aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorTakashi Iwai <tiwai@suse.de>2018-01-19 07:57:33 -0500
committerTakashi Iwai <tiwai@suse.de>2018-01-19 07:57:33 -0500
commitc86d95cb6b7ecda7b7d56e40c24b7d8b9bf9159a (patch)
tree3d868496a15be09dfb3a471e7b9dce24c0216d89
parent388fdb8f882af67ff8394d9420c1e0e42ba35619 (diff)
parent8f05b9c65d817dd288da5f1c6ecc134b66f8a190 (diff)
Merge tag 'asoc-v4.16-3' of https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound into for-next
ASoC: Updates for v4.16 Some final updates for the merge window, this brings in some improvements to the ACPI GPIO handling for Intel and a bunch of fixes.
-rw-r--r--Documentation/ABI/testing/sysfs-devices-system-cpu16
-rw-r--r--Documentation/admin-guide/kernel-parameters.txt58
-rw-r--r--Documentation/devicetree/bindings/sound/mxs-audio-sgtl5000.txt33
-rw-r--r--Documentation/filesystems/nilfs2.txt4
-rw-r--r--Documentation/kbuild/kconfig-language.txt23
-rw-r--r--Documentation/networking/index.rst2
-rw-r--r--Documentation/networking/msg_zerocopy.rst4
-rw-r--r--Documentation/usb/gadget-testing.txt2
-rw-r--r--Documentation/x86/pti.txt186
-rw-r--r--MAINTAINERS6
-rw-r--r--Makefile45
-rw-r--r--arch/ia64/kernel/time.c2
-rw-r--r--arch/mips/kernel/cps-vec.S2
-rw-r--r--arch/mips/kernel/process.c12
-rw-r--r--arch/mips/kernel/ptrace.c147
-rw-r--r--arch/powerpc/include/asm/exception-64e.h6
-rw-r--r--arch/powerpc/include/asm/exception-64s.h57
-rw-r--r--arch/powerpc/include/asm/feature-fixups.h13
-rw-r--r--arch/powerpc/include/asm/hvcall.h17
-rw-r--r--arch/powerpc/include/asm/paca.h10
-rw-r--r--arch/powerpc/include/asm/plpar_wrappers.h14
-rw-r--r--arch/powerpc/include/asm/setup.h13
-rw-r--r--arch/powerpc/kernel/asm-offsets.c5
-rw-r--r--arch/powerpc/kernel/entry_64.S44
-rw-r--r--arch/powerpc/kernel/exceptions-64s.S137
-rw-r--r--arch/powerpc/kernel/setup_64.c101
-rw-r--r--arch/powerpc/kernel/vmlinux.lds.S9
-rw-r--r--arch/powerpc/kvm/book3s_64_mmu.c1
-rw-r--r--arch/powerpc/kvm/book3s_64_mmu_hv.c90
-rw-r--r--arch/powerpc/kvm/book3s_hv_rmhandlers.S9
-rw-r--r--arch/powerpc/kvm/book3s_pr.c2
-rw-r--r--arch/powerpc/kvm/book3s_rmhandlers.S7
-rw-r--r--arch/powerpc/kvm/book3s_segment.S4
-rw-r--r--arch/powerpc/lib/feature-fixups.c41
-rw-r--r--arch/powerpc/platforms/powernv/setup.c49
-rw-r--r--arch/powerpc/platforms/pseries/dlpar.c21
-rw-r--r--arch/powerpc/platforms/pseries/pseries.h2
-rw-r--r--arch/powerpc/platforms/pseries/ras.c3
-rw-r--r--arch/powerpc/platforms/pseries/setup.c35
-rw-r--r--arch/riscv/configs/defconfig75
-rw-r--r--arch/riscv/include/asm/csr.h8
-rw-r--r--arch/riscv/include/asm/io.h4
-rw-r--r--arch/riscv/include/asm/irqflags.h10
-rw-r--r--arch/riscv/include/asm/pgtable.h4
-rw-r--r--arch/riscv/include/asm/ptrace.h2
-rw-r--r--arch/riscv/include/asm/tlbflush.h4
-rw-r--r--arch/riscv/include/asm/uaccess.h12
-rw-r--r--arch/riscv/include/asm/unistd.h1
-rw-r--r--arch/riscv/include/asm/vdso-syscalls.h28
-rw-r--r--arch/riscv/include/uapi/asm/syscalls.h26
-rw-r--r--arch/riscv/kernel/entry.S8
-rw-r--r--arch/riscv/kernel/process.c4
-rw-r--r--arch/riscv/kernel/syscall_table.c1
-rw-r--r--arch/riscv/kernel/vdso/flush_icache.S1
-rw-r--r--arch/riscv/mm/fault.c2
-rw-r--r--arch/sh/boards/mach-se/770x/setup.c24
-rw-r--r--arch/sh/include/mach-se/mach/se.h1
-rw-r--r--arch/x86/Kconfig15
-rw-r--r--arch/x86/Makefile8
-rw-r--r--arch/x86/crypto/aesni-intel_asm.S5
-rw-r--r--arch/x86/crypto/camellia-aesni-avx-asm_64.S3
-rw-r--r--arch/x86/crypto/camellia-aesni-avx2-asm_64.S3
-rw-r--r--arch/x86/crypto/crc32c-pcl-intel-asm_64.S3
-rw-r--r--arch/x86/entry/calling.h36
-rw-r--r--arch/x86/entry/entry_32.S5
-rw-r--r--arch/x86/entry/entry_64.S12
-rw-r--r--arch/x86/events/intel/bts.c18
-rw-r--r--arch/x86/include/asm/asm-prototypes.h25
-rw-r--r--arch/x86/include/asm/cpufeatures.h4
-rw-r--r--arch/x86/include/asm/mshyperv.h18
-rw-r--r--arch/x86/include/asm/msr-index.h3
-rw-r--r--arch/x86/include/asm/nospec-branch.h214
-rw-r--r--arch/x86/include/asm/pci_x86.h1
-rw-r--r--arch/x86/include/asm/processor-flags.h2
-rw-r--r--arch/x86/include/asm/tlbflush.h6
-rw-r--r--arch/x86/include/asm/xen/hypercall.h5
-rw-r--r--arch/x86/kernel/alternative.c7
-rw-r--r--arch/x86/kernel/cpu/amd.c28
-rw-r--r--arch/x86/kernel/cpu/bugs.c185
-rw-r--r--arch/x86/kernel/cpu/common.c3
-rw-r--r--arch/x86/kernel/cpu/microcode/intel.c13
-rw-r--r--arch/x86/kernel/ftrace_32.S6
-rw-r--r--arch/x86/kernel/ftrace_64.S8
-rw-r--r--arch/x86/kernel/irq_32.c9
-rw-r--r--arch/x86/kernel/tboot.c11
-rw-r--r--arch/x86/kvm/mmu.c19
-rw-r--r--arch/x86/kvm/svm.c13
-rw-r--r--arch/x86/kvm/vmx.c21
-rw-r--r--arch/x86/lib/Makefile1
-rw-r--r--arch/x86/lib/checksum_32.S7
-rw-r--r--arch/x86/lib/retpoline.S48
-rw-r--r--arch/x86/mm/pti.c32
-rw-r--r--arch/x86/pci/common.c5
-rw-r--r--arch/x86/pci/fixup.c29
-rw-r--r--arch/x86/platform/efi/efi_64.c2
-rw-r--r--arch/x86/platform/intel-mid/device_libs/platform_bt.c2
-rw-r--r--arch/x86/xen/mmu_pv.c8
-rw-r--r--arch/x86/xen/xen-ops.h2
-rw-r--r--block/blk-core.c9
-rw-r--r--block/blk-mq.c2
-rw-r--r--block/blk.h2
-rw-r--r--crypto/algapi.c12
-rw-r--r--drivers/acpi/utils.c41
-rw-r--r--drivers/base/Kconfig3
-rw-r--r--drivers/base/cpu.c48
-rw-r--r--drivers/block/loop.c10
-rw-r--r--drivers/block/rbd.c18
-rw-r--r--drivers/gpio/gpio-merrifield.c11
-rw-r--r--drivers/gpio/gpiolib.c35
-rw-r--r--drivers/gpu/drm/i915/gvt/cmd_parser.c4
-rw-r--r--drivers/gpu/drm/i915/gvt/gtt.c5
-rw-r--r--drivers/gpu/drm/i915/i915_gem.c2
-rw-r--r--drivers/gpu/drm/i915/i915_reg.h2
-rw-r--r--drivers/gpu/drm/i915/intel_engine_cs.c5
-rw-r--r--drivers/gpu/drm/i915/intel_lrc.c3
-rw-r--r--drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgf119.c1
-rw-r--r--drivers/gpu/drm/tegra/sor.c3
-rw-r--r--drivers/gpu/drm/vc4/vc4_irq.c3
-rw-r--r--drivers/gpu/drm/vc4/vc4_v3d.c3
-rw-r--r--drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c2
-rw-r--r--drivers/gpu/drm/vmwgfx/vmwgfx_kms.c6
-rw-r--r--drivers/gpu/drm/vmwgfx/vmwgfx_kms.h2
-rw-r--r--drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c41
-rw-r--r--drivers/iio/adc/stm32-dfsdm-adc.c13
-rw-r--r--drivers/iio/adc/stm32-dfsdm-core.c8
-rw-r--r--drivers/infiniband/core/core_priv.h2
-rw-r--r--drivers/infiniband/core/device.c18
-rw-r--r--drivers/infiniband/core/nldev.c54
-rw-r--r--drivers/infiniband/hw/mlx4/mr.c2
-rw-r--r--drivers/infiniband/ulp/ipoib/ipoib_main.c25
-rw-r--r--drivers/infiniband/ulp/ipoib/ipoib_multicast.c5
-rw-r--r--drivers/infiniband/ulp/srpt/ib_srpt.c5
-rw-r--r--drivers/mmc/host/renesas_sdhi_core.c3
-rw-r--r--drivers/mmc/host/s3cmci.c6
-rw-r--r--drivers/mux/core.c4
-rw-r--r--drivers/net/can/flexcan.c2
-rw-r--r--drivers/net/can/usb/ems_usb.c1
-rw-r--r--drivers/net/can/usb/gs_usb.c2
-rw-r--r--drivers/net/can/vxcan.c2
-rw-r--r--drivers/net/dsa/b53/b53_common.c9
-rw-r--r--drivers/net/ethernet/3com/3c59x.c90
-rw-r--r--drivers/net/ethernet/amazon/ena/ena_netdev.c45
-rw-r--r--drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c2
-rw-r--r--drivers/net/ethernet/broadcom/bnxt/bnxt_tc.c2
-rw-r--r--drivers/net/ethernet/chelsio/cxgb4/cxgb4.h1
-rw-r--r--drivers/net/ethernet/chelsio/cxgb4/t4_hw.c17
-rw-r--r--drivers/net/ethernet/freescale/fec_main.c7
-rw-r--r--drivers/net/ethernet/freescale/gianfar_ptp.c3
-rw-r--r--drivers/net/ethernet/intel/e1000/e1000.h3
-rw-r--r--drivers/net/ethernet/intel/e1000/e1000_main.c27
-rw-r--r--drivers/net/ethernet/intel/e1000e/ich8lan.c11
-rw-r--r--drivers/net/ethernet/intel/i40e/i40e_main.c37
-rw-r--r--drivers/net/ethernet/intel/i40e/i40e_txrx.c26
-rw-r--r--drivers/net/ethernet/intel/i40evf/i40e_txrx.c26
-rw-r--r--drivers/net/ethernet/mellanox/mlxsw/pci.c7
-rw-r--r--drivers/net/ethernet/mellanox/mlxsw/pci_hw.h1
-rw-r--r--drivers/net/ethernet/mellanox/mlxsw/spectrum.c11
-rw-r--r--drivers/net/ethernet/mellanox/mlxsw/spectrum.h2
-rw-r--r--drivers/net/ethernet/mellanox/mlxsw/spectrum_qdisc.c7
-rw-r--r--drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c2
-rw-r--r--drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c6
-rw-r--r--drivers/net/ethernet/netronome/nfp/nfp_net_common.c2
-rw-r--r--drivers/net/ethernet/renesas/sh_eth.c33
-rw-r--r--drivers/net/ethernet/stmicro/stmmac/stmmac_main.c6
-rw-r--r--drivers/net/geneve.c14
-rw-r--r--drivers/net/macvlan.c7
-rw-r--r--drivers/net/phy/mdio-sun4i.c6
-rw-r--r--drivers/net/phy/phylink.c5
-rw-r--r--drivers/net/phy/sfp-bus.c6
-rw-r--r--drivers/net/usb/qmi_wwan.c1
-rw-r--r--drivers/net/wireless/ath/wcn36xx/main.c23
-rw-r--r--drivers/net/wireless/ath/wcn36xx/pmc.c6
-rw-r--r--drivers/net/wireless/intel/iwlwifi/pcie/internal.h10
-rw-r--r--drivers/net/wireless/intel/iwlwifi/pcie/tx-gen2.c11
-rw-r--r--drivers/net/wireless/intel/iwlwifi/pcie/tx.c8
-rw-r--r--drivers/net/xen-netfront.c1
-rw-r--r--drivers/nvme/host/core.c7
-rw-r--r--drivers/nvme/host/fabrics.c1
-rw-r--r--drivers/nvme/host/nvme.h12
-rw-r--r--drivers/nvme/host/pci.c42
-rw-r--r--drivers/nvme/host/rdma.c14
-rw-r--r--drivers/nvme/target/fcloop.c2
-rw-r--r--drivers/of/of_mdio.c9
-rw-r--r--drivers/platform/x86/wmi.c2
-rw-r--r--drivers/staging/android/ashmem.c2
-rw-r--r--drivers/usb/gadget/udc/core.c28
-rw-r--r--drivers/usb/misc/usb3503.c2
-rw-r--r--drivers/usb/mon/mon_bin.c8
-rw-r--r--drivers/usb/serial/cp210x.c2
-rw-r--r--drivers/usb/storage/unusual_uas.h7
-rw-r--r--drivers/usb/usbip/usbip_common.c17
-rw-r--r--drivers/usb/usbip/vudc_rx.c19
-rw-r--r--drivers/usb/usbip/vudc_tx.c11
-rw-r--r--drivers/xen/gntdev.c8
-rw-r--r--include/acpi/acpi_bus.h3
-rw-r--r--include/linux/acpi.h6
-rw-r--r--include/linux/bpf.h26
-rw-r--r--include/linux/completion.h1
-rw-r--r--include/linux/cpu.h7
-rw-r--r--include/linux/crash_core.h2
-rw-r--r--include/linux/irqflags.h4
-rw-r--r--include/linux/lockdep.h2
-rw-r--r--include/linux/sh_eth.h1
-rw-r--r--include/net/sctp/structs.h2
-rw-r--r--include/net/vxlan.h2
-rw-r--r--include/sound/soc-acpi.h10
-rw-r--r--include/sound/soc.h3
-rw-r--r--include/uapi/linux/if_ether.h3
-rw-r--r--include/uapi/linux/libc-compat.h61
-rw-r--r--include/uapi/linux/netfilter/nf_conntrack_common.h2
-rw-r--r--init/Kconfig8
-rw-r--r--kernel/bpf/arraymap.c47
-rw-r--r--kernel/bpf/core.c19
-rw-r--r--kernel/bpf/sockmap.c11
-rw-r--r--kernel/bpf/verifier.c36
-rw-r--r--kernel/cgroup/cgroup-v1.c6
-rw-r--r--kernel/cgroup/cgroup.c20
-rw-r--r--kernel/crash_core.c2
-rw-r--r--kernel/sched/completion.c5
-rw-r--r--kernel/sched/membarrier.c2
-rw-r--r--lib/test_bpf.c11
-rw-r--r--mm/kmemleak.c2
-rw-r--r--net/8021q/vlan.c7
-rw-r--r--net/bluetooth/l2cap_core.c20
-rw-r--r--net/caif/caif_dev.c5
-rw-r--r--net/caif/caif_usb.c4
-rw-r--r--net/caif/cfcnfg.c10
-rw-r--r--net/caif/cfctrl.c4
-rw-r--r--net/core/dev.c14
-rw-r--r--net/core/ethtool.c15
-rw-r--r--net/core/filter.c6
-rw-r--r--net/core/rtnetlink.c10
-rw-r--r--net/core/sock_diag.c2
-rw-r--r--net/core/sysctl_net_core.c6
-rw-r--r--net/ipv4/raw.c4
-rw-r--r--net/ipv6/exthdrs.c9
-rw-r--r--net/ipv6/ip6_fib.c73
-rw-r--r--net/ipv6/ip6_output.c5
-rw-r--r--net/ipv6/ip6_tunnel.c15
-rw-r--r--net/mac80211/rx.c2
-rw-r--r--net/netfilter/nf_tables_api.c8
-rw-r--r--net/rds/rdma.c4
-rw-r--r--net/sched/act_gact.c2
-rw-r--r--net/sched/act_mirred.c2
-rw-r--r--net/sctp/input.c28
-rw-r--r--net/sctp/socket.c28
-rw-r--r--net/sctp/stream.c22
-rw-r--r--net/sctp/transport.c29
-rw-r--r--net/socket.c13
-rw-r--r--net/tipc/group.c22
-rw-r--r--net/wireless/nl80211.c3
-rw-r--r--scripts/genksyms/.gitignore1
-rw-r--r--scripts/kconfig/expr.c5
-rw-r--r--security/Kconfig2
-rw-r--r--security/apparmor/domain.c9
-rw-r--r--security/apparmor/include/perms.h3
-rw-r--r--security/apparmor/ipc.c53
-rw-r--r--sound/soc/au1x/ac97c.c6
-rw-r--r--sound/soc/bcm/bcm2835-i2s.c20
-rw-r--r--sound/soc/codecs/cs42l73.c2
-rw-r--r--sound/soc/codecs/cx20442.c26
-rw-r--r--sound/soc/codecs/sgtl5000.c5
-rw-r--r--sound/soc/codecs/tlv320dac33.c15
-rw-r--r--sound/soc/codecs/uda1380.c2
-rw-r--r--sound/soc/intel/Kconfig1
-rw-r--r--sound/soc/intel/boards/Kconfig1
-rw-r--r--sound/soc/intel/boards/bytcht_da7213.c4
-rw-r--r--sound/soc/intel/boards/bytcht_es8316.c26
-rw-r--r--sound/soc/intel/boards/bytcr_rt5640.c4
-rw-r--r--sound/soc/intel/boards/bytcr_rt5651.c4
-rw-r--r--sound/soc/intel/boards/cht_bsw_rt5645.c6
-rw-r--r--sound/soc/intel/boards/cht_bsw_rt5672.c4
-rw-r--r--sound/soc/intel/skylake/skl-topology.c1
-rw-r--r--sound/soc/mxs/mxs-sgtl5000.c25
-rw-r--r--sound/soc/soc-acpi.c33
-rw-r--r--sound/soc/soc-core.c112
-rw-r--r--sound/soc/soc-io.c6
-rw-r--r--tools/objtool/Makefile2
-rw-r--r--tools/objtool/check.c69
-rw-r--r--tools/objtool/check.h2
-rw-r--r--tools/testing/selftests/bpf/test_align.c22
-rw-r--r--tools/testing/selftests/x86/Makefile2
-rw-r--r--tools/testing/selftests/x86/test_vsyscall.c500
283 files changed, 3941 insertions, 1099 deletions
diff --git a/Documentation/ABI/testing/sysfs-devices-system-cpu b/Documentation/ABI/testing/sysfs-devices-system-cpu
index d6d862db3b5d..bfd29bc8d37a 100644
--- a/Documentation/ABI/testing/sysfs-devices-system-cpu
+++ b/Documentation/ABI/testing/sysfs-devices-system-cpu
@@ -375,3 +375,19 @@ Contact: Linux kernel mailing list <linux-kernel@vger.kernel.org>
375Description: information about CPUs heterogeneity. 375Description: information about CPUs heterogeneity.
376 376
377 cpu_capacity: capacity of cpu#. 377 cpu_capacity: capacity of cpu#.
378
379What: /sys/devices/system/cpu/vulnerabilities
380 /sys/devices/system/cpu/vulnerabilities/meltdown
381 /sys/devices/system/cpu/vulnerabilities/spectre_v1
382 /sys/devices/system/cpu/vulnerabilities/spectre_v2
383Date: January 2018
384Contact: Linux kernel mailing list <linux-kernel@vger.kernel.org>
385Description: Information about CPU vulnerabilities
386
387 The files are named after the code names of CPU
388 vulnerabilities. The output of those files reflects the
389 state of the CPUs in the system. Possible output values:
390
391 "Not affected" CPU is not affected by the vulnerability
392 "Vulnerable" CPU is affected and no mitigation in effect
393 "Mitigation: $M" CPU is affected and mitigation $M is in effect
diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt
index af7104aaffd9..46b26bfee27b 100644
--- a/Documentation/admin-guide/kernel-parameters.txt
+++ b/Documentation/admin-guide/kernel-parameters.txt
@@ -713,9 +713,6 @@
713 It will be ignored when crashkernel=X,high is not used 713 It will be ignored when crashkernel=X,high is not used
714 or memory reserved is below 4G. 714 or memory reserved is below 4G.
715 715
716 crossrelease_fullstack
717 [KNL] Allow to record full stack trace in cross-release
718
719 cryptomgr.notests 716 cryptomgr.notests
720 [KNL] Disable crypto self-tests 717 [KNL] Disable crypto self-tests
721 718
@@ -2626,6 +2623,11 @@
2626 nosmt [KNL,S390] Disable symmetric multithreading (SMT). 2623 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
2627 Equivalent to smt=1. 2624 Equivalent to smt=1.
2628 2625
2626 nospectre_v2 [X86] Disable all mitigations for the Spectre variant 2
2627 (indirect branch prediction) vulnerability. System may
2628 allow data leaks with this option, which is equivalent
2629 to spectre_v2=off.
2630
2629 noxsave [BUGS=X86] Disables x86 extended register state save 2631 noxsave [BUGS=X86] Disables x86 extended register state save
2630 and restore using xsave. The kernel will fallback to 2632 and restore using xsave. The kernel will fallback to
2631 enabling legacy floating-point and sse state. 2633 enabling legacy floating-point and sse state.
@@ -2712,8 +2714,6 @@
2712 steal time is computed, but won't influence scheduler 2714 steal time is computed, but won't influence scheduler
2713 behaviour 2715 behaviour
2714 2716
2715 nopti [X86-64] Disable kernel page table isolation
2716
2717 nolapic [X86-32,APIC] Do not enable or use the local APIC. 2717 nolapic [X86-32,APIC] Do not enable or use the local APIC.
2718 2718
2719 nolapic_timer [X86-32,APIC] Do not use the local APIC timer. 2719 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
@@ -3100,6 +3100,12 @@
3100 pcie_scan_all Scan all possible PCIe devices. Otherwise we 3100 pcie_scan_all Scan all possible PCIe devices. Otherwise we
3101 only look for one device below a PCIe downstream 3101 only look for one device below a PCIe downstream
3102 port. 3102 port.
3103 big_root_window Try to add a big 64bit memory window to the PCIe
3104 root complex on AMD CPUs. Some GFX hardware
3105 can resize a BAR to allow access to all VRAM.
3106 Adding the window is slightly risky (it may
3107 conflict with unreported devices), so this
3108 taints the kernel.
3103 3109
3104 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power 3110 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
3105 Management. 3111 Management.
@@ -3288,11 +3294,20 @@
3288 pt. [PARIDE] 3294 pt. [PARIDE]
3289 See Documentation/blockdev/paride.txt. 3295 See Documentation/blockdev/paride.txt.
3290 3296
3291 pti= [X86_64] 3297 pti= [X86_64] Control Page Table Isolation of user and
3292 Control user/kernel address space isolation: 3298 kernel address spaces. Disabling this feature
3293 on - enable 3299 removes hardening, but improves performance of
3294 off - disable 3300 system calls and interrupts.
3295 auto - default setting 3301
3302 on - unconditionally enable
3303 off - unconditionally disable
3304 auto - kernel detects whether your CPU model is
3305 vulnerable to issues that PTI mitigates
3306
3307 Not specifying this option is equivalent to pti=auto.
3308
3309 nopti [X86_64]
3310 Equivalent to pti=off
3296 3311
3297 pty.legacy_count= 3312 pty.legacy_count=
3298 [KNL] Number of legacy pty's. Overwrites compiled-in 3313 [KNL] Number of legacy pty's. Overwrites compiled-in
@@ -3943,6 +3958,29 @@
3943 sonypi.*= [HW] Sony Programmable I/O Control Device driver 3958 sonypi.*= [HW] Sony Programmable I/O Control Device driver
3944 See Documentation/laptops/sonypi.txt 3959 See Documentation/laptops/sonypi.txt
3945 3960
3961 spectre_v2= [X86] Control mitigation of Spectre variant 2
3962 (indirect branch speculation) vulnerability.
3963
3964 on - unconditionally enable
3965 off - unconditionally disable
3966 auto - kernel detects whether your CPU model is
3967 vulnerable
3968
3969 Selecting 'on' will, and 'auto' may, choose a
3970 mitigation method at run time according to the
3971 CPU, the available microcode, the setting of the
3972 CONFIG_RETPOLINE configuration option, and the
3973 compiler with which the kernel was built.
