info;
302	302
@@ -304,8 +304,9 @@ int qxl_alloc_surface_release_reserved(struct qxl_device *qdev, · 1 unmodified line
304	304	idr_ret = qxl_release_alloc(qdev, QXL_RELEASE_SURFACE_CMD, release);
305	305	if (idr_ret < 0)
306	306	return idr_ret;
307		− bo = to_qxl_bo(entry->tv.bo);
	307	+ bo = create_rel->release_bo;
308	308
	309	+ (*release)->release_bo = bo;
309	310	(*release)->release_offset = create_rel->release_offset + 64;
310	311
311	312	qxl_release_list_add(*release, bo);
@@ -365,6 +366,7 @@ int qxl_alloc_release_reserved(struct qxl_device *qdev, unsigned long size, · 53 unmodified lines
365	366
366	367	bo = qxl_bo_ref(qdev->current_release_bo[cur_idx]);
367	368
	369	+ (*release)->release_bo = bo;
368	370	(release)->release_offset = qdev->current_release_bo_offset[cur_idx] release_size_per_bo[cur_idx];
369	371	qdev->current_release_bo_offset[cur_idx]++;
370	372
@@ -408,13 +410,12 @@ union qxl_release_info qxl_release_map(struct qxl_device qdev, · 37 unmodified lines
408	410	{
409	411	void *ptr;
410	412	union qxl_release_info *info;
411		− struct qxl_bo_list *entry = list_first_entry(&release->bos, struct qxl_bo_list, tv.head);
412		− struct qxl_bo *bo = to_qxl_bo(entry->tv.bo);
	413	+ struct qxl_bo *bo = release->release_bo;
413	414
414		− ptr = qxl_bo_kmap_atomic_page(qdev, bo, release->release_offset & PAGE_SIZE);
	415	+ ptr = qxl_bo_kmap_atomic_page(qdev, bo, release->release_offset & PAGE_MASK);
415	416	if (!ptr)
416	417	return NULL;
417		− info = ptr + (release->release_offset & ~PAGE_SIZE);
	418	+ info = ptr + (release->release_offset & ~PAGE_MASK);
418	419	return info;
419	420	}
420	421
@@ -422,11 +423,10 @@ void qxl_release_unmap(struct qxl_device *qdev, · 1 unmodified line
422	423	struct qxl_release *release,
423	424	union qxl_release_info *info)
424	425	{
425		− struct qxl_bo_list *entry = list_first_entry(&release->bos, struct qxl_bo_list, tv.head);
426		− struct qxl_bo *bo = to_qxl_bo(entry->tv.bo);
	426	+ struct qxl_bo *bo = release->release_bo;
427	427	void *ptr;
428	428
429		− ptr = ((void *)info) - (release->release_offset & ~PAGE_SIZE);
	429	+ ptr = ((void *)info) - (release->release_offset & ~PAGE_MASK);
430	430	qxl_bo_kunmap_atomic_page(qdev, bo, ptr);
431	431	}
432	432

drivers/gpu/drm/sun4i/sun4i_lvds.c+0−55

@@ -94,64 +94,9 @@ static void sun4i_lvds_encoder_disable(struct drm_encoder *encoder) · 93 unmodified lines
94	94	}
95	95	}
96	96
97		−static enum drm_mode_status sun4i_lvds_encoder_mode_valid(struct drm_encoder *crtc,
98		− const struct drm_display_mode *mode)
99		−{
100		− struct sun4i_lvds *lvds = drm_encoder_to_sun4i_lvds(crtc);
101		− struct sun4i_tcon *tcon = lvds->tcon;
102		− u32 hsync = mode->hsync_end - mode->hsync_start;
103		− u32 vsync = mode->vsync_end - mode->vsync_start;
104		− unsigned long rate = mode->clock * 1000;
105		− long rounded_rate;
106		−
107		− DRM_DEBUG_DRIVER("Validating modes...\n");
108		−
109		− if (hsync < 1)
110		− return MODE_HSYNC_NARROW;
111		−
112		− if (hsync > 0x3ff)
113		− return MODE_HSYNC_WIDE;
114		−
115		− if ((mode->hdisplay < 1) \|\| (mode->htotal < 1))
116		− return MODE_H_ILLEGAL;
117		−
118		− if ((mode->hdisplay > 0x7ff) \|\| (mode->htotal > 0xfff))
119		− return MODE_BAD_HVALUE;
120		−
121		− DRM_DEBUG_DRIVER("Horizontal parameters OK\n");
122		−
123		− if (vsync < 1)
124		− return MODE_VSYNC_NARROW;
125		−
126		− if (vsync > 0x3ff)
127		− return MODE_VSYNC_WIDE;
128		−
129		− if ((mode->vdisplay < 1) \|\| (mode->vtotal < 1))
130		− return MODE_V_ILLEGAL;
131		−
132		− if ((mode->vdisplay > 0x7ff) \|\| (mode->vtotal > 0xfff))
133		− return MODE_BAD_VVALUE;
134		−
135		− DRM_DEBUG_DRIVER("Vertical parameters OK\n");
136		−
137		− tcon->dclk_min_div = 7;
138		− tcon->dclk_max_div = 7;
139		− rounded_rate = clk_round_rate(tcon->dclk, rate);
140		− if (rounded_rate < rate)
141		− return MODE_CLOCK_LOW;
142		−
143		− if (rounded_rate > rate)
144		− return MODE_CLOCK_HIGH;
145		−
146		− DRM_DEBUG_DRIVER("Clock rate OK\n");
147		−
148		− return MODE_OK;
149		−}
150		−
151	97	static const struct drm_encoder_helper_funcs sun4i_lvds_enc_helper_funcs = {
152	98	.disable = sun4i_lvds_encoder_disable,
153	99	.enable = sun4i_lvds_encoder_enable,
154		− .mode_valid = sun4i_lvds_encoder_mode_valid,
155	100	};
156	101
157	102	static const struct drm_encoder_funcs sun4i_lvds_enc_funcs = {

drivers/gpu/drm/vc4/vc4_bo.c+2−0

@@ -195,6 +195,7 @@ static void vc4_bo_destroy(struct vc4_bo *bo) · 194 unmodified lines
195	195	vc4_bo_set_label(obj, -1);
196	196
197	197	if (bo->validated_shader) {
	198	+ kfree(bo->validated_shader->uniform_addr_offsets);
198	199	kfree(bo->validated_shader->texture_samples);
199	200	kfree(bo->validated_shader);
200	201	bo->validated_shader = NULL;
@@ -591,6 +592,7 @@ void vc4_free_object(struct drm_gem_object *gem_bo) · 390 unmodified lines
591	592	}
592	593
593	594	if (bo->validated_shader) {
	595	+ kfree(bo->validated_shader->uniform_addr_offsets);
594	596	kfree(bo->validated_shader->texture_samples);
595	597	kfree(bo->validated_shader);
596	598	bo->validated_shader = NULL;

drivers/gpu/drm/vc4/vc4_crtc.c+45−1

@@ -760,6 +760,7 @@ static irqreturn_t vc4_crtc_irq_handler(int irq, void *data) · 759 unmodified lines
760	760	struct vc4_async_flip_state {
761	761	struct drm_crtc *crtc;
762	762	struct drm_framebuffer *fb;
	763	+ struct drm_framebuffer *old_fb;
763	764	struct drm_pending_vblank_event *event;
764	765
765	766	struct vc4_seqno_cb cb;
@@ -789,6 +790,23 @@ vc4_async_page_flip_complete(struct vc4_seqno_cb *cb) · 23 unmodified lines
789	790
790	791	drm_crtc_vblank_put(crtc);
791	792	drm_framebuffer_put(flip_state->fb);
	793	+
	794	+ /* Decrement the BO usecnt in order to keep the inc/dec calls balanced
	795	+ * when the planes are updated through the async update path.
	796	+ * FIXME: we should move to generic async-page-flip when it's
	797	+ * available, so that we can get rid of this hand-made cleanup_fb()
	798	+ * logic.
	799	+ */
	800	+ if (flip_state->old_fb) {
	801	+ struct drm_gem_cma_object *cma_bo;
	802	+ struct vc4_bo *bo;
	803	+
	804	+ cma_bo = drm_fb_cma_get_gem_obj(flip_state->old_fb, 0);
	805	+ bo = to_vc4_bo(&cma_bo->base);
	806	+ vc4_bo_dec_usecnt(bo);
	807	+ drm_framebuffer_put(flip_state->old_fb);
	808	+ }
	809	+
792	810	kfree(flip_state);
793	811
794	812	up(&vc4->async_modeset);
@@ -813,9 +831,22 @@ static int vc4_async_page_flip(struct drm_crtc *crtc, · 18 unmodified lines
813	831	struct drm_gem_cma_object *cma_bo = drm_fb_cma_get_gem_obj(fb, 0);
814	832	struct vc4_bo *bo = to_vc4_bo(&cma_bo->base);
815	833
	834	+ /* Increment the BO usecnt here, so that we never end up with an
	835	+ * unbalanced number of vc4_bo_{dec,inc}_usecnt() calls when the
	836	+ * plane is later updated through the non-async path.
	837	+ * FIXME: we should move to generic async-page-flip when it's
	838	+ * available, so that we can get rid of this hand-made prepare_fb()
	839	+ * logic.
	840	+ */
	841	+ ret = vc4_bo_inc_usecnt(bo);
	842	+ if (ret)
	843	+ return ret;
	844	+
816	845	flip_state = kzalloc(sizeof(*flip_state), GFP_KERNEL);
817		− if (!flip_state)
	846	+ if (!flip_state) {
	847	+ vc4_bo_dec_usecnt(bo);
818	848	return -ENOMEM;
	849	+ }
819	850
820	851	drm_framebuffer_get(fb);
821	852	flip_state->fb = fb;
@@ -826,10 +857,23 @@ static int vc4_async_page_flip(struct drm_crtc *crtc, · 4 unmodified lines
826	857	ret = down_interruptible(&vc4->async_modeset);
827	858	if (ret) {
828	859	drm_framebuffer_put(fb);
	860	+ vc4_bo_dec_usecnt(bo);
829	861	kfree(flip_state);
830	862	return ret;
831	863	}
832	864
	865	+ /* Save the current FB before it's replaced by the new one in
	866	+ * drm_atomic_set_fb_for_plane(). We'll need the old FB in
	867	+ * vc4_async_page_flip_complete() to decrement the BO usecnt and keep
	868	+ * it consistent.
	869	+ * FIXME: we should move to generic async-page-flip when it's
	870	+ * available, so that we can get rid of this hand-made cleanup_fb()
	871	+ * logic.
	872	+ */
	873	+ flip_state->old_fb = plane->state->fb;
	874	+ if (flip_state->old_fb)
	875	+ drm_framebuffer_get(flip_state->old_fb);
	876	+
833	877	WARN_ON(drm_crtc_vblank_get(crtc) != 0);
834	878
835	879	/* Immediately update the plane's legacy fb pointer, so that later

drivers/gpu/drm/vc4/vc4_validate_shaders.c+1−0

@@ -942,6 +942,7 @@ vc4_validate_shader(struct drm_gem_cma_object *shader_obj) · 941 unmodified lines
942	942	fail:
943	943	kfree(validation_state.branch_targets);
944	944	if (validated_shader) {
	945	+ kfree(validated_shader->uniform_addr_offsets);
945	946	kfree(validated_shader->texture_samples);
946	947	kfree(validated_shader);
947	948	}

drivers/gpu/drm/virtio/virtgpu_vq.c+2−2

@@ -293,7 +293,7 @@ retry: · 292 unmodified lines
293	293	ret = virtqueue_add_sgs(vq, sgs, outcnt, incnt, vbuf, GFP_ATOMIC);
294	294	if (ret == -ENOSPC) {
295	295	spin_unlock(&vgdev->ctrlq.qlock);
296		− wait_event(vgdev->ctrlq.ack_queue, vq->num_free);
	296	+ wait_event(vgdev->ctrlq.ack_queue, vq->num_free >= outcnt + incnt);
297	297	spin_lock(&vgdev->ctrlq.qlock);
298	298	goto retry;
299	299	} else {
@@ -368,7 +368,7 @@ retry: · 68 unmodified lines
368	368	ret = virtqueue_add_sgs(vq, sgs, outcnt, 0, vbuf, GFP_ATOMIC);
369	369	if (ret == -ENOSPC) {
370	370	spin_unlock(&vgdev->cursorq.qlock);
371		− wait_event(vgdev->cursorq.ack_queue, vq->num_free);
	371	+ wait_event(vgdev->cursorq.ack_queue, vq->num_free >= outcnt);
372	372	spin_lock(&vgdev->cursorq.qlock);
373	373	goto retry;
374	374	} else {

drivers/gpu/drm/vmwgfx/vmwgfx_fb.c+10−21

@@ -441,11 +441,11 @@ static int vmwgfx_set_config_internal(struct drm_mode_set *set) · 440 unmodified lines
441	441	struct drm_crtc *crtc = set->crtc;
442	442	struct drm_framebuffer *fb;
443	443	struct drm_crtc *tmp;
444		− struct drm_modeset_acquire_ctx *ctx;
445	444	struct drm_device *dev = set->crtc->dev;
	445	+ struct drm_modeset_acquire_ctx ctx;
446	446	int ret;
447	447
448		− ctx = dev->mode_config.acquire_ctx;
	448	+ drm_modeset_acquire_init(&ctx, 0);
449	449
450	450	restart:
451	451	/*
@@ -458,7 +458,7 @@ restart: · 6 unmodified lines
458	458
459	459	fb = set->fb;
460	460
461		− ret = crtc->funcs->set_config(set, ctx);
	461	+ ret = crtc->funcs->set_config(set, &ctx);
462	462	if (ret == 0) {
463	463	crtc->primary->crtc = crtc;
464	464	crtc->primary->fb = fb;
@@ -473,20 +473,13 @@ restart: · 8 unmodified lines
473	473	}
474	474
475	475	if (ret == -EDEADLK) {
476		− dev->mode_config.acquire_ctx = NULL;
477		−
478		−retry_locking:
479		− drm_modeset_backoff(ctx);
480		−
481		− ret = drm_modeset_lock_all_ctx(dev, ctx);
482		− if (ret)
483		− goto retry_locking;
484		−
485		− dev->mode_config.acquire_ctx = ctx;
486		−
	476	+ drm_modeset_backoff(&ctx);
487	477	goto restart;
488	478	}
489	479
	480	+ drm_modeset_drop_locks(&ctx);
	481	+ drm_modeset_acquire_fini(&ctx);
	482	+
490	483	return ret;
491	484	}
492	485
@@ -624,7 +617,6 @@ static int vmw_fb_set_par(struct fb_info *info) · 131 unmodified lines
624	617	}
625	618
626	619	mutex_lock(&par->bo_mutex);
627		− drm_modeset_lock_all(vmw_priv->dev);
628	620	ret = vmw_fb_kms_framebuffer(info);
629	621	if (ret)
630	622	goto out_unlock;
@@ -657,7 +649,6 @@ out_unlock: · 26 unmodified lines
657	649	drm_mode_destroy(vmw_priv->dev, old_mode);
658	650	par->set_mode = mode;
659	651
660		− drm_modeset_unlock_all(vmw_priv->dev);
661	652	mutex_unlock(&par->bo_mutex);
662	653
663	654	return ret;
@@ -713,18 +704,14 @@ int vmw_fb_init(struct vmw_private *vmw_priv) · 49 unmodified lines
713	704	par->max_width = fb_width;
714	705	par->max_height = fb_height;
715	706
716		− drm_modeset_lock_all(vmw_priv->dev);
717	707	ret = vmw_kms_fbdev_init_data(vmw_priv, 0, par->max_width,
718	708	par->max_height, &par->con,
719	709	&par->crtc, &init_mode);
720		− if (ret) {
721		− drm_modeset_unlock_all(vmw_priv->dev);
	710	+ if (ret)
722	711	goto err_kms;
723		− }
724	712
725	713	info->var.xres = init_mode->hdisplay;
726	714	info->var.yres = init_mode->vdisplay;
727		− drm_modeset_unlock_all(vmw_priv->dev);
728	715
729	716	/*
730	717	* Create buffers and alloc memory
@@ -832,7 +819,9 @@ int vmw_fb_close(struct vmw_private *vmw_priv) · 101 unmodified lines
832	819	cancel_delayed_work_sync(&par->local_work);
833	820	unregister_framebuffer(info);
834	821
	822	+ mutex_lock(&par->bo_mutex);
835	823	(void) vmw_fb_kms_detach(par, true, true);
	824	+ mutex_unlock(&par->bo_mutex);
836	825
837	826	vfree(par->vmalloc);
838	827	framebuffer_release(info);

drivers/gpu/drm/vmwgfx/vmwgfx_kms.c+11−3

@@ -2595,6 +2595,7 @@ void vmw_kms_helper_resource_finish(struct vmw_validation_ctx *ctx, · 2594 unmodified lines
2595	2595	vmw_kms_helper_buffer_finish(res->dev_priv, NULL, ctx->buf,
2596	2596	out_fence, NULL);
2597	2597
	2598	+ vmw_dmabuf_unreference(&ctx->buf);
2598	2599	vmw_resource_unreserve(res, false, NULL, 0);
2599	2600	mutex_unlock(&res->dev_priv->cmdbuf_mutex);
2600	2601	}
@@ -2680,7 +2681,9 @@ int vmw_kms_fbdev_init_data(struct vmw_private *dev_priv, · 79 unmodified lines
2680	2681	struct vmw_display_unit *du;
2681	2682	struct drm_display_mode *mode;
2682	2683	int i = 0;
	2684	+ int ret = 0;
2683	2685
	2686	+ mutex_lock(&dev_priv->dev->mode_config.mutex);
2684	2687	list_for_each_entry(con, &dev_priv->dev->mode_config.connector_list,
2685	2688	head) {
2686	2689	if (i == unit)
@@ -2691,7 +2694,8 @@ int vmw_kms_fbdev_init_data(struct vmw_private *dev_priv, · 4 unmodified lines
2691	2694
2692	2695	if (i != unit) {
2693	2696	DRM_ERROR("Could not find initial display unit.\n");
2694		− return -EINVAL;
	2697	+ ret = -EINVAL;
	2698	+ goto out_unlock;
2695	2699	}
2696	2700
2697	2701	if (list_empty(&con->modes))
@@ -2699,7 +2703,8 @@ int vmw_kms_fbdev_init_data(struct vmw_private *dev_priv, · 1 unmodified line
2699	2703
2700	2704	if (list_empty(&con->modes)) {
2701	2705	DRM_ERROR("Could not find initial display mode.\n");
2702		− return -EINVAL;
	2706	+ ret = -EINVAL;
	2707	+ goto out_unlock;
2703	2708	}
2704	2709
2705	2710	du = vmw_connector_to_du(con);
@@ -2720,7 +2725,10 @@ int vmw_kms_fbdev_init_data(struct vmw_private *dev_priv, · 14 unmodified lines
2720	2725	head);
2721	2726	}
2722	2727
2723		− return 0;
	2728	+ out_unlock:
	2729	+ mutex_unlock(&dev_priv->dev->mode_config.mutex);
	2730	+
	2731	+ return ret;
2724	2732	}
2725	2733
2726	2734	/**

drivers/hid/hid-ids.h+3−0

@@ -525,6 +525,9 @@ · 524 unmodified lines
525	525	#define I2C_VENDOR_ID_HANTICK 0x0911
526	526	#define I2C_PRODUCT_ID_HANTICK_5288 0x5288
527	527
	528	+#define I2C_VENDOR_ID_RAYD 0x2386
	529	+#define I2C_PRODUCT_ID_RAYD_3118 0x3118
	530	+
528	531	#define USB_VENDOR_ID_HANWANG 0x0b57
529	532	#define USB_DEVICE_ID_HANWANG_TABLET_FIRST 0x5000
530	533	#define USB_DEVICE_ID_HANWANG_TABLET_LAST 0x8fff

drivers/hid/hid-input.c+17−7

@@ -387,7 +387,8 @@ static int hidinput_get_battery_property(struct power_supply *psy, · 386 unmodified lines
387	387	break;
388	388
389	389	case POWER_SUPPLY_PROP_CAPACITY:
390		− if (dev->battery_report_type == HID_FEATURE_REPORT) {
	390	+ if (dev->battery_status != HID_BATTERY_REPORTED &&
	391	+ !dev->battery_avoid_query) {
391	392	value = hidinput_query_battery_capacity(dev);
392	393	if (value < 0)
393	394	return value;
@@ -403,17 +404,17 @@ static int hidinput_get_battery_property(struct power_supply *psy, · 9 unmodified lines
403	404	break;
404	405
405	406	case POWER_SUPPLY_PROP_STATUS:
406		− if (!dev->battery_reported &&
407		− dev->battery_report_type == HID_FEATURE_REPORT) {
	407	+ if (dev->battery_status != HID_BATTERY_REPORTED &&
	408	+ !dev->battery_avoid_query) {
408	409	value = hidinput_query_battery_capacity(dev);
409	410	if (value < 0)
410	411	return value;
411	412
412	413	dev->battery_capacity = value;
413		− dev->battery_reported = true;
	414	+ dev->battery_status = HID_BATTERY_QUERIED;
414	415	}
415	416
416		− if (!dev->battery_reported)
	417	+ if (dev->battery_status == HID_BATTERY_UNKNOWN)
417	418	val->intval = POWER_SUPPLY_STATUS_UNKNOWN;
418	419	else if (dev->battery_capacity == 100)
419	420	val->intval = POWER_SUPPLY_STATUS_FULL;
@@ -486,6 +487,14 @@ static int hidinput_setup_battery(struct hid_device *dev, unsigned report_type, · 66 unmodified lines
486	487	dev->battery_report_type = report_type;
487	488	dev->battery_report_id = field->report->id;
488	489
	490	+ /*
	491	+ * Stylus is normally not connected to the device and thus we
	492	+ * can't query the device and get meaningful battery strength.
	493	+ * We have to wait for the device to report it on its own.
	494	+ */
	495	+ dev->battery_avoid_query = report_type == HID_INPUT_REPORT &&
	496	+ field->physical == HID_DG_STYLUS;
	497	+
489	498	dev->battery = power_supply_register(&dev->dev, psy_desc, &psy_cfg);
490	499	if (IS_ERR(dev->battery)) {
491	500	error = PTR_ERR(dev->battery);
@@ -530,9 +539,10 @@ static void hidinput_update_battery(struct hid_device *dev, int value) · 38 unmodified lines
530	539
531	540	capacity = hidinput_scale_battery_capacity(dev, value);
532	541
533		− if (!dev->battery_reported \|\| capacity != dev->battery_capacity) {
	542	+ if (dev->battery_status != HID_BATTERY_REPORTED \|\|
	543	+ capacity != dev->battery_capacity) {
534	544	dev->battery_capacity = capacity;
535		− dev->battery_reported = true;
	545	+ dev->battery_status = HID_BATTERY_REPORTED;
536	546	power_supply_changed(dev->battery);
537	547	}
538	548	}

drivers/hid/hidraw.c+5−0

@@ -192,6 +192,11 @@ static ssize_t hidraw_get_report(struct file file, char __user buffer, size_t · 191 unmodified lines
192	192	int ret = 0, len;
193	193	unsigned char report_number;
194	194
	195	+ if (!hidraw_table[minor] \|\| !hidraw_table[minor]->exist) {
	196	+ ret = -ENODEV;
	197	+ goto out;
	198	+ }
	199	+
195	200	dev = hidraw_table[minor]->hid;
196	201
197	202	if (!dev->ll_driver->raw_request) {

drivers/hid/i2c-hid/i2c-hid.c+13−0

@@ -47,6 +47,7 @@ · 46 unmodified lines
47	47	/* quirks to control the device */
48	48	#define I2C_HID_QUIRK_SET_PWR_WAKEUP_DEV BIT(0)
49	49	#define I2C_HID_QUIRK_NO_IRQ_AFTER_RESET BIT(1)
	50	+#define I2C_HID_QUIRK_RESEND_REPORT_DESCR BIT(2)
50	51
51	52	/* flags */
52	53	#define I2C_HID_STARTED 0
@@ -171,6 +172,8 @@ static const struct i2c_hid_quirks { · 118 unmodified lines
171	172	I2C_HID_QUIRK_SET_PWR_WAKEUP_DEV },
172	173	{ I2C_VENDOR_ID_HANTICK, I2C_PRODUCT_ID_HANTICK_5288,
173	174	I2C_HID_QUIRK_NO_IRQ_AFTER_RESET },
	175	+ { I2C_VENDOR_ID_RAYD, I2C_PRODUCT_ID_RAYD_3118,
	176	+ I2C_HID_QUIRK_RESEND_REPORT_DESCR },
174	177	{ 0, 0 }
175	178	};
176	179
@@ -1220,6 +1223,16 @@ static int i2c_hid_resume(struct device *dev) · 1043 unmodified lines
1220	1223	if (ret)
1221	1224	return ret;
1222	1225
	1226	+ /* RAYDIUM device (2386:3118) need to re-send report descr cmd
	1227	+ * after resume, after this it will be back normal.
	1228	+ * otherwise it issues too many incomplete reports.
	1229	+ */
	1230	+ if (ihid->quirks & I2C_HID_QUIRK_RESEND_REPORT_DESCR) {
	1231	+ ret = i2c_hid_command(client, &hid_report_descr_cmd, NULL, 0);
	1232	+ if (ret)
	1233	+ return ret;
	1234	+ }
	1235	+
1223	1236	if (hid->driver && hid->driver->reset_resume) {
1224	1237	ret = hid->driver->reset_resume(hid);
1225	1238	return ret;

drivers/hid/wacom_wac.c+46−30

@@ -689,6 +689,45 @@ static int wacom_intuos_get_tool_type(int tool_id) · 688 unmodified lines
689	689	return tool_type;
690	690	}
691	691
	692	+static void wacom_exit_report(struct wacom_wac *wacom)
	693	+{
	694	+ struct input_dev *input = wacom->pen_input;
	695	+ struct wacom_features *features = &wacom->features;
	696	+ unsigned char *data = wacom->data;
	697	+ int idx = (features->type == INTUOS) ? (data[1] & 0x01) : 0;
	698	+
	699	+ /*
	700	+ * Reset all states otherwise we lose the initial states
	701	+ * when in-prox next time
	702	+ */
	703	+ input_report_abs(input, ABS_X, 0);
	704	+ input_report_abs(input, ABS_Y, 0);
	705	+ input_report_abs(input, ABS_DISTANCE, 0);
	706	+ input_report_abs(input, ABS_TILT_X, 0);
	707	+ input_report_abs(input, ABS_TILT_Y, 0);
	708	+ if (wacom->tool[idx] >= BTN_TOOL_MOUSE) {
	709	+ input_report_key(input, BTN_LEFT, 0);
	710	+ input_report_key(input, BTN_MIDDLE, 0);
	711	+ input_report_key(input, BTN_RIGHT, 0);
	712	+ input_report_key(input, BTN_SIDE, 0);
	713	+ input_report_key(input, BTN_EXTRA, 0);
	714	+ input_report_abs(input, ABS_THROTTLE, 0);
	715	+ input_report_abs(input, ABS_RZ, 0);
	716	+ } else {
	717	+ input_report_abs(input, ABS_PRESSURE, 0);
	718	+ input_report_key(input, BTN_STYLUS, 0);
	719	+ input_report_key(input, BTN_STYLUS2, 0);
	720	+ input_report_key(input, BTN_TOUCH, 0);
	721	+ input_report_abs(input, ABS_WHEEL, 0);
	722	+ if (features->type >= INTUOS3S)
	723	+ input_report_abs(input, ABS_Z, 0);
	724	+ }
	725	+ input_report_key(input, wacom->tool[idx], 0);
	726	+ input_report_abs(input, ABS_MISC, 0); /* reset tool id */
	727	+ input_event(input, EV_MSC, MSC_SERIAL, wacom->serial[idx]);
	728	+ wacom->id[idx] = 0;
	729	+}
	730	+
692	731	static int wacom_intuos_inout(struct wacom_wac *wacom)
693	732	{
694	733	struct wacom_features *features = &wacom->features;
@@ -741,36 +780,7 @@ static int wacom_intuos_inout(struct wacom_wac *wacom) · 46 unmodified lines
741	780	if (!wacom->id[idx])
742	781	return 1;
743	782
744		− /*
745		− * Reset all states otherwise we lose the initial states
746		− * when in-prox next time
747		− */
748		− input_report_abs(input, ABS_X, 0);
749		− input_report_abs(input, ABS_Y, 0);
750		− input_report_abs(input, ABS_DISTANCE, 0);
751		− input_report_abs(input, ABS_TILT_X, 0);
752		− input_report_abs(input, ABS_TILT_Y, 0);
753		− if (wacom->tool[idx] >= BTN_TOOL_MOUSE) {
754		− input_report_key(input, BTN_LEFT, 0);
755		− input_report_key(input, BTN_MIDDLE, 0);
756		− input_report_key(input, BTN_RIGHT, 0);
757		− input_report_key(input, BTN_SIDE, 0);
758		− input_report_key(input, BTN_EXTRA, 0);
759		− input_report_abs(input, ABS_THROTTLE, 0);
760		− input_report_abs(input, ABS_RZ, 0);
761		− } else {
762		− input_report_abs(input, ABS_PRESSURE, 0);
763		− input_report_key(input, BTN_STYLUS, 0);
764		− input_report_key(input, BTN_STYLUS2, 0);
765		− input_report_key(input, BTN_TOUCH, 0);
766		− input_report_abs(input, ABS_WHEEL, 0);
767		− if (features->type >= INTUOS3S)
768		− input_report_abs(input, ABS_Z, 0);
769		− }
770		− input_report_key(input, wacom->tool[idx], 0);
771		− input_report_abs(input, ABS_MISC, 0); /* reset tool id */
772		− input_event(input, EV_MSC, MSC_SERIAL, wacom->serial[idx]);
773		− wacom->id[idx] = 0;
	783	+ wacom_exit_report(wacom);
774	784	return 2;
775	785	}
776	786
@@ -1235,6 +1245,12 @@ static void wacom_intuos_pro2_bt_pen(struct wacom_wac *wacom) · 458 unmodified lines
1235	1245	if (!valid)
1236	1246	continue;
1237	1247
	1248	+ if (!prox) {
	1249	+ wacom->shared->stylus_in_proximity = false;
	1250	+ wacom_exit_report(wacom);
	1251	+ input_sync(pen_input);
	1252	+ return;
	1253	+ }
1238	1254	if (range) {
1239	1255	input_report_abs(pen_input, ABS_X, get_unaligned_le16(&frame[1]));
1240	1256	input_report_abs(pen_input, ABS_Y, get_unaligned_le16(&frame[3]));

drivers/hwmon/k10temp.c+14−3

@@ -40,6 +40,10 @@ static DEFINE_MUTEX(nb_smu_ind_mutex); · 39 unmodified lines
40	40	#define PCI_DEVICE_ID_AMD_17H_DF_F3 0x1463
41	41	#endif
42	42
	43	+#ifndef PCI_DEVICE_ID_AMD_17H_RR_NB
	44	+#define PCI_DEVICE_ID_AMD_17H_RR_NB 0x15d0
	45	+#endif
	46	+
43	47	/* CPUID function 0x80000001, ebx */
44	48	#define CPUID_PKGTYPE_MASK 0xf0000000
45	49	#define CPUID_PKGTYPE_F 0x00000000
@@ -72,6 +76,7 @@ struct k10temp_data { · 26 unmodified lines
72	76	struct pci_dev *pdev;
73	77	void (read_tempreg)(struct pci_dev pdev, u32 *regval);
74	78	int temp_offset;
	79	+ u32 temp_adjust_mask;
75	80	};
76	81
77	82	struct tctl_offset {
@@ -84,6 +89,7 @@ static const struct tctl_offset tctl_offset_table[] = { · 6 unmodified lines
84	89	{ 0x17, "AMD Ryzen 5 1600X", 20000 },
85	90	{ 0x17, "AMD Ryzen 7 1700X", 20000 },
86	91	{ 0x17, "AMD Ryzen 7 1800X", 20000 },
	92	+ { 0x17, "AMD Ryzen 7 2700X", 10000 },
87	93	{ 0x17, "AMD Ryzen Threadripper 1950X", 27000 },
88	94	{ 0x17, "AMD Ryzen Threadripper 1920X", 27000 },
89	95	{ 0x17, "AMD Ryzen Threadripper 1900X", 27000 },
@@ -129,6 +135,8 @@ static ssize_t temp1_input_show(struct device *dev, · 39 unmodified lines
129	135
130	136	data->read_tempreg(data->pdev, &regval);
131	137	temp = (regval >> 21) * 125;
	138	+ if (regval & data->temp_adjust_mask)
	139	+ temp -= 49000;
132	140	if (temp > data->temp_offset)
133	141	temp -= data->temp_offset;
134	142	else
@@ -259,12 +267,14 @@ static int k10temp_probe(struct pci_dev *pdev, · 124 unmodified lines
259	267	data->pdev = pdev;
260	268
261	269	if (boot_cpu_data.x86 == 0x15 && (boot_cpu_data.x86_model == 0x60 \|\|
262		− boot_cpu_data.x86_model == 0x70))
	270	+ boot_cpu_data.x86_model == 0x70)) {
263	271	data->read_tempreg = read_tempreg_nb_f15;
264		− else if (boot_cpu_data.x86 == 0x17)
	272	+ } else if (boot_cpu_data.x86 == 0x17) {
	273	+ data->temp_adjust_mask = 0x80000;
265	274	data->read_tempreg = read_tempreg_nb_f17;
266		− else
	275	+ } else {
267	276	data->read_tempreg = read_tempreg_pci;
	277	+ }
268	278
269	279	for (i = 0; i < ARRAY_SIZE(tctl_offset_table); i++) {
270	280	const struct tctl_offset *entry = &tctl_offset_table[i];
@@ -292,6 +302,7 @@ static const struct pci_device_id k10temp_id_table[] = { · 21 unmodified lines
292	302	{ PCI_VDEVICE(AMD, PCI_DEVICE_ID_AMD_16H_NB_F3) },
293	303	{ PCI_VDEVICE(AMD, PCI_DEVICE_ID_AMD_16H_M30H_NB_F3) },
294	304	{ PCI_VDEVICE(AMD, PCI_DEVICE_ID_AMD_17H_DF_F3) },
	305	+ { PCI_VDEVICE(AMD, PCI_DEVICE_ID_AMD_17H_RR_NB) },
295	306	{}
296	307	};
297	308	MODULE_DEVICE_TABLE(pci, k10temp_id_table);

drivers/hwmon/nct6683.c+2−2

@@ -1380,8 +1380,8 @@ static int __init nct6683_find(int sioaddr, struct nct6683_sio_data *sio_data) · 1379 unmodified lines
1380	1380	/* Activate logical device if needed */
1381	1381	val = superio_inb(sioaddr, SIO_REG_ENABLE);
1382	1382	if (!(val & 0x01)) {
1383		− pr_err("EC is disabled\n");
1384		− goto fail;
	1383	+ pr_warn("Forcibly enabling EC access. Data may be unusable.\n");
	1384	+ superio_outb(sioaddr, SIO_REG_ENABLE, val \| 0x01);
1385	1385	}
1386	1386
1387	1387	superio_exit(sioaddr);

drivers/hwmon/scmi-hwmon.c+4−1

@@ -170,7 +170,10 @@ static int scmi_hwmon_probe(struct scmi_device *sdev) · 169 unmodified lines
170	170	scmi_chip_info.info = ptr_scmi_ci;
171	171	chip_info = &scmi_chip_info;
172	172
173		− for (type = 0; type < hwmon_max && nr_count[type]; type++) {
	173	+ for (type = 0; type < hwmon_max; type++) {
	174	+ if (!nr_count[type])
	175	+ continue;
	176	+
174	177	scmi_hwmon_add_chan_info(scmi_hwmon_chan, dev, nr_count[type],
175	178	type, hwmon_attributes[type]);
176	179	*ptr_scmi_ci++ = scmi_hwmon_chan++;

drivers/i2c/busses/Kconfig+0−3

@@ -707,7 +707,6 @@ config I2C_MPC · 706 unmodified lines
707	707	config I2C_MT65XX
708	708	tristate "MediaTek I2C adapter"
709	709	depends on ARCH_MEDIATEK \|\| COMPILE_TEST
710		− depends on HAS_DMA
711	710	help
712	711	This selects the MediaTek(R) Integrated Inter Circuit bus driver
713	712	for MT65xx and MT81xx.
@@ -885,7 +884,6 @@ config I2C_SH7760 · 171 unmodified lines
885	884
886	885	config I2C_SH_MOBILE
887	886	tristate "SuperH Mobile I2C Controller"
888		− depends on HAS_DMA
889	887	depends on ARCH_SHMOBILE \|\| ARCH_RENESAS \|\| COMPILE_TEST
890	888	help
891	889	If you say yes to this option, support will be included for the
@@ -1098,7 +1096,6 @@ config I2C_XLP9XX · 206 unmodified lines
1098	1096
1099	1097	config I2C_RCAR
1100	1098	tristate "Renesas R-Car I2C Controller"
1101		− depends on HAS_DMA
1102	1099	depends on ARCH_RENESAS \|\| COMPILE_TEST
1103	1100	select I2C_SLAVE
1104	1101	help

drivers/i2c/busses/i2c-sprd.c+18−4

@@ -86,6 +86,7 @@ struct sprd_i2c { · 85 unmodified lines
86	86	u32 count;
87	87	int irq;
88	88	int err;
	89	+ bool is_suspended;
89	90	};
90	91
91	92	static void sprd_i2c_set_count(struct sprd_i2c *i2c_dev, u32 count)
@@ -283,6 +284,9 @@ static int sprd_i2c_master_xfer(struct i2c_adapter *i2c_adap, · 191 unmodified lines
283	284	struct sprd_i2c *i2c_dev = i2c_adap->algo_data;
284	285	int im, ret;
285	286
	287	+ if (i2c_dev->is_suspended)
	288	+ return -EBUSY;
	289	+
286	290	ret = pm_runtime_get_sync(i2c_dev->dev);
287	291	if (ret < 0)
288	292	return ret;
@@ -364,13 +368,12 @@ static irqreturn_t sprd_i2c_isr_thread(int irq, void *dev_id) · 75 unmodified lines
364	368	struct sprd_i2c *i2c_dev = dev_id;
365	369	struct i2c_msg *msg = i2c_dev->msg;
366	370	bool ack = !(readl(i2c_dev->base + I2C_STATUS) & I2C_RX_ACK);
367		− u32 i2c_count = readl(i2c_dev->base + I2C_COUNT);
368	371	u32 i2c_tran;
369	372
370	373	if (msg->flags & I2C_M_RD)
371	374	i2c_tran = i2c_dev->count >= I2C_FIFO_FULL_THLD;
372	375	else
373		− i2c_tran = i2c_count;
	376	+ i2c_tran = i2c_dev->count;
374	377
375	378	/*
376	379	* If we got one ACK from slave when writing data, and we did not
@@ -408,14 +411,13 @@ static irqreturn_t sprd_i2c_isr(int irq, void *dev_id) · 31 unmodified lines
408	411	{
409	412	struct sprd_i2c *i2c_dev = dev_id;
410	413	struct i2c_msg *msg = i2c_dev->msg;
411		− u32 i2c_count = readl(i2c_dev->base + I2C_COUNT);
412	414	bool ack = !(readl(i2c_dev->base + I2C_STATUS) & I2C_RX_ACK);
413	415	u32 i2c_tran;
414	416
415	417	if (msg->flags & I2C_M_RD)
416	418	i2c_tran = i2c_dev->count >= I2C_FIFO_FULL_THLD;
417	419	else
418		− i2c_tran = i2c_count;
	420	+ i2c_tran = i2c_dev->count;
419	421
420	422	/*
421	423	* If we did not get one ACK from slave when writing data, then we
@@ -586,11 +588,23 @@ static int sprd_i2c_remove(struct platform_device *pdev) · 164 unmodified lines
586	588
587	589	static int __maybe_unused sprd_i2c_suspend_noirq(struct device *pdev)
588	590	{
	591	+ struct sprd_i2c *i2c_dev = dev_get_drvdata(pdev);
	592	+
	593	+ i2c_lock_adapter(&i2c_dev->adap);
	594	+ i2c_dev->is_suspended = true;
	595	+ i2c_unlock_adapter(&i2c_dev->adap);
	596	+
589	597	return pm_runtime_force_suspend(pdev);
590	598	}
591	599
592	600	static int __maybe_unused sprd_i2c_resume_noirq(struct device *pdev)
593	601	{
	602	+ struct sprd_i2c *i2c_dev = dev_get_drvdata(pdev);
	603	+
	604	+ i2c_lock_adapter(&i2c_dev->adap);
	605	+ i2c_dev->is_suspended = false;
	606	+ i2c_unlock_adapter(&i2c_dev->adap);
	607	+
594	608	return pm_runtime_force_resume(pdev);
595	609	}
596	610

drivers/i2c/i2c-dev.c+1−1

@@ -280,7 +280,7 @@ static noinline int i2cdev_ioctl_rdwr(struct i2c_client *client, · 279 unmodified lines
280	280	*/
281	281	if (msgs[i].flags & I2C_M_RECV_LEN) {
282	282	if (!(msgs[i].flags & I2C_M_RD) \|\|
283		− msgs[i].buf[0] < 1 \|\|
	283	+ msgs[i].len < 1 \|\| msgs[i].buf[0] < 1 \|\|
284	284	msgs[i].len < msgs[i].buf[0] +
285	285	I2C_SMBUS_BLOCK_MAX) {
286	286	res = -EINVAL;

drivers/infiniband/Kconfig+4−1

@@ -61,9 +61,12 @@ config INFINIBAND_ON_DEMAND_PAGING · 60 unmodified lines
61	61	pages on demand instead.
62	62
63	63	config INFINIBAND_ADDR_TRANS
64		− bool
	64	+ bool "RDMA/CM"
65	65	depends on INFINIBAND
66	66	default y
	67	+ ---help---
	68	+ Support for RDMA communication manager (CM).
	69	+ This allows for a generic connection abstraction over RDMA.
67	70
68	71	config INFINIBAND_ADDR_TRANS_CONFIGFS
69	72	bool

drivers/infiniband/core/cache.c+35−20

@@ -291,14 +291,18 @@ static int find_gid(struct ib_gid_table table, const union ib_gid gid, · 290 unmodified lines
291	291	* so lookup free slot only if requested.
292	292	*/
293	293	if (pempty && empty < 0) {
294		− if (data->props & GID_TABLE_ENTRY_INVALID) {
295		− /* Found an invalid (free) entry; allocate it */
296		− if (data->props & GID_TABLE_ENTRY_DEFAULT) {
297		− if (default_gid)
298		− empty = curr_index;
299		− } else {
300		− empty = curr_index;
301		− }
	294	+ if (data->props & GID_TABLE_ENTRY_INVALID &&
	295	+ (default_gid ==
	296	+ !!(data->props & GID_TABLE_ENTRY_DEFAULT))) {
	297	+ /*
	298	+ * Found an invalid (free) entry; allocate it.
	299	+ * If default GID is requested, then our
	300	+ * found slot must be one of the DEFAULT
	301	+ * reserved slots or we fail.
	302	+ * This ensures that only DEFAULT reserved
	303	+ * slots are used for default property GIDs.
	304	+ */
	305	+ empty = curr_index;
302	306	}
303	307	}
304	308
@@ -420,8 +424,10 @@ int ib_cache_gid_add(struct ib_device *ib_dev, u8 port, · 115 unmodified lines
420	424	return ret;
421	425	}
422	426
423		−int ib_cache_gid_del(struct ib_device *ib_dev, u8 port,
424		− union ib_gid gid, struct ib_gid_attr attr)
	427	+static int
	428	+_ib_cache_gid_del(struct ib_device *ib_dev, u8 port,
	429	+ union ib_gid gid, struct ib_gid_attr attr,
	430	+ unsigned long mask, bool default_gid)
425	431	{
426	432	struct ib_gid_table *table;
427	433	int ret = 0;
@@ -431,11 +437,7 @@ int ib_cache_gid_del(struct ib_device *ib_dev, u8 port, · 3 unmodified lines
431	437
432	438	mutex_lock(&table->lock);
433	439
434		− ix = find_gid(table, gid, attr, false,
435		− GID_ATTR_FIND_MASK_GID \|
436		− GID_ATTR_FIND_MASK_GID_TYPE \|
437		− GID_ATTR_FIND_MASK_NETDEV,
438		− NULL);
	440	+ ix = find_gid(table, gid, attr, default_gid, mask, NULL);
439	441	if (ix < 0) {
440	442	ret = -EINVAL;
441	443	goto out_unlock;
@@ -452,6 +454,17 @@ out_unlock: · 10 unmodified lines
452	454	return ret;
453	455	}
454	456
	457	+int ib_cache_gid_del(struct ib_device *ib_dev, u8 port,
	458	+ union ib_gid gid, struct ib_gid_attr attr)
	459	+{
	460	+ unsigned long mask = GID_ATTR_FIND_MASK_GID \|
	461	+ GID_ATTR_FIND_MASK_GID_TYPE \|
	462	+ GID_ATTR_FIND_MASK_DEFAULT \|
	463	+ GID_ATTR_FIND_MASK_NETDEV;
	464	+
	465	+ return _ib_cache_gid_del(ib_dev, port, gid, attr, mask, false);
	466	+}
	467	+
455	468	int ib_cache_gid_del_all_netdev_gids(struct ib_device *ib_dev, u8 port,
456	469	struct net_device *ndev)
457	470	{
@@ -728,7 +741,7 @@ void ib_cache_gid_set_default_gid(struct ib_device *ib_dev, u8 port, · 270 unmodified lines
728	741	unsigned long gid_type_mask,
729	742	enum ib_cache_gid_default_mode mode)
730	743	{
731		− union ib_gid gid;
	744	+ union ib_gid gid = { };
732	745	struct ib_gid_attr gid_attr;
733	746	struct ib_gid_table *table;
734	747	unsigned int gid_type;
@@ -736,7 +749,9 @@ void ib_cache_gid_set_default_gid(struct ib_device *ib_dev, u8 port, · 1 unmodified line
736	749
737	750	table = ib_dev->cache.ports[port - rdma_start_port(ib_dev)].gid;
738	751
739		− make_default_gid(ndev, &gid);
	752	+ mask = GID_ATTR_FIND_MASK_GID_TYPE \|
	753	+ GID_ATTR_FIND_MASK_DEFAULT \|
	754	+ GID_ATTR_FIND_MASK_NETDEV;
740	755	memset(&gid_attr, 0, sizeof(gid_attr));
741	756	gid_attr.ndev = ndev;
742	757
@@ -747,12 +762,12 @@ void ib_cache_gid_set_default_gid(struct ib_device *ib_dev, u8 port, · 4 unmodified lines
747	762	gid_attr.gid_type = gid_type;
748	763
749	764	if (mode == IB_CACHE_GID_DEFAULT_MODE_SET) {
750		− mask = GID_ATTR_FIND_MASK_GID_TYPE \|
751		− GID_ATTR_FIND_MASK_DEFAULT;
	765	+ make_default_gid(ndev, &gid);
752	766	__ib_cache_gid_add(ib_dev, port, &gid,
753	767	&gid_attr, mask, true);
754	768	} else if (mode == IB_CACHE_GID_DEFAULT_MODE_DELETE) {
755		− ib_cache_gid_del(ib_dev, port, &gid, &gid_attr);
	769	+ _ib_cache_gid_del(ib_dev, port, &gid,
	770	+ &gid_attr, mask, true);
756	771	}
757	772	}
758	773	}

drivers/infiniband/core/cma.c+43−17

@@ -382,6 +382,8 @@ struct cma_hdr { · 381 unmodified lines
382	382	#define CMA_VERSION 0x00
383	383
384	384	struct cma_req_info {
	385	+ struct sockaddr_storage listen_addr_storage;
	386	+ struct sockaddr_storage src_addr_storage;
385	387	struct ib_device *device;
386	388	int port;
387	389	union ib_gid local_gid;
@@ -866,7 +868,6 @@ static int cma_modify_qp_rtr(struct rdma_id_private *id_priv, · 478 unmodified lines
866	868	{
867	869	struct ib_qp_attr qp_attr;
868	870	int qp_attr_mask, ret;
869		− union ib_gid sgid;
870	871
871	872	mutex_lock(&id_priv->qp_mutex);
872	873	if (!id_priv->id.qp) {
@@ -889,12 +890,6 @@ static int cma_modify_qp_rtr(struct rdma_id_private *id_priv, · 16 unmodified lines
889	890	if (ret)
890	891	goto out;
891	892
892		− ret = ib_query_gid(id_priv->id.device, id_priv->id.port_num,
893		− rdma_ah_read_grh(&qp_attr.ah_attr)->sgid_index,
894		− &sgid, NULL);
895		− if (ret)
896		− goto out;
897		−
898	893	BUG_ON(id_priv->cma_dev->device != id_priv->id.device);
899	894
900	895	if (conn_param)
@@ -1340,11 +1335,11 @@ static bool validate_net_dev(struct net_device *net_dev, · 439 unmodified lines
1340	1335	}
1341	1336
1342	1337	static struct net_device cma_get_net_dev(struct ib_cm_event ib_event,
1343		− const struct cma_req_info *req)
	1338	+ struct cma_req_info *req)
1344	1339	{
1345		− struct sockaddr_storage listen_addr_storage, src_addr_storage;
1346		− struct sockaddr listen_addr = (struct sockaddr )&listen_addr_storage,
1347		− src_addr = (struct sockaddr )&src_addr_storage;
	1340	+ struct sockaddr *listen_addr =
	1341	+ (struct sockaddr *)&req->listen_addr_storage;
	1342	+ struct sockaddr src_addr = (struct sockaddr )&req->src_addr_storage;
1348	1343	struct net_device *net_dev;
1349	1344	const union ib_gid *gid = req->has_gid ? &req->local_gid : NULL;
1350	1345	int err;
@@ -1359,11 +1354,6 @@ static struct net_device cma_get_net_dev(struct ib_cm_event ib_event, · 8 unmodified lines
1359	1354	if (!net_dev)
1360	1355	return ERR_PTR(-ENODEV);
1361	1356
1362		− if (!validate_net_dev(net_dev, listen_addr, src_addr)) {
1363		− dev_put(net_dev);
1364		− return ERR_PTR(-EHOSTUNREACH);
1365		− }
1366		−
1367	1357	return net_dev;
1368	1358	}
1369	1359
@@ -1490,15 +1480,51 @@ static struct rdma_id_private cma_id_from_event(struct ib_cm_id cm_id, · 120 unmodified lines
1490	1480	}
1491	1481	}
1492	1482
	1483	+ /*
	1484	+ * Net namespace might be getting deleted while route lookup,
	1485	+ * cm_id lookup is in progress. Therefore, perform netdevice
	1486	+ * validation, cm_id lookup under rcu lock.
	1487	+ * RCU lock along with netdevice state check, synchronizes with
	1488	+ * netdevice migrating to different net namespace and also avoids
	1489	+ * case where net namespace doesn't get deleted while lookup is in
	1490	+ * progress.
	1491	+ * If the device state is not IFF_UP, its properties such as ifindex
	1492	+ * and nd_net cannot be trusted to remain valid without rcu lock.
	1493	+ * net/core/dev.c change_net_namespace() ensures to synchronize with
	1494	+ * ongoing operations on net device after device is closed using
	1495	+ * synchronize_net().
	1496	+ */
	1497	+ rcu_read_lock();
	1498	+ if (*net_dev) {
	1499	+ /*
	1500	+ * If netdevice is down, it is likely that it is administratively
	1501	+ * down or it might be migrating to different namespace.
	1502	+ * In that case avoid further processing, as the net namespace
	1503	+ * or ifindex may change.
	1504	+ */
	1505	+ if (((*net_dev)->flags & IFF_UP) == 0) {
	1506	+ id_priv = ERR_PTR(-EHOSTUNREACH);
	1507	+ goto err;
	1508	+ }
	1509	+
	1510	+ if (!validate_net_dev(*net_dev,
	1511	+ (struct sockaddr *)&req.listen_addr_storage,
	1512	+ (struct sockaddr *)&req.src_addr_storage)) {
	1513	+ id_priv = ERR_PTR(-EHOSTUNREACH);
	1514	+ goto err;
	1515	+ }
	1516	+ }
	1517	+
1493	1518	bind_list = cma_ps_find(net_dev ? dev_net(net_dev) : &init_net,
1494	1519	rdma_ps_from_service_id(req.service_id),
1495	1520	cma_port_from_service_id(req.service_id));
1496	1521	id_priv = cma_find_listener(bind_list, cm_id, ib_event, &req, *net_dev);
	1522	+err:
	1523	+ rcu_read_unlock();
1497	1524	if (IS_ERR(id_priv) && *net_dev) {
1498	1525	dev_put(*net_dev);
1499	1526	*net_dev = NULL;
1500	1527	}
1501		−
1502	1528	return id_priv;
1503	1529	}
1504	1530

drivers/infiniband/core/iwpm_util.c+4−1

@@ -114,7 +114,7 @@ int iwpm_create_mapinfo(struct sockaddr_storage *local_sockaddr, · 113 unmodified lines
114	114	struct sockaddr_storage *mapped_sockaddr,
115	115	u8 nl_client)
116	116	{
117		− struct hlist_head *hash_bucket_head;
	117	+ struct hlist_head *hash_bucket_head = NULL;
118	118	struct iwpm_mapping_info *map_info;
119	119	unsigned long flags;
120	120	int ret = -EINVAL;
@@ -142,6 +142,9 @@ int iwpm_create_mapinfo(struct sockaddr_storage *local_sockaddr, · 21 unmodified lines
142	142	}
143	143	}
144	144	spin_unlock_irqrestore(&iwpm_mapinfo_lock, flags);
	145	+
	146	+ if (!hash_bucket_head)
	147	+ kfree(map_info);
145	148	return ret;
146	149	}
147	150

drivers/infiniband/core/mad.c+2−2

@@ -59,7 +59,7 @@ module_param_named(recv_queue_size, mad_recvq_size, int, 0444); · 58 unmodified lines
59	59	MODULE_PARM_DESC(recv_queue_size, "Size of receive queue in number of work requests");
60	60
61	61	static struct list_head ib_mad_port_list;
62		−static u32 ib_mad_client_id = 0;
	62	+static atomic_t ib_mad_client_id = ATOMIC_INIT(0);
63	63
64	64	/* Port list lock */
65	65	static DEFINE_SPINLOCK(ib_mad_port_list_lock);
@@ -377,7 +377,7 @@ struct ib_mad_agent ib_register_mad_agent(struct ib_device device, · 311 unmodified lines
377	377	}
378	378
379	379	spin_lock_irqsave(&port_priv->reg_lock, flags);
380		− mad_agent_priv->agent.hi_tid = ++ib_mad_client_id;
	380	+ mad_agent_priv->agent.hi_tid = atomic_inc_return(&ib_mad_client_id);
381	381
382	382	/*
383	383	* Make sure MAD registration (if supplied)

drivers/infiniband/core/roce_gid_mgmt.c+14−12

@@ -255,6 +255,7 @@ static void bond_delete_netdev_default_gids(struct ib_device *ib_dev, · 254 unmodified lines
255	255	struct net_device *rdma_ndev)
256	256	{
257	257	struct net_device *real_dev = rdma_vlan_dev_real_dev(event_ndev);
	258	+ unsigned long gid_type_mask;
258	259
259	260	if (!rdma_ndev)
260	261	return;
@@ -264,21 +265,22 @@ static void bond_delete_netdev_default_gids(struct ib_device *ib_dev, · 3 unmodified lines
264	265
265	266	rcu_read_lock();
266	267
267		− if (rdma_is_upper_dev_rcu(rdma_ndev, event_ndev) &&
268		− is_eth_active_slave_of_bonding_rcu(rdma_ndev, real_dev) ==
269		− BONDING_SLAVE_STATE_INACTIVE) {
270		− unsigned long gid_type_mask;
271		−
	268	+ if (((rdma_ndev != event_ndev &&
	269	+ !rdma_is_upper_dev_rcu(rdma_ndev, event_ndev)) \|\|
	270	+ is_eth_active_slave_of_bonding_rcu(rdma_ndev, real_dev)
	271	+ ==
	272	+ BONDING_SLAVE_STATE_INACTIVE)) {
272	273	rcu_read_unlock();
	274	+ return;
	275	+ }
273	276
274		− gid_type_mask = roce_gid_type_mask_support(ib_dev, port);
	277	+ rcu_read_unlock();
275	278
276		− ib_cache_gid_set_default_gid(ib_dev, port, rdma_ndev,
277		− gid_type_mask,
278		− IB_CACHE_GID_DEFAULT_MODE_DELETE);
279		− } else {
280		− rcu_read_unlock();
281		− }
	279	+ gid_type_mask = roce_gid_type_mask_support(ib_dev, port);
	280	+
	281	+ ib_cache_gid_set_default_gid(ib_dev, port, rdma_ndev,
	282	+ gid_type_mask,
	283	+ IB_CACHE_GID_DEFAULT_MODE_DELETE);
282	284	}
283	285
284	286	static void enum_netdev_ipv4_ips(struct ib_device *ib_dev,

drivers/infiniband/core/ucma.c+28−16

@@ -159,6 +159,23 @@ static void ucma_put_ctx(struct ucma_context *ctx) · 158 unmodified lines
159	159	complete(&ctx->comp);
160	160	}
161	161
	162	+/*
	163	+ * Same as ucm_get_ctx but requires that ->cm_id->device is valid, eg that the
	164	+ * CM_ID is bound.
	165	+ */
	166	+static struct ucma_context ucma_get_ctx_dev(struct ucma_file file, int id)
	167	+{
	168	+ struct ucma_context *ctx = ucma_get_ctx(file, id);
	169	+
	170	+ if (IS_ERR(ctx))
	171	+ return ctx;
	172	+ if (!ctx->cm_id->device) {
	173	+ ucma_put_ctx(ctx);
	174	+ return ERR_PTR(-EINVAL);
	175	+ }
	176	+ return ctx;
	177	+}
	178	+
162	179	static void ucma_close_event_id(struct work_struct *work)
163	180	{
164	181	struct ucma_event *uevent_close = container_of(work, struct ucma_event, close_work);
@@ -683,7 +700,7 @@ static ssize_t ucma_resolve_ip(struct ucma_file *file, · 518 unmodified lines
683	700	if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
684	701	return -EFAULT;
685	702
686		− if (!rdma_addr_size_in6(&cmd.src_addr) \|\|
	703	+ if ((cmd.src_addr.sin6_family && !rdma_addr_size_in6(&cmd.src_addr)) \|\|
687	704	!rdma_addr_size_in6(&cmd.dst_addr))
688	705	return -EINVAL;
689	706
@@ -734,7 +751,7 @@ static ssize_t ucma_resolve_route(struct ucma_file *file, · 44 unmodified lines
734	751	if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
735	752	return -EFAULT;
736	753
737		− ctx = ucma_get_ctx(file, cmd.id);
	754	+ ctx = ucma_get_ctx_dev(file, cmd.id);
738	755	if (IS_ERR(ctx))
739	756	return PTR_ERR(ctx);
740	757
@@ -1050,7 +1067,7 @@ static ssize_t ucma_connect(struct ucma_file file, const char __user inbuf, · 309 unmodified lines
1050	1067	if (!cmd.conn_param.valid)
1051	1068	return -EINVAL;
1052	1069
1053		− ctx = ucma_get_ctx(file, cmd.id);
	1070	+ ctx = ucma_get_ctx_dev(file, cmd.id);
1054	1071	if (IS_ERR(ctx))
1055	1072	return PTR_ERR(ctx);
1056	1073
@@ -1092,7 +1109,7 @@ static ssize_t ucma_accept(struct ucma_file file, const char __user inbuf, · 35 unmodified lines
1092	1109	if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
1093	1110	return -EFAULT;
1094	1111
1095		− ctx = ucma_get_ctx(file, cmd.id);
	1112	+ ctx = ucma_get_ctx_dev(file, cmd.id);
1096	1113	if (IS_ERR(ctx))
1097	1114	return PTR_ERR(ctx);
1098	1115
@@ -1120,7 +1137,7 @@ static ssize_t ucma_reject(struct ucma_file file, const char __user inbuf, · 21 unmodified lines
1120	1137	if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
1121	1138	return -EFAULT;
1122	1139
1123		− ctx = ucma_get_ctx(file, cmd.id);
	1140	+ ctx = ucma_get_ctx_dev(file, cmd.id);
1124	1141	if (IS_ERR(ctx))
1125	1142	return PTR_ERR(ctx);
1126	1143
@@ -1139,7 +1156,7 @@ static ssize_t ucma_disconnect(struct ucma_file file, const char __user inbuf, · 12 unmodified lines
1139	1156	if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
1140	1157	return -EFAULT;
1141	1158
1142		− ctx = ucma_get_ctx(file, cmd.id);
	1159	+ ctx = ucma_get_ctx_dev(file, cmd.id);
1143	1160	if (IS_ERR(ctx))
1144	1161	return PTR_ERR(ctx);
1145	1162
@@ -1167,15 +1184,10 @@ static ssize_t ucma_init_qp_attr(struct ucma_file *file, · 21 unmodified lines
1167	1184	if (cmd.qp_state > IB_QPS_ERR)
1168	1185	return -EINVAL;
1169	1186
1170		− ctx = ucma_get_ctx(file, cmd.id);
	1187	+ ctx = ucma_get_ctx_dev(file, cmd.id);
1171	1188	if (IS_ERR(ctx))
1172	1189	return PTR_ERR(ctx);
1173	1190
1174		− if (!ctx->cm_id->device) {
1175		− ret = -EINVAL;
1176		− goto out;
1177		− }
1178		−
1179	1191	resp.qp_attr_mask = 0;
1180	1192	memset(&qp_attr, 0, sizeof qp_attr);
1181	1193	qp_attr.qp_state = cmd.qp_state;
@@ -1316,13 +1328,13 @@ static ssize_t ucma_set_option(struct ucma_file file, const char __user inbuf, · 134 unmodified lines
1316	1328	if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
1317	1329	return -EFAULT;
1318	1330
	1331	+ if (unlikely(cmd.optlen > KMALLOC_MAX_SIZE))
	1332	+ return -EINVAL;
	1333	+
1319	1334	ctx = ucma_get_ctx(file, cmd.id);
1320	1335	if (IS_ERR(ctx))
1321	1336	return PTR_ERR(ctx);
1322	1337
1323		− if (unlikely(cmd.optlen > KMALLOC_MAX_SIZE))
1324		− return -EINVAL;
1325		−
1326	1338	optval = memdup_user(u64_to_user_ptr(cmd.optval),
1327	1339	cmd.optlen);
1328	1340	if (IS_ERR(optval)) {
@@ -1384,7 +1396,7 @@ static ssize_t ucma_process_join(struct ucma_file *file, · 55 unmodified lines
1384	1396	else
1385	1397	return -EINVAL;
1386	1398
1387		− ctx = ucma_get_ctx(file, cmd->id);
	1399	+ ctx = ucma_get_ctx_dev(file, cmd->id);
1388	1400	if (IS_ERR(ctx))
1389	1401	return PTR_ERR(ctx);
1390	1402

drivers/infiniband/core/uverbs_cmd.c+6−0

@@ -691,6 +691,7 @@ ssize_t ib_uverbs_reg_mr(struct ib_uverbs_file *file, · 690 unmodified lines
691	691
692	692	mr->device = pd->device;
693	693	mr->pd = pd;
	694	+ mr->dm = NULL;
694	695	mr->uobject = uobj;
695	696	atomic_inc(&pd->usecnt);
696	697	mr->res.type = RDMA_RESTRACK_MR;
@@ -765,6 +766,11 @@ ssize_t ib_uverbs_rereg_mr(struct ib_uverbs_file *file, · 68 unmodified lines
765	766
766	767	mr = uobj->object;
767	768
	769	+ if (mr->dm) {
	770	+ ret = -EINVAL;
	771	+ goto put_uobjs;
	772	+ }
	773	+
768	774	if (cmd.flags & IB_MR_REREG_ACCESS) {
769	775	ret = ib_check_mr_access(cmd.access_flags);
770	776	if (ret)

drivers/infiniband/core/uverbs_ioctl.c+9−0

@@ -234,6 +234,15 @@ static int uverbs_validate_kernel_mandatory(const struct uverbs_method_spec *met · 233 unmodified lines
234	234	return -EINVAL;
235	235	}
236	236
	237	+ for (; i < method_spec->num_buckets; i++) {
	238	+ struct uverbs_attr_spec_hash *attr_spec_bucket =
	239	+ method_spec->attr_buckets[i];
	240	+
	241	+ if (!bitmap_empty(attr_spec_bucket->mandatory_attrs_bitmask,
	242	+ attr_spec_bucket->num_attrs))
	243	+ return -EINVAL;
	244	+ }
	245	+
237	246	return 0;
238	247	}
239	248

drivers/infiniband/core/uverbs_std_types_flow_action.c+6−6

@@ -363,28 +363,28 @@ static int UVERBS_HANDLER(UVERBS_METHOD_FLOW_ACTION_ESP_MODIFY)(struct ib_device · 362 unmodified lines
363	363
364	364	static const struct uverbs_attr_spec uverbs_flow_action_esp_keymat[] = {
365	365	[IB_UVERBS_FLOW_ACTION_ESP_KEYMAT_AES_GCM] = {
366		− .ptr = {
	366	+ { .ptr = {
367	367	.type = UVERBS_ATTR_TYPE_PTR_IN,
368	368	UVERBS_ATTR_TYPE(struct ib_uverbs_flow_action_esp_keymat_aes_gcm),
369	369	.flags = UVERBS_ATTR_SPEC_F_MIN_SZ_OR_ZERO,
370		− },
	370	+ } },
371	371	},
372	372	};
373	373
374	374	static const struct uverbs_attr_spec uverbs_flow_action_esp_replay[] = {
375	375	[IB_UVERBS_FLOW_ACTION_ESP_REPLAY_NONE] = {
376		− .ptr = {
	376	+ { .ptr = {
377	377	.type = UVERBS_ATTR_TYPE_PTR_IN,
378	378	/* No need to specify any data */
379	379	.len = 0,
380		− }
	380	+ } }
381	381	},
382	382	[IB_UVERBS_FLOW_ACTION_ESP_REPLAY_BMP] = {
383		− .ptr = {
	383	+ { .ptr = {
384	384	.type = UVERBS_ATTR_TYPE_PTR_IN,
385	385	UVERBS_ATTR_STRUCT(struct ib_uverbs_flow_action_esp_replay_bmp, size),
386	386	.flags = UVERBS_ATTR_SPEC_F_MIN_SZ_OR_ZERO,
387		− }
	387	+ } }
388	388	},
389	389	};
390	390

drivers/infiniband/core/verbs.c+1−0

@@ -1656,6 +1656,7 @@ struct ib_mr ib_alloc_mr(struct ib_pd pd, · 1655 unmodified lines
1656	1656	if (!IS_ERR(mr)) {
1657	1657	mr->device = pd->device;
1658	1658	mr->pd = pd;
	1659	+ mr->dm = NULL;
1659	1660	mr->uobject = NULL;
1660	1661	atomic_inc(&pd->usecnt);
1661	1662	mr->need_inval = false;

drivers/infiniband/hw/cxgb4/cq.c+10−1

@@ -315,7 +315,7 @@ static void advance_oldest_read(struct t4_wq *wq) · 314 unmodified lines
315	315	* Deal with out-of-order and/or completions that complete
316	316	* prior unsignalled WRs.
317	317	*/
318		−void c4iw_flush_hw_cq(struct c4iw_cq *chp)
	318	+void c4iw_flush_hw_cq(struct c4iw_cq chp, struct c4iw_qp flush_qhp)
319	319	{
320	320	struct t4_cqe hw_cqe, swcqe, read_cqe;
321	321	struct c4iw_qp *qhp;
@@ -339,6 +339,13 @@ void c4iw_flush_hw_cq(struct c4iw_cq *chp) · 17 unmodified lines
339	339	if (qhp == NULL)
340	340	goto next_cqe;
341	341
	342	+ if (flush_qhp != qhp) {
	343	+ spin_lock(&qhp->lock);
	344	+
	345	+ if (qhp->wq.flushed == 1)
	346	+ goto next_cqe;
	347	+ }
	348	+
342	349	if (CQE_OPCODE(hw_cqe) == FW_RI_TERMINATE)
343	350	goto next_cqe;
344	351
@@ -390,6 +397,8 @@ void c4iw_flush_hw_cq(struct c4iw_cq *chp) · 45 unmodified lines
390	397	next_cqe:
391	398	t4_hwcq_consume(&chp->cq);
392	399	ret = t4_next_hw_cqe(&chp->cq, &hw_cqe);
	400	+ if (qhp && flush_qhp != qhp)
	401	+ spin_unlock(&qhp->lock);
393	402	}
394	403	}
395	404