3974
3975 Specific mitigations can also be selected manually:
3976
3977 retpoline - replace indirect branches
3978 retpoline,generic - google's original retpoline
3979 retpoline,amd - AMD-specific minimal thunk
3980
3981 Not specifying this option is equivalent to
3982 spectre_v2=auto.
3983
3946 spia_io_base= [HW,MTD] 3984 spia_io_base= [HW,MTD]
3947 spia_fio_base= 3985 spia_fio_base=
3948 spia_pedr= 3986 spia_pedr=
diff --git a/Documentation/devicetree/bindings/sound/mxs-audio-sgtl5000.txt b/Documentation/devicetree/bindings/sound/mxs-audio-sgtl5000.txt
index 601c518eddaa..4eb980bd0287 100644
--- a/Documentation/devicetree/bindings/sound/mxs-audio-sgtl5000.txt
+++ b/Documentation/devicetree/bindings/sound/mxs-audio-sgtl5000.txt
@@ -1,10 +1,31 @@
1* Freescale MXS audio complex with SGTL5000 codec 1* Freescale MXS audio complex with SGTL5000 codec
2 2
3Required properties: 3Required properties:
4- compatible: "fsl,mxs-audio-sgtl5000" 4- compatible : "fsl,mxs-audio-sgtl5000"
5- model: The user-visible name of this sound complex 5- model : The user-visible name of this sound complex
6- saif-controllers: The phandle list of the MXS SAIF controller 6- saif-controllers : The phandle list of the MXS SAIF controller
7- audio-codec: The phandle of the SGTL5000 audio codec 7- audio-codec : The phandle of the SGTL5000 audio codec
8- audio-routing : A list of the connections between audio components.
9 Each entry is a pair of strings, the first being the
10 connection's sink, the second being the connection's
11 source. Valid names could be power supplies, SGTL5000
12 pins, and the jacks on the board:
13
14 Power supplies:
15 * Mic Bias
16
17 SGTL5000 pins:
18 * MIC_IN
19 * LINE_IN
20 * HP_OUT
21 * LINE_OUT
22
23 Board connectors:
24 * Mic Jack
25 * Line In Jack
26 * Headphone Jack
27 * Line Out Jack
28 * Ext Spk
8 29
9Example: 30Example:
10 31
@@ -14,4 +35,8 @@ sound {
14 model = "imx28-evk-sgtl5000"; 35 model = "imx28-evk-sgtl5000";
15 saif-controllers = <&saif0 &saif1>; 36 saif-controllers = <&saif0 &saif1>;
16 audio-codec = <&sgtl5000>; 37 audio-codec = <&sgtl5000>;
38 audio-routing =
39 "MIC_IN", "Mic Jack",
40 "Mic Jack", "Mic Bias",
41 "Headphone Jack", "HP_OUT";
17}; 42};
diff --git a/Documentation/filesystems/nilfs2.txt b/Documentation/filesystems/nilfs2.txt
index c0727dc36271..f2f3f8592a6f 100644
--- a/Documentation/filesystems/nilfs2.txt
+++ b/Documentation/filesystems/nilfs2.txt
@@ -25,8 +25,8 @@ available from the following download page. At least "mkfs.nilfs2",
25cleaner or garbage collector) are required. Details on the tools are 25cleaner or garbage collector) are required. Details on the tools are
26described in the man pages included in the package. 26described in the man pages included in the package.
27 27
28Project web page: http://nilfs.sourceforge.net/ 28Project web page: https://nilfs.sourceforge.io/
29Download page: http://nilfs.sourceforge.net/en/download.html 29Download page: https://nilfs.sourceforge.io/en/download.html
30List info: http://vger.kernel.org/vger-lists.html#linux-nilfs 30List info: http://vger.kernel.org/vger-lists.html#linux-nilfs
31 31
32Caveats 32Caveats
diff --git a/Documentation/kbuild/kconfig-language.txt b/Documentation/kbuild/kconfig-language.txt
index 262722d8867b..c4a293a03c33 100644
--- a/Documentation/kbuild/kconfig-language.txt
+++ b/Documentation/kbuild/kconfig-language.txt
@@ -200,10 +200,14 @@ module state. Dependency expressions have the following syntax:
200<expr> ::= <symbol> (1) 200<expr> ::= <symbol> (1)
201 <symbol> '=' <symbol> (2) 201 <symbol> '=' <symbol> (2)
202 <symbol> '!=' <symbol> (3) 202 <symbol> '!=' <symbol> (3)
203 '(' <expr> ')' (4) 203 <symbol1> '<' <symbol2> (4)
204 '!' <expr> (5) 204 <symbol1> '>' <symbol2> (4)
205 <expr> '&&' <expr> (6) 205 <symbol1> '<=' <symbol2> (4)
206 <expr> '||' <expr> (7) 206 <symbol1> '>=' <symbol2> (4)
207 '(' <expr> ')' (5)
208 '!' <expr> (6)
209 <expr> '&&' <expr> (7)
210 <expr> '||' <expr> (8)
207 211
208Expressions are listed in decreasing order of precedence. 212Expressions are listed in decreasing order of precedence.
209 213
@@ -214,10 +218,13 @@ Expressions are listed in decreasing order of precedence.
214 otherwise 'n'. 218 otherwise 'n'.
215(3) If the values of both symbols are equal, it returns 'n', 219(3) If the values of both symbols are equal, it returns 'n',
216 otherwise 'y'. 220 otherwise 'y'.
217(4) Returns the value of the expression. Used to override precedence. 221(4) If value of <symbol1> is respectively lower, greater, lower-or-equal,
218(5) Returns the result of (2-/expr/). 222 or greater-or-equal than value of <symbol2>, it returns 'y',
219(6) Returns the result of min(/expr/, /expr/). 223 otherwise 'n'.
220(7) Returns the result of max(/expr/, /expr/). 224(5) Returns the value of the expression. Used to override precedence.
225(6) Returns the result of (2-/expr/).
226(7) Returns the result of min(/expr/, /expr/).
227(8) Returns the result of max(/expr/, /expr/).
221 228
222An expression can have a value of 'n', 'm' or 'y' (or 0, 1, 2 229An expression can have a value of 'n', 'm' or 'y' (or 0, 1, 2
223respectively for calculations). A menu entry becomes visible when its 230respectively for calculations). A menu entry becomes visible when its
diff --git a/Documentation/networking/index.rst b/Documentation/networking/index.rst
index 66e620866245..7d4b15977d61 100644
--- a/Documentation/networking/index.rst
+++ b/Documentation/networking/index.rst
@@ -9,6 +9,7 @@ Contents:
9 batman-adv 9 batman-adv
10 kapi 10 kapi
11 z8530book 11 z8530book
12 msg_zerocopy
12 13
13.. only:: subproject 14.. only:: subproject
14 15
@@ -16,4 +17,3 @@ Contents:
16 ======= 17 =======
17 18
18 * :ref:`genindex` 19 * :ref:`genindex`
19
diff --git a/Documentation/networking/msg_zerocopy.rst b/Documentation/networking/msg_zerocopy.rst
index 77f6d7e25cfd..291a01264967 100644
--- a/Documentation/networking/msg_zerocopy.rst
+++ b/Documentation/networking/msg_zerocopy.rst
@@ -72,6 +72,10 @@ this flag, a process must first signal intent by setting a socket option:
72 if (setsockopt(fd, SOL_SOCKET, SO_ZEROCOPY, &one, sizeof(one))) 72 if (setsockopt(fd, SOL_SOCKET, SO_ZEROCOPY, &one, sizeof(one)))
73 error(1, errno, "setsockopt zerocopy"); 73 error(1, errno, "setsockopt zerocopy");
74 74
75Setting the socket option only works when the socket is in its initial
76(TCP_CLOSED) state. Trying to set the option for a socket returned by accept(),
77for example, will lead to an EBUSY error. In this case, the option should be set
78to the listening socket and it will be inherited by the accepted sockets.
75 79
76Transmission 80Transmission
77------------ 81------------
diff --git a/Documentation/usb/gadget-testing.txt b/Documentation/usb/gadget-testing.txt
index 441a4b9b666f..5908a21fddb6 100644
--- a/Documentation/usb/gadget-testing.txt
+++ b/Documentation/usb/gadget-testing.txt
@@ -693,7 +693,7 @@ such specification consists of a number of lines with an inverval value
693in each line. The rules stated above are best illustrated with an example: 693in each line. The rules stated above are best illustrated with an example:
694 694
695# mkdir functions/uvc.usb0/control/header/h 695# mkdir functions/uvc.usb0/control/header/h
696# cd functions/uvc.usb0/control/header/h 696# cd functions/uvc.usb0/control/
697# ln -s header/h class/fs 697# ln -s header/h class/fs
698# ln -s header/h class/ss 698# ln -s header/h class/ss
699# mkdir -p functions/uvc.usb0/streaming/uncompressed/u/360p 699# mkdir -p functions/uvc.usb0/streaming/uncompressed/u/360p
diff --git a/Documentation/x86/pti.txt b/Documentation/x86/pti.txt
new file mode 100644
index 000000000000..d11eff61fc9a
--- /dev/null
+++ b/Documentation/x86/pti.txt
@@ -0,0 +1,186 @@
1Overview
2========
3
4Page Table Isolation (pti, previously known as KAISER[1]) is a
5countermeasure against attacks on the shared user/kernel address
6space such as the "Meltdown" approach[2].
7
8To mitigate this class of attacks, we create an independent set of
9page tables for use only when running userspace applications. When
10the kernel is entered via syscalls, interrupts or exceptions, the
11page tables are switched to the full "kernel" copy. When the system
12switches back to user mode, the user copy is used again.
13
14The userspace page tables contain only a minimal amount of kernel
15data: only what is needed to enter/exit the kernel such as the
16entry/exit functions themselves and the interrupt descriptor table
17(IDT). There are a few strictly unnecessary things that get mapped
18such as the first C function when entering an interrupt (see
19comments in pti.c).
20
21This approach helps to ensure that side-channel attacks leveraging
22the paging structures do not function when PTI is enabled. It can be
23enabled by setting CONFIG_PAGE_TABLE_ISOLATION=y at compile time.
24Once enabled at compile-time, it can be disabled at boot with the
25'nopti' or 'pti=' kernel parameters (see kernel-parameters.txt).
26
27Page Table Management
28=====================
29
30When PTI is enabled, the kernel manages two sets of page tables.
31The first set is very similar to the single set which is present in
32kernels without PTI. This includes a complete mapping of userspace
33that the kernel can use for things like copy_to_user().
34
35Although _complete_, the user portion of the kernel page tables is
36crippled by setting the NX bit in the top level. This ensures
37that any missed kernel->user CR3 switch will immediately crash
38userspace upon executing its first instruction.
39
40The userspace page tables map only the kernel data needed to enter
41and exit the kernel. This data is entirely contained in the 'struct
42cpu_entry_area' structure which is placed in the fixmap which gives
43each CPU's copy of the area a compile-time-fixed virtual address.
44
45For new userspace mappings, the kernel makes the entries in its
46page tables like normal. The only difference is when the kernel
47makes entries in the top (PGD) level. In addition to setting the
48entry in the main kernel PGD, a copy of the entry is made in the
49userspace page tables' PGD.
50
51This sharing at the PGD level also inherently shares all the lower
52layers of the page tables. This leaves a single, shared set of
53userspace page tables to manage. One PTE to lock, one set of
54accessed bits, dirty bits, etc...
55
56Overhead
57========
58
59Protection against side-channel attacks is important. But,
60this protection comes at a cost:
61
621. Increased Memory Use
63 a. Each process now needs an order-1 PGD instead of order-0.
64 (Consumes an additional 4k per process).
65 b. The 'cpu_entry_area' structure must be 2MB in size and 2MB
66 aligned so that it can be mapped by setting a single PMD
67 entry. This consumes nearly 2MB of RAM once the kernel
68 is decompressed, but no space in the kernel image itself.
69
702. Runtime Cost
71 a. CR3 manipulation to switch between the page table copies
72 must be done at interrupt, syscall, and exception entry
73 and exit (it can be skipped when the kernel is interrupted,
74 though.) Moves to CR3 are on the order of a hundred
75 cycles, and are required at every entry and exit.
76 b. A "trampoline" must be used for SYSCALL entry. This
77 trampoline depends on a smaller set of resources than the
78 non-PTI SYSCALL entry code, so requires mapping fewer
79 things into the userspace page tables. The downside is
80 that stacks must be switched at entry time.
81 d. Global pages are disabled for all kernel structures not
82 mapped into both kernel and userspace page tables. This
83 feature of the MMU allows different processes to share TLB
84 entries mapping the kernel. Losing the feature means more
85 TLB misses after a context switch. The actual loss of
86 performance is very small, however, never exceeding 1%.
87 d. Process Context IDentifiers (PCID) is a CPU feature that
88 allows us to skip flushing the entire TLB when switching page
89 tables by setting a special bit in CR3 when the page tables
90 are changed. This makes switching the page tables (at context
91 switch, or kernel entry/exit) cheaper. But, on systems with
92 PCID support, the context switch code must flush both the user
93 and kernel entries out of the TLB. The user PCID TLB flush is
94 deferred until the exit to userspace, minimizing the cost.
95 See intel.com/sdm for the gory PCID/INVPCID details.
96 e. The userspace page tables must be populated for each new
97 process. Even without PTI, the shared kernel mappings
98 are created by copying top-level (PGD) entries into each
99 new process. But, with PTI, there are now *two* kernel
100 mappings: one in the kernel page tables that maps everything
101 and one for the entry/exit structures. At fork(), we need to
102 copy both.
103 f. In addition to the fork()-time copying, there must also
104 be an update to the userspace PGD any time a set_pgd() is done
105 on a PGD used to map userspace. This ensures that the kernel
106 and userspace copies always map the same userspace
107 memory.
108 g. On systems without PCID support, each CR3 write flushes
109 the entire TLB. That means that each syscall, interrupt
110 or exception flushes the TLB.
111 h. INVPCID is a TLB-flushing instruction which allows flushing
112 of TLB entries for non-current PCIDs. Some systems support
113 PCIDs, but do not support INVPCID. On these systems, addresses
114 can only be flushed from the TLB for the current PCID. When
115 flushing a kernel address, we need to flush all PCIDs, so a
116 single kernel address flush will require a TLB-flushing CR3
117 write upon the next use of every PCID.
118
119Possible Future Work
120====================
1211. We can be more careful about not actually writing to CR3
122 unless its value is actually changed.
1232. Allow PTI to be enabled/disabled at runtime in addition to the
124 boot-time switching.
125
126Testing
127========
128
129To test stability of PTI, the following test procedure is recommended,
130ideally doing all of these in parallel:
131
1321. Set CONFIG_DEBUG_ENTRY=y
1332. Run several copies of all of the tools/testing/selftests/x86/ tests
134 (excluding MPX and protection_keys) in a loop on multiple CPUs for
135 several minutes. These tests frequently uncover corner cases in the
136 kernel entry code. In general, old kernels might cause these tests
137 themselves to crash, but they should never crash the kernel.
1383. Run the 'perf' tool in a mode (top or record) that generates many
139 frequent performance monitoring non-maskable interrupts (see "NMI"
140 in /proc/interrupts). This exercises the NMI entry/exit code which
141 is known to trigger bugs in code paths that did not expect to be
142 interrupted, including nested NMIs. Using "-c" boosts the rate of
143 NMIs, and using two -c with separate counters encourages nested NMIs
144 and less deterministic behavior.
145
146 while true; do perf record -c 10000 -e instructions,cycles -a sleep 10; done
147
1484. Launch a KVM virtual machine.
1495. Run 32-bit binaries on systems supporting the SYSCALL instruction.
150 This has been a lightly-tested code path and needs extra scrutiny.
151
152Debugging
153=========
154
155Bugs in PTI cause a few different signatures of crashes
156that are worth noting here.
157
158 * Failures of the selftests/x86 code. Usually a bug in one of the
159 more obscure corners of entry_64.S
160 * Crashes in early boot, especially around CPU bringup. Bugs
161 in the trampoline code or mappings cause these.
162 * Crashes at the first interrupt. Caused by bugs in entry_64.S,
163 like screwing up a page table switch. Also caused by
164 incorrectly mapping the IRQ handler entry code.
165 * Crashes at the first NMI. The NMI code is separate from main
166 interrupt handlers and can have bugs that do not affect
167 normal interrupts. Also caused by incorrectly mapping NMI
168 code. NMIs that interrupt the entry code must be very
169 careful and can be the cause of crashes that show up when
170 running perf.
171 * Kernel crashes at the first exit to userspace. entry_64.S
172 bugs, or failing to map some of the exit code.
173 * Crashes at first interrupt that interrupts userspace. The paths
174 in entry_64.S that return to userspace are sometimes separate
175 from the ones that return to the kernel.
176 * Double faults: overflowing the kernel stack because of page
177 faults upon page faults. Caused by touching non-pti-mapped
178 data in the entry code, or forgetting to switch to kernel
179 CR3 before calling into C functions which are not pti-mapped.
180 * Userspace segfaults early in boot, sometimes manifesting
181 as mount(8) failing to mount the rootfs. These have
182 tended to be TLB invalidation issues. Usually invalidating
183 the wrong PCID, or otherwise missing an invalidation.
184
1851. https://gruss.cc/files/kaiser.pdf
1862. https://meltdownattack.com/meltdown.pdf
diff --git a/MAINTAINERS b/MAINTAINERS
index f4800a4c5065..49f72c59fea6 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -9638,8 +9638,8 @@ F: include/uapi/linux/sunrpc/
9638NILFS2 FILESYSTEM 9638NILFS2 FILESYSTEM
9639M: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> 9639M: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
9640L: linux-nilfs@vger.kernel.org 9640L: linux-nilfs@vger.kernel.org
9641W: http://nilfs.sourceforge.net/ 9641W: https://nilfs.sourceforge.io/
9642W: http://nilfs.osdn.jp/ 9642W: https://nilfs.osdn.jp/
9643T: git git://github.com/konis/nilfs2.git 9643T: git git://github.com/konis/nilfs2.git
9644S: Supported 9644S: Supported
9645F: Documentation/filesystems/nilfs2.txt 9645F: Documentation/filesystems/nilfs2.txt
@@ -10135,7 +10135,7 @@ F: drivers/irqchip/irq-ompic.c
10135F: drivers/irqchip/irq-or1k-* 10135F: drivers/irqchip/irq-or1k-*
10136 10136
10137OPENVSWITCH 10137OPENVSWITCH
10138M: Pravin Shelar <pshelar@nicira.com> 10138M: Pravin B Shelar <pshelar@ovn.org>
10139L: netdev@vger.kernel.org 10139L: netdev@vger.kernel.org
10140L: dev@openvswitch.org 10140L: dev@openvswitch.org
10141W: http://openvswitch.org 10141W: http://openvswitch.org
diff --git a/Makefile b/Makefile
index eb59638035dd..bf5b8cbb9469 100644
--- a/Makefile
+++ b/Makefile
@@ -2,7 +2,7 @@
2VERSION = 4 2VERSION = 4
3PATCHLEVEL = 15 3PATCHLEVEL = 15
4SUBLEVEL = 0 4SUBLEVEL = 0
5EXTRAVERSION = -rc7 5EXTRAVERSION = -rc8
6NAME = Fearless Coyote 6NAME = Fearless Coyote
7 7
8# *DOCUMENTATION* 8# *DOCUMENTATION*
@@ -484,26 +484,6 @@ CLANG_GCC_TC := --gcc-toolchain=$(GCC_TOOLCHAIN)
484endif 484endif
485KBUILD_CFLAGS += $(CLANG_TARGET) $(CLANG_GCC_TC) 485KBUILD_CFLAGS += $(CLANG_TARGET) $(CLANG_GCC_TC)
486KBUILD_AFLAGS += $(CLANG_TARGET) $(CLANG_GCC_TC) 486KBUILD_AFLAGS += $(CLANG_TARGET) $(CLANG_GCC_TC)
487KBUILD_CPPFLAGS += $(call cc-option,-Qunused-arguments,)
488KBUILD_CFLAGS += $(call cc-disable-warning, unused-variable)
489KBUILD_CFLAGS += $(call cc-disable-warning, format-invalid-specifier)
490KBUILD_CFLAGS += $(call cc-disable-warning, gnu)
491KBUILD_CFLAGS += $(call cc-disable-warning, address-of-packed-member)
492# Quiet clang warning: comparison of unsigned expression < 0 is always false
493KBUILD_CFLAGS += $(call cc-disable-warning, tautological-compare)
494# CLANG uses a _MergedGlobals as optimization, but this breaks modpost, as the
495# source of a reference will be _MergedGlobals and not on of the whitelisted names.
496# See modpost pattern 2
497KBUILD_CFLAGS += $(call cc-option, -mno-global-merge,)
498KBUILD_CFLAGS += $(call cc-option, -fcatch-undefined-behavior)
499KBUILD_CFLAGS += $(call cc-option, -no-integrated-as)
500KBUILD_AFLAGS += $(call cc-option, -no-integrated-as)
501else
502
503# These warnings generated too much noise in a regular build.
504# Use make W=1 to enable them (see scripts/Makefile.extrawarn)
505KBUILD_CFLAGS += $(call cc-disable-warning, unused-but-set-variable)
506KBUILD_CFLAGS += $(call cc-disable-warning, unused-const-variable)
507endif 487endif
508 488
509ifeq ($(config-targets),1) 489ifeq ($(config-targets),1)
@@ -716,6 +696,29 @@ ifdef CONFIG_CC_STACKPROTECTOR
716endif 696endif
717KBUILD_CFLAGS += $(stackp-flag) 697KBUILD_CFLAGS += $(stackp-flag)
718 698
699ifeq ($(cc-name),clang)
700KBUILD_CPPFLAGS += $(call cc-option,-Qunused-arguments,)
701KBUILD_CFLAGS += $(call cc-disable-warning, unused-variable)
702KBUILD_CFLAGS += $(call cc-disable-warning, format-invalid-specifier)
703KBUILD_CFLAGS += $(call cc-disable-warning, gnu)
704KBUILD_CFLAGS += $(call cc-disable-warning, address-of-packed-member)
705# Quiet clang warning: comparison of unsigned expression < 0 is always false
706KBUILD_CFLAGS += $(call cc-disable-warning, tautological-compare)
707# CLANG uses a _MergedGlobals as optimization, but this breaks modpost, as the
708# source of a reference will be _MergedGlobals and not on of the whitelisted names.
709# See modpost pattern 2
710KBUILD_CFLAGS += $(call cc-option, -mno-global-merge,)
711KBUILD_CFLAGS += $(call cc-option, -fcatch-undefined-behavior)
712KBUILD_CFLAGS += $(call cc-option, -no-integrated-as)
713KBUILD_AFLAGS += $(call cc-option, -no-integrated-as)
714else
715
716# These warnings generated too much noise in a regular build.
717# Use make W=1 to enable them (see scripts/Makefile.extrawarn)
718KBUILD_CFLAGS += $(call cc-disable-warning, unused-but-set-variable)
719KBUILD_CFLAGS += $(call cc-disable-warning, unused-const-variable)
720endif
721
719ifdef CONFIG_FRAME_POINTER 722ifdef CONFIG_FRAME_POINTER
720KBUILD_CFLAGS += -fno-omit-frame-pointer -fno-optimize-sibling-calls 723KBUILD_CFLAGS += -fno-omit-frame-pointer -fno-optimize-sibling-calls
721else 724else
diff --git a/arch/ia64/kernel/time.c b/arch/ia64/kernel/time.c
index c6ecb97151a2..9025699049ca 100644
--- a/arch/ia64/kernel/time.c
+++ b/arch/ia64/kernel/time.c
@@ -88,7 +88,7 @@ void vtime_flush(struct task_struct *tsk)
88 } 88 }
89 89
90 if (ti->softirq_time) { 90 if (ti->softirq_time) {
91 delta = cycle_to_nsec(ti->softirq_time)); 91 delta = cycle_to_nsec(ti->softirq_time);
92 account_system_index_time(tsk, delta, CPUTIME_SOFTIRQ); 92 account_system_index_time(tsk, delta, CPUTIME_SOFTIRQ);
93 } 93 }
94 94
diff --git a/arch/mips/kernel/cps-vec.S b/arch/mips/kernel/cps-vec.S
index c7ed26029cbb..e68e6e04063a 100644
--- a/arch/mips/kernel/cps-vec.S
+++ b/arch/mips/kernel/cps-vec.S
@@ -235,6 +235,7 @@ LEAF(mips_cps_core_init)
235 has_mt t0, 3f 235 has_mt t0, 3f
236 236
237 .set push 237 .set push
238 .set MIPS_ISA_LEVEL_RAW
238 .set mt 239 .set mt
239 240
240 /* Only allow 1 TC per VPE to execute... */ 241 /* Only allow 1 TC per VPE to execute... */
@@ -388,6 +389,7 @@ LEAF(mips_cps_boot_vpes)
388#elif defined(CONFIG_MIPS_MT) 389#elif defined(CONFIG_MIPS_MT)
389 390
390 .set push 391 .set push
392 .set MIPS_ISA_LEVEL_RAW
391 .set mt 393 .set mt
392 394
393 /* If the core doesn't support MT then return */ 395 /* If the core doesn't support MT then return */
diff --git a/arch/mips/kernel/process.c b/arch/mips/kernel/process.c
index 45d0b6b037ee..57028d49c202 100644
--- a/arch/mips/kernel/process.c
+++ b/arch/mips/kernel/process.c
@@ -705,6 +705,18 @@ int mips_set_process_fp_mode(struct task_struct *task, unsigned int value)
705 struct task_struct *t; 705 struct task_struct *t;
706 int max_users; 706 int max_users;
707 707
708 /* If nothing to change, return right away, successfully. */
709 if (value == mips_get_process_fp_mode(task))
710 return 0;
711
712 /* Only accept a mode change if 64-bit FP enabled for o32. */
713 if (!IS_ENABLED(CONFIG_MIPS_O32_FP64_SUPPORT))
714 return -EOPNOTSUPP;
715
716 /* And only for o32 tasks. */
717 if (IS_ENABLED(CONFIG_64BIT) && !test_thread_flag(TIF_32BIT_REGS))
718 return -EOPNOTSUPP;
719
708 /* Check the value is valid */ 720 /* Check the value is valid */
709 if (value & ~known_bits) 721 if (value & ~known_bits)
710 return -EOPNOTSUPP; 722 return -EOPNOTSUPP;
diff --git a/arch/mips/kernel/ptrace.c b/arch/mips/kernel/ptrace.c
index efbd8df8b665..0b23b1ad99e6 100644
--- a/arch/mips/kernel/ptrace.c
+++ b/arch/mips/kernel/ptrace.c
@@ -419,63 +419,160 @@ static int gpr64_set(struct task_struct *target,
419 419
420#endif /* CONFIG_64BIT */ 420#endif /* CONFIG_64BIT */
421 421
422/*
423 * Copy the floating-point context to the supplied NT_PRFPREG buffer,
424 * !CONFIG_CPU_HAS_MSA variant. FP context's general register slots
425 * correspond 1:1 to buffer slots. Only general registers are copied.
426 */
427static int fpr_get_fpa(struct task_struct *target,
428 unsigned int *pos, unsigned int *count,
429 void **kbuf, void __user **ubuf)
430{
431 return user_regset_copyout(pos, count, kbuf, ubuf,
432 &target->thread.fpu,
433 0, NUM_FPU_REGS * sizeof(elf_fpreg_t));
434}
435
436/*
437 * Copy the floating-point context to the supplied NT_PRFPREG buffer,
438 * CONFIG_CPU_HAS_MSA variant. Only lower 64 bits of FP context's
439 * general register slots are copied to buffer slots. Only general
440 * registers are copied.
441 */
442static int fpr_get_msa(struct task_struct *target,
443 unsigned int *pos, unsigned int *count,
444 void **kbuf, void __user **ubuf)
445{
446 unsigned int i;
447 u64 fpr_val;
448 int err;
449
450 BUILD_BUG_ON(sizeof(fpr_val) != sizeof(elf_fpreg_t));
451 for (i = 0; i < NUM_FPU_REGS; i++) {
452 fpr_val = get_fpr64(&target->thread.fpu.fpr[i], 0);
453 err = user_regset_copyout(pos, count, kbuf, ubuf,
454 &fpr_val, i * sizeof(elf_fpreg_t),
455 (i + 1) * sizeof(elf_fpreg_t));
456 if (err)
457 return err;
458 }
459
460 return 0;
461}
462
463/*
464 * Copy the floating-point context to the supplied NT_PRFPREG buffer.
465 * Choose the appropriate helper for general registers, and then copy
466 * the FCSR register separately.
467 */
422static int fpr_get(struct task_struct *target, 468static int fpr_get(struct task_struct *target,
423 const struct user_regset *regset, 469 const struct user_regset *regset,
424 unsigned int pos, unsigned int count, 470 unsigned int pos, unsigned int count,
425 void *kbuf, void __user *ubuf) 471 void *kbuf, void __user *ubuf)
426{ 472{
427 unsigned i; 473 const int fcr31_pos = NUM_FPU_REGS * sizeof(elf_fpreg_t);
428 int err; 474 int err;
429 u64 fpr_val;
430 475
431 /* XXX fcr31 */ 476 if (sizeof(target->thread.fpu.fpr[0]) == sizeof(elf_fpreg_t))
477 err = fpr_get_fpa(target, &pos, &count, &kbuf, &ubuf);
478 else
479 err = fpr_get_msa(target, &pos, &count, &kbuf, &ubuf);
480 if (err)
481 return err;
432 482
433 if (sizeof(target->thread.fpu.fpr[i]) == sizeof(elf_fpreg_t)) 483 err = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
434 return user_regset_copyout(&pos, &count, &kbuf, &ubuf, 484 &target->thread.fpu.fcr31,
435 &target->thread.fpu, 485 fcr31_pos, fcr31_pos + sizeof(u32));
436 0, sizeof(elf_fpregset_t));
437 486
438 for (i = 0; i < NUM_FPU_REGS; i++) { 487 return err;
439 fpr_val = get_fpr64(&target->thread.fpu.fpr[i], 0); 488}
440 err = user_regset_copyout(&pos, &count, &kbuf, &ubuf, 489
441 &fpr_val, i * sizeof(elf_fpreg_t), 490/*
442 (i + 1) * sizeof(elf_fpreg_t)); 491 * Copy the supplied NT_PRFPREG buffer to the floating-point context,
492 * !CONFIG_CPU_HAS_MSA variant. Buffer slots correspond 1:1 to FP
493 * context's general register slots. Only general registers are copied.
494 */
495static int fpr_set_fpa(struct task_struct *target,
496 unsigned int *pos, unsigned int *count,
497 const void **kbuf, const void __user **ubuf)
498{
499 return user_regset_copyin(pos, count, kbuf, ubuf,
500 &target->thread.fpu,
501 0, NUM_FPU_REGS * sizeof(elf_fpreg_t));
502}
503
504/*
505 * Copy the supplied NT_PRFPREG buffer to the floating-point context,
506 * CONFIG_CPU_HAS_MSA variant. Buffer slots are copied to lower 64
507 * bits only of FP context's general register slots. Only general
508 * registers are copied.
509 */
510static int fpr_set_msa(struct task_struct *target,
511 unsigned int *pos, unsigned int *count,
512 const void **kbuf, const void __user **ubuf)
513{
514 unsigned int i;
515 u64 fpr_val;
516 int err;
517
518 BUILD_BUG_ON(sizeof(fpr_val) != sizeof(elf_fpreg_t));
519 for (i = 0; i < NUM_FPU_REGS && *count > 0; i++) {
520 err = user_regset_copyin(pos, count, kbuf, ubuf,
521 &fpr_val, i * sizeof(elf_fpreg_t),
522 (i + 1) * sizeof(elf_fpreg_t));
443 if (err) 523 if (err)
444 return err; 524 return err;
525 set_fpr64(&target->thread.fpu.fpr[i], 0, fpr_val);
445 } 526 }
446 527
447 return 0; 528 return 0;
448} 529}
449 530
531/*
532 * Copy the supplied NT_PRFPREG buffer to the floating-point context.
533 * Choose the appropriate helper for general registers, and then copy
534 * the FCSR register separately.
535 *
536 * We optimize for the case where `count % sizeof(elf_fpreg_t) == 0',
537 * which is supposed to have been guaranteed by the kernel before
538 * calling us, e.g. in `ptrace_regset'. We enforce that requirement,
539 * so that we can safely avoid preinitializing temporaries for
540 * partial register writes.
541 */
450static int fpr_set(struct task_struct *target, 542static int fpr_set(struct task_struct *target,
451 const struct user_regset *regset, 543 const struct user_regset *regset,
452 unsigned int pos, unsigned int count, 544 unsigned int pos, unsigned int count,
453 const void *kbuf, const void __user *ubuf) 545 const void *kbuf, const void __user *ubuf)
454{ 546{
455 unsigned i; 547 const int fcr31_pos = NUM_FPU_REGS * sizeof(elf_fpreg_t);
548 u32 fcr31;
456 int err; 549 int err;
457 u64 fpr_val;
458 550
459 /* XXX fcr31 */ 551 BUG_ON(count % sizeof(elf_fpreg_t));
552
553 if (pos + count > sizeof(elf_fpregset_t))
554 return -EIO;
460 555
461 init_fp_ctx(target); 556 init_fp_ctx(target);
462 557
463 if (sizeof(target->thread.fpu.fpr[i]) == sizeof(elf_fpreg_t)) 558 if (sizeof(target->thread.fpu.fpr[0]) == sizeof(elf_fpreg_t))
464 return user_regset_copyin(&pos, &count, &kbuf, &ubuf, 559 err = fpr_set_fpa(target, &pos, &count, &kbuf, &ubuf);
465 &target->thread.fpu, 560 else
466 0, sizeof(elf_fpregset_t)); 561 err = fpr_set_msa(target, &pos, &count, &kbuf, &ubuf);
562 if (err)
563 return err;
467 564
468 BUILD_BUG_ON(sizeof(fpr_val) != sizeof(elf_fpreg_t)); 565 if (count > 0) {
469 for (i = 0; i < NUM_FPU_REGS && count >= sizeof(elf_fpreg_t); i++) {
470 err = user_regset_copyin(&pos, &count, &kbuf, &ubuf, 566 err = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
471 &fpr_val, i * sizeof(elf_fpreg_t), 567 &fcr31,
472 (i + 1) * sizeof(elf_fpreg_t)); 568 fcr31_pos, fcr31_pos + sizeof(u32));
473 if (err) 569 if (err)
474 return err; 570 return err;
475 set_fpr64(&target->thread.fpu.fpr[i], 0, fpr_val); 571
572 ptrace_setfcr31(target, fcr31);
476 } 573 }
477 574
478 return 0; 575 return err;
479} 576}
480 577
481enum mips_regset { 578enum mips_regset {
diff --git a/arch/powerpc/include/asm/exception-64e.h b/arch/powerpc/include/asm/exception-64e.h
index a703452d67b6..555e22d5e07f 100644
--- a/arch/powerpc/include/asm/exception-64e.h
+++ b/arch/powerpc/include/asm/exception-64e.h
@@ -209,5 +209,11 @@ exc_##label##_book3e:
209 ori r3,r3,vector_offset@l; \ 209 ori r3,r3,vector_offset@l; \
210 mtspr SPRN_IVOR##vector_number,r3; 210 mtspr SPRN_IVOR##vector_number,r3;
211 211
212#define RFI_TO_KERNEL \
213 rfi
214
215#define RFI_TO_USER \
216 rfi
217
212#endif /* _ASM_POWERPC_EXCEPTION_64E_H */ 218#endif /* _ASM_POWERPC_EXCEPTION_64E_H */
213 219
diff --git a/arch/powerpc/include/asm/exception-64s.h b/arch/powerpc/include/asm/exception-64s.h
index b27205297e1d..7197b179c1b1 100644
--- a/arch/powerpc/include/asm/exception-64s.h
+++ b/arch/powerpc/include/asm/exception-64s.h
@@ -74,6 +74,59 @@
74 */ 74 */
75#define EX_R3 EX_DAR 75#define EX_R3 EX_DAR
76 76
77/*
78 * Macros for annotating the expected destination of (h)rfid
79 *
80 * The nop instructions allow us to insert one or more instructions to flush the
81 * L1-D cache when returning to userspace or a guest.
82 */
83#define RFI_FLUSH_SLOT \
84 RFI_FLUSH_FIXUP_SECTION; \
85 nop; \
86 nop; \
87 nop
88
89#define RFI_TO_KERNEL \
90 rfid
91
92#define RFI_TO_USER \
93 RFI_FLUSH_SLOT; \
94 rfid; \
95 b rfi_flush_fallback
96
97#define RFI_TO_USER_OR_KERNEL \
98 RFI_FLUSH_SLOT; \
99 rfid; \
100 b rfi_flush_fallback
101
102#define RFI_TO_GUEST \
103 RFI_FLUSH_SLOT; \
104 rfid; \
105 b rfi_flush_fallback
106
107#define HRFI_TO_KERNEL \
108 hrfid
109
110#define HRFI_TO_USER \
111 RFI_FLUSH_SLOT; \
112 hrfid; \
113 b hrfi_flush_fallback
114
115#define HRFI_TO_USER_OR_KERNEL \
116 RFI_FLUSH_SLOT; \
117 hrfid; \
118 b hrfi_flush_fallback
119
120#define HRFI_TO_GUEST \
121 RFI_FLUSH_SLOT; \
122 hrfid; \
123 b hrfi_flush_fallback
124
125#define HRFI_TO_UNKNOWN \
126 RFI_FLUSH_SLOT; \
127 hrfid; \
128 b hrfi_flush_fallback
129
77#ifdef CONFIG_RELOCATABLE 130#ifdef CONFIG_RELOCATABLE
78#define __EXCEPTION_RELON_PROLOG_PSERIES_1(label, h) \ 131#define __EXCEPTION_RELON_PROLOG_PSERIES_1(label, h) \
79 mfspr r11,SPRN_##h##SRR0; /* save SRR0 */ \ 132 mfspr r11,SPRN_##h##SRR0; /* save SRR0 */ \
@@ -218,7 +271,7 @@ END_FTR_SECTION_NESTED(ftr,ftr,943)
218 mtspr SPRN_##h##SRR0,r12; \ 271 mtspr SPRN_##h##SRR0,r12; \
219 mfspr r12,SPRN_##h##SRR1; /* and SRR1 */ \ 272 mfspr r12,SPRN_##h##SRR1; /* and SRR1 */ \
220 mtspr SPRN_##h##SRR1,r10; \ 273 mtspr SPRN_##h##SRR1,r10; \
221 h##rfid; \ 274 h##RFI_TO_KERNEL; \
222 b . /* prevent speculative execution */ 275 b . /* prevent speculative execution */
223#define EXCEPTION_PROLOG_PSERIES_1(label, h) \ 276#define EXCEPTION_PROLOG_PSERIES_1(label, h) \
224 __EXCEPTION_PROLOG_PSERIES_1(label, h) 277 __EXCEPTION_PROLOG_PSERIES_1(label, h)
@@ -232,7 +285,7 @@ END_FTR_SECTION_NESTED(ftr,ftr,943)
232 mtspr SPRN_##h##SRR0,r12; \ 285 mtspr SPRN_##h##SRR0,r12; \
233 mfspr r12,SPRN_##h##SRR1; /* and SRR1 */ \ 286 mfspr r12,SPRN_##h##SRR1; /* and SRR1 */ \
234 mtspr SPRN_##h##SRR1,r10; \ 287 mtspr SPRN_##h##SRR1,r10; \
235 h##rfid; \ 288 h##RFI_TO_KERNEL; \
236 b . /* prevent speculative execution */ 289 b . /* prevent speculative execution */
237 290
238#define EXCEPTION_PROLOG_PSERIES_1_NORI(label, h) \ 291#define EXCEPTION_PROLOG_PSERIES_1_NORI(label, h) \
diff --git a/arch/powerpc/include/asm/feature-fixups.h b/arch/powerpc/include/asm/feature-fixups.h
index 8f88f771cc55..1e82eb3caabd 100644
--- a/arch/powerpc/include/asm/feature-fixups.h
+++ b/arch/powerpc/include/asm/feature-fixups.h
@@ -187,7 +187,20 @@ label##3: \
187 FTR_ENTRY_OFFSET label##1b-label##3b; \ 187 FTR_ENTRY_OFFSET label##1b-label##3b; \
188 .popsection; 188 .popsection;
189 189
190#define RFI_FLUSH_FIXUP_SECTION \
191951: \
192 .pushsection __rfi_flush_fixup,"a"; \
193 .align 2; \
194952: \
195 FTR_ENTRY_OFFSET 951b-952b; \
196 .popsection;
197
198
190#ifndef __ASSEMBLY__ 199#ifndef __ASSEMBLY__
200#include <linux/types.h>
201
202extern long __start___rfi_flush_fixup, __stop___rfi_flush_fixup;
203
191void apply_feature_fixups(void); 204void apply_feature_fixups(void);
192void setup_feature_keys(void); 205void setup_feature_keys(void);
193#endif 206#endif
diff --git a/arch/powerpc/include/asm/hvcall.h b/arch/powerpc/include/asm/hvcall.h
index a409177be8bd..f0461618bf7b 100644
--- a/arch/powerpc/include/asm/hvcall.h
+++ b/arch/powerpc/include/asm/hvcall.h
@@ -241,6 +241,7 @@
241#define H_GET_HCA_INFO 0x1B8 241#define H_GET_HCA_INFO 0x1B8
242#define H_GET_PERF_COUNT 0x1BC 242#define H_GET_PERF_COUNT 0x1BC
243#define H_MANAGE_TRACE 0x1C0 243#define H_MANAGE_TRACE 0x1C0
244#define H_GET_CPU_CHARACTERISTICS 0x1C8
244#define H_FREE_LOGICAL_LAN_BUFFER 0x1D4 245#define H_FREE_LOGICAL_LAN_BUFFER 0x1D4
245#define H_QUERY_INT_STATE 0x1E4 246#define H_QUERY_INT_STATE 0x1E4
246#define H_POLL_PENDING 0x1D8 247#define H_POLL_PENDING 0x1D8
@@ -330,6 +331,17 @@
330#define H_SIGNAL_SYS_RESET_ALL_OTHERS -2 331#define H_SIGNAL_SYS_RESET_ALL_OTHERS -2
331/* >= 0 values are CPU number */ 332/* >= 0 values are CPU number */
332 333
334/* H_GET_CPU_CHARACTERISTICS return values */
335#define H_CPU_CHAR_SPEC_BAR_ORI31 (1ull << 63) // IBM bit 0
336#define H_CPU_CHAR_BCCTRL_SERIALISED (1ull << 62) // IBM bit 1
337#define H_CPU_CHAR_L1D_FLUSH_ORI30 (1ull << 61) // IBM bit 2
338#define H_CPU_CHAR_L1D_FLUSH_TRIG2 (1ull << 60) // IBM bit 3
339#define H_CPU_CHAR_L1D_THREAD_PRIV (1ull << 59) // IBM bit 4
340
341#define H_CPU_BEHAV_FAVOUR_SECURITY (1ull << 63) // IBM bit 0
342#define H_CPU_BEHAV_L1D_FLUSH_PR (1ull << 62) // IBM bit 1
343#define H_CPU_BEHAV_BNDS_CHK_SPEC_BAR (1ull << 61) // IBM bit 2
344
333/* Flag values used in H_REGISTER_PROC_TBL hcall */ 345/* Flag values used in H_REGISTER_PROC_TBL hcall */
334#define PROC_TABLE_OP_MASK 0x18 346#define PROC_TABLE_OP_MASK 0x18
335#define PROC_TABLE_DEREG 0x10 347#define PROC_TABLE_DEREG 0x10
@@ -436,6 +448,11 @@ static inline unsigned int get_longbusy_msecs(int longbusy_rc)
436 } 448 }
437} 449}
438 450
451struct h_cpu_char_result {
452 u64 character;
453 u64 behaviour;
454};
455
439#endif /* __ASSEMBLY__ */ 456#endif /* __ASSEMBLY__ */
440#endif /* __KERNEL__ */ 457#endif /* __KERNEL__ */
441#endif /* _ASM_POWERPC_HVCALL_H */ 458#endif /* _ASM_POWERPC_HVCALL_H */
diff --git a/arch/powerpc/include/asm/paca.h b/arch/powerpc/include/asm/paca.h
index 3892db93b837..23ac7fc0af23 100644
--- a/arch/powerpc/include/asm/paca.h
+++ b/arch/powerpc/include/asm/paca.h
@@ -232,6 +232,16 @@ struct paca_struct {
232 struct sibling_subcore_state *sibling_subcore_state; 232 struct sibling_subcore_state *sibling_subcore_state;
233#endif 233#endif
234#endif 234#endif
235#ifdef CONFIG_PPC_BOOK3S_64
236 /*
237 * rfi fallback flush must be in its own cacheline to prevent
238 * other paca data leaking into the L1d
239 */
240 u64 exrfi[EX_SIZE] __aligned(0x80);
241 void *rfi_flush_fallback_area;
242 u64 l1d_flush_congruence;
243 u64 l1d_flush_sets;
244#endif
235}; 245};
236 246
237extern void copy_mm_to_paca(struct mm_struct *mm); 247extern void copy_mm_to_paca(struct mm_struct *mm);
diff --git a/arch/powerpc/include/asm/plpar_wrappers.h b/arch/powerpc/include/asm/plpar_wrappers.h
index 7f01b22fa6cb..55eddf50d149 100644
--- a/arch/powerpc/include/asm/plpar_wrappers.h
+++ b/arch/powerpc/include/asm/plpar_wrappers.h
@@ -326,4 +326,18 @@ static inline long plapr_signal_sys_reset(long cpu)
326 return plpar_hcall_norets(H_SIGNAL_SYS_RESET, cpu); 326 return plpar_hcall_norets(H_SIGNAL_SYS_RESET, cpu);
327} 327}
328 328
329static inline long plpar_get_cpu_characteristics(struct h_cpu_char_result *p)
330{
331 unsigned long retbuf[PLPAR_HCALL_BUFSIZE];
332 long rc;
333
334 rc = plpar_hcall(H_GET_CPU_CHARACTERISTICS, retbuf);
335 if (rc == H_SUCCESS) {
336 p->character = retbuf[0];
337 p->behaviour = retbuf[1];
338 }
339
340 return rc;
341}
342
329#endif /* _ASM_POWERPC_PLPAR_WRAPPERS_H */ 343#endif /* _ASM_POWERPC_PLPAR_WRAPPERS_H */
diff --git a/arch/powerpc/include/asm/setup.h b/arch/powerpc/include/asm/setup.h
index cf00ec26303a..469b7fdc9be4 100644
--- a/arch/powerpc/include/asm/setup.h
+++ b/arch/powerpc/include/asm/setup.h
@@ -39,6 +39,19 @@ static inline void pseries_big_endian_exceptions(void) {}
39static inline void pseries_little_endian_exceptions(void) {} 39static inline void pseries_little_endian_exceptions(void) {}
40#endif /* CONFIG_PPC_PSERIES */ 40#endif /* CONFIG_PPC_PSERIES */
41 41
42void rfi_flush_enable(bool enable);
43
44/* These are bit flags */
45enum l1d_flush_type {
46 L1D_FLUSH_NONE = 0x1,
47 L1D_FLUSH_FALLBACK = 0x2,
48 L1D_FLUSH_ORI = 0x4,
49 L1D_FLUSH_MTTRIG = 0x8,
50};
51
52void __init setup_rfi_flush(enum l1d_flush_type, bool enable);
53void do_rfi_flush_fixups(enum l1d_flush_type types);
54
42#endif /* !__ASSEMBLY__ */ 55#endif /* !__ASSEMBLY__ */
43 56
44#endif /* _ASM_POWERPC_SETUP_H */ 57#endif /* _ASM_POWERPC_SETUP_H */
diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c
index 6b958414b4e0..f390d57cf2e1 100644
--- a/arch/powerpc/kernel/asm-offsets.c
+++ b/arch/powerpc/kernel/asm-offsets.c
@@ -237,6 +237,11 @@ int main(void)
237 OFFSET(PACA_NMI_EMERG_SP, paca_struct, nmi_emergency_sp); 237 OFFSET(PACA_NMI_EMERG_SP, paca_struct, nmi_emergency_sp);
238 OFFSET(PACA_IN_MCE, paca_struct, in_mce); 238 OFFSET(PACA_IN_MCE, paca_struct, in_mce);
239 OFFSET(PACA_IN_NMI, paca_struct, in_nmi); 239 OFFSET(PACA_IN_NMI, paca_struct, in_nmi);
240 OFFSET(PACA_RFI_FLUSH_FALLBACK_AREA, paca_struct, rfi_flush_fallback_area);
241 OFFSET(PACA_EXRFI, paca_struct, exrfi);
242 OFFSET(PACA_L1D_FLUSH_CONGRUENCE, paca_struct, l1d_flush_congruence);
243 OFFSET(PACA_L1D_FLUSH_SETS, paca_struct, l1d_flush_sets);
244
240#endif 245#endif
241 OFFSET(PACAHWCPUID, paca_struct, hw_cpu_id); 246 OFFSET(PACAHWCPUID, paca_struct, hw_cpu_id);
242 OFFSET(PACAKEXECSTATE, paca_struct, kexec_state); 247 OFFSET(PACAKEXECSTATE, paca_struct, kexec_state);
diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S
index 3320bcac7192..2748584b767d 100644
--- a/arch/powerpc/kernel/entry_64.S
+++ b/arch/powerpc/kernel/entry_64.S
@@ -37,6 +37,11 @@
37#include <asm/tm.h> 37#include <asm/tm.h>
38#include <asm/ppc-opcode.h> 38#include <asm/ppc-opcode.h>
39#include <asm/export.h> 39#include <asm/export.h>
40#ifdef CONFIG_PPC_BOOK3S
41#include <asm/exception-64s.h>
42#else
43#include <asm/exception-64e.h>
44#endif
40 45
41/* 46/*
42 * System calls. 47 * System calls.