drivers/infiniband/hw/cxgb4/device.c+8−1

@@ -875,6 +875,11 @@ static int c4iw_rdev_open(struct c4iw_rdev *rdev) · 874 unmodified lines
875	875
876	876	rdev->status_page->db_off = 0;
877	877
	878	+ init_completion(&rdev->rqt_compl);
	879	+ init_completion(&rdev->pbl_compl);
	880	+ kref_init(&rdev->rqt_kref);
	881	+ kref_init(&rdev->pbl_kref);
	882	+
878	883	return 0;
879	884	err_free_status_page_and_wr_log:
880	885	if (c4iw_wr_log && rdev->wr_log)
@@ -893,13 +898,15 @@ destroy_resource: · 12 unmodified lines
893	898
894	899	static void c4iw_rdev_close(struct c4iw_rdev *rdev)
895	900	{
896		− destroy_workqueue(rdev->free_workq);
897	901	kfree(rdev->wr_log);
898	902	c4iw_release_dev_ucontext(rdev, &rdev->uctx);
899	903	free_page((unsigned long)rdev->status_page);
900	904	c4iw_pblpool_destroy(rdev);
901	905	c4iw_rqtpool_destroy(rdev);
	906	+ wait_for_completion(&rdev->pbl_compl);
	907	+ wait_for_completion(&rdev->rqt_compl);
902	908	c4iw_ocqp_pool_destroy(rdev);
	909	+ destroy_workqueue(rdev->free_workq);
903	910	c4iw_destroy_resource(&rdev->resource);
904	911	}
905	912

drivers/infiniband/hw/cxgb4/iw_cxgb4.h+5−1

@@ -185,6 +185,10 @@ struct c4iw_rdev { · 184 unmodified lines
185	185	struct wr_log_entry *wr_log;
186	186	int wr_log_size;
187	187	struct workqueue_struct *free_workq;
	188	+ struct completion rqt_compl;
	189	+ struct completion pbl_compl;
	190	+ struct kref rqt_kref;
	191	+ struct kref pbl_kref;
188	192	};
189	193
190	194	static inline int c4iw_fatal_error(struct c4iw_rdev *rdev)
@@ -1049,7 +1053,7 @@ u32 c4iw_pblpool_alloc(struct c4iw_rdev *rdev, int size); · 858 unmodified lines
1049	1053	void c4iw_pblpool_free(struct c4iw_rdev *rdev, u32 addr, int size);
1050	1054	u32 c4iw_ocqp_pool_alloc(struct c4iw_rdev *rdev, int size);
1051	1055	void c4iw_ocqp_pool_free(struct c4iw_rdev *rdev, u32 addr, int size);
1052		−void c4iw_flush_hw_cq(struct c4iw_cq *chp);
	1056	+void c4iw_flush_hw_cq(struct c4iw_cq chp, struct c4iw_qp flush_qhp);
1053	1057	void c4iw_count_rcqes(struct t4_cq cq, struct t4_wq wq, int *count);
1054	1058	int c4iw_ep_disconnect(struct c4iw_ep *ep, int abrupt, gfp_t gfp);
1055	1059	int c4iw_flush_rq(struct t4_wq wq, struct t4_cq cq, int count);

drivers/infiniband/hw/cxgb4/qp.c+2−2

@@ -1343,12 +1343,12 @@ static void __flush_qp(struct c4iw_qp qhp, struct c4iw_cq rchp, · 1342 unmodified lines
1343	1343	qhp->wq.flushed = 1;
1344	1344	t4_set_wq_in_error(&qhp->wq);
1345	1345
1346		− c4iw_flush_hw_cq(rchp);
	1346	+ c4iw_flush_hw_cq(rchp, qhp);
1347	1347	c4iw_count_rcqes(&rchp->cq, &qhp->wq, &count);
1348	1348	rq_flushed = c4iw_flush_rq(&qhp->wq, &rchp->cq, count);
1349	1349
1350	1350	if (schp != rchp)
1351		− c4iw_flush_hw_cq(schp);
	1351	+ c4iw_flush_hw_cq(schp, qhp);
1352	1352	sq_flushed = c4iw_flush_sq(qhp);
1353	1353
1354	1354	spin_unlock(&qhp->lock);

drivers/infiniband/hw/cxgb4/resource.c+24−2

@@ -260,12 +260,22 @@ u32 c4iw_pblpool_alloc(struct c4iw_rdev *rdev, int size) · 259 unmodified lines
260	260	rdev->stats.pbl.cur += roundup(size, 1 << MIN_PBL_SHIFT);
261	261	if (rdev->stats.pbl.cur > rdev->stats.pbl.max)
262	262	rdev->stats.pbl.max = rdev->stats.pbl.cur;
	263	+ kref_get(&rdev->pbl_kref);
263	264	} else
264	265	rdev->stats.pbl.fail++;
265	266	mutex_unlock(&rdev->stats.lock);
266	267	return (u32)addr;
267	268	}
268	269
	270	+static void destroy_pblpool(struct kref *kref)
	271	+{
	272	+ struct c4iw_rdev *rdev;
	273	+
	274	+ rdev = container_of(kref, struct c4iw_rdev, pbl_kref);
	275	+ gen_pool_destroy(rdev->pbl_pool);
	276	+ complete(&rdev->pbl_compl);
	277	+}
	278	+
269	279	void c4iw_pblpool_free(struct c4iw_rdev *rdev, u32 addr, int size)
270	280	{
271	281	pr_debug("addr 0x%x size %d\n", addr, size);
@@ -273,6 +283,7 @@ void c4iw_pblpool_free(struct c4iw_rdev *rdev, u32 addr, int size) · 1 unmodified line
273	283	rdev->stats.pbl.cur -= roundup(size, 1 << MIN_PBL_SHIFT);
274	284	mutex_unlock(&rdev->stats.lock);
275	285	gen_pool_free(rdev->pbl_pool, (unsigned long)addr, size);
	286	+ kref_put(&rdev->pbl_kref, destroy_pblpool);
276	287	}
277	288
278	289	int c4iw_pblpool_create(struct c4iw_rdev *rdev)
@@ -310,7 +321,7 @@ int c4iw_pblpool_create(struct c4iw_rdev *rdev) · 31 unmodified lines
310	321
311	322	void c4iw_pblpool_destroy(struct c4iw_rdev *rdev)
312	323	{
313		− gen_pool_destroy(rdev->pbl_pool);
	324	+ kref_put(&rdev->pbl_kref, destroy_pblpool);
314	325	}
315	326
316	327	/*
@@ -331,12 +342,22 @@ u32 c4iw_rqtpool_alloc(struct c4iw_rdev *rdev, int size) · 14 unmodified lines
331	342	rdev->stats.rqt.cur += roundup(size << 6, 1 << MIN_RQT_SHIFT);
332	343	if (rdev->stats.rqt.cur > rdev->stats.rqt.max)
333	344	rdev->stats.rqt.max = rdev->stats.rqt.cur;
	345	+ kref_get(&rdev->rqt_kref);
334	346	} else
335	347	rdev->stats.rqt.fail++;
336	348	mutex_unlock(&rdev->stats.lock);
337	349	return (u32)addr;
338	350	}
339	351
	352	+static void destroy_rqtpool(struct kref *kref)
	353	+{
	354	+ struct c4iw_rdev *rdev;
	355	+
	356	+ rdev = container_of(kref, struct c4iw_rdev, rqt_kref);
	357	+ gen_pool_destroy(rdev->rqt_pool);
	358	+ complete(&rdev->rqt_compl);
	359	+}
	360	+
340	361	void c4iw_rqtpool_free(struct c4iw_rdev *rdev, u32 addr, int size)
341	362	{
342	363	pr_debug("addr 0x%x size %d\n", addr, size << 6);
@@ -344,6 +365,7 @@ void c4iw_rqtpool_free(struct c4iw_rdev *rdev, u32 addr, int size) · 1 unmodified line
344	365	rdev->stats.rqt.cur -= roundup(size << 6, 1 << MIN_RQT_SHIFT);
345	366	mutex_unlock(&rdev->stats.lock);
346	367	gen_pool_free(rdev->rqt_pool, (unsigned long)addr, size << 6);
	368	+ kref_put(&rdev->rqt_kref, destroy_rqtpool);
347	369	}
348	370
349	371	int c4iw_rqtpool_create(struct c4iw_rdev *rdev)
@@ -380,7 +402,7 @@ int c4iw_rqtpool_create(struct c4iw_rdev *rdev) · 30 unmodified lines
380	402
381	403	void c4iw_rqtpool_destroy(struct c4iw_rdev *rdev)
382	404	{
383		− gen_pool_destroy(rdev->rqt_pool);
	405	+ kref_put(&rdev->rqt_kref, destroy_rqtpool);
384	406	}
385	407
386	408	/*

drivers/infiniband/hw/hfi1/affinity.c+5−6

@@ -412,7 +412,6 @@ static void hfi1_cleanup_sdma_notifier(struct hfi1_msix_entry *msix) · 411 unmodified lines
412	412	static int get_irq_affinity(struct hfi1_devdata *dd,
413	413	struct hfi1_msix_entry *msix)
414	414	{
415		− int ret;
416	415	cpumask_var_t diff;
417	416	struct hfi1_affinity_node *entry;
418	417	struct cpu_mask_set *set = NULL;
@@ -424,10 +423,6 @@ static int get_irq_affinity(struct hfi1_devdata *dd, · 5 unmodified lines
424	423	extra[0] = '\0';
425	424	cpumask_clear(&msix->mask);
426	425
427		− ret = zalloc_cpumask_var(&diff, GFP_KERNEL);
428		− if (!ret)
429		− return -ENOMEM;
430		−
431	426	entry = node_affinity_lookup(dd->node);
432	427
433	428	switch (msix->type) {
@@ -458,6 +453,9 @@ static int get_irq_affinity(struct hfi1_devdata *dd, · 24 unmodified lines
458	453	* finds its CPU here.
459	454	*/
460	455	if (cpu == -1 && set) {
	456	+ if (!zalloc_cpumask_var(&diff, GFP_KERNEL))
	457	+ return -ENOMEM;
	458	+
461	459	if (cpumask_equal(&set->mask, &set->used)) {
462	460	/*
463	461	* We've used up all the CPUs, bump up the generation
@@ -469,6 +467,8 @@ static int get_irq_affinity(struct hfi1_devdata *dd, · 5 unmodified lines
469	467	cpumask_andnot(diff, &set->mask, &set->used);
470	468	cpu = cpumask_first(diff);
471	469	cpumask_set_cpu(cpu, &set->used);
	470	+
	471	+ free_cpumask_var(diff);
472	472	}
473	473
474	474	cpumask_set_cpu(cpu, &msix->mask);
@@ -482,7 +482,6 @@ static int get_irq_affinity(struct hfi1_devdata *dd, · 7 unmodified lines
482	482	hfi1_setup_sdma_notifier(msix);
483	483	}
484	484
485		− free_cpumask_var(diff);
486	485	return 0;
487	486	}
488	487

drivers/infiniband/hw/hfi1/driver.c+15−4

@@ -433,31 +433,43 @@ void hfi1_process_ecn_slowpath(struct rvt_qp qp, struct hfi1_packet pkt, · 432 unmodified lines
433	433	bool do_cnp)
434	434	{
435	435	struct hfi1_ibport *ibp = to_iport(qp->ibqp.device, qp->port_num);
	436	+ struct hfi1_pportdata *ppd = ppd_from_ibp(ibp);
436	437	struct ib_other_headers *ohdr = pkt->ohdr;
437	438	struct ib_grh *grh = pkt->grh;
438	439	u32 rqpn = 0, bth1;
439		− u16 pkey, rlid, dlid = ib_get_dlid(pkt->hdr);
	440	+ u16 pkey;
	441	+ u32 rlid, slid, dlid = 0;
440	442	u8 hdr_type, sc, svc_type;
441	443	bool is_mcast = false;
442	444
	445	+ /* can be called from prescan */
443	446	if (pkt->etype == RHF_RCV_TYPE_BYPASS) {
444	447	is_mcast = hfi1_is_16B_mcast(dlid);
445	448	pkey = hfi1_16B_get_pkey(pkt->hdr);
446	449	sc = hfi1_16B_get_sc(pkt->hdr);
	450	+ dlid = hfi1_16B_get_dlid(pkt->hdr);
	451	+ slid = hfi1_16B_get_slid(pkt->hdr);
447	452	hdr_type = HFI1_PKT_TYPE_16B;
448	453	} else {
449	454	is_mcast = (dlid > be16_to_cpu(IB_MULTICAST_LID_BASE)) &&
450	455	(dlid != be16_to_cpu(IB_LID_PERMISSIVE));
451	456	pkey = ib_bth_get_pkey(ohdr);
452	457	sc = hfi1_9B_get_sc5(pkt->hdr, pkt->rhf);
	458	+ dlid = ib_get_dlid(pkt->hdr);
	459	+ slid = ib_get_slid(pkt->hdr);
453	460	hdr_type = HFI1_PKT_TYPE_9B;
454	461	}
455	462
456	463	switch (qp->ibqp.qp_type) {
	464	+ case IB_QPT_UD:
	465	+ dlid = ppd->lid;
	466	+ rlid = slid;
	467	+ rqpn = ib_get_sqpn(pkt->ohdr);
	468	+ svc_type = IB_CC_SVCTYPE_UD;
	469	+ break;
457	470	case IB_QPT_SMI:
458	471	case IB_QPT_GSI:
459		− case IB_QPT_UD:
460		− rlid = ib_get_slid(pkt->hdr);
	472	+ rlid = slid;
461	473	rqpn = ib_get_sqpn(pkt->ohdr);
462	474	svc_type = IB_CC_SVCTYPE_UD;
463	475	break;
@@ -482,7 +494,6 @@ void hfi1_process_ecn_slowpath(struct rvt_qp qp, struct hfi1_packet pkt, · 18 unmodified lines
482	494	dlid, rlid, sc, grh);
483	495
484	496	if (!is_mcast && (bth1 & IB_BECN_SMASK)) {
485		− struct hfi1_pportdata *ppd = ppd_from_ibp(ibp);
486	497	u32 lqpn = bth1 & RVT_QPN_MASK;
487	498	u8 sl = ibp->sc_to_sl[sc];
488	499

drivers/infiniband/hw/hfi1/hfi.h+4−4

@@ -1537,13 +1537,13 @@ void set_link_ipg(struct hfi1_pportdata *ppd); · 1536 unmodified lines
1537	1537	void process_becn(struct hfi1_pportdata *ppd, u8 sl, u32 rlid, u32 lqpn,
1538	1538	u32 rqpn, u8 svc_type);
1539	1539	void return_cnp(struct hfi1_ibport ibp, struct rvt_qp qp, u32 remote_qpn,
1540		− u32 pkey, u32 slid, u32 dlid, u8 sc5,
	1540	+ u16 pkey, u32 slid, u32 dlid, u8 sc5,
1541	1541	const struct ib_grh *old_grh);
1542	1542	void return_cnp_16B(struct hfi1_ibport ibp, struct rvt_qp qp,
1543		− u32 remote_qpn, u32 pkey, u32 slid, u32 dlid,
	1543	+ u32 remote_qpn, u16 pkey, u32 slid, u32 dlid,
1544	1544	u8 sc5, const struct ib_grh *old_grh);
1545	1545	typedef void (hfi1_handle_cnp)(struct hfi1_ibport ibp, struct rvt_qp *qp,
1546		− u32 remote_qpn, u32 pkey, u32 slid, u32 dlid,
	1546	+ u32 remote_qpn, u16 pkey, u32 slid, u32 dlid,
1547	1547	u8 sc5, const struct ib_grh *old_grh);
1548	1548
1549	1549	#define PKEY_CHECK_INVALID -1
@@ -2437,7 +2437,7 @@ static inline void hfi1_make_16b_hdr(struct hfi1_16b_header *hdr, · 887 unmodified lines
2437	2437	((slid >> OPA_16B_SLID_SHIFT) << OPA_16B_SLID_HIGH_SHIFT);
2438	2438	lrh2 = (lrh2 & ~OPA_16B_DLID_MASK) \|
2439	2439	((dlid >> OPA_16B_DLID_SHIFT) << OPA_16B_DLID_HIGH_SHIFT);
2440		− lrh2 = (lrh2 & ~OPA_16B_PKEY_MASK) \| (pkey << OPA_16B_PKEY_SHIFT);
	2440	+ lrh2 = (lrh2 & ~OPA_16B_PKEY_MASK) \| ((u32)pkey << OPA_16B_PKEY_SHIFT);
2441	2441	lrh2 = (lrh2 & ~OPA_16B_L4_MASK) \| l4;
2442	2442
2443	2443	hdr->lrh[0] = lrh0;

drivers/infiniband/hw/hfi1/init.c+31−12

@@ -88,9 +88,9 @@ · 87 unmodified lines
88	88	* pio buffers per ctxt, etc.) Zero means use one user context per CPU.
89	89	*/
90	90	int num_user_contexts = -1;
91		−module_param_named(num_user_contexts, num_user_contexts, uint, S_IRUGO);
	91	+module_param_named(num_user_contexts, num_user_contexts, int, 0444);
92	92	MODULE_PARM_DESC(
93		− num_user_contexts, "Set max number of user contexts to use");
	93	+ num_user_contexts, "Set max number of user contexts to use (default: -1 will use the real (non-HT) CPU count)");
94	94
95	95	uint krcvqs[RXE_NUM_DATA_VL];
96	96	int krcvqsset;
@@ -1209,30 +1209,49 @@ static void finalize_asic_data(struct hfi1_devdata *dd, · 1112 unmodified lines
1209	1209	kfree(ad);
1210	1210	}
1211	1211
1212		−static void __hfi1_free_devdata(struct kobject *kobj)
	1212	+/**
	1213	+ * hfi1_clean_devdata - cleans up per-unit data structure
	1214	+ * @dd: pointer to a valid devdata structure
	1215	+ *
	1216	+ * It cleans up all data structures set up by
	1217	+ * by hfi1_alloc_devdata().
	1218	+ */
	1219	+static void hfi1_clean_devdata(struct hfi1_devdata *dd)
1213	1220	{
1214		− struct hfi1_devdata *dd =
1215		− container_of(kobj, struct hfi1_devdata, kobj);
1216	1221	struct hfi1_asic_data *ad;
1217	1222	unsigned long flags;
1218	1223
1219	1224	spin_lock_irqsave(&hfi1_devs_lock, flags);
1220		− idr_remove(&hfi1_unit_table, dd->unit);
1221		− list_del(&dd->list);
	1225	+ if (!list_empty(&dd->list)) {
	1226	+ idr_remove(&hfi1_unit_table, dd->unit);
	1227	+ list_del_init(&dd->list);
	1228	+ }
1222	1229	ad = release_asic_data(dd);
1223	1230	spin_unlock_irqrestore(&hfi1_devs_lock, flags);
1224		− if (ad)
1225		− finalize_asic_data(dd, ad);
	1231	+
	1232	+ finalize_asic_data(dd, ad);
1226	1233	free_platform_config(dd);
1227	1234	rcu_barrier(); /* wait for rcu callbacks to complete */
1228	1235	free_percpu(dd->int_counter);
1229	1236	free_percpu(dd->rcv_limit);
1230	1237	free_percpu(dd->send_schedule);
1231	1238	free_percpu(dd->tx_opstats);
	1239	+ dd->int_counter = NULL;
	1240	+ dd->rcv_limit = NULL;
	1241	+ dd->send_schedule = NULL;
	1242	+ dd->tx_opstats = NULL;
1232	1243	sdma_clean(dd, dd->num_sdma);
1233	1244	rvt_dealloc_device(&dd->verbs_dev.rdi);
1234	1245	}
1235	1246
	1247	+static void __hfi1_free_devdata(struct kobject *kobj)
	1248	+{
	1249	+ struct hfi1_devdata *dd =
	1250	+ container_of(kobj, struct hfi1_devdata, kobj);
	1251	+
	1252	+ hfi1_clean_devdata(dd);
	1253	+}
	1254	+
1236	1255	static struct kobj_type hfi1_devdata_type = {
1237	1256	.release = __hfi1_free_devdata,
1238	1257	};
@@ -1265,6 +1284,8 @@ struct hfi1_devdata hfi1_alloc_devdata(struct pci_dev pdev, size_t extra) · 26 unmodified lines
1265	1284	return ERR_PTR(-ENOMEM);
1266	1285	dd->num_pports = nports;
1267	1286	dd->pport = (struct hfi1_pportdata *)(dd + 1);
	1287	+ dd->pcidev = pdev;
	1288	+ pci_set_drvdata(pdev, dd);
1268	1289
1269	1290	INIT_LIST_HEAD(&dd->list);
1270	1291	idr_preload(GFP_KERNEL);
@@ -1331,9 +1352,7 @@ struct hfi1_devdata hfi1_alloc_devdata(struct pci_dev pdev, size_t extra) · 60 unmodified lines
1331	1352	return dd;
1332	1353
1333	1354	bail:
1334		− if (!list_empty(&dd->list))
1335		− list_del_init(&dd->list);
1336		− rvt_dealloc_device(&dd->verbs_dev.rdi);
	1355	+ hfi1_clean_devdata(dd);
1337	1356	return ERR_PTR(ret);
1338	1357	}
1339	1358

drivers/infiniband/hw/hfi1/pcie.c+0−3

@@ -163,9 +163,6 @@ int hfi1_pcie_ddinit(struct hfi1_devdata dd, struct pci_dev pdev) · 162 unmodified lines
163	163	resource_size_t addr;
164	164	int ret = 0;
165	165
166		− dd->pcidev = pdev;
167		− pci_set_drvdata(pdev, dd);
168		−
169	166	addr = pci_resource_start(pdev, 0);
170	167	len = pci_resource_len(pdev, 0);
171	168

drivers/infiniband/hw/hfi1/platform.c+1−0

@@ -199,6 +199,7 @@ void free_platform_config(struct hfi1_devdata *dd) · 198 unmodified lines
199	199	{
200	200	/* Release memory allocated for eprom or fallback file read. */
201	201	kfree(dd->platform_config.data);
	202	+ dd->platform_config.data = NULL;
202	203	}
203	204
204	205	void get_port_type(struct hfi1_pportdata *ppd)

drivers/infiniband/hw/hfi1/qsfp.c+2−0

@@ -204,6 +204,8 @@ static void clean_i2c_bus(struct hfi1_i2c_bus *bus) · 203 unmodified lines
204	204
205	205	void clean_up_i2c(struct hfi1_devdata dd, struct hfi1_asic_data ad)
206	206	{
	207	+ if (!ad)
	208	+ return;
207	209	clean_i2c_bus(ad->i2c_bus0);
208	210	ad->i2c_bus0 = NULL;
209	211	clean_i2c_bus(ad->i2c_bus1);

drivers/infiniband/hw/hfi1/ruc.c+40−10

@@ -733,6 +733,20 @@ static inline void hfi1_make_ruc_bth(struct rvt_qp *qp, · 732 unmodified lines
733	733	ohdr->bth[2] = cpu_to_be32(bth2);
734	734	}
735	735
	736	+/**
	737	+ * hfi1_make_ruc_header_16B - build a 16B header
	738	+ * @qp: the queue pair
	739	+ * @ohdr: a pointer to the destination header memory
	740	+ * @bth0: bth0 passed in from the RC/UC builder
	741	+ * @bth2: bth2 passed in from the RC/UC builder
	742	+ * @middle: non zero implies indicates ahg "could" be used
	743	+ * @ps: the current packet state
	744	+ *
	745	+ * This routine may disarm ahg under these situations:
	746	+ * - packet needs a GRH
	747	+ * - BECN needed
	748	+ * - migration state not IB_MIG_MIGRATED
	749	+ */
736	750	static inline void hfi1_make_ruc_header_16B(struct rvt_qp *qp,
737	751	struct ib_other_headers *ohdr,
738	752	u32 bth0, u32 bth2, int middle,
@@ -777,6 +791,12 @@ static inline void hfi1_make_ruc_header_16B(struct rvt_qp *qp, · 38 unmodified lines
777	791	else
778	792	middle = 0;
779	793
	794	+ if (qp->s_flags & RVT_S_ECN) {
	795	+ qp->s_flags &= ~RVT_S_ECN;
	796	+ /* we recently received a FECN, so return a BECN */
	797	+ becn = true;
	798	+ middle = 0;
	799	+ }
780	800	if (middle)
781	801	build_ahg(qp, bth2);
782	802	else
@@ -784,11 +804,6 @@ static inline void hfi1_make_ruc_header_16B(struct rvt_qp *qp, · 1 unmodified line
784	804
785	805	bth0 \|= pkey;
786	806	bth0 \|= extra_bytes << 20;
787		− if (qp->s_flags & RVT_S_ECN) {
788		− qp->s_flags &= ~RVT_S_ECN;
789		− /* we recently received a FECN, so return a BECN */
790		− becn = true;
791		− }
792	807	hfi1_make_ruc_bth(qp, ohdr, bth0, bth1, bth2);
793	808
794	809	if (!ppd->lid)
@@ -806,6 +821,20 @@ static inline void hfi1_make_ruc_header_16B(struct rvt_qp *qp, · 11 unmodified lines
806	821	pkey, becn, 0, l4, priv->s_sc);
807	822	}
808	823
	824	+/**
	825	+ * hfi1_make_ruc_header_9B - build a 9B header
	826	+ * @qp: the queue pair
	827	+ * @ohdr: a pointer to the destination header memory
	828	+ * @bth0: bth0 passed in from the RC/UC builder
	829	+ * @bth2: bth2 passed in from the RC/UC builder
	830	+ * @middle: non zero implies indicates ahg "could" be used
	831	+ * @ps: the current packet state
	832	+ *
	833	+ * This routine may disarm ahg under these situations:
	834	+ * - packet needs a GRH
	835	+ * - BECN needed
	836	+ * - migration state not IB_MIG_MIGRATED
	837	+ */
809	838	static inline void hfi1_make_ruc_header_9B(struct rvt_qp *qp,
810	839	struct ib_other_headers *ohdr,
811	840	u32 bth0, u32 bth2, int middle,
@@ -839,6 +868,12 @@ static inline void hfi1_make_ruc_header_9B(struct rvt_qp *qp, · 27 unmodified lines
839	868	else
840	869	middle = 0;
841	870
	871	+ if (qp->s_flags & RVT_S_ECN) {
	872	+ qp->s_flags &= ~RVT_S_ECN;
	873	+ /* we recently received a FECN, so return a BECN */
	874	+ bth1 \|= (IB_BECN_MASK << IB_BECN_SHIFT);
	875	+ middle = 0;
	876	+ }
842	877	if (middle)
843	878	build_ahg(qp, bth2);
844	879	else
@@ -846,11 +881,6 @@ static inline void hfi1_make_ruc_header_9B(struct rvt_qp *qp, · 1 unmodified line
846	881
847	882	bth0 \|= pkey;
848	883	bth0 \|= extra_bytes << 20;
849		− if (qp->s_flags & RVT_S_ECN) {
850		− qp->s_flags &= ~RVT_S_ECN;
851		− /* we recently received a FECN, so return a BECN */
852		− bth1 \|= (IB_BECN_MASK << IB_BECN_SHIFT);
853		− }
854	884	hfi1_make_ruc_bth(qp, ohdr, bth0, bth1, bth2);
855	885	hfi1_make_ib_hdr(&ps->s_txreq->phdr.hdr.ibh,
856	886	lrh0,

drivers/infiniband/hw/hfi1/ud.c+2−2

@@ -628,7 +628,7 @@ int hfi1_lookup_pkey_idx(struct hfi1_ibport *ibp, u16 pkey) · 627 unmodified lines
628	628	}
629	629
630	630	void return_cnp_16B(struct hfi1_ibport ibp, struct rvt_qp qp,
631		− u32 remote_qpn, u32 pkey, u32 slid, u32 dlid,
	631	+ u32 remote_qpn, u16 pkey, u32 slid, u32 dlid,
632	632	u8 sc5, const struct ib_grh *old_grh)
633	633	{
634	634	u64 pbc, pbc_flags = 0;
@@ -687,7 +687,7 @@ void return_cnp_16B(struct hfi1_ibport ibp, struct rvt_qp qp, · 52 unmodified lines
687	687	}
688	688
689	689	void return_cnp(struct hfi1_ibport ibp, struct rvt_qp qp, u32 remote_qpn,
690		− u32 pkey, u32 slid, u32 dlid, u8 sc5,
	690	+ u16 pkey, u32 slid, u32 dlid, u8 sc5,
691	691	const struct ib_grh *old_grh)
692	692	{
693	693	u64 pbc, pbc_flags = 0;

drivers/infiniband/hw/hns/hns_roce_hem.c+6−6

@@ -912,7 +912,7 @@ int hns_roce_init_hem_table(struct hns_roce_dev *hr_dev, · 911 unmodified lines
912	912	obj_per_chunk = buf_chunk_size / obj_size;
913	913	num_hem = (nobj + obj_per_chunk - 1) / obj_per_chunk;
914	914	bt_chunk_num = bt_chunk_size / 8;
915		− if (table->type >= HEM_TYPE_MTT)
	915	+ if (type >= HEM_TYPE_MTT)
916	916	num_bt_l0 = bt_chunk_num;
917	917
918	918	table->hem = kcalloc(num_hem, sizeof(*table->hem),
@@ -920,7 +920,7 @@ int hns_roce_init_hem_table(struct hns_roce_dev *hr_dev, · 1 unmodified line
920	920	if (!table->hem)
921	921	goto err_kcalloc_hem_buf;
922	922
923		− if (check_whether_bt_num_3(table->type, hop_num)) {
	923	+ if (check_whether_bt_num_3(type, hop_num)) {
924	924	unsigned long num_bt_l1;
925	925
926	926	num_bt_l1 = (num_hem + bt_chunk_num - 1) /
@@ -939,8 +939,8 @@ int hns_roce_init_hem_table(struct hns_roce_dev *hr_dev, · 12 unmodified lines
939	939	goto err_kcalloc_l1_dma;
940	940	}
941	941
942		− if (check_whether_bt_num_2(table->type, hop_num) \|\|
943		− check_whether_bt_num_3(table->type, hop_num)) {
	942	+ if (check_whether_bt_num_2(type, hop_num) \|\|
	943	+ check_whether_bt_num_3(type, hop_num)) {
944	944	table->bt_l0 = kcalloc(num_bt_l0, sizeof(*table->bt_l0),
945	945	GFP_KERNEL);
946	946	if (!table->bt_l0)
@@ -1039,14 +1039,14 @@ void hns_roce_cleanup_hem_table(struct hns_roce_dev *hr_dev, · 92 unmodified lines
1039	1039	void hns_roce_cleanup_hem(struct hns_roce_dev *hr_dev)
1040	1040	{
1041	1041	hns_roce_cleanup_hem_table(hr_dev, &hr_dev->cq_table.table);
1042		− hns_roce_cleanup_hem_table(hr_dev, &hr_dev->qp_table.irrl_table);
1043	1042	if (hr_dev->caps.trrl_entry_sz)
1044	1043	hns_roce_cleanup_hem_table(hr_dev,
1045	1044	&hr_dev->qp_table.trrl_table);
	1045	+ hns_roce_cleanup_hem_table(hr_dev, &hr_dev->qp_table.irrl_table);
1046	1046	hns_roce_cleanup_hem_table(hr_dev, &hr_dev->qp_table.qp_table);
1047	1047	hns_roce_cleanup_hem_table(hr_dev, &hr_dev->mr_table.mtpt_table);
1048		− hns_roce_cleanup_hem_table(hr_dev, &hr_dev->mr_table.mtt_table);
1049	1048	if (hns_roce_check_whether_mhop(hr_dev, HEM_TYPE_CQE))
1050	1049	hns_roce_cleanup_hem_table(hr_dev,
1051	1050	&hr_dev->mr_table.mtt_cqe_table);
	1051	+ hns_roce_cleanup_hem_table(hr_dev, &hr_dev->mr_table.mtt_table);
1052	1052	}

drivers/infiniband/hw/hns/hns_roce_hw_v2.c+29−20

@@ -71,6 +71,11 @@ static int set_rwqe_data_seg(struct ib_qp ibqp, struct ib_send_wr wr, · 70 unmodified lines
71	71	return -EINVAL;
72	72	}
73	73
	74	+ if (wr->opcode == IB_WR_RDMA_READ) {
	75	+ dev_err(hr_dev->dev, "Not support inline data!\n");
	76	+ return -EINVAL;
	77	+ }
	78	+
74	79	for (i = 0; i < wr->num_sge; i++) {
75	80	memcpy(wqe, ((void *)wr->sg_list[i].addr),
76	81	wr->sg_list[i].length);
@@ -148,7 +153,7 @@ static int hns_roce_v2_post_send(struct ib_qp ibqp, struct ib_send_wr wr, · 71 unmodified lines
148	153	ibqp->qp_type != IB_QPT_GSI &&
149	154	ibqp->qp_type != IB_QPT_UD)) {
150	155	dev_err(dev, "Not supported QP(0x%x)type!\n", ibqp->qp_type);
151		− *bad_wr = NULL;
	156	+ *bad_wr = wr;
152	157	return -EOPNOTSUPP;
153	158	}
154	159
@@ -182,7 +187,8 @@ static int hns_roce_v2_post_send(struct ib_qp ibqp, struct ib_send_wr wr, · 27 unmodified lines
182	187	qp->sq.wrid[(qp->sq.head + nreq) & (qp->sq.wqe_cnt - 1)] =
183	188	wr->wr_id;
184	189
185		− owner_bit = ~(qp->sq.head >> ilog2(qp->sq.wqe_cnt)) & 0x1;
	190	+ owner_bit =
	191	+ ~(((qp->sq.head + nreq) >> ilog2(qp->sq.wqe_cnt)) & 0x1);
186	192
187	193	/* Corresponding to the QP type, wqe process separately */
188	194	if (ibqp->qp_type == IB_QPT_GSI) {
@@ -456,6 +462,7 @@ static int hns_roce_v2_post_send(struct ib_qp ibqp, struct ib_send_wr wr, · 267 unmodified lines
456	462	} else {
457	463	dev_err(dev, "Illegal qp_type(0x%x)\n", ibqp->qp_type);
458	464	spin_unlock_irqrestore(&qp->sq.lock, flags);
	465	+ *bad_wr = wr;
459	466	return -EOPNOTSUPP;
460	467	}
461	468	}
@@ -2592,10 +2599,12 @@ static void modify_qp_init_to_init(struct ib_qp *ibqp, · 2130 unmodified lines
2592	2599	roce_set_field(qpc_mask->byte_4_sqpn_tst, V2_QPC_BYTE_4_SQPN_M,
2593	2600	V2_QPC_BYTE_4_SQPN_S, 0);
2594	2601
2595		− roce_set_field(context->byte_56_dqpn_err, V2_QPC_BYTE_56_DQPN_M,
2596		− V2_QPC_BYTE_56_DQPN_S, hr_qp->qpn);
2597		− roce_set_field(qpc_mask->byte_56_dqpn_err, V2_QPC_BYTE_56_DQPN_M,
2598		− V2_QPC_BYTE_56_DQPN_S, 0);
	2602	+ if (attr_mask & IB_QP_DEST_QPN) {
	2603	+ roce_set_field(context->byte_56_dqpn_err, V2_QPC_BYTE_56_DQPN_M,
	2604	+ V2_QPC_BYTE_56_DQPN_S, hr_qp->qpn);
	2605	+ roce_set_field(qpc_mask->byte_56_dqpn_err,
	2606	+ V2_QPC_BYTE_56_DQPN_M, V2_QPC_BYTE_56_DQPN_S, 0);
	2607	+ }
2599	2608	roce_set_field(context->byte_168_irrl_idx,
2600	2609	V2_QPC_BYTE_168_SQ_SHIFT_BAK_M,
2601	2610	V2_QPC_BYTE_168_SQ_SHIFT_BAK_S,
@@ -2650,8 +2659,7 @@ static int modify_qp_init_to_rtr(struct ib_qp *ibqp, · 48 unmodified lines
2650	2659	return -EINVAL;
2651	2660	}
2652	2661
2653		− if ((attr_mask & IB_QP_ALT_PATH) \|\| (attr_mask & IB_QP_ACCESS_FLAGS) \|\|
2654		− (attr_mask & IB_QP_PKEY_INDEX) \|\| (attr_mask & IB_QP_QKEY)) {
	2662	+ if (attr_mask & IB_QP_ALT_PATH) {
2655	2663	dev_err(dev, "INIT2RTR attr_mask (0x%x) error\n", attr_mask);
2656	2664	return -EINVAL;
2657	2665	}
@@ -2800,10 +2808,12 @@ static int modify_qp_init_to_rtr(struct ib_qp *ibqp, · 142 unmodified lines
2800	2808	V2_QPC_BYTE_140_RR_MAX_S, 0);
2801	2809	}
2802	2810
2803		− roce_set_field(context->byte_56_dqpn_err, V2_QPC_BYTE_56_DQPN_M,
2804		− V2_QPC_BYTE_56_DQPN_S, attr->dest_qp_num);
2805		− roce_set_field(qpc_mask->byte_56_dqpn_err, V2_QPC_BYTE_56_DQPN_M,
2806		− V2_QPC_BYTE_56_DQPN_S, 0);
	2811	+ if (attr_mask & IB_QP_DEST_QPN) {
	2812	+ roce_set_field(context->byte_56_dqpn_err, V2_QPC_BYTE_56_DQPN_M,
	2813	+ V2_QPC_BYTE_56_DQPN_S, attr->dest_qp_num);
	2814	+ roce_set_field(qpc_mask->byte_56_dqpn_err,
	2815	+ V2_QPC_BYTE_56_DQPN_M, V2_QPC_BYTE_56_DQPN_S, 0);
	2816	+ }
2807	2817
2808	2818	/* Configure GID index */
2809	2819	port_num = rdma_ah_get_port_num(&attr->ah_attr);
@@ -2845,7 +2855,7 @@ static int modify_qp_init_to_rtr(struct ib_qp *ibqp, · 35 unmodified lines
2845	2855	if (ibqp->qp_type == IB_QPT_GSI \|\| ibqp->qp_type == IB_QPT_UD)
2846	2856	roce_set_field(context->byte_24_mtu_tc, V2_QPC_BYTE_24_MTU_M,
2847	2857	V2_QPC_BYTE_24_MTU_S, IB_MTU_4096);
2848		− else
	2858	+ else if (attr_mask & IB_QP_PATH_MTU)
2849	2859	roce_set_field(context->byte_24_mtu_tc, V2_QPC_BYTE_24_MTU_M,
2850	2860	V2_QPC_BYTE_24_MTU_S, attr->path_mtu);
2851	2861
@@ -2922,11 +2932,9 @@ static int modify_qp_rtr_to_rts(struct ib_qp *ibqp, · 70 unmodified lines
2922	2932	return -EINVAL;
2923	2933	}
2924	2934
2925		− /* If exist optional param, return error */
2926		− if ((attr_mask & IB_QP_ALT_PATH) \|\| (attr_mask & IB_QP_ACCESS_FLAGS) \|\|
2927		− (attr_mask & IB_QP_QKEY) \|\| (attr_mask & IB_QP_PATH_MIG_STATE) \|\|
2928		− (attr_mask & IB_QP_CUR_STATE) \|\|
2929		− (attr_mask & IB_QP_MIN_RNR_TIMER)) {
	2935	+ /* Not support alternate path and path migration */
	2936	+ if ((attr_mask & IB_QP_ALT_PATH) \|\|
	2937	+ (attr_mask & IB_QP_PATH_MIG_STATE)) {
2930	2938	dev_err(dev, "RTR2RTS attr_mask (0x%x)error\n", attr_mask);
2931	2939	return -EINVAL;
2932	2940	}
@@ -3161,7 +3169,8 @@ static int hns_roce_v2_modify_qp(struct ib_qp *ibqp, · 228 unmodified lines
3161	3169	(cur_state == IB_QPS_RTR && new_state == IB_QPS_ERR) \|\|
3162	3170	(cur_state == IB_QPS_RTS && new_state == IB_QPS_ERR) \|\|
3163	3171	(cur_state == IB_QPS_SQD && new_state == IB_QPS_ERR) \|\|
3164		− (cur_state == IB_QPS_SQE && new_state == IB_QPS_ERR)) {
	3172	+ (cur_state == IB_QPS_SQE && new_state == IB_QPS_ERR) \|\|
	3173	+ (cur_state == IB_QPS_ERR && new_state == IB_QPS_ERR)) {
3165	3174	/* Nothing */
3166	3175	;
3167	3176	} else {
@@ -4478,7 +4487,7 @@ static int hns_roce_v2_create_eq(struct hns_roce_dev *hr_dev, · 1310 unmodified lines
4478	4487	ret = hns_roce_cmd_mbox(hr_dev, mailbox->dma, 0, eq->eqn, 0,
4479	4488	eq_cmd, HNS_ROCE_CMD_TIMEOUT_MSECS);
4480	4489	if (ret) {
4481		− dev_err(dev, "[mailbox cmd] creat eqc failed.\n");
	4490	+ dev_err(dev, "[mailbox cmd] create eqc failed.\n");
4482	4491	goto err_cmd_mbox;
4483	4492	}
4484	4493

drivers/infiniband/hw/hns/hns_roce_qp.c+1−1

@@ -620,7 +620,7 @@ static int hns_roce_create_qp_common(struct hns_roce_dev *hr_dev, · 619 unmodified lines
620	620	to_hr_ucontext(ib_pd->uobject->context),
621	621	ucmd.db_addr, &hr_qp->rdb);
622	622	if (ret) {
623		− dev_err(dev, "rp record doorbell map failed!\n");
	623	+ dev_err(dev, "rq record doorbell map failed!\n");
624	624	goto err_mtt;
625	625	}
626	626	}

drivers/infiniband/hw/mlx4/mr.c+1−1

@@ -346,7 +346,7 @@ int mlx4_ib_umem_calc_optimal_mtt_size(struct ib_umem *umem, u64 start_va, · 345 unmodified lines
346	346	/* Add to the first block the misalignment that it suffers from. */
347	347	total_len += (first_block_start & ((1ULL << block_shift) - 1ULL));
348	348	last_block_end = current_block_start + current_block_len;
349		− last_block_aligned_end = round_up(last_block_end, 1 << block_shift);
	349	+ last_block_aligned_end = round_up(last_block_end, 1ULL << block_shift);
350	350	total_len += (last_block_aligned_end - last_block_end);
351	351
352	352	if (total_len & ((1ULL << block_shift) - 1ULL))

drivers/infiniband/hw/mlx4/qp.c+2−1

@@ -673,7 +673,8 @@ static int set_qp_rss(struct mlx4_ib_dev dev, struct mlx4_ib_rss rss_ctx, · 672 unmodified lines
673	673	MLX4_IB_RX_HASH_SRC_PORT_TCP \|
674	674	MLX4_IB_RX_HASH_DST_PORT_TCP \|
675	675	MLX4_IB_RX_HASH_SRC_PORT_UDP \|
676		− MLX4_IB_RX_HASH_DST_PORT_UDP)) {
	676	+ MLX4_IB_RX_HASH_DST_PORT_UDP \|
	677	+ MLX4_IB_RX_HASH_INNER)) {
677	678	pr_debug("RX Hash fields_mask has unsupported mask (0x%llx)\n",
678	679	ucmd->rx_hash_fields_mask);
679	680	return (-EOPNOTSUPP);

drivers/infiniband/hw/mlx5/Kconfig+1−0

@@ -1,6 +1,7 @@
1	1	config MLX5_INFINIBAND
2	2	tristate "Mellanox Connect-IB HCA support"
3	3	depends on NETDEVICES && ETHERNET && PCI && MLX5_CORE
	4	+ depends on INFINIBAND_USER_ACCESS \|\| INFINIBAND_USER_ACCESS=n
4	5	---help---
5	6	This driver provides low-level InfiniBand support for
6	7	Mellanox Connect-IB PCI Express host channel adapters (HCAs).

drivers/infiniband/hw/mlx5/main.c+3−6

@@ -52,7 +52,6 @@ · 51 unmodified lines
52	52	#include <linux/mlx5/port.h>
53	53	#include <linux/mlx5/vport.h>
54	54	#include <linux/mlx5/fs.h>
55		−#include <linux/mlx5/fs_helpers.h>
56	55	#include <linux/list.h>
57	56	#include <rdma/ib_smi.h>
58	57	#include <rdma/ib_umem.h>
@@ -180,7 +179,7 @@ static int mlx5_netdev_event(struct notifier_block *this, · 121 unmodified lines
180	179	if (rep_ndev == ndev)
181	180	roce->netdev = (event == NETDEV_UNREGISTER) ?
182	181	NULL : ndev;
183		− } else if (ndev->dev.parent == &ibdev->mdev->pdev->dev) {
	182	+ } else if (ndev->dev.parent == &mdev->pdev->dev) {
184	183	roce->netdev = (event == NETDEV_UNREGISTER) ?
185	184	NULL : ndev;
186	185	}
@@ -4757,7 +4756,7 @@ mlx5_ib_get_vector_affinity(struct ib_device *ibdev, int comp_vector) · 4570 unmodified lines
4757	4756	{
4758	4757	struct mlx5_ib_dev *dev = to_mdev(ibdev);
4759	4758
4760		− return mlx5_get_vector_affinity(dev->mdev, comp_vector);
	4759	+ return mlx5_get_vector_affinity_hint(dev->mdev, comp_vector);
4761	4760	}
4762	4761
4763	4762	/* The mlx5_ib_multiport_mutex should be held when calling this function */
@@ -5427,9 +5426,7 @@ static void mlx5_ib_stage_cong_debugfs_cleanup(struct mlx5_ib_dev *dev) · 663 unmodified lines
5427	5426	static int mlx5_ib_stage_uar_init(struct mlx5_ib_dev *dev)
5428	5427	{
5429	5428	dev->mdev->priv.uar = mlx5_get_uars_page(dev->mdev);
5430		− if (!dev->mdev->priv.uar)
5431		− return -ENOMEM;
5432		− return 0;
	5429	+ return PTR_ERR_OR_ZERO(dev->mdev->priv.uar);
5433	5430	}
5434	5431
5435	5432	static void mlx5_ib_stage_uar_cleanup(struct mlx5_ib_dev *dev)

drivers/infiniband/hw/mlx5/mr.c+23−9

@@ -866,25 +866,28 @@ static int mr_umem_get(struct ib_pd *pd, u64 start, u64 length, · 865 unmodified lines
866	866	int *order)
867	867	{
868	868	struct mlx5_ib_dev *dev = to_mdev(pd->device);
	869	+ struct ib_umem *u;
869	870	int err;
870	871
871		− *umem = ib_umem_get(pd->uobject->context, start, length,
872		− access_flags, 0);
873		− err = PTR_ERR_OR_ZERO(*umem);
	872	+ *umem = NULL;
	873	+
	874	+ u = ib_umem_get(pd->uobject->context, start, length, access_flags, 0);
	875	+ err = PTR_ERR_OR_ZERO(u);
874	876	if (err) {
875		− *umem = NULL;
876		− mlx5_ib_err(dev, "umem get failed (%d)\n", err);
	877	+ mlx5_ib_dbg(dev, "umem get failed (%d)\n", err);
877	878	return err;
878	879	}
879	880
880		− mlx5_ib_cont_pages(*umem, start, MLX5_MKEY_PAGE_SHIFT_MASK, npages,
	881	+ mlx5_ib_cont_pages(u, start, MLX5_MKEY_PAGE_SHIFT_MASK, npages,
881	882	page_shift, ncont, order);
882	883	if (!*npages) {
883	884	mlx5_ib_warn(dev, "avoid zero region\n");
884		− ib_umem_release(*umem);
	885	+ ib_umem_release(u);
885	886	return -EINVAL;
886	887	}
887	888
	889	+ *umem = u;
	890	+
888	891	mlx5_ib_dbg(dev, "npages %d, ncont %d, order %d, page_shift %d\n",
889	892	npages, ncont, order, page_shift);
890	893
@@ -1458,13 +1461,12 @@ int mlx5_ib_rereg_user_mr(struct ib_mr *ib_mr, int flags, u64 start, · 567 unmodified lines
1458	1461	int access_flags = flags & IB_MR_REREG_ACCESS ?
1459	1462	new_access_flags :
1460	1463	mr->access_flags;
1461		− u64 addr = (flags & IB_MR_REREG_TRANS) ? virt_addr : mr->umem->address;
1462		− u64 len = (flags & IB_MR_REREG_TRANS) ? length : mr->umem->length;
1463	1464	int page_shift = 0;
1464	1465	int upd_flags = 0;
1465	1466	int npages = 0;
1466	1467	int ncont = 0;
1467	1468	int order = 0;
	1469	+ u64 addr, len;
1468	1470	int err;
1469	1471
1470	1472	mlx5_ib_dbg(dev, "start 0x%llx, virt_addr 0x%llx, length 0x%llx, access_flags 0x%x\n",
@@ -1472,6 +1474,17 @@ int mlx5_ib_rereg_user_mr(struct ib_mr *ib_mr, int flags, u64 start, · 1 unmodified line
1472	1474
1473	1475	atomic_sub(mr->npages, &dev->mdev->priv.reg_pages);
1474	1476
	1477	+ if (!mr->umem)
	1478	+ return -EINVAL;
	1479	+
	1480	+ if (flags & IB_MR_REREG_TRANS) {
	1481	+ addr = virt_addr;
	1482	+ len = length;
	1483	+ } else {
	1484	+ addr = mr->umem->address;
	1485	+ len = mr->umem->length;
	1486	+ }
	1487	+
1475	1488	if (flags != IB_MR_REREG_PD) {
1476	1489	/*
1477	1490	* Replace umem. This needs to be done whether or not UMR is
@@ -1479,6 +1492,7 @@ int mlx5_ib_rereg_user_mr(struct ib_mr *ib_mr, int flags, u64 start, · 1 unmodified line
1479	1492	*/
1480	1493	flags \|= IB_MR_REREG_TRANS;
1481	1494	ib_umem_release(mr->umem);
	1495	+ mr->umem = NULL;
1482	1496	err = mr_umem_get(pd, addr, len, access_flags, &mr->umem,
1483	1497	&npages, &page_shift, &ncont, &order);
1484	1498	if (err)

drivers/infiniband/hw/mlx5/qp.c+13−9

@@ -259,7 +259,11 @@ static int set_rq_size(struct mlx5_ib_dev dev, struct ib_qp_cap cap, · 258 unmodified lines
259	259	} else {
260	260	if (ucmd) {
261	261	qp->rq.wqe_cnt = ucmd->rq_wqe_count;
	262	+ if (ucmd->rq_wqe_shift > BITS_PER_BYTE * sizeof(ucmd->rq_wqe_shift))
	263	+ return -EINVAL;
262	264	qp->rq.wqe_shift = ucmd->rq_wqe_shift;
	265	+ if ((1 << qp->rq.wqe_shift) / sizeof(struct mlx5_wqe_data_seg) < qp->wq_sig)
	266	+ return -EINVAL;
263	267	qp->rq.max_gs = (1 << qp->rq.wqe_shift) / sizeof(struct mlx5_wqe_data_seg) - qp->wq_sig;
264	268	qp->rq.max_post = qp->rq.wqe_cnt;
265	269	} else {
@@ -2451,18 +2455,18 @@ enum { · 2185 unmodified lines
2451	2455
2452	2456	static int ib_rate_to_mlx5(struct mlx5_ib_dev *dev, u8 rate)
2453	2457	{
2454		− if (rate == IB_RATE_PORT_CURRENT) {
	2458	+ if (rate == IB_RATE_PORT_CURRENT)
2455	2459	return 0;
2456		− } else if (rate < IB_RATE_2_5_GBPS \|\| rate > IB_RATE_300_GBPS) {
	2460	+
	2461	+ if (rate < IB_RATE_2_5_GBPS \|\| rate > IB_RATE_300_GBPS)
2457	2462	return -EINVAL;
2458		− } else {
2459		− while (rate != IB_RATE_2_5_GBPS &&
2460		− !(1 << (rate + MLX5_STAT_RATE_OFFSET) &
2461		− MLX5_CAP_GEN(dev->mdev, stat_rate_support)))
2462		− --rate;
2463		− }
2464	2463
2465		− return rate + MLX5_STAT_RATE_OFFSET;
	2464	+ while (rate != IB_RATE_PORT_CURRENT &&
	2465	+ !(1 << (rate + MLX5_STAT_RATE_OFFSET) &
	2466	+ MLX5_CAP_GEN(dev->mdev, stat_rate_support)))
	2467	+ --rate;
	2468	+
	2469	+ return rate ? rate + MLX5_STAT_RATE_OFFSET : rate;
2466	2470	}
2467	2471
2468	2472	static int modify_raw_packet_eth_prio(struct mlx5_core_dev *dev,

drivers/infiniband/hw/nes/nes_nic.c+1−1

@@ -461,7 +461,7 @@ static bool nes_nic_send(struct sk_buff skb, struct net_device netdev) · 460 unmodified lines
461	461	/**
462	462	* nes_netdev_start_xmit
463	463	*/
464		−static int nes_netdev_start_xmit(struct sk_buff skb, struct net_device netdev)
	464	+static netdev_tx_t nes_netdev_start_xmit(struct sk_buff skb, struct net_device netdev)
465	465	{
466	466	struct nes_vnic *nesvnic = netdev_priv(netdev);
467	467	struct nes_device *nesdev = nesvnic->nesdev;

drivers/infiniband/sw/rxe/rxe_opcode.c+1−1

@@ -390,7 +390,7 @@ struct rxe_opcode_info rxe_opcode[RXE_NUM_OPCODE] = { · 389 unmodified lines
390	390	.name = "IB_OPCODE_RC_SEND_ONLY_INV",
391	391	.mask = RXE_IETH_MASK \| RXE_PAYLOAD_MASK \| RXE_REQ_MASK
392	392	\| RXE_COMP_MASK \| RXE_RWR_MASK \| RXE_SEND_MASK
393		− \| RXE_END_MASK,
	393	+ \| RXE_END_MASK \| RXE_START_MASK,
394	394	.length = RXE_BTH_BYTES + RXE_IETH_BYTES,
395	395	.offset = {
396	396	[RXE_BTH] = 0,

drivers/infiniband/sw/rxe/rxe_req.c+0−1

@@ -728,7 +728,6 @@ next_wqe: · 727 unmodified lines
728	728	rollback_state(wqe, qp, &rollback_wqe, rollback_psn);
729	729
730	730	if (ret == -EAGAIN) {
731		− kfree_skb(skb);
732	731	rxe_run_task(&qp->req.task, 1);
733	732	goto exit;
734	733	}

drivers/infiniband/sw/rxe/rxe_resp.c+1−5

@@ -742,7 +742,6 @@ static enum resp_states read_reply(struct rxe_qp *qp, · 741 unmodified lines
742	742	err = rxe_xmit_packet(rxe, qp, &ack_pkt, skb);
743	743	if (err) {
744	744	pr_err("Failed sending RDMA reply.\n");
745		− kfree_skb(skb);
746	745	return RESPST_ERR_RNR;
747	746	}
748	747
@@ -954,10 +953,8 @@ static int send_ack(struct rxe_qp qp, struct rxe_pkt_info pkt, · 205 unmodified lines
954	953	}
955	954
956	955	err = rxe_xmit_packet(rxe, qp, &ack_pkt, skb);
957		− if (err) {
	956	+ if (err)
958	957	pr_err_ratelimited("Failed sending ack\n");
959		− kfree_skb(skb);
960		− }
961	958
962	959	err1:
963	960	return err;
@@ -1141,7 +1138,6 @@ static enum resp_states duplicate_request(struct rxe_qp *qp, · 177 unmodified lines
1141	1138	if (rc) {
1142	1139	pr_err("Failed resending result. This flow is not handled - skb ignored\n");
1143	1140	rxe_drop_ref(qp);
1144		− kfree_skb(skb_copy);
1145	1141	rc = RESPST_CLEANUP;
1146	1142	goto out;
1147	1143	}

drivers/infiniband/ulp/ipoib/ipoib_main.c+1−1

@@ -1094,7 +1094,7 @@ drop_and_unlock: · 1093 unmodified lines
1094	1094	spin_unlock_irqrestore(&priv->lock, flags);
1095	1095	}
1096	1096
1097		−static int ipoib_start_xmit(struct sk_buff skb, struct net_device dev)
	1097	+static netdev_tx_t ipoib_start_xmit(struct sk_buff skb, struct net_device dev)
1098	1098	{
1099	1099	struct ipoib_dev_priv *priv = ipoib_priv(dev);
1100	1100	struct rdma_netdev *rn = netdev_priv(dev);

drivers/infiniband/ulp/srp/Kconfig+1−1

@@ -1,6 +1,6 @@
1	1	config INFINIBAND_SRP
2	2	tristate "InfiniBand SCSI RDMA Protocol"
3		− depends on SCSI
	3	+ depends on SCSI && INFINIBAND_ADDR_TRANS
4	4	select SCSI_SRP_ATTRS
5	5	---help---
6	6	Support for the SCSI RDMA Protocol over InfiniBand. This

drivers/infiniband/ulp/srpt/Kconfig+1−1

@@ -1,6 +1,6 @@
1	1	config INFINIBAND_SRPT
2	2	tristate "InfiniBand SCSI RDMA Protocol target support"
3		− depends on INFINIBAND && TARGET_CORE
	3	+ depends on INFINIBAND && INFINIBAND_ADDR_TRANS && TARGET_CORE
4	4	---help---
5	5
6	6	Support for the SCSI RDMA Protocol (SRP) Target driver. The

drivers/input/evdev.c+6−1

@@ -31,6 +31,7 @@ · 30 unmodified lines
31	31	enum evdev_clock_type {
32	32	EV_CLK_REAL = 0,
33	33	EV_CLK_MONO,
	34	+ EV_CLK_BOOT,
34	35	EV_CLK_MAX
35	36	};
36	37
@@ -197,10 +198,12 @@ static int evdev_set_clk_type(struct evdev_client *client, unsigned int clkid) · 160 unmodified lines
197	198	case CLOCK_REALTIME:
198	199	clk_type = EV_CLK_REAL;
199	200	break;
200		− case CLOCK_BOOTTIME:
201	201	case CLOCK_MONOTONIC:
202	202	clk_type = EV_CLK_MONO;
203	203	break;
	204	+ case CLOCK_BOOTTIME:
	205	+ clk_type = EV_CLK_BOOT;
	206	+ break;
204	207	default:
205	208	return -EINVAL;
206	209	}
@@ -311,6 +314,8 @@ static void evdev_events(struct input_handle *handle, · 104 unmodified lines
311	314
312	315	ev_time[EV_CLK_MONO] = ktime_get();
313	316	ev_time[EV_CLK_REAL] = ktime_mono_to_real(ev_time[EV_CLK_MONO]);
	317	+ ev_time[EV_CLK_BOOT] = ktime_mono_to_any(ev_time[EV_CLK_MONO],
	318	+ TK_OFFS_BOOT);
314	319
315	320	rcu_read_lock();
316	321

drivers/input/input-leds.c+4−4

@@ -88,6 +88,7 @@ static int input_leds_connect(struct input_handler *handler, · 87 unmodified lines
88	88	const struct input_device_id *id)
89	89	{
90	90	struct input_leds *leds;
	91	+ struct input_led *led;
91	92	unsigned int num_leds;
92	93	unsigned int led_code;
93	94	int led_no;
@@ -119,14 +120,13 @@ static int input_leds_connect(struct input_handler *handler, · 25 unmodified lines
119	120
120	121	led_no = 0;
121	122	for_each_set_bit(led_code, dev->ledbit, LED_CNT) {
122		− struct input_led *led = &leds->leds[led_no];
	123	+ if (!input_led_info[led_code].name)
	124	+ continue;
123	125
	126	+ led = &leds->leds[led_no];
124	127	led->handle = &leds->handle;
125	128	led->code = led_code;
126	129
127		− if (!input_led_info[led_code].name)
128		− continue;
129		−
130	130	led->cdev.name = kasprintf(GFP_KERNEL, "%s::%s",
131	131	dev_name(&dev->dev),
132	132	input_led_info[led_code].name);

drivers/input/mouse/alps.c+1−1

@@ -583,7 +583,7 @@ static void alps_process_trackstick_packet_v3(struct psmouse *psmouse) · 582 unmodified lines
583	583
584	584	x = (s8)(((packet[0] & 0x20) << 2) \| (packet[1] & 0x7f));
585	585	y = (s8)(((packet[0] & 0x10) << 3) \| (packet[2] & 0x7f));
586		− z = packet[4] & 0x7c;
	586	+ z = packet[4] & 0x7f;
587	587
588	588	/*
589	589	* The x and y values tend to be quite large, and when used

drivers/input/rmi4/rmi_spi.c+5−2

@@ -147,8 +147,11 @@ static int rmi_spi_xfer(struct rmi_spi_xport *rmi_spi, · 146 unmodified lines
147	147	if (len > RMI_SPI_XFER_SIZE_LIMIT)
148	148	return -EINVAL;
149	149
150		− if (rmi_spi->xfer_buf_size < len)
151		− rmi_spi_manage_pools(rmi_spi, len);
	150	+ if (rmi_spi->xfer_buf_size < len) {
	151	+ ret = rmi_spi_manage_pools(rmi_spi, len);
	152	+ if (ret < 0)
	153	+ return ret;
	154	+ }
152	155
153	156	if (addr == 0)
154	157	/*

drivers/input/touchscreen/Kconfig+1−1

@@ -362,7 +362,7 @@ config TOUCHSCREEN_HIDEEP · 361 unmodified lines
362	362
363	363	If unsure, say N.
364	364
365		− To compile this driver as a moudle, choose M here : the
	365	+ To compile this driver as a module, choose M here : the
366	366	module will be called hideep_ts.
367	367
368	368	config TOUCHSCREEN_ILI210X