@@ -262,13 +267,23 @@ BEGIN_FTR_SECTION
262END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR) 267END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR)
263 268
264 ld r13,GPR13(r1) /* only restore r13 if returning to usermode */ 269 ld r13,GPR13(r1) /* only restore r13 if returning to usermode */
270 ld r2,GPR2(r1)
271 ld r1,GPR1(r1)
272 mtlr r4
273 mtcr r5
274 mtspr SPRN_SRR0,r7
275 mtspr SPRN_SRR1,r8
276 RFI_TO_USER
277 b . /* prevent speculative execution */
278
279 /* exit to kernel */
2651: ld r2,GPR2(r1) 2801: ld r2,GPR2(r1)
266 ld r1,GPR1(r1) 281 ld r1,GPR1(r1)
267 mtlr r4 282 mtlr r4
268 mtcr r5 283 mtcr r5
269 mtspr SPRN_SRR0,r7 284 mtspr SPRN_SRR0,r7
270 mtspr SPRN_SRR1,r8 285 mtspr SPRN_SRR1,r8
271 RFI 286 RFI_TO_KERNEL
272 b . /* prevent speculative execution */ 287 b . /* prevent speculative execution */
273 288
274.Lsyscall_error: 289.Lsyscall_error:
@@ -397,8 +412,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR)
397 mtmsrd r10, 1 412 mtmsrd r10, 1
398 mtspr SPRN_SRR0, r11 413 mtspr SPRN_SRR0, r11
399 mtspr SPRN_SRR1, r12 414 mtspr SPRN_SRR1, r12
400 415 RFI_TO_USER
401 rfid
402 b . /* prevent speculative execution */ 416 b . /* prevent speculative execution */
403#endif 417#endif
404_ASM_NOKPROBE_SYMBOL(system_call_common); 418_ASM_NOKPROBE_SYMBOL(system_call_common);
@@ -878,7 +892,7 @@ BEGIN_FTR_SECTION
878END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR) 892END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR)
879 ACCOUNT_CPU_USER_EXIT(r13, r2, r4) 893 ACCOUNT_CPU_USER_EXIT(r13, r2, r4)
880 REST_GPR(13, r1) 894 REST_GPR(13, r1)
8811: 895
882 mtspr SPRN_SRR1,r3 896 mtspr SPRN_SRR1,r3
883 897
884 ld r2,_CCR(r1) 898 ld r2,_CCR(r1)
@@ -891,8 +905,22 @@ END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR)
891 ld r3,GPR3(r1) 905 ld r3,GPR3(r1)
892 ld r4,GPR4(r1) 906 ld r4,GPR4(r1)
893 ld r1,GPR1(r1) 907 ld r1,GPR1(r1)
908 RFI_TO_USER
909 b . /* prevent speculative execution */
894 910
895 rfid 9111: mtspr SPRN_SRR1,r3
912
913 ld r2,_CCR(r1)
914 mtcrf 0xFF,r2
915 ld r2,_NIP(r1)
916 mtspr SPRN_SRR0,r2
917
918 ld r0,GPR0(r1)
919 ld r2,GPR2(r1)
920 ld r3,GPR3(r1)
921 ld r4,GPR4(r1)
922 ld r1,GPR1(r1)
923 RFI_TO_KERNEL
896 b . /* prevent speculative execution */ 924 b . /* prevent speculative execution */
897 925
898#endif /* CONFIG_PPC_BOOK3E */ 926#endif /* CONFIG_PPC_BOOK3E */
@@ -1073,7 +1101,7 @@ __enter_rtas:
1073 1101
1074 mtspr SPRN_SRR0,r5 1102 mtspr SPRN_SRR0,r5
1075 mtspr SPRN_SRR1,r6 1103 mtspr SPRN_SRR1,r6
1076 rfid 1104 RFI_TO_KERNEL
1077 b . /* prevent speculative execution */ 1105 b . /* prevent speculative execution */
1078 1106
1079rtas_return_loc: 1107rtas_return_loc:
@@ -1098,7 +1126,7 @@ rtas_return_loc:
1098 1126
1099 mtspr SPRN_SRR0,r3 1127 mtspr SPRN_SRR0,r3
1100 mtspr SPRN_SRR1,r4 1128 mtspr SPRN_SRR1,r4
1101 rfid 1129 RFI_TO_KERNEL
1102 b . /* prevent speculative execution */ 1130 b . /* prevent speculative execution */
1103_ASM_NOKPROBE_SYMBOL(__enter_rtas) 1131_ASM_NOKPROBE_SYMBOL(__enter_rtas)
1104_ASM_NOKPROBE_SYMBOL(rtas_return_loc) 1132_ASM_NOKPROBE_SYMBOL(rtas_return_loc)
@@ -1171,7 +1199,7 @@ _GLOBAL(enter_prom)
1171 LOAD_REG_IMMEDIATE(r12, MSR_SF | MSR_ISF | MSR_LE) 1199 LOAD_REG_IMMEDIATE(r12, MSR_SF | MSR_ISF | MSR_LE)
1172 andc r11,r11,r12 1200 andc r11,r11,r12
1173 mtsrr1 r11 1201 mtsrr1 r11
1174 rfid 1202 RFI_TO_KERNEL
1175#endif /* CONFIG_PPC_BOOK3E */ 1203#endif /* CONFIG_PPC_BOOK3E */
1176 1204
11771: /* Return from OF */ 12051: /* Return from OF */
diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S
index e441b469dc8f..2dc10bf646b8 100644
--- a/arch/powerpc/kernel/exceptions-64s.S
+++ b/arch/powerpc/kernel/exceptions-64s.S
@@ -256,7 +256,7 @@ BEGIN_FTR_SECTION
256 LOAD_HANDLER(r12, machine_check_handle_early) 256 LOAD_HANDLER(r12, machine_check_handle_early)
2571: mtspr SPRN_SRR0,r12 2571: mtspr SPRN_SRR0,r12
258 mtspr SPRN_SRR1,r11 258 mtspr SPRN_SRR1,r11
259 rfid 259 RFI_TO_KERNEL
260 b . /* prevent speculative execution */ 260 b . /* prevent speculative execution */
2612: 2612:
262 /* Stack overflow. Stay on emergency stack and panic. 262 /* Stack overflow. Stay on emergency stack and panic.
@@ -445,7 +445,7 @@ EXC_COMMON_BEGIN(machine_check_handle_early)
445 li r3,MSR_ME 445 li r3,MSR_ME
446 andc r10,r10,r3 /* Turn off MSR_ME */ 446 andc r10,r10,r3 /* Turn off MSR_ME */
447 mtspr SPRN_SRR1,r10 447 mtspr SPRN_SRR1,r10
448 rfid 448 RFI_TO_KERNEL
449 b . 449 b .
4502: 4502:
451 /* 451 /*
@@ -463,7 +463,7 @@ EXC_COMMON_BEGIN(machine_check_handle_early)
463 */ 463 */
464 bl machine_check_queue_event 464 bl machine_check_queue_event
465 MACHINE_CHECK_HANDLER_WINDUP 465 MACHINE_CHECK_HANDLER_WINDUP
466 rfid 466 RFI_TO_USER_OR_KERNEL
4679: 4679:
468 /* Deliver the machine check to host kernel in V mode. */ 468 /* Deliver the machine check to host kernel in V mode. */
469 MACHINE_CHECK_HANDLER_WINDUP 469 MACHINE_CHECK_HANDLER_WINDUP
@@ -598,6 +598,9 @@ EXC_COMMON_BEGIN(slb_miss_common)
598 stw r9,PACA_EXSLB+EX_CCR(r13) /* save CR in exc. frame */ 598 stw r9,PACA_EXSLB+EX_CCR(r13) /* save CR in exc. frame */
599 std r10,PACA_EXSLB+EX_LR(r13) /* save LR */ 599 std r10,PACA_EXSLB+EX_LR(r13) /* save LR */
600 600
601 andi. r9,r11,MSR_PR // Check for exception from userspace
602 cmpdi cr4,r9,MSR_PR // And save the result in CR4 for later
603
601 /* 604 /*
602 * Test MSR_RI before calling slb_allocate_realmode, because the 605 * Test MSR_RI before calling slb_allocate_realmode, because the
603 * MSR in r11 gets clobbered. However we still want to allocate 606 * MSR in r11 gets clobbered. However we still want to allocate
@@ -624,9 +627,12 @@ END_MMU_FTR_SECTION_IFCLR(MMU_FTR_TYPE_RADIX)
624 627
625 /* All done -- return from exception. */ 628 /* All done -- return from exception. */
626 629
630 bne cr4,1f /* returning to kernel */
631
627.machine push 632.machine push
628.machine "power4" 633.machine "power4"
629 mtcrf 0x80,r9 634 mtcrf 0x80,r9
635 mtcrf 0x08,r9 /* MSR[PR] indication is in cr4 */
630 mtcrf 0x04,r9 /* MSR[RI] indication is in cr5 */ 636 mtcrf 0x04,r9 /* MSR[RI] indication is in cr5 */
631 mtcrf 0x02,r9 /* I/D indication is in cr6 */ 637 mtcrf 0x02,r9 /* I/D indication is in cr6 */
632 mtcrf 0x01,r9 /* slb_allocate uses cr0 and cr7 */ 638 mtcrf 0x01,r9 /* slb_allocate uses cr0 and cr7 */
@@ -640,9 +646,30 @@ END_MMU_FTR_SECTION_IFCLR(MMU_FTR_TYPE_RADIX)
640 ld r11,PACA_EXSLB+EX_R11(r13) 646 ld r11,PACA_EXSLB+EX_R11(r13)
641 ld r12,PACA_EXSLB+EX_R12(r13) 647 ld r12,PACA_EXSLB+EX_R12(r13)
642 ld r13,PACA_EXSLB+EX_R13(r13) 648 ld r13,PACA_EXSLB+EX_R13(r13)
643 rfid 649 RFI_TO_USER
650 b . /* prevent speculative execution */
6511:
652.machine push
653.machine "power4"
654 mtcrf 0x80,r9
655 mtcrf 0x08,r9 /* MSR[PR] indication is in cr4 */
656 mtcrf 0x04,r9 /* MSR[RI] indication is in cr5 */
657 mtcrf 0x02,r9 /* I/D indication is in cr6 */
658 mtcrf 0x01,r9 /* slb_allocate uses cr0 and cr7 */
659.machine pop
660
661 RESTORE_CTR(r9, PACA_EXSLB)
662 RESTORE_PPR_PACA(PACA_EXSLB, r9)
663 mr r3,r12
664 ld r9,PACA_EXSLB+EX_R9(r13)
665 ld r10,PACA_EXSLB+EX_R10(r13)
666 ld r11,PACA_EXSLB+EX_R11(r13)
667 ld r12,PACA_EXSLB+EX_R12(r13)
668 ld r13,PACA_EXSLB+EX_R13(r13)
669 RFI_TO_KERNEL
644 b . /* prevent speculative execution */ 670 b . /* prevent speculative execution */
645 671
672
6462: std r3,PACA_EXSLB+EX_DAR(r13) 6732: std r3,PACA_EXSLB+EX_DAR(r13)
647 mr r3,r12 674 mr r3,r12
648 mfspr r11,SPRN_SRR0 675 mfspr r11,SPRN_SRR0
@@ -651,7 +678,7 @@ END_MMU_FTR_SECTION_IFCLR(MMU_FTR_TYPE_RADIX)
651 mtspr SPRN_SRR0,r10 678 mtspr SPRN_SRR0,r10
652 ld r10,PACAKMSR(r13) 679 ld r10,PACAKMSR(r13)
653 mtspr SPRN_SRR1,r10 680 mtspr SPRN_SRR1,r10
654 rfid 681 RFI_TO_KERNEL
655 b . 682 b .
656 683
6578: std r3,PACA_EXSLB+EX_DAR(r13) 6848: std r3,PACA_EXSLB+EX_DAR(r13)
@@ -662,7 +689,7 @@ END_MMU_FTR_SECTION_IFCLR(MMU_FTR_TYPE_RADIX)
662 mtspr SPRN_SRR0,r10 689 mtspr SPRN_SRR0,r10
663 ld r10,PACAKMSR(r13) 690 ld r10,PACAKMSR(r13)
664 mtspr SPRN_SRR1,r10 691 mtspr SPRN_SRR1,r10
665 rfid 692 RFI_TO_KERNEL
666 b . 693 b .
667 694
668EXC_COMMON_BEGIN(unrecov_slb) 695EXC_COMMON_BEGIN(unrecov_slb)
@@ -901,7 +928,7 @@ EXC_COMMON(trap_0b_common, 0xb00, unknown_exception)
901 mtspr SPRN_SRR0,r10 ; \ 928 mtspr SPRN_SRR0,r10 ; \
902 ld r10,PACAKMSR(r13) ; \ 929 ld r10,PACAKMSR(r13) ; \
903 mtspr SPRN_SRR1,r10 ; \ 930 mtspr SPRN_SRR1,r10 ; \
904 rfid ; \ 931 RFI_TO_KERNEL ; \
905 b . ; /* prevent speculative execution */ 932 b . ; /* prevent speculative execution */
906 933
907#ifdef CONFIG_PPC_FAST_ENDIAN_SWITCH 934#ifdef CONFIG_PPC_FAST_ENDIAN_SWITCH
@@ -917,7 +944,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_REAL_LE) \
917 xori r12,r12,MSR_LE ; \ 944 xori r12,r12,MSR_LE ; \
918 mtspr SPRN_SRR1,r12 ; \ 945 mtspr SPRN_SRR1,r12 ; \
919 mr r13,r9 ; \ 946 mr r13,r9 ; \
920 rfid ; /* return to userspace */ \ 947 RFI_TO_USER ; /* return to userspace */ \
921 b . ; /* prevent speculative execution */ 948 b . ; /* prevent speculative execution */
922#else 949#else
923#define SYSCALL_FASTENDIAN_TEST 950#define SYSCALL_FASTENDIAN_TEST
@@ -1063,7 +1090,7 @@ TRAMP_REAL_BEGIN(hmi_exception_early)
1063 mtcr r11 1090 mtcr r11
1064 REST_GPR(11, r1) 1091 REST_GPR(11, r1)
1065 ld r1,GPR1(r1) 1092 ld r1,GPR1(r1)
1066 hrfid 1093 HRFI_TO_USER_OR_KERNEL
1067 1094
10681: mtcr r11 10951: mtcr r11
1069 REST_GPR(11, r1) 1096 REST_GPR(11, r1)
@@ -1314,7 +1341,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_CFAR)
1314 ld r11,PACA_EXGEN+EX_R11(r13) 1341 ld r11,PACA_EXGEN+EX_R11(r13)
1315 ld r12,PACA_EXGEN+EX_R12(r13) 1342 ld r12,PACA_EXGEN+EX_R12(r13)
1316 ld r13,PACA_EXGEN+EX_R13(r13) 1343 ld r13,PACA_EXGEN+EX_R13(r13)
1317 HRFID 1344 HRFI_TO_UNKNOWN
1318 b . 1345 b .
1319#endif 1346#endif
1320 1347
@@ -1418,10 +1445,94 @@ masked_##_H##interrupt: \
1418 ld r10,PACA_EXGEN+EX_R10(r13); \ 1445 ld r10,PACA_EXGEN+EX_R10(r13); \
1419 ld r11,PACA_EXGEN+EX_R11(r13); \ 1446 ld r11,PACA_EXGEN+EX_R11(r13); \
1420 /* returns to kernel where r13 must be set up, so don't restore it */ \ 1447 /* returns to kernel where r13 must be set up, so don't restore it */ \
1421 ##_H##rfid; \ 1448 ##_H##RFI_TO_KERNEL; \
1422 b .; \ 1449 b .; \
1423 MASKED_DEC_HANDLER(_H) 1450 MASKED_DEC_HANDLER(_H)
1424 1451
1452TRAMP_REAL_BEGIN(rfi_flush_fallback)
1453 SET_SCRATCH0(r13);
1454 GET_PACA(r13);
1455 std r9,PACA_EXRFI+EX_R9(r13)
1456 std r10,PACA_EXRFI+EX_R10(r13)
1457 std r11,PACA_EXRFI+EX_R11(r13)
1458 std r12,PACA_EXRFI+EX_R12(r13)
1459 std r8,PACA_EXRFI+EX_R13(r13)
1460 mfctr r9
1461 ld r10,PACA_RFI_FLUSH_FALLBACK_AREA(r13)
1462 ld r11,PACA_L1D_FLUSH_SETS(r13)
1463 ld r12,PACA_L1D_FLUSH_CONGRUENCE(r13)
1464 /*
1465 * The load adresses are at staggered offsets within cachelines,
1466 * which suits some pipelines better (on others it should not
1467 * hurt).
1468 */
1469 addi r12,r12,8
1470 mtctr r11
1471 DCBT_STOP_ALL_STREAM_IDS(r11) /* Stop prefetch streams */
1472
1473 /* order ld/st prior to dcbt stop all streams with flushing */
1474 sync
14751: li r8,0
1476 .rept 8 /* 8-way set associative */
1477 ldx r11,r10,r8
1478 add r8,r8,r12
1479 xor r11,r11,r11 // Ensure r11 is 0 even if fallback area is not
1480 add r8,r8,r11 // Add 0, this creates a dependency on the ldx
1481 .endr
1482 addi r10,r10,128 /* 128 byte cache line */
1483 bdnz 1b
1484
1485 mtctr r9
1486 ld r9,PACA_EXRFI+EX_R9(r13)
1487 ld r10,PACA_EXRFI+EX_R10(r13)
1488 ld r11,PACA_EXRFI+EX_R11(r13)
1489 ld r12,PACA_EXRFI+EX_R12(r13)
1490 ld r8,PACA_EXRFI+EX_R13(r13)
1491 GET_SCRATCH0(r13);
1492 rfid
1493
1494TRAMP_REAL_BEGIN(hrfi_flush_fallback)
1495 SET_SCRATCH0(r13);
1496 GET_PACA(r13);
1497 std r9,PACA_EXRFI+EX_R9(r13)
1498 std r10,PACA_EXRFI+EX_R10(r13)
1499 std r11,PACA_EXRFI+EX_R11(r13)
1500 std r12,PACA_EXRFI+EX_R12(r13)
1501 std r8,PACA_EXRFI+EX_R13(r13)
1502 mfctr r9
1503 ld r10,PACA_RFI_FLUSH_FALLBACK_AREA(r13)
1504 ld r11,PACA_L1D_FLUSH_SETS(r13)
1505 ld r12,PACA_L1D_FLUSH_CONGRUENCE(r13)
1506 /*
1507 * The load adresses are at staggered offsets within cachelines,
1508 * which suits some pipelines better (on others it should not
1509 * hurt).
1510 */
1511 addi r12,r12,8
1512 mtctr r11
1513 DCBT_STOP_ALL_STREAM_IDS(r11) /* Stop prefetch streams */
1514
1515 /* order ld/st prior to dcbt stop all streams with flushing */
1516 sync
15171: li r8,0
1518 .rept 8 /* 8-way set associative */
1519 ldx r11,r10,r8
1520 add r8,r8,r12
1521 xor r11,r11,r11 // Ensure r11 is 0 even if fallback area is not
1522 add r8,r8,r11 // Add 0, this creates a dependency on the ldx
1523 .endr
1524 addi r10,r10,128 /* 128 byte cache line */
1525 bdnz 1b
1526
1527 mtctr r9
1528 ld r9,PACA_EXRFI+EX_R9(r13)
1529 ld r10,PACA_EXRFI+EX_R10(r13)
1530 ld r11,PACA_EXRFI+EX_R11(r13)
1531 ld r12,PACA_EXRFI+EX_R12(r13)
1532 ld r8,PACA_EXRFI+EX_R13(r13)
1533 GET_SCRATCH0(r13);
1534 hrfid
1535
1425/* 1536/*
1426 * Real mode exceptions actually use this too, but alternate 1537 * Real mode exceptions actually use this too, but alternate
1427 * instruction code patches (which end up in the common .text area) 1538 * instruction code patches (which end up in the common .text area)
@@ -1441,7 +1552,7 @@ TRAMP_REAL_BEGIN(kvmppc_skip_interrupt)
1441 addi r13, r13, 4 1552 addi r13, r13, 4
1442 mtspr SPRN_SRR0, r13 1553 mtspr SPRN_SRR0, r13
1443 GET_SCRATCH0(r13) 1554 GET_SCRATCH0(r13)
1444 rfid 1555 RFI_TO_KERNEL
1445 b . 1556 b .
1446 1557
1447TRAMP_REAL_BEGIN(kvmppc_skip_Hinterrupt) 1558TRAMP_REAL_BEGIN(kvmppc_skip_Hinterrupt)
@@ -1453,7 +1564,7 @@ TRAMP_REAL_BEGIN(kvmppc_skip_Hinterrupt)
1453 addi r13, r13, 4 1564 addi r13, r13, 4
1454 mtspr SPRN_HSRR0, r13 1565 mtspr SPRN_HSRR0, r13
1455 GET_SCRATCH0(r13) 1566 GET_SCRATCH0(r13)
1456 hrfid 1567 HRFI_TO_KERNEL
1457 b . 1568 b .
1458#endif 1569#endif
1459 1570
diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c
index 8956a9856604..491be4179ddd 100644
--- a/arch/powerpc/kernel/setup_64.c
+++ b/arch/powerpc/kernel/setup_64.c
@@ -801,3 +801,104 @@ static int __init disable_hardlockup_detector(void)
801 return 0; 801 return 0;
802} 802}
803early_initcall(disable_hardlockup_detector); 803early_initcall(disable_hardlockup_detector);
804
805#ifdef CONFIG_PPC_BOOK3S_64
806static enum l1d_flush_type enabled_flush_types;
807static void *l1d_flush_fallback_area;
808static bool no_rfi_flush;
809bool rfi_flush;
810
811static int __init handle_no_rfi_flush(char *p)
812{
813 pr_info("rfi-flush: disabled on command line.");
814 no_rfi_flush = true;
815 return 0;
816}
817early_param("no_rfi_flush", handle_no_rfi_flush);
818
819/*
820 * The RFI flush is not KPTI, but because users will see doco that says to use
821 * nopti we hijack that option here to also disable the RFI flush.
822 */
823static int __init handle_no_pti(char *p)
824{
825 pr_info("rfi-flush: disabling due to 'nopti' on command line.\n");
826 handle_no_rfi_flush(NULL);
827 return 0;
828}
829early_param("nopti", handle_no_pti);
830
831static void do_nothing(void *unused)
832{
833 /*
834 * We don't need to do the flush explicitly, just enter+exit kernel is
835 * sufficient, the RFI exit handlers will do the right thing.
836 */
837}
838
839void rfi_flush_enable(bool enable)
840{
841 if (rfi_flush == enable)
842 return;
843
844 if (enable) {
845 do_rfi_flush_fixups(enabled_flush_types);
846 on_each_cpu(do_nothing, NULL, 1);
847 } else
848 do_rfi_flush_fixups(L1D_FLUSH_NONE);
849
850 rfi_flush = enable;
851}
852
853static void init_fallback_flush(void)
854{
855 u64 l1d_size, limit;
856 int cpu;
857
858 l1d_size = ppc64_caches.l1d.size;
859 limit = min(safe_stack_limit(), ppc64_rma_size);
860
861 /*
862 * Align to L1d size, and size it at 2x L1d size, to catch possible
863 * hardware prefetch runoff. We don't have a recipe for load patterns to
864 * reliably avoid the prefetcher.
865 */
866 l1d_flush_fallback_area = __va(memblock_alloc_base(l1d_size * 2, l1d_size, limit));
867 memset(l1d_flush_fallback_area, 0, l1d_size * 2);
868
869 for_each_possible_cpu(cpu) {
870 /*
871 * The fallback flush is currently coded for 8-way
872 * associativity. Different associativity is possible, but it
873 * will be treated as 8-way and may not evict the lines as
874 * effectively.
875 *
876 * 128 byte lines are mandatory.