drivers/input/touchscreen/atmel_mxt_ts.c+124−76

@@ -280,7 +280,8 @@ struct mxt_data { · 279 unmodified lines
280	280	struct input_dev *input_dev;
281	281	char phys[64]; /* device physical location */
282	282	struct mxt_object *object_table;
283		− struct mxt_info info;
	283	+ struct mxt_info *info;
	284	+ void *raw_info_block;
284	285	unsigned int irq;
285	286	unsigned int max_x;
286	287	unsigned int max_y;
@@ -460,12 +461,13 @@ static int mxt_lookup_bootloader_address(struct mxt_data *data, bool retry) · 173 unmodified lines
460	461	{
461	462	u8 appmode = data->client->addr;
462	463	u8 bootloader;
	464	+ u8 family_id = data->info ? data->info->family_id : 0;
463	465
464	466	switch (appmode) {
465	467	case 0x4a:
466	468	case 0x4b:
467	469	/* Chips after 1664S use different scheme */
468		− if (retry \|\| data->info.family_id >= 0xa2) {
	470	+ if (retry \|\| family_id >= 0xa2) {
469	471	bootloader = appmode - 0x24;
470	472	break;
471	473	}
@@ -692,7 +694,7 @@ mxt_get_object(struct mxt_data *data, u8 type) · 220 unmodified lines
692	694	struct mxt_object *object;
693	695	int i;
694	696
695		− for (i = 0; i < data->info.object_num; i++) {
	697	+ for (i = 0; i < data->info->object_num; i++) {
696	698	object = data->object_table + i;
697	699	if (object->type == type)
698	700	return object;
@@ -1462,12 +1464,12 @@ static int mxt_update_cfg(struct mxt_data data, const struct firmware cfg) · 763 unmodified lines
1462	1464	data_pos += offset;
1463	1465	}
1464	1466
1465		− if (cfg_info.family_id != data->info.family_id) {
	1467	+ if (cfg_info.family_id != data->info->family_id) {
1466	1468	dev_err(dev, "Family ID mismatch!\n");
1467	1469	return -EINVAL;
1468	1470	}
1469	1471
1470		− if (cfg_info.variant_id != data->info.variant_id) {
	1472	+ if (cfg_info.variant_id != data->info->variant_id) {
1471	1473	dev_err(dev, "Variant ID mismatch!\n");
1472	1474	return -EINVAL;
1473	1475	}
@@ -1512,7 +1514,7 @@ static int mxt_update_cfg(struct mxt_data data, const struct firmware cfg) · 38 unmodified lines
1512	1514
1513	1515	/* Malloc memory to store configuration */
1514	1516	cfg_start_ofs = MXT_OBJECT_START +
1515		− data->info.object_num * sizeof(struct mxt_object) +
	1517	+ data->info->object_num * sizeof(struct mxt_object) +
1516	1518	MXT_INFO_CHECKSUM_SIZE;
1517	1519	config_mem_size = data->mem_size - cfg_start_ofs;
1518	1520	config_mem = kzalloc(config_mem_size, GFP_KERNEL);
@@ -1563,20 +1565,6 @@ release_mem: · 44 unmodified lines
1563	1565	return ret;
1564	1566	}
1565	1567
1566		−static int mxt_get_info(struct mxt_data *data)
1567		−{
1568		− struct i2c_client *client = data->client;
1569		− struct mxt_info *info = &data->info;
1570		− int error;
1571		−
1572		− /* Read 7-byte info block starting at address 0 */
1573		− error = __mxt_read_reg(client, 0, sizeof(*info), info);
1574		− if (error)
1575		− return error;
1576		−
1577		− return 0;
1578		−}
1579		−
1580	1568	static void mxt_free_input_device(struct mxt_data *data)
1581	1569	{
1582	1570	if (data->input_dev) {
@@ -1591,9 +1579,10 @@ static void mxt_free_object_table(struct mxt_data *data) · 8 unmodified lines
1591	1579	video_unregister_device(&data->dbg.vdev);
1592	1580	v4l2_device_unregister(&data->dbg.v4l2);
1593	1581	#endif
1594		−
1595		− kfree(data->object_table);
1596	1582	data->object_table = NULL;
	1583	+ data->info = NULL;
	1584	+ kfree(data->raw_info_block);
	1585	+ data->raw_info_block = NULL;
1597	1586	kfree(data->msg_buf);
1598	1587	data->msg_buf = NULL;
1599	1588	data->T5_address = 0;
@@ -1609,34 +1598,18 @@ static void mxt_free_object_table(struct mxt_data *data) · 9 unmodified lines
1609	1598	data->max_reportid = 0;
1610	1599	}
1611	1600
1612		−static int mxt_get_object_table(struct mxt_data *data)
	1601	+static int mxt_parse_object_table(struct mxt_data *data,
	1602	+ struct mxt_object *object_table)
1613	1603	{
1614	1604	struct i2c_client *client = data->client;
1615		− size_t table_size;
1616		− struct mxt_object *object_table;
1617		− int error;
1618	1605	int i;
1619	1606	u8 reportid;
1620	1607	u16 end_address;
1621	1608
1622		− table_size = data->info.object_num * sizeof(struct mxt_object);
1623		− object_table = kzalloc(table_size, GFP_KERNEL);
1624		− if (!object_table) {
1625		− dev_err(&data->client->dev, "Failed to allocate memory\n");
1626		− return -ENOMEM;
1627		− }
1628		−
1629		− error = __mxt_read_reg(client, MXT_OBJECT_START, table_size,
1630		− object_table);
1631		− if (error) {
1632		− kfree(object_table);
1633		− return error;
1634		− }
1635		−
1636	1609	/* Valid Report IDs start counting from 1 */
1637	1610	reportid = 1;
1638	1611	data->mem_size = 0;
1639		− for (i = 0; i < data->info.object_num; i++) {
	1612	+ for (i = 0; i < data->info->object_num; i++) {
1640	1613	struct mxt_object *object = object_table + i;
1641	1614	u8 min_id, max_id;
1642	1615
@@ -1660,8 +1633,8 @@ static int mxt_get_object_table(struct mxt_data *data) · 17 unmodified lines
1660	1633
1661	1634	switch (object->type) {
1662	1635	case MXT_GEN_MESSAGE_T5:
1663		− if (data->info.family_id == 0x80 &&
1664		− data->info.version < 0x20) {
	1636	+ if (data->info->family_id == 0x80 &&
	1637	+ data->info->version < 0x20) {
1665	1638	/*
1666	1639	* On mXT224 firmware versions prior to V2.0
1667	1640	* read and discard unused CRC byte otherwise
@@ -1716,24 +1689,102 @@ static int mxt_get_object_table(struct mxt_data *data) · 48 unmodified lines
1716	1689	/* If T44 exists, T5 position has to be directly after */
1717	1690	if (data->T44_address && (data->T5_address != data->T44_address + 1)) {
1718	1691	dev_err(&client->dev, "Invalid T44 position\n");
1719		− error = -EINVAL;
1720		− goto free_object_table;
	1692	+ return -EINVAL;
1721	1693	}
1722	1694
1723	1695	data->msg_buf = kcalloc(data->max_reportid,
1724	1696	data->T5_msg_size, GFP_KERNEL);
1725		− if (!data->msg_buf) {
1726		− dev_err(&client->dev, "Failed to allocate message buffer\n");
	1697	+ if (!data->msg_buf)
	1698	+ return -ENOMEM;
	1699	+
	1700	+ return 0;
	1701	+}
	1702	+
	1703	+static int mxt_read_info_block(struct mxt_data *data)
	1704	+{
	1705	+ struct i2c_client *client = data->client;
	1706	+ int error;
	1707	+ size_t size;
	1708	+ void id_buf, buf;
	1709	+ uint8_t num_objects;
	1710	+ u32 calculated_crc;
	1711	+ u8 *crc_ptr;
	1712	+
	1713	+ /* If info block already allocated, free it */
	1714	+ if (data->raw_info_block)
	1715	+ mxt_free_object_table(data);
	1716	+
	1717	+ /* Read 7-byte ID information block starting at address 0 */
	1718	+ size = sizeof(struct mxt_info);
	1719	+ id_buf = kzalloc(size, GFP_KERNEL);
	1720	+ if (!id_buf)
	1721	+ return -ENOMEM;
	1722	+
	1723	+ error = __mxt_read_reg(client, 0, size, id_buf);
	1724	+ if (error)
	1725	+ goto err_free_mem;
	1726	+
	1727	+ /* Resize buffer to give space for rest of info block */
	1728	+ num_objects = ((struct mxt_info *)id_buf)->object_num;
	1729	+ size += (num_objects * sizeof(struct mxt_object))
	1730	+ + MXT_INFO_CHECKSUM_SIZE;
	1731	+
	1732	+ buf = krealloc(id_buf, size, GFP_KERNEL);
	1733	+ if (!buf) {
1727	1734	error = -ENOMEM;
1728		− goto free_object_table;
	1735	+ goto err_free_mem;
	1736	+ }
	1737	+ id_buf = buf;
	1738	+
	1739	+ /* Read rest of info block */
	1740	+ error = __mxt_read_reg(client, MXT_OBJECT_START,
	1741	+ size - MXT_OBJECT_START,
	1742	+ id_buf + MXT_OBJECT_START);
	1743	+ if (error)
	1744	+ goto err_free_mem;
	1745	+
	1746	+ /* Extract & calculate checksum */
	1747	+ crc_ptr = id_buf + size - MXT_INFO_CHECKSUM_SIZE;
	1748	+ data->info_crc = crc_ptr[0] \| (crc_ptr[1] << 8) \| (crc_ptr[2] << 16);
	1749	+
	1750	+ calculated_crc = mxt_calculate_crc(id_buf, 0,
	1751	+ size - MXT_INFO_CHECKSUM_SIZE);
	1752	+
	1753	+ /*
	1754	+ * CRC mismatch can be caused by data corruption due to I2C comms
	1755	+ * issue or else device is not using Object Based Protocol (eg i2c-hid)
	1756	+ */
	1757	+ if ((data->info_crc == 0) \|\| (data->info_crc != calculated_crc)) {
	1758	+ dev_err(&client->dev,
	1759	+ "Info Block CRC error calculated=0x%06X read=0x%06X\n",
	1760	+ calculated_crc, data->info_crc);
	1761	+ error = -EIO;
	1762	+ goto err_free_mem;
	1763	+ }
	1764	+
	1765	+ data->raw_info_block = id_buf;
	1766	+ data->info = (struct mxt_info *)id_buf;
	1767	+
	1768	+ dev_info(&client->dev,
	1769	+ "Family: %u Variant: %u Firmware V%u.%u.%02X Objects: %u\n",
	1770	+ data->info->family_id, data->info->variant_id,
	1771	+ data->info->version >> 4, data->info->version & 0xf,
	1772	+ data->info->build, data->info->object_num);
	1773	+
	1774	+ /* Parse object table information */
	1775	+ error = mxt_parse_object_table(data, id_buf + MXT_OBJECT_START);
	1776	+ if (error) {
	1777	+ dev_err(&client->dev, "Error %d parsing object table\n", error);
	1778	+ mxt_free_object_table(data);
	1779	+ goto err_free_mem;
1729	1780	}
1730	1781
1731		− data->object_table = object_table;
	1782	+ data->object_table = (struct mxt_object *)(id_buf + MXT_OBJECT_START);
1732	1783
1733	1784	return 0;
1734	1785
1735		−free_object_table:
1736		− mxt_free_object_table(data);
	1786	+err_free_mem:
	1787	+ kfree(id_buf);
1737	1788	return error;
1738	1789	}
1739	1790
@@ -2046,7 +2097,7 @@ static int mxt_initialize(struct mxt_data *data) · 306 unmodified lines
2046	2097	int error;
2047	2098
2048	2099	while (1) {
2049		− error = mxt_get_info(data);
	2100	+ error = mxt_read_info_block(data);
2050	2101	if (!error)
2051	2102	break;
2052	2103
@@ -2077,16 +2128,9 @@ static int mxt_initialize(struct mxt_data *data) · 24 unmodified lines
2077	2128	msleep(MXT_FW_RESET_TIME);
2078	2129	}
2079	2130
2080		− /* Get object table information */
2081		− error = mxt_get_object_table(data);
2082		− if (error) {
2083		− dev_err(&client->dev, "Error %d reading object table\n", error);
2084		− return error;
2085		− }
2086		−
2087	2131	error = mxt_acquire_irq(data);
2088	2132	if (error)
2089		− goto err_free_object_table;
	2133	+ return error;
2090	2134
2091	2135	error = request_firmware_nowait(THIS_MODULE, true, MXT_CFG_NAME,
2092	2136	&client->dev, GFP_KERNEL, data,
@@ -2094,14 +2138,10 @@ static int mxt_initialize(struct mxt_data *data) · 1 unmodified line
2094	2138	if (error) {
2095	2139	dev_err(&client->dev, "Failed to invoke firmware loader: %d\n",
2096	2140	error);
2097		− goto err_free_object_table;
	2141	+ return error;
2098	2142	}
2099	2143
2100	2144	return 0;
2101		−
2102		−err_free_object_table:
2103		− mxt_free_object_table(data);
2104		− return error;
2105	2145	}
2106	2146
2107	2147	static int mxt_set_t7_power_cfg(struct mxt_data *data, u8 sleep)
@@ -2162,7 +2202,7 @@ recheck: · 54 unmodified lines
2162	2202	static u16 mxt_get_debug_value(struct mxt_data *data, unsigned int x,
2163	2203	unsigned int y)
2164	2204	{
2165		− struct mxt_info *info = &data->info;
	2205	+ struct mxt_info *info = data->info;
2166	2206	struct mxt_dbg *dbg = &data->dbg;
2167	2207	unsigned int ofs, page;
2168	2208	unsigned int col = 0;
@@ -2490,7 +2530,7 @@ static const struct video_device mxt_video_device = { · 321 unmodified lines
2490	2530
2491	2531	static void mxt_debug_init(struct mxt_data *data)
2492	2532	{
2493		− struct mxt_info *info = &data->info;
	2533	+ struct mxt_info *info = data->info;
2494	2534	struct mxt_dbg *dbg = &data->dbg;
2495	2535	struct mxt_object *object;
2496	2536	int error;
@@ -2576,7 +2616,6 @@ static int mxt_configure_objects(struct mxt_data *data, · 79 unmodified lines
2576	2616	const struct firmware *cfg)
2577	2617	{
2578	2618	struct device *dev = &data->client->dev;
2579		− struct mxt_info *info = &data->info;
2580	2619	int error;
2581	2620
2582	2621	error = mxt_init_t7_power_cfg(data);
@@ -2601,11 +2640,6 @@ static int mxt_configure_objects(struct mxt_data *data, · 18 unmodified lines
2601	2640
2602	2641	mxt_debug_init(data);
2603	2642
2604		− dev_info(dev,
2605		− "Family: %u Variant: %u Firmware V%u.%u.%02X Objects: %u\n",
2606		− info->family_id, info->variant_id, info->version >> 4,
2607		− info->version & 0xf, info->build, info->object_num);
2608		−
2609	2643	return 0;
2610	2644	}
2611	2645
@@ -2614,7 +2648,7 @@ static ssize_t mxt_fw_version_show(struct device *dev, · 2 unmodified lines
2614	2648	struct device_attribute attr, char buf)
2615	2649	{
2616	2650	struct mxt_data *data = dev_get_drvdata(dev);
2617		− struct mxt_info *info = &data->info;
	2651	+ struct mxt_info *info = data->info;
2618	2652	return scnprintf(buf, PAGE_SIZE, "%u.%u.%02X\n",
2619	2653	info->version >> 4, info->version & 0xf, info->build);
2620	2654	}
@@ -2624,7 +2658,7 @@ static ssize_t mxt_hw_version_show(struct device *dev, · 3 unmodified lines
2624	2658	struct device_attribute attr, char buf)
2625	2659	{
2626	2660	struct mxt_data *data = dev_get_drvdata(dev);
2627		− struct mxt_info *info = &data->info;
	2661	+ struct mxt_info *info = data->info;
2628	2662	return scnprintf(buf, PAGE_SIZE, "%u.%u\n",
2629	2663	info->family_id, info->variant_id);
2630	2664	}
@@ -2663,7 +2697,7 @@ static ssize_t mxt_object_show(struct device *dev, · 32 unmodified lines
2663	2697	return -ENOMEM;
2664	2698
2665	2699	error = 0;
2666		− for (i = 0; i < data->info.object_num; i++) {
	2700	+ for (i = 0; i < data->info->object_num; i++) {
2667	2701	object = data->object_table + i;
2668	2702
2669	2703	if (!mxt_object_readable(object->type))
@@ -3034,6 +3068,15 @@ static const struct dmi_system_id mxt_dmi_table[] = { · 364 unmodified lines
3034	3068	},
3035	3069	.driver_data = samus_platform_data,
3036	3070	},
	3071	+ {
	3072	+ /* Samsung Chromebook Pro */
	3073	+ .ident = "Samsung Chromebook Pro",
	3074	+ .matches = {
	3075	+ DMI_MATCH(DMI_SYS_VENDOR, "Google"),
	3076	+ DMI_MATCH(DMI_PRODUCT_NAME, "Caroline"),
	3077	+ },
	3078	+ .driver_data = samus_platform_data,
	3079	+ },
3037	3080	{
3038	3081	/* Other Google Chromebooks */
3039	3082	.ident = "Chromebook",
@@ -3254,6 +3297,11 @@ static SIMPLE_DEV_PM_OPS(mxt_pm_ops, mxt_suspend, mxt_resume); · 214 unmodified lines
3254	3297
3255	3298	static const struct of_device_id mxt_of_match[] = {
3256	3299	{ .compatible = "atmel,maxtouch", },
	3300	+ /* Compatibles listed below are deprecated */
	3301	+ { .compatible = "atmel,qt602240_ts", },
	3302	+ { .compatible = "atmel,atmel_mxt_ts", },
	3303	+ { .compatible = "atmel,atmel_mxt_tp", },
	3304	+ { .compatible = "atmel,mXT224", },
3257	3305	{},
3258	3306	};
3259	3307	MODULE_DEVICE_TABLE(of, mxt_of_match);

drivers/iommu/amd_iommu.c+1−1

@@ -83,7 +83,6 @@ · 82 unmodified lines
83	83
84	84	static DEFINE_SPINLOCK(amd_iommu_devtable_lock);
85	85	static DEFINE_SPINLOCK(pd_bitmap_lock);
86		−static DEFINE_SPINLOCK(iommu_table_lock);
87	86
88	87	/* List of all available dev_data structures */
89	88	static LLIST_HEAD(dev_data_list);
@@ -3562,6 +3561,7 @@ EXPORT_SYMBOL(amd_iommu_device_info); · 3472 unmodified lines
3562	3561	*****************************************************************************/
3563	3562
3564	3563	static struct irq_chip amd_ir_chip;
	3564	+static DEFINE_SPINLOCK(iommu_table_lock);
3565	3565
3566	3566	static void set_dte_irq_entry(u16 devid, struct irq_remap_table *table)
3567	3567	{

drivers/iommu/dma-iommu.c+25−29

@@ -167,40 +167,16 @@ EXPORT_SYMBOL(iommu_put_dma_cookie); · 166 unmodified lines
167	167	* @list: Reserved region list from iommu_get_resv_regions()
168	168	*
169	169	* IOMMU drivers can use this to implement their .get_resv_regions callback
170		− * for general non-IOMMU-specific reservations. Currently, this covers host
171		− * bridge windows for PCI devices and GICv3 ITS region reservation on ACPI
172		− * based ARM platforms that may require HW MSI reservation.
	170	+ * for general non-IOMMU-specific reservations. Currently, this covers GICv3
	171	+ * ITS region reservation on ACPI based ARM platforms that may require HW MSI
	172	+ * reservation.
173	173	*/
174	174	void iommu_dma_get_resv_regions(struct device dev, struct list_head list)
175	175	{
176		− struct pci_host_bridge *bridge;
177		− struct resource_entry *window;
178		−
179		− if (!is_of_node(dev->iommu_fwspec->iommu_fwnode) &&
180		− iort_iommu_msi_get_resv_regions(dev, list) < 0)
181		− return;
182		−
183		− if (!dev_is_pci(dev))
184		− return;
185		−
186		− bridge = pci_find_host_bridge(to_pci_dev(dev)->bus);
187		− resource_list_for_each_entry(window, &bridge->windows) {
188		− struct iommu_resv_region *region;
189		− phys_addr_t start;
190		− size_t length;
191		−
192		− if (resource_type(window->res) != IORESOURCE_MEM)
193		− continue;
194	176
195		− start = window->res->start - window->offset;
196		− length = window->res->end - window->res->start + 1;
197		− region = iommu_alloc_resv_region(start, length, 0,
198		− IOMMU_RESV_RESERVED);
199		− if (!region)
200		− return;
	177	+ if (!is_of_node(dev->iommu_fwspec->iommu_fwnode))
	178	+ iort_iommu_msi_get_resv_regions(dev, list);
201	179
202		− list_add_tail(&region->list, list);
203		− }
204	180	}
205	181	EXPORT_SYMBOL(iommu_dma_get_resv_regions);
206	182
@@ -229,6 +205,23 @@ static int cookie_init_hw_msi_region(struct iommu_dma_cookie *cookie, · 22 unmodified lines
229	205	return 0;
230	206	}
231	207
	208	+static void iova_reserve_pci_windows(struct pci_dev *dev,
	209	+ struct iova_domain *iovad)
	210	+{
	211	+ struct pci_host_bridge *bridge = pci_find_host_bridge(dev->bus);
	212	+ struct resource_entry *window;
	213	+ unsigned long lo, hi;
	214	+
	215	+ resource_list_for_each_entry(window, &bridge->windows) {
	216	+ if (resource_type(window->res) != IORESOURCE_MEM)
	217	+ continue;
	218	+
	219	+ lo = iova_pfn(iovad, window->res->start - window->offset);
	220	+ hi = iova_pfn(iovad, window->res->end - window->offset);
	221	+ reserve_iova(iovad, lo, hi);
	222	+ }
	223	+}
	224	+
232	225	static int iova_reserve_iommu_regions(struct device *dev,
233	226	struct iommu_domain *domain)
234	227	{
@@ -238,6 +231,9 @@ static int iova_reserve_iommu_regions(struct device *dev, · 3 unmodified lines
238	231	LIST_HEAD(resv_regions);
239	232	int ret = 0;
240	233
	234	+ if (dev_is_pci(dev))
	235	+ iova_reserve_pci_windows(to_pci_dev(dev), iovad);
	236	+
241	237	iommu_get_resv_regions(dev, &resv_regions);
242	238	list_for_each_entry(region, &resv_regions, list) {
243	239	unsigned long lo, hi;

drivers/iommu/dmar.c+1−1

@@ -1345,7 +1345,7 @@ void qi_flush_dev_iotlb(struct intel_iommu *iommu, u16 sid, u16 qdep, · 1344 unmodified lines
1345	1345	struct qi_desc desc;
1346	1346
1347	1347	if (mask) {
1348		− BUG_ON(addr & ((1 << (VTD_PAGE_SHIFT + mask)) - 1));
	1348	+ WARN_ON_ONCE(addr & ((1ULL << (VTD_PAGE_SHIFT + mask)) - 1));
1349	1349	addr \|= (1ULL << (VTD_PAGE_SHIFT + mask - 1)) - 1;
1350	1350	desc.high = QI_DEV_IOTLB_ADDR(addr) \| QI_DEV_IOTLB_SIZE;
1351	1351	} else

drivers/iommu/intel_irq_remapping.c+1−1

@@ -1136,7 +1136,7 @@ static void intel_ir_reconfigure_irte(struct irq_data *irqd, bool force) · 1135 unmodified lines
1136	1136	irte->dest_id = IRTE_DEST(cfg->dest_apicid);
1137	1137
1138	1138	/* Update the hardware only if the interrupt is in remapped mode. */
1139		− if (!force \|\| ir_data->irq_2_iommu.mode == IRQ_REMAPPING)
	1139	+ if (force \|\| ir_data->irq_2_iommu.mode == IRQ_REMAPPING)
1140	1140	modify_irte(&ir_data->irq_2_iommu, irte);
1141	1141	}
1142	1142

drivers/iommu/rockchip-iommu.c+9−2

@@ -1098,7 +1098,7 @@ static int rk_iommu_of_xlate(struct device *dev, · 1097 unmodified lines
1098	1098	data->iommu = platform_get_drvdata(iommu_dev);
1099	1099	dev->archdata.iommu = data;
1100	1100
1101		− of_dev_put(iommu_dev);
	1101	+ platform_device_put(iommu_dev);
1102	1102
1103	1103	return 0;
1104	1104	}
@@ -1175,8 +1175,15 @@ static int rk_iommu_probe(struct platform_device *pdev) · 70 unmodified lines
1175	1175	for (i = 0; i < iommu->num_clocks; ++i)
1176	1176	iommu->clocks[i].id = rk_iommu_clocks[i];
1177	1177
	1178	+ /*
	1179	+ * iommu clocks should be present for all new devices and devicetrees
	1180	+ * but there are older devicetrees without clocks out in the wild.
	1181	+ * So clocks as optional for the time being.
	1182	+ */
1178	1183	err = devm_clk_bulk_get(iommu->dev, iommu->num_clocks, iommu->clocks);
1179		− if (err)
	1184	+ if (err == -ENOENT)
	1185	+ iommu->num_clocks = 0;
	1186	+ else if (err)
1180	1187	return err;
1181	1188
1182	1189	err = clk_bulk_prepare(iommu->num_clocks, iommu->clocks);

drivers/irqchip/qcom-irq-combiner.c+2−2

@@ -1,4 +1,4 @@
1		−/* Copyright (c) 2015-2016, The Linux Foundation. All rights reserved.
	1	+/* Copyright (c) 2015-2018, The Linux Foundation. All rights reserved.
2	2	*
3	3	* This program is free software; you can redistribute it and/or modify
4	4	* it under the terms of the GNU General Public License version 2 and
@@ -68,7 +68,7 @@ static void combiner_handle_irq(struct irq_desc *desc) · 63 unmodified lines
68	68
69	69	bit = readl_relaxed(combiner->regs[reg].addr);
70	70	status = bit & combiner->regs[reg].enabled;
71		− if (!status)
	71	+ if (bit && !status)
72	72	pr_warn_ratelimited("Unexpected IRQ on CPU%d: (%08x %08lx %p)\n",
73	73	smp_processor_id(), bit,
74	74	combiner->regs[reg].enabled,

drivers/isdn/mISDN/dsp_hwec.c+5−3

@@ -68,12 +68,12 @@ void dsp_hwec_enable(struct dsp dsp, const char arg) · 67 unmodified lines
68	68	goto _do;
69	69
70	70	{
71		− char _dup[len + 1];
72	71	char dup, tok, name, val;
73	72	int tmp;
74	73
75		− strcpy(_dup, arg);
76		− dup = _dup;
	74	+ dup = kstrdup(arg, GFP_ATOMIC);
	75	+ if (!dup)
	76	+ return;
77	77
78	78	while ((tok = strsep(&dup, ","))) {
79	79	if (!strlen(tok))
@@ -89,6 +89,8 @@ void dsp_hwec_enable(struct dsp dsp, const char arg) · 9 unmodified lines
89	89	deftaps = tmp;
90	90	}
91	91	}
	92	+
	93	+ kfree(dup);
92	94	}
93	95
94	96	_do:

drivers/isdn/mISDN/l1oip_core.c+11−3

@@ -279,7 +279,7 @@ l1oip_socket_send(struct l1oip *hc, u8 localcodec, u8 channel, u32 chanmask, · 278 unmodified lines
279	279	u16 timebase, u8 *buf, int len)
280	280	{
281	281	u8 *p;
282		− u8 frame[len + 32];
	282	+ u8 frame[MAX_DFRAME_LEN_L1 + 32];
283	283	struct socket *socket = NULL;
284	284
285	285	if (debug & DEBUG_L1OIP_MSG)
@@ -902,7 +902,11 @@ handle_dmsg(struct mISDNchannel ch, struct sk_buff skb) · 616 unmodified lines
902	902	p = skb->data;
903	903	l = skb->len;
904	904	while (l) {
905		− ll = (l < L1OIP_MAX_PERFRAME) ? l : L1OIP_MAX_PERFRAME;
	905	+ /*
	906	+ * This is technically bounded by L1OIP_MAX_PERFRAME but
	907	+ * MAX_DFRAME_LEN_L1 < L1OIP_MAX_PERFRAME
	908	+ */
	909	+ ll = (l < MAX_DFRAME_LEN_L1) ? l : MAX_DFRAME_LEN_L1;
906	910	l1oip_socket_send(hc, 0, dch->slot, 0,
907	911	hc->chan[dch->slot].tx_counter++, p, ll);
908	912	p += ll;
@@ -1140,7 +1144,11 @@ handle_bmsg(struct mISDNchannel ch, struct sk_buff skb) · 231 unmodified lines
1140	1144	p = skb->data;
1141	1145	l = skb->len;
1142	1146	while (l) {
1143		− ll = (l < L1OIP_MAX_PERFRAME) ? l : L1OIP_MAX_PERFRAME;
	1147	+ /*
	1148	+ * This is technically bounded by L1OIP_MAX_PERFRAME but
	1149	+ * MAX_DFRAME_LEN_L1 < L1OIP_MAX_PERFRAME
	1150	+ */
	1151	+ ll = (l < MAX_DFRAME_LEN_L1) ? l : MAX_DFRAME_LEN_L1;
1144	1152	l1oip_socket_send(hc, hc->codec, bch->slot, 0,
1145	1153	hc->chan[bch->slot].tx_counter, p, ll);
1146	1154	hc->chan[bch->slot].tx_counter += ll;

drivers/md/bcache/alloc.c+4−1

@@ -290,7 +290,7 @@ do { \ · 289 unmodified lines
290	290	if (kthread_should_stop() \|\| \
291	291	test_bit(CACHE_SET_IO_DISABLE, &ca->set->flags)) { \
292	292	set_current_state(TASK_RUNNING); \
293		− return 0; \
	293	+ goto out; \
294	294	} \
295	295	\
296	296	schedule(); \
@@ -378,6 +378,9 @@ retry_invalidate: · 81 unmodified lines
378	378	bch_prio_write(ca);
379	379	}
380	380	}
	381	+out:
	382	+ wait_for_kthread_stop();
	383	+ return 0;
381	384	}
382	385
383	386	/* Allocation */

drivers/md/bcache/bcache.h+4−0

@@ -392,6 +392,8 @@ struct cached_dev { · 391 unmodified lines
392	392	#define DEFAULT_CACHED_DEV_ERROR_LIMIT 64
393	393	atomic_t io_errors;
394	394	unsigned error_limit;
	395	+
	396	+ char backing_dev_name[BDEVNAME_SIZE];
395	397	};
396	398
397	399	enum alloc_reserve {
@@ -464,6 +466,8 @@ struct cache { · 66 unmodified lines
464	466	atomic_long_t meta_sectors_written;
465	467	atomic_long_t btree_sectors_written;
466	468	atomic_long_t sectors_written;
	469	+
	470	+ char cache_dev_name[BDEVNAME_SIZE];
467	471	};
468	472
469	473	struct gc_stat {

drivers/md/bcache/debug.c+1−2

@@ -106,7 +106,6 @@ void bch_btree_verify(struct btree *b) · 105 unmodified lines
106	106
107	107	void bch_data_verify(struct cached_dev dc, struct bio bio)
108	108	{
109		− char name[BDEVNAME_SIZE];
110	109	struct bio *check;
111	110	struct bio_vec bv, cbv;
112	111	struct bvec_iter iter, citer = { 0 };
@@ -134,7 +133,7 @@ void bch_data_verify(struct cached_dev dc, struct bio bio) · 21 unmodified lines
134	133	bv.bv_len),
135	134	dc->disk.c,
136	135	"verify failed at dev %s sector %llu",
137		− bdevname(dc->bdev, name),
	136	+ dc->backing_dev_name,
138	137	(uint64_t) bio->bi_iter.bi_sector);
139	138
140	139	kunmap_atomic(p1);

drivers/md/bcache/io.c+3−5

@@ -52,7 +52,6 @@ void bch_submit_bbio(struct bio bio, struct cache_set c, · 51 unmodified lines
52	52	/* IO errors */
53	53	void bch_count_backing_io_errors(struct cached_dev dc, struct bio bio)
54	54	{
55		− char buf[BDEVNAME_SIZE];
56	55	unsigned errors;
57	56
58	57	WARN_ONCE(!dc, "NULL pointer of struct cached_dev");
@@ -60,7 +59,7 @@ void bch_count_backing_io_errors(struct cached_dev dc, struct bio bio) · 1 unmodified line
60	59	errors = atomic_add_return(1, &dc->io_errors);
61	60	if (errors < dc->error_limit)
62	61	pr_err("%s: IO error on backing device, unrecoverable",
63		− bio_devname(bio, buf));
	62	+ dc->backing_dev_name);
64	63	else
65	64	bch_cached_dev_error(dc);
66	65	}
@@ -105,19 +104,18 @@ void bch_count_io_errors(struct cache *ca, · 38 unmodified lines
105	104	}
106	105
107	106	if (error) {
108		− char buf[BDEVNAME_SIZE];
109	107	unsigned errors = atomic_add_return(1 << IO_ERROR_SHIFT,
110	108	&ca->io_errors);
111	109	errors >>= IO_ERROR_SHIFT;
112	110
113	111	if (errors < ca->set->error_limit)
114	112	pr_err("%s: IO error on %s%s",
115		− bdevname(ca->bdev, buf), m,
	113	+ ca->cache_dev_name, m,
116	114	is_read ? ", recovering." : ".");
117	115	else
118	116	bch_cache_set_error(ca->set,
119	117	"%s: too many IO errors %s",
120		− bdevname(ca->bdev, buf), m);
	118	+ ca->cache_dev_name, m);
121	119	}
122	120	}
123	121

drivers/md/bcache/request.c+1−4

@@ -649,11 +649,8 @@ static void backing_request_endio(struct bio *bio) · 648 unmodified lines
649	649	*/
650	650	if (unlikely(s->iop.writeback &&
651	651	bio->bi_opf & REQ_PREFLUSH)) {
652		− char buf[BDEVNAME_SIZE];
653		−
654		− bio_devname(bio, buf);
655	652	pr_err("Can't flush %s: returned bi_status %i",
656		− buf, bio->bi_status);
	653	+ dc->backing_dev_name, bio->bi_status);
657	654	} else {
658	655	/* set to orig_bio->bi_status in bio_complete() */
659	656	s->iop.status = bio->bi_status;

drivers/md/bcache/super.c+52−23

@@ -936,7 +936,6 @@ static void cancel_writeback_rate_update_dwork(struct cached_dev *dc) · 935 unmodified lines
936	936	static void cached_dev_detach_finish(struct work_struct *w)
937	937	{
938	938	struct cached_dev *dc = container_of(w, struct cached_dev, detach);
939		− char buf[BDEVNAME_SIZE];
940	939	struct closure cl;
941	940	closure_init_stack(&cl);
942	941
@@ -967,7 +966,7 @@ static void cached_dev_detach_finish(struct work_struct *w) · 24 unmodified lines
967	966
968	967	mutex_unlock(&bch_register_lock);
969	968
970		− pr_info("Caching disabled for %s", bdevname(dc->bdev, buf));
	969	+ pr_info("Caching disabled for %s", dc->backing_dev_name);
971	970
972	971	/* Drop ref we took in cached_dev_detach() */
973	972	closure_put(&dc->disk.cl);
@@ -999,29 +998,28 @@ int bch_cached_dev_attach(struct cached_dev dc, struct cache_set c, · 25 unmodified lines
999	998	{
1000	999	uint32_t rtime = cpu_to_le32(get_seconds());
1001	1000	struct uuid_entry *u;
1002		− char buf[BDEVNAME_SIZE];
1003	1001	struct cached_dev exist_dc, t;
1004	1002
1005		− bdevname(dc->bdev, buf);
1006		−
1007	1003	if ((set_uuid && memcmp(set_uuid, c->sb.set_uuid, 16)) \|\|
1008	1004	(!set_uuid && memcmp(dc->sb.set_uuid, c->sb.set_uuid, 16)))
1009	1005	return -ENOENT;
1010	1006
1011	1007	if (dc->disk.c) {
1012		− pr_err("Can't attach %s: already attached", buf);
	1008	+ pr_err("Can't attach %s: already attached",
	1009	+ dc->backing_dev_name);
1013	1010	return -EINVAL;
1014	1011	}
1015	1012
1016	1013	if (test_bit(CACHE_SET_STOPPING, &c->flags)) {
1017		− pr_err("Can't attach %s: shutting down", buf);
	1014	+ pr_err("Can't attach %s: shutting down",
	1015	+ dc->backing_dev_name);
1018	1016	return -EINVAL;
1019	1017	}
1020	1018
1021	1019	if (dc->sb.block_size < c->sb.block_size) {
1022	1020	/* Will die */
1023	1021	pr_err("Couldn't attach %s: block size less than set's block size",
1024		− buf);
	1022	+ dc->backing_dev_name);
1025	1023	return -EINVAL;
1026	1024	}
1027	1025
@@ -1029,7 +1027,7 @@ int bch_cached_dev_attach(struct cached_dev dc, struct cache_set c, · 1 unmodified line
1029	1027	list_for_each_entry_safe(exist_dc, t, &c->cached_devs, list) {
1030	1028	if (!memcmp(dc->sb.uuid, exist_dc->sb.uuid, 16)) {
1031	1029	pr_err("Tried to attach %s but duplicate UUID already attached",
1032		− buf);
	1030	+ dc->backing_dev_name);
1033	1031
1034	1032	return -EINVAL;
1035	1033	}
@@ -1047,13 +1045,15 @@ int bch_cached_dev_attach(struct cached_dev dc, struct cache_set c, · 11 unmodified lines
1047	1045
1048	1046	if (!u) {
1049	1047	if (BDEV_STATE(&dc->sb) == BDEV_STATE_DIRTY) {
1050		− pr_err("Couldn't find uuid for %s in set", buf);
	1048	+ pr_err("Couldn't find uuid for %s in set",
	1049	+ dc->backing_dev_name);
1051	1050	return -ENOENT;
1052	1051	}
1053	1052
1054	1053	u = uuid_find_empty(c);
1055	1054	if (!u) {
1056		− pr_err("Not caching %s, no room for UUID", buf);
	1055	+ pr_err("Not caching %s, no room for UUID",
	1056	+ dc->backing_dev_name);
1057	1057	return -EINVAL;
1058	1058	}
1059	1059	}
@@ -1112,7 +1112,8 @@ int bch_cached_dev_attach(struct cached_dev dc, struct cache_set c, · 52 unmodified lines
1112	1112	up_write(&dc->writeback_lock);
1113	1113
1114	1114	pr_info("Caching %s as %s on set %pU",
1115		− bdevname(dc->bdev, buf), dc->disk.disk->disk_name,
	1115	+ dc->backing_dev_name,
	1116	+ dc->disk.disk->disk_name,
1116	1117	dc->disk.c->sb.set_uuid);
1117	1118	return 0;
1118	1119	}
@@ -1225,10 +1226,10 @@ static void register_bdev(struct cache_sb sb, struct page sb_page, · 106 unmodified lines
1225	1226	struct block_device *bdev,
1226	1227	struct cached_dev *dc)
1227	1228	{
1228		− char name[BDEVNAME_SIZE];
1229	1229	const char *err = "cannot allocate memory";
1230	1230	struct cache_set *c;
1231	1231
	1232	+ bdevname(bdev, dc->backing_dev_name);
1232	1233	memcpy(&dc->sb, sb, sizeof(struct cache_sb));
1233	1234	dc->bdev = bdev;
1234	1235	dc->bdev->bd_holder = dc;
@@ -1237,6 +1238,7 @@ static void register_bdev(struct cache_sb sb, struct page sb_page, · 2 unmodified lines
1237	1238	bio_first_bvec_all(&dc->sb_bio)->bv_page = sb_page;
1238	1239	get_page(sb_page);
1239	1240
	1241	+
1240	1242	if (cached_dev_init(dc, sb->block_size << 9))
1241	1243	goto err;
1242	1244
@@ -1247,7 +1249,7 @@ static void register_bdev(struct cache_sb sb, struct page sb_page, · 4 unmodified lines
1247	1249	if (bch_cache_accounting_add_kobjs(&dc->accounting, &dc->disk.kobj))
1248	1250	goto err;
1249	1251
1250		− pr_info("registered backing device %s", bdevname(bdev, name));
	1252	+ pr_info("registered backing device %s", dc->backing_dev_name);
1251	1253
1252	1254	list_add(&dc->list, &uncached_devices);
1253	1255	list_for_each_entry(c, &bch_cache_sets, list)
@@ -1259,7 +1261,7 @@ static void register_bdev(struct cache_sb sb, struct page sb_page, · 5 unmodified lines
1259	1261
1260	1262	return;
1261	1263	err:
1262		− pr_notice("error %s: %s", bdevname(bdev, name), err);
	1264	+ pr_notice("error %s: %s", dc->backing_dev_name, err);
1263	1265	bcache_device_stop(&dc->disk);
1264	1266	}
1265	1267
@@ -1367,7 +1369,7 @@ int bch_flash_dev_create(struct cache_set *c, uint64_t size) · 101 unmodified lines
1367	1369
1368	1370	bool bch_cached_dev_error(struct cached_dev *dc)
1369	1371	{
1370		− char name[BDEVNAME_SIZE];
	1372	+ struct cache_set *c;
1371	1373
1372	1374	if (!dc \|\| test_bit(BCACHE_DEV_CLOSING, &dc->disk.flags))
1373	1375	return false;
@@ -1377,7 +1379,22 @@ bool bch_cached_dev_error(struct cached_dev *dc) · 3 unmodified lines
1377	1379	smp_mb();
1378	1380
1379	1381	pr_err("stop %s: too many IO errors on backing device %s\n",
1380		− dc->disk.disk->disk_name, bdevname(dc->bdev, name));
	1382	+ dc->disk.disk->disk_name, dc->backing_dev_name);
	1383	+
	1384	+ /*
	1385	+ * If the cached device is still attached to a cache set,
	1386	+ * even dc->io_disable is true and no more I/O requests
	1387	+ * accepted, cache device internal I/O (writeback scan or
	1388	+ * garbage collection) may still prevent bcache device from
	1389	+ * being stopped. So here CACHE_SET_IO_DISABLE should be
	1390	+ * set to c->flags too, to make the internal I/O to cache
	1391	+ * device rejected and stopped immediately.
	1392	+ * If c is NULL, that means the bcache device is not attached
	1393	+ * to any cache set, then no CACHE_SET_IO_DISABLE bit to set.
	1394	+ */
	1395	+ c = dc->disk.c;
	1396	+ if (c && test_and_set_bit(CACHE_SET_IO_DISABLE, &c->flags))
	1397	+ pr_info("CACHE_SET_IO_DISABLE already set");
1381	1398
1382	1399	bcache_device_stop(&dc->disk);
1383	1400	return true;
@@ -1395,7 +1412,7 @@ bool bch_cache_set_error(struct cache_set c, const char fmt, ...) · 11 unmodified lines
1395	1412	return false;
1396	1413
1397	1414	if (test_and_set_bit(CACHE_SET_IO_DISABLE, &c->flags))
1398		− pr_warn("CACHE_SET_IO_DISABLE already set");
	1415	+ pr_info("CACHE_SET_IO_DISABLE already set");
1399	1416
1400	1417	/* XXX: we can be called from atomic context
1401	1418	acquire_console_sem();
@@ -1539,6 +1556,20 @@ static void conditional_stop_bcache_device(struct cache_set *c, · 137 unmodified lines
1539	1556	*/
1540	1557	pr_warn("stop_when_cache_set_failed of %s is \"auto\" and cache is dirty, stop it to avoid potential data corruption.",
1541	1558	d->disk->disk_name);
	1559	+ /*
	1560	+ * There might be a small time gap that cache set is
	1561	+ * released but bcache device is not. Inside this time
	1562	+ * gap, regular I/O requests will directly go into
	1563	+ * backing device as no cache set attached to. This
	1564	+ * behavior may also introduce potential inconsistence
	1565	+ * data in writeback mode while cache is dirty.
	1566	+ * Therefore before calling bcache_device_stop() due
	1567	+ * to a broken cache device, dc->io_disable should be
	1568	+ * explicitly set to true.
	1569	+ */
	1570	+ dc->io_disable = true;
	1571	+ /* make others know io_disable is true earlier */
	1572	+ smp_mb();
1542	1573	bcache_device_stop(d);
1543	1574	} else {
1544	1575	/*
@@ -2003,12 +2034,10 @@ static int cache_alloc(struct cache *ca) · 458 unmodified lines
2003	2034	static int register_cache(struct cache_sb sb, struct page sb_page,
2004	2035	struct block_device bdev, struct cache ca)
2005	2036	{
2006		− char name[BDEVNAME_SIZE];
2007	2037	const char err = NULL; / must be set for any error case */
2008	2038	int ret = 0;
2009	2039
2010		− bdevname(bdev, name);
2011		−
	2040	+ bdevname(bdev, ca->cache_dev_name);
2012	2041	memcpy(&ca->sb, sb, sizeof(struct cache_sb));
2013	2042	ca->bdev = bdev;
2014	2043	ca->bdev->bd_holder = ca;
@@ -2045,14 +2074,14 @@ static int register_cache(struct cache_sb sb, struct page sb_page, · 30 unmodified lines
2045	2074	goto out;
2046	2075	}
2047	2076
2048		− pr_info("registered cache device %s", name);
	2077	+ pr_info("registered cache device %s", ca->cache_dev_name);
2049	2078
2050	2079	out:
2051	2080	kobject_put(&ca->kobj);
2052	2081
2053	2082	err:
2054	2083	if (err)
2055		− pr_notice("error %s: %s", name, err);
	2084	+ pr_notice("error %s: %s", ca->cache_dev_name, err);
2056	2085
2057	2086	return ret;
2058	2087	}

drivers/md/bcache/writeback.c+3−1

@@ -244,8 +244,10 @@ static void dirty_endio(struct bio *bio) · 243 unmodified lines
244	244	struct keybuf_key *w = bio->bi_private;
245	245	struct dirty_io *io = w->private;
246	246
247		− if (bio->bi_status)
	247	+ if (bio->bi_status) {
248	248	SET_KEY_DIRTY(&w->key, false);
	249	+ bch_count_backing_io_errors(io->dc, bio);
	250	+ }
249	251
250	252	closure_put(&io->cl);
251	253	}

drivers/md/md.c+4−2

@@ -9256,8 +9256,10 @@ void md_reload_sb(struct mddev *mddev, int nr) · 9255 unmodified lines
9256	9256	check_sb_changes(mddev, rdev);
9257	9257
9258	9258	/* Read all rdev's to update recovery_offset */
9259		− rdev_for_each_rcu(rdev, mddev)
9260		− read_rdev(mddev, rdev);
	9259	+ rdev_for_each_rcu(rdev, mddev) {
	9260	+ if (!test_bit(Faulty, &rdev->flags))
	9261	+ read_rdev(mddev, rdev);
	9262	+ }
9261	9263	}
9262	9264	EXPORT_SYMBOL(md_reload_sb);
9263	9265

drivers/md/raid1.c+20−5

@@ -854,7 +854,7 @@ static void flush_pending_writes(struct r1conf *conf) · 853 unmodified lines
854	854	* there is no normal IO happeing. It must arrange to call
855	855	* lower_barrier when the particular background IO completes.
856	856	*/
857		−static void raise_barrier(struct r1conf *conf, sector_t sector_nr)
	857	+static sector_t raise_barrier(struct r1conf *conf, sector_t sector_nr)
858	858	{
859	859	int idx = sector_to_idx(sector_nr);
860	860
@@ -885,13 +885,23 @@ static void raise_barrier(struct r1conf *conf, sector_t sector_nr) · 24 unmodified lines
885	885	* max resync count which allowed on current I/O barrier bucket.
886	886	*/
887	887	wait_event_lock_irq(conf->wait_barrier,
888		− !conf->array_frozen &&
	888	+ (!conf->array_frozen &&
889	889	!atomic_read(&conf->nr_pending[idx]) &&
890		− atomic_read(&conf->barrier[idx]) < RESYNC_DEPTH,
	890	+ atomic_read(&conf->barrier[idx]) < RESYNC_DEPTH) \|\|
	891	+ test_bit(MD_RECOVERY_INTR, &conf->mddev->recovery),
891	892	conf->resync_lock);
892	893
	894	+ if (test_bit(MD_RECOVERY_INTR, &conf->mddev->recovery)) {
	895	+ atomic_dec(&conf->barrier[idx]);
	896	+ spin_unlock_irq(&conf->resync_lock);
	897	+ wake_up(&conf->wait_barrier);
	898	+ return -EINTR;
	899	+ }
	900	+
893	901	atomic_inc(&conf->nr_sync_pending);
894	902	spin_unlock_irq(&conf->resync_lock);
	903	+
	904	+ return 0;
895	905	}
896	906
897	907	static void lower_barrier(struct r1conf *conf, sector_t sector_nr)
@@ -1092,6 +1102,8 @@ static void alloc_behind_master_bio(struct r1bio *r1_bio, · 194 unmodified lines
1092	1102	goto skip_copy;
1093	1103	}
1094	1104
	1105	+ behind_bio->bi_write_hint = bio->bi_write_hint;
	1106	+
1095	1107	while (i < vcnt && size) {
1096	1108	struct page *page;
1097	1109	int len = min_t(int, PAGE_SIZE, size);
@@ -2662,9 +2674,12 @@ static sector_t raid1_sync_request(struct mddev *mddev, sector_t sector_nr, · 1564 unmodified lines
2662	2674
2663	2675	bitmap_cond_end_sync(mddev->bitmap, sector_nr,
2664	2676	mddev_is_clustered(mddev) && (sector_nr + 2 * RESYNC_SECTORS > conf->cluster_sync_high));
2665		− r1_bio = raid1_alloc_init_r1buf(conf);
2666	2677
2667		− raise_barrier(conf, sector_nr);
	2678	+
	2679	+ if (raise_barrier(conf, sector_nr))
	2680	+ return 0;
	2681	+
	2682	+ r1_bio = raid1_alloc_init_r1buf(conf);
2668	2683
2669	2684	rcu_read_lock();
2670	2685	/*

drivers/media/i2c/saa7115.c+1−1

@@ -20,7 +20,7 @@ · 19 unmodified lines
20	20	//
21	21	// VBI support (2004) and cleanups (2005) by Hans Verkuil <hverkuil@xs4all.nl>
22	22	//
23		−// Copyright (c) 2005-2006 Mauro Carvalho Chehab <mchehab@infradead.org>
	23	+// Copyright (c) 2005-2006 Mauro Carvalho Chehab <mchehab@kernel.org>
24	24	// SAA7111, SAA7113 and SAA7118 support
25	25
26	26	#include "saa711x_regs.h"

drivers/media/i2c/saa711x_regs.h+1−1

@@ -2,7 +2,7 @@ · 1 unmodified line
2	2	* SPDX-License-Identifier: GPL-2.0+
3	3	* saa711x - Philips SAA711x video decoder register specifications
4	4	*
5		− * Copyright (c) 2006 Mauro Carvalho Chehab <mchehab@infradead.org>
	5	+ * Copyright (c) 2006 Mauro Carvalho Chehab <mchehab@kernel.org>
6	6	*/
7	7
8	8	#define R_00_CHIP_VERSION 0x00

drivers/media/i2c/tda7432.c+1−1

@@ -8,7 +8,7 @@ · 7 unmodified lines
8	8	* Muting and tone control by Jonathan Isom <jisom@ematic.com>
9	9	*
10	10	* Copyright (c) 2000 Eric Sandeen <eric_sandeen@bigfoot.com>
11		− * Copyright (c) 2006 Mauro Carvalho Chehab <mchehab@infradead.org>
	11	+ * Copyright (c) 2006 Mauro Carvalho Chehab <mchehab@kernel.org>
12	12	* This code is placed under the terms of the GNU General Public License
13	13	* Based on tda9855.c by Steve VanDeBogart (vandebo@uclink.berkeley.edu)
14	14	* Which was based on tda8425.c by Greg Alexander (c) 1998

drivers/media/i2c/tvp5150.c+1−1

@@ -2,7 +2,7 @@ · 1 unmodified line
2	2	//
3	3	// tvp5150 - Texas Instruments TVP5150A/AM1 and TVP5151 video decoder driver
4	4	//
5		−// Copyright (c) 2005,2006 Mauro Carvalho Chehab <mchehab@infradead.org>
	5	+// Copyright (c) 2005,2006 Mauro Carvalho Chehab <mchehab@kernel.org>
6	6
7	7	#include <dt-bindings/media/tvp5150.h>
8	8	#include <linux/i2c.h>

drivers/media/i2c/tvp5150_reg.h+1−1

@@ -3,7 +3,7 @@ · 2 unmodified lines
3	3	*
4	4	* tvp5150 - Texas Instruments TVP5150A/AM1 video decoder registers
5	5	*
6		− * Copyright (c) 2005,2006 Mauro Carvalho Chehab <mchehab@infradead.org>
	6	+ * Copyright (c) 2005,2006 Mauro Carvalho Chehab <mchehab@kernel.org>
7	7	*/
8	8
9	9	#define TVP5150_VD_IN_SRC_SEL_1 0x00 /* Video input source selection #1 */

drivers/media/i2c/tvp7002.c+1−1

@@ -5,7 +5,7 @@ · 4 unmodified lines
5	5	* Author: Santiago Nunez-Corrales <santiago.nunez@ridgerun.com>
6	6	*
7	7	* This code is partially based upon the TVP5150 driver
8		− * written by Mauro Carvalho Chehab (mchehab@infradead.org),
	8	+ * written by Mauro Carvalho Chehab <mchehab@kernel.org>,
9	9	* the TVP514x driver written by Vaibhav Hiremath <hvaibhav@ti.com>
10	10	* and the TVP7002 driver in the TI LSP 2.10.00.14. Revisions by
11	11	* Muralidharan Karicheri and Snehaprabha Narnakaje (TI).

drivers/media/i2c/tvp7002_reg.h+1−1

@@ -5,7 +5,7 @@ · 4 unmodified lines
5	5	* Author: Santiago Nunez-Corrales <santiago.nunez@ridgerun.com>
6	6	*
7	7	* This code is partially based upon the TVP5150 driver
8		− * written by Mauro Carvalho Chehab (mchehab@infradead.org),
	8	+ * written by Mauro Carvalho Chehab <mchehab@kernel.org>,
9	9	* the TVP514x driver written by Vaibhav Hiremath <hvaibhav@ti.com>
10	10	* and the TVP7002 driver in the TI LSP 2.10.00.14
11	11	*

drivers/media/media-devnode.c+1−1

@@ -4,7 +4,7 @@ · 3 unmodified lines
4	4	* Copyright (C) 2010 Nokia Corporation
5	5	*
6	6	* Based on drivers/media/video/v4l2_dev.c code authored by
7		− * Mauro Carvalho Chehab <mchehab@infradead.org> (version 2)
	7	+ * Mauro Carvalho Chehab <mchehab@kernel.org> (version 2)
8	8	* Alan Cox, <alan@lxorguk.ukuu.org.uk> (version 1)
9	9	*
10	10	* Contacts: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

drivers/media/pci/bt8xx/bttv-audio-hook.c+1−1

@@ -1,7 +1,7 @@
1	1	/*
2	2	* Handlers for board audio hooks, splitted from bttv-cards
3	3	*
4		− * Copyright (c) 2006 Mauro Carvalho Chehab (mchehab@infradead.org)
	4	+ * Copyright (c) 2006 Mauro Carvalho Chehab <mchehab@kernel.org>
5	5	* This code is placed under the terms of the GNU General Public License
6	6	*/
7	7

drivers/media/pci/bt8xx/bttv-audio-hook.h+1−1

@@ -1,7 +1,7 @@
1	1	/*
2	2	* Handlers for board audio hooks, splitted from bttv-cards
3	3	*
4		− * Copyright (c) 2006 Mauro Carvalho Chehab (mchehab@infradead.org)
	4	+ * Copyright (c) 2006 Mauro Carvalho Chehab <mchehab@kernel.org>
5	5	* This code is placed under the terms of the GNU General Public License
6	6	*/
7	7

drivers/media/pci/bt8xx/bttv-cards.c+2−2

@@ -2447,7 +2447,7 @@ struct tvcard bttv_tvcards[] = { · 2446 unmodified lines
2447	2447	},
2448	2448	/* ---- card 0x88---------------------------------- */
2449	2449	[BTTV_BOARD_ACORP_Y878F] = {
2450		− /* Mauro Carvalho Chehab <mchehab@infradead.org> */
	2450	+ /* Mauro Carvalho Chehab <mchehab@kernel.org> */
2451	2451	.name = "Acorp Y878F",
2452	2452	.video_inputs = 3,
2453	2453	/* .audio_inputs= 1, */
@@ -2688,7 +2688,7 @@ struct tvcard bttv_tvcards[] = { · 234 unmodified lines
2688	2688	},
2689	2689	[BTTV_BOARD_ENLTV_FM_2] = {
2690	2690	/* Encore TV Tuner Pro ENL TV-FM-2
2691		− Mauro Carvalho Chehab <mchehab@infradead.org */
	2691	+ Mauro Carvalho Chehab <mchehab@kernel.org> */
2692	2692	.name = "Encore ENL TV-FM-2",
2693	2693	.video_inputs = 3,
2694	2694	/* .audio_inputs= 1, */

drivers/media/pci/bt8xx/bttv-driver.c+1−1

@@ -13,7 +13,7 @@ · 12 unmodified lines
13	13	(c) 2005-2006 Nickolay V. Shmyrev <nshmyrev@yandex.ru>
14	14
15	15	Fixes to be fully V4L2 compliant by
16		− (c) 2006 Mauro Carvalho Chehab <mchehab@infradead.org>
	16	+ (c) 2006 Mauro Carvalho Chehab <mchehab@kernel.org>
17	17
18	18	Cropping and overscan support
19	19	Copyright (C) 2005, 2006 Michael H. Schimek <mschimek@gmx.at>

drivers/media/pci/bt8xx/bttv-i2c.c+1−1

@@ -8,7 +8,7 @@ · 7 unmodified lines
8	8	& Marcus Metzler (mocm@thp.uni-koeln.de)
9	9	(c) 1999-2003 Gerd Knorr <kraxel@bytesex.org>
10	10
11		− (c) 2005 Mauro Carvalho Chehab <mchehab@infradead.org>
	11	+ (c) 2005 Mauro Carvalho Chehab <mchehab@kernel.org>
12	12	- Multituner support and i2c address binding
13	13
14	14	This program is free software; you can redistribute it and/or modify

drivers/media/pci/cx23885/cx23885-input.c+1−1

@@ -13,7 +13,7 @@ · 12 unmodified lines
13	13	* Copyright (C) 2008 <srinivasa.deevi at conexant dot com>
14	14	* Copyright (C) 2005 Ludovico Cavedon <cavedon@sssup.it>
15	15	* Markus Rechberger <mrechberger@gmail.com>
16		− * Mauro Carvalho Chehab <mchehab@infradead.org>
	16	+ * Mauro Carvalho Chehab <mchehab@kernel.org>
17	17	* Sascha Sommer <saschasommer@freenet.de>
18	18	* Copyright (C) 2004, 2005 Chris Pascoe
19	19	* Copyright (C) 2003, 2004 Gerd Knorr

drivers/media/pci/cx88/cx88-alsa.c+2−2

@@ -4,7 +4,7 @@ · 3 unmodified lines
4	4	*
5	5	* (c) 2007 Trent Piepho <xyzzy@speakeasy.org>
6	6	* (c) 2005,2006 Ricardo Cerqueira <v4l@cerqueira.org>
7		− * (c) 2005 Mauro Carvalho Chehab <mchehab@infradead.org>
	7	+ * (c) 2005 Mauro Carvalho Chehab <mchehab@kernel.org>
8	8	* Based on a dummy cx88 module by Gerd Knorr <kraxel@bytesex.org>
9	9	* Based on dummy.c by Jaroslav Kysela <perex@perex.cz>
10	10	*
@@ -103,7 +103,7 @@ MODULE_PARM_DESC(index, "Index value for cx88x capture interface(s)."); · 92 unmodified lines
103	103
104	104	MODULE_DESCRIPTION("ALSA driver module for cx2388x based TV cards");
105	105	MODULE_AUTHOR("Ricardo Cerqueira");
106		−MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@infradead.org>");
	106	+MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@kernel.org>");
107	107	MODULE_LICENSE("GPL");
108	108	MODULE_VERSION(CX88_VERSION);
109	109

drivers/media/pci/cx88/cx88-blackbird.c+1−1

@@ -5,7 +5,7 @@ · 4 unmodified lines
5	5	* (c) 2004 Jelle Foks <jelle@foks.us>
6	6	* (c) 2004 Gerd Knorr <kraxel@bytesex.org>
7	7	*
8		− * (c) 2005-2006 Mauro Carvalho Chehab <mchehab@infradead.org>
	8	+ * (c) 2005-2006 Mauro Carvalho Chehab <mchehab@kernel.org>
9	9	* - video_ioctl2 conversion
10	10	*
11	11	* Includes parts from the ivtv driver <http://sourceforge.net/projects/ivtv/>

drivers/media/pci/cx88/cx88-core.c+1−1

@@ -4,7 +4,7 @@ · 3 unmodified lines
4	4	*
5	5	* (c) 2003 Gerd Knorr <kraxel@bytesex.org> [SuSE Labs]
6	6	*
7		− * (c) 2005-2006 Mauro Carvalho Chehab <mchehab@infradead.org>
	7	+ * (c) 2005-2006 Mauro Carvalho Chehab <mchehab@kernel.org>
8	8	* - Multituner support
9	9	* - video_ioctl2 conversion
10	10	* - PAL/M fixes

drivers/media/pci/cx88/cx88-i2c.c+1−1

@@ -8,7 +8,7 @@ · 7 unmodified lines
8	8	* (c) 2002 Yurij Sysoev <yurij@naturesoft.net>
9	9	* (c) 1999-2003 Gerd Knorr <kraxel@bytesex.org>
10	10	*
11		− * (c) 2005 Mauro Carvalho Chehab <mchehab@infradead.org>
	11	+ * (c) 2005 Mauro Carvalho Chehab <mchehab@kernel.org>
12	12	* - Multituner support and i2c address binding
13	13	*
14	14	* This program is free software; you can redistribute it and/or modify

drivers/media/pci/cx88/cx88-video.c+1−1

@@ -5,7 +5,7 @@ · 4 unmodified lines
5	5	*
6	6	* (c) 2003-04 Gerd Knorr <kraxel@bytesex.org> [SuSE Labs]
7	7	*
8		− * (c) 2005-2006 Mauro Carvalho Chehab <mchehab@infradead.org>
	8	+ * (c) 2005-2006 Mauro Carvalho Chehab <mchehab@kernel.org>
9	9	* - Multituner support
10	10	* - video_ioctl2 conversion
11	11	* - PAL/M fixes

drivers/media/radio/radio-aimslab.c+1−1

@@ -4,7 +4,7 @@ · 3 unmodified lines
4	4	* Copyright 1997 M. Kirkwood
5	5	*
6	6	* Converted to the radio-isa framework by Hans Verkuil <hans.verkuil@cisco.com>
7		− * Converted to V4L2 API by Mauro Carvalho Chehab <mchehab@infradead.org>
	7	+ * Converted to V4L2 API by Mauro Carvalho Chehab <mchehab@kernel.org>
8	8	* Converted to new API by Alan Cox <alan@lxorguk.ukuu.org.uk>
9	9	* Various bugfixes and enhancements by Russell Kroll <rkroll@exploits.org>
10	10	*

drivers/media/radio/radio-aztech.c+1−1

@@ -2,7 +2,7 @@ · 1 unmodified line
2	2	* radio-aztech.c - Aztech radio card driver
3	3	*
4	4	* Converted to the radio-isa framework by Hans Verkuil <hans.verkuil@xs4all.nl>
5		− * Converted to V4L2 API by Mauro Carvalho Chehab <mchehab@infradead.org>
	5	+ * Converted to V4L2 API by Mauro Carvalho Chehab <mchehab@kernel.org>
6	6	* Adapted to support the Video for Linux API by
7	7	* Russell Kroll <rkroll@exploits.org>. Based on original tuner code by:
8	8	*

drivers/media/radio/radio-gemtek.c+1−1

@@ -15,7 +15,7 @@ · 14 unmodified lines
15	15	* Various bugfixes and enhancements by Russell Kroll <rkroll@exploits.org>
16	16	*
17	17	* Converted to the radio-isa framework by Hans Verkuil <hans.verkuil@cisco.com>
18		− * Converted to V4L2 API by Mauro Carvalho Chehab <mchehab@infradead.org>
	18	+ * Converted to V4L2 API by Mauro Carvalho Chehab <mchehab@kernel.org>
19	19	*
20	20	* Note: this card seems to swap the left and right audio channels!
21	21	*

drivers/media/radio/radio-maxiradio.c+1−1

@@ -27,7 +27,7 @@ · 26 unmodified lines
27	27	* BUGS:
28	28	* - card unmutes if you change frequency
29	29	*
30		− * (c) 2006, 2007 by Mauro Carvalho Chehab <mchehab@infradead.org>:
	30	+ * (c) 2006, 2007 by Mauro Carvalho Chehab <mchehab@kernel.org>:
31	31	* - Conversion to V4L2 API
32	32	* - Uses video_ioctl2 for parsing and to add debug support
33	33	*/

drivers/media/radio/radio-rtrack2.c+1−1

@@ -7,7 +7,7 @@ · 6 unmodified lines
7	7	* Various bugfixes and enhancements by Russell Kroll <rkroll@exploits.org>
8	8	*
9	9	* Converted to the radio-isa framework by Hans Verkuil <hans.verkuil@cisco.com>
10		− * Converted to V4L2 API by Mauro Carvalho Chehab <mchehab@infradead.org>
	10	+ * Converted to V4L2 API by Mauro Carvalho Chehab <mchehab@kernel.org>
11	11	*
12	12	* Fully tested with actual hardware and the v4l2-compliance tool.
13	13	*/

drivers/media/radio/radio-sf16fmi.c+1−1

@@ -13,7 +13,7 @@ · 12 unmodified lines
13	13	* No volume control - only mute/unmute - you have to use line volume
14	14	* control on SB-part of SF16-FMI/SF16-FMP/SF16-FMD
15	15	*
16		− * Converted to V4L2 API by Mauro Carvalho Chehab <mchehab@infradead.org>
	16	+ * Converted to V4L2 API by Mauro Carvalho Chehab <mchehab@kernel.org>
17	17	*/
18	18
19	19	#include <linux/kernel.h> /* __setup */

drivers/media/radio/radio-terratec.c+1−1

@@ -17,7 +17,7 @@ · 16 unmodified lines
17	17	* Volume Control is done digitally
18	18	*
19	19	* Converted to the radio-isa framework by Hans Verkuil <hans.verkuil@cisco.com>
20		− * Converted to V4L2 API by Mauro Carvalho Chehab <mchehab@infradead.org>
	20	+ * Converted to V4L2 API by Mauro Carvalho Chehab <mchehab@kernel.org>
21	21	*/
22	22
23	23	#include <linux/module.h> /* Modules */

drivers/media/radio/radio-trust.c+1−1

@@ -12,7 +12,7 @@ · 11 unmodified lines
12	12	* Scott McGrath (smcgrath@twilight.vtc.vsc.edu)
13	13	* William McGrath (wmcgrath@twilight.vtc.vsc.edu)
14	14	*
15		− * Converted to V4L2 API by Mauro Carvalho Chehab <mchehab@infradead.org>
	15	+ * Converted to V4L2 API by Mauro Carvalho Chehab <mchehab@kernel.org>
16	16	*/
17	17
18	18	#include <stdarg.h>

drivers/media/radio/radio-typhoon.c+1−1

@@ -25,7 +25,7 @@ · 24 unmodified lines
25	25	* The frequency change is necessary since the card never seems to be
26	26	* completely silent.
27	27	*
28		− * Converted to V4L2 API by Mauro Carvalho Chehab <mchehab@infradead.org>
	28	+ * Converted to V4L2 API by Mauro Carvalho Chehab <mchehab@kernel.org>
29	29	*/
30	30
31	31	#include <linux/module.h> /* Modules */

drivers/media/radio/radio-zoltrix.c+1−1

@@ -27,7 +27,7 @@ · 26 unmodified lines
27	27	* 2002-07-15 - Fix Stereo typo
28	28	*
29	29	* 2006-07-24 - Converted to V4L2 API
30		− * by Mauro Carvalho Chehab <mchehab@infradead.org>
	30	+ * by Mauro Carvalho Chehab <mchehab@kernel.org>
31	31	*
32	32	* Converted to the radio-isa framework by Hans Verkuil <hans.verkuil@cisco.com>
33	33	*

drivers/media/rc/keymaps/rc-avermedia-m135a.c+1−1

@@ -12,7 +12,7 @@ · 11 unmodified lines
12	12	*
13	13	* On Avermedia M135A with IR model RM-JX, the same codes exist on both
14	14	* Positivo (BR) and original IR, initial version and remote control codes
15		− * added by Mauro Carvalho Chehab <mchehab@infradead.org>
	15	+ * added by Mauro Carvalho Chehab <mchehab@kernel.org>
16	16	*
17	17	* Positivo also ships Avermedia M135A with model RM-K6, extra control
18	18	* codes added by Herton Ronaldo Krzesinski <herton@mandriva.com.br>

drivers/media/rc/keymaps/rc-encore-enltv-fm53.c+1−1

@@ -9,7 +9,7 @@ · 8 unmodified lines
9	9	#include <linux/module.h>
10	10
11	11	/* Encore ENLTV-FM v5.3
12		− Mauro Carvalho Chehab <mchehab@infradead.org>
	12	+ Mauro Carvalho Chehab <mchehab@kernel.org>
13	13	*/
14	14
15	15	static struct rc_map_table encore_enltv_fm53[] = {

drivers/media/rc/keymaps/rc-encore-enltv2.c+1−1

@@ -9,7 +9,7 @@ · 8 unmodified lines
9	9	#include <linux/module.h>
10	10
11	11	/* Encore ENLTV2-FM - silver plastic - "Wand Media" written at the botton
12		− Mauro Carvalho Chehab <mchehab@infradead.org> */
	12	+ Mauro Carvalho Chehab <mchehab@kernel.org> */
13	13
14	14	static struct rc_map_table encore_enltv2[] = {
15	15	{ 0x4c, KEY_POWER2 },

drivers/media/rc/keymaps/rc-kaiomy.c+1−1

@@ -9,7 +9,7 @@ · 8 unmodified lines
9	9	#include <linux/module.h>
10	10
11	11	/* Kaiomy TVnPC U2
12		− Mauro Carvalho Chehab <mchehab@infradead.org>
	12	+ Mauro Carvalho Chehab <mchehab@kernel.org>
13	13	*/
14	14
15	15	static struct rc_map_table kaiomy[] = {

drivers/media/rc/keymaps/rc-kworld-plus-tv-analog.c+1−1

@@ -9,7 +9,7 @@ · 8 unmodified lines
9	9	#include <linux/module.h>
10	10
11	11	/* Kworld Plus TV Analog Lite PCI IR
12		− Mauro Carvalho Chehab <mchehab@infradead.org>
	12	+ Mauro Carvalho Chehab <mchehab@kernel.org>
13	13	*/
14	14
15	15	static struct rc_map_table kworld_plus_tv_analog[] = {

drivers/media/rc/keymaps/rc-pixelview-new.c+1−1

@@ -9,7 +9,7 @@ · 8 unmodified lines
9	9	#include <linux/module.h>
10	10
11	11	/*
12		− Mauro Carvalho Chehab <mchehab@infradead.org>
	12	+ Mauro Carvalho Chehab <mchehab@kernel.org>
13	13	present on PV MPEG 8000GT
14	14	*/
15	15

drivers/media/tuners/tea5761.c+2−2

@@ -2,7 +2,7 @@ · 1 unmodified line
2	2	// For Philips TEA5761 FM Chip
3	3	// I2C address is always 0x20 (0x10 at 7-bit mode).
4	4	//
5		−// Copyright (c) 2005-2007 Mauro Carvalho Chehab (mchehab@infradead.org)
	5	+// Copyright (c) 2005-2007 Mauro Carvalho Chehab <mchehab@kernel.org>
6	6
7	7	#include <linux/i2c.h>
8	8	#include <linux/slab.h>
@@ -337,5 +337,5 @@ EXPORT_SYMBOL_GPL(tea5761_attach); · 328 unmodified lines
337	337	EXPORT_SYMBOL_GPL(tea5761_autodetection);
338	338
339	339	MODULE_DESCRIPTION("Philips TEA5761 FM tuner driver");
340		−MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@infradead.org>");
	340	+MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@kernel.org>");
341	341	MODULE_LICENSE("GPL v2");

drivers/media/tuners/tea5767.c+2−2

@@ -2,7 +2,7 @@ · 1 unmodified line
2	2	// For Philips TEA5767 FM Chip used on some TV Cards like Prolink Pixelview
3	3	// I2C address is always 0xC0.
4	4	//
5		−// Copyright (c) 2005 Mauro Carvalho Chehab (mchehab@infradead.org)
	5	+// Copyright (c) 2005 Mauro Carvalho Chehab <mchehab@kernel.org>
6	6	//
7	7	// tea5767 autodetection thanks to Torsten Seeboth and Atsushi Nakagawa
8	8	// from their contributions on DScaler.
@@ -469,5 +469,5 @@ EXPORT_SYMBOL_GPL(tea5767_attach); · 460 unmodified lines
469	469	EXPORT_SYMBOL_GPL(tea5767_autodetection);
470	470
471	471	MODULE_DESCRIPTION("Philips TEA5767 FM tuner driver");
472		−MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@infradead.org>");
	472	+MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@kernel.org>");
473	473	MODULE_LICENSE("GPL v2");