877 */
878 u64 c = l1d_size / 8;
879
880 paca[cpu].rfi_flush_fallback_area = l1d_flush_fallback_area;
881 paca[cpu].l1d_flush_congruence = c;
882 paca[cpu].l1d_flush_sets = c / 128;
883 }
884}
885
886void __init setup_rfi_flush(enum l1d_flush_type types, bool enable)
887{
888 if (types & L1D_FLUSH_FALLBACK) {
889 pr_info("rfi-flush: Using fallback displacement flush\n");
890 init_fallback_flush();
891 }
892
893 if (types & L1D_FLUSH_ORI)
894 pr_info("rfi-flush: Using ori type flush\n");
895
896 if (types & L1D_FLUSH_MTTRIG)
897 pr_info("rfi-flush: Using mttrig type flush\n");
898
899 enabled_flush_types = types;
900
901 if (!no_rfi_flush)
902 rfi_flush_enable(enable);
903}
904#endif /* CONFIG_PPC_BOOK3S_64 */
diff --git a/arch/powerpc/kernel/vmlinux.lds.S b/arch/powerpc/kernel/vmlinux.lds.S
index 0494e1566ee2..307843d23682 100644
--- a/arch/powerpc/kernel/vmlinux.lds.S
+++ b/arch/powerpc/kernel/vmlinux.lds.S
@@ -132,6 +132,15 @@ SECTIONS
132 /* Read-only data */ 132 /* Read-only data */
133 RO_DATA(PAGE_SIZE) 133 RO_DATA(PAGE_SIZE)
134 134
135#ifdef CONFIG_PPC64
136 . = ALIGN(8);
137 __rfi_flush_fixup : AT(ADDR(__rfi_flush_fixup) - LOAD_OFFSET) {
138 __start___rfi_flush_fixup = .;
139 *(__rfi_flush_fixup)
140 __stop___rfi_flush_fixup = .;
141 }
142#endif
143
135 EXCEPTION_TABLE(0) 144 EXCEPTION_TABLE(0)
136 145
137 NOTES :kernel :notes 146 NOTES :kernel :notes
diff --git a/arch/powerpc/kvm/book3s_64_mmu.c b/arch/powerpc/kvm/book3s_64_mmu.c
index 29ebe2fd5867..a93d719edc90 100644
--- a/arch/powerpc/kvm/book3s_64_mmu.c
+++ b/arch/powerpc/kvm/book3s_64_mmu.c
@@ -235,6 +235,7 @@ static int kvmppc_mmu_book3s_64_xlate(struct kvm_vcpu *vcpu, gva_t eaddr,
235 gpte->may_read = true; 235 gpte->may_read = true;
236 gpte->may_write = true; 236 gpte->may_write = true;
237 gpte->page_size = MMU_PAGE_4K; 237 gpte->page_size = MMU_PAGE_4K;
238 gpte->wimg = HPTE_R_M;
238 239
239 return 0; 240 return 0;
240 } 241 }
diff --git a/arch/powerpc/kvm/book3s_64_mmu_hv.c b/arch/powerpc/kvm/book3s_64_mmu_hv.c
index 966097232d21..b73dbc9e797d 100644
--- a/arch/powerpc/kvm/book3s_64_mmu_hv.c
+++ b/arch/powerpc/kvm/book3s_64_mmu_hv.c
@@ -65,11 +65,17 @@ struct kvm_resize_hpt {
65 u32 order; 65 u32 order;
66 66
67 /* These fields protected by kvm->lock */ 67 /* These fields protected by kvm->lock */
68
69 /* Possible values and their usage:
70 * <0 an error occurred during allocation,
71 * -EBUSY allocation is in the progress,
72 * 0 allocation made successfuly.
73 */
68 int error; 74 int error;
69 bool prepare_done;
70 75
71 /* Private to the work thread, until prepare_done is true, 76 /* Private to the work thread, until error != -EBUSY,
72 * then protected by kvm->resize_hpt_sem */ 77 * then protected by kvm->lock.
78 */
73 struct kvm_hpt_info hpt; 79 struct kvm_hpt_info hpt;
74}; 80};
75 81
@@ -159,8 +165,6 @@ long kvmppc_alloc_reset_hpt(struct kvm *kvm, int order)
159 * Reset all the reverse-mapping chains for all memslots 165 * Reset all the reverse-mapping chains for all memslots
160 */ 166 */
161 kvmppc_rmap_reset(kvm); 167 kvmppc_rmap_reset(kvm);
162 /* Ensure that each vcpu will flush its TLB on next entry. */
163 cpumask_setall(&kvm->arch.need_tlb_flush);
164 err = 0; 168 err = 0;
165 goto out; 169 goto out;
166 } 170 }
@@ -176,6 +180,10 @@ long kvmppc_alloc_reset_hpt(struct kvm *kvm, int order)
176 kvmppc_set_hpt(kvm, &info); 180 kvmppc_set_hpt(kvm, &info);
177 181
178out: 182out:
183 if (err == 0)
184 /* Ensure that each vcpu will flush its TLB on next entry. */
185 cpumask_setall(&kvm->arch.need_tlb_flush);
186
179 mutex_unlock(&kvm->lock); 187 mutex_unlock(&kvm->lock);
180 return err; 188 return err;
181} 189}
@@ -1413,16 +1421,20 @@ static void resize_hpt_pivot(struct kvm_resize_hpt *resize)
1413 1421
1414static void resize_hpt_release(struct kvm *kvm, struct kvm_resize_hpt *resize) 1422static void resize_hpt_release(struct kvm *kvm, struct kvm_resize_hpt *resize)
1415{ 1423{
1416 BUG_ON(kvm->arch.resize_hpt != resize); 1424 if (WARN_ON(!mutex_is_locked(&kvm->lock)))
1425 return;
1417 1426
1418 if (!resize) 1427 if (!resize)
1419 return; 1428 return;
1420 1429
1421 if (resize->hpt.virt) 1430 if (resize->error != -EBUSY) {
1422 kvmppc_free_hpt(&resize->hpt); 1431 if (resize->hpt.virt)
1432 kvmppc_free_hpt(&resize->hpt);
1433 kfree(resize);
1434 }
1423 1435
1424 kvm->arch.resize_hpt = NULL; 1436 if (kvm->arch.resize_hpt == resize)
1425 kfree(resize); 1437 kvm->arch.resize_hpt = NULL;
1426} 1438}
1427 1439
1428static void resize_hpt_prepare_work(struct work_struct *work) 1440static void resize_hpt_prepare_work(struct work_struct *work)
@@ -1431,17 +1443,41 @@ static void resize_hpt_prepare_work(struct work_struct *work)
1431 struct kvm_resize_hpt, 1443 struct kvm_resize_hpt,
1432 work); 1444 work);
1433 struct kvm *kvm = resize->kvm; 1445 struct kvm *kvm = resize->kvm;
1434 int err; 1446 int err = 0;
1435 1447
1436 resize_hpt_debug(resize, "resize_hpt_prepare_work(): order = %d\n", 1448 if (WARN_ON(resize->error != -EBUSY))
1437 resize->order); 1449 return;
1438
1439 err = resize_hpt_allocate(resize);
1440 1450
1441 mutex_lock(&kvm->lock); 1451 mutex_lock(&kvm->lock);
1442 1452
1453 /* Request is still current? */
1454 if (kvm->arch.resize_hpt == resize) {
1455 /* We may request large allocations here:
1456 * do not sleep with kvm->lock held for a while.
1457 */
1458 mutex_unlock(&kvm->lock);
1459
1460 resize_hpt_debug(resize, "resize_hpt_prepare_work(): order = %d\n",
1461 resize->order);
1462
1463 err = resize_hpt_allocate(resize);
1464
1465 /* We have strict assumption about -EBUSY
1466 * when preparing for HPT resize.
1467 */
1468 if (WARN_ON(err == -EBUSY))
1469 err = -EINPROGRESS;
1470
1471 mutex_lock(&kvm->lock);
1472 /* It is possible that kvm->arch.resize_hpt != resize
1473 * after we grab kvm->lock again.
1474 */
1475 }
1476
1443 resize->error = err; 1477 resize->error = err;
1444 resize->prepare_done = true; 1478
1479 if (kvm->arch.resize_hpt != resize)
1480 resize_hpt_release(kvm, resize);
1445 1481
1446 mutex_unlock(&kvm->lock); 1482 mutex_unlock(&kvm->lock);
1447} 1483}
@@ -1466,14 +1502,12 @@ long kvm_vm_ioctl_resize_hpt_prepare(struct kvm *kvm,
1466 1502
1467 if (resize) { 1503 if (resize) {
1468 if (resize->order == shift) { 1504 if (resize->order == shift) {
1469 /* Suitable resize in progress */ 1505 /* Suitable resize in progress? */
1470 if (resize->prepare_done) { 1506 ret = resize->error;
1471 ret = resize->error; 1507 if (ret == -EBUSY)
1472 if (ret != 0)
1473 resize_hpt_release(kvm, resize);
1474 } else {
1475 ret = 100; /* estimated time in ms */ 1508 ret = 100; /* estimated time in ms */
1476 } 1509 else if (ret)
1510 resize_hpt_release(kvm, resize);
1477 1511
1478 goto out; 1512 goto out;
1479 } 1513 }
@@ -1493,6 +1527,8 @@ long kvm_vm_ioctl_resize_hpt_prepare(struct kvm *kvm,
1493 ret = -ENOMEM; 1527 ret = -ENOMEM;
1494 goto out; 1528 goto out;
1495 } 1529 }
1530
1531 resize->error = -EBUSY;
1496 resize->order = shift; 1532 resize->order = shift;
1497 resize->kvm = kvm; 1533 resize->kvm = kvm;
1498 INIT_WORK(&resize->work, resize_hpt_prepare_work); 1534 INIT_WORK(&resize->work, resize_hpt_prepare_work);
@@ -1547,16 +1583,12 @@ long kvm_vm_ioctl_resize_hpt_commit(struct kvm *kvm,
1547 if (!resize || (resize->order != shift)) 1583 if (!resize || (resize->order != shift))
1548 goto out; 1584 goto out;
1549 1585
1550 ret = -EBUSY;
1551 if (!resize->prepare_done)
1552 goto out;
1553
1554 ret = resize->error; 1586 ret = resize->error;
1555 if (ret != 0) 1587 if (ret)
1556 goto out; 1588 goto out;
1557 1589
1558 ret = resize_hpt_rehash(resize); 1590 ret = resize_hpt_rehash(resize);
1559 if (ret != 0) 1591 if (ret)
1560 goto out; 1592 goto out;
1561 1593
1562 resize_hpt_pivot(resize); 1594 resize_hpt_pivot(resize);
diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
index 2659844784b8..9c61f736c75b 100644
--- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S
+++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
@@ -79,7 +79,7 @@ _GLOBAL_TOC(kvmppc_hv_entry_trampoline)
79 mtmsrd r0,1 /* clear RI in MSR */ 79 mtmsrd r0,1 /* clear RI in MSR */
80 mtsrr0 r5 80 mtsrr0 r5
81 mtsrr1 r6 81 mtsrr1 r6
82 RFI 82 RFI_TO_KERNEL
83 83
84kvmppc_call_hv_entry: 84kvmppc_call_hv_entry:
85BEGIN_FTR_SECTION 85BEGIN_FTR_SECTION
@@ -199,7 +199,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S)
199 mtmsrd r6, 1 /* Clear RI in MSR */ 199 mtmsrd r6, 1 /* Clear RI in MSR */
200 mtsrr0 r8 200 mtsrr0 r8
201 mtsrr1 r7 201 mtsrr1 r7
202 RFI 202 RFI_TO_KERNEL
203 203
204 /* Virtual-mode return */ 204 /* Virtual-mode return */
205.Lvirt_return: 205.Lvirt_return:
@@ -1167,8 +1167,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300)
1167 1167
1168 ld r0, VCPU_GPR(R0)(r4) 1168 ld r0, VCPU_GPR(R0)(r4)
1169 ld r4, VCPU_GPR(R4)(r4) 1169 ld r4, VCPU_GPR(R4)(r4)
1170 1170 HRFI_TO_GUEST
1171 hrfid
1172 b . 1171 b .
1173 1172
1174secondary_too_late: 1173secondary_too_late:
@@ -3320,7 +3319,7 @@ END_MMU_FTR_SECTION_IFSET(MMU_FTR_TYPE_RADIX)
3320 ld r4, PACAKMSR(r13) 3319 ld r4, PACAKMSR(r13)
3321 mtspr SPRN_SRR0, r3 3320 mtspr SPRN_SRR0, r3
3322 mtspr SPRN_SRR1, r4 3321 mtspr SPRN_SRR1, r4
3323 rfid 3322 RFI_TO_KERNEL
33249: addi r3, r1, STACK_FRAME_OVERHEAD 33239: addi r3, r1, STACK_FRAME_OVERHEAD
3325 bl kvmppc_bad_interrupt 3324 bl kvmppc_bad_interrupt
3326 b 9b 3325 b 9b
diff --git a/arch/powerpc/kvm/book3s_pr.c b/arch/powerpc/kvm/book3s_pr.c
index d0dc8624198f..7deaeeb14b93 100644
--- a/arch/powerpc/kvm/book3s_pr.c
+++ b/arch/powerpc/kvm/book3s_pr.c
@@ -60,6 +60,7 @@ static void kvmppc_giveup_fac(struct kvm_vcpu *vcpu, ulong fac);
60#define MSR_USER32 MSR_USER 60#define MSR_USER32 MSR_USER
61#define MSR_USER64 MSR_USER 61#define MSR_USER64 MSR_USER
62#define HW_PAGE_SIZE PAGE_SIZE 62#define HW_PAGE_SIZE PAGE_SIZE
63#define HPTE_R_M _PAGE_COHERENT
63#endif 64#endif
64 65
65static bool kvmppc_is_split_real(struct kvm_vcpu *vcpu) 66static bool kvmppc_is_split_real(struct kvm_vcpu *vcpu)
@@ -557,6 +558,7 @@ int kvmppc_handle_pagefault(struct kvm_run *run, struct kvm_vcpu *vcpu,
557 pte.eaddr = eaddr; 558 pte.eaddr = eaddr;
558 pte.vpage = eaddr >> 12; 559 pte.vpage = eaddr >> 12;
559 pte.page_size = MMU_PAGE_64K; 560 pte.page_size = MMU_PAGE_64K;
561 pte.wimg = HPTE_R_M;
560 } 562 }
561 563
562 switch (kvmppc_get_msr(vcpu) & (MSR_DR|MSR_IR)) { 564 switch (kvmppc_get_msr(vcpu) & (MSR_DR|MSR_IR)) {
diff --git a/arch/powerpc/kvm/book3s_rmhandlers.S b/arch/powerpc/kvm/book3s_rmhandlers.S
index 42a4b237df5f..34a5adeff084 100644
--- a/arch/powerpc/kvm/book3s_rmhandlers.S
+++ b/arch/powerpc/kvm/book3s_rmhandlers.S
@@ -46,6 +46,9 @@
46 46
47#define FUNC(name) name 47#define FUNC(name) name
48 48
49#define RFI_TO_KERNEL RFI
50#define RFI_TO_GUEST RFI
51
49.macro INTERRUPT_TRAMPOLINE intno 52.macro INTERRUPT_TRAMPOLINE intno
50 53
51.global kvmppc_trampoline_\intno 54.global kvmppc_trampoline_\intno
@@ -141,7 +144,7 @@ kvmppc_handler_skip_ins:
141 GET_SCRATCH0(r13) 144 GET_SCRATCH0(r13)
142 145
143 /* And get back into the code */ 146 /* And get back into the code */
144 RFI 147 RFI_TO_KERNEL
145#endif 148#endif
146 149
147/* 150/*
@@ -164,6 +167,6 @@ _GLOBAL_TOC(kvmppc_entry_trampoline)
164 ori r5, r5, MSR_EE 167 ori r5, r5, MSR_EE
165 mtsrr0 r7 168 mtsrr0 r7
166 mtsrr1 r6 169 mtsrr1 r6
167 RFI 170 RFI_TO_KERNEL
168 171
169#include "book3s_segment.S" 172#include "book3s_segment.S"
diff --git a/arch/powerpc/kvm/book3s_segment.S b/arch/powerpc/kvm/book3s_segment.S
index 2a2b96d53999..93a180ceefad 100644
--- a/arch/powerpc/kvm/book3s_segment.S
+++ b/arch/powerpc/kvm/book3s_segment.S
@@ -156,7 +156,7 @@ no_dcbz32_on:
156 PPC_LL r9, SVCPU_R9(r3) 156 PPC_LL r9, SVCPU_R9(r3)
157 PPC_LL r3, (SVCPU_R3)(r3) 157 PPC_LL r3, (SVCPU_R3)(r3)
158 158
159 RFI 159 RFI_TO_GUEST
160kvmppc_handler_trampoline_enter_end: 160kvmppc_handler_trampoline_enter_end:
161 161
162 162
@@ -407,5 +407,5 @@ END_FTR_SECTION_IFSET(CPU_FTR_HVMODE)
407 cmpwi r12, BOOK3S_INTERRUPT_DOORBELL 407 cmpwi r12, BOOK3S_INTERRUPT_DOORBELL
408 beqa BOOK3S_INTERRUPT_DOORBELL 408 beqa BOOK3S_INTERRUPT_DOORBELL
409 409
410 RFI 410 RFI_TO_KERNEL
411kvmppc_handler_trampoline_exit_end: 411kvmppc_handler_trampoline_exit_end:
diff --git a/arch/powerpc/lib/feature-fixups.c b/arch/powerpc/lib/feature-fixups.c
index 41cf5ae273cf..a95ea007d654 100644
--- a/arch/powerpc/lib/feature-fixups.c
+++ b/arch/powerpc/lib/feature-fixups.c
@@ -116,6 +116,47 @@ void do_feature_fixups(unsigned long value, void *fixup_start, void *fixup_end)
116 } 116 }
117} 117}
118 118
119#ifdef CONFIG_PPC_BOOK3S_64
120void do_rfi_flush_fixups(enum l1d_flush_type types)
121{
122 unsigned int instrs[3], *dest;
123 long *start, *end;
124 int i;
125
126 start = PTRRELOC(&__start___rfi_flush_fixup),
127 end = PTRRELOC(&__stop___rfi_flush_fixup);
128
129 instrs[0] = 0x60000000; /* nop */
130 instrs[1] = 0x60000000; /* nop */
131 instrs[2] = 0x60000000; /* nop */
132
133 if (types & L1D_FLUSH_FALLBACK)
134 /* b .+16 to fallback flush */
135 instrs[0] = 0x48000010;
136
137 i = 0;
138 if (types & L1D_FLUSH_ORI) {
139 instrs[i++] = 0x63ff0000; /* ori 31,31,0 speculation barrier */
140 instrs[i++] = 0x63de0000; /* ori 30,30,0 L1d flush*/
141 }
142
143 if (types & L1D_FLUSH_MTTRIG)
144 instrs[i++] = 0x7c12dba6; /* mtspr TRIG2,r0 (SPR #882) */
145
146 for (i = 0; start < end; start++, i++) {
147 dest = (void *)start + *start;
148
149 pr_devel("patching dest %lx\n", (unsigned long)dest);
150
151 patch_instruction(dest, instrs[0]);
152 patch_instruction(dest + 1, instrs[1]);
153 patch_instruction(dest + 2, instrs[2]);
154 }
155
156 printk(KERN_DEBUG "rfi-flush: patched %d locations\n", i);
157}
158#endif /* CONFIG_PPC_BOOK3S_64 */
159
119void do_lwsync_fixups(unsigned long value, void *fixup_start, void *fixup_end) 160void do_lwsync_fixups(unsigned long value, void *fixup_start, void *fixup_end)
120{ 161{
121 long *start, *end; 162 long *start, *end;
diff --git a/arch/powerpc/platforms/powernv/setup.c b/arch/powerpc/platforms/powernv/setup.c
index 1edfbc1e40f4..4fb21e17504a 100644
--- a/arch/powerpc/platforms/powernv/setup.c
+++ b/arch/powerpc/platforms/powernv/setup.c
@@ -37,13 +37,62 @@
37#include <asm/kexec.h> 37#include <asm/kexec.h>
38#include <asm/smp.h> 38#include <asm/smp.h>
39#include <asm/tm.h> 39#include <asm/tm.h>
40#include <asm/setup.h>
40 41
41#include "powernv.h" 42#include "powernv.h"
42 43
44static void pnv_setup_rfi_flush(void)
45{
46 struct device_node *np, *fw_features;
47 enum l1d_flush_type type;
48 int enable;
49
50 /* Default to fallback in case fw-features are not available */
51 type = L1D_FLUSH_FALLBACK;
52 enable = 1;
53
54 np = of_find_node_by_name(NULL, "ibm,opal");
55 fw_features = of_get_child_by_name(np, "fw-features");
56 of_node_put(np);
57
58 if (fw_features) {
59 np = of_get_child_by_name(fw_features, "inst-l1d-flush-trig2");
60 if (np && of_property_read_bool(np, "enabled"))
61 type = L1D_FLUSH_MTTRIG;
62
63 of_node_put(np);
64
65 np = of_get_child_by_name(fw_features, "inst-l1d-flush-ori30,30,0");
66 if (np && of_property_read_bool(np, "enabled"))
67 type = L1D_FLUSH_ORI;
68
69 of_node_put(np);
70
71 /* Enable unless firmware says NOT to */
72 enable = 2;
73 np = of_get_child_by_name(fw_features, "needs-l1d-flush-msr-hv-1-to-0");
74 if (np && of_property_read_bool(np, "disabled"))
75 enable--;
76
77 of_node_put(np);
78
79 np = of_get_child_by_name(fw_features, "needs-l1d-flush-msr-pr-0-to-1");
80 if (np && of_property_read_bool(np, "disabled"))
81 enable--;
82
83 of_node_put(np);
84 of_node_put(fw_features);
85 }
86
87 setup_rfi_flush(type, enable > 0);
88}
89
43static void __init pnv_setup_arch(void) 90static void __init pnv_setup_arch(void)
44{ 91{
45 set_arch_panic_timeout(10, ARCH_PANIC_TIMEOUT); 92 set_arch_panic_timeout(10, ARCH_PANIC_TIMEOUT);
46 93
94 pnv_setup_rfi_flush();
95
47 /* Initialize SMP */ 96 /* Initialize SMP */
48 pnv_smp_init(); 97 pnv_smp_init();
49 98
diff --git a/arch/powerpc/platforms/pseries/dlpar.c b/arch/powerpc/platforms/pseries/dlpar.c
index 6e35780c5962..a0b20c03f078 100644
--- a/arch/powerpc/platforms/pseries/dlpar.c
+++ b/arch/powerpc/platforms/pseries/dlpar.c
@@ -574,11 +574,26 @@ static ssize_t dlpar_show(struct class *class, struct class_attribute *attr,
574 574
575static CLASS_ATTR_RW(dlpar); 575static CLASS_ATTR_RW(dlpar);
576 576
577static int __init pseries_dlpar_init(void) 577int __init dlpar_workqueue_init(void)
578{ 578{
579 if (pseries_hp_wq)
580 return 0;
581
579 pseries_hp_wq = alloc_workqueue("pseries hotplug workqueue", 582 pseries_hp_wq = alloc_workqueue("pseries hotplug workqueue",
580 WQ_UNBOUND, 1); 583 WQ_UNBOUND, 1);
584
585 return pseries_hp_wq ? 0 : -ENOMEM;
586}
587
588static int __init dlpar_sysfs_init(void)
589{
590 int rc;
591
592 rc = dlpar_workqueue_init();
593 if (rc)
594 return rc;
595
581 return sysfs_create_file(kernel_kobj, &class_attr_dlpar.attr); 596 return sysfs_create_file(kernel_kobj, &class_attr_dlpar.attr);
582} 597}
583machine_device_initcall(pseries, pseries_dlpar_init); 598machine_device_initcall(pseries, dlpar_sysfs_init);
584 599
diff --git a/arch/powerpc/platforms/pseries/pseries.h b/arch/powerpc/platforms/pseries/pseries.h
index 4470a3194311..1ae1d9f4dbe9 100644
--- a/arch/powerpc/platforms/pseries/pseries.h
+++ b/arch/powerpc/platforms/pseries/pseries.h
@@ -98,4 +98,6 @@ static inline unsigned long cmo_get_page_size(void)
98 return CMO_PageSize; 98 return CMO_PageSize;
99} 99}
100 100
101int dlpar_workqueue_init(void);
102
101#endif /* _PSERIES_PSERIES_H */ 103#endif /* _PSERIES_PSERIES_H */
diff --git a/arch/powerpc/platforms/pseries/ras.c b/arch/powerpc/platforms/pseries/ras.c
index 4923ffe230cf..81d8614e7379 100644
--- a/arch/powerpc/platforms/pseries/ras.c
+++ b/arch/powerpc/platforms/pseries/ras.c
@@ -69,7 +69,8 @@ static int __init init_ras_IRQ(void)
69 /* Hotplug Events */ 69 /* Hotplug Events */
70 np = of_find_node_by_path("/event-sources/hot-plug-events"); 70 np = of_find_node_by_path("/event-sources/hot-plug-events");
71 if (np != NULL) { 71 if (np != NULL) {
72 request_event_sources_irqs(np, ras_hotplug_interrupt, 72 if (dlpar_workqueue_init() == 0)
73 request_event_sources_irqs(np, ras_hotplug_interrupt,
73 "RAS_HOTPLUG"); 74 "RAS_HOTPLUG");
74 of_node_put(np); 75 of_node_put(np);
75 } 76 }
diff --git a/arch/powerpc/platforms/pseries/setup.c b/arch/powerpc/platforms/pseries/setup.c
index a8531e012658..ae4f596273b5 100644
--- a/arch/powerpc/platforms/pseries/setup.c
+++ b/arch/powerpc/platforms/pseries/setup.c
@@ -459,6 +459,39 @@ static void __init find_and_init_phbs(void)
459 of_pci_check_probe_only(); 459 of_pci_check_probe_only();
460} 460}
461 461
462static void pseries_setup_rfi_flush(void)
463{
464 struct h_cpu_char_result result;
465 enum l1d_flush_type types;
466 bool enable;
467 long rc;
468
469 /* Enable by default */
470 enable = true;
471
472 rc = plpar_get_cpu_characteristics(&result);
473 if (rc == H_SUCCESS) {
474 types = L1D_FLUSH_NONE;
475
476 if (result.character & H_CPU_CHAR_L1D_FLUSH_TRIG2)
477 types |= L1D_FLUSH_MTTRIG;
478 if (result.character & H_CPU_CHAR_L1D_FLUSH_ORI30)
479 types |= L1D_FLUSH_ORI;
480
481 /* Use fallback if nothing set in hcall */
482 if (types == L1D_FLUSH_NONE)
483 types = L1D_FLUSH_FALLBACK;
484
485 if (!(result.behaviour & H_CPU_BEHAV_L1D_FLUSH_PR))
486 enable = false;
487 } else {
488 /* Default to fallback if case hcall is not available */
489 types = L1D_FLUSH_FALLBACK;
490 }
491
492 setup_rfi_flush(types, enable);
493}
494
462static void __init pSeries_setup_arch(void) 495static void __init pSeries_setup_arch(void)
463{ 496{
464 set_arch_panic_timeout(10, ARCH_PANIC_TIMEOUT); 497 set_arch_panic_timeout(10, ARCH_PANIC_TIMEOUT);
@@ -476,6 +509,8 @@ static void __init pSeries_setup_arch(void)
476 509
477 fwnmi_init(); 510 fwnmi_init();
478 511
512 pseries_setup_rfi_flush();
513
479 /* By default, only probe PCI (can be overridden by rtas_pci) */ 514 /* By default, only probe PCI (can be overridden by rtas_pci) */
480 pci_add_flags(PCI_PROBE_ONLY); 515 pci_add_flags(PCI_PROBE_ONLY);
481 516
diff --git a/arch/riscv/configs/defconfig b/arch/riscv/configs/defconfig
index e69de29bb2d1..47dacf06c679 100644
--- a/arch/riscv/configs/defconfig
+++ b/arch/riscv/configs/defconfig
@@ -0,0 +1,75 @@
1CONFIG_SMP=y
2CONFIG_PCI=y
3CONFIG_PCIE_XILINX=y
4CONFIG_SYSVIPC=y
5CONFIG_POSIX_MQUEUE=y
6CONFIG_IKCONFIG=y
7CONFIG_IKCONFIG_PROC=y
8CONFIG_CGROUPS=y
9CONFIG_CGROUP_SCHED=y
10CONFIG_CFS_BANDWIDTH=y
11CONFIG_CGROUP_BPF=y
12CONFIG_NAMESPACES=y
13CONFIG_USER_NS=y
14CONFIG_BLK_DEV_INITRD=y
15CONFIG_EXPERT=y
16CONFIG_CHECKPOINT_RESTORE=y
17CONFIG_BPF_SYSCALL=y
18CONFIG_NET=y
19CONFIG_PACKET=y
20CONFIG_UNIX=y
21CONFIG_INET=y
22CONFIG_IP_MULTICAST=y
23CONFIG_IP_ADVANCED_ROUTER=y
24CONFIG_IP_PNP=y
25CONFIG_IP_PNP_DHCP=y
26CONFIG_IP_PNP_BOOTP=y
27CONFIG_IP_PNP_RARP=y
28CONFIG_NETLINK_DIAG=y
29CONFIG_DEVTMPFS=y
30CONFIG_BLK_DEV_LOOP=y
31CONFIG_VIRTIO_BLK=y
32CONFIG_BLK_DEV_SD=y
33CONFIG_BLK_DEV_SR=y
34CONFIG_ATA=y
35CONFIG_SATA_AHCI=y
36CONFIG_SATA_AHCI_PLATFORM=y
37CONFIG_NETDEVICES=y
38CONFIG_VIRTIO_NET=y
39CONFIG_MACB=y
40CONFIG_E1000E=y
41CONFIG_R8169=y
42CONFIG_MICROSEMI_PHY=y
43CONFIG_INPUT_MOUSEDEV=y
44CONFIG_SERIAL_8250=y
45CONFIG_SERIAL_8250_CONSOLE=y
46CONFIG_SERIAL_OF_PLATFORM=y
47# CONFIG_PTP_1588_CLOCK is not set
48CONFIG_DRM=y
49CONFIG_DRM_RADEON=y
50CONFIG_FRAMEBUFFER_CONSOLE=y
51CONFIG_USB=y
52CONFIG_USB_XHCI_HCD=y
53CONFIG_USB_XHCI_PLATFORM=y
54CONFIG_USB_EHCI_HCD=y
55CONFIG_USB_EHCI_HCD_PLATFORM=y
56CONFIG_USB_OHCI_HCD=y
57CONFIG_USB_OHCI_HCD_PLATFORM=y
58CONFIG_USB_STORAGE=y
59CONFIG_USB_UAS=y
60CONFIG_VIRTIO_MMIO=y
61CONFIG_RAS=y
62CONFIG_EXT4_FS=y
63CONFIG_EXT4_FS_POSIX_ACL=y
64CONFIG_AUTOFS4_FS=y
65CONFIG_MSDOS_FS=y
66CONFIG_VFAT_FS=y
67CONFIG_TMPFS=y
68CONFIG_TMPFS_POSIX_ACL=y
69CONFIG_NFS_FS=y
70CONFIG_NFS_V4=y
71CONFIG_NFS_V4_1=y
72CONFIG_NFS_V4_2=y
73CONFIG_ROOT_NFS=y
74# CONFIG_RCU_TRACE is not set
75CONFIG_CRYPTO_USER_API_HASH=y
diff --git a/arch/riscv/include/asm/csr.h b/arch/riscv/include/asm/csr.h
index 0d64bc9f4f91..3c7a2c97e377 100644
--- a/arch/riscv/include/asm/csr.h
+++ b/arch/riscv/include/asm/csr.h
@@ -17,10 +17,10 @@
17#include <linux/const.h> 17#include <linux/const.h>
18 18
19/* Status register flags */ 19/* Status register flags */
20#define SR_IE _AC(0x00000002, UL) /* Interrupt Enable */ 20#define SR_SIE _AC(0x00000002, UL) /* Supervisor Interrupt Enable */
21#define SR_PIE _AC(0x00000020, UL) /* Previous IE */ 21#define SR_SPIE _AC(0x00000020, UL) /* Previous Supervisor IE */
22#define SR_PS _AC(0x00000100, UL) /* Previously Supervisor */ 22#define SR_SPP _AC(0x00000100, UL) /* Previously Supervisor */
23#define SR_SUM _AC(0x00040000, UL) /* Supervisor may access User Memory */ 23#define SR_SUM _AC(0x00040000, UL) /* Supervisor may access User Memory */
24 24
25#define SR_FS _AC(0x00006000, UL) /* Floating-point Status */ 25#define SR_FS _AC(0x00006000, UL) /* Floating-point Status */
26#define SR_FS_OFF _AC(0x00000000, UL) 26#define SR_FS_OFF _AC(0x00000000, UL)
diff --git a/arch/riscv/include/asm/io.h b/arch/riscv/include/asm/io.h
index a82ce599b639..b269451e7e85 100644
--- a/arch/riscv/include/asm/io.h
+++ b/arch/riscv/include/asm/io.h
@@ -21,8 +21,6 @@
21 21
22#include <linux/types.h> 22#include <linux/types.h>
23 23
24#ifdef CONFIG_MMU
25
26extern void __iomem *ioremap(phys_addr_t offset, unsigned long size); 24extern void __iomem *ioremap(phys_addr_t offset, unsigned long size);
27 25
28/* 26/*
@@ -36,8 +34,6 @@ extern void __iomem *ioremap(phys_addr_t offset, unsigned long size);
36 34
37extern void iounmap(volatile void __iomem *addr); 35extern void iounmap(volatile void __iomem *addr);
38 36
39#endif /* CONFIG_MMU */
40
41/* Generic IO read/write. These perform native-endian accesses. */ 37/* Generic IO read/write. These perform native-endian accesses. */
42#define __raw_writeb __raw_writeb 38#define __raw_writeb __raw_writeb
43static inline void __raw_writeb(u8 val, volatile void __iomem *addr) 39static inline void __raw_writeb(u8 val, volatile void __iomem *addr)
diff --git a/arch/riscv/include/asm/irqflags.h b/arch/riscv/include/asm/irqflags.h
index 6fdc860d7f84..07a3c6d5706f 100644
--- a/arch/riscv/include/asm/irqflags.h
+++ b/arch/riscv/include/asm/irqflags.h
@@ -27,25 +27,25 @@ static inline unsigned long arch_local_save_flags(void)
27/* unconditionally enable interrupts */ 27/* unconditionally enable interrupts */
28static inline void arch_local_irq_enable(void) 28static inline void arch_local_irq_enable(void)
29{ 29{
30 csr_set(sstatus, SR_IE); 30 csr_set(sstatus, SR_SIE);
31} 31}
32 32
33/* unconditionally disable interrupts */ 33/* unconditionally disable interrupts */
34static inline void arch_local_irq_disable(void) 34static inline void arch_local_irq_disable(void)
35{ 35{
36 csr_clear(sstatus, SR_IE); 36 csr_clear(sstatus, SR_SIE);
37} 37}
38 38
39/* get status and disable interrupts */ 39/* get status and disable interrupts */
40static inline unsigned long arch_local_irq_save(void) 40static inline unsigned long arch_local_irq_save(void)
41{ 41{
42 return csr_read_clear(sstatus, SR_IE); 42 return csr_read_clear(sstatus, SR_SIE);
43} 43}
44 44
45/* test flags */ 45/* test flags */
46static inline int arch_irqs_disabled_flags(unsigned long flags) 46static inline int arch_irqs_disabled_flags(unsigned long flags)
47{ 47{
48 return !(flags & SR_IE); 48 return !(flags & SR_SIE);
49} 49}
50 50
51/* test hardware interrupt enable bit */ 51/* test hardware interrupt enable bit */
@@ -57,7 +57,7 @@ static inline int arch_irqs_disabled(void)
57/* set interrupt enabled status */ 57/* set interrupt enabled status */
58static inline void arch_local_irq_restore(unsigned long flags) 58static inline void arch_local_irq_restore(unsigned long flags)
59{ 59{
60 csr_set(sstatus, flags & SR_IE); 60 csr_set(sstatus, flags & SR_SIE);
61} 61}
62 62
63#endif /* _ASM_RISCV_IRQFLAGS_H */ 63#endif /* _ASM_RISCV_IRQFLAGS_H */
diff --git a/arch/riscv/include/asm/pgtable.h b/arch/riscv/include/asm/pgtable.h
index 2cbd92ed1629..16301966d65b 100644
--- a/arch/riscv/include/asm/pgtable.h
+++ b/arch/riscv/include/asm/pgtable.h
@@ -20,8 +20,6 @@
20 20
21#ifndef __ASSEMBLY__ 21#ifndef __ASSEMBLY__
22 22
23#ifdef CONFIG_MMU
24
25/* Page Upper Directory not used in RISC-V */ 23/* Page Upper Directory not used in RISC-V */
26#include <asm-generic/pgtable-nopud.h> 24#include <asm-generic/pgtable-nopud.h>
27#include <asm/page.h> 25#include <asm/page.h>
@@ -413,8 +411,6 @@ static inline void pgtable_cache_init(void)
413 /* No page table caches to initialize */ 411 /* No page table caches to initialize */
414} 412}
415 413
416#endif /* CONFIG_MMU */
417
418#define VMALLOC_SIZE (KERN_VIRT_SIZE >> 1) 414#define VMALLOC_SIZE (KERN_VIRT_SIZE >> 1)
419#define VMALLOC_END (PAGE_OFFSET - 1) 415#define VMALLOC_END (PAGE_OFFSET - 1)
420#define VMALLOC_START (PAGE_OFFSET - VMALLOC_SIZE) 416#define VMALLOC_START (PAGE_OFFSET - VMALLOC_SIZE)
diff --git a/arch/riscv/include/asm/ptrace.h b/arch/riscv/include/asm/ptrace.h
index 93b8956e25e4..2c5df945d43c 100644
--- a/arch/riscv/include/asm/ptrace.h
+++ b/arch/riscv/include/asm/ptrace.h
@@ -66,7 +66,7 @@ struct pt_regs {
66#define REG_FMT "%08lx" 66#define REG_FMT "%08lx"
67#endif 67#endif
68 68
69#define user_mode(regs) (((regs)->sstatus & SR_PS) == 0) 69#define user_mode(regs) (((regs)->sstatus & SR_SPP) == 0)
70 70
71 71
72/* Helpers for working with the instruction pointer */ 72/* Helpers for working with the instruction pointer */
diff --git a/arch/riscv/include/asm/tlbflush.h b/arch/riscv/include/asm/tlbflush.h
index 715b0f10af58..7b9c24ebdf52 100644
--- a/arch/riscv/include/asm/tlbflush.h
+++ b/arch/riscv/include/asm/tlbflush.h
@@ -15,8 +15,6 @@
15#ifndef _ASM_RISCV_TLBFLUSH_H 15#ifndef _ASM_RISCV_TLBFLUSH_H
16#define _ASM_RISCV_TLBFLUSH_H 16#define _ASM_RISCV_TLBFLUSH_H
17 17
18#ifdef CONFIG_MMU
19
20#include <linux/mm_types.h> 18#include <linux/mm_types.h>
21 19
22/* 20/*
@@ -64,6 +62,4 @@ static inline void flush_tlb_kernel_range(unsigned long start,
64 flush_tlb_all(); 62 flush_tlb_all();
65} 63}
66 64
67#endif /* CONFIG_MMU */
68
69#endif /* _ASM_RISCV_TLBFLUSH_H */ 65#endif /* _ASM_RISCV_TLBFLUSH_H */
diff --git a/arch/riscv/include/asm/uaccess.h b/arch/riscv/include/asm/uaccess.h
index 27b90d64814b..14b0b22fb578 100644
--- a/arch/riscv/include/asm/uaccess.h
+++ b/arch/riscv/include/asm/uaccess.h
@@ -127,7 +127,6 @@ extern int fixup_exception(struct pt_regs *state);
127 * call. 127 * call.
128 */ 128 */
129 129
130#ifdef CONFIG_MMU
131#define __get_user_asm(insn, x, ptr, err) \ 130#define __get_user_asm(insn, x, ptr, err) \
132do { \ 131do { \
133 uintptr_t __tmp; \ 132 uintptr_t __tmp; \
@@ -153,13 +152,11 @@ do { \
153 __disable_user_access(); \ 152 __disable_user_access(); \
154 (x) = __x; \ 153 (x) = __x; \
155} while (0) 154} while (0)
156#endif /* CONFIG_MMU */
157 155
158#ifdef CONFIG_64BIT 156#ifdef CONFIG_64BIT
159#define __get_user_8(x, ptr, err) \ 157#define __get_user_8(x, ptr, err) \
160 __get_user_asm("ld", x, ptr, err) 158 __get_user_asm("ld", x, ptr, err)
161#else /* !CONFIG_64BIT */ 159#else /* !CONFIG_64BIT */
162#ifdef CONFIG_MMU
163#define __get_user_8(x, ptr, err) \ 160#define __get_user_8(x, ptr, err) \
164do { \ 161do { \
165 u32 __user *__ptr = (u32 __user *)(ptr); \ 162 u32 __user *__ptr = (u32 __user *)(ptr); \
@@ -193,7 +190,6 @@ do { \
193 (x) = (__typeof__(x))((__typeof__((x)-(x)))( \ 190 (x) = (__typeof__(x))((__typeof__((x)-(x)))( \
194 (((u64)__hi << 32) | __lo))); \ 191 (((u64)__hi << 32) | __lo))); \
195} while (0) 192} while (0)
196#endif /* CONFIG_MMU */
197#endif /* CONFIG_64BIT */ 193#endif /* CONFIG_64BIT */
198 194
199 195
@@ -267,8 +263,6 @@ do { \
267 ((x) = 0, -EFAULT); \ 263 ((x) = 0, -EFAULT); \
268}) 264})
269 265
270
271#ifdef CONFIG_MMU
272#define __put_user_asm(insn, x, ptr, err) \ 266#define __put_user_asm(insn, x, ptr, err) \
273do { \ 267do { \
274 uintptr_t __tmp; \ 268 uintptr_t __tmp; \
@@ -292,14 +286,11 @@ do { \
292 : "rJ" (__x), "i" (-EFAULT)); \ 286 : "rJ" (__x), "i" (-EFAULT)); \
293 __disable_user_access(); \ 287 __disable_user_access(); \
294} while (0) 288} while (0)
295#endif /* CONFIG_MMU */
296
297 289
298#ifdef CONFIG_64BIT 290#ifdef CONFIG_64BIT
299#define __put_user_8(x, ptr, err) \ 291#define __put_user_8(x, ptr, err) \
300 __put_user_asm("sd", x, ptr, err) 292 __put_user_asm("sd", x, ptr, err)
301#else /* !CONFIG_64BIT */ 293#else /* !CONFIG_64BIT */
302#ifdef CONFIG_MMU
303#define __put_user_8(x, ptr, err) \ 294#define __put_user_8(x, ptr, err) \
304do { \ 295do { \
305 u32 __user *__ptr = (u32 __user *)(ptr); \ 296 u32 __user *__ptr = (u32 __user *)(ptr); \
@@ -329,7 +320,6 @@ do { \
329 : "rJ" (__x), "rJ" (__x >> 32), "i" (-EFAULT)); \ 320 : "rJ" (__x), "rJ" (__x >> 32), "i" (-EFAULT)); \
330 __disable_user_access(); \ 321 __disable_user_access(); \
331} while (0) 322} while (0)
332#endif /* CONFIG_MMU */
333#endif /* CONFIG_64BIT */ 323#endif /* CONFIG_64BIT */
334 324
335 325
@@ -438,7 +428,6 @@ unsigned long __must_check clear_user(void __user *to, unsigned long n)
438 * will set "err" to -EFAULT, while successful accesses return the previous 428 * will set "err" to -EFAULT, while successful accesses return the previous
439 * value. 429 * value.
440 */ 430 */
441#ifdef CONFIG_MMU
442#define __cmpxchg_user(ptr, old, new, err, size, lrb, scb) \ 431#define __cmpxchg_user(ptr, old, new, err, size, lrb, scb) \
443({ \ 432({ \
444 __typeof__(ptr) __ptr = (ptr); \ 433 __typeof__(ptr) __ptr = (ptr); \
@@ -508,6 +497,5 @@ unsigned long __must_check clear_user(void __user *to, unsigned long n)
508 (err) = __err; \ 497 (err) = __err; \
509 __ret; \ 498 __ret; \
510}) 499})
511#endif /* CONFIG_MMU */
512 500
513#endif /* _ASM_RISCV_UACCESS_H */ 501#endif /* _ASM_RISCV_UACCESS_H */
diff --git a/arch/riscv/include/asm/unistd.h b/arch/riscv/include/asm/unistd.h
index 9f250ed007cd..2f704a5c4196 100644
--- a/arch/riscv/include/asm/unistd.h
+++ b/arch/riscv/include/asm/unistd.h
@@ -14,3 +14,4 @@
14#define __ARCH_HAVE_MMU 14#define __ARCH_HAVE_MMU
15#define __ARCH_WANT_SYS_CLONE 15#define __ARCH_WANT_SYS_CLONE
16#include <uapi/asm/unistd.h> 16#include <uapi/asm/unistd.h>
17#include <uapi/asm/syscalls.h>
diff --git a/arch/riscv/include/asm/vdso-syscalls.h b/arch/riscv/include/asm/vdso-syscalls.h
deleted file mode 100644
index a2ccf1894929..000000000000
--- a/arch/riscv/include/asm/vdso-syscalls.h
+++ /dev/null
@@ -1,28 +0,0 @@
1/*
2 * Copyright (C) 2017 SiFive
3 *
4 * This program is free software; you can redistribute it and/or modify
5 * it under the terms of the GNU General Public License version 2 as
6 * published by the Free Software Foundation.
7 *
8 * This program is distributed in the hope that it will be useful,
9 * but WITHOUT ANY WARRANTY; without even the implied warranty of
10 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11 * GNU General Public License for more details.
12 *
13 * You should have received a copy of the GNU General Public License
14 * along with this program. If not, see <http://www.gnu.org/licenses/>.
15 */
16
17#ifndef _ASM_RISCV_VDSO_SYSCALLS_H
18#define _ASM_RISCV_VDSO_SYSCALLS_H
19
20#ifdef CONFIG_SMP
21
22/* These syscalls are only used by the vDSO and are not in the uapi. */
23#define __NR_riscv_flush_icache (__NR_arch_specific_syscall + 15)
24__SYSCALL(__NR_riscv_flush_icache, sys_riscv_flush_icache)
25
26#endif
27
28#endif /* _ASM_RISCV_VDSO_H */
diff --git a/arch/riscv/include/uapi/asm/syscalls.h b/arch/riscv/include/uapi/asm/syscalls.h
new file mode 100644
index 000000000000..818655b0d535
--- /dev/null
+++ b/arch/riscv/include/uapi/asm/syscalls.h
@@ -0,0 +1,26 @@
1/* SPDX-License-Identifier: GPL-2.0 */
2/*
3 * Copyright (C) 2017 SiFive
4 */
5
6#ifndef _ASM__UAPI__SYSCALLS_H
7#define _ASM__UAPI__SYSCALLS_H
8
9/*
10 * Allows the instruction cache to be flushed from userspace. Despite RISC-V
11 * having a direct 'fence.i' instruction available to userspace (which we
12 * can't trap!), that's not actually viable when running on Linux because the
13 * kernel might schedule a process on another hart. There is no way for
14 * userspace to handle this without invoking the kernel (as it doesn't know the
15 * thread->hart mappings), so we've defined a RISC-V specific system call to
16 * flush the instruction cache.
17 *
18 * __NR_riscv_flush_icache is defined to flush the instruction cache over an
19 * address range, with the flush applying to either all threads or just the
20 * caller. We don't currently do anything with the address range, that's just
21 * in there for forwards compatibility.