drivers/media/tuners/tuner-xc2028-types.h+1−1

@@ -5,7 +5,7 @@ · 4 unmodified lines
5	5	* This file includes internal tipes to be used inside tuner-xc2028.
6	6	* Shouldn't be included outside tuner-xc2028
7	7	*
8		− * Copyright (c) 2007-2008 Mauro Carvalho Chehab (mchehab@infradead.org)
	8	+ * Copyright (c) 2007-2008 Mauro Carvalho Chehab <mchehab@kernel.org>
9	9	*/
10	10
11	11	/* xc3028 firmware types */

drivers/media/tuners/tuner-xc2028.c+2−2

@@ -1,7 +1,7 @@
1	1	// SPDX-License-Identifier: GPL-2.0
2	2	// tuner-xc2028
3	3	//
4		−// Copyright (c) 2007-2008 Mauro Carvalho Chehab (mchehab@infradead.org)
	4	+// Copyright (c) 2007-2008 Mauro Carvalho Chehab <mchehab@kernel.org>
5	5	//
6	6	// Copyright (c) 2007 Michel Ludwig (michel.ludwig@gmail.com)
7	7	// - frontend interface
@@ -1518,7 +1518,7 @@ EXPORT_SYMBOL(xc2028_attach); · 1510 unmodified lines
1518	1518
1519	1519	MODULE_DESCRIPTION("Xceive xc2028/xc3028 tuner driver");
1520	1520	MODULE_AUTHOR("Michel Ludwig <michel.ludwig@gmail.com>");
1521		−MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@infradead.org>");
	1521	+MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@kernel.org>");
1522	1522	MODULE_LICENSE("GPL v2");
1523	1523	MODULE_FIRMWARE(XC2028_DEFAULT_FIRMWARE);
1524	1524	MODULE_FIRMWARE(XC3028L_DEFAULT_FIRMWARE);

drivers/media/tuners/tuner-xc2028.h+1−1

@@ -2,7 +2,7 @@ · 1 unmodified line
2	2	* SPDX-License-Identifier: GPL-2.0
3	3	* tuner-xc2028
4	4	*
5		− * Copyright (c) 2007-2008 Mauro Carvalho Chehab (mchehab@infradead.org)
	5	+ * Copyright (c) 2007-2008 Mauro Carvalho Chehab <mchehab@kernel.org>
6	6	*/
7	7
8	8	#ifndef __TUNER_XC2028_H__

drivers/media/usb/em28xx/em28xx-camera.c+1−1

@@ -2,7 +2,7 @@ · 1 unmodified line
2	2	//
3	3	// em28xx-camera.c - driver for Empia EM25xx/27xx/28xx USB video capture devices
4	4	//
5		−// Copyright (C) 2009 Mauro Carvalho Chehab <mchehab@infradead.org>
	5	+// Copyright (C) 2009 Mauro Carvalho Chehab <mchehab@kernel.org>
6	6	// Copyright (C) 2013 Frank Schäfer <fschaefer.oss@googlemail.com>
7	7	//
8	8	// This program is free software; you can redistribute it and/or modify

drivers/media/usb/em28xx/em28xx-cards.c+1−1

@@ -5,7 +5,7 @@ · 4 unmodified lines
5	5	//
6	6	// Copyright (C) 2005 Ludovico Cavedon <cavedon@sssup.it>
7	7	// Markus Rechberger <mrechberger@gmail.com>
8		−// Mauro Carvalho Chehab <mchehab@infradead.org>
	8	+// Mauro Carvalho Chehab <mchehab@kernel.org>
9	9	// Sascha Sommer <saschasommer@freenet.de>
10	10	// Copyright (C) 2012 Frank Schäfer <fschaefer.oss@googlemail.com>
11	11	//

drivers/media/usb/em28xx/em28xx-core.c+2−2

@@ -4,7 +4,7 @@ · 3 unmodified lines
4	4	//
5	5	// Copyright (C) 2005 Ludovico Cavedon <cavedon@sssup.it>
6	6	// Markus Rechberger <mrechberger@gmail.com>
7		−// Mauro Carvalho Chehab <mchehab@infradead.org>
	7	+// Mauro Carvalho Chehab <mchehab@kernel.org>
8	8	// Sascha Sommer <saschasommer@freenet.de>
9	9	// Copyright (C) 2012 Frank Schäfer <fschaefer.oss@googlemail.com>
10	10	//
@@ -32,7 +32,7 @@ · 21 unmodified lines
32	32
33	33	#define DRIVER_AUTHOR "Ludovico Cavedon <cavedon@sssup.it>, " \
34	34	"Markus Rechberger <mrechberger@gmail.com>, " \
35		− "Mauro Carvalho Chehab <mchehab@infradead.org>, " \
	35	+ "Mauro Carvalho Chehab <mchehab@kernel.org>, " \
36	36	"Sascha Sommer <saschasommer@freenet.de>"
37	37
38	38	MODULE_AUTHOR(DRIVER_AUTHOR);

drivers/media/usb/em28xx/em28xx-dvb.c+2−2

@@ -2,7 +2,7 @@ · 1 unmodified line
2	2	//
3	3	// DVB device driver for em28xx
4	4	//
5		−// (c) 2008-2011 Mauro Carvalho Chehab <mchehab@infradead.org>
	5	+// (c) 2008-2011 Mauro Carvalho Chehab <mchehab@kernel.org>
6	6	//
7	7	// (c) 2008 Devin Heitmueller <devin.heitmueller@gmail.com>
8	8	// - Fixes for the driver to properly work with HVR-950
@@ -63,7 +63,7 @@ · 54 unmodified lines
63	63	#include "tc90522.h"
64	64	#include "qm1d1c0042.h"
65	65
66		−MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@infradead.org>");
	66	+MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@kernel.org>");
67	67	MODULE_LICENSE("GPL v2");
68	68	MODULE_DESCRIPTION(DRIVER_DESC " - digital TV interface");
69	69	MODULE_VERSION(EM28XX_VERSION);

drivers/media/usb/em28xx/em28xx-i2c.c+1−1

@@ -4,7 +4,7 @@ · 3 unmodified lines
4	4	//
5	5	// Copyright (C) 2005 Ludovico Cavedon <cavedon@sssup.it>
6	6	// Markus Rechberger <mrechberger@gmail.com>
7		−// Mauro Carvalho Chehab <mchehab@infradead.org>
	7	+// Mauro Carvalho Chehab <mchehab@kernel.org>
8	8	// Sascha Sommer <saschasommer@freenet.de>
9	9	// Copyright (C) 2013 Frank Schäfer <fschaefer.oss@googlemail.com>
10	10	//

drivers/media/usb/em28xx/em28xx-input.c+1−1

@@ -4,7 +4,7 @@ · 3 unmodified lines
4	4	//
5	5	// Copyright (C) 2005 Ludovico Cavedon <cavedon@sssup.it>
6	6	// Markus Rechberger <mrechberger@gmail.com>
7		−// Mauro Carvalho Chehab <mchehab@infradead.org>
	7	+// Mauro Carvalho Chehab <mchehab@kernel.org>
8	8	// Sascha Sommer <saschasommer@freenet.de>
9	9	//
10	10	// This program is free software; you can redistribute it and/or modify

drivers/media/usb/em28xx/em28xx-video.c+2−2

@@ -5,7 +5,7 @@ · 4 unmodified lines
5	5	//
6	6	// Copyright (C) 2005 Ludovico Cavedon <cavedon@sssup.it>
7	7	// Markus Rechberger <mrechberger@gmail.com>
8		−// Mauro Carvalho Chehab <mchehab@infradead.org>
	8	+// Mauro Carvalho Chehab <mchehab@kernel.org>
9	9	// Sascha Sommer <saschasommer@freenet.de>
10	10	// Copyright (C) 2012 Frank Schäfer <fschaefer.oss@googlemail.com>
11	11	//
@@ -44,7 +44,7 @@ · 32 unmodified lines
44	44
45	45	#define DRIVER_AUTHOR "Ludovico Cavedon <cavedon@sssup.it>, " \
46	46	"Markus Rechberger <mrechberger@gmail.com>, " \
47		− "Mauro Carvalho Chehab <mchehab@infradead.org>, " \
	47	+ "Mauro Carvalho Chehab <mchehab@kernel.org>, " \
48	48	"Sascha Sommer <saschasommer@freenet.de>"
49	49
50	50	static unsigned int isoc_debug;

drivers/media/usb/em28xx/em28xx.h+1−1

@@ -4,7 +4,7 @@ · 3 unmodified lines
4	4	*
5	5	* Copyright (C) 2005 Markus Rechberger <mrechberger@gmail.com>
6	6	* Ludovico Cavedon <cavedon@sssup.it>
7		− * Mauro Carvalho Chehab <mchehab@infradead.org>
	7	+ * Mauro Carvalho Chehab <mchehab@kernel.org>
8	8	* Copyright (C) 2012 Frank Schäfer <fschaefer.oss@googlemail.com>
9	9	*
10	10	* Based on the em2800 driver from Sascha Sommer <saschasommer@freenet.de>

drivers/media/usb/gspca/zc3xx-reg.h+1−1

@@ -1,7 +1,7 @@
1	1	/*
2	2	* zc030x registers
3	3	*
4		− * Copyright (c) 2008 Mauro Carvalho Chehab <mchehab@infradead.org>
	4	+ * Copyright (c) 2008 Mauro Carvalho Chehab <mchehab@kernel.org>
5	5	*
6	6	* The register aliases used here came from this driver:
7	7	* http://zc0302.sourceforge.net/zc0302.php

drivers/media/usb/tm6000/tm6000-cards.c+1−1

@@ -1,7 +1,7 @@
1	1	// SPDX-License-Identifier: GPL-2.0
2	2	// tm6000-cards.c - driver for TM5600/TM6000/TM6010 USB video capture devices
3	3	//
4		−// Copyright (c) 2006-2007 Mauro Carvalho Chehab <mchehab@infradead.org>
	4	+// Copyright (c) 2006-2007 Mauro Carvalho Chehab <mchehab@kernel.org>
5	5
6	6	#include <linux/init.h>
7	7	#include <linux/module.h>

drivers/media/usb/tm6000/tm6000-core.c+1−1

@@ -1,7 +1,7 @@
1	1	// SPDX-License-Identifier: GPL-2.0
2	2	// tm6000-core.c - driver for TM5600/TM6000/TM6010 USB video capture devices
3	3	//
4		−// Copyright (c) 2006-2007 Mauro Carvalho Chehab <mchehab@infradead.org>
	4	+// Copyright (c) 2006-2007 Mauro Carvalho Chehab <mchehab@kernel.org>
5	5	//
6	6	// Copyright (c) 2007 Michel Ludwig <michel.ludwig@gmail.com>
7	7	// - DVB-T support

drivers/media/usb/tm6000/tm6000-i2c.c+1−1

@@ -1,7 +1,7 @@
1	1	// SPDX-License-Identifier: GPL-2.0
2	2	// tm6000-i2c.c - driver for TM5600/TM6000/TM6010 USB video capture devices
3	3	//
4		−// Copyright (c) 2006-2007 Mauro Carvalho Chehab <mchehab@infradead.org>
	4	+// Copyright (c) 2006-2007 Mauro Carvalho Chehab <mchehab@kernel.org>
5	5	//
6	6	// Copyright (c) 2007 Michel Ludwig <michel.ludwig@gmail.com>
7	7	// - Fix SMBus Read Byte command

drivers/media/usb/tm6000/tm6000-regs.h+1−1

@@ -2,7 +2,7 @@ · 1 unmodified line
2	2	* SPDX-License-Identifier: GPL-2.0
3	3	* tm6000-regs.h - driver for TM5600/TM6000/TM6010 USB video capture devices
4	4	*
5		− * Copyright (c) 2006-2007 Mauro Carvalho Chehab <mchehab@infradead.org>
	5	+ * Copyright (c) 2006-2007 Mauro Carvalho Chehab <mchehab@kernel.org>
6	6	*/
7	7
8	8	/*

drivers/media/usb/tm6000/tm6000-usb-isoc.h+1−1

@@ -2,7 +2,7 @@ · 1 unmodified line
2	2	* SPDX-License-Identifier: GPL-2.0
3	3	* tm6000-buf.c - driver for TM5600/TM6000/TM6010 USB video capture devices
4	4	*
5		− * Copyright (c) 2006-2007 Mauro Carvalho Chehab <mchehab@infradead.org>
	5	+ * Copyright (c) 2006-2007 Mauro Carvalho Chehab <mchehab@kernel.org>
6	6	*/
7	7
8	8	#include <linux/videodev2.h>

drivers/media/usb/tm6000/tm6000-video.c+1−1

@@ -1,7 +1,7 @@
1	1	// SPDX-License-Identifier: GPL-2.0
2	2	// tm6000-video.c - driver for TM5600/TM6000/TM6010 USB video capture devices
3	3	//
4		−// Copyright (c) 2006-2007 Mauro Carvalho Chehab <mchehab@infradead.org>
	4	+// Copyright (c) 2006-2007 Mauro Carvalho Chehab <mchehab@kernel.org>
5	5	//
6	6	// Copyright (c) 2007 Michel Ludwig <michel.ludwig@gmail.com>
7	7	// - Fixed module load/unload

drivers/media/usb/tm6000/tm6000.h+1−1

@@ -2,7 +2,7 @@ · 1 unmodified line
2	2	* SPDX-License-Identifier: GPL-2.0
3	3	* tm6000.h - driver for TM5600/TM6000/TM6010 USB video capture devices
4	4	*
5		− * Copyright (c) 2006-2007 Mauro Carvalho Chehab <mchehab@infradead.org>
	5	+ * Copyright (c) 2006-2007 Mauro Carvalho Chehab <mchehab@kernel.org>
6	6	*
7	7	* Copyright (c) 2007 Michel Ludwig <michel.ludwig@gmail.com>
8	8	* - DVB-T support

drivers/media/v4l2-core/v4l2-dev.c+2−2

@@ -10,7 +10,7 @@ · 9 unmodified lines
10	10	* 2 of the License, or (at your option) any later version.
11	11	*
12	12	* Authors: Alan Cox, <alan@lxorguk.ukuu.org.uk> (version 1)
13		− * Mauro Carvalho Chehab <mchehab@infradead.org> (version 2)
	13	+ * Mauro Carvalho Chehab <mchehab@kernel.org> (version 2)
14	14	*
15	15	* Fixes: 20000516 Claudio Matsuoka <claudio@conectiva.com>
16	16	* - Added procfs support
@@ -1079,7 +1079,7 @@ static void __exit videodev_exit(void) · 1062 unmodified lines
1079	1079	subsys_initcall(videodev_init);
1080	1080	module_exit(videodev_exit)
1081	1081
1082		−MODULE_AUTHOR("Alan Cox, Mauro Carvalho Chehab <mchehab@infradead.org>");
	1082	+MODULE_AUTHOR("Alan Cox, Mauro Carvalho Chehab <mchehab@kernel.org>");
1083	1083	MODULE_DESCRIPTION("Device registrar for Video4Linux drivers v2");
1084	1084	MODULE_LICENSE("GPL");
1085	1085	MODULE_ALIAS_CHARDEV_MAJOR(VIDEO_MAJOR);

drivers/media/v4l2-core/v4l2-ioctl.c+1−1

@@ -9,7 +9,7 @@ · 8 unmodified lines
9	9	* 2 of the License, or (at your option) any later version.
10	10	*
11	11	* Authors: Alan Cox, <alan@lxorguk.ukuu.org.uk> (version 1)
12		− * Mauro Carvalho Chehab <mchehab@infradead.org> (version 2)
	12	+ * Mauro Carvalho Chehab <mchehab@kernel.org> (version 2)
13	13	*/
14	14
15	15	#include <linux/mm.h>

drivers/media/v4l2-core/videobuf-core.c+3−3

@@ -1,11 +1,11 @@
1	1	/*
2	2	* generic helper functions for handling video4linux capture buffers
3	3	*
4		− * (c) 2007 Mauro Carvalho Chehab, <mchehab@infradead.org>
	4	+ * (c) 2007 Mauro Carvalho Chehab, <mchehab@kernel.org>
5	5	*
6	6	* Highly based on video-buf written originally by:
7	7	* (c) 2001,02 Gerd Knorr <kraxel@bytesex.org>
8		− * (c) 2006 Mauro Carvalho Chehab, <mchehab@infradead.org>
	8	+ * (c) 2006 Mauro Carvalho Chehab, <mchehab@kernel.org>
9	9	* (c) 2006 Ted Walther and John Sokol
10	10	*
11	11	* This program is free software; you can redistribute it and/or modify
@@ -38,7 +38,7 @@ static int debug; · 26 unmodified lines
38	38	module_param(debug, int, 0644);
39	39
40	40	MODULE_DESCRIPTION("helper module to manage video4linux buffers");
41		−MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@infradead.org>");
	41	+MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@kernel.org>");
42	42	MODULE_LICENSE("GPL");
43	43
44	44	#define dprintk(level, fmt, arg...) \

drivers/media/v4l2-core/videobuf-dma-contig.c+1−1

@@ -7,7 +7,7 @@ · 6 unmodified lines
7	7	* Copyright (c) 2008 Magnus Damm
8	8	*
9	9	* Based on videobuf-vmalloc.c,
10		− * (c) 2007 Mauro Carvalho Chehab, <mchehab@infradead.org>
	10	+ * (c) 2007 Mauro Carvalho Chehab, <mchehab@kernel.org>
11	11	*
12	12	* This program is free software; you can redistribute it and/or modify
13	13	* it under the terms of the GNU General Public License as published by

drivers/media/v4l2-core/videobuf-dma-sg.c+3−3

@@ -6,11 +6,11 @@ · 5 unmodified lines
6	6	* into PAGE_SIZE chunks). They also assume the driver does not need
7	7	* to touch the video data.
8	8	*
9		− * (c) 2007 Mauro Carvalho Chehab, <mchehab@infradead.org>
	9	+ * (c) 2007 Mauro Carvalho Chehab, <mchehab@kernel.org>
10	10	*
11	11	* Highly based on video-buf written originally by:
12	12	* (c) 2001,02 Gerd Knorr <kraxel@bytesex.org>
13		− * (c) 2006 Mauro Carvalho Chehab, <mchehab@infradead.org>
	13	+ * (c) 2006 Mauro Carvalho Chehab, <mchehab@kernel.org>
14	14	* (c) 2006 Ted Walther and John Sokol
15	15	*
16	16	* This program is free software; you can redistribute it and/or modify
@@ -48,7 +48,7 @@ static int debug; · 31 unmodified lines
48	48	module_param(debug, int, 0644);
49	49
50	50	MODULE_DESCRIPTION("helper module to manage video4linux dma sg buffers");
51		−MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@infradead.org>");
	51	+MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@kernel.org>");
52	52	MODULE_LICENSE("GPL");
53	53
54	54	#define dprintk(level, fmt, arg...) \

drivers/media/v4l2-core/videobuf-vmalloc.c+2−2

@@ -6,7 +6,7 @@ · 5 unmodified lines
6	6	* into PAGE_SIZE chunks). They also assume the driver does not need
7	7	* to touch the video data.
8	8	*
9		− * (c) 2007 Mauro Carvalho Chehab, <mchehab@infradead.org>
	9	+ * (c) 2007 Mauro Carvalho Chehab, <mchehab@kernel.org>
10	10	*
11	11	* This program is free software; you can redistribute it and/or modify
12	12	* it under the terms of the GNU General Public License as published by
@@ -41,7 +41,7 @@ static int debug; · 28 unmodified lines
41	41	module_param(debug, int, 0644);
42	42
43	43	MODULE_DESCRIPTION("helper module to manage video4linux vmalloc buffers");
44		−MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@infradead.org>");
	44	+MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@kernel.org>");
45	45	MODULE_LICENSE("GPL");
46	46
47	47	#define dprintk(level, fmt, arg...) \

drivers/memory/emif-asm-offsets.c+1−71

@@ -16,77 +16,7 @@ · 15 unmodified lines
16	16
17	17	int main(void)
18	18	{
19		− DEFINE(EMIF_SDCFG_VAL_OFFSET,
20		− offsetof(struct emif_regs_amx3, emif_sdcfg_val));
21		− DEFINE(EMIF_TIMING1_VAL_OFFSET,
22		− offsetof(struct emif_regs_amx3, emif_timing1_val));
23		− DEFINE(EMIF_TIMING2_VAL_OFFSET,
24		− offsetof(struct emif_regs_amx3, emif_timing2_val));
25		− DEFINE(EMIF_TIMING3_VAL_OFFSET,
26		− offsetof(struct emif_regs_amx3, emif_timing3_val));
27		− DEFINE(EMIF_REF_CTRL_VAL_OFFSET,
28		− offsetof(struct emif_regs_amx3, emif_ref_ctrl_val));
29		− DEFINE(EMIF_ZQCFG_VAL_OFFSET,
30		− offsetof(struct emif_regs_amx3, emif_zqcfg_val));
31		− DEFINE(EMIF_PMCR_VAL_OFFSET,
32		− offsetof(struct emif_regs_amx3, emif_pmcr_val));
33		− DEFINE(EMIF_PMCR_SHDW_VAL_OFFSET,
34		− offsetof(struct emif_regs_amx3, emif_pmcr_shdw_val));
35		− DEFINE(EMIF_RD_WR_LEVEL_RAMP_CTRL_OFFSET,
36		− offsetof(struct emif_regs_amx3, emif_rd_wr_level_ramp_ctrl));
37		− DEFINE(EMIF_RD_WR_EXEC_THRESH_OFFSET,
38		− offsetof(struct emif_regs_amx3, emif_rd_wr_exec_thresh));
39		− DEFINE(EMIF_COS_CONFIG_OFFSET,
40		− offsetof(struct emif_regs_amx3, emif_cos_config));
41		− DEFINE(EMIF_PRIORITY_TO_COS_MAPPING_OFFSET,
42		− offsetof(struct emif_regs_amx3, emif_priority_to_cos_mapping));
43		− DEFINE(EMIF_CONNECT_ID_SERV_1_MAP_OFFSET,
44		− offsetof(struct emif_regs_amx3, emif_connect_id_serv_1_map));
45		− DEFINE(EMIF_CONNECT_ID_SERV_2_MAP_OFFSET,
46		− offsetof(struct emif_regs_amx3, emif_connect_id_serv_2_map));
47		− DEFINE(EMIF_OCP_CONFIG_VAL_OFFSET,
48		− offsetof(struct emif_regs_amx3, emif_ocp_config_val));
49		− DEFINE(EMIF_LPDDR2_NVM_TIM_OFFSET,
50		− offsetof(struct emif_regs_amx3, emif_lpddr2_nvm_tim));
51		− DEFINE(EMIF_LPDDR2_NVM_TIM_SHDW_OFFSET,
52		− offsetof(struct emif_regs_amx3, emif_lpddr2_nvm_tim_shdw));
53		− DEFINE(EMIF_DLL_CALIB_CTRL_VAL_OFFSET,
54		− offsetof(struct emif_regs_amx3, emif_dll_calib_ctrl_val));
55		− DEFINE(EMIF_DLL_CALIB_CTRL_VAL_SHDW_OFFSET,
56		− offsetof(struct emif_regs_amx3, emif_dll_calib_ctrl_val_shdw));
57		− DEFINE(EMIF_DDR_PHY_CTLR_1_OFFSET,
58		− offsetof(struct emif_regs_amx3, emif_ddr_phy_ctlr_1));
59		− DEFINE(EMIF_EXT_PHY_CTRL_VALS_OFFSET,
60		− offsetof(struct emif_regs_amx3, emif_ext_phy_ctrl_vals));
61		− DEFINE(EMIF_REGS_AMX3_SIZE, sizeof(struct emif_regs_amx3));
62		−
63		− BLANK();
64		−
65		− DEFINE(EMIF_PM_BASE_ADDR_VIRT_OFFSET,
66		− offsetof(struct ti_emif_pm_data, ti_emif_base_addr_virt));
67		− DEFINE(EMIF_PM_BASE_ADDR_PHYS_OFFSET,
68		− offsetof(struct ti_emif_pm_data, ti_emif_base_addr_phys));
69		− DEFINE(EMIF_PM_CONFIG_OFFSET,
70		− offsetof(struct ti_emif_pm_data, ti_emif_sram_config));
71		− DEFINE(EMIF_PM_REGS_VIRT_OFFSET,
72		− offsetof(struct ti_emif_pm_data, regs_virt));
73		− DEFINE(EMIF_PM_REGS_PHYS_OFFSET,
74		− offsetof(struct ti_emif_pm_data, regs_phys));
75		− DEFINE(EMIF_PM_DATA_SIZE, sizeof(struct ti_emif_pm_data));
76		−
77		− BLANK();
78		−
79		− DEFINE(EMIF_PM_SAVE_CONTEXT_OFFSET,
80		− offsetof(struct ti_emif_pm_functions, save_context));
81		− DEFINE(EMIF_PM_RESTORE_CONTEXT_OFFSET,
82		− offsetof(struct ti_emif_pm_functions, restore_context));
83		− DEFINE(EMIF_PM_ENTER_SR_OFFSET,
84		− offsetof(struct ti_emif_pm_functions, enter_sr));
85		− DEFINE(EMIF_PM_EXIT_SR_OFFSET,
86		− offsetof(struct ti_emif_pm_functions, exit_sr));
87		− DEFINE(EMIF_PM_ABORT_SR_OFFSET,
88		− offsetof(struct ti_emif_pm_functions, abort_sr));
89		− DEFINE(EMIF_PM_FUNCTIONS_SIZE, sizeof(struct ti_emif_pm_functions));
	19	+ ti_emif_asm_offsets();
90	20
91	21	return 0;
92	22	}

drivers/message/fusion/mptsas.c+1−0

@@ -1994,6 +1994,7 @@ static struct scsi_host_template mptsas_driver_template = { · 1993 unmodified lines
1994	1994	.cmd_per_lun = 7,
1995	1995	.use_clustering = ENABLE_CLUSTERING,
1996	1996	.shost_attrs = mptscsih_host_attrs,
	1997	+ .no_write_same = 1,
1997	1998	};
1998	1999
1999	2000	static int mptsas_get_linkerrors(struct sas_phy *phy)

drivers/mmc/host/renesas_sdhi_internal_dmac.c+33−6

@@ -9,6 +9,7 @@ · 8 unmodified lines
9	9	* published by the Free Software Foundation.
10	10	*/
11	11
	12	+#include <linux/bitops.h>
12	13	#include <linux/device.h>
13	14	#include <linux/dma-mapping.h>
14	15	#include <linux/io-64-nonatomic-hi-lo.h>
@@ -62,6 +63,17 @@ · 47 unmodified lines
62	63	* need a custom accessor.
63	64	*/
64	65
	66	+static unsigned long global_flags;
	67	+/*
	68	+ * Workaround for avoiding to use RX DMAC by multiple channels.
	69	+ * On R-Car H3 ES1.* and M3-W ES1.0, when multiple SDHI channels use
	70	+ * RX DMAC simultaneously, sometimes hundreds of bytes data are not
	71	+ * stored into the system memory even if the DMAC interrupt happened.
	72	+ * So, this driver then uses one RX DMAC channel only.
	73	+ */
	74	+#define SDHI_INTERNAL_DMAC_ONE_RX_ONLY 0
	75	+#define SDHI_INTERNAL_DMAC_RX_IN_USE 1
	76	+
65	77	/* Definitions for sampling clocks */
66	78	static struct renesas_sdhi_scc rcar_gen3_scc_taps[] = {
67	79	{
@@ -126,6 +138,9 @@ renesas_sdhi_internal_dmac_abort_dma(struct tmio_mmc_host *host) { · 58 unmodified lines
126	138	renesas_sdhi_internal_dmac_dm_write(host, DM_CM_RST,
127	139	RST_RESERVED_BITS \| val);
128	140
	141	+ if (host->data && host->data->flags & MMC_DATA_READ)
	142	+ clear_bit(SDHI_INTERNAL_DMAC_RX_IN_USE, &global_flags);
	143	+
129	144	renesas_sdhi_internal_dmac_enable_dma(host, true);
130	145	}
131	146
@@ -155,6 +170,9 @@ renesas_sdhi_internal_dmac_start_dma(struct tmio_mmc_host *host, · 23 unmodified lines
155	170	if (data->flags & MMC_DATA_READ) {
156	171	dtran_mode \|= DTRAN_MODE_CH_NUM_CH1;
157	172	dir = DMA_FROM_DEVICE;
	173	+ if (test_bit(SDHI_INTERNAL_DMAC_ONE_RX_ONLY, &global_flags) &&
	174	+ test_and_set_bit(SDHI_INTERNAL_DMAC_RX_IN_USE, &global_flags))
	175	+ goto force_pio;
158	176	} else {
159	177	dtran_mode \|= DTRAN_MODE_CH_NUM_CH0;
160	178	dir = DMA_TO_DEVICE;
@@ -208,6 +226,9 @@ static void renesas_sdhi_internal_dmac_complete_tasklet_fn(unsigned long arg) · 47 unmodified lines
208	226	renesas_sdhi_internal_dmac_enable_dma(host, false);
209	227	dma_unmap_sg(&host->pdev->dev, host->sg_ptr, host->sg_len, dir);
210	228
	229	+ if (dir == DMA_FROM_DEVICE)
	230	+ clear_bit(SDHI_INTERNAL_DMAC_RX_IN_USE, &global_flags);
	231	+
211	232	tmio_mmc_do_data_irq(host);
212	233	out:
213	234	spin_unlock_irq(&host->lock);
@@ -251,18 +272,24 @@ static const struct tmio_mmc_dma_ops renesas_sdhi_internal_dmac_dma_ops = { · 37 unmodified lines
251	272	* implementation as others may use a different implementation.
252	273	*/
253	274	static const struct soc_device_attribute gen3_soc_whitelist[] = {
254		− { .soc_id = "r8a7795", .revision = "ES1.*" },
255		− { .soc_id = "r8a7795", .revision = "ES2.0" },
256		− { .soc_id = "r8a7796", .revision = "ES1.0" },
257		− { .soc_id = "r8a77995", .revision = "ES1.0" },
258		− { /* sentinel */ }
	275	+ { .soc_id = "r8a7795", .revision = "ES1.*",
	276	+ .data = (void *)BIT(SDHI_INTERNAL_DMAC_ONE_RX_ONLY) },
	277	+ { .soc_id = "r8a7795", .revision = "ES2.0" },
	278	+ { .soc_id = "r8a7796", .revision = "ES1.0",
	279	+ .data = (void *)BIT(SDHI_INTERNAL_DMAC_ONE_RX_ONLY) },
	280	+ { .soc_id = "r8a77995", .revision = "ES1.0" },
	281	+ { /* sentinel */ }
259	282	};
260	283
261	284	static int renesas_sdhi_internal_dmac_probe(struct platform_device *pdev)
262	285	{
263		− if (!soc_device_match(gen3_soc_whitelist))
	286	+ const struct soc_device_attribute *soc = soc_device_match(gen3_soc_whitelist);
	287	+
	288	+ if (!soc)
264	289	return -ENODEV;
265	290
	291	+ global_flags \|= (unsigned long)soc->data;
	292	+
266	293	return renesas_sdhi_probe(pdev, &renesas_sdhi_internal_dmac_dma_ops);
267	294	}
268	295

drivers/mmc/host/sdhci-pci-core.c+23−2

@@ -1312,7 +1312,7 @@ static void amd_enable_manual_tuning(struct pci_dev *pdev) · 1311 unmodified lines
1312	1312	pci_write_config_dword(pdev, AMD_SD_MISC_CONTROL, val);
1313	1313	}
1314	1314
1315		−static int amd_execute_tuning(struct sdhci_host *host, u32 opcode)
	1315	+static int amd_execute_tuning_hs200(struct sdhci_host *host, u32 opcode)
1316	1316	{
1317	1317	struct sdhci_pci_slot *slot = sdhci_priv(host);
1318	1318	struct pci_dev *pdev = slot->chip->pdev;
@@ -1351,6 +1351,27 @@ static int amd_execute_tuning(struct sdhci_host *host, u32 opcode) · 32 unmodified lines
1351	1351	return 0;
1352	1352	}
1353	1353
	1354	+static int amd_execute_tuning(struct mmc_host *mmc, u32 opcode)
	1355	+{
	1356	+ struct sdhci_host *host = mmc_priv(mmc);
	1357	+
	1358	+ /* AMD requires custom HS200 tuning */
	1359	+ if (host->timing == MMC_TIMING_MMC_HS200)
	1360	+ return amd_execute_tuning_hs200(host, opcode);
	1361	+
	1362	+ /* Otherwise perform standard SDHCI tuning */
	1363	+ return sdhci_execute_tuning(mmc, opcode);
	1364	+}
	1365	+
	1366	+static int amd_probe_slot(struct sdhci_pci_slot *slot)
	1367	+{
	1368	+ struct mmc_host_ops *ops = &slot->host->mmc_host_ops;
	1369	+
	1370	+ ops->execute_tuning = amd_execute_tuning;
	1371	+
	1372	+ return 0;
	1373	+}
	1374	+
1354	1375	static int amd_probe(struct sdhci_pci_chip *chip)
1355	1376	{
1356	1377	struct pci_dev *smbus_dev;
@@ -1385,12 +1406,12 @@ static const struct sdhci_ops amd_sdhci_pci_ops = { · 28 unmodified lines
1385	1406	.set_bus_width = sdhci_set_bus_width,
1386	1407	.reset = sdhci_reset,
1387	1408	.set_uhs_signaling = sdhci_set_uhs_signaling,
1388		− .platform_execute_tuning = amd_execute_tuning,
1389	1409	};
1390	1410
1391	1411	static const struct sdhci_pci_fixes sdhci_amd = {
1392	1412	.probe = amd_probe,
1393	1413	.ops = &amd_sdhci_pci_ops,
	1414	+ .probe_slot = amd_probe_slot,
1394	1415	};
1395	1416
1396	1417	static const struct pci_device_id pci_ids[] = {

drivers/mtd/chips/cfi_cmdset_0001.c+28−5

@@ -45,6 +45,7 @@ · 44 unmodified lines
45	45	#define I82802AB 0x00ad
46	46	#define I82802AC 0x00ac
47	47	#define PF38F4476 0x881c
	48	+#define M28F00AP30 0x8963
48	49	/* STMicroelectronics chips */
49	50	#define M50LPW080 0x002F
50	51	#define M50FLW080A 0x0080
@@ -375,6 +376,17 @@ static void cfi_fixup_major_minor(struct cfi_private *cfi, · 324 unmodified lines
375	376	extp->MinorVersion = '1';
376	377	}
377	378
	379	+static int cfi_is_micron_28F00AP30(struct cfi_private cfi, struct flchip chip)
	380	+{
	381	+ /*
	382	+ * Micron(was Numonyx) 1Gbit bottom boot are buggy w.r.t
	383	+ * Erase Supend for their small Erase Blocks(0x8000)
	384	+ */
	385	+ if (cfi->mfr == CFI_MFR_INTEL && cfi->id == M28F00AP30)
	386	+ return 1;
	387	+ return 0;
	388	+}
	389	+
378	390	static inline struct cfi_pri_intelext *
379	391	read_pri_intelext(struct map_info *map, __u16 adr)
380	392	{
@@ -831,21 +843,30 @@ static int chip_ready (struct map_info map, struct flchip chip, unsigned long · 450 unmodified lines
831	843	(mode == FL_WRITING && (cfip->SuspendCmdSupport & 1))))
832	844	goto sleep;
833	845
	846	+ /* Do not allow suspend iff read/write to EB address */
	847	+ if ((adr & chip->in_progress_block_mask) ==
	848	+ chip->in_progress_block_addr)
	849	+ goto sleep;
	850	+
	851	+ /* do not suspend small EBs, buggy Micron Chips */
	852	+ if (cfi_is_micron_28F00AP30(cfi, chip) &&
	853	+ (chip->in_progress_block_mask == ~(0x8000-1)))
	854	+ goto sleep;
834	855
835	856	/* Erase suspend */
836		− map_write(map, CMD(0xB0), adr);
	857	+ map_write(map, CMD(0xB0), chip->in_progress_block_addr);
837	858
838	859	/* If the flash has finished erasing, then 'erase suspend'
839	860	* appears to make some (28F320) flash devices switch to
840	861	* 'read' mode. Make sure that we switch to 'read status'
841	862	* mode so we get the right data. --rmk
842	863	*/
843		− map_write(map, CMD(0x70), adr);
	864	+ map_write(map, CMD(0x70), chip->in_progress_block_addr);
844	865	chip->oldstate = FL_ERASING;
845	866	chip->state = FL_ERASE_SUSPENDING;
846	867	chip->erase_suspended = 1;
847	868	for (;;) {
848		− status = map_read(map, adr);
	869	+ status = map_read(map, chip->in_progress_block_addr);
849	870	if (map_word_andequal(map, status, status_OK, status_OK))
850	871	break;
851	872
@@ -1041,8 +1062,8 @@ static void put_chip(struct map_info map, struct flchip chip, unsigned long ad · 189 unmodified lines
1041	1062	sending the 0x70 (Read Status) command to an erasing
1042	1063	chip and expecting it to be ignored, that's what we
1043	1064	do. */
1044		− map_write(map, CMD(0xd0), adr);
1045		− map_write(map, CMD(0x70), adr);
	1065	+ map_write(map, CMD(0xd0), chip->in_progress_block_addr);
	1066	+ map_write(map, CMD(0x70), chip->in_progress_block_addr);
1046	1067	chip->oldstate = FL_READY;
1047	1068	chip->state = FL_ERASING;
1048	1069	break;
@@ -1933,6 +1954,8 @@ static int __xipram do_erase_oneblock(struct map_info map, struct flchip chip, · 884 unmodified lines
1933	1954	map_write(map, CMD(0xD0), adr);
1934	1955	chip->state = FL_ERASING;
1935	1956	chip->erase_suspended = 0;
	1957	+ chip->in_progress_block_addr = adr;
	1958	+ chip->in_progress_block_mask = ~(len - 1);
1936	1959
1937	1960	ret = INVAL_CACHE_AND_WAIT(map, chip, adr,
1938	1961	adr, len,

drivers/mtd/chips/cfi_cmdset_0002.c+6−3

@@ -816,9 +816,10 @@ static int get_chip(struct map_info map, struct flchip chip, unsigned long adr · 815 unmodified lines
816	816	(mode == FL_WRITING && (cfip->EraseSuspend & 0x2))))
817	817	goto sleep;
818	818
819		− /* We could check to see if we're trying to access the sector
820		− * that is currently being erased. However, no user will try
821		− * anything like that so we just wait for the timeout. */
	819	+ /* Do not allow suspend iff read/write to EB address */
	820	+ if ((adr & chip->in_progress_block_mask) ==
	821	+ chip->in_progress_block_addr)
	822	+ goto sleep;
822	823
823	824	/* Erase suspend */
824	825	/* It's harmless to issue the Erase-Suspend and Erase-Resume
@@ -2267,6 +2268,7 @@ static int __xipram do_erase_chip(struct map_info map, struct flchip chip) · 1442 unmodified lines
2267	2268	chip->state = FL_ERASING;
2268	2269	chip->erase_suspended = 0;
2269	2270	chip->in_progress_block_addr = adr;
	2271	+ chip->in_progress_block_mask = ~(map->size - 1);
2270	2272
2271	2273	INVALIDATE_CACHE_UDELAY(map, chip,
2272	2274	adr, map->size,
@@ -2356,6 +2358,7 @@ static int __xipram do_erase_oneblock(struct map_info map, struct flchip chip, · 83 unmodified lines
2356	2358	chip->state = FL_ERASING;
2357	2359	chip->erase_suspended = 0;
2358	2360	chip->in_progress_block_addr = adr;
	2361	+ chip->in_progress_block_mask = ~(len - 1);
2359	2362
2360	2363	INVALIDATE_CACHE_UDELAY(map, chip,
2361	2364	adr, len,

drivers/mtd/nand/core.c+0−3

@@ -162,7 +162,6 @@ int nanddev_mtd_erase(struct mtd_info mtd, struct erase_info einfo) · 161 unmodified lines
162	162	ret = nanddev_erase(nand, &pos);
163	163	if (ret) {
164	164	einfo->fail_addr = nanddev_pos_to_offs(nand, &pos);
165		− einfo->state = MTD_ERASE_FAILED;
166	165
167	166	return ret;
168	167	}
@@ -170,8 +169,6 @@ int nanddev_mtd_erase(struct mtd_info mtd, struct erase_info einfo) · 1 unmodified line
170	169	nanddev_pos_next_eraseblock(nand, &pos);
171	170	}
172	171
173		− einfo->state = MTD_ERASE_DONE;
174		−
175	172	return 0;
176	173	}
177	174	EXPORT_SYMBOL_GPL(nanddev_mtd_erase);

drivers/mtd/nand/raw/marvell_nand.c+8−17

@@ -2299,29 +2299,20 @@ static int marvell_nand_chip_init(struct device dev, struct marvell_nfc nfc, · 2298 unmodified lines
2299	2299	/*
2300	2300	* The legacy "num-cs" property indicates the number of CS on the only
2301	2301	* chip connected to the controller (legacy bindings does not support
2302		− * more than one chip). CS are only incremented one by one while the RB
2303		− * pin is always the #0.
	2302	+ * more than one chip). The CS and RB pins are always the #0.
2304	2303	*
2305	2304	* When not using legacy bindings, a couple of "reg" and "nand-rb"
2306	2305	* properties must be filled. For each chip, expressed as a subnode,
2307	2306	* "reg" points to the CS lines and "nand-rb" to the RB line.
2308	2307	*/
2309		− if (pdata) {
	2308	+ if (pdata \|\| nfc->caps->legacy_of_bindings) {
2310	2309	nsels = 1;
2311		− } else if (nfc->caps->legacy_of_bindings &&
2312		− !of_get_property(np, "num-cs", &nsels)) {
2313		− dev_err(dev, "missing num-cs property\n");
2314		− return -EINVAL;
2315		− } else if (!of_get_property(np, "reg", &nsels)) {
2316		− dev_err(dev, "missing reg property\n");
2317		− return -EINVAL;
2318		− }
2319		−
2320		− if (!pdata)
2321		− nsels /= sizeof(u32);
2322		− if (!nsels) {
2323		− dev_err(dev, "invalid reg property size\n");
2324		− return -EINVAL;
	2310	+ } else {
	2311	+ nsels = of_property_count_elems_of_size(np, "reg", sizeof(u32));
	2312	+ if (nsels <= 0) {
	2313	+ dev_err(dev, "missing/invalid reg property\n");
	2314	+ return -EINVAL;
	2315	+ }
2325	2316	}
2326	2317
2327	2318	/* Alloc the nand chip structure */

drivers/mtd/nand/raw/tango_nand.c+1−1

@@ -645,7 +645,7 @@ static int tango_nand_probe(struct platform_device *pdev) · 644 unmodified lines
645	645
646	646	writel_relaxed(MODE_RAW, nfc->pbus_base + PBUS_PAD_MODE);
647	647
648		− clk = clk_get(&pdev->dev, NULL);
	648	+ clk = devm_clk_get(&pdev->dev, NULL);
649	649	if (IS_ERR(clk))
650	650	return PTR_ERR(clk);
651	651

drivers/mtd/spi-nor/cadence-quadspi.c+17−2

@@ -501,7 +501,9 @@ static int cqspi_indirect_read_execute(struct spi_nor nor, u8 rxbuf, · 500 unmodified lines
501	501	void __iomem *reg_base = cqspi->iobase;
502	502	void __iomem *ahb_base = cqspi->ahb_base;
503	503	unsigned int remaining = n_rx;
	504	+ unsigned int mod_bytes = n_rx % 4;
504	505	unsigned int bytes_to_read = 0;
	506	+ u8 *rxbuf_end = rxbuf + n_rx;
505	507	int ret = 0;
506	508
507	509	writel(from_addr, reg_base + CQSPI_REG_INDIRECTRDSTARTADDR);
@@ -530,11 +532,24 @@ static int cqspi_indirect_read_execute(struct spi_nor nor, u8 rxbuf, · 22 unmodified lines
530	532	}
531	533
532	534	while (bytes_to_read != 0) {
	535	+ unsigned int word_remain = round_down(remaining, 4);
	536	+
533	537	bytes_to_read *= cqspi->fifo_width;
534	538	bytes_to_read = bytes_to_read > remaining ?
535	539	remaining : bytes_to_read;
536		− ioread32_rep(ahb_base, rxbuf,
537		− DIV_ROUND_UP(bytes_to_read, 4));
	540	+ bytes_to_read = round_down(bytes_to_read, 4);
	541	+ /* Read 4 byte word chunks then single bytes */
	542	+ if (bytes_to_read) {
	543	+ ioread32_rep(ahb_base, rxbuf,
	544	+ (bytes_to_read / 4));
	545	+ } else if (!word_remain && mod_bytes) {
	546	+ unsigned int temp = ioread32(ahb_base);
	547	+
	548	+ bytes_to_read = mod_bytes;
	549	+ memcpy(rxbuf, &temp, min((unsigned int)
	550	+ (rxbuf_end - rxbuf),
	551	+ bytes_to_read));
	552	+ }
538	553	rxbuf += bytes_to_read;
539	554	remaining -= bytes_to_read;
540	555	bytes_to_read = cqspi_get_rd_sram_level(cqspi);

drivers/net/bonding/bond_main.c+1−2

@@ -1660,8 +1660,7 @@ int bond_enslave(struct net_device bond_dev, struct net_device slave_dev, · 1659 unmodified lines
1660	1660	} /* switch(bond_mode) */
1661	1661
1662	1662	#ifdef CONFIG_NET_POLL_CONTROLLER
1663		− slave_dev->npinfo = bond->dev->npinfo;
1664		− if (slave_dev->npinfo) {
	1663	+ if (bond->dev->npinfo) {
1665	1664	if (slave_enable_netpoll(new_slave)) {
1666	1665	netdev_info(bond_dev, "master_dev is using netpoll, but new slave device does not support netpoll\n");
1667	1666	res = -EBUSY;

drivers/net/dsa/mv88e6xxx/hwtstamp.c+10−2

@@ -285,10 +285,18 @@ static void mv88e6xxx_get_rxts(struct mv88e6xxx_chip *chip, · 284 unmodified lines
285	285	struct sk_buff_head *rxq)
286	286	{
287	287	u16 buf[4] = { 0 }, status, seq_id;
288		− u64 ns, timelo, timehi;
289	288	struct skb_shared_hwtstamps *shwt;
	289	+ struct sk_buff_head received;
	290	+ u64 ns, timelo, timehi;
	291	+ unsigned long flags;
290	292	int err;
291	293
	294	+ /* The latched timestamp belongs to one of the received frames. */
	295	+ __skb_queue_head_init(&received);
	296	+ spin_lock_irqsave(&rxq->lock, flags);
	297	+ skb_queue_splice_tail_init(rxq, &received);
	298	+ spin_unlock_irqrestore(&rxq->lock, flags);
	299	+
292	300	mutex_lock(&chip->reg_lock);
293	301	err = mv88e6xxx_port_ptp_read(chip, ps->port_id,
294	302	reg, buf, ARRAY_SIZE(buf));
@@ -311,7 +319,7 @@ static void mv88e6xxx_get_rxts(struct mv88e6xxx_chip *chip, · 16 unmodified lines
311	319	/* Since the device can only handle one time stamp at a time,
312	320	* we purge any extra frames from the queue.
313	321	*/
314		− for ( ; skb; skb = skb_dequeue(rxq)) {
	322	+ for ( ; skb; skb = __skb_dequeue(&received)) {
315	323	if (mv88e6xxx_ts_valid(status) && seq_match(skb, seq_id)) {
316	324	ns = timehi << 16 \| timelo;
317	325

drivers/net/ethernet/amd/xgbe/xgbe-common.h+8−0

@@ -1321,6 +1321,10 @@ · 1320 unmodified lines
1321	1321	#define MDIO_VEND2_AN_STAT 0x8002
1322	1322	#endif
1323	1323
	1324	+#ifndef MDIO_VEND2_PMA_CDR_CONTROL
	1325	+#define MDIO_VEND2_PMA_CDR_CONTROL 0x8056
	1326	+#endif
	1327	+
1324	1328	#ifndef MDIO_CTRL1_SPEED1G
1325	1329	#define MDIO_CTRL1_SPEED1G (MDIO_CTRL1_SPEED10G & ~BMCR_SPEED100)
1326	1330	#endif
@@ -1369,6 +1373,10 @@ · 42 unmodified lines
1369	1373	#define XGBE_AN_CL37_TX_CONFIG_MASK 0x08
1370	1374	#define XGBE_AN_CL37_MII_CTRL_8BIT 0x0100
1371	1375
	1376	+#define XGBE_PMA_CDR_TRACK_EN_MASK 0x01
	1377	+#define XGBE_PMA_CDR_TRACK_EN_OFF 0x00
	1378	+#define XGBE_PMA_CDR_TRACK_EN_ON 0x01
	1379	+
1372	1380	/* Bit setting and getting macros
1373	1381	* The get macro will extract the current bit field value from within
1374	1382	* the variable

drivers/net/ethernet/amd/xgbe/xgbe-debugfs.c+16−0

@@ -519,6 +519,22 @@ void xgbe_debugfs_init(struct xgbe_prv_data *pdata) · 518 unmodified lines
519	519	"debugfs_create_file failed\n");
520	520	}
521	521
	522	+ if (pdata->vdata->an_cdr_workaround) {
	523	+ pfile = debugfs_create_bool("an_cdr_workaround", 0600,
	524	+ pdata->xgbe_debugfs,
	525	+ &pdata->debugfs_an_cdr_workaround);
	526	+ if (!pfile)
	527	+ netdev_err(pdata->netdev,
	528	+ "debugfs_create_bool failed\n");
	529	+
	530	+ pfile = debugfs_create_bool("an_cdr_track_early", 0600,
	531	+ pdata->xgbe_debugfs,
	532	+ &pdata->debugfs_an_cdr_track_early);
	533	+ if (!pfile)
	534	+ netdev_err(pdata->netdev,
	535	+ "debugfs_create_bool failed\n");
	536	+ }
	537	+
522	538	kfree(buf);
523	539	}
524	540

drivers/net/ethernet/amd/xgbe/xgbe-main.c+1−0

@@ -349,6 +349,7 @@ int xgbe_config_netdev(struct xgbe_prv_data *pdata) · 348 unmodified lines
349	349	XGMAC_SET_BITS(pdata->rss_options, MAC_RSSCR, UDP4TE, 1);
350	350
351	351	/* Call MDIO/PHY initialization routine */
	352	+ pdata->debugfs_an_cdr_workaround = pdata->vdata->an_cdr_workaround;
352	353	ret = pdata->phy_if.phy_init(pdata);
353	354	if (ret)
354	355	return ret;

drivers/net/ethernet/amd/xgbe/xgbe-mdio.c+19−5

@@ -432,11 +432,16 @@ static void xgbe_an73_disable(struct xgbe_prv_data *pdata) · 431 unmodified lines
432	432	xgbe_an73_set(pdata, false, false);
433	433	xgbe_an73_disable_interrupts(pdata);
434	434
	435	+ pdata->an_start = 0;
	436	+
435	437	netif_dbg(pdata, link, pdata->netdev, "CL73 AN disabled\n");
436	438	}
437	439
438	440	static void xgbe_an_restart(struct xgbe_prv_data *pdata)
439	441	{
	442	+ if (pdata->phy_if.phy_impl.an_pre)
	443	+ pdata->phy_if.phy_impl.an_pre(pdata);
	444	+
440	445	switch (pdata->an_mode) {
441	446	case XGBE_AN_MODE_CL73:
442	447	case XGBE_AN_MODE_CL73_REDRV:
@@ -453,6 +458,9 @@ static void xgbe_an_restart(struct xgbe_prv_data *pdata) · 10 unmodified lines
453	458
454	459	static void xgbe_an_disable(struct xgbe_prv_data *pdata)
455	460	{
	461	+ if (pdata->phy_if.phy_impl.an_post)
	462	+ pdata->phy_if.phy_impl.an_post(pdata);
	463	+
456	464	switch (pdata->an_mode) {
457	465	case XGBE_AN_MODE_CL73:
458	466	case XGBE_AN_MODE_CL73_REDRV:
@@ -505,11 +513,11 @@ static enum xgbe_an xgbe_an73_tx_training(struct xgbe_prv_data *pdata, · 46 unmodified lines
505	513	XMDIO_WRITE(pdata, MDIO_MMD_PMAPMD, MDIO_PMA_10GBR_PMD_CTRL,
506	514	reg);
507	515
508		− if (pdata->phy_if.phy_impl.kr_training_post)
509		− pdata->phy_if.phy_impl.kr_training_post(pdata);
510		−
511	516	netif_dbg(pdata, link, pdata->netdev,
512	517	"KR training initiated\n");
	518	+
	519	+ if (pdata->phy_if.phy_impl.kr_training_post)
	520	+ pdata->phy_if.phy_impl.kr_training_post(pdata);
513	521	}
514	522
515	523	return XGBE_AN_PAGE_RECEIVED;
@@ -637,11 +645,11 @@ static enum xgbe_an xgbe_an73_incompat_link(struct xgbe_prv_data *pdata) · 121 unmodified lines
637	645	return XGBE_AN_NO_LINK;
638	646	}
639	647
640		− xgbe_an73_disable(pdata);
	648	+ xgbe_an_disable(pdata);
641	649
642	650	xgbe_switch_mode(pdata);
643	651
644		− xgbe_an73_restart(pdata);
	652	+ xgbe_an_restart(pdata);
645	653
646	654	return XGBE_AN_INCOMPAT_LINK;
647	655	}
@@ -820,6 +828,9 @@ static void xgbe_an37_state_machine(struct xgbe_prv_data *pdata) · 172 unmodified lines
820	828	pdata->an_result = pdata->an_state;
821	829	pdata->an_state = XGBE_AN_READY;
822	830
	831	+ if (pdata->phy_if.phy_impl.an_post)
	832	+ pdata->phy_if.phy_impl.an_post(pdata);
	833	+
823	834	netif_dbg(pdata, link, pdata->netdev, "CL37 AN result: %s\n",
824	835	xgbe_state_as_string(pdata->an_result));
825	836	}
@@ -903,6 +914,9 @@ again: · 77 unmodified lines
903	914	pdata->kx_state = XGBE_RX_BPA;
904	915	pdata->an_start = 0;
905	916
	917	+ if (pdata->phy_if.phy_impl.an_post)
	918	+ pdata->phy_if.phy_impl.an_post(pdata);
	919	+
906	920	netif_dbg(pdata, link, pdata->netdev, "CL73 AN result: %s\n",
907	921	xgbe_state_as_string(pdata->an_result));
908	922	}

drivers/net/ethernet/amd/xgbe/xgbe-pci.c+2−0

@@ -456,6 +456,7 @@ static const struct xgbe_version_data xgbe_v2a = { · 455 unmodified lines
456	456	.irq_reissue_support = 1,
457	457	.tx_desc_prefetch = 5,
458	458	.rx_desc_prefetch = 5,
	459	+ .an_cdr_workaround = 1,
459	460	};
460	461
461	462	static const struct xgbe_version_data xgbe_v2b = {
@@ -470,6 +471,7 @@ static const struct xgbe_version_data xgbe_v2b = { · 8 unmodified lines
470	471	.irq_reissue_support = 1,
471	472	.tx_desc_prefetch = 5,
472	473	.rx_desc_prefetch = 5,
	474	+ .an_cdr_workaround = 1,
473	475	};
474	476
475	477	static const struct pci_device_id xgbe_pci_table[] = {

drivers/net/ethernet/amd/xgbe/xgbe-phy-v2.c+178−18

@@ -147,6 +147,14 @@ · 146 unmodified lines
147	147	/* Rate-change complete wait/retry count */
148	148	#define XGBE_RATECHANGE_COUNT 500
149	149
	150	+/* CDR delay values for KR support (in usec) */
	151	+#define XGBE_CDR_DELAY_INIT 10000
	152	+#define XGBE_CDR_DELAY_INC 10000
	153	+#define XGBE_CDR_DELAY_MAX 100000
	154	+
	155	+/* RRC frequency during link status check */
	156	+#define XGBE_RRC_FREQUENCY 10
	157	+
150	158	enum xgbe_port_mode {
151	159	XGBE_PORT_MODE_RSVD = 0,
152	160	XGBE_PORT_MODE_BACKPLANE,
@@ -245,6 +253,10 @@ enum xgbe_sfp_speed { · 92 unmodified lines
245	253	#define XGBE_SFP_BASE_VENDOR_SN 4
246	254	#define XGBE_SFP_BASE_VENDOR_SN_LEN 16
247	255
	256	+#define XGBE_SFP_EXTD_OPT1 1
	257	+#define XGBE_SFP_EXTD_OPT1_RX_LOS BIT(1)
	258	+#define XGBE_SFP_EXTD_OPT1_TX_FAULT BIT(3)
	259	+
248	260	#define XGBE_SFP_EXTD_DIAG 28
249	261	#define XGBE_SFP_EXTD_DIAG_ADDR_CHANGE BIT(2)
250	262
@@ -324,6 +336,7 @@ struct xgbe_phy_data { · 73 unmodified lines
324	336
325	337	unsigned int sfp_gpio_address;
326	338	unsigned int sfp_gpio_mask;
	339	+ unsigned int sfp_gpio_inputs;
327	340	unsigned int sfp_gpio_rx_los;
328	341	unsigned int sfp_gpio_tx_fault;
329	342	unsigned int sfp_gpio_mod_absent;
@@ -355,6 +368,10 @@ struct xgbe_phy_data { · 25 unmodified lines
355	368	unsigned int redrv_addr;
356	369	unsigned int redrv_lane;
357	370	unsigned int redrv_model;
	371	+
	372	+ /* KR AN support */
	373	+ unsigned int phy_cdr_notrack;
	374	+ unsigned int phy_cdr_delay;
358	375	};
359	376
360	377	/* I2C, MDIO and GPIO lines are muxed, so only one device at a time */
@@ -974,6 +991,49 @@ static void xgbe_phy_sfp_external_phy(struct xgbe_prv_data *pdata) · 613 unmodified lines
974	991	phy_data->sfp_phy_avail = 1;
975	992	}
976	993
	994	+static bool xgbe_phy_check_sfp_rx_los(struct xgbe_phy_data *phy_data)
	995	+{
	996	+ u8 *sfp_extd = phy_data->sfp_eeprom.extd;
	997	+
	998	+ if (!(sfp_extd[XGBE_SFP_EXTD_OPT1] & XGBE_SFP_EXTD_OPT1_RX_LOS))
	999	+ return false;
	1000	+
	1001	+ if (phy_data->sfp_gpio_mask & XGBE_GPIO_NO_RX_LOS)
	1002	+ return false;
	1003	+
	1004	+ if (phy_data->sfp_gpio_inputs & (1 << phy_data->sfp_gpio_rx_los))
	1005	+ return true;
	1006	+
	1007	+ return false;
	1008	+}
	1009	+
	1010	+static bool xgbe_phy_check_sfp_tx_fault(struct xgbe_phy_data *phy_data)
	1011	+{
	1012	+ u8 *sfp_extd = phy_data->sfp_eeprom.extd;
	1013	+
	1014	+ if (!(sfp_extd[XGBE_SFP_EXTD_OPT1] & XGBE_SFP_EXTD_OPT1_TX_FAULT))
	1015	+ return false;
	1016	+
	1017	+ if (phy_data->sfp_gpio_mask & XGBE_GPIO_NO_TX_FAULT)
	1018	+ return false;
	1019	+
	1020	+ if (phy_data->sfp_gpio_inputs & (1 << phy_data->sfp_gpio_tx_fault))
	1021	+ return true;
	1022	+
	1023	+ return false;
	1024	+}
	1025	+
	1026	+static bool xgbe_phy_check_sfp_mod_absent(struct xgbe_phy_data *phy_data)
	1027	+{
	1028	+ if (phy_data->sfp_gpio_mask & XGBE_GPIO_NO_MOD_ABSENT)
	1029	+ return false;
	1030	+
	1031	+ if (phy_data->sfp_gpio_inputs & (1 << phy_data->sfp_gpio_mod_absent))
	1032	+ return true;
	1033	+
	1034	+ return false;
	1035	+}
	1036	+
977	1037	static bool xgbe_phy_belfuse_parse_quirks(struct xgbe_prv_data *pdata)
978	1038	{
979	1039	struct xgbe_phy_data *phy_data = pdata->phy_data;
@@ -1019,6 +1079,10 @@ static void xgbe_phy_sfp_parse_eeprom(struct xgbe_prv_data *pdata) · 39 unmodified lines
1019	1079	if (sfp_base[XGBE_SFP_BASE_EXT_ID] != XGBE_SFP_EXT_ID_SFP)
1020	1080	return;
1021	1081
	1082	+ /* Update transceiver signals (eeprom extd/options) */
	1083	+ phy_data->sfp_tx_fault = xgbe_phy_check_sfp_tx_fault(phy_data);
	1084	+ phy_data->sfp_rx_los = xgbe_phy_check_sfp_rx_los(phy_data);
	1085	+
1022	1086	if (xgbe_phy_sfp_parse_quirks(pdata))
1023	1087	return;
1024	1088
@@ -1184,7 +1248,6 @@ put: · 159 unmodified lines
1184	1248	static void xgbe_phy_sfp_signals(struct xgbe_prv_data *pdata)
1185	1249	{
1186	1250	struct xgbe_phy_data *phy_data = pdata->phy_data;
1187		− unsigned int gpio_input;
1188	1251	u8 gpio_reg, gpio_ports[2];
1189	1252	int ret;
1190	1253
@@ -1199,23 +1262,9 @@ static void xgbe_phy_sfp_signals(struct xgbe_prv_data *pdata) · 8 unmodified lines
1199	1262	return;
1200	1263	}
1201	1264
1202		− gpio_input = (gpio_ports[1] << 8) \| gpio_ports[0];
1203		−
1204		− if (phy_data->sfp_gpio_mask & XGBE_GPIO_NO_MOD_ABSENT) {
1205		− /* No GPIO, just assume the module is present for now */
1206		− phy_data->sfp_mod_absent = 0;
1207		− } else {
1208		− if (!(gpio_input & (1 << phy_data->sfp_gpio_mod_absent)))
1209		− phy_data->sfp_mod_absent = 0;
1210		− }
1211		−
1212		− if (!(phy_data->sfp_gpio_mask & XGBE_GPIO_NO_RX_LOS) &&
1213		− (gpio_input & (1 << phy_data->sfp_gpio_rx_los)))
1214		− phy_data->sfp_rx_los = 1;
	1265	+ phy_data->sfp_gpio_inputs = (gpio_ports[1] << 8) \| gpio_ports[0];
1215	1266
1216		− if (!(phy_data->sfp_gpio_mask & XGBE_GPIO_NO_TX_FAULT) &&
1217		− (gpio_input & (1 << phy_data->sfp_gpio_tx_fault)))
1218		− phy_data->sfp_tx_fault = 1;
	1267	+ phy_data->sfp_mod_absent = xgbe_phy_check_sfp_mod_absent(phy_data);
1219	1268	}
1220	1269
1221	1270	static void xgbe_phy_sfp_mod_absent(struct xgbe_prv_data *pdata)
@@ -2361,7 +2410,7 @@ static int xgbe_phy_link_status(struct xgbe_prv_data pdata, int an_restart) · 1139 unmodified lines
2361	2410	return 1;
2362	2411
2363	2412	/* No link, attempt a receiver reset cycle */
2364		− if (phy_data->rrc_count++) {
	2413	+ if (phy_data->rrc_count++ > XGBE_RRC_FREQUENCY) {
2365	2414	phy_data->rrc_count = 0;
2366	2415	xgbe_phy_rrc(pdata);
2367	2416	}
@@ -2669,6 +2718,103 @@ static bool xgbe_phy_port_enabled(struct xgbe_prv_data *pdata) · 301 unmodified lines
2669	2718	return true;
2670	2719	}
2671	2720
	2721	+static void xgbe_phy_cdr_track(struct xgbe_prv_data *pdata)
	2722	+{
	2723	+ struct xgbe_phy_data *phy_data = pdata->phy_data;
	2724	+
	2725	+ if (!pdata->debugfs_an_cdr_workaround)
	2726	+ return;
	2727	+
	2728	+ if (!phy_data->phy_cdr_notrack)
	2729	+ return;
	2730	+
	2731	+ usleep_range(phy_data->phy_cdr_delay,
	2732	+ phy_data->phy_cdr_delay + 500);
	2733	+
	2734	+ XMDIO_WRITE_BITS(pdata, MDIO_MMD_PMAPMD, MDIO_VEND2_PMA_CDR_CONTROL,
	2735	+ XGBE_PMA_CDR_TRACK_EN_MASK,
	2736	+ XGBE_PMA_CDR_TRACK_EN_ON);
	2737	+
	2738	+ phy_data->phy_cdr_notrack = 0;
	2739	+}
	2740	+
	2741	+static void xgbe_phy_cdr_notrack(struct xgbe_prv_data *pdata)
	2742	+{
	2743	+ struct xgbe_phy_data *phy_data = pdata->phy_data;
	2744	+
	2745	+ if (!pdata->debugfs_an_cdr_workaround)
	2746	+ return;
	2747	+
	2748	+ if (phy_data->phy_cdr_notrack)
	2749	+ return;
	2750	+
	2751	+ XMDIO_WRITE_BITS(pdata, MDIO_MMD_PMAPMD, MDIO_VEND2_PMA_CDR_CONTROL,
	2752	+ XGBE_PMA_CDR_TRACK_EN_MASK,
	2753	+ XGBE_PMA_CDR_TRACK_EN_OFF);
	2754	+
	2755	+ xgbe_phy_rrc(pdata);
	2756	+
	2757	+ phy_data->phy_cdr_notrack = 1;
	2758	+}
	2759	+
	2760	+static void xgbe_phy_kr_training_post(struct xgbe_prv_data *pdata)
	2761	+{
	2762	+ if (!pdata->debugfs_an_cdr_track_early)
	2763	+ xgbe_phy_cdr_track(pdata);
	2764	+}
	2765	+
	2766	+static void xgbe_phy_kr_training_pre(struct xgbe_prv_data *pdata)
	2767	+{
	2768	+ if (pdata->debugfs_an_cdr_track_early)
	2769	+ xgbe_phy_cdr_track(pdata);
	2770	+}
	2771	+
	2772	+static void xgbe_phy_an_post(struct xgbe_prv_data *pdata)
	2773	+{
	2774	+ struct xgbe_phy_data *phy_data = pdata->phy_data;
	2775	+
	2776	+ switch (pdata->an_mode) {
	2777	+ case XGBE_AN_MODE_CL73:
	2778	+ case XGBE_AN_MODE_CL73_REDRV:
	2779	+ if (phy_data->cur_mode != XGBE_MODE_KR)
	2780	+ break;
	2781	+
	2782	+ xgbe_phy_cdr_track(pdata);
	2783	+
	2784	+ switch (pdata->an_result) {
	2785	+ case XGBE_AN_READY:
	2786	+ case XGBE_AN_COMPLETE:
	2787	+ break;
	2788	+ default:
	2789	+ if (phy_data->phy_cdr_delay < XGBE_CDR_DELAY_MAX)
	2790	+ phy_data->phy_cdr_delay += XGBE_CDR_DELAY_INC;
	2791	+ else
	2792	+ phy_data->phy_cdr_delay = XGBE_CDR_DELAY_INIT;
	2793	+ break;
	2794	+ }
	2795	+ break;
	2796	+ default:
	2797	+ break;
	2798	+ }
	2799	+}
	2800	+
	2801	+static void xgbe_phy_an_pre(struct xgbe_prv_data *pdata)
	2802	+{
	2803	+ struct xgbe_phy_data *phy_data = pdata->phy_data;
	2804	+
	2805	+ switch (pdata->an_mode) {
	2806	+ case XGBE_AN_MODE_CL73:
	2807	+ case XGBE_AN_MODE_CL73_REDRV:
	2808	+ if (phy_data->cur_mode != XGBE_MODE_KR)
	2809	+ break;
	2810	+
	2811	+ xgbe_phy_cdr_notrack(pdata);
	2812	+ break;
	2813	+ default:
	2814	+ break;
	2815	+ }
	2816	+}
	2817	+
2672	2818	static void xgbe_phy_stop(struct xgbe_prv_data *pdata)
2673	2819	{
2674	2820	struct xgbe_phy_data *phy_data = pdata->phy_data;
@@ -2680,6 +2826,9 @@ static void xgbe_phy_stop(struct xgbe_prv_data *pdata) · 5 unmodified lines
2680	2826	xgbe_phy_sfp_reset(phy_data);
2681	2827	xgbe_phy_sfp_mod_absent(pdata);
2682	2828
	2829	+ /* Reset CDR support */
	2830	+ xgbe_phy_cdr_track(pdata);
	2831	+
2683	2832	/* Power off the PHY */
2684	2833	xgbe_phy_power_off(pdata);
2685	2834
@@ -2712,6 +2861,9 @@ static int xgbe_phy_start(struct xgbe_prv_data *pdata) · 26 unmodified lines
2712	2861	/* Start in highest supported mode */
2713	2862	xgbe_phy_set_mode(pdata, phy_data->start_mode);
2714	2863
	2864	+ /* Reset CDR support */
	2865	+ xgbe_phy_cdr_track(pdata);
	2866	+
2715	2867	/* After starting the I2C controller, we can check for an SFP */
2716	2868	switch (phy_data->port_mode) {
2717	2869	case XGBE_PORT_MODE_SFP:
@@ -3019,6 +3171,8 @@ static int xgbe_phy_init(struct xgbe_prv_data *pdata) · 301 unmodified lines
3019	3171	}
3020	3172	}
3021	3173
	3174	+ phy_data->phy_cdr_delay = XGBE_CDR_DELAY_INIT;
	3175	+
3022	3176	/* Register for driving external PHYs */
3023	3177	mii = devm_mdiobus_alloc(pdata->dev);
3024	3178	if (!mii) {
@@ -3071,4 +3225,10 @@ void xgbe_init_function_ptrs_phy_v2(struct xgbe_phy_if *phy_if) · 46 unmodified lines
3071	3225	phy_impl->an_advertising = xgbe_phy_an_advertising;
3072	3226
3073	3227	phy_impl->an_outcome = xgbe_phy_an_outcome;
	3228	+
	3229	+ phy_impl->an_pre = xgbe_phy_an_pre;
	3230	+ phy_impl->an_post = xgbe_phy_an_post;
	3231	+
	3232	+ phy_impl->kr_training_pre = xgbe_phy_kr_training_pre;
	3233	+ phy_impl->kr_training_post = xgbe_phy_kr_training_post;
3074	3234	}

drivers/net/ethernet/amd/xgbe/xgbe.h+9−0

@@ -833,6 +833,7 @@ struct xgbe_hw_if { · 832 unmodified lines
833	833	/* This structure represents implementation specific routines for an
834	834	* implementation of a PHY. All routines are required unless noted below.
835	835	* Optional routines:
	836	+ * an_pre, an_post
836	837	* kr_training_pre, kr_training_post
837	838	*/
838	839	struct xgbe_phy_impl_if {
@@ -875,6 +876,10 @@ struct xgbe_phy_impl_if { · 36 unmodified lines
875	876	/* Process results of auto-negotiation */
876	877	enum xgbe_mode (an_outcome)(struct xgbe_prv_data );
877	878
	879	+ /* Pre/Post auto-negotiation support */
	880	+ void (an_pre)(struct xgbe_prv_data );
	881	+ void (an_post)(struct xgbe_prv_data );
	882	+
878	883	/* Pre/Post KR training enablement support */
879	884	void (kr_training_pre)(struct xgbe_prv_data );
880	885	void (kr_training_post)(struct xgbe_prv_data );
@@ -989,6 +994,7 @@ struct xgbe_version_data { · 108 unmodified lines
989	994	unsigned int irq_reissue_support;
990	995	unsigned int tx_desc_prefetch;
991	996	unsigned int rx_desc_prefetch;
	997	+ unsigned int an_cdr_workaround;
992	998	};
993	999
994	1000	struct xgbe_vxlan_data {
@@ -1257,6 +1263,9 @@ struct xgbe_prv_data { · 262 unmodified lines
1257	1263	unsigned int debugfs_xprop_reg;
1258	1264
1259	1265	unsigned int debugfs_xi2c_reg;
	1266	+
	1267	+ bool debugfs_an_cdr_workaround;
	1268	+ bool debugfs_an_cdr_track_early;
1260	1269	};
1261	1270
1262	1271	/* Function prototypes*/

drivers/net/ethernet/broadcom/bcmsysport.c+13−5

@@ -2144,14 +2144,21 @@ static const struct net_device_ops bcm_sysport_netdev_ops = { · 2143 unmodified lines
2144	2144	.ndo_select_queue = bcm_sysport_select_queue,
2145	2145	};
2146	2146
2147		−static int bcm_sysport_map_queues(struct net_device *dev,
	2147	+static int bcm_sysport_map_queues(struct notifier_block *nb,
2148	2148	struct dsa_notifier_register_info *info)
2149	2149	{
2150		− struct bcm_sysport_priv *priv = netdev_priv(dev);
2151	2150	struct bcm_sysport_tx_ring *ring;
	2151	+ struct bcm_sysport_priv *priv;
2152	2152	struct net_device *slave_dev;
2153	2153	unsigned int num_tx_queues;
2154	2154	unsigned int q, start, port;
	2155	+ struct net_device *dev;
	2156	+
	2157	+ priv = container_of(nb, struct bcm_sysport_priv, dsa_notifier);
	2158	+ if (priv->netdev != info->master)
	2159	+ return 0;
	2160	+
	2161	+ dev = info->master;
2155	2162
2156	2163	/* We can't be setting up queue inspection for non directly attached
2157	2164	* switches
@@ -2174,11 +2181,12 @@ static int bcm_sysport_map_queues(struct net_device *dev, · 16 unmodified lines
2174	2181	if (priv->is_lite)
2175	2182	netif_set_real_num_tx_queues(slave_dev,
2176	2183	slave_dev->num_tx_queues / 2);
	2184	+
2177	2185	num_tx_queues = slave_dev->real_num_tx_queues;
2178	2186
2179	2187	if (priv->per_port_num_tx_queues &&
2180	2188	priv->per_port_num_tx_queues != num_tx_queues)
2181		− netdev_warn(slave_dev, "asymetric number of per-port queues\n");
	2189	+ netdev_warn(slave_dev, "asymmetric number of per-port queues\n");
2182	2190
2183	2191	priv->per_port_num_tx_queues = num_tx_queues;
2184	2192
@@ -2201,7 +2209,7 @@ static int bcm_sysport_map_queues(struct net_device *dev, · 16 unmodified lines
2201	2209	return 0;
2202	2210	}
2203	2211
2204		−static int bcm_sysport_dsa_notifier(struct notifier_block *unused,
	2212	+static int bcm_sysport_dsa_notifier(struct notifier_block *nb,
2205	2213	unsigned long event, void *ptr)
2206	2214	{
2207	2215	struct dsa_notifier_register_info *info;
@@ -2211,7 +2219,7 @@ static int bcm_sysport_dsa_notifier(struct notifier_block *unused, · 3 unmodified lines
2211	2219
2212	2220	info = ptr;
2213	2221
2214		− return notifier_from_errno(bcm_sysport_map_queues(info->master, info));
	2222	+ return notifier_from_errno(bcm_sysport_map_queues(nb, info));
2215	2223	}
2216	2224
2217	2225	#define REV_FMT "v%2x.%02x"

drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c+27−22

@@ -1927,22 +1927,39 @@ static char bnxt_parse_pkglog(int desired_field, u8 data, size_t datalen) · 1926 unmodified lines
1927	1927	return retval;
1928	1928	}
1929	1929
1930		−static char bnxt_get_pkgver(struct net_device dev, char *buf, size_t buflen)
	1930	+static void bnxt_get_pkgver(struct net_device *dev)
1931	1931	{
	1932	+ struct bnxt *bp = netdev_priv(dev);
1932	1933	u16 index = 0;
1933		− u32 datalen;
	1934	+ char *pkgver;
	1935	+ u32 pkglen;
	1936	+ u8 *pkgbuf;
	1937	+ int len;
1934	1938
1935	1939	if (bnxt_find_nvram_item(dev, BNX_DIR_TYPE_PKG_LOG,
1936	1940	BNX_DIR_ORDINAL_FIRST, BNX_DIR_EXT_NONE,
1937		− &index, NULL, &datalen) != 0)
1938		− return NULL;
	1941	+ &index, NULL, &pkglen) != 0)
	1942	+ return;
1939	1943
1940		− memset(buf, 0, buflen);
1941		− if (bnxt_get_nvram_item(dev, index, 0, datalen, buf) != 0)
1942		− return NULL;
	1944	+ pkgbuf = kzalloc(pkglen, GFP_KERNEL);
	1945	+ if (!pkgbuf) {
	1946	+ dev_err(&bp->pdev->dev, "Unable to allocate memory for pkg version, length = %u\n",
	1947	+ pkglen);
	1948	+ return;
	1949	+ }
	1950	+
	1951	+ if (bnxt_get_nvram_item(dev, index, 0, pkglen, pkgbuf))
	1952	+ goto err;
1943	1953
1944		− return bnxt_parse_pkglog(BNX_PKG_LOG_FIELD_IDX_PKG_VERSION, buf,
1945		− datalen);
	1954	+ pkgver = bnxt_parse_pkglog(BNX_PKG_LOG_FIELD_IDX_PKG_VERSION, pkgbuf,
	1955	+ pkglen);
	1956	+ if (pkgver && pkgver != 0 && isdigit(pkgver)) {
	1957	+ len = strlen(bp->fw_ver_str);
	1958	+ snprintf(bp->fw_ver_str + len, FW_VER_STR_LEN - len - 1,
	1959	+ "/pkg %s", pkgver);
	1960	+ }
	1961	+err:
	1962	+ kfree(pkgbuf);
1946	1963	}
1947	1964
1948	1965	static int bnxt_get_eeprom(struct net_device *dev,
@@ -2615,22 +2632,10 @@ void bnxt_ethtool_init(struct bnxt *bp) · 666 unmodified lines
2615	2632	struct hwrm_selftest_qlist_input req = {0};
2616	2633	struct bnxt_test_info *test_info;
2617	2634	struct net_device *dev = bp->dev;
2618		− char *pkglog;
2619	2635	int i, rc;
2620	2636
2621		− pkglog = kzalloc(BNX_PKG_LOG_MAX_LENGTH, GFP_KERNEL);
2622		− if (pkglog) {
2623		− char *pkgver;
2624		− int len;
	2637	+ bnxt_get_pkgver(dev);
2625	2638
2626		− pkgver = bnxt_get_pkgver(dev, pkglog, BNX_PKG_LOG_MAX_LENGTH);
2627		− if (pkgver && pkgver != 0 && isdigit(pkgver)) {
2628		− len = strlen(bp->fw_ver_str);
2629		− snprintf(bp->fw_ver_str + len, FW_VER_STR_LEN - len - 1,
2630		− "/pkg %s", pkgver);
2631		− }
2632		− kfree(pkglog);
2633		− }
2634	2639	if (bp->hwrm_spec_code < 0x10704 \|\| !BNXT_SINGLE_PF(bp))
2635	2640	return;
2636	2641

drivers/net/ethernet/broadcom/bnxt/bnxt_nvm_defs.h+0−2

@@ -59,8 +59,6 @@ enum bnxt_nvm_directory_type { · 58 unmodified lines
59	59	#define BNX_DIR_ATTR_NO_CHKSUM (1 << 0)
60	60	#define BNX_DIR_ATTR_PROP_STREAM (1 << 1)
61	61
62		−#define BNX_PKG_LOG_MAX_LENGTH 4096
63		−
64	62	enum bnxnvm_pkglog_field_index {
65	63	BNX_PKG_LOG_FIELD_IDX_INSTALLED_TIMESTAMP = 0,
66	64	BNX_PKG_LOG_FIELD_IDX_PKG_DESCRIPTION = 1,

drivers/net/ethernet/freescale/ucc_geth_ethtool.c+1−1

@@ -61,7 +61,7 @@ static const char hw_stat_gstrings[][ETH_GSTRING_LEN] = { · 60 unmodified lines
61	61	static const char tx_fw_stat_gstrings[][ETH_GSTRING_LEN] = {
62	62	"tx-single-collision",
63	63	"tx-multiple-collision",
64		− "tx-late-collsion",
	64	+ "tx-late-collision",
65	65	"tx-aborted-frames",
66	66	"tx-lost-frames",
67	67	"tx-carrier-sense-errors",

drivers/net/ethernet/hisilicon/hns/hnae.h+1−1

@@ -87,7 +87,7 @@ do { \ · 86 unmodified lines
87	87
88	88	#define HNAE_AE_REGISTER 0x1
89	89
90		−#define RCB_RING_NAME_LEN 16
	90	+#define RCB_RING_NAME_LEN (IFNAMSIZ + 4)
91	91
92	92	#define HNAE_LOWEST_LATENCY_COAL_PARAM 30
93	93	#define HNAE_LOW_LATENCY_COAL_PARAM 80

drivers/net/ethernet/ibm/ibmvnic.c+56−33

@@ -794,46 +794,61 @@ static int ibmvnic_login(struct net_device *netdev) · 793 unmodified lines
794	794	{
795	795	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
796	796	unsigned long timeout = msecs_to_jiffies(30000);
797		− struct device *dev = &adapter->vdev->dev;
	797	+ int retry_count = 0;
798	798	int rc;
799	799
800	800	do {
801		− if (adapter->renegotiate) {
802		− adapter->renegotiate = false;
	801	+ if (retry_count > IBMVNIC_MAX_QUEUES) {
	802	+ netdev_warn(netdev, "Login attempts exceeded\n");
	803	+ return -1;
	804	+ }
	805	+
	806	+ adapter->init_done_rc = 0;
	807	+ reinit_completion(&adapter->init_done);
	808	+ rc = send_login(adapter);
	809	+ if (rc) {
	810	+ netdev_warn(netdev, "Unable to login\n");
	811	+ return rc;
	812	+ }
	813	+
	814	+ if (!wait_for_completion_timeout(&adapter->init_done,
	815	+ timeout)) {
	816	+ netdev_warn(netdev, "Login timed out\n");
	817	+ return -1;
	818	+ }
	819	+
	820	+ if (adapter->init_done_rc == PARTIALSUCCESS) {
	821	+ retry_count++;
803	822	release_sub_crqs(adapter, 1);
804	823
	824	+ adapter->init_done_rc = 0;
805	825	reinit_completion(&adapter->init_done);
806	826	send_cap_queries(adapter);
807	827	if (!wait_for_completion_timeout(&adapter->init_done,
808	828	timeout)) {
809		− dev_err(dev, "Capabilities query timeout\n");
	829	+ netdev_warn(netdev,
	830	+ "Capabilities query timed out\n");
810	831	return -1;
811	832	}
	833	+
812	834	rc = init_sub_crqs(adapter);
813	835	if (rc) {
814		− dev_err(dev,
815		− "Initialization of SCRQ's failed\n");
	836	+ netdev_warn(netdev,
	837	+ "SCRQ initialization failed\n");
816	838	return -1;
817	839	}
	840	+
818	841	rc = init_sub_crq_irqs(adapter);
819	842	if (rc) {
820		− dev_err(dev,
821		− "Initialization of SCRQ's irqs failed\n");
	843	+ netdev_warn(netdev,
	844	+ "SCRQ irq initialization failed\n");
822	845	return -1;
823	846	}
824		− }
825		−
826		− reinit_completion(&adapter->init_done);
827		− rc = send_login(adapter);
828		− if (rc) {
829		− dev_err(dev, "Unable to attempt device login\n");
830		− return rc;
831		− } else if (!wait_for_completion_timeout(&adapter->init_done,
832		− timeout)) {
833		− dev_err(dev, "Login timeout\n");
	847	+ } else if (adapter->init_done_rc) {
	848	+ netdev_warn(netdev, "Adapter login failed\n");
834	849	return -1;
835	850	}
836		− } while (adapter->renegotiate);
	851	+ } while (adapter->init_done_rc == PARTIALSUCCESS);
837	852
838	853	/* handle pending MAC address changes after successful login */
839	854	if (adapter->mac_change_pending) {
@@ -1034,16 +1049,14 @@ static int __ibmvnic_open(struct net_device *netdev) · 194 unmodified lines
1034	1049	netdev_dbg(netdev, "Enabling rx_scrq[%d] irq\n", i);
1035	1050	if (prev_state == VNIC_CLOSED)
1036	1051	enable_irq(adapter->rx_scrq[i]->irq);
1037		− else
1038		− enable_scrq_irq(adapter, adapter->rx_scrq[i]);
	1052	+ enable_scrq_irq(adapter, adapter->rx_scrq[i]);
1039	1053	}
1040	1054
1041	1055	for (i = 0; i < adapter->req_tx_queues; i++) {
1042	1056	netdev_dbg(netdev, "Enabling tx_scrq[%d] irq\n", i);
1043	1057	if (prev_state == VNIC_CLOSED)
1044	1058	enable_irq(adapter->tx_scrq[i]->irq);
1045		− else
1046		− enable_scrq_irq(adapter, adapter->tx_scrq[i]);
	1059	+ enable_scrq_irq(adapter, adapter->tx_scrq[i]);
1047	1060	}
1048	1061
1049	1062	rc = set_link_state(adapter, IBMVNIC_LOGICAL_LNK_UP);
@@ -1115,7 +1128,7 @@ static void clean_rx_pools(struct ibmvnic_adapter *adapter) · 65 unmodified lines
1115	1128	if (!adapter->rx_pool)
1116	1129	return;
1117	1130
1118		− rx_scrqs = be32_to_cpu(adapter->login_rsp_buf->num_rxadd_subcrqs);
	1131	+ rx_scrqs = adapter->num_active_rx_pools;
1119	1132	rx_entries = adapter->req_rx_add_entries_per_subcrq;
1120	1133
1121	1134	/* Free any remaining skbs in the rx buffer pools */
@@ -1164,7 +1177,7 @@ static void clean_tx_pools(struct ibmvnic_adapter *adapter) · 42 unmodified lines
1164	1177	if (!adapter->tx_pool \|\| !adapter->tso_pool)
1165	1178	return;
1166	1179
1167		− tx_scrqs = be32_to_cpu(adapter->login_rsp_buf->num_txsubm_subcrqs);
	1180	+ tx_scrqs = adapter->num_active_tx_pools;
1168	1181
1169	1182	/* Free any remaining skbs in the tx buffer pools */
1170	1183	for (i = 0; i < tx_scrqs; i++) {
@@ -1184,6 +1197,7 @@ static void ibmvnic_disable_irqs(struct ibmvnic_adapter *adapter) · 13 unmodified lines
1184	1197	if (adapter->tx_scrq[i]->irq) {
1185	1198	netdev_dbg(netdev,
1186	1199	"Disabling tx_scrq[%d] irq\n", i);
	1200	+ disable_scrq_irq(adapter, adapter->tx_scrq[i]);
1187	1201	disable_irq(adapter->tx_scrq[i]->irq);
1188	1202	}
1189	1203	}
@@ -1193,6 +1207,7 @@ static void ibmvnic_disable_irqs(struct ibmvnic_adapter *adapter) · 3 unmodified lines
1193	1207	if (adapter->rx_scrq[i]->irq) {
1194	1208	netdev_dbg(netdev,
1195	1209	"Disabling rx_scrq[%d] irq\n", i);
	1210	+ disable_scrq_irq(adapter, adapter->rx_scrq[i]);
1196	1211	disable_irq(adapter->rx_scrq[i]->irq);
1197	1212	}
1198	1213	}
@@ -1828,7 +1843,8 @@ static int do_reset(struct ibmvnic_adapter *adapter, · 629 unmodified lines
1828	1843	for (i = 0; i < adapter->req_rx_queues; i++)
1829	1844	napi_schedule(&adapter->napi[i]);
1830	1845
1831		− if (adapter->reset_reason != VNIC_RESET_FAILOVER)
	1846	+ if (adapter->reset_reason != VNIC_RESET_FAILOVER &&
	1847	+ adapter->reset_reason != VNIC_RESET_CHANGE_PARAM)
1832	1848	netdev_notify_peers(netdev);
1833	1849
1834	1850	netif_carrier_on(netdev);
@@ -2601,12 +2617,19 @@ static int enable_scrq_irq(struct ibmvnic_adapter *adapter, · 766 unmodified lines
2601	2617	{
2602	2618	struct device *dev = &adapter->vdev->dev;
2603	2619	unsigned long rc;
	2620	+ u64 val;
2604	2621
2605	2622	if (scrq->hw_irq > 0x100000000ULL) {
2606	2623	dev_err(dev, "bad hw_irq = %lx\n", scrq->hw_irq);
2607	2624	return 1;
2608	2625	}
2609	2626
	2627	+ val = (0xff000000) \| scrq->hw_irq;
	2628	+ rc = plpar_hcall_norets(H_EOI, val);
	2629	+ if (rc)
	2630	+ dev_err(dev, "H_EOI FAILED irq 0x%llx. rc=%ld\n",
	2631	+ val, rc);
	2632	+
2610	2633	rc = plpar_hcall_norets(H_VIOCTL, adapter->vdev->unit_address,
2611	2634	H_ENABLE_VIO_INTERRUPT, scrq->hw_irq, 0, 0);
2612	2635	if (rc)
@@ -3170,7 +3193,7 @@ static int send_version_xchg(struct ibmvnic_adapter *adapter) · 557 unmodified lines
3170	3193	struct vnic_login_client_data {
3171	3194	u8 type;
3172	3195	__be16 len;
3173		− char name;
	3196	+ char name[];
3174	3197	} __packed;
3175	3198
3176	3199	static int vnic_client_data_len(struct ibmvnic_adapter *adapter)
@@ -3199,21 +3222,21 @@ static void vnic_add_client_data(struct ibmvnic_adapter *adapter, · 22 unmodified lines
3199	3222	vlcd->type = 1;
3200	3223	len = strlen(os_name) + 1;
3201	3224	vlcd->len = cpu_to_be16(len);
3202		− strncpy(&vlcd->name, os_name, len);
3203		− vlcd = (struct vnic_login_client_data )((char )&vlcd->name + len);
	3225	+ strncpy(vlcd->name, os_name, len);
	3226	+ vlcd = (struct vnic_login_client_data *)(vlcd->name + len);
3204	3227
3205	3228	/* Type 2 - LPAR name */
3206	3229	vlcd->type = 2;
3207	3230	len = strlen(utsname()->nodename) + 1;
3208	3231	vlcd->len = cpu_to_be16(len);
3209		− strncpy(&vlcd->name, utsname()->nodename, len);
3210		− vlcd = (struct vnic_login_client_data )((char )&vlcd->name + len);
	3232	+ strncpy(vlcd->name, utsname()->nodename, len);
	3233	+ vlcd = (struct vnic_login_client_data *)(vlcd->name + len);
3211	3234
3212	3235	/* Type 3 - device name */
3213	3236	vlcd->type = 3;
3214	3237	len = strlen(adapter->netdev->name) + 1;
3215	3238	vlcd->len = cpu_to_be16(len);
3216		− strncpy(&vlcd->name, adapter->netdev->name, len);
	3239	+ strncpy(vlcd->name, adapter->netdev->name, len);
3217	3240	}
3218	3241
3219	3242	static int send_login(struct ibmvnic_adapter *adapter)
@@ -3942,7 +3965,7 @@ static int handle_login_rsp(union ibmvnic_crq *login_rsp_crq, · 722 unmodified lines
3942	3965	* to resend the login buffer with fewer queues requested.
3943	3966	*/
3944	3967	if (login_rsp_crq->generic.rc.code) {
3945		− adapter->renegotiate = true;
	3968	+ adapter->init_done_rc = login_rsp_crq->generic.rc.code;
3946	3969	complete(&adapter->init_done);
3947	3970	return 0;
3948	3971	}

drivers/net/ethernet/ibm/ibmvnic.h+0−1

@@ -1035,7 +1035,6 @@ struct ibmvnic_adapter { · 1034 unmodified lines
1035	1035
1036	1036	struct ibmvnic_sub_crq_queue **tx_scrq;
1037	1037	struct ibmvnic_sub_crq_queue **rx_scrq;
1038		− bool renegotiate;
1039	1038
1040	1039	/* rx structs */
1041	1040	struct napi_struct *napi;

drivers/net/ethernet/intel/ice/ice_adminq_cmd.h+1−1

@@ -586,7 +586,7 @@ struct ice_sw_rule_lg_act { · 585 unmodified lines
586	586	#define ICE_LG_ACT_MIRROR_VSI_ID_S 3
587	587	#define ICE_LG_ACT_MIRROR_VSI_ID_M (0x3FF << ICE_LG_ACT_MIRROR_VSI_ID_S)
588	588
589		− /* Action type = 5 - Large Action */
	589	+ /* Action type = 5 - Generic Value */
590	590	#define ICE_LG_ACT_GENERIC 0x5
591	591	#define ICE_LG_ACT_GENERIC_VALUE_S 3
592	592	#define ICE_LG_ACT_GENERIC_VALUE_M (0xFFFF << ICE_LG_ACT_GENERIC_VALUE_S)

drivers/net/ethernet/intel/ice/ice_common.c+17−5

@@ -78,6 +78,7 @@ ice_aq_manage_mac_read(struct ice_hw hw, void buf, u16 buf_size, · 77 unmodified lines
78	78	struct ice_aq_desc desc;
79	79	enum ice_status status;
80	80	u16 flags;
	81	+ u8 i;
81	82
82	83	cmd = &desc.params.mac_read;
83	84
@@ -98,8 +99,16 @@ ice_aq_manage_mac_read(struct ice_hw hw, void buf, u16 buf_size, · 14 unmodified lines
98	99	return ICE_ERR_CFG;
99	100	}
100	101
101		− ether_addr_copy(hw->port_info->mac.lan_addr, resp->mac_addr);
102		− ether_addr_copy(hw->port_info->mac.perm_addr, resp->mac_addr);
	102	+ /* A single port can report up to two (LAN and WoL) addresses */
	103	+ for (i = 0; i < cmd->num_addr; i++)
	104	+ if (resp[i].addr_type == ICE_AQC_MAN_MAC_ADDR_TYPE_LAN) {
	105	+ ether_addr_copy(hw->port_info->mac.lan_addr,
	106	+ resp[i].mac_addr);
	107	+ ether_addr_copy(hw->port_info->mac.perm_addr,
	108	+ resp[i].mac_addr);
	109	+ break;
	110	+ }
	111	+
103	112	return 0;
104	113	}
105	114
@@ -464,9 +473,12 @@ enum ice_status ice_init_hw(struct ice_hw *hw) · 358 unmodified lines
464	473	if (status)
465	474	goto err_unroll_sched;
466	475
467		− /* Get port MAC information */
468		− mac_buf_len = sizeof(struct ice_aqc_manage_mac_read_resp);
469		− mac_buf = devm_kzalloc(ice_hw_to_dev(hw), mac_buf_len, GFP_KERNEL);
	476	+ /* Get MAC information */
	477	+ /* A single port can report up to two (LAN and WoL) addresses */
	478	+ mac_buf = devm_kcalloc(ice_hw_to_dev(hw), 2,
	479	+ sizeof(struct ice_aqc_manage_mac_read_resp),
	480	+ GFP_KERNEL);
	481	+ mac_buf_len = 2 * sizeof(struct ice_aqc_manage_mac_read_resp);
470	482
471	483	if (!mac_buf) {
472	484	status = ICE_ERR_NO_MEMORY;

drivers/net/ethernet/intel/ice/ice_hw_autogen.h+0−2

@@ -121,8 +121,6 @@ · 120 unmodified lines
121	121	#define PFINT_FW_CTL_CAUSE_ENA_S 30
122	122	#define PFINT_FW_CTL_CAUSE_ENA_M BIT(PFINT_FW_CTL_CAUSE_ENA_S)
123	123	#define PFINT_OICR 0x0016CA00
124		−#define PFINT_OICR_INTEVENT_S 0
125		−#define PFINT_OICR_INTEVENT_M BIT(PFINT_OICR_INTEVENT_S)
126	124	#define PFINT_OICR_HLP_RDY_S 14
127	125	#define PFINT_OICR_HLP_RDY_M BIT(PFINT_OICR_HLP_RDY_S)
128	126	#define PFINT_OICR_CPM_RDY_S 15

drivers/net/ethernet/intel/ice/ice_main.c+0−4

@@ -1722,9 +1722,6 @@ static irqreturn_t ice_misc_intr(int __always_unused irq, void *data) · 1721 unmodified lines
1722	1722	oicr = rd32(hw, PFINT_OICR);
1723	1723	ena_mask = rd32(hw, PFINT_OICR_ENA);
1724	1724
1725		− if (!(oicr & PFINT_OICR_INTEVENT_M))
1726		− goto ena_intr;
1727		−
1728	1725	if (oicr & PFINT_OICR_GRST_M) {
1729	1726	u32 reset;
1730	1727	/* we have a reset warning */
@@ -1782,7 +1779,6 @@ static irqreturn_t ice_misc_intr(int __always_unused irq, void *data) · 51 unmodified lines
1782	1779	}
1783	1780	ret = IRQ_HANDLED;
1784	1781
1785		−ena_intr:
1786	1782	/* re-enable interrupt causes that are not handled during this pass */
1787	1783	wr32(hw, PFINT_OICR_ENA, ena_mask);
1788	1784	if (!test_bit(__ICE_DOWN, pf->state)) {

drivers/net/ethernet/intel/ice/ice_sched.c+2−2

@@ -751,14 +751,14 @@ ice_sched_add_nodes_to_layer(struct ice_port_info *pi, · 750 unmodified lines
751	751	u16 num_added = 0;
752	752	u32 temp;
753	753
	754	+ *num_nodes_added = 0;
	755	+
754	756	if (!num_nodes)
755	757	return status;
756	758
757	759	if (!parent \|\| layer < hw->sw_entry_point_layer)
758	760	return ICE_ERR_PARAM;
759	761
760		− *num_nodes_added = 0;
761		−
762	762	/* max children per node per layer */
763	763	max_child_nodes =
764	764	le16_to_cpu(hw->layer_info[parent->tx_sched_layer].max_children);

drivers/net/ethernet/intel/igb/igb_main.c+16−1

@@ -1700,7 +1700,22 @@ static void igb_configure_cbs(struct igb_adapter *adapter, int queue, · 1699 unmodified lines
1700	1700	WARN_ON(hw->mac.type != e1000_i210);
1701	1701	WARN_ON(queue < 0 \|\| queue > 1);
1702	1702
1703		− if (enable) {
	1703	+ if (enable \|\| queue == 0) {
	1704	+ /* i210 does not allow the queue 0 to be in the Strict
	1705	+ * Priority mode while the Qav mode is enabled, so,
	1706	+ * instead of disabling strict priority mode, we give
	1707	+ * queue 0 the maximum of credits possible.
	1708	+ *
	1709	+ * See section 8.12.19 of the i210 datasheet, "Note:
	1710	+ * Queue0 QueueMode must be set to 1b when
	1711	+ * TransmitMode is set to Qav."
	1712	+ */
	1713	+ if (queue == 0 && !enable) {
	1714	+ /* max "linkspeed" idleslope in kbps */
	1715	+ idleslope = 1000000;
	1716	+ hicredit = ETH_FRAME_LEN;
	1717	+ }
	1718	+
1704	1719	set_tx_desc_fetch_prio(hw, queue, TX_QUEUE_PRIO_HIGH);
1705	1720	set_queue_mode(hw, queue, QUEUE_MODE_STREAM_RESERVATION);
1706	1721

drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c+1−1

@@ -3420,7 +3420,7 @@ static int ixgbevf_setup_all_tx_resources(struct ixgbevf_adapter *adapter) · 3419 unmodified lines
3420	3420	if (!err)
3421	3421	continue;
3422	3422	hw_dbg(&adapter->hw, "Allocation for XDP Queue %u failed\n", j);
3423		− break;
	3423	+ goto err_setup_tx;
3424	3424	}
3425	3425
3426	3426	return 0;

drivers/net/ethernet/marvell/mvpp2.c+31−13

@@ -663,7 +663,7 @@ enum mvpp2_tag_type { · 662 unmodified lines
663	663	#define MVPP2_PE_VID_FILT_RANGE_END (MVPP2_PRS_TCAM_SRAM_SIZE - 31)
664	664	#define MVPP2_PE_VID_FILT_RANGE_START (MVPP2_PE_VID_FILT_RANGE_END - \
665	665	MVPP2_PRS_VLAN_FILT_RANGE_SIZE + 1)
666		−#define MVPP2_PE_LAST_FREE_TID (MVPP2_PE_VID_FILT_RANGE_START - 1)
	666	+#define MVPP2_PE_LAST_FREE_TID (MVPP2_PE_MAC_RANGE_START - 1)
667	667	#define MVPP2_PE_IP6_EXT_PROTO_UN (MVPP2_PRS_TCAM_SRAM_SIZE - 30)
668	668	#define MVPP2_PE_IP6_ADDR_UN (MVPP2_PRS_TCAM_SRAM_SIZE - 29)
669	669	#define MVPP2_PE_IP4_ADDR_UN (MVPP2_PRS_TCAM_SRAM_SIZE - 28)
@@ -916,6 +916,8 @@ static struct { · 246 unmodified lines
916	916
917	917	#define MVPP2_MIB_COUNTERS_STATS_DELAY (1 * HZ)
918	918
	919	+#define MVPP2_DESC_DMA_MASK DMA_BIT_MASK(40)
	920	+
919	921	/* Definitions */
920	922
921	923	/* Shared Packet Processor resources */
@@ -940,6 +942,7 @@ struct mvpp2 { · 18 unmodified lines
940	942	struct clk *pp_clk;
941	943	struct clk *gop_clk;
942	944	struct clk *mg_clk;
	945	+ struct clk *mg_core_clk;
943	946	struct clk *axi_clk;
944	947
945	948	/* List of pointers to port structures */
@@ -1429,7 +1432,7 @@ static dma_addr_t mvpp2_txdesc_dma_addr_get(struct mvpp2_port *port, · 483 unmodified lines
1429	1432	if (port->priv->hw_version == MVPP21)
1430	1433	return tx_desc->pp21.buf_dma_addr;
1431	1434	else
1432		− return tx_desc->pp22.buf_dma_addr_ptp & GENMASK_ULL(40, 0);
	1435	+ return tx_desc->pp22.buf_dma_addr_ptp & MVPP2_DESC_DMA_MASK;
1433	1436	}
1434	1437
1435	1438	static void mvpp2_txdesc_dma_addr_set(struct mvpp2_port *port,
@@ -1447,7 +1450,7 @@ static void mvpp2_txdesc_dma_addr_set(struct mvpp2_port *port, · 11 unmodified lines
1447	1450	} else {
1448	1451	u64 val = (u64)addr;
1449	1452
1450		− tx_desc->pp22.buf_dma_addr_ptp &= ~GENMASK_ULL(40, 0);
	1453	+ tx_desc->pp22.buf_dma_addr_ptp &= ~MVPP2_DESC_DMA_MASK;
1451	1454	tx_desc->pp22.buf_dma_addr_ptp \|= val;
1452	1455	tx_desc->pp22.packet_offset = offset;
1453	1456	}
@@ -1507,7 +1510,7 @@ static dma_addr_t mvpp2_rxdesc_dma_addr_get(struct mvpp2_port *port, · 53 unmodified lines
1507	1510	if (port->priv->hw_version == MVPP21)
1508	1511	return rx_desc->pp21.buf_dma_addr;
1509	1512	else
1510		− return rx_desc->pp22.buf_dma_addr_key_hash & GENMASK_ULL(40, 0);
	1513	+ return rx_desc->pp22.buf_dma_addr_key_hash & MVPP2_DESC_DMA_MASK;
1511	1514	}
1512	1515
1513	1516	static unsigned long mvpp2_rxdesc_cookie_get(struct mvpp2_port *port,
@@ -1516,7 +1519,7 @@ static unsigned long mvpp2_rxdesc_cookie_get(struct mvpp2_port *port, · 2 unmodified lines
1516	1519	if (port->priv->hw_version == MVPP21)
1517	1520	return rx_desc->pp21.buf_cookie;
1518	1521	else
1519		− return rx_desc->pp22.buf_cookie_misc & GENMASK_ULL(40, 0);
	1522	+ return rx_desc->pp22.buf_cookie_misc & MVPP2_DESC_DMA_MASK;
1520	1523	}
1521	1524
1522	1525	static size_t mvpp2_rxdesc_size_get(struct mvpp2_port *port,
@@ -8766,18 +8769,27 @@ static int mvpp2_probe(struct platform_device *pdev) · 7243 unmodified lines
8766	8769	err = clk_prepare_enable(priv->mg_clk);
8767	8770	if (err < 0)
8768	8771	goto err_gop_clk;
	8772	+
	8773	+ priv->mg_core_clk = devm_clk_get(&pdev->dev, "mg_core_clk");
	8774	+ if (IS_ERR(priv->mg_core_clk)) {
	8775	+ priv->mg_core_clk = NULL;
	8776	+ } else {
	8777	+ err = clk_prepare_enable(priv->mg_core_clk);
	8778	+ if (err < 0)
	8779	+ goto err_mg_clk;
	8780	+ }
8769	8781	}
8770	8782
8771	8783	priv->axi_clk = devm_clk_get(&pdev->dev, "axi_clk");
8772	8784	if (IS_ERR(priv->axi_clk)) {
8773	8785	err = PTR_ERR(priv->axi_clk);
8774	8786	if (err == -EPROBE_DEFER)
8775		− goto err_gop_clk;
	8787	+ goto err_mg_core_clk;
8776	8788	priv->axi_clk = NULL;
8777	8789	} else {
8778	8790	err = clk_prepare_enable(priv->axi_clk);
8779	8791	if (err < 0)
8780		− goto err_gop_clk;
	8792	+ goto err_mg_core_clk;
8781	8793	}
8782	8794
8783	8795	/* Get system's tclk rate */
@@ -8789,9 +8801,9 @@ static int mvpp2_probe(struct platform_device *pdev) · 5 unmodified lines
8789	8801	}
8790	8802
8791	8803	if (priv->hw_version == MVPP22) {
8792		− err = dma_set_mask(&pdev->dev, DMA_BIT_MASK(40));
	8804	+ err = dma_set_mask(&pdev->dev, MVPP2_DESC_DMA_MASK);
8793	8805	if (err)
8794		− goto err_mg_clk;
	8806	+ goto err_axi_clk;
8795	8807	/* Sadly, the BM pools all share the same register to
8796	8808	* store the high 32 bits of their address. So they
8797	8809	* must all have the same high 32 bits, which forces
@@ -8799,14 +8811,14 @@ static int mvpp2_probe(struct platform_device *pdev) · 1 unmodified line
8799	8811	*/
8800	8812	err = dma_set_coherent_mask(&pdev->dev, DMA_BIT_MASK(32));
8801	8813	if (err)
8802		− goto err_mg_clk;
	8814	+ goto err_axi_clk;
8803	8815	}
8804	8816
8805	8817	/* Initialize network controller */
8806	8818	err = mvpp2_init(pdev, priv);
8807	8819	if (err < 0) {
8808	8820	dev_err(&pdev->dev, "failed to initialize controller\n");
8809		− goto err_mg_clk;
	8821	+ goto err_axi_clk;
8810	8822	}
8811	8823
8812	8824	/* Initialize ports */
@@ -8819,7 +8831,7 @@ static int mvpp2_probe(struct platform_device *pdev) · 6 unmodified lines
8819	8831	if (priv->port_count == 0) {
8820	8832	dev_err(&pdev->dev, "no ports enabled\n");
8821	8833	err = -ENODEV;
8822		− goto err_mg_clk;
	8834	+ goto err_axi_clk;
8823	8835	}
8824	8836
8825	8837	/* Statistics must be gathered regularly because some of them (like
@@ -8847,8 +8859,13 @@ err_port_probe: · 21 unmodified lines
8847	8859	mvpp2_port_remove(priv->port_list[i]);
8848	8860	i++;
8849	8861	}
8850		−err_mg_clk:
	8862	+err_axi_clk:
8851	8863	clk_disable_unprepare(priv->axi_clk);
	8864	+
	8865	+err_mg_core_clk:
	8866	+ if (priv->hw_version == MVPP22)
	8867	+ clk_disable_unprepare(priv->mg_core_clk);
	8868	+err_mg_clk:
8852	8869	if (priv->hw_version == MVPP22)
8853	8870	clk_disable_unprepare(priv->mg_clk);
8854	8871	err_gop_clk:
@@ -8895,6 +8912,7 @@ static int mvpp2_remove(struct platform_device *pdev) · 40 unmodified lines
8895	8912	return 0;
8896	8913
8897	8914	clk_disable_unprepare(priv->axi_clk);
	8915	+ clk_disable_unprepare(priv->mg_core_clk);
8898	8916	clk_disable_unprepare(priv->mg_clk);
8899	8917	clk_disable_unprepare(priv->pp_clk);
8900	8918	clk_disable_unprepare(priv->gop_clk);

drivers/net/ethernet/mellanox/mlx4/main.c+1−1

@@ -1317,7 +1317,7 @@ static int mlx4_mf_unbond(struct mlx4_dev *dev) · 1316 unmodified lines
1317	1317
1318	1318	ret = mlx4_unbond_fs_rules(dev);
1319	1319	if (ret)
1320		− mlx4_warn(dev, "multifunction unbond for flow rules failedi (%d)\n", ret);
	1320	+ mlx4_warn(dev, "multifunction unbond for flow rules failed (%d)\n", ret);
1321	1321	ret1 = mlx4_unbond_mac_table(dev);
1322	1322	if (ret1) {
1323	1323	mlx4_warn(dev, "multifunction unbond for MAC table failed (%d)\n", ret1);

drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c+5−3

@@ -1007,12 +1007,14 @@ static void mlx5e_trust_update_sq_inline_mode(struct mlx5e_priv *priv) · 1006 unmodified lines
1007	1007
1008	1008	mutex_lock(&priv->state_lock);
1009	1009
1010		− if (!test_bit(MLX5E_STATE_OPENED, &priv->state))
1011		− goto out;
1012		−
1013	1010	new_channels.params = priv->channels.params;
1014	1011	mlx5e_trust_update_tx_min_inline_mode(priv, &new_channels.params);
1015	1012
	1013	+ if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
	1014	+ priv->channels.params = new_channels.params;
	1015	+ goto out;
	1016	+ }
	1017	+
1016	1018	/* Skip if tx_min_inline is the same */
1017	1019	if (new_channels.params.tx_min_inline_mode ==
1018	1020	priv->channels.params.tx_min_inline_mode)

drivers/net/ethernet/mellanox/mlx5/core/en_rep.c+3−2

@@ -877,13 +877,14 @@ static const struct net_device_ops mlx5e_netdev_ops_rep = { · 876 unmodified lines
877	877	};
878	878
879	879	static void mlx5e_build_rep_params(struct mlx5_core_dev *mdev,
880		− struct mlx5e_params *params)
	880	+ struct mlx5e_params *params, u16 mtu)
881	881	{
882	882	u8 cq_period_mode = MLX5_CAP_GEN(mdev, cq_period_start_from_cqe) ?
883	883	MLX5_CQ_PERIOD_MODE_START_FROM_CQE :
884	884	MLX5_CQ_PERIOD_MODE_START_FROM_EQE;
885	885
886	886	params->hard_mtu = MLX5E_ETH_HARD_MTU;
	887	+ params->sw_mtu = mtu;
887	888	params->log_sq_size = MLX5E_REP_PARAMS_LOG_SQ_SIZE;
888	889	params->rq_wq_type = MLX5_WQ_TYPE_LINKED_LIST;
889	890	params->log_rq_mtu_frames = MLX5E_REP_PARAMS_LOG_RQ_SIZE;
@@ -931,7 +932,7 @@ static void mlx5e_init_rep(struct mlx5_core_dev *mdev, · 41 unmodified lines
931	932
932	933	priv->channels.params.num_channels = profile->max_nch(mdev);
933	934
934		− mlx5e_build_rep_params(mdev, &priv->channels.params);
	935	+ mlx5e_build_rep_params(mdev, &priv->channels.params, netdev->mtu);
935	936	mlx5e_build_rep_netdev(netdev);
936	937
937	938	mlx5e_timestamp_init(priv);

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

@@ -290,7 +290,7 @@ static int mlx5e_test_loopback(struct mlx5e_priv *priv) · 289 unmodified lines
290	290
291	291	if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
292	292	netdev_err(priv->netdev,
293		− "\tCan't perform loobpack test while device is down\n");
	293	+ "\tCan't perform loopback test while device is down\n");
294	294	return -ENODEV;
295	295	}
296	296

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

@@ -1864,7 +1864,8 @@ static bool modify_header_match_supported(struct mlx5_flow_spec *spec, · 1863 unmodified lines
1864	1864	}
1865	1865
1866	1866	ip_proto = MLX5_GET(fte_match_set_lyr_2_4, headers_v, ip_protocol);
1867		− if (modify_ip_header && ip_proto != IPPROTO_TCP && ip_proto != IPPROTO_UDP) {
	1867	+ if (modify_ip_header && ip_proto != IPPROTO_TCP &&
	1868	+ ip_proto != IPPROTO_UDP && ip_proto != IPPROTO_ICMP) {
1868	1869	pr_info("can't offload re-write of ip proto %d\n", ip_proto);
1869	1870	return false;
1870	1871	}

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

@@ -255,7 +255,7 @@ mlx5e_txwqe_build_dsegs(struct mlx5e_txqsq sq, struct sk_buff skb, · 254 unmodified lines
255	255	dma_addr = dma_map_single(sq->pdev, skb_data, headlen,
256	256	DMA_TO_DEVICE);
257	257	if (unlikely(dma_mapping_error(sq->pdev, dma_addr)))
258		− return -ENOMEM;
	258	+ goto dma_unmap_wqe_err;
259	259
260	260	dseg->addr = cpu_to_be64(dma_addr);
261	261	dseg->lkey = sq->mkey_be;
@@ -273,7 +273,7 @@ mlx5e_txwqe_build_dsegs(struct mlx5e_txqsq sq, struct sk_buff skb, · 11 unmodified lines
273	273	dma_addr = skb_frag_dma_map(sq->pdev, frag, 0, fsz,
274	274	DMA_TO_DEVICE);
275	275	if (unlikely(dma_mapping_error(sq->pdev, dma_addr)))
276		− return -ENOMEM;
	276	+ goto dma_unmap_wqe_err;
277	277
278	278	dseg->addr = cpu_to_be64(dma_addr);
279	279	dseg->lkey = sq->mkey_be;
@@ -285,6 +285,10 @@ mlx5e_txwqe_build_dsegs(struct mlx5e_txqsq sq, struct sk_buff skb, · 5 unmodified lines
285	285	}
286	286
287	287	return num_dma;
	288	+
	289	+dma_unmap_wqe_err:
	290	+ mlx5e_dma_unmap_wqe_err(sq, num_dma);
	291	+ return -ENOMEM;
288	292	}
289	293
290	294	static inline void
@@ -380,17 +384,15 @@ static netdev_tx_t mlx5e_sq_xmit(struct mlx5e_txqsq sq, struct sk_buff skb, · 89 unmodified lines
380	384	num_dma = mlx5e_txwqe_build_dsegs(sq, skb, skb_data, headlen,
381	385	(struct mlx5_wqe_data_seg *)cseg + ds_cnt);
382	386	if (unlikely(num_dma < 0))
383		− goto dma_unmap_wqe_err;
	387	+ goto err_drop;
384	388
385	389	mlx5e_txwqe_complete(sq, skb, opcode, ds_cnt + num_dma,
386	390	num_bytes, num_dma, wi, cseg);
387	391
388	392	return NETDEV_TX_OK;
389	393
390		−dma_unmap_wqe_err:
	394	+err_drop:
391	395	sq->stats.dropped++;
392		− mlx5e_dma_unmap_wqe_err(sq, wi->num_dma);
393		−
394	396	dev_kfree_skb_any(skb);
395	397
396	398	return NETDEV_TX_OK;
@@ -645,17 +647,15 @@ netdev_tx_t mlx5i_sq_xmit(struct mlx5e_txqsq sq, struct sk_buff skb, · 248 unmodified lines
645	647	num_dma = mlx5e_txwqe_build_dsegs(sq, skb, skb_data, headlen,
646	648	(struct mlx5_wqe_data_seg *)cseg + ds_cnt);
647	649	if (unlikely(num_dma < 0))
648		− goto dma_unmap_wqe_err;
	650	+ goto err_drop;
649	651
650	652	mlx5e_txwqe_complete(sq, skb, opcode, ds_cnt + num_dma,
651	653	num_bytes, num_dma, wi, cseg);
652	654
653	655	return NETDEV_TX_OK;
654	656
655		−dma_unmap_wqe_err:
	657	+err_drop:
656	658	sq->stats.dropped++;
657		− mlx5e_dma_unmap_wqe_err(sq, wi->num_dma);
658		−
659	659	dev_kfree_skb_any(skb);
660	660
661	661	return NETDEV_TX_OK;

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

@@ -187,6 +187,7 @@ static void del_sw_ns(struct fs_node *node); · 186 unmodified lines
187	187	static void del_sw_hw_rule(struct fs_node *node);
188	188	static bool mlx5_flow_dests_cmp(struct mlx5_flow_destination *d1,
189	189	struct mlx5_flow_destination *d2);
	190	+static void cleanup_root_ns(struct mlx5_flow_root_namespace *root_ns);
190	191	static struct mlx5_flow_rule *
191	192	find_flow_rule(struct fs_fte *fte,
192	193	struct mlx5_flow_destination *dest);
@@ -481,7 +482,8 @@ static void del_sw_hw_rule(struct fs_node *node) · 288 unmodified lines
481	482
482	483	if (rule->dest_attr.type == MLX5_FLOW_DESTINATION_TYPE_COUNTER &&
483	484	--fte->dests_size) {
484		− modify_mask = BIT(MLX5_SET_FTE_MODIFY_ENABLE_MASK_ACTION);
	485	+ modify_mask = BIT(MLX5_SET_FTE_MODIFY_ENABLE_MASK_ACTION) \|
	486	+ BIT(MLX5_SET_FTE_MODIFY_ENABLE_MASK_FLOW_COUNTERS);
485	487	fte->action.action &= ~MLX5_FLOW_CONTEXT_ACTION_COUNT;
486	488	update_fte = true;
487	489	goto out;
@@ -2351,23 +2353,27 @@ static int create_anchor_flow_table(struct mlx5_flow_steering *steering) · 1863 unmodified lines
2351	2353
2352	2354	static int init_root_ns(struct mlx5_flow_steering *steering)
2353	2355	{
	2356	+ int err;
	2357	+
2354	2358	steering->root_ns = create_root_ns(steering, FS_FT_NIC_RX);
2355	2359	if (!steering->root_ns)
2356		− goto cleanup;
	2360	+ return -ENOMEM;
2357	2361
2358		− if (init_root_tree(steering, &root_fs, &steering->root_ns->ns.node))
2359		− goto cleanup;
	2362	+ err = init_root_tree(steering, &root_fs, &steering->root_ns->ns.node);
	2363	+ if (err)
	2364	+ goto out_err;
2360	2365
2361	2366	set_prio_attrs(steering->root_ns);
2362		−
2363		− if (create_anchor_flow_table(steering))
2364		− goto cleanup;
	2367	+ err = create_anchor_flow_table(steering);
	2368	+ if (err)
	2369	+ goto out_err;
2365	2370
2366	2371	return 0;
2367	2372
2368		−cleanup:
2369		− mlx5_cleanup_fs(steering->dev);
2370		− return -ENOMEM;
	2373	+out_err:
	2374	+ cleanup_root_ns(steering->root_ns);
	2375	+ steering->root_ns = NULL;
	2376	+ return err;
2371	2377	}
2372	2378
2373	2379	static void clean_tree(struct fs_node *node)

drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c+5−7

@@ -1718,13 +1718,11 @@ __mlxsw_sp_port_mdb_del(struct mlxsw_sp_port *mlxsw_sp_port, · 1717 unmodified lines
1718	1718	struct net_device *dev = mlxsw_sp_port->dev;
1719	1719	int err;
1720	1720
1721		− if (bridge_port->bridge_device->multicast_enabled) {
1722		− if (bridge_port->bridge_device->multicast_enabled) {
1723		− err = mlxsw_sp_port_smid_set(mlxsw_sp_port, mid->mid,
1724		− false);
1725		− if (err)
1726		− netdev_err(dev, "Unable to remove port from SMID\n");
1727		− }
	1721	+ if (bridge_port->bridge_device->multicast_enabled &&
	1722	+ !bridge_port->mrouter) {
	1723	+ err = mlxsw_sp_port_smid_set(mlxsw_sp_port, mid->mid, false);
	1724	+ if (err)
	1725	+ netdev_err(dev, "Unable to remove port from SMID\n");
1728	1726	}
1729	1727
1730	1728	err = mlxsw_sp_port_remove_from_mid(mlxsw_sp_port, mid);

drivers/net/ethernet/netronome/nfp/flower/action.c+8−2

@@ -183,17 +183,21 @@ static int · 182 unmodified lines
183	183	nfp_fl_set_ipv4_udp_tun(struct nfp_fl_set_ipv4_udp_tun *set_tun,
184	184	const struct tc_action *action,
185	185	struct nfp_fl_pre_tunnel *pre_tun,
186		− enum nfp_flower_tun_type tun_type)
	186	+ enum nfp_flower_tun_type tun_type,
	187	+ struct net_device *netdev)
187	188	{
188	189	size_t act_size = sizeof(struct nfp_fl_set_ipv4_udp_tun);
189	190	struct ip_tunnel_info *ip_tun = tcf_tunnel_info(action);
190	191	u32 tmp_set_ip_tun_type_index = 0;
191	192	/* Currently support one pre-tunnel so index is always 0. */
192	193	int pretun_idx = 0;
	194	+ struct net *net;
193	195
194	196	if (ip_tun->options_len)
195	197	return -EOPNOTSUPP;
196	198
	199	+ net = dev_net(netdev);
	200	+
197	201	set_tun->head.jump_id = NFP_FL_ACTION_OPCODE_SET_IPV4_TUNNEL;
198	202	set_tun->head.len_lw = act_size >> NFP_FL_LW_SIZ;
199	203
@@ -204,6 +208,7 @@ nfp_fl_set_ipv4_udp_tun(struct nfp_fl_set_ipv4_udp_tun *set_tun, · 4 unmodified lines
204	208
205	209	set_tun->tun_type_index = cpu_to_be32(tmp_set_ip_tun_type_index);
206	210	set_tun->tun_id = ip_tun->key.tun_id;
	211	+ set_tun->ttl = net->ipv4.sysctl_ip_default_ttl;
207	212
208	213	/* Complete pre_tunnel action. */
209	214	pre_tun->ipv4_dst = ip_tun->key.u.ipv4.dst;
@@ -511,7 +516,8 @@ nfp_flower_loop_action(const struct tc_action *a, · 301 unmodified lines
511	516	*a_len += sizeof(struct nfp_fl_pre_tunnel);
512	517
513	518	set_tun = (void )&nfp_fl->action_data[a_len];
514		− err = nfp_fl_set_ipv4_udp_tun(set_tun, a, pre_tun, *tun_type);
	519	+ err = nfp_fl_set_ipv4_udp_tun(set_tun, a, pre_tun, *tun_type,
	520	+ netdev);
515	521	if (err)
516	522	return err;
517	523	*a_len += sizeof(struct nfp_fl_set_ipv4_udp_tun);

drivers/net/ethernet/netronome/nfp/flower/cmsg.c+37−7

@@ -258,9 +258,6 @@ nfp_flower_cmsg_process_one_rx(struct nfp_app app, struct sk_buff skb) · 257 unmodified lines
258	258	case NFP_FLOWER_CMSG_TYPE_ACTIVE_TUNS:
259	259	nfp_tunnel_keep_alive(app, skb);
260	260	break;
261		− case NFP_FLOWER_CMSG_TYPE_TUN_NEIGH:
262		− /* Acks from the NFP that the route is added - ignore. */
263		− break;
264	261	default:
265	262	nfp_flower_cmsg_warn(app, "Cannot handle invalid repr control type %u\n",
266	263	type);
@@ -275,18 +272,49 @@ out: · 8 unmodified lines
275	272
276	273	void nfp_flower_cmsg_process_rx(struct work_struct *work)
277	274	{
	275	+ struct sk_buff_head cmsg_joined;
278	276	struct nfp_flower_priv *priv;
279	277	struct sk_buff *skb;
280	278
281	279	priv = container_of(work, struct nfp_flower_priv, cmsg_work);
	280	+ skb_queue_head_init(&cmsg_joined);
	281	+
	282	+ spin_lock_bh(&priv->cmsg_skbs_high.lock);
	283	+ skb_queue_splice_tail_init(&priv->cmsg_skbs_high, &cmsg_joined);
	284	+ spin_unlock_bh(&priv->cmsg_skbs_high.lock);
282	285
283		− while ((skb = skb_dequeue(&priv->cmsg_skbs)))
	286	+ spin_lock_bh(&priv->cmsg_skbs_low.lock);
	287	+ skb_queue_splice_tail_init(&priv->cmsg_skbs_low, &cmsg_joined);
	288	+ spin_unlock_bh(&priv->cmsg_skbs_low.lock);
	289	+
	290	+ while ((skb = __skb_dequeue(&cmsg_joined)))
284	291	nfp_flower_cmsg_process_one_rx(priv->app, skb);
285	292	}
286	293
287		−void nfp_flower_cmsg_rx(struct nfp_app app, struct sk_buff skb)
	294	+static void
	295	+nfp_flower_queue_ctl_msg(struct nfp_app app, struct sk_buff skb, int type)
288	296	{
289	297	struct nfp_flower_priv *priv = app->priv;
	298	+ struct sk_buff_head *skb_head;
	299	+
	300	+ if (type == NFP_FLOWER_CMSG_TYPE_PORT_REIFY \|\|
	301	+ type == NFP_FLOWER_CMSG_TYPE_PORT_MOD)
	302	+ skb_head = &priv->cmsg_skbs_high;
	303	+ else
	304	+ skb_head = &priv->cmsg_skbs_low;
	305	+
	306	+ if (skb_queue_len(skb_head) >= NFP_FLOWER_WORKQ_MAX_SKBS) {
	307	+ nfp_flower_cmsg_warn(app, "Dropping queued control messages\n");
	308	+ dev_kfree_skb_any(skb);
	309	+ return;
	310	+ }
	311	+
	312	+ skb_queue_tail(skb_head, skb);
	313	+ schedule_work(&priv->cmsg_work);
	314	+}
	315	+
	316	+void nfp_flower_cmsg_rx(struct nfp_app app, struct sk_buff skb)
	317	+{
290	318	struct nfp_flower_cmsg_hdr *cmsg_hdr;
291	319
292	320	cmsg_hdr = nfp_flower_cmsg_get_hdr(skb);
@@ -306,8 +334,10 @@ void nfp_flower_cmsg_rx(struct nfp_app app, struct sk_buff skb) · 13 unmodified lines
306	334	nfp_flower_process_mtu_ack(app, skb)) {
307	335	/* Handle MTU acks outside wq to prevent RTNL conflict. */
308	336	dev_consume_skb_any(skb);
	337	+ } else if (cmsg_hdr->type == NFP_FLOWER_CMSG_TYPE_TUN_NEIGH) {
	338	+ /* Acks from the NFP that the route is added - ignore. */
	339	+ dev_consume_skb_any(skb);
309	340	} else {
310		− skb_queue_tail(&priv->cmsg_skbs, skb);
311		− schedule_work(&priv->cmsg_work);
	341	+ nfp_flower_queue_ctl_msg(app, skb, cmsg_hdr->type);
312	342	}
313	343	}

drivers/net/ethernet/netronome/nfp/flower/cmsg.h+6−1

@@ -108,6 +108,8 @@ · 107 unmodified lines
108	108	#define NFP_FL_IPV4_TUNNEL_TYPE GENMASK(7, 4)
109	109	#define NFP_FL_IPV4_PRE_TUN_INDEX GENMASK(2, 0)
110	110
	111	+#define NFP_FLOWER_WORKQ_MAX_SKBS 30000
	112	+
111	113	#define nfp_flower_cmsg_warn(app, fmt, args...) \
112	114	do { \
113	115	if (net_ratelimit()) \
@@ -188,7 +190,10 @@ struct nfp_fl_set_ipv4_udp_tun { · 74 unmodified lines
188	190	__be16 reserved;
189	191	__be64 tun_id __packed;
190	192	__be32 tun_type_index;
191		− __be32 extra[3];
	193	+ __be16 reserved2;
	194	+ u8 ttl;
	195	+ u8 reserved3;
	196	+ __be32 extra[2];
192	197	};
193	198
194	199	/* Metadata with L2 (1W/4B)

drivers/net/ethernet/netronome/nfp/flower/main.c+5−3

@@ -360,7 +360,7 @@ nfp_flower_spawn_phy_reprs(struct nfp_app app, struct nfp_flower_priv priv) · 359 unmodified lines
360	360	}
361	361
362	362	SET_NETDEV_DEV(repr, &priv->nn->pdev->dev);
363		− nfp_net_get_mac_addr(app->pf, port);
	363	+ nfp_net_get_mac_addr(app->pf, repr, port);
364	364
365	365	cmsg_port_id = nfp_flower_cmsg_phys_port(phys_port);
366	366	err = nfp_repr_init(app, repr,
@@ -519,7 +519,8 @@ static int nfp_flower_init(struct nfp_app *app) · 152 unmodified lines
519	519
520	520	app->priv = app_priv;
521	521	app_priv->app = app;
522		− skb_queue_head_init(&app_priv->cmsg_skbs);
	522	+ skb_queue_head_init(&app_priv->cmsg_skbs_high);
	523	+ skb_queue_head_init(&app_priv->cmsg_skbs_low);
523	524	INIT_WORK(&app_priv->cmsg_work, nfp_flower_cmsg_process_rx);
524	525	init_waitqueue_head(&app_priv->reify_wait_queue);
525	526
@@ -549,7 +550,8 @@ static void nfp_flower_clean(struct nfp_app *app) · 23 unmodified lines
549	550	{
550	551	struct nfp_flower_priv *app_priv = app->priv;
551	552
552		− skb_queue_purge(&app_priv->cmsg_skbs);
	553	+ skb_queue_purge(&app_priv->cmsg_skbs_high);
	554	+ skb_queue_purge(&app_priv->cmsg_skbs_low);
553	555	flush_work(&app_priv->cmsg_work);
554	556
555	557	nfp_flower_metadata_cleanup(app);

drivers/net/ethernet/netronome/nfp/flower/main.h+6−2

@@ -107,7 +107,10 @@ struct nfp_mtu_conf { · 106 unmodified lines
107	107	* @mask_table: Hash table used to store masks
108	108	* @flow_table: Hash table used to store flower rules
109	109	* @cmsg_work: Workqueue for control messages processing
110		− * @cmsg_skbs: List of skbs for control message processing
	110	+ * @cmsg_skbs_high: List of higher priority skbs for control message
	111	+ * processing
	112	+ * @cmsg_skbs_low: List of lower priority skbs for control message
	113	+ * processing
111	114	* @nfp_mac_off_list: List of MAC addresses to offload
112	115	* @nfp_mac_index_list: List of unique 8-bit indexes for non NFP netdevs
113	116	* @nfp_ipv4_off_list: List of IPv4 addresses to offload
@@ -136,7 +139,8 @@ struct nfp_flower_priv { · 22 unmodified lines
136	139	DECLARE_HASHTABLE(mask_table, NFP_FLOWER_MASK_HASH_BITS);
137	140	DECLARE_HASHTABLE(flow_table, NFP_FLOWER_HASH_BITS);
138	141	struct work_struct cmsg_work;
139		− struct sk_buff_head cmsg_skbs;
	142	+ struct sk_buff_head cmsg_skbs_high;
	143	+ struct sk_buff_head cmsg_skbs_low;
140	144	struct list_head nfp_mac_off_list;
141	145	struct list_head nfp_mac_index_list;
142	146	struct list_head nfp_ipv4_off_list;

drivers/net/ethernet/netronome/nfp/nfp_app_nic.c+1−1

@@ -69,7 +69,7 @@ int nfp_app_nic_vnic_alloc(struct nfp_app app, struct nfp_net nn, · 68 unmodified lines
69	69	if (err)
70	70	return err < 0 ? err : 0;
71	71
72		− nfp_net_get_mac_addr(app->pf, nn->port);
	72	+ nfp_net_get_mac_addr(app->pf, nn->dp.netdev, nn->port);
73	73
74	74	return 0;
75	75	}

drivers/net/ethernet/netronome/nfp/nfp_main.h+3−1

@@ -171,7 +171,9 @@ void nfp_net_pci_remove(struct nfp_pf *pf); · 170 unmodified lines
171	171	int nfp_hwmon_register(struct nfp_pf *pf);
172	172	void nfp_hwmon_unregister(struct nfp_pf *pf);
173	173
174		−void nfp_net_get_mac_addr(struct nfp_pf pf, struct nfp_port port);
	174	+void
	175	+nfp_net_get_mac_addr(struct nfp_pf pf, struct net_device netdev,
	176	+ struct nfp_port *port);
175	177
176	178	bool nfp_ctrl_tx(struct nfp_net nn, struct sk_buff skb);
177	179

drivers/net/ethernet/netronome/nfp/nfp_net_main.c+18−13

@@ -67,23 +67,26 @@ · 66 unmodified lines
67	67	/**
68	68	* nfp_net_get_mac_addr() - Get the MAC address.
69	69	* @pf: NFP PF handle
	70	+ * @netdev: net_device to set MAC address on
70	71	* @port: NFP port structure
71	72	*
72	73	* First try to get the MAC address from NSP ETH table. If that
73	74	* fails generate a random address.
74	75	*/
75		−void nfp_net_get_mac_addr(struct nfp_pf pf, struct nfp_port port)
	76	+void
	77	+nfp_net_get_mac_addr(struct nfp_pf pf, struct net_device netdev,
	78	+ struct nfp_port *port)
76	79	{
77	80	struct nfp_eth_table_port *eth_port;
78	81
79	82	eth_port = __nfp_port_get_eth_port(port);
80	83	if (!eth_port) {
81		− eth_hw_addr_random(port->netdev);
	84	+ eth_hw_addr_random(netdev);
82	85	return;
83	86	}
84	87
85		− ether_addr_copy(port->netdev->dev_addr, eth_port->mac_addr);
86		− ether_addr_copy(port->netdev->perm_addr, eth_port->mac_addr);
	88	+ ether_addr_copy(netdev->dev_addr, eth_port->mac_addr);
	89	+ ether_addr_copy(netdev->perm_addr, eth_port->mac_addr);
87	90	}
88	91
89	92	static struct nfp_eth_table_port *
@@ -511,16 +514,18 @@ static int nfp_net_pci_map_mem(struct nfp_pf *pf) · 421 unmodified lines
511	514	return PTR_ERR(mem);
512	515	}
513	516
514		− min_size = NFP_MAC_STATS_SIZE * (pf->eth_tbl->max_index + 1);
515		− pf->mac_stats_mem = nfp_rtsym_map(pf->rtbl, "_mac_stats",
516		− "net.macstats", min_size,
517		− &pf->mac_stats_bar);
518		− if (IS_ERR(pf->mac_stats_mem)) {
519		− if (PTR_ERR(pf->mac_stats_mem) != -ENOENT) {
520		− err = PTR_ERR(pf->mac_stats_mem);
521		− goto err_unmap_ctrl;
	517	+ if (pf->eth_tbl) {
	518	+ min_size = NFP_MAC_STATS_SIZE * (pf->eth_tbl->max_index + 1);
	519	+ pf->mac_stats_mem = nfp_rtsym_map(pf->rtbl, "_mac_stats",
	520	+ "net.macstats", min_size,
	521	+ &pf->mac_stats_bar);
	522	+ if (IS_ERR(pf->mac_stats_mem)) {
	523	+ if (PTR_ERR(pf->mac_stats_mem) != -ENOENT) {
	524	+ err = PTR_ERR(pf->mac_stats_mem);
	525	+ goto err_unmap_ctrl;
	526	+ }
	527	+ pf->mac_stats_mem = NULL;
522	528	}
523		− pf->mac_stats_mem = NULL;
524	529	}
525	530
526	531	pf->vf_cfg_mem = nfp_net_pf_map_rtsym(pf, "net.vfcfg",