22 */
23#define __NR_riscv_flush_icache (__NR_arch_specific_syscall + 15)
24__SYSCALL(__NR_riscv_flush_icache, sys_riscv_flush_icache)
25
26#endif
diff --git a/arch/riscv/kernel/entry.S b/arch/riscv/kernel/entry.S
index 20ee86f782a9..7404ec222406 100644
--- a/arch/riscv/kernel/entry.S
+++ b/arch/riscv/kernel/entry.S
@@ -196,7 +196,7 @@ handle_syscall:
196 addi s2, s2, 0x4 196 addi s2, s2, 0x4
197 REG_S s2, PT_SEPC(sp) 197 REG_S s2, PT_SEPC(sp)
198 /* System calls run with interrupts enabled */ 198 /* System calls run with interrupts enabled */
199 csrs sstatus, SR_IE 199 csrs sstatus, SR_SIE
200 /* Trace syscalls, but only if requested by the user. */ 200 /* Trace syscalls, but only if requested by the user. */
201 REG_L t0, TASK_TI_FLAGS(tp) 201 REG_L t0, TASK_TI_FLAGS(tp)
202 andi t0, t0, _TIF_SYSCALL_TRACE 202 andi t0, t0, _TIF_SYSCALL_TRACE
@@ -224,8 +224,8 @@ ret_from_syscall:
224 224
225ret_from_exception: 225ret_from_exception:
226 REG_L s0, PT_SSTATUS(sp) 226 REG_L s0, PT_SSTATUS(sp)
227 csrc sstatus, SR_IE 227 csrc sstatus, SR_SIE
228 andi s0, s0, SR_PS 228 andi s0, s0, SR_SPP
229 bnez s0, restore_all 229 bnez s0, restore_all
230 230
231resume_userspace: 231resume_userspace:
@@ -255,7 +255,7 @@ work_pending:
255 bnez s1, work_resched 255 bnez s1, work_resched
256work_notifysig: 256work_notifysig:
257 /* Handle pending signals and notify-resume requests */ 257 /* Handle pending signals and notify-resume requests */
258 csrs sstatus, SR_IE /* Enable interrupts for do_notify_resume() */ 258 csrs sstatus, SR_SIE /* Enable interrupts for do_notify_resume() */
259 move a0, sp /* pt_regs */ 259 move a0, sp /* pt_regs */
260 move a1, s0 /* current_thread_info->flags */ 260 move a1, s0 /* current_thread_info->flags */
261 tail do_notify_resume 261 tail do_notify_resume
diff --git a/arch/riscv/kernel/process.c b/arch/riscv/kernel/process.c
index 0d90dcc1fbd3..d74d4adf2d54 100644
--- a/arch/riscv/kernel/process.c
+++ b/arch/riscv/kernel/process.c
@@ -76,7 +76,7 @@ void show_regs(struct pt_regs *regs)
76void start_thread(struct pt_regs *regs, unsigned long pc, 76void start_thread(struct pt_regs *regs, unsigned long pc,
77 unsigned long sp) 77 unsigned long sp)
78{ 78{
79 regs->sstatus = SR_PIE /* User mode, irqs on */ | SR_FS_INITIAL; 79 regs->sstatus = SR_SPIE /* User mode, irqs on */ | SR_FS_INITIAL;
80 regs->sepc = pc; 80 regs->sepc = pc;
81 regs->sp = sp; 81 regs->sp = sp;
82 set_fs(USER_DS); 82 set_fs(USER_DS);
@@ -110,7 +110,7 @@ int copy_thread(unsigned long clone_flags, unsigned long usp,
110 const register unsigned long gp __asm__ ("gp"); 110 const register unsigned long gp __asm__ ("gp");
111 memset(childregs, 0, sizeof(struct pt_regs)); 111 memset(childregs, 0, sizeof(struct pt_regs));
112 childregs->gp = gp; 112 childregs->gp = gp;
113 childregs->sstatus = SR_PS | SR_PIE; /* Supervisor, irqs on */ 113 childregs->sstatus = SR_SPP | SR_SPIE; /* Supervisor, irqs on */
114 114
115 p->thread.ra = (unsigned long)ret_from_kernel_thread; 115 p->thread.ra = (unsigned long)ret_from_kernel_thread;
116 p->thread.s[0] = usp; /* fn */ 116 p->thread.s[0] = usp; /* fn */
diff --git a/arch/riscv/kernel/syscall_table.c b/arch/riscv/kernel/syscall_table.c
index a5bd6401f95e..ade52b903a43 100644
--- a/arch/riscv/kernel/syscall_table.c
+++ b/arch/riscv/kernel/syscall_table.c
@@ -23,5 +23,4 @@
23void *sys_call_table[__NR_syscalls] = { 23void *sys_call_table[__NR_syscalls] = {
24 [0 ... __NR_syscalls - 1] = sys_ni_syscall, 24 [0 ... __NR_syscalls - 1] = sys_ni_syscall,
25#include <asm/unistd.h> 25#include <asm/unistd.h>
26#include <asm/vdso-syscalls.h>
27}; 26};
diff --git a/arch/riscv/kernel/vdso/flush_icache.S b/arch/riscv/kernel/vdso/flush_icache.S
index b0fbad74e873..023e4d4aef58 100644
--- a/arch/riscv/kernel/vdso/flush_icache.S
+++ b/arch/riscv/kernel/vdso/flush_icache.S
@@ -13,7 +13,6 @@
13 13
14#include <linux/linkage.h> 14#include <linux/linkage.h>
15#include <asm/unistd.h> 15#include <asm/unistd.h>
16#include <asm/vdso-syscalls.h>
17 16
18 .text 17 .text
19/* int __vdso_flush_icache(void *start, void *end, unsigned long flags); */ 18/* int __vdso_flush_icache(void *start, void *end, unsigned long flags); */
diff --git a/arch/riscv/mm/fault.c b/arch/riscv/mm/fault.c
index df2ca3c65048..0713f3c67ab4 100644
--- a/arch/riscv/mm/fault.c
+++ b/arch/riscv/mm/fault.c
@@ -63,7 +63,7 @@ asmlinkage void do_page_fault(struct pt_regs *regs)
63 goto vmalloc_fault; 63 goto vmalloc_fault;
64 64
65 /* Enable interrupts if they were enabled in the parent context. */ 65 /* Enable interrupts if they were enabled in the parent context. */
66 if (likely(regs->sstatus & SR_PIE)) 66 if (likely(regs->sstatus & SR_SPIE))
67 local_irq_enable(); 67 local_irq_enable();
68 68
69 /* 69 /*
diff --git a/arch/sh/boards/mach-se/770x/setup.c b/arch/sh/boards/mach-se/770x/setup.c
index 77c35350ee77..412326d59e6f 100644
--- a/arch/sh/boards/mach-se/770x/setup.c
+++ b/arch/sh/boards/mach-se/770x/setup.c
@@ -9,6 +9,7 @@
9 */ 9 */
10#include <linux/init.h> 10#include <linux/init.h>
11#include <linux/platform_device.h> 11#include <linux/platform_device.h>
12#include <linux/sh_eth.h>
12#include <mach-se/mach/se.h> 13#include <mach-se/mach/se.h>
13#include <mach-se/mach/mrshpc.h> 14#include <mach-se/mach/mrshpc.h>
14#include <asm/machvec.h> 15#include <asm/machvec.h>
@@ -115,13 +116,23 @@ static struct platform_device heartbeat_device = {
115#if defined(CONFIG_CPU_SUBTYPE_SH7710) ||\ 116#if defined(CONFIG_CPU_SUBTYPE_SH7710) ||\
116 defined(CONFIG_CPU_SUBTYPE_SH7712) 117 defined(CONFIG_CPU_SUBTYPE_SH7712)
117/* SH771X Ethernet driver */ 118/* SH771X Ethernet driver */
119static struct sh_eth_plat_data sh_eth_plat = {
120 .phy = PHY_ID,
121 .phy_interface = PHY_INTERFACE_MODE_MII,
122};
123
118static struct resource sh_eth0_resources[] = { 124static struct resource sh_eth0_resources[] = {
119 [0] = { 125 [0] = {
120 .start = SH_ETH0_BASE, 126 .start = SH_ETH0_BASE,
121 .end = SH_ETH0_BASE + 0x1B8, 127 .end = SH_ETH0_BASE + 0x1B8 - 1,
122 .flags = IORESOURCE_MEM, 128 .flags = IORESOURCE_MEM,
123 }, 129 },
124 [1] = { 130 [1] = {
131 .start = SH_TSU_BASE,
132 .end = SH_TSU_BASE + 0x200 - 1,
133 .flags = IORESOURCE_MEM,
134 },
135 [2] = {
125 .start = SH_ETH0_IRQ, 136 .start = SH_ETH0_IRQ,
126 .end = SH_ETH0_IRQ, 137 .end = SH_ETH0_IRQ,
127 .flags = IORESOURCE_IRQ, 138 .flags = IORESOURCE_IRQ,
@@ -132,7 +143,7 @@ static struct platform_device sh_eth0_device = {
132 .name = "sh771x-ether", 143 .name = "sh771x-ether",
133 .id = 0, 144 .id = 0,
134 .dev = { 145 .dev = {
135 .platform_data = PHY_ID, 146 .platform_data = &sh_eth_plat,
136 }, 147 },
137 .num_resources = ARRAY_SIZE(sh_eth0_resources), 148 .num_resources = ARRAY_SIZE(sh_eth0_resources),
138 .resource = sh_eth0_resources, 149 .resource = sh_eth0_resources,
@@ -141,10 +152,15 @@ static struct platform_device sh_eth0_device = {
141static struct resource sh_eth1_resources[] = { 152static struct resource sh_eth1_resources[] = {
142 [0] = { 153 [0] = {
143 .start = SH_ETH1_BASE, 154 .start = SH_ETH1_BASE,
144 .end = SH_ETH1_BASE + 0x1B8, 155 .end = SH_ETH1_BASE + 0x1B8 - 1,
145 .flags = IORESOURCE_MEM, 156 .flags = IORESOURCE_MEM,
146 }, 157 },
147 [1] = { 158 [1] = {
159 .start = SH_TSU_BASE,
160 .end = SH_TSU_BASE + 0x200 - 1,
161 .flags = IORESOURCE_MEM,
162 },
163 [2] = {
148 .start = SH_ETH1_IRQ, 164 .start = SH_ETH1_IRQ,
149 .end = SH_ETH1_IRQ, 165 .end = SH_ETH1_IRQ,
150 .flags = IORESOURCE_IRQ, 166 .flags = IORESOURCE_IRQ,
@@ -155,7 +171,7 @@ static struct platform_device sh_eth1_device = {
155 .name = "sh771x-ether", 171 .name = "sh771x-ether",
156 .id = 1, 172 .id = 1,
157 .dev = { 173 .dev = {
158 .platform_data = PHY_ID, 174 .platform_data = &sh_eth_plat,
159 }, 175 },
160 .num_resources = ARRAY_SIZE(sh_eth1_resources), 176 .num_resources = ARRAY_SIZE(sh_eth1_resources),
161 .resource = sh_eth1_resources, 177 .resource = sh_eth1_resources,
diff --git a/arch/sh/include/mach-se/mach/se.h b/arch/sh/include/mach-se/mach/se.h
index 4246ef9b07a3..aa83fe1ff0b1 100644
--- a/arch/sh/include/mach-se/mach/se.h
+++ b/arch/sh/include/mach-se/mach/se.h
@@ -100,6 +100,7 @@
100/* Base address */ 100/* Base address */
101#define SH_ETH0_BASE 0xA7000000 101#define SH_ETH0_BASE 0xA7000000
102#define SH_ETH1_BASE 0xA7000400 102#define SH_ETH1_BASE 0xA7000400
103#define SH_TSU_BASE 0xA7000800
103/* PHY ID */ 104/* PHY ID */
104#if defined(CONFIG_CPU_SUBTYPE_SH7710) 105#if defined(CONFIG_CPU_SUBTYPE_SH7710)
105# define PHY_ID 0x00 106# define PHY_ID 0x00
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index d4fc98c50378..20da391b5f32 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -55,7 +55,6 @@ config X86
55 select ARCH_HAS_GCOV_PROFILE_ALL 55 select ARCH_HAS_GCOV_PROFILE_ALL
56 select ARCH_HAS_KCOV if X86_64 56 select ARCH_HAS_KCOV if X86_64
57 select ARCH_HAS_PMEM_API if X86_64 57 select ARCH_HAS_PMEM_API if X86_64
58 # Causing hangs/crashes, see the commit that added this change for details.
59 select ARCH_HAS_REFCOUNT 58 select ARCH_HAS_REFCOUNT
60 select ARCH_HAS_UACCESS_FLUSHCACHE if X86_64 59 select ARCH_HAS_UACCESS_FLUSHCACHE if X86_64
61 select ARCH_HAS_SET_MEMORY 60 select ARCH_HAS_SET_MEMORY
@@ -89,6 +88,7 @@ config X86
89 select GENERIC_CLOCKEVENTS_MIN_ADJUST 88 select GENERIC_CLOCKEVENTS_MIN_ADJUST
90 select GENERIC_CMOS_UPDATE 89 select GENERIC_CMOS_UPDATE
91 select GENERIC_CPU_AUTOPROBE 90 select GENERIC_CPU_AUTOPROBE
91 select GENERIC_CPU_VULNERABILITIES
92 select GENERIC_EARLY_IOREMAP 92 select GENERIC_EARLY_IOREMAP
93 select GENERIC_FIND_FIRST_BIT 93 select GENERIC_FIND_FIRST_BIT
94 select GENERIC_IOMAP 94 select GENERIC_IOMAP
@@ -429,6 +429,19 @@ config GOLDFISH
429 def_bool y 429 def_bool y
430 depends on X86_GOLDFISH 430 depends on X86_GOLDFISH
431 431
432config RETPOLINE
433 bool "Avoid speculative indirect branches in kernel"
434 default y
435 help
436 Compile kernel with the retpoline compiler options to guard against
437 kernel-to-user data leaks by avoiding speculative indirect
438 branches. Requires a compiler with -mindirect-branch=thunk-extern
439 support for full protection. The kernel may run slower.
440
441 Without compiler support, at least indirect branches in assembler
442 code are eliminated. Since this includes the syscall entry path,
443 it is not entirely pointless.
444
432config INTEL_RDT 445config INTEL_RDT
433 bool "Intel Resource Director Technology support" 446 bool "Intel Resource Director Technology support"
434 default n 447 default n
diff --git a/arch/x86/Makefile b/arch/x86/Makefile
index 3e73bc255e4e..fad55160dcb9 100644
--- a/arch/x86/Makefile
+++ b/arch/x86/Makefile
@@ -230,6 +230,14 @@ KBUILD_CFLAGS += -Wno-sign-compare
230# 230#
231KBUILD_CFLAGS += -fno-asynchronous-unwind-tables 231KBUILD_CFLAGS += -fno-asynchronous-unwind-tables
232 232
233# Avoid indirect branches in kernel to deal with Spectre
234ifdef CONFIG_RETPOLINE
235 RETPOLINE_CFLAGS += $(call cc-option,-mindirect-branch=thunk-extern -mindirect-branch-register)
236 ifneq ($(RETPOLINE_CFLAGS),)
237 KBUILD_CFLAGS += $(RETPOLINE_CFLAGS) -DRETPOLINE
238 endif
239endif
240
233archscripts: scripts_basic 241archscripts: scripts_basic
234 $(Q)$(MAKE) $(build)=arch/x86/tools relocs 242 $(Q)$(MAKE) $(build)=arch/x86/tools relocs
235 243
diff --git a/arch/x86/crypto/aesni-intel_asm.S b/arch/x86/crypto/aesni-intel_asm.S
index 16627fec80b2..3d09e3aca18d 100644
--- a/arch/x86/crypto/aesni-intel_asm.S
+++ b/arch/x86/crypto/aesni-intel_asm.S
@@ -32,6 +32,7 @@
32#include <linux/linkage.h> 32#include <linux/linkage.h>
33#include <asm/inst.h> 33#include <asm/inst.h>
34#include <asm/frame.h> 34#include <asm/frame.h>
35#include <asm/nospec-branch.h>
35 36
36/* 37/*
37 * The following macros are used to move an (un)aligned 16 byte value to/from 38 * The following macros are used to move an (un)aligned 16 byte value to/from
@@ -2884,7 +2885,7 @@ ENTRY(aesni_xts_crypt8)
2884 pxor INC, STATE4 2885 pxor INC, STATE4
2885 movdqu IV, 0x30(OUTP) 2886 movdqu IV, 0x30(OUTP)
2886 2887
2887 call *%r11 2888 CALL_NOSPEC %r11
2888 2889
2889 movdqu 0x00(OUTP), INC 2890 movdqu 0x00(OUTP), INC
2890 pxor INC, STATE1 2891 pxor INC, STATE1
@@ -2929,7 +2930,7 @@ ENTRY(aesni_xts_crypt8)
2929 _aesni_gf128mul_x_ble() 2930 _aesni_gf128mul_x_ble()
2930 movups IV, (IVP) 2931 movups IV, (IVP)
2931 2932
2932 call *%r11 2933 CALL_NOSPEC %r11
2933 2934
2934 movdqu 0x40(OUTP), INC 2935 movdqu 0x40(OUTP), INC
2935 pxor INC, STATE1 2936 pxor INC, STATE1
diff --git a/arch/x86/crypto/camellia-aesni-avx-asm_64.S b/arch/x86/crypto/camellia-aesni-avx-asm_64.S
index f7c495e2863c..a14af6eb09cb 100644
--- a/arch/x86/crypto/camellia-aesni-avx-asm_64.S
+++ b/arch/x86/crypto/camellia-aesni-avx-asm_64.S
@@ -17,6 +17,7 @@
17 17
18#include <linux/linkage.h> 18#include <linux/linkage.h>
19#include <asm/frame.h> 19#include <asm/frame.h>
20#include <asm/nospec-branch.h>
20 21
21#define CAMELLIA_TABLE_BYTE_LEN 272 22#define CAMELLIA_TABLE_BYTE_LEN 272
22 23
@@ -1227,7 +1228,7 @@ camellia_xts_crypt_16way:
1227 vpxor 14 * 16(%rax), %xmm15, %xmm14; 1228 vpxor 14 * 16(%rax), %xmm15, %xmm14;
1228 vpxor 15 * 16(%rax), %xmm15, %xmm15; 1229 vpxor 15 * 16(%rax), %xmm15, %xmm15;
1229 1230
1230 call *%r9; 1231 CALL_NOSPEC %r9;
1231 1232
1232 addq $(16 * 16), %rsp; 1233 addq $(16 * 16), %rsp;
1233 1234
diff --git a/arch/x86/crypto/camellia-aesni-avx2-asm_64.S b/arch/x86/crypto/camellia-aesni-avx2-asm_64.S
index eee5b3982cfd..b66bbfa62f50 100644
--- a/arch/x86/crypto/camellia-aesni-avx2-asm_64.S
+++ b/arch/x86/crypto/camellia-aesni-avx2-asm_64.S
@@ -12,6 +12,7 @@
12 12
13#include <linux/linkage.h> 13#include <linux/linkage.h>
14#include <asm/frame.h> 14#include <asm/frame.h>
15#include <asm/nospec-branch.h>
15 16
16#define CAMELLIA_TABLE_BYTE_LEN 272 17#define CAMELLIA_TABLE_BYTE_LEN 272
17 18
@@ -1343,7 +1344,7 @@ camellia_xts_crypt_32way:
1343 vpxor 14 * 32(%rax), %ymm15, %ymm14; 1344 vpxor 14 * 32(%rax), %ymm15, %ymm14;
1344 vpxor 15 * 32(%rax), %ymm15, %ymm15; 1345 vpxor 15 * 32(%rax), %ymm15, %ymm15;
1345 1346
1346 call *%r9; 1347 CALL_NOSPEC %r9;
1347 1348
1348 addq $(16 * 32), %rsp; 1349 addq $(16 * 32), %rsp;
1349 1350
diff --git a/arch/x86/crypto/crc32c-pcl-intel-asm_64.S b/arch/x86/crypto/crc32c-pcl-intel-asm_64.S
index 7a7de27c6f41..d9b734d0c8cc 100644
--- a/arch/x86/crypto/crc32c-pcl-intel-asm_64.S
+++ b/arch/x86/crypto/crc32c-pcl-intel-asm_64.S
@@ -45,6 +45,7 @@
45 45
46#include <asm/inst.h> 46#include <asm/inst.h>
47#include <linux/linkage.h> 47#include <linux/linkage.h>
48#include <asm/nospec-branch.h>
48 49
49## ISCSI CRC 32 Implementation with crc32 and pclmulqdq Instruction 50## ISCSI CRC 32 Implementation with crc32 and pclmulqdq Instruction
50 51
@@ -172,7 +173,7 @@ continue_block:
172 movzxw (bufp, %rax, 2), len 173 movzxw (bufp, %rax, 2), len
173 lea crc_array(%rip), bufp 174 lea crc_array(%rip), bufp
174 lea (bufp, len, 1), bufp 175 lea (bufp, len, 1), bufp
175 jmp *bufp 176 JMP_NOSPEC bufp
176 177
177 ################################################################ 178 ################################################################
178 ## 2a) PROCESS FULL BLOCKS: 179 ## 2a) PROCESS FULL BLOCKS:
diff --git a/arch/x86/entry/calling.h b/arch/x86/entry/calling.h
index 45a63e00a6af..3f48f695d5e6 100644
--- a/arch/x86/entry/calling.h
+++ b/arch/x86/entry/calling.h
@@ -198,8 +198,11 @@ For 32-bit we have the following conventions - kernel is built with
198 * PAGE_TABLE_ISOLATION PGDs are 8k. Flip bit 12 to switch between the two 198 * PAGE_TABLE_ISOLATION PGDs are 8k. Flip bit 12 to switch between the two
199 * halves: 199 * halves:
200 */ 200 */
201#define PTI_SWITCH_PGTABLES_MASK (1<<PAGE_SHIFT) 201#define PTI_USER_PGTABLE_BIT PAGE_SHIFT
202#define PTI_SWITCH_MASK (PTI_SWITCH_PGTABLES_MASK|(1<<X86_CR3_PTI_SWITCH_BIT)) 202#define PTI_USER_PGTABLE_MASK (1 << PTI_USER_PGTABLE_BIT)
203#define PTI_USER_PCID_BIT X86_CR3_PTI_PCID_USER_BIT
204#define PTI_USER_PCID_MASK (1 << PTI_USER_PCID_BIT)
205#define PTI_USER_PGTABLE_AND_PCID_MASK (PTI_USER_PCID_MASK | PTI_USER_PGTABLE_MASK)
203 206
204.macro SET_NOFLUSH_BIT reg:req 207.macro SET_NOFLUSH_BIT reg:req
205 bts $X86_CR3_PCID_NOFLUSH_BIT, \reg 208 bts $X86_CR3_PCID_NOFLUSH_BIT, \reg
@@ -208,7 +211,7 @@ For 32-bit we have the following conventions - kernel is built with
208.macro ADJUST_KERNEL_CR3 reg:req 211.macro ADJUST_KERNEL_CR3 reg:req
209 ALTERNATIVE "", "SET_NOFLUSH_BIT \reg", X86_FEATURE_PCID 212 ALTERNATIVE "", "SET_NOFLUSH_BIT \reg", X86_FEATURE_PCID
210 /* Clear PCID and "PAGE_TABLE_ISOLATION bit", point CR3 at kernel pagetables: */ 213 /* Clear PCID and "PAGE_TABLE_ISOLATION bit", point CR3 at kernel pagetables: */
211 andq $(~PTI_SWITCH_MASK), \reg 214 andq $(~PTI_USER_PGTABLE_AND_PCID_MASK), \reg
212.endm 215.endm
213 216
214.macro SWITCH_TO_KERNEL_CR3 scratch_reg:req 217.macro SWITCH_TO_KERNEL_CR3 scratch_reg:req
@@ -239,15 +242,19 @@ For 32-bit we have the following conventions - kernel is built with
239 /* Flush needed, clear the bit */ 242 /* Flush needed, clear the bit */
240 btr \scratch_reg, THIS_CPU_user_pcid_flush_mask 243 btr \scratch_reg, THIS_CPU_user_pcid_flush_mask
241 movq \scratch_reg2, \scratch_reg 244 movq \scratch_reg2, \scratch_reg
242 jmp .Lwrcr3_\@ 245 jmp .Lwrcr3_pcid_\@
243 246
244.Lnoflush_\@: 247.Lnoflush_\@:
245 movq \scratch_reg2, \scratch_reg 248 movq \scratch_reg2, \scratch_reg
246 SET_NOFLUSH_BIT \scratch_reg 249 SET_NOFLUSH_BIT \scratch_reg
247 250
251.Lwrcr3_pcid_\@:
252 /* Flip the ASID to the user version */
253 orq $(PTI_USER_PCID_MASK), \scratch_reg
254
248.Lwrcr3_\@: 255.Lwrcr3_\@:
249 /* Flip the PGD and ASID to the user version */ 256 /* Flip the PGD to the user version */
250 orq $(PTI_SWITCH_MASK), \scratch_reg 257 orq $(PTI_USER_PGTABLE_MASK), \scratch_reg
251 mov \scratch_reg, %cr3 258 mov \scratch_reg, %cr3
252.Lend_\@: 259.Lend_\@:
253.endm 260.endm
@@ -263,17 +270,12 @@ For 32-bit we have the following conventions - kernel is built with
263 movq %cr3, \scratch_reg 270 movq %cr3, \scratch_reg
264 movq \scratch_reg, \save_reg 271 movq \scratch_reg, \save_reg
265 /* 272 /*
266 * Is the "switch mask" all zero? That means that both of 273 * Test the user pagetable bit. If set, then the user page tables
267 * these are zero: 274 * are active. If clear CR3 already has the kernel page table
268 * 275 * active.
269 * 1. The user/kernel PCID bit, and
270 * 2. The user/kernel "bit" that points CR3 to the
271 * bottom half of the 8k PGD
272 *
273 * That indicates a kernel CR3 value, not a user CR3.
274 */ 276 */
275 testq $(PTI_SWITCH_MASK), \scratch_reg 277 bt $PTI_USER_PGTABLE_BIT, \scratch_reg
276 jz .Ldone_\@ 278 jnc .Ldone_\@
277 279
278 ADJUST_KERNEL_CR3 \scratch_reg 280 ADJUST_KERNEL_CR3 \scratch_reg
279 movq \scratch_reg, %cr3 281 movq \scratch_reg, %cr3
@@ -290,7 +292,7 @@ For 32-bit we have the following conventions - kernel is built with
290 * KERNEL pages can always resume with NOFLUSH as we do 292 * KERNEL pages can always resume with NOFLUSH as we do
291 * explicit flushes. 293 * explicit flushes.
292 */ 294 */
293 bt $X86_CR3_PTI_SWITCH_BIT, \save_reg 295 bt $PTI_USER_PGTABLE_BIT, \save_reg
294 jnc .Lnoflush_\@ 296 jnc .Lnoflush_\@
295 297
296 /* 298 /*
diff --git a/arch/x86/entry/entry_32.S b/arch/x86/entry/entry_32.S
index ace8f321a5a1..a1f28a54f23a 100644
--- a/arch/x86/entry/entry_32.S
+++ b/arch/x86/entry/entry_32.S
@@ -44,6 +44,7 @@
44#include <asm/asm.h> 44#include <asm/asm.h>
45#include <asm/smap.h> 45#include <asm/smap.h>
46#include <asm/frame.h> 46#include <asm/frame.h>
47#include <asm/nospec-branch.h>
47 48
48 .section .entry.text, "ax" 49 .section .entry.text, "ax"
49 50
@@ -290,7 +291,7 @@ ENTRY(ret_from_fork)
290 291
291 /* kernel thread */ 292 /* kernel thread */
2921: movl %edi, %eax 2931: movl %edi, %eax
293 call *%ebx 294 CALL_NOSPEC %ebx
294 /* 295 /*
295 * A kernel thread is allowed to return here after successfully 296 * A kernel thread is allowed to return here after successfully
296 * calling do_execve(). Exit to userspace to complete the execve() 297 * calling do_execve(). Exit to userspace to complete the execve()
@@ -919,7 +920,7 @@ common_exception:
919 movl %ecx, %es 920 movl %ecx, %es
920 TRACE_IRQS_OFF 921 TRACE_IRQS_OFF
921 movl %esp, %eax # pt_regs pointer 922 movl %esp, %eax # pt_regs pointer
922 call *%edi 923 CALL_NOSPEC %edi
923 jmp ret_from_exception 924 jmp ret_from_exception
924END(common_exception) 925END(common_exception)
925 926
diff --git a/arch/x86/entry/entry_64.S b/arch/x86/entry/entry_64.S
index f048e384ff54..4f8e1d35a97c 100644
--- a/arch/x86/entry/entry_64.S
+++ b/arch/x86/entry/entry_64.S
@@ -37,6 +37,7 @@
37#include <asm/pgtable_types.h> 37#include <asm/pgtable_types.h>
38#include <asm/export.h> 38#include <asm/export.h>
39#include <asm/frame.h> 39#include <asm/frame.h>
40#include <asm/nospec-branch.h>
40#include <linux/err.h> 41#include <linux/err.h>
41 42
42#include "calling.h" 43#include "calling.h"
@@ -191,7 +192,7 @@ ENTRY(entry_SYSCALL_64_trampoline)
191 */ 192 */
192 pushq %rdi 193 pushq %rdi
193 movq $entry_SYSCALL_64_stage2, %rdi 194 movq $entry_SYSCALL_64_stage2, %rdi
194 jmp *%rdi 195 JMP_NOSPEC %rdi
195END(entry_SYSCALL_64_trampoline) 196END(entry_SYSCALL_64_trampoline)
196 197
197 .popsection 198 .popsection
@@ -270,7 +271,12 @@ entry_SYSCALL_64_fastpath:
270 * It might end up jumping to the slow path. If it jumps, RAX 271 * It might end up jumping to the slow path. If it jumps, RAX
271 * and all argument registers are clobbered. 272 * and all argument registers are clobbered.