drivers/net/ethernet/netronome/nfp/nfpcore/nfp_mutex.c+4−1

@@ -211,8 +211,11 @@ int nfp_cpp_mutex_lock(struct nfp_cpp_mutex *mutex) · 210 unmodified lines
211	211	break;
212	212
213	213	err = msleep_interruptible(timeout_ms);
214		− if (err != 0)
	214	+ if (err != 0) {
	215	+ nfp_info(mutex->cpp,
	216	+ "interrupted waiting for NFP mutex\n");
215	217	return -ERESTARTSYS;
	218	+ }
216	219
217	220	if (time_is_before_eq_jiffies(warn_at)) {
218	221	warn_at = jiffies + NFP_MUTEX_WAIT_NEXT_WARN * HZ;

drivers/net/ethernet/netronome/nfp/nfpcore/nfp_nsp.c+1−2

@@ -281,8 +281,7 @@ nfp_nsp_wait_reg(struct nfp_cpp cpp, u64 reg, u32 nsp_cpp, u64 addr, · 280 unmodified lines
281	281	if ((*reg & mask) == val)
282	282	return 0;
283	283
284		− if (msleep_interruptible(25))
285		− return -ERESTARTSYS;
	284	+ msleep(25);
286	285
287	286	if (time_after(start_time, wait_until))
288	287	return -ETIMEDOUT;

drivers/net/ethernet/qlogic/qed/qed_ll2.c+1−1

@@ -2370,7 +2370,7 @@ static int qed_ll2_start_xmit(struct qed_dev cdev, struct sk_buff skb) · 2369 unmodified lines
2370	2370	u8 flags = 0;
2371	2371
2372	2372	if (unlikely(skb->ip_summed != CHECKSUM_NONE)) {
2373		− DP_INFO(cdev, "Cannot transmit a checksumed packet\n");
	2373	+ DP_INFO(cdev, "Cannot transmit a checksummed packet\n");
2374	2374	return -EINVAL;
2375	2375	}
2376	2376

drivers/net/ethernet/qlogic/qed/qed_roce.c+1−1

@@ -848,7 +848,7 @@ int qed_roce_query_qp(struct qed_hwfn *p_hwfn, · 847 unmodified lines
848	848
849	849	if (!(qp->resp_offloaded)) {
850	850	DP_NOTICE(p_hwfn,
851		− "The responder's qp should be offloded before requester's\n");
	851	+ "The responder's qp should be offloaded before requester's\n");
852	852	return -EINVAL;
853	853	}
854	854

drivers/net/ethernet/qualcomm/rmnet/rmnet_config.c+6−5

@@ -350,15 +350,16 @@ static int rmnet_fill_info(struct sk_buff skb, const struct net_device dev) · 349 unmodified lines
350	350
351	351	real_dev = priv->real_dev;
352	352
353		− if (!rmnet_is_real_dev_registered(real_dev))
354		− return -ENODEV;
355		−
356	353	if (nla_put_u16(skb, IFLA_RMNET_MUX_ID, priv->mux_id))
357	354	goto nla_put_failure;
358	355
359		− port = rmnet_get_port_rtnl(real_dev);
	356	+ if (rmnet_is_real_dev_registered(real_dev)) {
	357	+ port = rmnet_get_port_rtnl(real_dev);
	358	+ f.flags = port->data_format;
	359	+ } else {
	360	+ f.flags = 0;
	361	+ }
360	362
361		− f.flags = port->data_format;
362	363	f.mask = ~0;
363	364
364	365	if (nla_put(skb, IFLA_RMNET_FLAGS, sizeof(f), &f))

drivers/net/ethernet/realtek/8139too.c+1−1

@@ -2224,7 +2224,7 @@ static void rtl8139_poll_controller(struct net_device *dev) · 2223 unmodified lines
2224	2224	struct rtl8139_private *tp = netdev_priv(dev);
2225	2225	const int irq = tp->pci_dev->irq;
2226	2226
2227		− disable_irq(irq);
	2227	+ disable_irq_nosync(irq);
2228	2228	rtl8139_interrupt(irq, dev);
2229	2229	enable_irq(irq);
2230	2230	}

drivers/net/ethernet/sfc/ef10.c+50−38

@@ -3999,29 +3999,6 @@ static void efx_ef10_prepare_flr(struct efx_nic *efx) · 3998 unmodified lines
3999	3999	atomic_set(&efx->active_queues, 0);
4000	4000	}
4001	4001
4002		−static bool efx_ef10_filter_equal(const struct efx_filter_spec *left,
4003		− const struct efx_filter_spec *right)
4004		−{
4005		− if ((left->match_flags ^ right->match_flags) \|
4006		− ((left->flags ^ right->flags) &
4007		− (EFX_FILTER_FLAG_RX \| EFX_FILTER_FLAG_TX)))
4008		− return false;
4009		−
4010		− return memcmp(&left->outer_vid, &right->outer_vid,
4011		− sizeof(struct efx_filter_spec) -
4012		− offsetof(struct efx_filter_spec, outer_vid)) == 0;
4013		−}
4014		−
4015		−static unsigned int efx_ef10_filter_hash(const struct efx_filter_spec *spec)
4016		−{
4017		− BUILD_BUG_ON(offsetof(struct efx_filter_spec, outer_vid) & 3);
4018		− return jhash2((const u32 *)&spec->outer_vid,
4019		− (sizeof(struct efx_filter_spec) -
4020		− offsetof(struct efx_filter_spec, outer_vid)) / 4,
4021		− 0);
4022		− /* XXX should we randomise the initval? */
4023		−}
4024		−
4025	4002	/* Decide whether a filter should be exclusive or else should allow
4026	4003	* delivery to additional recipients. Currently we decide that
4027	4004	* filters for specific local unicast MAC and IP addresses are
@@ -4346,7 +4323,7 @@ static s32 efx_ef10_filter_insert(struct efx_nic *efx, · 318 unmodified lines
4346	4323	goto out_unlock;
4347	4324	match_pri = rc;
4348	4325
4349		− hash = efx_ef10_filter_hash(spec);
	4326	+ hash = efx_filter_spec_hash(spec);
4350	4327	is_mc_recip = efx_filter_is_mc_recipient(spec);
4351	4328	if (is_mc_recip)
4352	4329	bitmap_zero(mc_rem_map, EFX_EF10_FILTER_SEARCH_LIMIT);
@@ -4378,7 +4355,7 @@ static s32 efx_ef10_filter_insert(struct efx_nic *efx, · 25 unmodified lines
4378	4355	if (!saved_spec) {
4379	4356	if (ins_index < 0)
4380	4357	ins_index = i;
4381		− } else if (efx_ef10_filter_equal(spec, saved_spec)) {
	4358	+ } else if (efx_filter_spec_equal(spec, saved_spec)) {
4382	4359	if (spec->priority < saved_spec->priority &&
4383	4360	spec->priority != EFX_FILTER_PRI_AUTO) {
4384	4361	rc = -EPERM;
@@ -4762,28 +4739,63 @@ static s32 efx_ef10_filter_get_rx_ids(struct efx_nic *efx, · 377 unmodified lines
4762	4739	static bool efx_ef10_filter_rfs_expire_one(struct efx_nic *efx, u32 flow_id,
4763	4740	unsigned int filter_idx)
4764	4741	{
	4742	+ struct efx_filter_spec *spec, saved_spec;
4765	4743	struct efx_ef10_filter_table *table;
4766		− struct efx_filter_spec *spec;
4767		− bool ret;
	4744	+ struct efx_arfs_rule *rule = NULL;
	4745	+ bool ret = true, force = false;
	4746	+ u16 arfs_id;
4768	4747
4769	4748	down_read(&efx->filter_sem);
4770	4749	table = efx->filter_state;
4771	4750	down_write(&table->lock);
4772	4751	spec = efx_ef10_filter_entry_spec(table, filter_idx);
4773	4752
4774		− if (!spec \|\| spec->priority != EFX_FILTER_PRI_HINT) {
4775		− ret = true;
	4753	+ if (!spec \|\| spec->priority != EFX_FILTER_PRI_HINT)
4776	4754	goto out_unlock;
4777		− }
4778	4755
4779		− if (!rps_may_expire_flow(efx->net_dev, spec->dmaq_id,
4780		− flow_id, filter_idx)) {
	4756	+ spin_lock_bh(&efx->rps_hash_lock);
	4757	+ if (!efx->rps_hash_table) {
	4758	+ /* In the absence of the table, we always return 0 to ARFS. */
	4759	+ arfs_id = 0;
	4760	+ } else {
	4761	+ rule = efx_rps_hash_find(efx, spec);
	4762	+ if (!rule)
	4763	+ /* ARFS table doesn't know of this filter, so remove it */
	4764	+ goto expire;
	4765	+ arfs_id = rule->arfs_id;
	4766	+ ret = efx_rps_check_rule(rule, filter_idx, &force);
	4767	+ if (force)
	4768	+ goto expire;
	4769	+ if (!ret) {
	4770	+ spin_unlock_bh(&efx->rps_hash_lock);
	4771	+ goto out_unlock;
	4772	+ }
	4773	+ }
	4774	+ if (!rps_may_expire_flow(efx->net_dev, spec->dmaq_id, flow_id, arfs_id))
4781	4775	ret = false;
4782		− goto out_unlock;
	4776	+ else if (rule)
	4777	+ rule->filter_id = EFX_ARFS_FILTER_ID_REMOVING;
	4778	+expire:
	4779	+ saved_spec = spec; / remove operation will kfree spec */
	4780	+ spin_unlock_bh(&efx->rps_hash_lock);
	4781	+ /* At this point (since we dropped the lock), another thread might queue
	4782	+ * up a fresh insertion request (but the actual insertion will be held
	4783	+ * up by our possession of the filter table lock). In that case, it
	4784	+ * will set rule->filter_id to EFX_ARFS_FILTER_ID_PENDING, meaning that
	4785	+ * the rule is not removed by efx_rps_hash_del() below.
	4786	+ */
	4787	+ if (ret)
	4788	+ ret = efx_ef10_filter_remove_internal(efx, 1U << spec->priority,
	4789	+ filter_idx, true) == 0;
	4790	+ /* While we can't safely dereference rule (we dropped the lock), we can
	4791	+ * still test it for NULL.
	4792	+ */
	4793	+ if (ret && rule) {
	4794	+ /* Expiring, so remove entry from ARFS table */
	4795	+ spin_lock_bh(&efx->rps_hash_lock);
	4796	+ efx_rps_hash_del(efx, &saved_spec);
	4797	+ spin_unlock_bh(&efx->rps_hash_lock);
4783	4798	}
4784		−
4785		− ret = efx_ef10_filter_remove_internal(efx, 1U << spec->priority,
4786		− filter_idx, true) == 0;
4787	4799	out_unlock:
4788	4800	up_write(&table->lock);
4789	4801	up_read(&efx->filter_sem);
@@ -5265,7 +5277,7 @@ static int efx_ef10_filter_insert_addr_list(struct efx_nic *efx, · 475 unmodified lines
5265	5277	ids = vlan->uc;
5266	5278	}
5267	5279
5268		− filter_flags = efx_rss_enabled(efx) ? EFX_FILTER_FLAG_RX_RSS : 0;
	5280	+ filter_flags = efx_rss_active(&efx->rss_context) ? EFX_FILTER_FLAG_RX_RSS : 0;
5269	5281
5270	5282	/* Insert/renew filters */
5271	5283	for (i = 0; i < addr_count; i++) {
@@ -5334,7 +5346,7 @@ static int efx_ef10_filter_insert_def(struct efx_nic *efx, · 62 unmodified lines
5334	5346	int rc;
5335	5347	u16 *id;
5336	5348
5337		− filter_flags = efx_rss_enabled(efx) ? EFX_FILTER_FLAG_RX_RSS : 0;
	5349	+ filter_flags = efx_rss_active(&efx->rss_context) ? EFX_FILTER_FLAG_RX_RSS : 0;
5338	5350
5339	5351	efx_filter_init_rx(&spec, EFX_FILTER_PRI_AUTO, filter_flags, 0);
5340	5352

drivers/net/ethernet/sfc/efx.c+143−0

@@ -3027,6 +3027,10 @@ static int efx_init_struct(struct efx_nic *efx, · 3026 unmodified lines
3027	3027	mutex_init(&efx->mac_lock);
3028	3028	#ifdef CONFIG_RFS_ACCEL
3029	3029	mutex_init(&efx->rps_mutex);
	3030	+ spin_lock_init(&efx->rps_hash_lock);
	3031	+ /* Failure to allocate is not fatal, but may degrade ARFS performance */
	3032	+ efx->rps_hash_table = kcalloc(EFX_ARFS_HASH_TABLE_SIZE,
	3033	+ sizeof(*efx->rps_hash_table), GFP_KERNEL);
3030	3034	#endif
3031	3035	efx->phy_op = &efx_dummy_phy_operations;
3032	3036	efx->mdio.dev = net_dev;
@@ -3070,6 +3074,10 @@ static void efx_fini_struct(struct efx_nic *efx) · 37 unmodified lines
3070	3074	{
3071	3075	int i;
3072	3076
	3077	+#ifdef CONFIG_RFS_ACCEL
	3078	+ kfree(efx->rps_hash_table);
	3079	+#endif
	3080	+
3073	3081	for (i = 0; i < EFX_MAX_CHANNELS; i++)
3074	3082	kfree(efx->channel[i]);
3075	3083
@@ -3092,6 +3100,141 @@ void efx_update_sw_stats(struct efx_nic efx, u64 stats) · 16 unmodified lines
3092	3100	stats[GENERIC_STAT_rx_noskb_drops] = atomic_read(&efx->n_rx_noskb_drops);
3093	3101	}
3094	3102
	3103	+bool efx_filter_spec_equal(const struct efx_filter_spec *left,
	3104	+ const struct efx_filter_spec *right)
	3105	+{
	3106	+ if ((left->match_flags ^ right->match_flags) \|
	3107	+ ((left->flags ^ right->flags) &
	3108	+ (EFX_FILTER_FLAG_RX \| EFX_FILTER_FLAG_TX)))
	3109	+ return false;
	3110	+
	3111	+ return memcmp(&left->outer_vid, &right->outer_vid,
	3112	+ sizeof(struct efx_filter_spec) -
	3113	+ offsetof(struct efx_filter_spec, outer_vid)) == 0;
	3114	+}
	3115	+
	3116	+u32 efx_filter_spec_hash(const struct efx_filter_spec *spec)
	3117	+{
	3118	+ BUILD_BUG_ON(offsetof(struct efx_filter_spec, outer_vid) & 3);
	3119	+ return jhash2((const u32 *)&spec->outer_vid,
	3120	+ (sizeof(struct efx_filter_spec) -
	3121	+ offsetof(struct efx_filter_spec, outer_vid)) / 4,
	3122	+ 0);
	3123	+}
	3124	+
	3125	+#ifdef CONFIG_RFS_ACCEL
	3126	+bool efx_rps_check_rule(struct efx_arfs_rule *rule, unsigned int filter_idx,
	3127	+ bool *force)
	3128	+{
	3129	+ if (rule->filter_id == EFX_ARFS_FILTER_ID_PENDING) {
	3130	+ /* ARFS is currently updating this entry, leave it */
	3131	+ return false;
	3132	+ }
	3133	+ if (rule->filter_id == EFX_ARFS_FILTER_ID_ERROR) {
	3134	+ /* ARFS tried and failed to update this, so it's probably out
	3135	+ * of date. Remove the filter and the ARFS rule entry.
	3136	+ */
	3137	+ rule->filter_id = EFX_ARFS_FILTER_ID_REMOVING;
	3138	+ *force = true;
	3139	+ return true;
	3140	+ } else if (WARN_ON(rule->filter_id != filter_idx)) { /* can't happen */
	3141	+ /* ARFS has moved on, so old filter is not needed. Since we did
	3142	+ * not mark the rule with EFX_ARFS_FILTER_ID_REMOVING, it will
	3143	+ * not be removed by efx_rps_hash_del() subsequently.
	3144	+ */
	3145	+ *force = true;
	3146	+ return true;
	3147	+ }
	3148	+ /* Remove it iff ARFS wants to. */
	3149	+ return true;
	3150	+}
	3151	+
	3152	+struct hlist_head efx_rps_hash_bucket(struct efx_nic efx,
	3153	+ const struct efx_filter_spec *spec)
	3154	+{
	3155	+ u32 hash = efx_filter_spec_hash(spec);
	3156	+
	3157	+ WARN_ON(!spin_is_locked(&efx->rps_hash_lock));
	3158	+ if (!efx->rps_hash_table)
	3159	+ return NULL;
	3160	+ return &efx->rps_hash_table[hash % EFX_ARFS_HASH_TABLE_SIZE];
	3161	+}
	3162	+
	3163	+struct efx_arfs_rule efx_rps_hash_find(struct efx_nic efx,
	3164	+ const struct efx_filter_spec *spec)
	3165	+{
	3166	+ struct efx_arfs_rule *rule;
	3167	+ struct hlist_head *head;
	3168	+ struct hlist_node *node;
	3169	+
	3170	+ head = efx_rps_hash_bucket(efx, spec);
	3171	+ if (!head)
	3172	+ return NULL;
	3173	+ hlist_for_each(node, head) {
	3174	+ rule = container_of(node, struct efx_arfs_rule, node);
	3175	+ if (efx_filter_spec_equal(spec, &rule->spec))
	3176	+ return rule;
	3177	+ }
	3178	+ return NULL;
	3179	+}
	3180	+
	3181	+struct efx_arfs_rule efx_rps_hash_add(struct efx_nic efx,
	3182	+ const struct efx_filter_spec *spec,
	3183	+ bool *new)
	3184	+{
	3185	+ struct efx_arfs_rule *rule;
	3186	+ struct hlist_head *head;
	3187	+ struct hlist_node *node;
	3188	+
	3189	+ head = efx_rps_hash_bucket(efx, spec);
	3190	+ if (!head)
	3191	+ return NULL;
	3192	+ hlist_for_each(node, head) {
	3193	+ rule = container_of(node, struct efx_arfs_rule, node);
	3194	+ if (efx_filter_spec_equal(spec, &rule->spec)) {
	3195	+ *new = false;
	3196	+ return rule;
	3197	+ }
	3198	+ }
	3199	+ rule = kmalloc(sizeof(*rule), GFP_ATOMIC);
	3200	+ *new = true;
	3201	+ if (rule) {
	3202	+ memcpy(&rule->spec, spec, sizeof(rule->spec));
	3203	+ hlist_add_head(&rule->node, head);
	3204	+ }
	3205	+ return rule;
	3206	+}
	3207	+
	3208	+void efx_rps_hash_del(struct efx_nic efx, const struct efx_filter_spec spec)
	3209	+{
	3210	+ struct efx_arfs_rule *rule;
	3211	+ struct hlist_head *head;
	3212	+ struct hlist_node *node;
	3213	+
	3214	+ head = efx_rps_hash_bucket(efx, spec);
	3215	+ if (WARN_ON(!head))
	3216	+ return;
	3217	+ hlist_for_each(node, head) {
	3218	+ rule = container_of(node, struct efx_arfs_rule, node);
	3219	+ if (efx_filter_spec_equal(spec, &rule->spec)) {
	3220	+ /* Someone already reused the entry. We know that if
	3221	+ * this check doesn't fire (i.e. filter_id == REMOVING)
	3222	+ * then the REMOVING mark was put there by our caller,
	3223	+ * because caller is holding a lock on filter table and
	3224	+ * only holders of that lock set REMOVING.
	3225	+ */
	3226	+ if (rule->filter_id != EFX_ARFS_FILTER_ID_REMOVING)
	3227	+ return;
	3228	+ hlist_del(node);
	3229	+ kfree(rule);
	3230	+ return;
	3231	+ }
	3232	+ }
	3233	+ /* We didn't find it. */
	3234	+ WARN_ON(1);
	3235	+}
	3236	+#endif
	3237	+
3095	3238	/* RSS contexts. We're using linked lists and crappy O(n) algorithms, because
3096	3239	* (a) this is an infrequent control-plane operation and (b) n is small (max 64)
3097	3240	*/

drivers/net/ethernet/sfc/efx.h+21−0

@@ -186,6 +186,27 @@ static inline void efx_filter_rfs_expire(struct work_struct *data) {} · 185 unmodified lines
186	186	#endif
187	187	bool efx_filter_is_mc_recipient(const struct efx_filter_spec *spec);
188	188
	189	+bool efx_filter_spec_equal(const struct efx_filter_spec *left,
	190	+ const struct efx_filter_spec *right);
	191	+u32 efx_filter_spec_hash(const struct efx_filter_spec *spec);
	192	+
	193	+#ifdef CONFIG_RFS_ACCEL
	194	+bool efx_rps_check_rule(struct efx_arfs_rule *rule, unsigned int filter_idx,
	195	+ bool *force);
	196	+
	197	+struct efx_arfs_rule efx_rps_hash_find(struct efx_nic efx,
	198	+ const struct efx_filter_spec *spec);
	199	+
	200	+/* @new is written to indicate if entry was newly added (true) or if an old
	201	+ * entry was found and returned (false).
	202	+ */
	203	+struct efx_arfs_rule efx_rps_hash_add(struct efx_nic efx,
	204	+ const struct efx_filter_spec *spec,
	205	+ bool *new);
	206	+
	207	+void efx_rps_hash_del(struct efx_nic efx, const struct efx_filter_spec spec);
	208	+#endif
	209	+
189	210	/* RSS contexts */
190	211	struct efx_rss_context efx_alloc_rss_context_entry(struct efx_nic efx);
191	212	struct efx_rss_context efx_find_rss_context_entry(struct efx_nic efx, u32 id);

drivers/net/ethernet/sfc/farch.c+34−7

@@ -2905,18 +2905,45 @@ bool efx_farch_filter_rfs_expire_one(struct efx_nic *efx, u32 flow_id, · 2904 unmodified lines
2905	2905	{
2906	2906	struct efx_farch_filter_state *state = efx->filter_state;
2907	2907	struct efx_farch_filter_table *table;
2908		− bool ret = false;
	2908	+ bool ret = false, force = false;
	2909	+ u16 arfs_id;
2909	2910
2910	2911	down_write(&state->lock);
	2912	+ spin_lock_bh(&efx->rps_hash_lock);
2911	2913	table = &state->table[EFX_FARCH_FILTER_TABLE_RX_IP];
2912	2914	if (test_bit(index, table->used_bitmap) &&
2913		− table->spec[index].priority == EFX_FILTER_PRI_HINT &&
2914		− rps_may_expire_flow(efx->net_dev, table->spec[index].dmaq_id,
2915		− flow_id, index)) {
2916		− efx_farch_filter_table_clear_entry(efx, table, index);
2917		− ret = true;
	2915	+ table->spec[index].priority == EFX_FILTER_PRI_HINT) {
	2916	+ struct efx_arfs_rule *rule = NULL;
	2917	+ struct efx_filter_spec spec;
	2918	+
	2919	+ efx_farch_filter_to_gen_spec(&spec, &table->spec[index]);
	2920	+ if (!efx->rps_hash_table) {
	2921	+ /* In the absence of the table, we always returned 0 to
	2922	+ * ARFS, so use the same to query it.
	2923	+ */
	2924	+ arfs_id = 0;
	2925	+ } else {
	2926	+ rule = efx_rps_hash_find(efx, &spec);
	2927	+ if (!rule) {
	2928	+ /* ARFS table doesn't know of this filter, remove it */
	2929	+ force = true;
	2930	+ } else {
	2931	+ arfs_id = rule->arfs_id;
	2932	+ if (!efx_rps_check_rule(rule, index, &force))
	2933	+ goto out_unlock;
	2934	+ }
	2935	+ }
	2936	+ if (force \|\| rps_may_expire_flow(efx->net_dev, spec.dmaq_id,
	2937	+ flow_id, arfs_id)) {
	2938	+ if (rule)
	2939	+ rule->filter_id = EFX_ARFS_FILTER_ID_REMOVING;
	2940	+ efx_rps_hash_del(efx, &spec);
	2941	+ efx_farch_filter_table_clear_entry(efx, table, index);
	2942	+ ret = true;
	2943	+ }
2918	2944	}
2919		−
	2945	+out_unlock:
	2946	+ spin_unlock_bh(&efx->rps_hash_lock);
2920	2947	up_write(&state->lock);
2921	2948	return ret;
2922	2949	}

drivers/net/ethernet/sfc/net_driver.h+61−0

@@ -733,6 +733,56 @@ struct efx_rss_context { · 732 unmodified lines
733	733	u32 rx_indir_table[128];
734	734	};
735	735
	736	+#ifdef CONFIG_RFS_ACCEL
	737	+/* Order of these is important, since filter_id >= %EFX_ARFS_FILTER_ID_PENDING
	738	+ * is used to test if filter does or will exist.
	739	+ */
	740	+#define EFX_ARFS_FILTER_ID_PENDING -1
	741	+#define EFX_ARFS_FILTER_ID_ERROR -2
	742	+#define EFX_ARFS_FILTER_ID_REMOVING -3
	743	+/**
	744	+ * struct efx_arfs_rule - record of an ARFS filter and its IDs
	745	+ * @node: linkage into hash table
	746	+ * @spec: details of the filter (used as key for hash table). Use efx->type to
	747	+ * determine which member to use.
	748	+ * @rxq_index: channel to which the filter will steer traffic.
	749	+ * @arfs_id: filter ID which was returned to ARFS
	750	+ * @filter_id: index in software filter table. May be
	751	+ * %EFX_ARFS_FILTER_ID_PENDING if filter was not inserted yet,
	752	+ * %EFX_ARFS_FILTER_ID_ERROR if filter insertion failed, or
	753	+ * %EFX_ARFS_FILTER_ID_REMOVING if expiry is currently removing the filter.
	754	+ */
	755	+struct efx_arfs_rule {
	756	+ struct hlist_node node;
	757	+ struct efx_filter_spec spec;
	758	+ u16 rxq_index;
	759	+ u16 arfs_id;
	760	+ s32 filter_id;
	761	+};
	762	+
	763	+/* Size chosen so that the table is one page (4kB) */
	764	+#define EFX_ARFS_HASH_TABLE_SIZE 512
	765	+
	766	+/**
	767	+ * struct efx_async_filter_insertion - Request to asynchronously insert a filter
	768	+ * @net_dev: Reference to the netdevice
	769	+ * @spec: The filter to insert
	770	+ * @work: Workitem for this request
	771	+ * @rxq_index: Identifies the channel for which this request was made
	772	+ * @flow_id: Identifies the kernel-side flow for which this request was made
	773	+ */
	774	+struct efx_async_filter_insertion {
	775	+ struct net_device *net_dev;
	776	+ struct efx_filter_spec spec;
	777	+ struct work_struct work;
	778	+ u16 rxq_index;
	779	+ u32 flow_id;
	780	+};
	781	+
	782	+/* Maximum number of ARFS workitems that may be in flight on an efx_nic */
	783	+#define EFX_RPS_MAX_IN_FLIGHT 8
	784	+#endif /* CONFIG_RFS_ACCEL */
	785	+
736	786	/**
737	787	* struct efx_nic - an Efx NIC
738	788	* @name: Device name (net device name or bus id before net device registered)
@@ -850,6 +900,12 @@ struct efx_rss_context { · 111 unmodified lines
850	900	* @rps_expire_channel: Next channel to check for expiry
851	901	* @rps_expire_index: Next index to check for expiry in
852	902	* @rps_expire_channel's @rps_flow_id
	903	+ * @rps_slot_map: bitmap of in-flight entries in @rps_slot
	904	+ * @rps_slot: array of ARFS insertion requests for efx_filter_rfs_work()
	905	+ * @rps_hash_lock: Protects ARFS filter mapping state (@rps_hash_table and
	906	+ * @rps_next_id).
	907	+ * @rps_hash_table: Mapping between ARFS filters and their various IDs
	908	+ * @rps_next_id: next arfs_id for an ARFS filter
853	909	* @active_queues: Count of RX and TX queues that haven't been flushed and drained.
854	910	* @rxq_flush_pending: Count of number of receive queues that need to be flushed.
855	911	* Decremented when the efx_flush_rx_queue() is called.
@@ -1004,6 +1060,11 @@ struct efx_nic { · 148 unmodified lines
1004	1060	struct mutex rps_mutex;
1005	1061	unsigned int rps_expire_channel;
1006	1062	unsigned int rps_expire_index;
	1063	+ unsigned long rps_slot_map;
	1064	+ struct efx_async_filter_insertion rps_slot[EFX_RPS_MAX_IN_FLIGHT];
	1065	+ spinlock_t rps_hash_lock;
	1066	+ struct hlist_head *rps_hash_table;
	1067	+ u32 rps_next_id;
1007	1068	#endif
1008	1069
1009	1070	atomic_t active_queues;

drivers/net/ethernet/sfc/rx.c+90−34

@@ -827,31 +827,38 @@ MODULE_PARM_DESC(rx_refill_threshold, · 826 unmodified lines
827	827
828	828	#ifdef CONFIG_RFS_ACCEL
829	829
830		−/**
831		− * struct efx_async_filter_insertion - Request to asynchronously insert a filter
832		− * @net_dev: Reference to the netdevice
833		− * @spec: The filter to insert
834		− * @work: Workitem for this request
835		− * @rxq_index: Identifies the channel for which this request was made
836		− * @flow_id: Identifies the kernel-side flow for which this request was made
837		− */
838		−struct efx_async_filter_insertion {
839		− struct net_device *net_dev;
840		− struct efx_filter_spec spec;
841		− struct work_struct work;
842		− u16 rxq_index;
843		− u32 flow_id;
844		−};
845		−
846	830	static void efx_filter_rfs_work(struct work_struct *data)
847	831	{
848	832	struct efx_async_filter_insertion *req = container_of(data, struct efx_async_filter_insertion,
849	833	work);
850	834	struct efx_nic *efx = netdev_priv(req->net_dev);
851	835	struct efx_channel *channel = efx_get_channel(efx, req->rxq_index);
	836	+ int slot_idx = req - efx->rps_slot;
	837	+ struct efx_arfs_rule *rule;
	838	+ u16 arfs_id = 0;
852	839	int rc;
853	840
854		− rc = efx->type->filter_insert(efx, &req->spec, false);
	841	+ rc = efx->type->filter_insert(efx, &req->spec, true);
	842	+ if (rc >= 0)
	843	+ rc %= efx->type->max_rx_ip_filters;
	844	+ if (efx->rps_hash_table) {
	845	+ spin_lock_bh(&efx->rps_hash_lock);
	846	+ rule = efx_rps_hash_find(efx, &req->spec);
	847	+ /* The rule might have already gone, if someone else's request
	848	+ * for the same spec was already worked and then expired before
	849	+ * we got around to our work. In that case we have nothing
	850	+ * tying us to an arfs_id, meaning that as soon as the filter
	851	+ * is considered for expiry it will be removed.
	852	+ */
	853	+ if (rule) {
	854	+ if (rc < 0)
	855	+ rule->filter_id = EFX_ARFS_FILTER_ID_ERROR;
	856	+ else
	857	+ rule->filter_id = rc;
	858	+ arfs_id = rule->arfs_id;
	859	+ }
	860	+ spin_unlock_bh(&efx->rps_hash_lock);
	861	+ }
855	862	if (rc >= 0) {
856	863	/* Remember this so we can check whether to expire the filter
857	864	* later.
@@ -863,23 +870,23 @@ static void efx_filter_rfs_work(struct work_struct *data) · 5 unmodified lines
863	870
864	871	if (req->spec.ether_type == htons(ETH_P_IP))
865	872	netif_info(efx, rx_status, efx->net_dev,
866		− "steering %s %pI4:%u:%pI4:%u to queue %u [flow %u filter %d]\n",
	873	+ "steering %s %pI4:%u:%pI4:%u to queue %u [flow %u filter %d id %u]\n",
867	874	(req->spec.ip_proto == IPPROTO_TCP) ? "TCP" : "UDP",
868	875	req->spec.rem_host, ntohs(req->spec.rem_port),
869	876	req->spec.loc_host, ntohs(req->spec.loc_port),
870		− req->rxq_index, req->flow_id, rc);
	877	+ req->rxq_index, req->flow_id, rc, arfs_id);
871	878	else
872	879	netif_info(efx, rx_status, efx->net_dev,
873		− "steering %s [%pI6]:%u:[%pI6]:%u to queue %u [flow %u filter %d]\n",
	880	+ "steering %s [%pI6]:%u:[%pI6]:%u to queue %u [flow %u filter %d id %u]\n",
874	881	(req->spec.ip_proto == IPPROTO_TCP) ? "TCP" : "UDP",
875	882	req->spec.rem_host, ntohs(req->spec.rem_port),
876	883	req->spec.loc_host, ntohs(req->spec.loc_port),
877		− req->rxq_index, req->flow_id, rc);
	884	+ req->rxq_index, req->flow_id, rc, arfs_id);
878	885	}
879	886
880	887	/* Release references */
	888	+ clear_bit(slot_idx, &efx->rps_slot_map);
881	889	dev_put(req->net_dev);
882		− kfree(req);
883	890	}
884	891
885	892	int efx_filter_rfs(struct net_device net_dev, const struct sk_buff skb,
@@ -887,23 +894,39 @@ int efx_filter_rfs(struct net_device net_dev, const struct sk_buff skb, · 1 unmodified line
887	894	{
888	895	struct efx_nic *efx = netdev_priv(net_dev);
889	896	struct efx_async_filter_insertion *req;
	897	+ struct efx_arfs_rule *rule;
890	898	struct flow_keys fk;
	899	+ int slot_idx;
	900	+ bool new;
	901	+ int rc;
891	902
892		− if (flow_id == RPS_FLOW_ID_INVALID)
893		− return -EINVAL;
	903	+ /* find a free slot */
	904	+ for (slot_idx = 0; slot_idx < EFX_RPS_MAX_IN_FLIGHT; slot_idx++)
	905	+ if (!test_and_set_bit(slot_idx, &efx->rps_slot_map))
	906	+ break;
	907	+ if (slot_idx >= EFX_RPS_MAX_IN_FLIGHT)
	908	+ return -EBUSY;
894	909
895		− if (!skb_flow_dissect_flow_keys(skb, &fk, 0))
896		− return -EPROTONOSUPPORT;
	910	+ if (flow_id == RPS_FLOW_ID_INVALID) {
	911	+ rc = -EINVAL;
	912	+ goto out_clear;
	913	+ }
897	914
898		− if (fk.basic.n_proto != htons(ETH_P_IP) && fk.basic.n_proto != htons(ETH_P_IPV6))
899		− return -EPROTONOSUPPORT;
900		− if (fk.control.flags & FLOW_DIS_IS_FRAGMENT)
901		− return -EPROTONOSUPPORT;
	915	+ if (!skb_flow_dissect_flow_keys(skb, &fk, 0)) {
	916	+ rc = -EPROTONOSUPPORT;
	917	+ goto out_clear;
	918	+ }
902	919
903		− req = kmalloc(sizeof(*req), GFP_ATOMIC);
904		− if (!req)
905		− return -ENOMEM;
	920	+ if (fk.basic.n_proto != htons(ETH_P_IP) && fk.basic.n_proto != htons(ETH_P_IPV6)) {
	921	+ rc = -EPROTONOSUPPORT;
	922	+ goto out_clear;
	923	+ }
	924	+ if (fk.control.flags & FLOW_DIS_IS_FRAGMENT) {
	925	+ rc = -EPROTONOSUPPORT;
	926	+ goto out_clear;
	927	+ }
906	928
	929	+ req = efx->rps_slot + slot_idx;
907	930	efx_filter_init_rx(&req->spec, EFX_FILTER_PRI_HINT,
908	931	efx->rx_scatter ? EFX_FILTER_FLAG_RX_SCATTER : 0,
909	932	rxq_index);
@@ -927,12 +950,45 @@ int efx_filter_rfs(struct net_device net_dev, const struct sk_buff skb, · 17 unmodified lines
927	950	req->spec.rem_port = fk.ports.src;
928	951	req->spec.loc_port = fk.ports.dst;
929	952
	953	+ if (efx->rps_hash_table) {
	954	+ /* Add it to ARFS hash table */
	955	+ spin_lock(&efx->rps_hash_lock);
	956	+ rule = efx_rps_hash_add(efx, &req->spec, &new);
	957	+ if (!rule) {
	958	+ rc = -ENOMEM;
	959	+ goto out_unlock;
	960	+ }
	961	+ if (new)
	962	+ rule->arfs_id = efx->rps_next_id++ % RPS_NO_FILTER;
	963	+ rc = rule->arfs_id;
	964	+ /* Skip if existing or pending filter already does the right thing */
	965	+ if (!new && rule->rxq_index == rxq_index &&
	966	+ rule->filter_id >= EFX_ARFS_FILTER_ID_PENDING)
	967	+ goto out_unlock;
	968	+ rule->rxq_index = rxq_index;
	969	+ rule->filter_id = EFX_ARFS_FILTER_ID_PENDING;
	970	+ spin_unlock(&efx->rps_hash_lock);
	971	+ } else {
	972	+ /* Without an ARFS hash table, we just use arfs_id 0 for all
	973	+ * filters. This means if multiple flows hash to the same
	974	+ * flow_id, all but the most recently touched will be eligible
	975	+ * for expiry.
	976	+ */
	977	+ rc = 0;
	978	+ }
	979	+
	980	+ /* Queue the request */
930	981	dev_hold(req->net_dev = net_dev);
931	982	INIT_WORK(&req->work, efx_filter_rfs_work);
932	983	req->rxq_index = rxq_index;
933	984	req->flow_id = flow_id;
934	985	schedule_work(&req->work);
935		− return 0;
	986	+ return rc;
	987	+out_unlock:
	988	+ spin_unlock(&efx->rps_hash_lock);
	989	+out_clear:
	990	+ clear_bit(slot_idx, &efx->rps_slot_map);
	991	+ return rc;
936	992	}
937	993
938	994	bool __efx_filter_rfs_expire(struct efx_nic *efx, unsigned int quota)

drivers/net/ethernet/stmicro/stmmac/dwmac4.h+1−1

@@ -347,7 +347,7 @@ enum power_event { · 346 unmodified lines
347	347	#define MTL_RX_OVERFLOW_INT BIT(16)
348	348
349	349	/* Default operating mode of the MAC */
350		−#define GMAC_CORE_INIT (GMAC_CONFIG_JD \| GMAC_CONFIG_PS \| GMAC_CONFIG_ACS \| \
	350	+#define GMAC_CORE_INIT (GMAC_CONFIG_JD \| GMAC_CONFIG_PS \| \
351	351	GMAC_CONFIG_BE \| GMAC_CONFIG_DCRS)
352	352
353	353	/* To dump the core regs excluding the Address Registers */

drivers/net/ethernet/stmicro/stmmac/dwmac4_core.c+0−7

@@ -31,13 +31,6 @@ static void dwmac4_core_init(struct mac_device_info *hw, · 30 unmodified lines
31	31
32	32	value \|= GMAC_CORE_INIT;
33	33
34		− /* Clear ACS bit because Ethernet switch tagging formats such as
35		− * Broadcom tags can look like invalid LLC/SNAP packets and cause the
36		− * hardware to truncate packets on reception.
37		− */
38		− if (netdev_uses_dsa(dev))
39		− value &= ~GMAC_CONFIG_ACS;
40		−
41	34	if (mtu > 1500)
42	35	value \|= GMAC_CONFIG_2K;
43	36	if (mtu > 2000)

drivers/net/ethernet/stmicro/stmmac/stmmac_main.c+6−1

@@ -3495,8 +3495,13 @@ static int stmmac_rx(struct stmmac_priv *priv, int limit, u32 queue) · 3494 unmodified lines
3495	3495
3496	3496	/* ACS is set; GMAC core strips PAD/FCS for IEEE 802.3
3497	3497	* Type frames (LLC/LLC-SNAP)
	3498	+ *
	3499	+ * llc_snap is never checked in GMAC >= 4, so this ACS
	3500	+ * feature is always disabled and packets need to be
	3501	+ * stripped manually.
3498	3502	*/
3499		− if (unlikely(status != llc_snap))
	3503	+ if (unlikely(priv->synopsys_id >= DWMAC_CORE_4_00) \|\|
	3504	+ unlikely(status != llc_snap))
3500	3505	frame_len -= ETH_FCS_LEN;
3501	3506
3502	3507	if (netif_msg_rx_status(priv)) {

drivers/net/ethernet/ti/cpsw.c+3−1

@@ -129,7 +129,7 @@ do { \ · 128 unmodified lines
129	129
130	130	#define RX_PRIORITY_MAPPING 0x76543210
131	131	#define TX_PRIORITY_MAPPING 0x33221100
132		−#define CPDMA_TX_PRIORITY_MAP 0x01234567
	132	+#define CPDMA_TX_PRIORITY_MAP 0x76543210
133	133
134	134	#define CPSW_VLAN_AWARE BIT(1)
135	135	#define CPSW_RX_VLAN_ENCAP BIT(2)
@@ -1340,6 +1340,8 @@ static inline void cpsw_add_dual_emac_def_ale_entries( · 1204 unmodified lines
1340	1340	cpsw_ale_add_ucast(cpsw->ale, priv->mac_addr,
1341	1341	HOST_PORT_NUM, ALE_VLAN \|
1342	1342	ALE_SECURE, slave->port_vlan);
	1343	+ cpsw_ale_control_set(cpsw->ale, slave_port,
	1344	+ ALE_PORT_DROP_UNKNOWN_VLAN, 1);
1343	1345	}
1344	1346
1345	1347	static void soft_reset_slave(struct cpsw_slave *slave)

drivers/net/macsec.c+2−3

@@ -3277,7 +3277,7 @@ static int macsec_newlink(struct net net, struct net_device dev, · 3276 unmodified lines
3277	3277
3278	3278	err = netdev_upper_dev_link(real_dev, dev, extack);
3279	3279	if (err < 0)
3280		− goto put_dev;
	3280	+ goto unregister;
3281	3281
3282	3282	/* need to be already registered so that ->init has run and
3283	3283	* the MAC addr is set
@@ -3316,8 +3316,7 @@ del_dev: · 32 unmodified lines
3316	3316	macsec_del_dev(macsec);
3317	3317	unlink:
3318	3318	netdev_upper_dev_unlink(real_dev, dev);
3319		−put_dev:
3320		− dev_put(real_dev);
	3319	+unregister:
3321	3320	unregister_netdevice(dev);
3322	3321	return err;
3323	3322	}

drivers/net/phy/marvell.c+9−0

@@ -1393,6 +1393,15 @@ static int m88e1318_set_wol(struct phy_device *phydev, · 1392 unmodified lines
1393	1393	if (err < 0)
1394	1394	goto error;
1395	1395
	1396	+ /* If WOL event happened once, the LED[2] interrupt pin
	1397	+ * will not be cleared unless we reading the interrupt status
	1398	+ * register. If interrupts are in use, the normal interrupt
	1399	+ * handling will clear the WOL event. Clear the WOL event
	1400	+ * before enabling it if !phy_interrupt_is_valid()
	1401	+ */
	1402	+ if (!phy_interrupt_is_valid(phydev))
	1403	+ phy_read(phydev, MII_M1011_IEVENT);
	1404	+
1396	1405	/* Enable the WOL interrupt */
1397	1406	err = __phy_modify(phydev, MII_88E1318S_PHY_CSIER, 0,
1398	1407	MII_88E1318S_PHY_CSIER_WOL_EIE);

drivers/net/phy/microchip.c+177−1

@@ -20,6 +20,7 @@ · 19 unmodified lines
20	20	#include <linux/ethtool.h>
21	21	#include <linux/phy.h>
22	22	#include <linux/microchipphy.h>
	23	+#include <linux/delay.h>
23	24
24	25	#define DRIVER_AUTHOR "WOOJUNG HUH <woojung.huh@microchip.com>"
25	26	#define DRIVER_DESC "Microchip LAN88XX PHY driver"
@@ -30,6 +31,16 @@ struct lan88xx_priv { · 4 unmodified lines
30	31	__u32 wolopts;
31	32	};
32	33
	34	+static int lan88xx_read_page(struct phy_device *phydev)
	35	+{
	36	+ return __phy_read(phydev, LAN88XX_EXT_PAGE_ACCESS);
	37	+}
	38	+
	39	+static int lan88xx_write_page(struct phy_device *phydev, int page)
	40	+{
	41	+ return __phy_write(phydev, LAN88XX_EXT_PAGE_ACCESS, page);
	42	+}
	43	+
33	44	static int lan88xx_phy_config_intr(struct phy_device *phydev)
34	45	{
35	46	int rc;
@@ -66,6 +77,150 @@ static int lan88xx_suspend(struct phy_device *phydev) · 30 unmodified lines
66	77	return 0;
67	78	}
68	79
	80	+static int lan88xx_TR_reg_set(struct phy_device *phydev, u16 regaddr,
	81	+ u32 data)
	82	+{
	83	+ int val, save_page, ret = 0;
	84	+ u16 buf;
	85	+
	86	+ /* Save current page */
	87	+ save_page = phy_save_page(phydev);
	88	+ if (save_page < 0) {
	89	+ pr_warn("Failed to get current page\n");
	90	+ goto err;
	91	+ }
	92	+
	93	+ /* Switch to TR page */
	94	+ lan88xx_write_page(phydev, LAN88XX_EXT_PAGE_ACCESS_TR);
	95	+
	96	+ ret = __phy_write(phydev, LAN88XX_EXT_PAGE_TR_LOW_DATA,
	97	+ (data & 0xFFFF));
	98	+ if (ret < 0) {
	99	+ pr_warn("Failed to write TR low data\n");
	100	+ goto err;
	101	+ }
	102	+
	103	+ ret = __phy_write(phydev, LAN88XX_EXT_PAGE_TR_HIGH_DATA,
	104	+ (data & 0x00FF0000) >> 16);
	105	+ if (ret < 0) {
	106	+ pr_warn("Failed to write TR high data\n");
	107	+ goto err;
	108	+ }
	109	+
	110	+ /* Config control bits [15:13] of register */
	111	+ buf = (regaddr & ~(0x3 << 13));/* Clr [14:13] to write data in reg */
	112	+ buf \|= 0x8000; /* Set [15] to Packet transmit */
	113	+
	114	+ ret = __phy_write(phydev, LAN88XX_EXT_PAGE_TR_CR, buf);
	115	+ if (ret < 0) {
	116	+ pr_warn("Failed to write data in reg\n");
	117	+ goto err;
	118	+ }
	119	+
	120	+ usleep_range(1000, 2000);/* Wait for Data to be written */
	121	+ val = __phy_read(phydev, LAN88XX_EXT_PAGE_TR_CR);
	122	+ if (!(val & 0x8000))
	123	+ pr_warn("TR Register[0x%X] configuration failed\n", regaddr);
	124	+err:
	125	+ return phy_restore_page(phydev, save_page, ret);
	126	+}
	127	+
	128	+static void lan88xx_config_TR_regs(struct phy_device *phydev)
	129	+{
	130	+ int err;
	131	+
	132	+ /* Get access to Channel 0x1, Node 0xF , Register 0x01.
	133	+ * Write 24-bit value 0x12B00A to register. Setting MrvlTrFix1000Kf,
	134	+ * MrvlTrFix1000Kp, MasterEnableTR bits.
	135	+ */
	136	+ err = lan88xx_TR_reg_set(phydev, 0x0F82, 0x12B00A);
	137	+ if (err < 0)
	138	+ pr_warn("Failed to Set Register[0x0F82]\n");
	139	+
	140	+ /* Get access to Channel b'10, Node b'1101, Register 0x06.
	141	+ * Write 24-bit value 0xD2C46F to register. Setting SSTrKf1000Slv,
	142	+ * SSTrKp1000Mas bits.
	143	+ */
	144	+ err = lan88xx_TR_reg_set(phydev, 0x168C, 0xD2C46F);
	145	+ if (err < 0)
	146	+ pr_warn("Failed to Set Register[0x168C]\n");
	147	+
	148	+ /* Get access to Channel b'10, Node b'1111, Register 0x11.
	149	+ * Write 24-bit value 0x620 to register. Setting rem_upd_done_thresh
	150	+ * bits
	151	+ */
	152	+ err = lan88xx_TR_reg_set(phydev, 0x17A2, 0x620);
	153	+ if (err < 0)
	154	+ pr_warn("Failed to Set Register[0x17A2]\n");
	155	+
	156	+ /* Get access to Channel b'10, Node b'1101, Register 0x10.
	157	+ * Write 24-bit value 0xEEFFDD to register. Setting
	158	+ * eee_TrKp1Long_1000, eee_TrKp2Long_1000, eee_TrKp3Long_1000,
	159	+ * eee_TrKp1Short_1000,eee_TrKp2Short_1000, eee_TrKp3Short_1000 bits.
	160	+ */
	161	+ err = lan88xx_TR_reg_set(phydev, 0x16A0, 0xEEFFDD);
	162	+ if (err < 0)
	163	+ pr_warn("Failed to Set Register[0x16A0]\n");
	164	+
	165	+ /* Get access to Channel b'10, Node b'1101, Register 0x13.
	166	+ * Write 24-bit value 0x071448 to register. Setting
	167	+ * slv_lpi_tr_tmr_val1, slv_lpi_tr_tmr_val2 bits.
	168	+ */
	169	+ err = lan88xx_TR_reg_set(phydev, 0x16A6, 0x071448);
	170	+ if (err < 0)
	171	+ pr_warn("Failed to Set Register[0x16A6]\n");
	172	+
	173	+ /* Get access to Channel b'10, Node b'1101, Register 0x12.
	174	+ * Write 24-bit value 0x13132F to register. Setting
	175	+ * slv_sigdet_timer_val1, slv_sigdet_timer_val2 bits.
	176	+ */
	177	+ err = lan88xx_TR_reg_set(phydev, 0x16A4, 0x13132F);
	178	+ if (err < 0)
	179	+ pr_warn("Failed to Set Register[0x16A4]\n");
	180	+
	181	+ /* Get access to Channel b'10, Node b'1101, Register 0x14.
	182	+ * Write 24-bit value 0x0 to register. Setting eee_3level_delay,
	183	+ * eee_TrKf_freeze_delay bits.
	184	+ */
	185	+ err = lan88xx_TR_reg_set(phydev, 0x16A8, 0x0);
	186	+ if (err < 0)
	187	+ pr_warn("Failed to Set Register[0x16A8]\n");
	188	+
	189	+ /* Get access to Channel b'01, Node b'1111, Register 0x34.
	190	+ * Write 24-bit value 0x91B06C to register. Setting
	191	+ * FastMseSearchThreshLong1000, FastMseSearchThreshShort1000,
	192	+ * FastMseSearchUpdGain1000 bits.
	193	+ */
	194	+ err = lan88xx_TR_reg_set(phydev, 0x0FE8, 0x91B06C);
	195	+ if (err < 0)
	196	+ pr_warn("Failed to Set Register[0x0FE8]\n");
	197	+
	198	+ /* Get access to Channel b'01, Node b'1111, Register 0x3E.
	199	+ * Write 24-bit value 0xC0A028 to register. Setting
	200	+ * FastMseKp2ThreshLong1000, FastMseKp2ThreshShort1000,
	201	+ * FastMseKp2UpdGain1000, FastMseKp2ExitEn1000 bits.
	202	+ */
	203	+ err = lan88xx_TR_reg_set(phydev, 0x0FFC, 0xC0A028);
	204	+ if (err < 0)
	205	+ pr_warn("Failed to Set Register[0x0FFC]\n");
	206	+
	207	+ /* Get access to Channel b'01, Node b'1111, Register 0x35.
	208	+ * Write 24-bit value 0x041600 to register. Setting
	209	+ * FastMseSearchPhShNum1000, FastMseSearchClksPerPh1000,
	210	+ * FastMsePhChangeDelay1000 bits.
	211	+ */
	212	+ err = lan88xx_TR_reg_set(phydev, 0x0FEA, 0x041600);
	213	+ if (err < 0)
	214	+ pr_warn("Failed to Set Register[0x0FEA]\n");
	215	+
	216	+ /* Get access to Channel b'10, Node b'1101, Register 0x03.
	217	+ * Write 24-bit value 0x000004 to register. Setting TrFreeze bits.
	218	+ */
	219	+ err = lan88xx_TR_reg_set(phydev, 0x1686, 0x000004);
	220	+ if (err < 0)
	221	+ pr_warn("Failed to Set Register[0x1686]\n");
	222	+}
	223	+
69	224	static int lan88xx_probe(struct phy_device *phydev)
70	225	{
71	226	struct device *dev = &phydev->mdio.dev;
@@ -132,6 +287,25 @@ static void lan88xx_set_mdix(struct phy_device *phydev) · 60 unmodified lines
132	287	phy_write(phydev, LAN88XX_EXT_PAGE_ACCESS, LAN88XX_EXT_PAGE_SPACE_0);
133	288	}
134	289
	290	+static int lan88xx_config_init(struct phy_device *phydev)
	291	+{
	292	+ int val;
	293	+
	294	+ genphy_config_init(phydev);
	295	+ /Zerodetect delay enable /
	296	+ val = phy_read_mmd(phydev, MDIO_MMD_PCS,
	297	+ PHY_ARDENNES_MMD_DEV_3_PHY_CFG);
	298	+ val \|= PHY_ARDENNES_MMD_DEV_3_PHY_CFG_ZD_DLY_EN_;
	299	+
	300	+ phy_write_mmd(phydev, MDIO_MMD_PCS, PHY_ARDENNES_MMD_DEV_3_PHY_CFG,
	301	+ val);
	302	+
	303	+ /* Config DSP registers */
	304	+ lan88xx_config_TR_regs(phydev);
	305	+
	306	+ return 0;
	307	+}
	308	+
135	309	static int lan88xx_config_aneg(struct phy_device *phydev)
136	310	{
137	311	lan88xx_set_mdix(phydev);
@@ -151,7 +325,7 @@ static struct phy_driver microchip_phy_driver[] = { · 13 unmodified lines
151	325	.probe = lan88xx_probe,
152	326	.remove = lan88xx_remove,
153	327
154		− .config_init = genphy_config_init,
	328	+ .config_init = lan88xx_config_init,
155	329	.config_aneg = lan88xx_config_aneg,
156	330
157	331	.ack_interrupt = lan88xx_phy_ack_interrupt,
@@ -160,6 +334,8 @@ static struct phy_driver microchip_phy_driver[] = { · 2 unmodified lines
160	334	.suspend = lan88xx_suspend,
161	335	.resume = genphy_resume,
162	336	.set_wol = lan88xx_set_wol,
	337	+ .read_page = lan88xx_read_page,
	338	+ .write_page = lan88xx_write_page,
163	339	} };
164	340
165	341	module_phy_driver(microchip_phy_driver);

drivers/net/phy/phy_device.c+10−1

@@ -535,8 +535,17 @@ static int get_phy_id(struct mii_bus bus, int addr, u32 phy_id, · 534 unmodified lines
535	535
536	536	/* Grab the bits from PHYIR1, and put them in the upper half */
537	537	phy_reg = mdiobus_read(bus, addr, MII_PHYSID1);
538		− if (phy_reg < 0)
	538	+ if (phy_reg < 0) {
	539	+ /* if there is no device, return without an error so scanning
	540	+ * the bus works properly
	541	+ */
	542	+ if (phy_reg == -EIO \|\| phy_reg == -ENODEV) {
	543	+ *phy_id = 0xffffffff;
	544	+ return 0;
	545	+ }
	546	+
539	547	return -EIO;
	548	+ }
540	549
541	550	*phy_id = (phy_reg & 0xffff) << 16;
542	551

drivers/net/ppp/pppoe.c+4−0

@@ -620,6 +620,10 @@ static int pppoe_connect(struct socket sock, struct sockaddr uservaddr, · 619 unmodified lines
620	620	lock_sock(sk);
621	621
622	622	error = -EINVAL;
	623	+
	624	+ if (sockaddr_len != sizeof(struct sockaddr_pppox))
	625	+ goto end;
	626	+
623	627	if (sp->sa_protocol != PX_PROTO_OE)
624	628	goto end;
625	629

drivers/net/team/team.c+31−7

@@ -261,6 +261,17 @@ static void __team_option_inst_mark_removed_port(struct team *team, · 260 unmodified lines
261	261	}
262	262	}
263	263
	264	+static bool __team_option_inst_tmp_find(const struct list_head *opts,
	265	+ const struct team_option_inst *needle)
	266	+{
	267	+ struct team_option_inst *opt_inst;
	268	+
	269	+ list_for_each_entry(opt_inst, opts, tmp_list)
	270	+ if (opt_inst == needle)
	271	+ return true;
	272	+ return false;
	273	+}
	274	+
264	275	static int __team_options_register(struct team *team,
265	276	const struct team_option *option,
266	277	size_t option_count)
@@ -1061,14 +1072,11 @@ static void team_port_leave(struct team team, struct team_port port) · 794 unmodified lines
1061	1072	}
1062	1073
1063	1074	#ifdef CONFIG_NET_POLL_CONTROLLER
1064		−static int team_port_enable_netpoll(struct team team, struct team_port port)
	1075	+static int __team_port_enable_netpoll(struct team_port *port)
1065	1076	{
1066	1077	struct netpoll *np;
1067	1078	int err;
1068	1079
1069		− if (!team->dev->npinfo)
1070		− return 0;
1071		−
1072	1080	np = kzalloc(sizeof(*np), GFP_KERNEL);
1073	1081	if (!np)
1074	1082	return -ENOMEM;
@@ -1082,6 +1090,14 @@ static int team_port_enable_netpoll(struct team team, struct team_port port) · 7 unmodified lines
1082	1090	return err;
1083	1091	}
1084	1092
	1093	+static int team_port_enable_netpoll(struct team_port *port)
	1094	+{
	1095	+ if (!port->team->dev->npinfo)
	1096	+ return 0;
	1097	+
	1098	+ return __team_port_enable_netpoll(port);
	1099	+}
	1100	+
1085	1101	static void team_port_disable_netpoll(struct team_port *port)
1086	1102	{
1087	1103	struct netpoll *np = port->np;
@@ -1096,7 +1112,7 @@ static void team_port_disable_netpoll(struct team_port *port) · 8 unmodified lines
1096	1112	kfree(np);
1097	1113	}
1098	1114	#else
1099		−static int team_port_enable_netpoll(struct team team, struct team_port port)
	1115	+static int team_port_enable_netpoll(struct team_port *port)
1100	1116	{
1101	1117	return 0;
1102	1118	}
@@ -1210,7 +1226,7 @@ static int team_port_add(struct team team, struct net_device port_dev, · 107 unmodified lines
1210	1226	goto err_vids_add;
1211	1227	}
1212	1228
1213		− err = team_port_enable_netpoll(team, port);
	1229	+ err = team_port_enable_netpoll(port);
1214	1230	if (err) {
1215	1231	netdev_err(dev, "Failed to enable netpoll on device %s\n",
1216	1232	portname);
@@ -1907,7 +1923,7 @@ static int team_netpoll_setup(struct net_device *dev, · 690 unmodified lines
1907	1923
1908	1924	mutex_lock(&team->lock);
1909	1925	list_for_each_entry(port, &team->port_list, list) {
1910		− err = team_port_enable_netpoll(team, port);
	1926	+ err = __team_port_enable_netpoll(port);
1911	1927	if (err) {
1912	1928	__team_netpoll_cleanup(team);
1913	1929	break;
@@ -2568,6 +2584,14 @@ static int team_nl_cmd_options_set(struct sk_buff skb, struct genl_info info) · 654 unmodified lines
2568	2584	if (err)
2569	2585	goto team_put;
2570	2586	opt_inst->changed = true;
	2587	+
	2588	+ /* dumb/evil user-space can send us duplicate opt,
	2589	+ * keep only the last one
	2590	+ */
	2591	+ if (__team_option_inst_tmp_find(&opt_inst_list,
	2592	+ opt_inst))
	2593	+ continue;
	2594	+
2571	2595	list_add(&opt_inst->tmp_list, &opt_inst_list);
2572	2596	}
2573	2597	if (!opt_found) {

drivers/net/tun.c+1−6

@@ -1102,12 +1102,7 @@ static netdev_tx_t tun_net_xmit(struct sk_buff skb, struct net_device dev) · 1101 unmodified lines
1102	1102	goto drop;
1103	1103
1104	1104	len = run_ebpf_filter(tun, skb, len);
1105		−
1106		− /* Trim extra bytes since we may insert vlan proto & TCI
1107		− * in tun_put_user().
1108		− */
1109		− len -= skb_vlan_tag_present(skb) ? sizeof(struct veth) : 0;
1110		− if (len <= 0 \|\| pskb_trim(skb, len))
	1105	+ if (len == 0 \|\| pskb_trim(skb, len))
1111	1106	goto drop;
1112	1107
1113	1108	if (unlikely(skb_orphan_frags_rx(skb, GFP_ATOMIC)))

drivers/net/usb/qmi_wwan.c+14−0

@@ -1098,6 +1098,7 @@ static const struct usb_device_id products[] = { · 1097 unmodified lines
1098	1098	{QMI_FIXED_INTF(0x05c6, 0x9080, 8)},
1099	1099	{QMI_FIXED_INTF(0x05c6, 0x9083, 3)},
1100	1100	{QMI_FIXED_INTF(0x05c6, 0x9084, 4)},
	1101	+ {QMI_FIXED_INTF(0x05c6, 0x90b2, 3)}, /* ublox R410M */
1101	1102	{QMI_FIXED_INTF(0x05c6, 0x920d, 0)},
1102	1103	{QMI_FIXED_INTF(0x05c6, 0x920d, 5)},
1103	1104	{QMI_QUIRK_SET_DTR(0x05c6, 0x9625, 4)}, /* YUGA CLM920-NC5 */
@@ -1107,6 +1108,7 @@ static const struct usb_device_id products[] = { · 3 unmodified lines
1107	1108	{QMI_FIXED_INTF(0x1435, 0xd181, 3)}, /* Wistron NeWeb D18Q1 */
1108	1109	{QMI_FIXED_INTF(0x1435, 0xd181, 4)}, /* Wistron NeWeb D18Q1 */
1109	1110	{QMI_FIXED_INTF(0x1435, 0xd181, 5)}, /* Wistron NeWeb D18Q1 */
	1111	+ {QMI_FIXED_INTF(0x1435, 0xd191, 4)}, /* Wistron NeWeb D19Q1 */
1110	1112	{QMI_FIXED_INTF(0x16d8, 0x6003, 0)}, /* CMOTech 6003 */
1111	1113	{QMI_FIXED_INTF(0x16d8, 0x6007, 0)}, /* CMOTech CHE-628S */
1112	1114	{QMI_FIXED_INTF(0x16d8, 0x6008, 0)}, /* CMOTech CMU-301 */
@@ -1342,6 +1344,18 @@ static int qmi_wwan_probe(struct usb_interface *intf, · 229 unmodified lines
1342	1344	id->driver_info = (unsigned long)&qmi_wwan_info;
1343	1345	}
1344	1346
	1347	+ /* There are devices where the same interface number can be
	1348	+ * configured as different functions. We should only bind to
	1349	+ * vendor specific functions when matching on interface number
	1350	+ */
	1351	+ if (id->match_flags & USB_DEVICE_ID_MATCH_INT_NUMBER &&
	1352	+ desc->bInterfaceClass != USB_CLASS_VENDOR_SPEC) {
	1353	+ dev_dbg(&intf->dev,
	1354	+ "Rejecting interface number match for class %02x\n",
	1355	+ desc->bInterfaceClass);
	1356	+ return -ENODEV;
	1357	+ }
	1358	+
1345	1359	/* Quectel EC20 quirk where we've QMI on interface 4 instead of 0 */
1346	1360	if (quectel_ec20_detected(intf) && desc->bInterfaceNumber == 0) {
1347	1361	dev_dbg(&intf->dev, "Quectel EC20 quirk, skipping interface 0\n");

drivers/net/virtio_net.c+48−31

@@ -147,6 +147,17 @@ struct receive_queue { · 146 unmodified lines
147	147	struct xdp_rxq_info xdp_rxq;
148	148	};
149	149
	150	+/* Control VQ buffers: protected by the rtnl lock */
	151	+struct control_buf {
	152	+ struct virtio_net_ctrl_hdr hdr;
	153	+ virtio_net_ctrl_ack status;
	154	+ struct virtio_net_ctrl_mq mq;
	155	+ u8 promisc;
	156	+ u8 allmulti;
	157	+ __virtio16 vid;
	158	+ __virtio64 offloads;
	159	+};
	160	+
150	161	struct virtnet_info {
151	162	struct virtio_device *vdev;
152	163	struct virtqueue *cvq;
@@ -192,14 +203,7 @@ struct virtnet_info { · 39 unmodified lines
192	203	struct hlist_node node;
193	204	struct hlist_node node_dead;
194	205
195		− /* Control VQ buffers: protected by the rtnl lock */
196		− struct virtio_net_ctrl_hdr ctrl_hdr;
197		− virtio_net_ctrl_ack ctrl_status;
198		− struct virtio_net_ctrl_mq ctrl_mq;
199		− u8 ctrl_promisc;
200		− u8 ctrl_allmulti;
201		− u16 ctrl_vid;
202		− u64 ctrl_offloads;
	206	+ struct control_buf *ctrl;
203	207
204	208	/* Ethtool settings */
205	209	u8 duplex;
@@ -1269,7 +1273,9 @@ static int virtnet_poll(struct napi_struct *napi, int budget) · 1063 unmodified lines
1269	1273	{
1270	1274	struct receive_queue *rq =
1271	1275	container_of(napi, struct receive_queue, napi);
1272		− unsigned int received;
	1276	+ struct virtnet_info *vi = rq->vq->vdev->priv;
	1277	+ struct send_queue *sq;
	1278	+ unsigned int received, qp;
1273	1279	bool xdp_xmit = false;
1274	1280
1275	1281	virtnet_poll_cleantx(rq);
@@ -1280,8 +1286,13 @@ static int virtnet_poll(struct napi_struct *napi, int budget) · 4 unmodified lines
1280	1286	if (received < budget)
1281	1287	virtqueue_napi_complete(napi, rq->vq, received);
1282	1288
1283		− if (xdp_xmit)
	1289	+ if (xdp_xmit) {
	1290	+ qp = vi->curr_queue_pairs - vi->xdp_queue_pairs +
	1291	+ smp_processor_id();
	1292	+ sq = &vi->sq[qp];
	1293	+ virtqueue_kick(sq->vq);
1284	1294	xdp_do_flush_map();
	1295	+ }
1285	1296
1286	1297	return received;
1287	1298	}
@@ -1454,25 +1465,25 @@ static bool virtnet_send_command(struct virtnet_info *vi, u8 class, u8 cmd, · 166 unmodified lines
1454	1465	/* Caller should know better */
1455	1466	BUG_ON(!virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ));
1456	1467
1457		− vi->ctrl_status = ~0;
1458		− vi->ctrl_hdr.class = class;
1459		− vi->ctrl_hdr.cmd = cmd;
	1468	+ vi->ctrl->status = ~0;
	1469	+ vi->ctrl->hdr.class = class;
	1470	+ vi->ctrl->hdr.cmd = cmd;
1460	1471	/* Add header */
1461		− sg_init_one(&hdr, &vi->ctrl_hdr, sizeof(vi->ctrl_hdr));
	1472	+ sg_init_one(&hdr, &vi->ctrl->hdr, sizeof(vi->ctrl->hdr));
1462	1473	sgs[out_num++] = &hdr;
1463	1474
1464	1475	if (out)
1465	1476	sgs[out_num++] = out;
1466	1477
1467	1478	/* Add return status. */
1468		− sg_init_one(&stat, &vi->ctrl_status, sizeof(vi->ctrl_status));
	1479	+ sg_init_one(&stat, &vi->ctrl->status, sizeof(vi->ctrl->status));
1469	1480	sgs[out_num] = &stat;
1470	1481
1471	1482	BUG_ON(out_num + 1 > ARRAY_SIZE(sgs));
1472	1483	virtqueue_add_sgs(vi->cvq, sgs, out_num, 1, vi, GFP_ATOMIC);
1473	1484
1474	1485	if (unlikely(!virtqueue_kick(vi->cvq)))
1475		− return vi->ctrl_status == VIRTIO_NET_OK;
	1486	+ return vi->ctrl->status == VIRTIO_NET_OK;
1476	1487
1477	1488	/* Spin for a response, the kick causes an ioport write, trapping
1478	1489	* into the hypervisor, so the request should be handled immediately.
@@ -1481,7 +1492,7 @@ static bool virtnet_send_command(struct virtnet_info *vi, u8 class, u8 cmd, · 2 unmodified lines
1481	1492	!virtqueue_is_broken(vi->cvq))
1482	1493	cpu_relax();
1483	1494
1484		− return vi->ctrl_status == VIRTIO_NET_OK;
	1495	+ return vi->ctrl->status == VIRTIO_NET_OK;
1485	1496	}
1486	1497
1487	1498	static int virtnet_set_mac_address(struct net_device dev, void p)
@@ -1593,8 +1604,8 @@ static int _virtnet_set_queues(struct virtnet_info *vi, u16 queue_pairs) · 105 unmodified lines
1593	1604	if (!vi->has_cvq \|\| !virtio_has_feature(vi->vdev, VIRTIO_NET_F_MQ))
1594	1605	return 0;
1595	1606
1596		− vi->ctrl_mq.virtqueue_pairs = cpu_to_virtio16(vi->vdev, queue_pairs);
1597		− sg_init_one(&sg, &vi->ctrl_mq, sizeof(vi->ctrl_mq));
	1607	+ vi->ctrl->mq.virtqueue_pairs = cpu_to_virtio16(vi->vdev, queue_pairs);
	1608	+ sg_init_one(&sg, &vi->ctrl->mq, sizeof(vi->ctrl->mq));
1598	1609
1599	1610	if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_MQ,
1600	1611	VIRTIO_NET_CTRL_MQ_VQ_PAIRS_SET, &sg)) {
@@ -1653,22 +1664,22 @@ static void virtnet_set_rx_mode(struct net_device *dev) · 52 unmodified lines
1653	1664	if (!virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_RX))
1654	1665	return;
1655	1666
1656		− vi->ctrl_promisc = ((dev->flags & IFF_PROMISC) != 0);
1657		− vi->ctrl_allmulti = ((dev->flags & IFF_ALLMULTI) != 0);
	1667	+ vi->ctrl->promisc = ((dev->flags & IFF_PROMISC) != 0);
	1668	+ vi->ctrl->allmulti = ((dev->flags & IFF_ALLMULTI) != 0);
1658	1669
1659		− sg_init_one(sg, &vi->ctrl_promisc, sizeof(vi->ctrl_promisc));
	1670	+ sg_init_one(sg, &vi->ctrl->promisc, sizeof(vi->ctrl->promisc));
1660	1671
1661	1672	if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_RX,
1662	1673	VIRTIO_NET_CTRL_RX_PROMISC, sg))
1663	1674	dev_warn(&dev->dev, "Failed to %sable promisc mode.\n",
1664		− vi->ctrl_promisc ? "en" : "dis");
	1675	+ vi->ctrl->promisc ? "en" : "dis");
1665	1676
1666		− sg_init_one(sg, &vi->ctrl_allmulti, sizeof(vi->ctrl_allmulti));
	1677	+ sg_init_one(sg, &vi->ctrl->allmulti, sizeof(vi->ctrl->allmulti));
1667	1678
1668	1679	if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_RX,
1669	1680	VIRTIO_NET_CTRL_RX_ALLMULTI, sg))
1670	1681	dev_warn(&dev->dev, "Failed to %sable allmulti mode.\n",
1671		− vi->ctrl_allmulti ? "en" : "dis");
	1682	+ vi->ctrl->allmulti ? "en" : "dis");
1672	1683
1673	1684	uc_count = netdev_uc_count(dev);
1674	1685	mc_count = netdev_mc_count(dev);
@@ -1714,8 +1725,8 @@ static int virtnet_vlan_rx_add_vid(struct net_device *dev, · 39 unmodified lines
1714	1725	struct virtnet_info *vi = netdev_priv(dev);
1715	1726	struct scatterlist sg;
1716	1727
1717		− vi->ctrl_vid = vid;
1718		− sg_init_one(&sg, &vi->ctrl_vid, sizeof(vi->ctrl_vid));
	1728	+ vi->ctrl->vid = cpu_to_virtio16(vi->vdev, vid);
	1729	+ sg_init_one(&sg, &vi->ctrl->vid, sizeof(vi->ctrl->vid));
1719	1730
1720	1731	if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_VLAN,
1721	1732	VIRTIO_NET_CTRL_VLAN_ADD, &sg))
@@ -1729,8 +1740,8 @@ static int virtnet_vlan_rx_kill_vid(struct net_device *dev, · 7 unmodified lines
1729	1740	struct virtnet_info *vi = netdev_priv(dev);
1730	1741	struct scatterlist sg;
1731	1742
1732		− vi->ctrl_vid = vid;
1733		− sg_init_one(&sg, &vi->ctrl_vid, sizeof(vi->ctrl_vid));
	1743	+ vi->ctrl->vid = cpu_to_virtio16(vi->vdev, vid);
	1744	+ sg_init_one(&sg, &vi->ctrl->vid, sizeof(vi->ctrl->vid));
1734	1745
1735	1746	if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_VLAN,
1736	1747	VIRTIO_NET_CTRL_VLAN_DEL, &sg))
@@ -2126,9 +2137,9 @@ static int virtnet_restore_up(struct virtio_device *vdev) · 389 unmodified lines
2126	2137	static int virtnet_set_guest_offloads(struct virtnet_info *vi, u64 offloads)
2127	2138	{
2128	2139	struct scatterlist sg;
2129		− vi->ctrl_offloads = cpu_to_virtio64(vi->vdev, offloads);
	2140	+ vi->ctrl->offloads = cpu_to_virtio64(vi->vdev, offloads);
2130	2141
2131		− sg_init_one(&sg, &vi->ctrl_offloads, sizeof(vi->ctrl_offloads));
	2142	+ sg_init_one(&sg, &vi->ctrl->offloads, sizeof(vi->ctrl->offloads));
2132	2143
2133	2144	if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_GUEST_OFFLOADS,
2134	2145	VIRTIO_NET_CTRL_GUEST_OFFLOADS_SET, &sg)) {
@@ -2351,6 +2362,7 @@ static void virtnet_free_queues(struct virtnet_info *vi) · 216 unmodified lines
2351	2362
2352	2363	kfree(vi->rq);
2353	2364	kfree(vi->sq);
	2365	+ kfree(vi->ctrl);
2354	2366	}
2355	2367
2356	2368	static void _free_receive_bufs(struct virtnet_info *vi)
@@ -2543,6 +2555,9 @@ static int virtnet_alloc_queues(struct virtnet_info *vi) · 186 unmodified lines
2543	2555	{
2544	2556	int i;
2545	2557
	2558	+ vi->ctrl = kzalloc(sizeof(*vi->ctrl), GFP_KERNEL);
	2559	+ if (!vi->ctrl)
	2560	+ goto err_ctrl;
2546	2561	vi->sq = kzalloc(sizeof(vi->sq) vi->max_queue_pairs, GFP_KERNEL);
2547	2562	if (!vi->sq)
2548	2563	goto err_sq;
@@ -2571,6 +2586,8 @@ static int virtnet_alloc_queues(struct virtnet_info *vi) · 22 unmodified lines
2571	2586	err_rq:
2572	2587	kfree(vi->sq);
2573	2588	err_sq:
	2589	+ kfree(vi->ctrl);
	2590	+err_ctrl:
2574	2591	return -ENOMEM;
2575	2592	}
2576	2593

drivers/net/vmxnet3/vmxnet3_drv.c+13−4

@@ -1218,6 +1218,7 @@ vmxnet3_get_hdr_len(struct vmxnet3_adapter adapter, struct sk_buff skb, · 1217 unmodified lines
1218	1218	union {
1219	1219	void *ptr;
1220	1220	struct ethhdr *eth;
	1221	+ struct vlan_ethhdr *veth;
1221	1222	struct iphdr *ipv4;
1222	1223	struct ipv6hdr *ipv6;
1223	1224	struct tcphdr *tcp;
@@ -1228,16 +1229,24 @@ vmxnet3_get_hdr_len(struct vmxnet3_adapter adapter, struct sk_buff skb, · 4 unmodified lines
1228	1229	if (unlikely(sizeof(struct iphdr) + sizeof(struct tcphdr) > maplen))
1229	1230	return 0;
1230	1231
	1232	+ if (skb->protocol == cpu_to_be16(ETH_P_8021Q) \|\|
	1233	+ skb->protocol == cpu_to_be16(ETH_P_8021AD))
	1234	+ hlen = sizeof(struct vlan_ethhdr);
	1235	+ else
	1236	+ hlen = sizeof(struct ethhdr);
	1237	+
1231	1238	hdr.eth = eth_hdr(skb);
1232	1239	if (gdesc->rcd.v4) {
1233		− BUG_ON(hdr.eth->h_proto != htons(ETH_P_IP));
1234		− hdr.ptr += sizeof(struct ethhdr);
	1240	+ BUG_ON(hdr.eth->h_proto != htons(ETH_P_IP) &&
	1241	+ hdr.veth->h_vlan_encapsulated_proto != htons(ETH_P_IP));
	1242	+ hdr.ptr += hlen;
1235	1243	BUG_ON(hdr.ipv4->protocol != IPPROTO_TCP);
1236	1244	hlen = hdr.ipv4->ihl << 2;
1237	1245	hdr.ptr += hdr.ipv4->ihl << 2;
1238	1246	} else if (gdesc->rcd.v6) {
1239		− BUG_ON(hdr.eth->h_proto != htons(ETH_P_IPV6));
1240		− hdr.ptr += sizeof(struct ethhdr);
	1247	+ BUG_ON(hdr.eth->h_proto != htons(ETH_P_IPV6) &&
	1248	+ hdr.veth->h_vlan_encapsulated_proto != htons(ETH_P_IPV6));
	1249	+ hdr.ptr += hlen;
1241	1250	/* Use an estimated value, since we also need to handle
1242	1251	* TSO case.
1243	1252	*/

drivers/net/vmxnet3/vmxnet3_int.h+2−2

@@ -69,10 +69,10 @@ · 68 unmodified lines
69	69	/*
70	70	* Version numbers
71	71	*/
72		−#define VMXNET3_DRIVER_VERSION_STRING "1.4.13.0-k"
	72	+#define VMXNET3_DRIVER_VERSION_STRING "1.4.14.0-k"
73	73
74	74	/* a 32-bit int, each byte encode a verion number in VMXNET3_DRIVER_VERSION */
75		−#define VMXNET3_DRIVER_VERSION_NUM 0x01040d00
	75	+#define VMXNET3_DRIVER_VERSION_NUM 0x01040e00
76	76
77	77	#if defined(CONFIG_PCI_MSI)
78	78	/* RSS only makes sense if MSI-X is supported. */

drivers/net/wireless/broadcom/brcm80211/brcmfmac/firmware.c+20−16

@@ -459,7 +459,7 @@ static void brcmf_fw_free_request(struct brcmf_fw_request *req) · 458 unmodified lines
459	459	kfree(req);
460	460	}
461	461
462		−static void brcmf_fw_request_nvram_done(const struct firmware fw, void ctx)
	462	+static int brcmf_fw_request_nvram_done(const struct firmware fw, void ctx)
463	463	{
464	464	struct brcmf_fw *fwctx = ctx;
465	465	struct brcmf_fw_item *cur;
@@ -498,13 +498,10 @@ static void brcmf_fw_request_nvram_done(const struct firmware fw, void ctx) · 32 unmodified lines
498	498	brcmf_dbg(TRACE, "nvram %p len %d\n", nvram, nvram_length);
499	499	cur->nv_data.data = nvram;
500	500	cur->nv_data.len = nvram_length;
501		− return;
	501	+ return 0;
502	502
503	503	fail:
504		− brcmf_dbg(TRACE, "failed: dev=%s\n", dev_name(fwctx->dev));
505		− fwctx->done(fwctx->dev, -ENOENT, NULL);
506		− brcmf_fw_free_request(fwctx->req);
507		− kfree(fwctx);
	504	+ return -ENOENT;
508	505	}
509	506
510	507	static int brcmf_fw_request_next_item(struct brcmf_fw *fwctx, bool async)
@@ -553,20 +550,27 @@ static void brcmf_fw_request_done(const struct firmware fw, void ctx) · 42 unmodified lines
553	550	brcmf_dbg(TRACE, "enter: firmware %s %sfound\n", cur->path,
554	551	fw ? "" : "not ");
555	552
556		− if (fw) {
557		− if (cur->type == BRCMF_FW_TYPE_BINARY)
558		− cur->binary = fw;
559		− else if (cur->type == BRCMF_FW_TYPE_NVRAM)
560		− brcmf_fw_request_nvram_done(fw, fwctx);
561		− else
562		− release_firmware(fw);
563		− } else if (cur->type == BRCMF_FW_TYPE_NVRAM) {
564		− brcmf_fw_request_nvram_done(NULL, fwctx);
565		− } else if (!(cur->flags & BRCMF_FW_REQF_OPTIONAL)) {
	553	+ if (!fw)
566	554	ret = -ENOENT;
	555	+
	556	+ switch (cur->type) {
	557	+ case BRCMF_FW_TYPE_NVRAM:
	558	+ ret = brcmf_fw_request_nvram_done(fw, fwctx);
	559	+ break;
	560	+ case BRCMF_FW_TYPE_BINARY:
	561	+ cur->binary = fw;
	562	+ break;
	563	+ default:
	564	+ /* something fishy here so bail out early */
	565	+ brcmf_err("unknown fw type: %d\n", cur->type);
	566	+ release_firmware(fw);
	567	+ ret = -EINVAL;
567	568	goto fail;
568	569	}
569	570
	571	+ if (ret < 0 && !(cur->flags & BRCMF_FW_REQF_OPTIONAL))
	572	+ goto fail;
	573	+
570	574	do {
571	575	if (++fwctx->curpos == fwctx->req->n_items) {
572	576	ret = 0;

drivers/net/wireless/intel/iwlwifi/fw/api/scan.h+5−8

@@ -8,6 +8,7 @@ · 7 unmodified lines
8	8	* Copyright(c) 2012 - 2014 Intel Corporation. All rights reserved.
9	9	* Copyright(c) 2013 - 2015 Intel Mobile Communications GmbH
10	10	* Copyright(c) 2016 - 2017 Intel Deutschland GmbH
	11	+ * Copyright(c) 2018 Intel Corporation
11	12	*
12	13	* This program is free software; you can redistribute it and/or modify
13	14	* it under the terms of version 2 of the GNU General Public License as
@@ -30,7 +31,7 @@ · 16 unmodified lines
30	31	* Copyright(c) 2012 - 2014 Intel Corporation. All rights reserved.
31	32	* Copyright(c) 2013 - 2015 Intel Mobile Communications GmbH
32	33	* Copyright(c) 2016 - 2017 Intel Deutschland GmbH
33		− * Copyright(c) 2018 Intel Corporation
	34	+ * Copyright(c) 2018 Intel Corporation
34	35	* All rights reserved.
35	36	*
36	37	* Redistribution and use in source and binary forms, with or without
@@ -749,13 +750,9 @@ struct iwl_scan_req_umac { · 712 unmodified lines
749	750	} __packed;
750	751
751	752	#define IWL_SCAN_REQ_UMAC_SIZE_V8 sizeof(struct iwl_scan_req_umac)
752		−#define IWL_SCAN_REQ_UMAC_SIZE_V7 (sizeof(struct iwl_scan_req_umac) - \
753		− 4 * sizeof(u8))
754		−#define IWL_SCAN_REQ_UMAC_SIZE_V6 (sizeof(struct iwl_scan_req_umac) - \
755		− 2 * sizeof(u8) - sizeof(__le16))
756		−#define IWL_SCAN_REQ_UMAC_SIZE_V1 (sizeof(struct iwl_scan_req_umac) - \
757		− 2 * sizeof(__le32) - 2 * sizeof(u8) - \
758		− sizeof(__le16))
	753	+#define IWL_SCAN_REQ_UMAC_SIZE_V7 48
	754	+#define IWL_SCAN_REQ_UMAC_SIZE_V6 44
	755	+#define IWL_SCAN_REQ_UMAC_SIZE_V1 36
759	756
760	757	/**
761	758	* struct iwl_umac_scan_abort

drivers/net/wireless/intel/iwlwifi/iwl-nvm-parse.c+95−16

@@ -76,6 +76,7 @@ · 75 unmodified lines
76	76	#include "iwl-io.h"
77	77	#include "iwl-csr.h"
78	78	#include "fw/acpi.h"
	79	+#include "fw/api/nvm-reg.h"
79	80
80	81	/* NVM offsets (in words) definitions */
81	82	enum nvm_offsets {
@@ -146,8 +147,8 @@ static const u8 iwl_ext_nvm_channels[] = { · 64 unmodified lines
146	147	149, 153, 157, 161, 165, 169, 173, 177, 181
147	148	};
148	149
149		−#define IWL_NUM_CHANNELS ARRAY_SIZE(iwl_nvm_channels)
150		−#define IWL_NUM_CHANNELS_EXT ARRAY_SIZE(iwl_ext_nvm_channels)
	150	+#define IWL_NVM_NUM_CHANNELS ARRAY_SIZE(iwl_nvm_channels)
	151	+#define IWL_NVM_NUM_CHANNELS_EXT ARRAY_SIZE(iwl_ext_nvm_channels)
151	152	#define NUM_2GHZ_CHANNELS 14
152	153	#define NUM_2GHZ_CHANNELS_EXT 14
153	154	#define FIRST_2GHZ_HT_MINUS 5
@@ -301,11 +302,11 @@ static int iwl_init_channel_map(struct device dev, const struct iwl_cfg cfg, · 147 unmodified lines
301	302	const u8 *nvm_chan;
302	303
303	304	if (cfg->nvm_type != IWL_NVM_EXT) {
304		− num_of_ch = IWL_NUM_CHANNELS;
	305	+ num_of_ch = IWL_NVM_NUM_CHANNELS;
305	306	nvm_chan = &iwl_nvm_channels[0];
306	307	num_2ghz_channels = NUM_2GHZ_CHANNELS;
307	308	} else {
308		− num_of_ch = IWL_NUM_CHANNELS_EXT;
	309	+ num_of_ch = IWL_NVM_NUM_CHANNELS_EXT;
309	310	nvm_chan = &iwl_ext_nvm_channels[0];
310	311	num_2ghz_channels = NUM_2GHZ_CHANNELS_EXT;
311	312	}
@@ -720,12 +721,12 @@ iwl_parse_nvm_data(struct iwl_trans trans, const struct iwl_cfg cfg, · 408 unmodified lines
720	721	if (cfg->nvm_type != IWL_NVM_EXT)
721	722	data = kzalloc(sizeof(*data) +
722	723	sizeof(struct ieee80211_channel) *
723		− IWL_NUM_CHANNELS,
	724	+ IWL_NVM_NUM_CHANNELS,
724	725	GFP_KERNEL);
725	726	else
726	727	data = kzalloc(sizeof(*data) +
727	728	sizeof(struct ieee80211_channel) *
728		− IWL_NUM_CHANNELS_EXT,
	729	+ IWL_NVM_NUM_CHANNELS_EXT,
729	730	GFP_KERNEL);
730	731	if (!data)
731	732	return NULL;
@@ -842,24 +843,34 @@ static u32 iwl_nvm_get_regdom_bw_flags(const u8 *nvm_chan, · 110 unmodified lines
842	843	return flags;
843	844	}
844	845
	846	+struct regdb_ptrs {
	847	+ struct ieee80211_wmm_rule *rule;
	848	+ u32 token;
	849	+};
	850	+
845	851	struct ieee80211_regdomain *
846	852	iwl_parse_nvm_mcc_info(struct device dev, const struct iwl_cfg cfg,
847		− int num_of_ch, __le32 *channels, u16 fw_mcc)
	853	+ int num_of_ch, __le32 *channels, u16 fw_mcc,
	854	+ u16 geo_info)
848	855	{
849	856	int ch_idx;
850	857	u16 ch_flags;
851	858	u32 reg_rule_flags, prev_reg_rule_flags = 0;
852	859	const u8 *nvm_chan = cfg->nvm_type == IWL_NVM_EXT ?
853	860	iwl_ext_nvm_channels : iwl_nvm_channels;
854		− struct ieee80211_regdomain *regd;
855		− int size_of_regd;
	861	+ struct ieee80211_regdomain regd, copy_rd;
	862	+ int size_of_regd, regd_to_copy, wmms_to_copy;
	863	+ int size_of_wmms = 0;
856	864	struct ieee80211_reg_rule *rule;
	865	+ struct ieee80211_wmm_rule wmm_rule, d_wmm, *s_wmm;
	866	+ struct regdb_ptrs *regdb_ptrs;
857	867	enum nl80211_band band;
858	868	int center_freq, prev_center_freq = 0;
859		− int valid_rules = 0;
	869	+ int valid_rules = 0, n_wmms = 0;
	870	+ int i;
860	871	bool new_rule;
861	872	int max_num_ch = cfg->nvm_type == IWL_NVM_EXT ?
862		− IWL_NUM_CHANNELS_EXT : IWL_NUM_CHANNELS;
	873	+ IWL_NVM_NUM_CHANNELS_EXT : IWL_NVM_NUM_CHANNELS;
863	874
864	875	if (WARN_ON_ONCE(num_of_ch > NL80211_MAX_SUPP_REG_RULES))
865	876	return ERR_PTR(-EINVAL);
@@ -875,10 +886,26 @@ iwl_parse_nvm_mcc_info(struct device dev, const struct iwl_cfg cfg, · 9 unmodified lines
875	886	sizeof(struct ieee80211_regdomain) +
876	887	num_of_ch * sizeof(struct ieee80211_reg_rule);
877	888
878		− regd = kzalloc(size_of_regd, GFP_KERNEL);
	889	+ if (geo_info & GEO_WMM_ETSI_5GHZ_INFO)
	890	+ size_of_wmms =
	891	+ num_of_ch * sizeof(struct ieee80211_wmm_rule);
	892	+
	893	+ regd = kzalloc(size_of_regd + size_of_wmms, GFP_KERNEL);
879	894	if (!regd)
880	895	return ERR_PTR(-ENOMEM);
881	896
	897	+ regdb_ptrs = kcalloc(num_of_ch, sizeof(*regdb_ptrs), GFP_KERNEL);
	898	+ if (!regdb_ptrs) {
	899	+ copy_rd = ERR_PTR(-ENOMEM);
	900	+ goto out;
	901	+ }
	902	+
	903	+ /* set alpha2 from FW. */
	904	+ regd->alpha2[0] = fw_mcc >> 8;
	905	+ regd->alpha2[1] = fw_mcc & 0xff;
	906	+
	907	+ wmm_rule = (struct ieee80211_wmm_rule )((u8 )regd + size_of_regd);
	908	+
882	909	for (ch_idx = 0; ch_idx < num_of_ch; ch_idx++) {
883	910	ch_flags = (u16)__le32_to_cpup(channels + ch_idx);
884	911	band = (ch_idx < NUM_2GHZ_CHANNELS) ?
@@ -927,14 +954,66 @@ iwl_parse_nvm_mcc_info(struct device dev, const struct iwl_cfg cfg, · 42 unmodified lines
927	954
928	955	iwl_nvm_print_channel_flags(dev, IWL_DL_LAR,
929	956	nvm_chan[ch_idx], ch_flags);
	957	+
	958	+ if (!(geo_info & GEO_WMM_ETSI_5GHZ_INFO) \|\|
	959	+ band == NL80211_BAND_2GHZ)
	960	+ continue;
	961	+
	962	+ if (!reg_query_regdb_wmm(regd->alpha2, center_freq,
	963	+ &regdb_ptrs[n_wmms].token, wmm_rule)) {
	964	+ /* Add only new rules */
	965	+ for (i = 0; i < n_wmms; i++) {
	966	+ if (regdb_ptrs[i].token ==
	967	+ regdb_ptrs[n_wmms].token) {
	968	+ rule->wmm_rule = regdb_ptrs[i].rule;
	969	+ break;
	970	+ }
	971	+ }
	972	+ if (i == n_wmms) {
	973	+ rule->wmm_rule = wmm_rule;
	974	+ regdb_ptrs[n_wmms++].rule = wmm_rule;
	975	+ wmm_rule++;
	976	+ }
	977	+ }
930	978	}
931	979
932	980	regd->n_reg_rules = valid_rules;
	981	+ regd->n_wmm_rules = n_wmms;
933	982
934		− /* set alpha2 from FW. */
935		− regd->alpha2[0] = fw_mcc >> 8;
936		− regd->alpha2[1] = fw_mcc & 0xff;
	983	+ /*
	984	+ * Narrow down regdom for unused regulatory rules to prevent hole
	985	+ * between reg rules to wmm rules.
	986	+ */
	987	+ regd_to_copy = sizeof(struct ieee80211_regdomain) +
	988	+ valid_rules * sizeof(struct ieee80211_reg_rule);
	989	+
	990	+ wmms_to_copy = sizeof(struct ieee80211_wmm_rule) * n_wmms;
	991	+
	992	+ copy_rd = kzalloc(regd_to_copy + wmms_to_copy, GFP_KERNEL);
	993	+ if (!copy_rd) {
	994	+ copy_rd = ERR_PTR(-ENOMEM);
	995	+ goto out;
	996	+ }
	997	+
	998	+ memcpy(copy_rd, regd, regd_to_copy);
	999	+ memcpy((u8 )copy_rd + regd_to_copy, (u8 )regd + size_of_regd,
	1000	+ wmms_to_copy);
	1001	+
	1002	+ d_wmm = (struct ieee80211_wmm_rule )((u8 )copy_rd + regd_to_copy);
	1003	+ s_wmm = (struct ieee80211_wmm_rule )((u8 )regd + size_of_regd);
	1004	+
	1005	+ for (i = 0; i < regd->n_reg_rules; i++) {
	1006	+ if (!regd->reg_rules[i].wmm_rule)
	1007	+ continue;
	1008	+
	1009	+ copy_rd->reg_rules[i].wmm_rule = d_wmm +
	1010	+ (regd->reg_rules[i].wmm_rule - s_wmm) /
	1011	+ sizeof(struct ieee80211_wmm_rule);
	1012	+ }
937	1013
938		− return regd;
	1014	+out:
	1015	+ kfree(regdb_ptrs);
	1016	+ kfree(regd);
	1017	+ return copy_rd;
939	1018	}
940	1019	IWL_EXPORT_SYMBOL(iwl_parse_nvm_mcc_info);

drivers/net/wireless/intel/iwlwifi/iwl-nvm-parse.h+4−2

@@ -101,12 +101,14 @@ void iwl_init_sbands(struct device dev, const struct iwl_cfg cfg, · 100 unmodified lines
101	101	*
102	102	* This function parses the regulatory channel data received as a
103	103	* MCC_UPDATE_CMD command. It returns a newly allocation regulatory domain,
104		− * to be fed into the regulatory core. An ERR_PTR is returned on error.
	104	+ * to be fed into the regulatory core. In case the geo_info is set handle
	105	+ * accordingly. An ERR_PTR is returned on error.
105	106	* If not given to the regulatory core, the user is responsible for freeing
106	107	* the regdomain returned here with kfree.
107	108	*/
108	109	struct ieee80211_regdomain *
109	110	iwl_parse_nvm_mcc_info(struct device dev, const struct iwl_cfg cfg,
110		− int num_of_ch, __le32 *channels, u16 fw_mcc);
	111	+ int num_of_ch, __le32 *channels, u16 fw_mcc,
	112	+ u16 geo_info);
111	113
112	114	#endif /* __iwl_nvm_parse_h__ */

drivers/net/wireless/intel/iwlwifi/mvm/mac80211.c+2−1

@@ -311,7 +311,8 @@ struct ieee80211_regdomain iwl_mvm_get_regdomain(struct wiphy wiphy, · 310 unmodified lines
311	311	regd = iwl_parse_nvm_mcc_info(mvm->trans->dev, mvm->cfg,
312	312	__le32_to_cpu(resp->n_channels),
313	313	resp->channels,
314		− __le16_to_cpu(resp->mcc));
	314	+ __le16_to_cpu(resp->mcc),
	315	+ __le16_to_cpu(resp->geo_info));
315	316	/* Store the return source id */
316	317	src_id = resp->source_id;
317	318	kfree(resp);

drivers/net/wireless/realtek/rtlwifi/btcoexist/halbtcoutsrc.c+0−15

@@ -158,16 +158,6 @@ static u8 halbtc_get_wifi_central_chnl(struct btc_coexist *btcoexist) · 157 unmodified lines
158	158
159	159	static u8 rtl_get_hwpg_single_ant_path(struct rtl_priv *rtlpriv)
160	160	{
161		− struct rtl_mod_params *mod_params = rtlpriv->cfg->mod_params;
162		−
163		− /* override ant_num / ant_path */
164		− if (mod_params->ant_sel) {
165		− rtlpriv->btcoexist.btc_info.ant_num =
166		− (mod_params->ant_sel == 1 ? ANT_X2 : ANT_X1);
167		−
168		− rtlpriv->btcoexist.btc_info.single_ant_path =
169		− (mod_params->ant_sel == 1 ? 0 : 1);
170		− }
171	161	return rtlpriv->btcoexist.btc_info.single_ant_path;
172	162	}
173	163
@@ -178,7 +168,6 @@ static u8 rtl_get_hwpg_bt_type(struct rtl_priv *rtlpriv) · 4 unmodified lines
178	168
179	169	static u8 rtl_get_hwpg_ant_num(struct rtl_priv *rtlpriv)
180	170	{
181		− struct rtl_mod_params *mod_params = rtlpriv->cfg->mod_params;
182	171	u8 num;
183	172
184	173	if (rtlpriv->btcoexist.btc_info.ant_num == ANT_X2)
@@ -186,10 +175,6 @@ static u8 rtl_get_hwpg_ant_num(struct rtl_priv *rtlpriv) · 1 unmodified line
186	175	else
187	176	num = 1;
188	177
189		− /* override ant_num / ant_path */
190		− if (mod_params->ant_sel)
191		− num = (mod_params->ant_sel == 1 ? ANT_X2 : ANT_X1) + 1;
192		−
193	178	return num;
194	179	}
195	180

drivers/net/wireless/realtek/rtlwifi/rtl8723be/hw.c+7−4

@@ -848,6 +848,9 @@ static bool _rtl8723be_init_mac(struct ieee80211_hw *hw) · 847 unmodified lines
848	848	return false;
849	849	}
850	850
	851	+ if (rtlpriv->cfg->ops->get_btc_status())
	852	+ rtlpriv->btcoexist.btc_ops->btc_power_on_setting(rtlpriv);
	853	+
851	854	bytetmp = rtl_read_byte(rtlpriv, REG_MULTI_FUNC_CTRL);
852	855	rtl_write_byte(rtlpriv, REG_MULTI_FUNC_CTRL, bytetmp \| BIT(3));
853	856
@@ -2696,21 +2699,21 @@ void rtl8723be_read_bt_coexist_info_from_hwpg(struct ieee80211_hw *hw, · 1842 unmodified lines
2696	2699	rtlpriv->btcoexist.btc_info.bt_type = BT_RTL8723B;
2697	2700	rtlpriv->btcoexist.btc_info.ant_num = (value & 0x1);
2698	2701	rtlpriv->btcoexist.btc_info.single_ant_path =
2699		− (value & 0x40); /0xc3[6]/
	2702	+ (value & 0x40 ? ANT_AUX : ANT_MAIN); /0xc3[6]/
2700	2703	} else {
2701	2704	rtlpriv->btcoexist.btc_info.btcoexist = 0;
2702	2705	rtlpriv->btcoexist.btc_info.bt_type = BT_RTL8723B;
2703	2706	rtlpriv->btcoexist.btc_info.ant_num = ANT_X2;
2704		− rtlpriv->btcoexist.btc_info.single_ant_path = 0;
	2707	+ rtlpriv->btcoexist.btc_info.single_ant_path = ANT_MAIN;
2705	2708	}
2706	2709
2707	2710	/* override ant_num / ant_path */
2708	2711	if (mod_params->ant_sel) {
2709	2712	rtlpriv->btcoexist.btc_info.ant_num =
2710		− (mod_params->ant_sel == 1 ? ANT_X2 : ANT_X1);
	2713	+ (mod_params->ant_sel == 1 ? ANT_X1 : ANT_X2);
2711	2714
2712	2715	rtlpriv->btcoexist.btc_info.single_ant_path =
2713		− (mod_params->ant_sel == 1 ? 0 : 1);
	2716	+ (mod_params->ant_sel == 1 ? ANT_AUX : ANT_MAIN);
2714	2717	}
2715	2718	}
2716	2719

drivers/net/wireless/realtek/rtlwifi/wifi.h+5−0

@@ -2823,6 +2823,11 @@ enum bt_ant_num { · 2822 unmodified lines
2823	2823	ANT_X1 = 1,
2824	2824	};
2825	2825
	2826	+enum bt_ant_path {
	2827	+ ANT_MAIN = 0,
	2828	+ ANT_AUX = 1,
	2829	+};
	2830	+
2826	2831	enum bt_co_type {
2827	2832	BT_2WIRE = 0,
2828	2833	BT_ISSC_3WIRE = 1,

drivers/nvdimm/Kconfig+1−2

@@ -103,8 +103,7 @@ config NVDIMM_DAX · 102 unmodified lines
103	103	Select Y if unsure
104	104
105	105	config OF_PMEM
106		− # FIXME: make tristate once OF_NUMA dependency removed
107		− bool "Device-tree support for persistent memory regions"
	106	+ tristate "Device-tree support for persistent memory regions"
108	107	depends on OF
109	108	default LIBNVDIMM
110	109	help

drivers/nvdimm/dimm_devs.c+12−10

@@ -88,9 +88,9 @@ int nvdimm_init_nsarea(struct nvdimm_drvdata *ndd) · 87 unmodified lines
88	88	int nvdimm_init_config_data(struct nvdimm_drvdata *ndd)
89	89	{
90	90	struct nvdimm_bus *nvdimm_bus = walk_to_nvdimm_bus(ndd->dev);
	91	+ int rc = validate_dimm(ndd), cmd_rc = 0;
91	92	struct nd_cmd_get_config_data_hdr *cmd;
92	93	struct nvdimm_bus_descriptor *nd_desc;
93		− int rc = validate_dimm(ndd);
94	94	u32 max_cmd_size, config_size;
95	95	size_t offset;
96	96
@@ -124,9 +124,11 @@ int nvdimm_init_config_data(struct nvdimm_drvdata *ndd) · 27 unmodified lines
124	124	cmd->in_offset = offset;
125	125	rc = nd_desc->ndctl(nd_desc, to_nvdimm(ndd->dev),
126	126	ND_CMD_GET_CONFIG_DATA, cmd,
127		− cmd->in_length + sizeof(*cmd), NULL);
128		− if (rc \|\| cmd->status) {
129		− rc = -ENXIO;
	127	+ cmd->in_length + sizeof(*cmd), &cmd_rc);
	128	+ if (rc < 0)
	129	+ break;
	130	+ if (cmd_rc < 0) {
	131	+ rc = cmd_rc;
130	132	break;
131	133	}
132	134	memcpy(ndd->data + offset, cmd->out_buf, cmd->in_length);
@@ -140,9 +142,9 @@ int nvdimm_init_config_data(struct nvdimm_drvdata *ndd) · 7 unmodified lines
140	142	int nvdimm_set_config_data(struct nvdimm_drvdata *ndd, size_t offset,
141	143	void *buf, size_t len)
142	144	{
143		− int rc = validate_dimm(ndd);
144	145	size_t max_cmd_size, buf_offset;
145	146	struct nd_cmd_set_config_hdr *cmd;
	147	+ int rc = validate_dimm(ndd), cmd_rc = 0;
146	148	struct nvdimm_bus *nvdimm_bus = walk_to_nvdimm_bus(ndd->dev);
147	149	struct nvdimm_bus_descriptor *nd_desc = nvdimm_bus->nd_desc;
148	150
@@ -164,7 +166,6 @@ int nvdimm_set_config_data(struct nvdimm_drvdata *ndd, size_t offset, · 15 unmodified lines
164	166	for (buf_offset = 0; len; len -= cmd->in_length,
165	167	buf_offset += cmd->in_length) {
166	168	size_t cmd_size;
167		− u32 *status;
168	169
169	170	cmd->in_offset = offset + buf_offset;
170	171	cmd->in_length = min(max_cmd_size, len);
@@ -172,12 +173,13 @@ int nvdimm_set_config_data(struct nvdimm_drvdata *ndd, size_t offset, · 1 unmodified line
172	173
173	174	/* status is output in the last 4-bytes of the command buffer */
174	175	cmd_size = sizeof(*cmd) + cmd->in_length + sizeof(u32);
175		− status = ((void *) cmd) + cmd_size - sizeof(u32);
176	176
177	177	rc = nd_desc->ndctl(nd_desc, to_nvdimm(ndd->dev),
178		− ND_CMD_SET_CONFIG_DATA, cmd, cmd_size, NULL);
179		− if (rc \|\| *status) {
180		− rc = rc ? rc : -ENXIO;
	178	+ ND_CMD_SET_CONFIG_DATA, cmd, cmd_size, &cmd_rc);
	179	+ if (rc < 0)
	180	+ break;
	181	+ if (cmd_rc < 0) {
	182	+ rc = cmd_rc;
181	183	break;
182	184	}
183	185	}

drivers/nvdimm/of_pmem.c+1−1

@@ -67,7 +67,7 @@ static int of_pmem_region_probe(struct platform_device *pdev) · 66 unmodified lines
67	67	*/
68	68	memset(&ndr_desc, 0, sizeof(ndr_desc));
69	69	ndr_desc.attr_groups = region_attr_groups;
70		− ndr_desc.numa_node = of_node_to_nid(np);
	70	+ ndr_desc.numa_node = dev_to_node(&pdev->dev);
71	71	ndr_desc.res = &pdev->resource[i];
72	72	ndr_desc.of_node = np;
73	73	set_bit(ND_REGION_PAGEMAP, &ndr_desc.flags);

drivers/nvme/host/Kconfig+1−1

@@ -27,7 +27,7 @@ config NVME_FABRICS · 26 unmodified lines
27	27
28	28	config NVME_RDMA
29	29	tristate "NVM Express over Fabrics RDMA host driver"
30		− depends on INFINIBAND && BLOCK
	30	+ depends on INFINIBAND && INFINIBAND_ADDR_TRANS && BLOCK
31	31	select NVME_CORE
32	32	select NVME_FABRICS
33	33	select SG_POOL

drivers/nvme/host/core.c+2−25

@@ -764,6 +764,7 @@ static int nvme_submit_user_cmd(struct request_queue *q, · 763 unmodified lines
764	764	ret = PTR_ERR(meta);
765	765	goto out_unmap;
766	766	}
	767	+ req->cmd_flags \|= REQ_INTEGRITY;
767	768	}
768	769	}
769	770
@@ -2997,31 +2998,7 @@ static void nvme_alloc_ns(struct nvme_ctrl *ctrl, unsigned nsid) · 2227 unmodified lines
2997	2998	if (nvme_init_ns_head(ns, nsid, id))
2998	2999	goto out_free_id;
2999	3000	nvme_setup_streams_ns(ctrl, ns);
3000		−
3001		−#ifdef CONFIG_NVME_MULTIPATH
3002		− /*
3003		− * If multipathing is enabled we need to always use the subsystem
3004		− * instance number for numbering our devices to avoid conflicts
3005		− * between subsystems that have multiple controllers and thus use
3006		− * the multipath-aware subsystem node and those that have a single
3007		− * controller and use the controller node directly.
3008		− */
3009		− if (ns->head->disk) {
3010		− sprintf(disk_name, "nvme%dc%dn%d", ctrl->subsys->instance,
3011		− ctrl->cntlid, ns->head->instance);
3012		− flags = GENHD_FL_HIDDEN;
3013		− } else {
3014		− sprintf(disk_name, "nvme%dn%d", ctrl->subsys->instance,
3015		− ns->head->instance);
3016		− }
3017		−#else
3018		− /*
3019		− * But without the multipath code enabled, multiple controller per
3020		− * subsystems are visible as devices and thus we cannot use the
3021		− * subsystem instance.
3022		− */
3023		− sprintf(disk_name, "nvme%dn%d", ctrl->instance, ns->head->instance);
3024		−#endif
	3001	+ nvme_set_disk_name(disk_name, ns, ctrl, &flags);
3025	3002
3026	3003	if ((ctrl->quirks & NVME_QUIRK_LIGHTNVM) && id->vs[0] == 0x1) {
3027	3004	if (nvme_nvm_register(ns, disk_name, node)) {

drivers/nvme/host/fabrics.c+6−0

@@ -668,6 +668,7 @@ static int nvmf_parse_options(struct nvmf_ctrl_options *opts, · 667 unmodified lines
668	668	ret = -ENOMEM;
669	669	goto out;
670	670	}
	671	+ kfree(opts->transport);
671	672	opts->transport = p;
672	673	break;
673	674	case NVMF_OPT_NQN:
@@ -676,6 +677,7 @@ static int nvmf_parse_options(struct nvmf_ctrl_options *opts, · 2 unmodified lines
676	677	ret = -ENOMEM;
677	678	goto out;
678	679	}
	680	+ kfree(opts->subsysnqn);
679	681	opts->subsysnqn = p;
680	682	nqnlen = strlen(opts->subsysnqn);
681	683	if (nqnlen >= NVMF_NQN_SIZE) {
@@ -698,6 +700,7 @@ static int nvmf_parse_options(struct nvmf_ctrl_options *opts, · 16 unmodified lines
698	700	ret = -ENOMEM;
699	701	goto out;
700	702	}
	703	+ kfree(opts->traddr);
701	704	opts->traddr = p;
702	705	break;
703	706	case NVMF_OPT_TRSVCID:
@@ -706,6 +709,7 @@ static int nvmf_parse_options(struct nvmf_ctrl_options *opts, · 2 unmodified lines
706	709	ret = -ENOMEM;
707	710	goto out;
708	711	}
	712	+ kfree(opts->trsvcid);
709	713	opts->trsvcid = p;
710	714	break;
711	715	case NVMF_OPT_QUEUE_SIZE:
@@ -792,6 +796,7 @@ static int nvmf_parse_options(struct nvmf_ctrl_options *opts, · 80 unmodified lines
792	796	ret = -EINVAL;
793	797	goto out;
794	798	}
	799	+ nvmf_host_put(opts->host);
795	800	opts->host = nvmf_host_add(p);
796	801	kfree(p);
797	802	if (!opts->host) {
@@ -817,6 +822,7 @@ static int nvmf_parse_options(struct nvmf_ctrl_options *opts, · 19 unmodified lines
817	822	ret = -ENOMEM;
818	823	goto out;
819	824	}
	825	+ kfree(opts->host_traddr);
820	826	opts->host_traddr = p;
821	827	break;
822	828	case NVMF_OPT_HOST_ID:

drivers/nvme/host/multipath.c+23−1

@@ -15,10 +15,32 @@ · 14 unmodified lines
15	15	#include "nvme.h"
16	16
17	17	static bool multipath = true;
18		−module_param(multipath, bool, 0644);
	18	+module_param(multipath, bool, 0444);
19	19	MODULE_PARM_DESC(multipath,
20	20	"turn on native support for multiple controllers per subsystem");
21	21
	22	+/*
	23	+ * If multipathing is enabled we need to always use the subsystem instance
	24	+ * number for numbering our devices to avoid conflicts between subsystems that
	25	+ * have multiple controllers and thus use the multipath-aware subsystem node
	26	+ * and those that have a single controller and use the controller node
	27	+ * directly.
	28	+ */
	29	+void nvme_set_disk_name(char disk_name, struct nvme_ns ns,
	30	+ struct nvme_ctrl ctrl, int flags)
	31	+{
	32	+ if (!multipath) {
	33	+ sprintf(disk_name, "nvme%dn%d", ctrl->instance, ns->head->instance);
	34	+ } else if (ns->head->disk) {
	35	+ sprintf(disk_name, "nvme%dc%dn%d", ctrl->subsys->instance,
	36	+ ctrl->cntlid, ns->head->instance);
	37	+ *flags = GENHD_FL_HIDDEN;
	38	+ } else {
	39	+ sprintf(disk_name, "nvme%dn%d", ctrl->subsys->instance,
	40	+ ns->head->instance);
	41	+ }
	42	+}
	43	+
22	44	void nvme_failover_req(struct request *req)
23	45	{
24	46	struct nvme_ns *ns = req->q->queuedata;

drivers/nvme/host/nvme.h+12−0

@@ -436,6 +436,8 @@ extern const struct attribute_group nvme_ns_id_attr_group; · 435 unmodified lines
436	436	extern const struct block_device_operations nvme_ns_head_ops;
437	437
438	438	#ifdef CONFIG_NVME_MULTIPATH
	439	+void nvme_set_disk_name(char disk_name, struct nvme_ns ns,
	440	+ struct nvme_ctrl ctrl, int flags);
439	441	void nvme_failover_req(struct request *req);
440	442	bool nvme_req_needs_failover(struct request *req, blk_status_t error);
441	443	void nvme_kick_requeue_lists(struct nvme_ctrl *ctrl);
@@ -461,6 +463,16 @@ static inline void nvme_mpath_check_last_path(struct nvme_ns *ns) · 19 unmodified lines
461	463	}
462	464
463	465	#else
	466	+/*
	467	+ * Without the multipath code enabled, multiple controller per subsystems are
	468	+ * visible as devices and thus we cannot use the subsystem instance.
	469	+ */
	470	+static inline void nvme_set_disk_name(char disk_name, struct nvme_ns ns,
	471	+ struct nvme_ctrl ctrl, int flags)
	472	+{
	473	+ sprintf(disk_name, "nvme%dn%d", ctrl->instance, ns->head->instance);
	474	+}
	475	+
464	476	static inline void nvme_failover_req(struct request *req)
465	477	{
466	478	}

drivers/nvme/target/Kconfig+1−1

@@ -27,7 +27,7 @@ config NVME_TARGET_LOOP · 26 unmodified lines
27	27
28	28	config NVME_TARGET_RDMA
29	29	tristate "NVMe over Fabrics RDMA target support"
30		− depends on INFINIBAND
	30	+ depends on INFINIBAND && INFINIBAND_ADDR_TRANS
31	31	depends on NVME_TARGET
32	32	select SGL_ALLOC
33	33	help

drivers/nvme/target/loop.c+6−0

@@ -469,6 +469,12 @@ static void nvme_loop_reset_ctrl_work(struct work_struct *work) · 468 unmodified lines
469	469	nvme_stop_ctrl(&ctrl->ctrl);
470	470	nvme_loop_shutdown_ctrl(ctrl);
471	471
	472	+ if (!nvme_change_ctrl_state(&ctrl->ctrl, NVME_CTRL_CONNECTING)) {
	473	+ /* state change failure should never happen */
	474	+ WARN_ON_ONCE(1);
	475	+ return;
	476	+ }
	477	+
472	478	ret = nvme_loop_configure_admin_queue(ctrl);
473	479	if (ret)
474	480	goto out_disable;

drivers/of/fdt.c+5−2

@@ -942,7 +942,7 @@ int __init early_init_dt_scan_chosen_stdout(void) · 941 unmodified lines
942	942	int offset;
943	943	const char p, q, *options = NULL;
944	944	int l;
945		− const struct earlycon_id *match;
	945	+ const struct earlycon_id **p_match;
946	946	const void *fdt = initial_boot_params;
947	947
948	948	offset = fdt_path_offset(fdt, "/chosen");
@@ -969,7 +969,10 @@ int __init early_init_dt_scan_chosen_stdout(void) · 20 unmodified lines
969	969	return 0;
970	970	}
971	971
972		− for (match = __earlycon_table; match < __earlycon_table_end; match++) {
	972	+ for (p_match = __earlycon_table; p_match < __earlycon_table_end;
	973	+ p_match++) {
	974	+ const struct earlycon_id match = p_match;
	975	+
973	976	if (!match->compatible[0])
974	977	continue;
975	978

drivers/parisc/ccio-dma.c+1−1

@@ -1263,7 +1263,7 @@ static struct parisc_driver ccio_driver __refdata = { · 1262 unmodified lines
1263	1263	* I/O Page Directory, the resource map, and initalizing the
1264	1264	* U2/Uturn chip into virtual mode.
1265	1265	*/
1266		−static void
	1266	+static void __init
1267	1267	ccio_ioc_init(struct ioc *ioc)
1268	1268	{
1269	1269	int i;

drivers/pci/dwc/pcie-kirin.c+1−1

@@ -486,7 +486,7 @@ static int kirin_pcie_probe(struct platform_device *pdev) · 485 unmodified lines
486	486	return ret;
487	487
488	488	kirin_pcie->gpio_id_reset = of_get_named_gpio(dev->of_node,
489		− "reset-gpio", 0);
	489	+ "reset-gpios", 0);
490	490	if (kirin_pcie->gpio_id_reset < 0)
491	491	return -ENODEV;
492	492

drivers/pci/host/pci-aardvark.c+30−23

@@ -29,6 +29,7 @@ · 28 unmodified lines
29	29	#define PCIE_CORE_DEV_CTRL_STATS_MAX_PAYLOAD_SZ_SHIFT 5
30	30	#define PCIE_CORE_DEV_CTRL_STATS_SNOOP_DISABLE (0 << 11)
31	31	#define PCIE_CORE_DEV_CTRL_STATS_MAX_RD_REQ_SIZE_SHIFT 12
	32	+#define PCIE_CORE_DEV_CTRL_STATS_MAX_RD_REQ_SZ 0x2
32	33	#define PCIE_CORE_LINK_CTRL_STAT_REG 0xd0
33	34	#define PCIE_CORE_LINK_L0S_ENTRY BIT(0)
34	35	#define PCIE_CORE_LINK_TRAINING BIT(5)
@@ -100,7 +101,8 @@ · 65 unmodified lines
100	101	#define PCIE_ISR1_MASK_REG (CONTROL_BASE_ADDR + 0x4C)
101	102	#define PCIE_ISR1_POWER_STATE_CHANGE BIT(4)
102	103	#define PCIE_ISR1_FLUSH BIT(5)
103		−#define PCIE_ISR1_ALL_MASK GENMASK(5, 4)
	104	+#define PCIE_ISR1_INTX_ASSERT(val) BIT(8 + (val))
	105	+#define PCIE_ISR1_ALL_MASK GENMASK(11, 4)
104	106	#define PCIE_MSI_ADDR_LOW_REG (CONTROL_BASE_ADDR + 0x50)
105	107	#define PCIE_MSI_ADDR_HIGH_REG (CONTROL_BASE_ADDR + 0x54)
106	108	#define PCIE_MSI_STATUS_REG (CONTROL_BASE_ADDR + 0x58)
@@ -172,8 +174,6 @@ · 65 unmodified lines
172	174	#define PCIE_CONFIG_WR_TYPE0 0xa
173	175	#define PCIE_CONFIG_WR_TYPE1 0xb
174	176
175		−/* PCI_BDF shifts 8bit, so we need extra 4bit shift */
176		−#define PCIE_BDF(dev) (dev << 4)
177	177	#define PCIE_CONF_BUS(bus) (((bus) & 0xff) << 20)
178	178	#define PCIE_CONF_DEV(dev) (((dev) & 0x1f) << 15)
179	179	#define PCIE_CONF_FUNC(fun) (((fun) & 0x7) << 12)
@@ -296,7 +296,8 @@ static void advk_pcie_setup_hw(struct advk_pcie *pcie) · 116 unmodified lines
296	296	reg = PCIE_CORE_DEV_CTRL_STATS_RELAX_ORDER_DISABLE \|
297	297	(7 << PCIE_CORE_DEV_CTRL_STATS_MAX_PAYLOAD_SZ_SHIFT) \|
298	298	PCIE_CORE_DEV_CTRL_STATS_SNOOP_DISABLE \|
299		− PCIE_CORE_DEV_CTRL_STATS_MAX_RD_REQ_SIZE_SHIFT;
	299	+ (PCIE_CORE_DEV_CTRL_STATS_MAX_RD_REQ_SZ <<
	300	+ PCIE_CORE_DEV_CTRL_STATS_MAX_RD_REQ_SIZE_SHIFT);
300	301	advk_writel(pcie, reg, PCIE_CORE_DEV_CTRL_STATS_REG);
301	302
302	303	/* Program PCIe Control 2 to disable strict ordering */
@@ -437,7 +438,7 @@ static int advk_pcie_rd_conf(struct pci_bus *bus, u32 devfn, · 134 unmodified lines
437	438	u32 reg;
438	439	int ret;
439	440
440		− if (PCI_SLOT(devfn) != 0) {
	441	+ if ((bus->number == pcie->root_bus_nr) && PCI_SLOT(devfn) != 0) {
441	442	*val = 0xffffffff;
442	443	return PCIBIOS_DEVICE_NOT_FOUND;
443	444	}
@@ -456,7 +457,7 @@ static int advk_pcie_rd_conf(struct pci_bus *bus, u32 devfn, · 12 unmodified lines
456	457	advk_writel(pcie, reg, PIO_CTRL);
457	458
458	459	/* Program the address registers */
459		− reg = PCIE_BDF(devfn) \| PCIE_CONF_REG(where);
	460	+ reg = PCIE_CONF_ADDR(bus->number, devfn, where);
460	461	advk_writel(pcie, reg, PIO_ADDR_LS);
461	462	advk_writel(pcie, 0, PIO_ADDR_MS);
462	463
@@ -491,7 +492,7 @@ static int advk_pcie_wr_conf(struct pci_bus *bus, u32 devfn, · 28 unmodified lines
491	492	int offset;
492	493	int ret;
493	494
494		− if (PCI_SLOT(devfn) != 0)
	495	+ if ((bus->number == pcie->root_bus_nr) && PCI_SLOT(devfn) != 0)
495	496	return PCIBIOS_DEVICE_NOT_FOUND;
496	497
497	498	if (where % size)
@@ -609,9 +610,9 @@ static void advk_pcie_irq_mask(struct irq_data *d) · 111 unmodified lines
609	610	irq_hw_number_t hwirq = irqd_to_hwirq(d);
610	611	u32 mask;
611	612
612		− mask = advk_readl(pcie, PCIE_ISR0_MASK_REG);
613		− mask \|= PCIE_ISR0_INTX_ASSERT(hwirq);
614		− advk_writel(pcie, mask, PCIE_ISR0_MASK_REG);
	613	+ mask = advk_readl(pcie, PCIE_ISR1_MASK_REG);
	614	+ mask \|= PCIE_ISR1_INTX_ASSERT(hwirq);
	615	+ advk_writel(pcie, mask, PCIE_ISR1_MASK_REG);
615	616	}
616	617
617	618	static void advk_pcie_irq_unmask(struct irq_data *d)
@@ -620,9 +621,9 @@ static void advk_pcie_irq_unmask(struct irq_data *d) · 2 unmodified lines
620	621	irq_hw_number_t hwirq = irqd_to_hwirq(d);
621	622	u32 mask;
622	623
623		− mask = advk_readl(pcie, PCIE_ISR0_MASK_REG);
624		− mask &= ~PCIE_ISR0_INTX_ASSERT(hwirq);
625		− advk_writel(pcie, mask, PCIE_ISR0_MASK_REG);
	624	+ mask = advk_readl(pcie, PCIE_ISR1_MASK_REG);
	625	+ mask &= ~PCIE_ISR1_INTX_ASSERT(hwirq);
	626	+ advk_writel(pcie, mask, PCIE_ISR1_MASK_REG);
626	627	}
627	628
628	629	static int advk_pcie_irq_map(struct irq_domain *h,
@@ -765,29 +766,35 @@ static void advk_pcie_handle_msi(struct advk_pcie *pcie) · 136 unmodified lines
765	766
766	767	static void advk_pcie_handle_int(struct advk_pcie *pcie)
767	768	{
768		− u32 val, mask, status;
	769	+ u32 isr0_val, isr0_mask, isr0_status;
	770	+ u32 isr1_val, isr1_mask, isr1_status;
769	771	int i, virq;
770	772
771		− val = advk_readl(pcie, PCIE_ISR0_REG);
772		− mask = advk_readl(pcie, PCIE_ISR0_MASK_REG);
773		− status = val & ((~mask) & PCIE_ISR0_ALL_MASK);
	773	+ isr0_val = advk_readl(pcie, PCIE_ISR0_REG);
	774	+ isr0_mask = advk_readl(pcie, PCIE_ISR0_MASK_REG);
	775	+ isr0_status = isr0_val & ((~isr0_mask) & PCIE_ISR0_ALL_MASK);
	776	+
	777	+ isr1_val = advk_readl(pcie, PCIE_ISR1_REG);
	778	+ isr1_mask = advk_readl(pcie, PCIE_ISR1_MASK_REG);
	779	+ isr1_status = isr1_val & ((~isr1_mask) & PCIE_ISR1_ALL_MASK);
774	780
775		− if (!status) {
776		− advk_writel(pcie, val, PCIE_ISR0_REG);
	781	+ if (!isr0_status && !isr1_status) {
	782	+ advk_writel(pcie, isr0_val, PCIE_ISR0_REG);
	783	+ advk_writel(pcie, isr1_val, PCIE_ISR1_REG);
777	784	return;
778	785	}
779	786
780	787	/* Process MSI interrupts */
781		− if (status & PCIE_ISR0_MSI_INT_PENDING)
	788	+ if (isr0_status & PCIE_ISR0_MSI_INT_PENDING)
782	789	advk_pcie_handle_msi(pcie);
783	790
784	791	/* Process legacy interrupts */
785	792	for (i = 0; i < PCI_NUM_INTX; i++) {
786		− if (!(status & PCIE_ISR0_INTX_ASSERT(i)))
	793	+ if (!(isr1_status & PCIE_ISR1_INTX_ASSERT(i)))
787	794	continue;
788	795
789		− advk_writel(pcie, PCIE_ISR0_INTX_ASSERT(i),
790		− PCIE_ISR0_REG);
	796	+ advk_writel(pcie, PCIE_ISR1_INTX_ASSERT(i),
	797	+ PCIE_ISR1_REG);
791	798
792	799	virq = irq_find_mapping(pcie->irq_domain, i);
793	800	generic_handle_irq(virq);

drivers/pci/pci-driver.c+3−2

@@ -958,10 +958,11 @@ static int pci_pm_freeze(struct device *dev) · 957 unmodified lines
958	958	* devices should not be touched during freeze/thaw transitions,
959	959	* however.
960	960	*/
961		− if (!dev_pm_test_driver_flags(dev, DPM_FLAG_SMART_SUSPEND))
	961	+ if (!dev_pm_smart_suspend_and_suspended(dev)) {
962	962	pm_runtime_resume(dev);
	963	+ pci_dev->state_saved = false;
	964	+ }
963	965
964		− pci_dev->state_saved = false;
965	966	if (pm->freeze) {
966	967	int error;
967	968

drivers/pci/pci.c+2−2

@@ -5273,11 +5273,11 @@ void pcie_print_link_status(struct pci_dev *dev) · 5272 unmodified lines
5273	5273	bw_avail = pcie_bandwidth_available(dev, &limiting_dev, &speed, &width);
5274	5274
5275	5275	if (bw_avail >= bw_cap)
5276		− pci_info(dev, "%u.%03u Gb/s available bandwidth (%s x%d link)\n",
	5276	+ pci_info(dev, "%u.%03u Gb/s available PCIe bandwidth (%s x%d link)\n",
5277	5277	bw_cap / 1000, bw_cap % 1000,
5278	5278	PCIE_SPEED2STR(speed_cap), width_cap);
5279	5279	else
5280		− pci_info(dev, "%u.%03u Gb/s available bandwidth, limited by %s x%d link at %s (capable of %u.%03u Gb/s with %s x%d link)\n",
	5280	+ pci_info(dev, "%u.%03u Gb/s available PCIe bandwidth, limited by %s x%d link at %s (capable of %u.%03u Gb/s with %s x%d link)\n",
5281	5281	bw_avail / 1000, bw_avail % 1000,
5282	5282	PCIE_SPEED2STR(speed), width,
5283	5283	limiting_dev ? pci_name(limiting_dev) : "<unknown>",

drivers/platform/x86/Kconfig+1−1

@@ -154,7 +154,7 @@ config DELL_LAPTOP · 153 unmodified lines
154	154	depends on ACPI_VIDEO \|\| ACPI_VIDEO = n
155	155	depends on RFKILL \|\| RFKILL = n
156	156	depends on SERIO_I8042
157		− select DELL_SMBIOS
	157	+ depends on DELL_SMBIOS
158	158	select POWER_SUPPLY
159	159	select LEDS_CLASS
160	160	select NEW_LEDS

drivers/platform/x86/asus-wireless.c+3−1

@@ -178,8 +178,10 @@ static int asus_wireless_remove(struct acpi_device *adev) · 177 unmodified lines
178	178	{
179	179	struct asus_wireless_data *data = acpi_driver_data(adev);
180	180
181		− if (data->wq)
	181	+ if (data->wq) {
	182	+ devm_led_classdev_unregister(&adev->dev, &data->led);
182	183	destroy_workqueue(data->wq);
	184	+ }
183	185	return 0;
184	186	}
185	187

drivers/rapidio/devices/rio_mport_cdev.c+9−10

@@ -740,10 +740,7 @@ static int do_dma_request(struct mport_dma_req *req, · 739 unmodified lines
740	740	tx->callback = dma_xfer_callback;
741	741	tx->callback_param = req;
742	742
743		− req->dmach = chan;
744		− req->sync = sync;
745	743	req->status = DMA_IN_PROGRESS;
746		− init_completion(&req->req_comp);
747	744	kref_get(&req->refcount);
748	745
749	746	cookie = dmaengine_submit(tx);
@@ -831,13 +828,20 @@ rio_dma_transfer(struct file *filp, u32 transfer_mode, · 81 unmodified lines
831	828	if (!req)
832	829	return -ENOMEM;
833	830
834		− kref_init(&req->refcount);
835		−
836	831	ret = get_dma_channel(priv);
837	832	if (ret) {
838	833	kfree(req);
839	834	return ret;
840	835	}
	836	+ chan = priv->dmach;
	837	+
	838	+ kref_init(&req->refcount);
	839	+ init_completion(&req->req_comp);
	840	+ req->dir = dir;
	841	+ req->filp = filp;
	842	+ req->priv = priv;
	843	+ req->dmach = chan;
	844	+ req->sync = sync;
841	845
842	846	/*
843	847	* If parameter loc_addr != NULL, we are transferring data from/to
@@ -925,11 +929,6 @@ rio_dma_transfer(struct file *filp, u32 transfer_mode, · 81 unmodified lines
925	929	xfer->offset, xfer->length);
926	930	}
927	931
928		− req->dir = dir;
929		− req->filp = filp;
930		− req->priv = priv;
931		− chan = priv->dmach;
932		−
933	932	nents = dma_map_sg(chan->device->dev,
934	933	req->sgt.sgl, req->sgt.nents, dir);
935	934	if (nents == 0) {

drivers/remoteproc/qcom_q6v5_pil.c+2−0

@@ -1083,6 +1083,7 @@ static int q6v5_alloc_memory_region(struct q6v5 *qproc) · 1082 unmodified lines
1083	1083	dev_err(qproc->dev, "unable to resolve mba region\n");
1084	1084	return ret;
1085	1085	}
	1086	+ of_node_put(node);
1086	1087
1087	1088	qproc->mba_phys = r.start;
1088	1089	qproc->mba_size = resource_size(&r);
@@ -1100,6 +1101,7 @@ static int q6v5_alloc_memory_region(struct q6v5 *qproc) · 11 unmodified lines
1100	1101	dev_err(qproc->dev, "unable to resolve mpss region\n");
1101	1102	return ret;
1102	1103	}
	1104	+ of_node_put(node);
1103	1105
1104	1106	qproc->mpss_phys = qproc->mpss_reloc = r.start;
1105	1107	qproc->mpss_size = resource_size(&r);

drivers/remoteproc/remoteproc_core.c+2−2

@@ -1163,7 +1163,7 @@ int rproc_trigger_recovery(struct rproc *rproc) · 1162 unmodified lines
1163	1163	if (ret)
1164	1164	return ret;
1165	1165
1166		− ret = rproc_stop(rproc, false);
	1166	+ ret = rproc_stop(rproc, true);
1167	1167	if (ret)
1168	1168	goto unlock_mutex;
1169	1169
@@ -1316,7 +1316,7 @@ void rproc_shutdown(struct rproc *rproc) · 146 unmodified lines
1316	1316	if (!atomic_dec_and_test(&rproc->power))
1317	1317	goto out;
1318	1318
1319		− ret = rproc_stop(rproc, true);
	1319	+ ret = rproc_stop(rproc, false);
1320	1320	if (ret) {
1321	1321	atomic_inc(&rproc->power);
1322	1322	goto out;

drivers/rpmsg/rpmsg_char.c+2−0

@@ -581,4 +581,6 @@ static void rpmsg_chrdev_exit(void) · 580 unmodified lines
581	581	unregister_chrdev_region(rpmsg_major, RPMSG_DEV_MAX);
582	582	}
583	583	module_exit(rpmsg_chrdev_exit);
	584	+
	585	+MODULE_ALIAS("rpmsg:rpmsg_chrdev");
584	586	MODULE_LICENSE("GPL v2");

drivers/rtc/rtc-opal.c+23−14

@@ -57,7 +57,7 @@ static void tm_to_opal(struct rtc_time tm, u32 y_m_d, u64 *h_m_s_ms) · 56 unmodified lines
57	57
58	58	static int opal_get_rtc_time(struct device dev, struct rtc_time tm)
59	59	{
60		− long rc = OPAL_BUSY;
	60	+ s64 rc = OPAL_BUSY;
61	61	int retries = 10;
62	62	u32 y_m_d;
63	63	u64 h_m_s_ms;
@@ -66,13 +66,17 @@ static int opal_get_rtc_time(struct device dev, struct rtc_time tm) · 2 unmodified lines
66	66
67	67	while (rc == OPAL_BUSY \|\| rc == OPAL_BUSY_EVENT) {
68	68	rc = opal_rtc_read(&__y_m_d, &__h_m_s_ms);
69		− if (rc == OPAL_BUSY_EVENT)
	69	+ if (rc == OPAL_BUSY_EVENT) {
	70	+ msleep(OPAL_BUSY_DELAY_MS);
70	71	opal_poll_events(NULL);
71		− else if (retries-- && (rc == OPAL_HARDWARE
72		− \|\| rc == OPAL_INTERNAL_ERROR))
73		− msleep(10);
74		− else if (rc != OPAL_BUSY && rc != OPAL_BUSY_EVENT)
75		− break;
	72	+ } else if (rc == OPAL_BUSY) {
	73	+ msleep(OPAL_BUSY_DELAY_MS);
	74	+ } else if (rc == OPAL_HARDWARE \|\| rc == OPAL_INTERNAL_ERROR) {
	75	+ if (retries--) {
	76	+ msleep(10); /* Wait 10ms before retry */
	77	+ rc = OPAL_BUSY; /* go around again */
	78	+ }
	79	+ }
76	80	}
77	81
78	82	if (rc != OPAL_SUCCESS)
@@ -87,21 +91,26 @@ static int opal_get_rtc_time(struct device dev, struct rtc_time tm) · 8 unmodified lines
87	91
88	92	static int opal_set_rtc_time(struct device dev, struct rtc_time tm)
89	93	{
90		− long rc = OPAL_BUSY;
	94	+ s64 rc = OPAL_BUSY;
91	95	int retries = 10;
92	96	u32 y_m_d = 0;
93	97	u64 h_m_s_ms = 0;
94	98
95	99	tm_to_opal(tm, &y_m_d, &h_m_s_ms);
	100	+
96	101	while (rc == OPAL_BUSY \|\| rc == OPAL_BUSY_EVENT) {
97	102	rc = opal_rtc_write(y_m_d, h_m_s_ms);
98		− if (rc == OPAL_BUSY_EVENT)
	103	+ if (rc == OPAL_BUSY_EVENT) {
	104	+ msleep(OPAL_BUSY_DELAY_MS);
99	105	opal_poll_events(NULL);
100		− else if (retries-- && (rc == OPAL_HARDWARE
101		− \|\| rc == OPAL_INTERNAL_ERROR))
102		− msleep(10);
103		− else if (rc != OPAL_BUSY && rc != OPAL_BUSY_EVENT)
104		− break;
	106	+ } else if (rc == OPAL_BUSY) {
	107	+ msleep(OPAL_BUSY_DELAY_MS);
	108	+ } else if (rc == OPAL_HARDWARE \|\| rc == OPAL_INTERNAL_ERROR) {
	109	+ if (retries--) {
	110	+ msleep(10); /* Wait 10ms before retry */
	111	+ rc = OPAL_BUSY; /* go around again */
	112	+ }
	113	+ }
105	114	}
106	115
107	116	return rc == OPAL_SUCCESS ? 0 : -EIO;

drivers/s390/block/dasd_alias.c+11−2

@@ -592,13 +592,22 @@ static int _schedule_lcu_update(struct alias_lcu *lcu, · 591 unmodified lines
592	592	int dasd_alias_add_device(struct dasd_device *device)
593	593	{
594	594	struct dasd_eckd_private *private = device->private;
595		− struct alias_lcu *lcu;
	595	+ __u8 uaddr = private->uid.real_unit_addr;
	596	+ struct alias_lcu *lcu = private->lcu;
596	597	unsigned long flags;
597	598	int rc;
598	599
599		− lcu = private->lcu;
600	600	rc = 0;
601	601	spin_lock_irqsave(&lcu->lock, flags);
	602	+ /*
	603	+ * Check if device and lcu type differ. If so, the uac data may be
	604	+ * outdated and needs to be updated.
	605	+ */
	606	+ if (private->uid.type != lcu->uac->unit[uaddr].ua_type) {
	607	+ lcu->flags \|= UPDATE_PENDING;
	608	+ DBF_DEV_EVENT(DBF_WARNING, device, "%s",
	609	+ "uid type mismatch - trigger rescan");
	610	+ }
602	611	if (!(lcu->flags & UPDATE_PENDING)) {
603	612	rc = _add_device_to_lcu(lcu, device, device);
604	613	if (rc)

drivers/s390/block/dasd_diag.c+0−1

@@ -27,7 +27,6 @@ · 26 unmodified lines
27	27	#include <asm/io.h>
28	28	#include <asm/irq.h>
29	29	#include <asm/vtoc.h>
30		−#include <asm/diag.h>
31	30
32	31	#include "dasd_int.h"
33	32	#include "dasd_diag.h"

drivers/s390/char/sclp_early_core.c+1−1

@@ -18,7 +18,7 @@ int sclp_init_state __section(.data) = sclp_init_state_uninitialized; · 17 unmodified lines
18	18	* Used to keep track of the size of the event masks. Qemu until version 2.11
19	19	* only supports 4 and needs a workaround.
20	20	*/
21		−bool sclp_mask_compat_mode;
	21	+bool sclp_mask_compat_mode __section(.data);
22	22
23	23	void sclp_early_wait_irq(void)
24	24	{

drivers/s390/cio/chsc.c+11−3

@@ -452,6 +452,7 @@ static void chsc_process_sei_link_incident(struct chsc_sei_nt0_area *sei_area) · 451 unmodified lines
452	452
453	453	static void chsc_process_sei_res_acc(struct chsc_sei_nt0_area *sei_area)
454	454	{
	455	+ struct channel_path *chp;
455	456	struct chp_link link;
456	457	struct chp_id chpid;
457	458	int status;
@@ -464,10 +465,17 @@ static void chsc_process_sei_res_acc(struct chsc_sei_nt0_area *sei_area) · 6 unmodified lines
464	465	chpid.id = sei_area->rsid;
465	466	/* allocate a new channel path structure, if needed */
466	467	status = chp_get_status(chpid);
467		− if (status < 0)
468		− chp_new(chpid);
469		− else if (!status)
	468	+ if (!status)
470	469	return;
	470	+
	471	+ if (status < 0) {
	472	+ chp_new(chpid);
	473	+ } else {
	474	+ chp = chpid_to_chp(chpid);
	475	+ mutex_lock(&chp->lock);
	476	+ chp_update_desc(chp);
	477	+ mutex_unlock(&chp->lock);
	478	+ }
471	479	memset(&link, 0, sizeof(struct chp_link));
472	480	link.chpid = chpid;
473	481	if ((sei_area->vf & 0xc0) != 0) {

drivers/s390/cio/vfio_ccw_fsm.c+12−7

@@ -20,12 +20,12 @@ static int fsm_io_helper(struct vfio_ccw_private *private) · 19 unmodified lines
20	20	int ccode;
21	21	__u8 lpm;
22	22	unsigned long flags;
	23	+ int ret;
23	24
24	25	sch = private->sch;
25	26
26	27	spin_lock_irqsave(sch->lock, flags);
27	28	private->state = VFIO_CCW_STATE_BUSY;
28		− spin_unlock_irqrestore(sch->lock, flags);
29	29
30	30	orb = cp_get_orb(&private->cp, (u32)(addr_t)sch, sch->lpm);
31	31
@@ -38,10 +38,12 @@ static int fsm_io_helper(struct vfio_ccw_private *private) · 6 unmodified lines
38	38	* Initialize device status information
39	39	*/
40	40	sch->schib.scsw.cmd.actl \|= SCSW_ACTL_START_PEND;
41		− return 0;
	41	+ ret = 0;
	42	+ break;
42	43	case 1: /* Status pending */
43	44	case 2: /* Busy */
44		− return -EBUSY;
	45	+ ret = -EBUSY;
	46	+ break;
45	47	case 3: /* Device/path not operational */
46	48	{
47	49	lpm = orb->cmd.lpm;
@@ -51,13 +53,16 @@ static int fsm_io_helper(struct vfio_ccw_private *private) · 3 unmodified lines
51	53	sch->lpm = 0;
52	54
53	55	if (cio_update_schib(sch))
54		− return -ENODEV;
55		−
56		− return sch->lpm ? -EACCES : -ENODEV;
	56	+ ret = -ENODEV;
	57	+ else
	58	+ ret = sch->lpm ? -EACCES : -ENODEV;
	59	+ break;
57	60	}
58	61	default:
59		− return ccode;
	62	+ ret = ccode;
60	63	}
	64	+ spin_unlock_irqrestore(sch->lock, flags);
	65	+ return ret;
61	66	}
62	67
63	68	static void fsm_notoper(struct vfio_ccw_private *private,

drivers/s390/net/qeth_core.h+0−2

@@ -557,7 +557,6 @@ enum qeth_prot_versions { · 556 unmodified lines
557	557	enum qeth_cmd_buffer_state {
558	558	BUF_STATE_FREE,
559	559	BUF_STATE_LOCKED,
560		− BUF_STATE_PROCESSED,
561	560	};
562	561
563	562	enum qeth_cq {
@@ -601,7 +600,6 @@ struct qeth_channel { · 37 unmodified lines
601	600	struct qeth_cmd_buffer iob[QETH_CMD_BUFFER_NO];
602	601	atomic_t irq_pending;
603	602	int io_buf_no;
604		− int buf_no;
605	603	};
606	604
607	605	/**

drivers/s390/net/qeth_core_main.c+71−87

@@ -706,7 +706,6 @@ void qeth_clear_ipacmd_list(struct qeth_card *card) · 705 unmodified lines
706	706	qeth_put_reply(reply);
707	707	}
708	708	spin_unlock_irqrestore(&card->lock, flags);
709		− atomic_set(&card->write.irq_pending, 0);
710	709	}
711	710	EXPORT_SYMBOL_GPL(qeth_clear_ipacmd_list);
712	711
@@ -818,7 +817,6 @@ void qeth_clear_cmd_buffers(struct qeth_channel *channel) · 105 unmodified lines
818	817
819	818	for (cnt = 0; cnt < QETH_CMD_BUFFER_NO; cnt++)
820	819	qeth_release_buffer(channel, &channel->iob[cnt]);
821		− channel->buf_no = 0;
822	820	channel->io_buf_no = 0;
823	821	}
824	822	EXPORT_SYMBOL_GPL(qeth_clear_cmd_buffers);
@@ -924,7 +922,6 @@ static int qeth_setup_channel(struct qeth_channel *channel) · 99 unmodified lines
924	922	kfree(channel->iob[cnt].data);
925	923	return -ENOMEM;
926	924	}
927		− channel->buf_no = 0;
928	925	channel->io_buf_no = 0;
929	926	atomic_set(&channel->irq_pending, 0);
930	927	spin_lock_init(&channel->iob_lock);
@@ -1100,16 +1097,9 @@ static void qeth_irq(struct ccw_device *cdev, unsigned long intparm, · 169 unmodified lines
1100	1097	{
1101	1098	int rc;
1102	1099	int cstat, dstat;
1103		− struct qeth_cmd_buffer *buffer;
	1100	+ struct qeth_cmd_buffer *iob = NULL;
1104	1101	struct qeth_channel *channel;
1105	1102	struct qeth_card *card;
1106		− struct qeth_cmd_buffer *iob;
1107		− __u8 index;
1108		−
1109		− if (__qeth_check_irb_error(cdev, intparm, irb))
1110		− return;
1111		− cstat = irb->scsw.cmd.cstat;
1112		− dstat = irb->scsw.cmd.dstat;
1113	1103
1114	1104	card = CARD_FROM_CDEV(cdev);
1115	1105	if (!card)
@@ -1127,6 +1117,19 @@ static void qeth_irq(struct ccw_device *cdev, unsigned long intparm, · 11 unmodified lines
1127	1117	channel = &card->data;
1128	1118	QETH_CARD_TEXT(card, 5, "data");
1129	1119	}
	1120	+
	1121	+ if (qeth_intparm_is_iob(intparm))
	1122	+ iob = (struct qeth_cmd_buffer *) __va((addr_t)intparm);
	1123	+
	1124	+ if (__qeth_check_irb_error(cdev, intparm, irb)) {
	1125	+ /* IO was terminated, free its resources. */
	1126	+ if (iob)
	1127	+ qeth_release_buffer(iob->channel, iob);
	1128	+ atomic_set(&channel->irq_pending, 0);
	1129	+ wake_up(&card->wait_q);
	1130	+ return;
	1131	+ }
	1132	+
1130	1133	atomic_set(&channel->irq_pending, 0);
1131	1134
1132	1135	if (irb->scsw.cmd.fctl & (SCSW_FCTL_CLEAR_FUNC))
@@ -1150,6 +1153,10 @@ static void qeth_irq(struct ccw_device *cdev, unsigned long intparm, · 17 unmodified lines
1150	1153	/* we don't have to handle this further */
1151	1154	intparm = 0;
1152	1155	}
	1156	+
	1157	+ cstat = irb->scsw.cmd.cstat;
	1158	+ dstat = irb->scsw.cmd.dstat;
	1159	+
1153	1160	if ((dstat & DEV_STAT_UNIT_EXCEP) \|\|
1154	1161	(dstat & DEV_STAT_UNIT_CHECK) \|\|
1155	1162	(cstat)) {
@@ -1182,25 +1189,15 @@ static void qeth_irq(struct ccw_device *cdev, unsigned long intparm, · 26 unmodified lines
1182	1189	channel->state = CH_STATE_RCD_DONE;
1183	1190	goto out;
1184	1191	}
1185		− if (intparm) {
1186		− buffer = (struct qeth_cmd_buffer *) __va((addr_t)intparm);
1187		− buffer->state = BUF_STATE_PROCESSED;
1188		− }
1189	1192	if (channel == &card->data)
1190	1193	return;
1191	1194	if (channel == &card->read &&
1192	1195	channel->state == CH_STATE_UP)
1193	1196	__qeth_issue_next_read(card);
1194	1197
1195		− iob = channel->iob;
1196		− index = channel->buf_no;
1197		− while (iob[index].state == BUF_STATE_PROCESSED) {
1198		− if (iob[index].callback != NULL)
1199		− iob[index].callback(channel, iob + index);
	1198	+ if (iob && iob->callback)
	1199	+ iob->callback(iob->channel, iob);
1200	1200
1201		− index = (index + 1) % QETH_CMD_BUFFER_NO;
1202		− }
1203		− channel->buf_no = index;
1204	1201	out:
1205	1202	wake_up(&card->wait_q);
1206	1203	return;
@@ -1870,8 +1867,8 @@ static int qeth_idx_activate_get_answer(struct qeth_channel *channel, · 663 unmodified lines
1870	1867	atomic_cmpxchg(&channel->irq_pending, 0, 1) == 0);
1871	1868	QETH_DBF_TEXT(SETUP, 6, "noirqpnd");
1872	1869	spin_lock_irqsave(get_ccwdev_lock(channel->ccwdev), flags);
1873		− rc = ccw_device_start(channel->ccwdev,
1874		− &channel->ccw, (addr_t) iob, 0, 0);
	1870	+ rc = ccw_device_start_timeout(channel->ccwdev, &channel->ccw,
	1871	+ (addr_t) iob, 0, 0, QETH_TIMEOUT);
1875	1872	spin_unlock_irqrestore(get_ccwdev_lock(channel->ccwdev), flags);
1876	1873
1877	1874	if (rc) {
@@ -1888,7 +1885,6 @@ static int qeth_idx_activate_get_answer(struct qeth_channel *channel, · 10 unmodified lines
1888	1885	if (channel->state != CH_STATE_UP) {
1889	1886	rc = -ETIME;
1890	1887	QETH_DBF_TEXT_(SETUP, 2, "3err%d", rc);
1891		− qeth_clear_cmd_buffers(channel);
1892	1888	} else
1893	1889	rc = 0;
1894	1890	return rc;
@@ -1942,8 +1938,8 @@ static int qeth_idx_activate_channel(struct qeth_channel *channel, · 47 unmodified lines
1942	1938	atomic_cmpxchg(&channel->irq_pending, 0, 1) == 0);
1943	1939	QETH_DBF_TEXT(SETUP, 6, "noirqpnd");
1944	1940	spin_lock_irqsave(get_ccwdev_lock(channel->ccwdev), flags);
1945		− rc = ccw_device_start(channel->ccwdev,
1946		− &channel->ccw, (addr_t) iob, 0, 0);
	1941	+ rc = ccw_device_start_timeout(channel->ccwdev, &channel->ccw,
	1942	+ (addr_t) iob, 0, 0, QETH_TIMEOUT);
1947	1943	spin_unlock_irqrestore(get_ccwdev_lock(channel->ccwdev), flags);
1948	1944
1949	1945	if (rc) {
@@ -1964,7 +1960,6 @@ static int qeth_idx_activate_channel(struct qeth_channel *channel, · 14 unmodified lines
1964	1960	QETH_DBF_MESSAGE(2, "%s IDX activate timed out\n",
1965	1961	dev_name(&channel->ccwdev->dev));
1966	1962	QETH_DBF_TEXT_(SETUP, 2, "2err%d", -ETIME);
1967		− qeth_clear_cmd_buffers(channel);
1968	1963	return -ETIME;
1969	1964	}
1970	1965	return qeth_idx_activate_get_answer(channel, idx_reply_cb);
@@ -2166,8 +2161,8 @@ int qeth_send_control_data(struct qeth_card *card, int len, · 195 unmodified lines
2166	2161
2167	2162	QETH_CARD_TEXT(card, 6, "noirqpnd");
2168	2163	spin_lock_irqsave(get_ccwdev_lock(card->write.ccwdev), flags);
2169		− rc = ccw_device_start(card->write.ccwdev, &card->write.ccw,
2170		− (addr_t) iob, 0, 0);
	2164	+ rc = ccw_device_start_timeout(CARD_WDEV(card), &card->write.ccw,
	2165	+ (addr_t) iob, 0, 0, event_timeout);
2171	2166	spin_unlock_irqrestore(get_ccwdev_lock(card->write.ccwdev), flags);
2172	2167	if (rc) {
2173	2168	QETH_DBF_MESSAGE(2, "%s qeth_send_control_data: "
@@ -2199,8 +2194,6 @@ int qeth_send_control_data(struct qeth_card *card, int len, · 25 unmodified lines
2199	2194	}
2200	2195	}
2201	2196
2202		− if (reply->rc == -EIO)
2203		− goto error;
2204	2197	rc = reply->rc;
2205	2198	qeth_put_reply(reply);
2206	2199	return rc;
@@ -2211,10 +2204,6 @@ time_err: · 4 unmodified lines
2211	2204	list_del_init(&reply->list);
2212	2205	spin_unlock_irqrestore(&reply->card->lock, flags);
2213	2206	atomic_inc(&reply->received);
2214		−error:
2215		− atomic_set(&card->write.irq_pending, 0);
2216		− qeth_release_buffer(iob->channel, iob);
2217		− card->write.buf_no = (card->write.buf_no + 1) % QETH_CMD_BUFFER_NO;
2218	2207	rc = reply->rc;
2219	2208	qeth_put_reply(reply);
2220	2209	return rc;
@@ -3033,28 +3022,23 @@ static int qeth_send_startlan(struct qeth_card *card) · 812 unmodified lines
3033	3022	return rc;
3034	3023	}
3035	3024
3036		−static int qeth_default_setadapterparms_cb(struct qeth_card *card,
3037		− struct qeth_reply *reply, unsigned long data)
	3025	+static int qeth_setadpparms_inspect_rc(struct qeth_ipa_cmd *cmd)
3038	3026	{
3039		− struct qeth_ipa_cmd *cmd;
3040		−
3041		− QETH_CARD_TEXT(card, 4, "defadpcb");
3042		−
3043		− cmd = (struct qeth_ipa_cmd *) data;
3044		− if (cmd->hdr.return_code == 0)
	3027	+ if (!cmd->hdr.return_code)
3045	3028	cmd->hdr.return_code =
3046	3029	cmd->data.setadapterparms.hdr.return_code;
3047		− return 0;
	3030	+ return cmd->hdr.return_code;
3048	3031	}
3049	3032
3050	3033	static int qeth_query_setadapterparms_cb(struct qeth_card *card,
3051	3034	struct qeth_reply *reply, unsigned long data)
3052	3035	{
3053		− struct qeth_ipa_cmd *cmd;
	3036	+ struct qeth_ipa_cmd cmd = (struct qeth_ipa_cmd ) data;
3054	3037
3055	3038	QETH_CARD_TEXT(card, 3, "quyadpcb");
	3039	+ if (qeth_setadpparms_inspect_rc(cmd))
	3040	+ return 0;
3056	3041
3057		− cmd = (struct qeth_ipa_cmd *) data;
3058	3042	if (cmd->data.setadapterparms.data.query_cmds_supp.lan_type & 0x7f) {
3059	3043	card->info.link_type =
3060	3044	cmd->data.setadapterparms.data.query_cmds_supp.lan_type;
@@ -3062,7 +3046,7 @@ static int qeth_query_setadapterparms_cb(struct qeth_card *card, · 1 unmodified line
3062	3046	}
3063	3047	card->options.adp.supported_funcs =
3064	3048	cmd->data.setadapterparms.data.query_cmds_supp.supported_cmds;
3065		− return qeth_default_setadapterparms_cb(card, reply, (unsigned long)cmd);
	3049	+ return 0;
3066	3050	}
3067	3051
3068	3052	static struct qeth_cmd_buffer qeth_get_adapter_cmd(struct qeth_card card,
@@ -3154,22 +3138,20 @@ EXPORT_SYMBOL_GPL(qeth_query_ipassists); · 85 unmodified lines
3154	3138	static int qeth_query_switch_attributes_cb(struct qeth_card *card,
3155	3139	struct qeth_reply *reply, unsigned long data)
3156	3140	{
3157		− struct qeth_ipa_cmd *cmd;
3158		− struct qeth_switch_info *sw_info;
	3141	+ struct qeth_ipa_cmd cmd = (struct qeth_ipa_cmd ) data;
3159	3142	struct qeth_query_switch_attributes *attrs;
	3143	+ struct qeth_switch_info *sw_info;
3160	3144
3161	3145	QETH_CARD_TEXT(card, 2, "qswiatcb");
3162		− cmd = (struct qeth_ipa_cmd *) data;
3163		− sw_info = (struct qeth_switch_info *)reply->param;
3164		− if (cmd->data.setadapterparms.hdr.return_code == 0) {
3165		− attrs = &cmd->data.setadapterparms.data.query_switch_attributes;
3166		− sw_info->capabilities = attrs->capabilities;
3167		− sw_info->settings = attrs->settings;
3168		− QETH_CARD_TEXT_(card, 2, "%04x%04x", sw_info->capabilities,
3169		− sw_info->settings);
3170		− }
3171		− qeth_default_setadapterparms_cb(card, reply, (unsigned long) cmd);
	3146	+ if (qeth_setadpparms_inspect_rc(cmd))
	3147	+ return 0;
3172	3148
	3149	+ sw_info = (struct qeth_switch_info *)reply->param;
	3150	+ attrs = &cmd->data.setadapterparms.data.query_switch_attributes;
	3151	+ sw_info->capabilities = attrs->capabilities;
	3152	+ sw_info->settings = attrs->settings;
	3153	+ QETH_CARD_TEXT_(card, 2, "%04x%04x", sw_info->capabilities,
	3154	+ sw_info->settings);
3173	3155	return 0;
3174	3156	}
3175	3157
@@ -4207,16 +4189,13 @@ EXPORT_SYMBOL_GPL(qeth_do_send_packet); · 1031 unmodified lines
4207	4189	static int qeth_setadp_promisc_mode_cb(struct qeth_card *card,
4208	4190	struct qeth_reply *reply, unsigned long data)
4209	4191	{
4210		− struct qeth_ipa_cmd *cmd;
	4192	+ struct qeth_ipa_cmd cmd = (struct qeth_ipa_cmd ) data;
4211	4193	struct qeth_ipacmd_setadpparms *setparms;
4212	4194
4213	4195	QETH_CARD_TEXT(card, 4, "prmadpcb");
4214	4196
4215		− cmd = (struct qeth_ipa_cmd *) data;

138 more lines aren’t shown here.

drivers/s390/net/qeth_core_mpc.h+12−0

This diff is large: its 19 lines aren’t shown here.

drivers/s390/net/qeth_l2_main.c+33−27

This diff is large: its 115 lines aren’t shown here.

drivers/s390/net/smsgiucv.c+1−1

This diff is large: its 9 lines aren’t shown here.

drivers/sbus/char/oradax.c+1−1

This diff is large: its 9 lines aren’t shown here.

drivers/scsi/fnic/fnic_trace.c+1−1

This diff is large: its 9 lines aren’t shown here.

drivers/scsi/isci/port_config.c+1−2

This diff is large: its 17 lines aren’t shown here.

drivers/scsi/megaraid/megaraid_sas_fusion.c+3−3

This diff is large: its 16 lines aren’t shown here.

drivers/scsi/scsi_debug.c+24−9

This diff is large: its 119 lines aren’t shown here.

drivers/scsi/scsi_transport_iscsi.c+18−11

This diff is large: its 93 lines aren’t shown here.

drivers/scsi/sd.c+2−0

This diff is large: its 9 lines aren’t shown here.

drivers/scsi/sd_zbc.c+82−58

This diff is large: its 283 lines aren’t shown here.

drivers/scsi/storvsc_drv.c+5−2

This diff is large: its 17 lines aren’t shown here.

drivers/scsi/ufs/ufshcd.c+40−0

This diff is large: its 87 lines aren’t shown here.

drivers/slimbus/messaging.c+1−1

This diff is large: its 9 lines aren’t shown here.

drivers/soc/bcm/raspberrypi-power.c+1−1

This diff is large: its 9 lines aren’t shown here.

drivers/staging/media/imx/imx-media-csi.c+1−1

This diff is large: its 9 lines aren’t shown here.

drivers/staging/wilc1000/host_interface.c+1−1

This diff is large: its 9 lines aren’t shown here.

drivers/target/target_core_iblock.c+4−4

This diff is large: its 24 lines aren’t shown here.

drivers/target/target_core_pscsi.c+2−0

This diff is large: its 16 lines aren’t shown here.

drivers/tty/n_gsm.c+22−1

This diff is large: its 56 lines aren’t shown here.

drivers/tty/serial/earlycon.c+4−2

This diff is large: its 20 lines aren’t shown here.

drivers/tty/serial/imx.c+18−1

This diff is large: its 40 lines aren’t shown here.

drivers/tty/serial/mvebu-uart.c+0−1

This diff is large: its 8 lines aren’t shown here.

drivers/tty/serial/qcom_geni_serial.c+6−4

This diff is large: its 25 lines aren’t shown here.

drivers/tty/serial/xilinx_uartps.c+1−1

This diff is large: its 9 lines aren’t shown here.

drivers/tty/tty_io.c+4−1

This diff is large: its 12 lines aren’t shown here.

drivers/tty/tty_ldisc.c+13−16

This diff is large: its 59 lines aren’t shown here.

drivers/uio/uio_hv_generic.c+23−49

This diff is large: its 136 lines aren’t shown here.

drivers/usb/Kconfig+1−0

This diff is large: its 7 lines aren’t shown here.

drivers/usb/core/config.c+3−1

This diff is large: its 11 lines aren’t shown here.

drivers/usb/core/hcd.c+13−6

This diff is large: its 65 lines aren’t shown here.

drivers/usb/core/hub.c+9−1

This diff is large: its 31 lines aren’t shown here.

drivers/usb/core/phy.c+66−27

This diff is large: its 152 lines aren’t shown here.

drivers/usb/core/phy.h+21−1

This diff is large: its 29 lines aren’t shown here.

drivers/usb/core/quirks.c+3−0

This diff is large: its 10 lines aren’t shown here.

drivers/usb/dwc2/core.h+2−0

This diff is large: its 16 lines aren’t shown here.

drivers/usb/dwc2/gadget.c+21−0

This diff is large: its 28 lines aren’t shown here.

drivers/usb/dwc2/hcd.c+8−5

This diff is large: its 28 lines aren’t shown here.

drivers/usb/dwc2/pci.c+3−1

This diff is large: its 12 lines aren’t shown here.

drivers/usb/dwc3/gadget.c+2−2

This diff is large: its 18 lines aren’t shown here.

drivers/usb/gadget/function/f_phonet.c+1−1

This diff is large: its 9 lines aren’t shown here.

drivers/usb/host/ehci-mem.c+2−1

This diff is large: its 12 lines aren’t shown here.

drivers/usb/host/ehci-sched.c+4−2

This diff is large: its 34 lines aren’t shown here.

drivers/usb/host/xhci-dbgtty.c+5−3

This diff is large: its 15 lines aren’t shown here.

drivers/usb/host/xhci-pci.c+4−1

This diff is large: its 12 lines aren’t shown here.

drivers/usb/host/xhci-plat.c+23−9

This diff is large: its 88 lines aren’t shown here.

drivers/usb/host/xhci.c+1−0

This diff is large: its 8 lines aren’t shown here.

drivers/usb/host/xhci.h+2−1

This diff is large: its 11 lines aren’t shown here.

drivers/usb/musb/musb_dsps.c+0−2

This diff is large: its 16 lines aren’t shown here.

drivers/usb/musb/musb_gadget.c+2−1

This diff is large: its 17 lines aren’t shown here.

drivers/usb/musb/musb_host.c+4−1

This diff is large: its 19 lines aren’t shown here.

drivers/usb/serial/Kconfig+1−0

This diff is large: its 8 lines aren’t shown here.

drivers/usb/serial/cp210x.c+1−0

This diff is large: its 8 lines aren’t shown here.

drivers/usb/serial/ftdi_sio.c+2−1

This diff is large: its 10 lines aren’t shown here.

drivers/usb/serial/option.c+5−0

This diff is large: its 19 lines aren’t shown here.

drivers/usb/serial/usb-serial-simple.c+7−0

This diff is large: its 28 lines aren’t shown here.

drivers/usb/serial/visor.c+35−34

This diff is large: its 83 lines aren’t shown here.

drivers/usb/typec/tcpm.c+1−0

This diff is large: its 8 lines aren’t shown here.

drivers/usb/typec/tps6598x.c+39−8

This diff is large: its 96 lines aren’t shown here.

drivers/usb/typec/ucsi/Makefile+1−1

This diff is large: its 8 lines aren’t shown here.

drivers/usb/typec/ucsi/ucsi.c+1−1

This diff is large: its 9 lines aren’t shown here.

drivers/usb/usbip/stub_main.c+5−0

This diff is large: its 13 lines aren’t shown here.

drivers/usb/usbip/usbip_common.h+1−1

This diff is large: its 9 lines aren’t shown here.

drivers/usb/usbip/usbip_event.c+0−4

This diff is large: its 11 lines aren’t shown here.

drivers/usb/usbip/vhci_hcd.c+13−0

This diff is large: its 53 lines aren’t shown here.

drivers/virt/vboxguest/vboxguest_core.c+38−32

This diff is large: its 190 lines aren’t shown here.

drivers/virt/vboxguest/vboxguest_core.h+9−0

This diff is large: its 14 lines aren’t shown here.

drivers/virt/vboxguest/vboxguest_linux.c+16−3

This diff is large: its 40 lines aren’t shown here.

drivers/virt/vboxguest/vboxguest_utils.c+13−4

This diff is large: its 53 lines aren’t shown here.

drivers/watchdog/aspeed_wdt.c+7−0

This diff is large: its 28 lines aren’t shown here.

drivers/watchdog/renesas_wdt.c+4−2

This diff is large: its 22 lines aren’t shown here.

drivers/watchdog/sch311x_wdt.c+1−1

This diff is large: its 9 lines aren’t shown here.

drivers/watchdog/w83977f_wdt.c+1−1

This diff is large: its 9 lines aren’t shown here.

drivers/watchdog/wafer5823wdt.c+1−1

This diff is large: its 9 lines aren’t shown here.

drivers/xen/xen-pciback/conf_space_quirks.c+1−1

This diff is large: its 9 lines aren’t shown here.

drivers/xen/xen-pciback/pci_stub.c+4−4

This diff is large: its 36 lines aren’t shown here.

drivers/xen/xenbus/xenbus_dev_frontend.c+2−1

This diff is large: its 17 lines aren’t shown here.

fs/afs/server.c+8−1

This diff is large: its 17 lines aren’t shown here.

fs/autofs4/root.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/binfmt_elf.c+4−4

This diff is large: its 15 lines aren’t shown here.

fs/btrfs/ctree.h+25−0

This diff is large: its 39 lines aren’t shown here.

fs/btrfs/delayed-inode.c+16−4

This diff is large: its 55 lines aren’t shown here.

fs/btrfs/delayed-ref.c+14−5

This diff is large: its 64 lines aren’t shown here.

fs/btrfs/delayed-ref.h+1−0

This diff is large: its 8 lines aren’t shown here.

fs/btrfs/disk-io.c+1−0

This diff is large: its 8 lines aren’t shown here.

fs/btrfs/extent-tree.c+64−16

This diff is large: its 232 lines aren’t shown here.

fs/btrfs/file.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/btrfs/inode.c+12−8

This diff is large: its 46 lines aren’t shown here.

fs/btrfs/print-tree.c+15−10

This diff is large: its 57 lines aren’t shown here.

fs/btrfs/print-tree.h+1−1

This diff is large: its 8 lines aren’t shown here.

fs/btrfs/qgroup.c+41−2

This diff is large: its 72 lines aren’t shown here.

fs/btrfs/relocation.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/btrfs/send.c+4−0

This diff is large: its 11 lines aren’t shown here.

fs/btrfs/transaction.c+1−0

This diff is large: its 8 lines aren’t shown here.

fs/btrfs/transaction.h+14−0

This diff is large: its 21 lines aren’t shown here.

fs/ceph/inode.c+7−3

This diff is large: its 36 lines aren’t shown here.

fs/ceph/xattr.c+25−3

This diff is large: its 55 lines aren’t shown here.

fs/cifs/Kconfig+1−1

This diff is large: its 9 lines aren’t shown here.

fs/cifs/cifs_debug.h+1−1

This diff is large: its 9 lines aren’t shown here.

fs/cifs/cifssmb.c+3−0

This diff is large: its 10 lines aren’t shown here.

fs/cifs/connect.c+16−16

This diff is large: its 46 lines aren’t shown here.

fs/cifs/dir.c+5−4

This diff is large: its 31 lines aren’t shown here.

fs/cifs/file.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/cifs/smb2ops.c+16−6

This diff is large: its 52 lines aren’t shown here.

fs/cifs/smb2pdu.c+7−6

This diff is large: its 41 lines aren’t shown here.

fs/cifs/smb2pdu.h+1−1

This diff is large: its 9 lines aren’t shown here.

fs/cifs/smbdirect.c+18−24

This diff is large: its 87 lines aren’t shown here.

fs/cifs/transport.c+6−3

This diff is large: its 31 lines aren’t shown here.

fs/ecryptfs/crypto.c+28−13

This diff is large: its 62 lines aren’t shown here.

fs/ecryptfs/file.c+16−5

This diff is large: its 34 lines aren’t shown here.

fs/ecryptfs/inode.c+1−2

This diff is large: its 10 lines aren’t shown here.

fs/ecryptfs/keystore.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/ext2/file.c+2−2

This diff is large: its 14 lines aren’t shown here.

fs/ext4/balloc.c+5−4

This diff is large: its 37 lines aren’t shown here.

fs/ext4/extents.c+11−5

This diff is large: its 30 lines aren’t shown here.

fs/ext4/super.c+1−0

This diff is large: its 7 lines aren’t shown here.

fs/fs-writeback.c+5−4

This diff is large: its 25 lines aren’t shown here.

fs/isofs/compress.c+16−3

This diff is large: its 68 lines aren’t shown here.

fs/isofs/inode.c+3−0

This diff is large: its 11 lines aren’t shown here.

fs/jbd2/transaction.c+1−0

This diff is large: its 8 lines aren’t shown here.

fs/jffs2/super.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/namespace.c+3−2

This diff is large: its 19 lines aren’t shown here.

fs/notify/fanotify/fanotify.c+15−19

This diff is large: its 48 lines aren’t shown here.

fs/notify/fsnotify.c+11−14

This diff is large: its 50 lines aren’t shown here.

fs/orangefs/super.c+5−0

This diff is large: its 12 lines aren’t shown here.

fs/proc/base.c+6−0

This diff is large: its 13 lines aren’t shown here.

fs/proc/loadavg.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/proc/task_mmu.c+5−1

This diff is large: its 22 lines aren’t shown here.

fs/quota/dquot.c+1−1

This diff is large: its 9 lines aren’t shown here.

fs/super.c+4−5

This diff is large: its 30 lines aren’t shown here.

fs/udf/unicode.c+6−0

This diff is large: its 20 lines aren’t shown here.

fs/xfs/libxfs/xfs_attr.c+8−1

This diff is large: its 16 lines aren’t shown here.

fs/xfs/libxfs/xfs_bmap.c+4−0

This diff is large: its 17 lines aren’t shown here.

fs/xfs/libxfs/xfs_inode_buf.c+21−0

This diff is large: its 41 lines aren’t shown here.

fs/xfs/xfs_file.c+19−5

This diff is large: its 51 lines aren’t shown here.

include/asm-generic/vmlinux.lds.h+1−1

This diff is large: its 9 lines aren’t shown here.

include/drm/drm_hdcp.h+1−1

This diff is large: its 9 lines aren’t shown here.

include/dt-bindings/clock/stm32mp1-clks.h+2−2

This diff is large: its 25 lines aren’t shown here.

include/kvm/arm_psci.h+14−2

This diff is large: its 30 lines aren’t shown here.

include/kvm/arm_vgic.h+1−0

This diff is large: its 8 lines aren’t shown here.

include/linux/backing-dev-defs.h+5−0

This diff is large: its 12 lines aren’t shown here.

include/linux/backing-dev.h+16−14

This diff is large: its 74 lines aren’t shown here.

include/linux/blk-mq.h+3−0

This diff is large: its 10 lines aren’t shown here.

include/linux/blkdev.h+6−0

This diff is large: its 20 lines aren’t shown here.

include/linux/bpf.h+5−3

This diff is large: its 43 lines aren’t shown here.

include/linux/clk-provider.h+3−0

This diff is large: its 10 lines aren’t shown here.

include/linux/compiler-clang.h+3−0

This diff is large: its 10 lines aren’t shown here.

include/linux/coresight-pmu.h+1−12

This diff is large: its 20 lines aren’t shown here.

include/linux/device.h+4−2

This diff is large: its 20 lines aren’t shown here.

include/linux/ethtool.h+2−0

This diff is large: its 9 lines aren’t shown here.

include/linux/fsnotify_backend.h+2−4

This diff is large: its 23 lines aren’t shown here.

include/linux/genhd.h+3−1

This diff is large: its 11 lines aren’t shown here.

include/linux/hid.h+8−1

This diff is large: its 23 lines aren’t shown here.

include/linux/hrtimer.h+2−0

This diff is large: its 12 lines aren’t shown here.

include/linux/if_vlan.h+5−2

This diff is large: its 28 lines aren’t shown here.

include/linux/livepatch.h+13−6

This diff is large: its 27 lines aren’t shown here.

include/linux/microchipphy.h+8−0

This diff is large: its 13 lines aren’t shown here.

include/linux/mlx5/driver.h+3−9

This diff is large: its 29 lines aren’t shown here.

include/linux/mtd/flashchip.h+1−0

This diff is large: its 8 lines aren’t shown here.

include/linux/remoteproc.h+1−1

This diff is large: its 9 lines aren’t shown here.

include/linux/serial_core.h+14−7

This diff is large: its 38 lines aren’t shown here.

include/linux/shrinker.h+5−2

This diff is large: its 12 lines aren’t shown here.

include/linux/stringhash.h+2−2

This diff is large: its 12 lines aren’t shown here.

include/linux/textsearch.h+2−2

This diff is large: its 18 lines aren’t shown here.

include/linux/thread_info.h+1−5

This diff is large: its 13 lines aren’t shown here.

include/linux/ti-emif-sram.h+75−0

This diff is large: its 82 lines aren’t shown here.

include/linux/timekeeper_internal.h+0−2

This diff is large: its 16 lines aren’t shown here.

include/linux/timekeeping.h+25−12

This diff is large: its 88 lines aren’t shown here.

include/linux/timekeeping32.h+0−3

This diff is large: its 10 lines aren’t shown here.

include/linux/timer.h+0−2

This diff is large: its 9 lines aren’t shown here.

include/linux/tty.h+1−1

This diff is large: its 9 lines aren’t shown here.

include/linux/usb/composite.h+1−1

This diff is large: its 9 lines aren’t shown here.

include/linux/vbox_utils.h+0−23

This diff is large: its 37 lines aren’t shown here.

include/linux/virtio.h+3−0

This diff is large: its 10 lines aren’t shown here.

include/linux/wait_bit.h+17−0

This diff is large: its 22 lines aren’t shown here.

include/media/i2c/tvp7002.h+1−1

This diff is large: its 9 lines aren’t shown here.

include/media/videobuf-core.h+2−2

This diff is large: its 14 lines aren’t shown here.

include/media/videobuf-dma-sg.h+2−2

This diff is large: its 14 lines aren’t shown here.

include/media/videobuf-vmalloc.h+1−1

This diff is large: its 9 lines aren’t shown here.

include/net/ife.h+2−1

This diff is large: its 10 lines aren’t shown here.

include/net/llc_conn.h+1−0

This diff is large: its 8 lines aren’t shown here.

include/net/tls.h+1−0

This diff is large: its 8 lines aren’t shown here.

include/scsi/scsi_dbg.h+0−2

This diff is large: its 9 lines aren’t shown here.

include/soc/bcm2835/raspberrypi-firmware.h+2−2

This diff is large: its 16 lines aren’t shown here.

include/sound/control.h+5−2

This diff is large: its 25 lines aren’t shown here.

include/trace/events/initcall.h+11−3

This diff is large: its 28 lines aren’t shown here.

include/trace/events/ufs.h+27−0

This diff is large: its 34 lines aren’t shown here.

include/trace/events/workqueue.h+2−0

This diff is large: its 9 lines aren’t shown here.

include/uapi/linux/if_infiniband.h+1−1

This diff is large: its 6 lines aren’t shown here.

include/uapi/linux/kvm.h+7−0

This diff is large: its 14 lines aren’t shown here.

include/uapi/linux/perf_event.h+15−3

This diff is large: its 27 lines aren’t shown here.

include/uapi/linux/random.h+3−0

This diff is large: its 10 lines aren’t shown here.

include/uapi/linux/rds.h+1−1

This diff is large: its 6 lines aren’t shown here.

include/uapi/linux/sysctl.h+0−18

This diff is large: its 25 lines aren’t shown here.

include/uapi/linux/time.h+0−1

This diff is large: its 8 lines aren’t shown here.

include/uapi/linux/tls.h+1−1

This diff is large: its 6 lines aren’t shown here.

include/uapi/linux/virtio_balloon.h+15−0

This diff is large: its 22 lines aren’t shown here.

include/uapi/rdma/cxgb3-abi.h+1−1

This diff is large: its 6 lines aren’t shown here.

include/uapi/rdma/cxgb4-abi.h+1−1

This diff is large: its 6 lines aren’t shown here.

include/uapi/rdma/hns-abi.h+1−1

This diff is large: its 6 lines aren’t shown here.

include/uapi/rdma/ib_user_cm.h+1−1

This diff is large: its 6 lines aren’t shown here.

include/uapi/rdma/ib_user_ioctl_verbs.h+1−1

This diff is large: its 6 lines aren’t shown here.

include/uapi/rdma/ib_user_mad.h+1−1

This diff is large: its 6 lines aren’t shown here.

include/uapi/rdma/ib_user_sa.h+1−1

This diff is large: its 6 lines aren’t shown here.

include/uapi/rdma/ib_user_verbs.h+1−1

This diff is large: its 6 lines aren’t shown here.

include/uapi/rdma/mlx4-abi.h+1−1

This diff is large: its 6 lines aren’t shown here.

include/uapi/rdma/mlx5-abi.h+1−1

This diff is large: its 6 lines aren’t shown here.

include/uapi/rdma/mthca-abi.h+1−1

This diff is large: its 6 lines aren’t shown here.

include/uapi/rdma/nes-abi.h+1−1

This diff is large: its 6 lines aren’t shown here.

include/uapi/rdma/qedr-abi.h+1−1

This diff is large: its 6 lines aren’t shown here.

include/uapi/rdma/rdma_user_cm.h+1−1

This diff is large: its 6 lines aren’t shown here.

include/uapi/rdma/rdma_user_ioctl.h+1−1

This diff is large: its 6 lines aren’t shown here.

include/uapi/rdma/rdma_user_rxe.h+1−1

This diff is large: its 6 lines aren’t shown here.

include/xen/interface/io/sndif.h+306−16

This diff is large: its 483 lines aren’t shown here.

kernel/bpf/arraymap.c+2−1

This diff is large: its 17 lines aren’t shown here.

kernel/bpf/core.c+29−16

This diff is large: its 83 lines aren’t shown here.

kernel/bpf/sockmap.c+73−29

This diff is large: its 234 lines aren’t shown here.

kernel/bpf/syscall.c+2−2

This diff is large: its 11 lines aren’t shown here.

kernel/events/callchain.c+11−14

This diff is large: its 35 lines aren’t shown here.

kernel/events/core.c+6−2

This diff is large: its 23 lines aren’t shown here.

kernel/events/uprobes.c+3−4

This diff is large: its 35 lines aren’t shown here.

kernel/fork.c+1−2

This diff is large: its 12 lines aren’t shown here.

kernel/kprobes.c+1−1

This diff is large: its 9 lines aren’t shown here.

kernel/livepatch/shadow.c+71−37

This diff is large: its 203 lines aren’t shown here.

kernel/module.c+2−1

This diff is large: its 10 lines aren’t shown here.

kernel/sysctl_binary.c+1−19

This diff is large: its 34 lines aren’t shown here.

kernel/time/clocksource.c+44−19

This diff is large: its 189 lines aren’t shown here.

kernel/time/hrtimer.c+14−2

This diff is large: its 49 lines aren’t shown here.

kernel/time/posix-cpu-timers.c+3−1

This diff is large: its 14 lines aren’t shown here.

kernel/time/posix-stubs.c+0−2

This diff is large: its 9 lines aren’t shown here.

kernel/time/posix-timers.c+17−9

This diff is large: its 56 lines aren’t shown here.

kernel/time/tick-common.c+0−15

This diff is large: its 29 lines aren’t shown here.

kernel/time/tick-internal.h+0−6

This diff is large: its 13 lines aren’t shown here.

kernel/time/tick-oneshot.c+5−6

This diff is large: its 22 lines aren’t shown here.

kernel/time/tick-sched.c+5−14

This diff is large: its 34 lines aren’t shown here.

kernel/time/timekeeping.c+37−48

This diff is large: its 155 lines aren’t shown here.

kernel/time/timekeeping.h+1−0

This diff is large: its 8 lines aren’t shown here.

kernel/trace/bpf_trace.c+21−4

This diff is large: its 45 lines aren’t shown here.

kernel/trace/ftrace.c+2−2

This diff is large: its 13 lines aren’t shown here.

kernel/trace/trace.c+1−1

This diff is large: its 9 lines aren’t shown here.

kernel/trace/trace_entries.h+1−1

This diff is large: its 9 lines aren’t shown here.

kernel/trace/trace_events_filter.c+7−7

This diff is large: its 22 lines aren’t shown here.

kernel/trace/trace_events_hist.c+12−0

This diff is large: its 33 lines aren’t shown here.

kernel/trace/trace_kprobe.c+0−2

This diff is large: its 9 lines aren’t shown here.

kernel/trace/trace_stack.c+1−1

This diff is large: its 9 lines aren’t shown here.

kernel/trace/trace_uprobe.c+14−21

This diff is large: its 129 lines aren’t shown here.

kernel/tracepoint.c+2−2

This diff is large: its 18 lines aren’t shown here.

lib/dma-direct.c+2−1

This diff is large: its 10 lines aren’t shown here.

lib/errseq.c+9−14

This diff is large: its 37 lines aren’t shown here.

lib/kobject.c+5−6

This diff is large: its 19 lines aren’t shown here.

lib/swiotlb.c+1−1

This diff is large: its 9 lines aren’t shown here.

lib/textsearch.c+23−17

This diff is large: its 101 lines aren’t shown here.

mm/backing-dev.c+2−1

This diff is large: its 17 lines aren’t shown here.

mm/filemap.c+4−5

This diff is large: its 37 lines aren’t shown here.

mm/huge_memory.c+4−1

This diff is large: its 12 lines aren’t shown here.

mm/memcontrol.c+1−1

This diff is large: its 9 lines aren’t shown here.

mm/migrate.c+19−3

This diff is large: its 56 lines aren’t shown here.

mm/mmap.c+10−1

This diff is large: its 22 lines aren’t shown here.

mm/page-writeback.c+9−9

This diff is large: its 62 lines aren’t shown here.

mm/rmap.c+0−3

This diff is large: its 10 lines aren’t shown here.

mm/vmscan.c+20−1

This diff is large: its 39 lines aren’t shown here.

net/bridge/br_if.c+2−2

This diff is large: its 11 lines aren’t shown here.

net/bridge/netfilter/ebtables.c+6−5

This diff is large: its 20 lines aren’t shown here.

net/caif/chnl_net.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/ceph/messenger.c+7−0

This diff is large: its 21 lines aren’t shown here.

net/ceph/mon_client.c+11−3

This diff is large: its 35 lines aren’t shown here.

net/compat.c+4−2

This diff is large: its 20 lines aren’t shown here.

net/core/dev.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/core/dev_addr_lists.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/core/ethtool.c+5−0

This diff is large: its 12 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+26−14

This diff is large: its 83 lines aren’t shown here.

net/dccp/ccids/ccid2.c+12−2

This diff is large: its 35 lines aren’t shown here.

net/dccp/timer.c+1−1

This diff is large: its 16 lines aren’t shown here.

net/dns_resolver/dns_key.c+5−7

This diff is large: its 26 lines aren’t shown here.

net/ife/ife.c+36−2

This diff is large: its 56 lines aren’t shown here.

net/ipv4/ip_output.c+5−3

This diff is large: its 22 lines aren’t shown here.

net/ipv4/route.c+53−65

This diff is large: its 175 lines aren’t shown here.

net/ipv4/tcp.c+9−6

This diff is large: its 57 lines aren’t shown here.

net/ipv4/tcp_bbr.c+3−1

This diff is large: its 11 lines aren’t shown here.

net/ipv4/tcp_input.c+2−5

This diff is large: its 14 lines aren’t shown here.

net/ipv6/netfilter/Kconfig+28−27

This diff is large: its 75 lines aren’t shown here.

net/ipv6/route.c+8−1

This diff is large: its 34 lines aren’t shown here.

net/ipv6/seg6_iptunnel.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/l2tp/l2tp_core.c+20−20

This diff is large: its 54 lines aren’t shown here.

net/l2tp/l2tp_core.h+2−1

This diff is large: its 17 lines aren’t shown here.

net/l2tp/l2tp_debugfs.c+16−2

This diff is large: its 32 lines aren’t shown here.

net/l2tp/l2tp_netlink.c+8−3

This diff is large: its 45 lines aren’t shown here.

net/l2tp/l2tp_ppp.c+27−7

This diff is large: its 59 lines aren’t shown here.

net/llc/af_llc.c+12−2

This diff is large: its 22 lines aren’t shown here.

net/llc/llc_c_ac.c+1−8

This diff is large: its 16 lines aren’t shown here.

net/llc/llc_conn.c+21−1

This diff is large: its 36 lines aren’t shown here.

net/netfilter/Kconfig+1−0

This diff is large: its 8 lines aren’t shown here.

net/netfilter/ipvs/ip_vs_ctl.c+0−8

This diff is large: its 25 lines aren’t shown here.

net/netfilter/ipvs/ip_vs_sync.c+80−75

This diff is large: its 380 lines aren’t shown here.

net/netfilter/nf_conntrack_expect.c+4−1

This diff is large: its 19 lines aren’t shown here.

net/netfilter/nf_conntrack_extend.c+2−0

This diff is large: its 16 lines aren’t shown here.

net/netfilter/nf_conntrack_sip.c+12−4

This diff is large: its 24 lines aren’t shown here.

net/netfilter/nf_tables_api.c+38−31

This diff is large: its 107 lines aren’t shown here.

net/netfilter/xt_connmark.c+29−20

This diff is large: its 94 lines aren’t shown here.

net/packet/af_packet.c+58−25

This diff is large: its 202 lines aren’t shown here.

net/packet/internal.h+5−5

This diff is large: its 25 lines aren’t shown here.

net/qrtr/qrtr.c+1−0

This diff is large: its 5 lines aren’t shown here.

net/rds/ib_cm.c+2−1

This diff is large: its 17 lines aren’t shown here.

net/rds/recv.c+1−0

This diff is large: its 8 lines aren’t shown here.

net/sched/act_ife.c+7−2

This diff is large: its 23 lines aren’t shown here.

net/sched/sch_fq.c+25−12

This diff is large: its 65 lines aren’t shown here.

net/sctp/inqueue.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/sctp/ipv6.c+32−29

This diff is large: its 99 lines aren’t shown here.

net/sctp/sm_statefuns.c+7−1

This diff is large: its 29 lines aren’t shown here.

net/sctp/stream.c+2−0

This diff is large: its 9 lines aren’t shown here.

net/smc/af_smc.c+33−38

This diff is large: its 145 lines aren’t shown here.

net/smc/smc_core.c+19−3

This diff is large: its 36 lines aren’t shown here.

net/smc/smc_core.h+2−1

This diff is large: its 10 lines aren’t shown here.

net/strparser/strparser.c+4−5

This diff is large: its 39 lines aren’t shown here.

net/sunrpc/rpc_pipe.c+1−0

This diff is large: its 8 lines aren’t shown here.

net/tipc/monitor.c+1−1

This diff is large: its 9 lines aren’t shown here.

net/tipc/name_table.c+21−13

This diff is large: its 103 lines aren’t shown here.

net/tipc/name_table.h+1−1

This diff is large: its 9 lines aren’t shown here.

net/tipc/net.c+2−0

This diff is large: its 9 lines aren’t shown here.

net/tipc/netlink.c+4−1

This diff is large: its 12 lines aren’t shown here.

net/tipc/node.c+5−8

This diff is large: its 31 lines aren’t shown here.

net/tipc/socket.c+3−1

This diff is large: its 18 lines aren’t shown here.

net/tipc/subscr.c+4−1

This diff is large: its 12 lines aren’t shown here.

net/tls/tls_main.c+7−0

This diff is large: its 28 lines aren’t shown here.

net/tls/tls_sw.c+9−1

This diff is large: its 31 lines aren’t shown here.

net/vmw_vsock/af_vsock.c+6−0

This diff is large: its 14 lines aren’t shown here.

samples/livepatch/livepatch-shadow-fix1.c+32−11

This diff is large: its 81 lines aren’t shown here.

samples/livepatch/livepatch-shadow-fix2.c+18−15

This diff is large: its 74 lines aren’t shown here.

samples/sockmap/Makefile+5−2

This diff is large: its 17 lines aren’t shown here.

scripts/Makefile.gcc-plugins+1−1

This diff is large: its 9 lines aren’t shown here.

scripts/Makefile.lib+1−1

This diff is large: its 9 lines aren’t shown here.

scripts/extract_xc3028.pl+1−1

This diff is large: its 8 lines aren’t shown here.

scripts/genksyms/Makefile+2−2

This diff is large: its 17 lines aren’t shown here.

scripts/mod/sumversion.c+1−8

This diff is large: its 16 lines aren’t shown here.

scripts/split-man.pl+1−1

This diff is large: its 9 lines aren’t shown here.

security/commoncap.c+2−0

This diff is large: its 9 lines aren’t shown here.

sound/core/control.c+1−1

This diff is large: its 9 lines aren’t shown here.

sound/core/pcm_compat.c+6−3

This diff is large: its 24 lines aren’t shown here.

sound/core/pcm_native.c+15−15

This diff is large: its 88 lines aren’t shown here.

sound/core/rawmidi_compat.c+12−6

This diff is large: its 64 lines aren’t shown here.

sound/core/seq/oss/seq_oss_event.c+9−6

This diff is large: its 68 lines aren’t shown here.

sound/core/seq/oss/seq_oss_midi.c+2−0

This diff is large: its 16 lines aren’t shown here.

sound/core/seq/oss/seq_oss_synth.c+49−36

This diff is large: its 191 lines aren’t shown here.

sound/core/seq/oss/seq_oss_synth.h+2−1

This diff is large: its 10 lines aren’t shown here.

sound/core/seq/seq_virmidi.c+2−2

This diff is large: its 15 lines aren’t shown here.

sound/drivers/aloop.c+15−2

This diff is large: its 84 lines aren’t shown here.

sound/drivers/opl3/opl3_synth.c+5−2

This diff is large: its 28 lines aren’t shown here.

sound/firewire/amdtp-stream.c+3−2

This diff is large: its 19 lines aren’t shown here.

sound/firewire/dice/dice-stream.c+1−1

This diff is large: its 9 lines aren’t shown here.

sound/firewire/dice/dice.c+1−1

This diff is large: its 9 lines aren’t shown here.

sound/pci/asihpi/hpimsginit.c+9−4

This diff is large: its 38 lines aren’t shown here.

sound/pci/asihpi/hpioctl.c+3−1

This diff is large: its 18 lines aren’t shown here.

sound/pci/hda/hda_hwdep.c+11−1

This diff is large: its 26 lines aren’t shown here.

sound/pci/hda/hda_intel.c+2−1

This diff is large: its 10 lines aren’t shown here.

sound/pci/hda/patch_hdmi.c+8−1

This diff is large: its 24 lines aren’t shown here.

sound/pci/hda/patch_realtek.c+9−1

This diff is large: its 47 lines aren’t shown here.

sound/pci/rme9652/hdspm.c+14−10

This diff is large: its 62 lines aren’t shown here.

sound/pci/rme9652/rme9652.c+4−2

This diff is large: its 21 lines aren’t shown here.

sound/soc/amd/acp-da7219-max98357a.c+1−1

This diff is large: its 9 lines aren’t shown here.

sound/soc/codecs/adau17x1.c+20−6

This diff is large: its 55 lines aren’t shown here.

sound/soc/codecs/adau17x1.h+2−1

This diff is large: its 10 lines aren’t shown here.

sound/soc/codecs/msm8916-wcd-analog.c+6−3

This diff is large: its 30 lines aren’t shown here.

sound/soc/codecs/rt5514.c+3−0

This diff is large: its 24 lines aren’t shown here.

sound/soc/fsl/fsl_esai.c+7−0

This diff is large: its 14 lines aren’t shown here.

sound/soc/fsl/fsl_ssi.c+11−3

This diff is large: its 35 lines aren’t shown here.

sound/soc/intel/Kconfig+13−9

This diff is large: its 38 lines aren’t shown here.

sound/soc/omap/omap-dmic.c+11−3

This diff is large: its 41 lines aren’t shown here.

sound/soc/sh/rcar/core.c+2−2

This diff is large: its 18 lines aren’t shown here.

sound/soc/soc-topology.c+9−5

This diff is large: its 51 lines aren’t shown here.

sound/usb/line6/midi.c+1−1

This diff is large: its 9 lines aren’t shown here.

sound/usb/mixer.c+4−3

This diff is large: its 28 lines aren’t shown here.

sound/usb/mixer_maps.c+3−0

This diff is large: its 12 lines aren’t shown here.

sound/usb/stream.c+1−1

This diff is large: its 9 lines aren’t shown here.

sound/usb/usx2y/us122l.c+1−1

This diff is large: its 9 lines aren’t shown here.

sound/usb/usx2y/usX2Yhwdep.c+1−1

This diff is large: its 9 lines aren’t shown here.

sound/usb/usx2y/usx2yhwdeppcm.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/arch/arm/include/uapi/asm/kvm.h+9−0

This diff is large: its 16 lines aren’t shown here.

tools/arch/x86/include/asm/required-features.h+1−7

This diff is large: its 22 lines aren’t shown here.

tools/arch/x86/include/uapi/asm/kvm.h+18−1

This diff is large: its 27 lines aren’t shown here.

tools/bpf/Makefile+2−0

This diff is large: its 9 lines aren’t shown here.

tools/bpf/bpf_dbg.c+5−2

This diff is large: its 21 lines aren’t shown here.

tools/include/linux/compiler.h+15−5

This diff is large: its 27 lines aren’t shown here.

tools/include/linux/coresight-pmu.h+1−12

This diff is large: its 20 lines aren’t shown here.

tools/include/uapi/asm-generic/mman-common.h+3−0

This diff is large: its 10 lines aren’t shown here.

tools/include/uapi/linux/bpf.h+1−0

This diff is large: its 8 lines aren’t shown here.

tools/include/uapi/linux/if_link.h+39−0

This diff is large: its 44 lines aren’t shown here.

tools/include/uapi/linux/kvm.h+20−1

This diff is large: its 54 lines aren’t shown here.

tools/include/uapi/linux/perf_event.h+15−3

This diff is large: its 27 lines aren’t shown here.

tools/include/uapi/sound/asound.h+1−0

This diff is large: its 8 lines aren’t shown here.

tools/lib/subcmd/parse-options.c+3−3

This diff is large: its 27 lines aren’t shown here.

tools/objtool/Makefile+2−2

This diff is large: its 11 lines aren’t shown here.

tools/perf/Documentation/perf-config.txt+5−0

This diff is large: its 12 lines aren’t shown here.

tools/perf/Documentation/perf-mem.txt+31−11

This diff is large: its 74 lines aren’t shown here.

tools/perf/Documentation/perf-sched.txt+2−2

This diff is large: its 11 lines aren’t shown here.

tools/perf/Documentation/perf-script.txt+9−8

This diff is large: its 24 lines aren’t shown here.

tools/perf/Documentation/perf-stat.txt+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/Makefile.config+2−2

This diff is large: its 18 lines aren’t shown here.

tools/perf/arch/arm/include/arch-tests.h+12−0

This diff is large: its 13 lines aren’t shown here.

tools/perf/arch/arm/tests/Build+2−0

This diff is large: its 5 lines aren’t shown here.

tools/perf/arch/arm/tests/arch-tests.c+16−0

This diff is large: its 17 lines aren’t shown here.

tools/perf/arch/arm/util/auxtrace.c+1−12

This diff is large: its 20 lines aren’t shown here.

tools/perf/arch/arm/util/cs-etm.c+1−12

This diff is large: its 20 lines aren’t shown here.

tools/perf/arch/arm/util/cs-etm.h+1−12

This diff is large: its 20 lines aren’t shown here.

tools/perf/arch/arm/util/pmu.c+1−12

This diff is large: its 20 lines aren’t shown here.

tools/perf/arch/s390/util/auxtrace.c+1−0

This diff is large: its 8 lines aren’t shown here.

tools/perf/arch/s390/util/header.c+0−18

This diff is large: its 22 lines aren’t shown here.

tools/perf/arch/x86/Makefile+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/arch/x86/annotate/instructions.c+66−1

This diff is large: its 114 lines aren’t shown here.

tools/perf/arch/x86/entry/syscalls/syscall_64.tbl+358−354

This diff is large: its 738 lines aren’t shown here.

tools/perf/builtin-help.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/builtin-mem.c+2−2

This diff is large: its 18 lines aren’t shown here.

tools/perf/builtin-script.c+6−3

This diff is large: its 27 lines aren’t shown here.

tools/perf/builtin-stat.c+40−14

This diff is large: its 111 lines aren’t shown here.

tools/perf/builtin-version.c+3−0

This diff is large: its 11 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/s390/mapfile.csv+5−5

This diff is large: its 12 lines aren’t shown here.

tools/perf/tests/attr/test-record-group-sampling+3−0

This diff is large: its 7 lines aren’t shown here.

tools/perf/tests/bpf-script-example.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/tests/bpf-script-test-kbuild.c+0−1

This diff is large: its 8 lines aren’t shown here.

tools/perf/tests/builtin-test.c+1−0

This diff is large: its 8 lines aren’t shown here.

tools/perf/tests/mmap-basic.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/tests/shell/record+probe_libc_inet_pton.sh+2−4

This diff is large: its 15 lines aren’t shown here.

tools/perf/trace/beauty/mmap.c+3−0

This diff is large: its 10 lines aren’t shown here.

tools/perf/ui/browsers/annotate.c+5−0

This diff is large: its 19 lines aren’t shown here.

tools/perf/ui/browsers/hists.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/util/annotate.c+21−5

This diff is large: its 66 lines aren’t shown here.

tools/perf/util/annotate.h+9−0

This diff is large: its 18 lines aren’t shown here.

tools/perf/util/cs-etm-decoder/cs-etm-decoder.c+1−2

This diff is large: its 8 lines aren’t shown here.

tools/perf/util/cs-etm.c+1−2

This diff is large: its 8 lines aren’t shown here.

tools/perf/util/cs-etm.h+1−12

This diff is large: its 20 lines aren’t shown here.

tools/perf/util/event.c+3−1

This diff is large: its 11 lines aren’t shown here.

tools/perf/util/evsel.c+16−8

This diff is large: its 60 lines aren’t shown here.

tools/perf/util/evsel.h+1−0

This diff is large: its 8 lines aren’t shown here.

tools/perf/util/generate-cmdlist.sh+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/util/header.c+2−1

This diff is large: its 10 lines aren’t shown here.

tools/perf/util/machine.c+18−12

This diff is large: its 53 lines aren’t shown here.

tools/perf/util/parse-events.y+4−4

This diff is large: its 28 lines aren’t shown here.

tools/perf/util/pmu.c+14−14

This diff is large: its 70 lines aren’t shown here.

tools/perf/util/symbol.c+3−5

This diff is large: its 20 lines aren’t shown here.

tools/perf/util/syscalltbl.c+3−3

This diff is large: its 24 lines aren’t shown here.

tools/perf/util/trace-event-scripting.c+2−2

This diff is large: its 18 lines aren’t shown here.

tools/power/acpi/Makefile.config+1−0

This diff is large: its 8 lines aren’t shown here.

tools/testing/nvdimm/test/nfit.c+65−19

This diff is large: its 164 lines aren’t shown here.

tools/testing/selftests/bpf/.gitignore+3−0

This diff is large: its 7 lines aren’t shown here.

tools/testing/selftests/bpf/test_progs.c+2−2

This diff is large: its 18 lines aren’t shown here.

tools/testing/selftests/bpf/test_sock.c+1−0

This diff is large: its 8 lines aren’t shown here.

tools/testing/selftests/bpf/test_sock_addr.c+1−0

This diff is large: its 8 lines aren’t shown here.

tools/testing/selftests/bpf/test_sock_addr.sh+2−2

This diff is large: its 18 lines aren’t shown here.

tools/testing/selftests/filesystems/Makefile+3−5

This diff is large: its 12 lines aren’t shown here.

tools/testing/selftests/firmware/Makefile+1−0

This diff is large: its 8 lines aren’t shown here.

tools/testing/selftests/firmware/fw_lib.sh+6−4

This diff is large: its 18 lines aren’t shown here.

tools/testing/selftests/firmware/fw_run_tests.sh+1−1

This diff is large: its 7 lines aren’t shown here.

tools/testing/selftests/ftrace/test.d/trigger/inter-event/trigger-extended-error-support.tc+1−1

This diff is large: its 9 lines aren’t shown here.

tools/testing/selftests/ftrace/test.d/trigger/inter-event/trigger-multi-actions-accept.tc+44−0

This diff is large: its 45 lines aren’t shown here.

tools/testing/selftests/kvm/Makefile+3−2

This diff is large: its 21 lines aren’t shown here.

tools/testing/selftests/kvm/include/kvm_util.h+9−6

This diff is large: its 34 lines aren’t shown here.

tools/testing/selftests/kvm/include/vmx.h+494−0

This diff is large: its 495 lines aren’t shown here.

tools/testing/selftests/kvm/lib/kvm_util.c+13−7

This diff is large: its 67 lines aren’t shown here.

tools/testing/selftests/kvm/lib/sparsebit.c+2−2

This diff is large: its 18 lines aren’t shown here.

tools/testing/selftests/kvm/lib/vmx.c+243−0

This diff is large: its 244 lines aren’t shown here.

tools/testing/selftests/kvm/vmx_tsc_adjust_test.c+231−0

This diff is large: its 232 lines aren’t shown here.

tools/testing/selftests/lib.mk+4−4

This diff is large: its 15 lines aren’t shown here.

tools/testing/selftests/net/Makefile+1−0

This diff is large: its 8 lines aren’t shown here.

tools/testing/selftests/x86/test_syscall_vdso.c+21−14

This diff is large: its 66 lines aren’t shown here.

virt/kvm/arm/arm.c+10−5

This diff is large: its 46 lines aren’t shown here.

virt/kvm/arm/psci.c+60−0

This diff is large: its 71 lines aren’t shown here.

virt/kvm/arm/vgic/vgic-init.c+1−1

This diff is large: its 9 lines aren’t shown here.

virt/kvm/arm/vgic/vgic-mmio-v2.c+5−0

This diff is large: its 19 lines aren’t shown here.

virt/kvm/arm/vgic/vgic-mmio.c+8−2

This diff is large: its 19 lines aren’t shown here.

virt/kvm/arm/vgic/vgic-v2.c+22−16

This diff is large: its 86 lines aren’t shown here.

virt/kvm/arm/vgic/vgic-v3.c+29−20

This diff is large: its 100 lines aren’t shown here.

virt/kvm/arm/vgic/vgic.c+25−27

This diff is large: its 131 lines aren’t shown here.

virt/kvm/arm/vgic/vgic.h+14−0

This diff is large: its 21 lines aren’t shown here.