272 */ 273 */
274#ifdef CONFIG_RETPOLINE
275 movq sys_call_table(, %rax, 8), %rax
276 call __x86_indirect_thunk_rax
277#else
273 call *sys_call_table(, %rax, 8) 278 call *sys_call_table(, %rax, 8)
279#endif
274.Lentry_SYSCALL_64_after_fastpath_call: 280.Lentry_SYSCALL_64_after_fastpath_call:
275 281
276 movq %rax, RAX(%rsp) 282 movq %rax, RAX(%rsp)
@@ -442,7 +448,7 @@ ENTRY(stub_ptregs_64)
442 jmp entry_SYSCALL64_slow_path 448 jmp entry_SYSCALL64_slow_path
443 449
4441: 4501:
445 jmp *%rax /* Called from C */ 451 JMP_NOSPEC %rax /* Called from C */
446END(stub_ptregs_64) 452END(stub_ptregs_64)
447 453
448.macro ptregs_stub func 454.macro ptregs_stub func
@@ -521,7 +527,7 @@ ENTRY(ret_from_fork)
5211: 5271:
522 /* kernel thread */ 528 /* kernel thread */
523 movq %r12, %rdi 529 movq %r12, %rdi
524 call *%rbx 530 CALL_NOSPEC %rbx
525 /* 531 /*
526 * A kernel thread is allowed to return here after successfully 532 * A kernel thread is allowed to return here after successfully
527 * calling do_execve(). Exit to userspace to complete the execve() 533 * calling do_execve(). Exit to userspace to complete the execve()
diff --git a/arch/x86/events/intel/bts.c b/arch/x86/events/intel/bts.c
index 141e07b06216..24ffa1e88cf9 100644
--- a/arch/x86/events/intel/bts.c
+++ b/arch/x86/events/intel/bts.c
@@ -582,6 +582,24 @@ static __init int bts_init(void)
582 if (!boot_cpu_has(X86_FEATURE_DTES64) || !x86_pmu.bts) 582 if (!boot_cpu_has(X86_FEATURE_DTES64) || !x86_pmu.bts)
583 return -ENODEV; 583 return -ENODEV;
584 584
585 if (boot_cpu_has(X86_FEATURE_PTI)) {
586 /*
587 * BTS hardware writes through a virtual memory map we must
588 * either use the kernel physical map, or the user mapping of
589 * the AUX buffer.
590 *
591 * However, since this driver supports per-CPU and per-task inherit
592 * we cannot use the user mapping since it will not be availble
593 * if we're not running the owning process.
594 *
595 * With PTI we can't use the kernal map either, because its not
596 * there when we run userspace.
597 *
598 * For now, disable this driver when using PTI.
599 */
600 return -ENODEV;
601 }
602
585 bts_pmu.capabilities = PERF_PMU_CAP_AUX_NO_SG | PERF_PMU_CAP_ITRACE | 603 bts_pmu.capabilities = PERF_PMU_CAP_AUX_NO_SG | PERF_PMU_CAP_ITRACE |
586 PERF_PMU_CAP_EXCLUSIVE; 604 PERF_PMU_CAP_EXCLUSIVE;
587 bts_pmu.task_ctx_nr = perf_sw_context; 605 bts_pmu.task_ctx_nr = perf_sw_context;
diff --git a/arch/x86/include/asm/asm-prototypes.h b/arch/x86/include/asm/asm-prototypes.h
index ff700d81e91e..0927cdc4f946 100644
--- a/arch/x86/include/asm/asm-prototypes.h
+++ b/arch/x86/include/asm/asm-prototypes.h
@@ -11,7 +11,32 @@
11#include <asm/pgtable.h> 11#include <asm/pgtable.h>
12#include <asm/special_insns.h> 12#include <asm/special_insns.h>
13#include <asm/preempt.h> 13#include <asm/preempt.h>
14#include <asm/asm.h>
14 15
15#ifndef CONFIG_X86_CMPXCHG64 16#ifndef CONFIG_X86_CMPXCHG64
16extern void cmpxchg8b_emu(void); 17extern void cmpxchg8b_emu(void);
17#endif 18#endif
19
20#ifdef CONFIG_RETPOLINE
21#ifdef CONFIG_X86_32
22#define INDIRECT_THUNK(reg) extern asmlinkage void __x86_indirect_thunk_e ## reg(void);
23#else
24#define INDIRECT_THUNK(reg) extern asmlinkage void __x86_indirect_thunk_r ## reg(void);
25INDIRECT_THUNK(8)
26INDIRECT_THUNK(9)
27INDIRECT_THUNK(10)
28INDIRECT_THUNK(11)
29INDIRECT_THUNK(12)
30INDIRECT_THUNK(13)
31INDIRECT_THUNK(14)
32INDIRECT_THUNK(15)
33#endif
34INDIRECT_THUNK(ax)
35INDIRECT_THUNK(bx)
36INDIRECT_THUNK(cx)
37INDIRECT_THUNK(dx)
38INDIRECT_THUNK(si)
39INDIRECT_THUNK(di)
40INDIRECT_THUNK(bp)
41INDIRECT_THUNK(sp)
42#endif /* CONFIG_RETPOLINE */
diff --git a/arch/x86/include/asm/cpufeatures.h b/arch/x86/include/asm/cpufeatures.h
index 21ac898df2d8..f275447862f4 100644
--- a/arch/x86/include/asm/cpufeatures.h
+++ b/arch/x86/include/asm/cpufeatures.h
@@ -203,6 +203,8 @@
203#define X86_FEATURE_PROC_FEEDBACK ( 7*32+ 9) /* AMD ProcFeedbackInterface */ 203#define X86_FEATURE_PROC_FEEDBACK ( 7*32+ 9) /* AMD ProcFeedbackInterface */
204#define X86_FEATURE_SME ( 7*32+10) /* AMD Secure Memory Encryption */ 204#define X86_FEATURE_SME ( 7*32+10) /* AMD Secure Memory Encryption */
205#define X86_FEATURE_PTI ( 7*32+11) /* Kernel Page Table Isolation enabled */ 205#define X86_FEATURE_PTI ( 7*32+11) /* Kernel Page Table Isolation enabled */
206#define X86_FEATURE_RETPOLINE ( 7*32+12) /* Generic Retpoline mitigation for Spectre variant 2 */
207#define X86_FEATURE_RETPOLINE_AMD ( 7*32+13) /* AMD Retpoline mitigation for Spectre variant 2 */
206#define X86_FEATURE_INTEL_PPIN ( 7*32+14) /* Intel Processor Inventory Number */ 208#define X86_FEATURE_INTEL_PPIN ( 7*32+14) /* Intel Processor Inventory Number */
207#define X86_FEATURE_INTEL_PT ( 7*32+15) /* Intel Processor Trace */ 209#define X86_FEATURE_INTEL_PT ( 7*32+15) /* Intel Processor Trace */
208#define X86_FEATURE_AVX512_4VNNIW ( 7*32+16) /* AVX-512 Neural Network Instructions */ 210#define X86_FEATURE_AVX512_4VNNIW ( 7*32+16) /* AVX-512 Neural Network Instructions */
@@ -342,5 +344,7 @@
342#define X86_BUG_MONITOR X86_BUG(12) /* IPI required to wake up remote CPU */ 344#define X86_BUG_MONITOR X86_BUG(12) /* IPI required to wake up remote CPU */
343#define X86_BUG_AMD_E400 X86_BUG(13) /* CPU is among the affected by Erratum 400 */ 345#define X86_BUG_AMD_E400 X86_BUG(13) /* CPU is among the affected by Erratum 400 */
344#define X86_BUG_CPU_MELTDOWN X86_BUG(14) /* CPU is affected by meltdown attack and needs kernel page table isolation */ 346#define X86_BUG_CPU_MELTDOWN X86_BUG(14) /* CPU is affected by meltdown attack and needs kernel page table isolation */
347#define X86_BUG_SPECTRE_V1 X86_BUG(15) /* CPU is affected by Spectre variant 1 attack with conditional branches */
348#define X86_BUG_SPECTRE_V2 X86_BUG(16) /* CPU is affected by Spectre variant 2 attack with indirect branches */
345 349
346#endif /* _ASM_X86_CPUFEATURES_H */ 350#endif /* _ASM_X86_CPUFEATURES_H */
diff --git a/arch/x86/include/asm/mshyperv.h b/arch/x86/include/asm/mshyperv.h
index 5400add2885b..8bf450b13d9f 100644
--- a/arch/x86/include/asm/mshyperv.h
+++ b/arch/x86/include/asm/mshyperv.h
@@ -7,6 +7,7 @@
7#include <linux/nmi.h> 7#include <linux/nmi.h>
8#include <asm/io.h> 8#include <asm/io.h>
9#include <asm/hyperv.h> 9#include <asm/hyperv.h>
10#include <asm/nospec-branch.h>
10 11
11/* 12/*
12 * The below CPUID leaves are present if VersionAndFeatures.HypervisorPresent 13 * The below CPUID leaves are present if VersionAndFeatures.HypervisorPresent
@@ -186,10 +187,11 @@ static inline u64 hv_do_hypercall(u64 control, void *input, void *output)
186 return U64_MAX; 187 return U64_MAX;
187 188
188 __asm__ __volatile__("mov %4, %%r8\n" 189 __asm__ __volatile__("mov %4, %%r8\n"
189 "call *%5" 190 CALL_NOSPEC
190 : "=a" (hv_status), ASM_CALL_CONSTRAINT, 191 : "=a" (hv_status), ASM_CALL_CONSTRAINT,
191 "+c" (control), "+d" (input_address) 192 "+c" (control), "+d" (input_address)
192 : "r" (output_address), "m" (hv_hypercall_pg) 193 : "r" (output_address),
194 THUNK_TARGET(hv_hypercall_pg)
193 : "cc", "memory", "r8", "r9", "r10", "r11"); 195 : "cc", "memory", "r8", "r9", "r10", "r11");
194#else 196#else
195 u32 input_address_hi = upper_32_bits(input_address); 197 u32 input_address_hi = upper_32_bits(input_address);
@@ -200,13 +202,13 @@ static inline u64 hv_do_hypercall(u64 control, void *input, void *output)
200 if (!hv_hypercall_pg) 202 if (!hv_hypercall_pg)
201 return U64_MAX; 203 return U64_MAX;
202 204
203 __asm__ __volatile__("call *%7" 205 __asm__ __volatile__(CALL_NOSPEC
204 : "=A" (hv_status), 206 : "=A" (hv_status),
205 "+c" (input_address_lo), ASM_CALL_CONSTRAINT 207 "+c" (input_address_lo), ASM_CALL_CONSTRAINT
206 : "A" (control), 208 : "A" (control),
207 "b" (input_address_hi), 209 "b" (input_address_hi),
208 "D"(output_address_hi), "S"(output_address_lo), 210 "D"(output_address_hi), "S"(output_address_lo),
209 "m" (hv_hypercall_pg) 211 THUNK_TARGET(hv_hypercall_pg)
210 : "cc", "memory"); 212 : "cc", "memory");
211#endif /* !x86_64 */ 213#endif /* !x86_64 */
212 return hv_status; 214 return hv_status;
@@ -227,10 +229,10 @@ static inline u64 hv_do_fast_hypercall8(u16 code, u64 input1)
227 229
228#ifdef CONFIG_X86_64 230#ifdef CONFIG_X86_64
229 { 231 {
230 __asm__ __volatile__("call *%4" 232 __asm__ __volatile__(CALL_NOSPEC
231 : "=a" (hv_status), ASM_CALL_CONSTRAINT, 233 : "=a" (hv_status), ASM_CALL_CONSTRAINT,
232 "+c" (control), "+d" (input1) 234 "+c" (control), "+d" (input1)
233 : "m" (hv_hypercall_pg) 235 : THUNK_TARGET(hv_hypercall_pg)
234 : "cc", "r8", "r9", "r10", "r11"); 236 : "cc", "r8", "r9", "r10", "r11");
235 } 237 }
236#else 238#else
@@ -238,13 +240,13 @@ static inline u64 hv_do_fast_hypercall8(u16 code, u64 input1)
238 u32 input1_hi = upper_32_bits(input1); 240 u32 input1_hi = upper_32_bits(input1);
239 u32 input1_lo = lower_32_bits(input1); 241 u32 input1_lo = lower_32_bits(input1);
240 242
241 __asm__ __volatile__ ("call *%5" 243 __asm__ __volatile__ (CALL_NOSPEC
242 : "=A"(hv_status), 244 : "=A"(hv_status),
243 "+c"(input1_lo), 245 "+c"(input1_lo),
244 ASM_CALL_CONSTRAINT 246 ASM_CALL_CONSTRAINT
245 : "A" (control), 247 : "A" (control),
246 "b" (input1_hi), 248 "b" (input1_hi),
247 "m" (hv_hypercall_pg) 249 THUNK_TARGET(hv_hypercall_pg)
248 : "cc", "edi", "esi"); 250 : "cc", "edi", "esi");
249 } 251 }
250#endif 252#endif
diff --git a/arch/x86/include/asm/msr-index.h b/arch/x86/include/asm/msr-index.h
index 34c4922bbc3f..e7b983a35506 100644
--- a/arch/x86/include/asm/msr-index.h
+++ b/arch/x86/include/asm/msr-index.h
@@ -355,6 +355,9 @@
355#define FAM10H_MMIO_CONF_BASE_MASK 0xfffffffULL 355#define FAM10H_MMIO_CONF_BASE_MASK 0xfffffffULL
356#define FAM10H_MMIO_CONF_BASE_SHIFT 20 356#define FAM10H_MMIO_CONF_BASE_SHIFT 20
357#define MSR_FAM10H_NODE_ID 0xc001100c 357#define MSR_FAM10H_NODE_ID 0xc001100c
358#define MSR_F10H_DECFG 0xc0011029
359#define MSR_F10H_DECFG_LFENCE_SERIALIZE_BIT 1
360#define MSR_F10H_DECFG_LFENCE_SERIALIZE BIT_ULL(MSR_F10H_DECFG_LFENCE_SERIALIZE_BIT)
358 361
359/* K8 MSRs */ 362/* K8 MSRs */
360#define MSR_K8_TOP_MEM1 0xc001001a 363#define MSR_K8_TOP_MEM1 0xc001001a
diff --git a/arch/x86/include/asm/nospec-branch.h b/arch/x86/include/asm/nospec-branch.h
new file mode 100644
index 000000000000..402a11c803c3
--- /dev/null
+++ b/arch/x86/include/asm/nospec-branch.h
@@ -0,0 +1,214 @@
1/* SPDX-License-Identifier: GPL-2.0 */
2
3#ifndef __NOSPEC_BRANCH_H__
4#define __NOSPEC_BRANCH_H__
5
6#include <asm/alternative.h>
7#include <asm/alternative-asm.h>
8#include <asm/cpufeatures.h>
9
10/*
11 * Fill the CPU return stack buffer.
12 *
13 * Each entry in the RSB, if used for a speculative 'ret', contains an
14 * infinite 'pause; jmp' loop to capture speculative execution.
15 *
16 * This is required in various cases for retpoline and IBRS-based
17 * mitigations for the Spectre variant 2 vulnerability. Sometimes to
18 * eliminate potentially bogus entries from the RSB, and sometimes
19 * purely to ensure that it doesn't get empty, which on some CPUs would
20 * allow predictions from other (unwanted!) sources to be used.
21 *
22 * We define a CPP macro such that it can be used from both .S files and
23 * inline assembly. It's possible to do a .macro and then include that
24 * from C via asm(".include <asm/nospec-branch.h>") but let's not go there.
25 */
26
27#define RSB_CLEAR_LOOPS 32 /* To forcibly overwrite all entries */
28#define RSB_FILL_LOOPS 16 /* To avoid underflow */
29
30/*
31 * Google experimented with loop-unrolling and this turned out to be
32 * the optimal version — two calls, each with their own speculation
33 * trap should their return address end up getting used, in a loop.
34 */
35#define __FILL_RETURN_BUFFER(reg, nr, sp) \
36 mov $(nr/2), reg; \
37771: \
38 call 772f; \
39773: /* speculation trap */ \
40 pause; \
41 jmp 773b; \
42772: \
43 call 774f; \
44775: /* speculation trap */ \
45 pause; \
46 jmp 775b; \
47774: \
48 dec reg; \
49 jnz 771b; \
50 add $(BITS_PER_LONG/8) * nr, sp;
51
52#ifdef __ASSEMBLY__
53
54/*
55 * This should be used immediately before a retpoline alternative. It tells
56 * objtool where the retpolines are so that it can make sense of the control
57 * flow by just reading the original instruction(s) and ignoring the
58 * alternatives.
59 */
60.macro ANNOTATE_NOSPEC_ALTERNATIVE
61 .Lannotate_\@:
62 .pushsection .discard.nospec
63 .long .Lannotate_\@ - .
64 .popsection
65.endm
66
67/*
68 * These are the bare retpoline primitives for indirect jmp and call.
69 * Do not use these directly; they only exist to make the ALTERNATIVE
70 * invocation below less ugly.
71 */
72.macro RETPOLINE_JMP reg:req
73 call .Ldo_rop_\@
74.Lspec_trap_\@:
75 pause
76 jmp .Lspec_trap_\@
77.Ldo_rop_\@:
78 mov \reg, (%_ASM_SP)
79 ret
80.endm
81
82/*
83 * This is a wrapper around RETPOLINE_JMP so the called function in reg
84 * returns to the instruction after the macro.
85 */
86.macro RETPOLINE_CALL reg:req
87 jmp .Ldo_call_\@
88.Ldo_retpoline_jmp_\@:
89 RETPOLINE_JMP \reg
90.Ldo_call_\@:
91 call .Ldo_retpoline_jmp_\@
92.endm
93
94/*
95 * JMP_NOSPEC and CALL_NOSPEC macros can be used instead of a simple
96 * indirect jmp/call which may be susceptible to the Spectre variant 2
97 * attack.
98 */
99.macro JMP_NOSPEC reg:req
100#ifdef CONFIG_RETPOLINE
101 ANNOTATE_NOSPEC_ALTERNATIVE
102 ALTERNATIVE_2 __stringify(jmp *\reg), \
103 __stringify(RETPOLINE_JMP \reg), X86_FEATURE_RETPOLINE, \
104 __stringify(lfence; jmp *\reg), X86_FEATURE_RETPOLINE_AMD
105#else
106 jmp *\reg
107#endif
108.endm
109
110.macro CALL_NOSPEC reg:req
111#ifdef CONFIG_RETPOLINE
112 ANNOTATE_NOSPEC_ALTERNATIVE
113 ALTERNATIVE_2 __stringify(call *\reg), \
114 __stringify(RETPOLINE_CALL \reg), X86_FEATURE_RETPOLINE,\
115 __stringify(lfence; call *\reg), X86_FEATURE_RETPOLINE_AMD
116#else
117 call *\reg
118#endif
119.endm
120
121 /*
122 * A simpler FILL_RETURN_BUFFER macro. Don't make people use the CPP
123 * monstrosity above, manually.
124 */
125.macro FILL_RETURN_BUFFER reg:req nr:req ftr:req
126#ifdef CONFIG_RETPOLINE
127 ANNOTATE_NOSPEC_ALTERNATIVE
128 ALTERNATIVE "jmp .Lskip_rsb_\@", \
129 __stringify(__FILL_RETURN_BUFFER(\reg,\nr,%_ASM_SP)) \
130 \ftr
131.Lskip_rsb_\@:
132#endif
133.endm
134
135#else /* __ASSEMBLY__ */
136
137#define ANNOTATE_NOSPEC_ALTERNATIVE \
138 "999:\n\t" \
139 ".pushsection .discard.nospec\n\t" \
140 ".long 999b - .\n\t" \
141 ".popsection\n\t"
142
143#if defined(CONFIG_X86_64) && defined(RETPOLINE)
144
145/*
146 * Since the inline asm uses the %V modifier which is only in newer GCC,
147 * the 64-bit one is dependent on RETPOLINE not CONFIG_RETPOLINE.
148 */
149# define CALL_NOSPEC \
150 ANNOTATE_NOSPEC_ALTERNATIVE \
151 ALTERNATIVE( \
152 "call *%[thunk_target]\n", \
153 "call __x86_indirect_thunk_%V[thunk_target]\n", \
154 X86_FEATURE_RETPOLINE)
155# define THUNK_TARGET(addr) [thunk_target] "r" (addr)
156
157#elif defined(CONFIG_X86_32) && defined(CONFIG_RETPOLINE)
158/*
159 * For i386 we use the original ret-equivalent retpoline, because
160 * otherwise we'll run out of registers. We don't care about CET
161 * here, anyway.
162 */
163# define CALL_NOSPEC ALTERNATIVE("call *%[thunk_target]\n", \
164 " jmp 904f;\n" \
165 " .align 16\n" \
166 "901: call 903f;\n" \
167 "902: pause;\n" \
168 " jmp 902b;\n" \
169 " .align 16\n" \
170 "903: addl $4, %%esp;\n" \
171 " pushl %[thunk_target];\n" \
172 " ret;\n" \
173 " .align 16\n" \
174 "904: call 901b;\n", \
175 X86_FEATURE_RETPOLINE)
176
177# define THUNK_TARGET(addr) [thunk_target] "rm" (addr)
178#else /* No retpoline for C / inline asm */
179# define CALL_NOSPEC "call *%[thunk_target]\n"
180# define THUNK_TARGET(addr) [thunk_target] "rm" (addr)
181#endif
182
183/* The Spectre V2 mitigation variants */
184enum spectre_v2_mitigation {
185 SPECTRE_V2_NONE,
186 SPECTRE_V2_RETPOLINE_MINIMAL,
187 SPECTRE_V2_RETPOLINE_MINIMAL_AMD,
188 SPECTRE_V2_RETPOLINE_GENERIC,
189 SPECTRE_V2_RETPOLINE_AMD,
190 SPECTRE_V2_IBRS,
191};
192
193/*
194 * On VMEXIT we must ensure that no RSB predictions learned in the guest
195 * can be followed in the host, by overwriting the RSB completely. Both
196 * retpoline and IBRS mitigations for Spectre v2 need this; only on future
197 * CPUs with IBRS_ATT *might* it be avoided.
198 */
199static inline void vmexit_fill_RSB(void)
200{
201#ifdef CONFIG_RETPOLINE
202 unsigned long loops = RSB_CLEAR_LOOPS / 2;
203
204 asm volatile (ANNOTATE_NOSPEC_ALTERNATIVE
205 ALTERNATIVE("jmp 910f",
206 __stringify(__FILL_RETURN_BUFFER(%0, RSB_CLEAR_LOOPS, %1)),
207 X86_FEATURE_RETPOLINE)
208 "910:"
209 : "=&r" (loops), ASM_CALL_CONSTRAINT
210 : "r" (loops) : "memory" );
211#endif
212}
213#endif /* __ASSEMBLY__ */
214#endif /* __NOSPEC_BRANCH_H__ */
diff --git a/arch/x86/include/asm/pci_x86.h b/arch/x86/include/asm/pci_x86.h
index 7a5d6695abd3..eb66fa9cd0fc 100644
--- a/arch/x86/include/asm/pci_x86.h
+++ b/arch/x86/include/asm/pci_x86.h
@@ -38,6 +38,7 @@ do { \
38#define PCI_NOASSIGN_ROMS 0x80000 38#define PCI_NOASSIGN_ROMS 0x80000
39#define PCI_ROOT_NO_CRS 0x100000 39#define PCI_ROOT_NO_CRS 0x100000
40#define PCI_NOASSIGN_BARS 0x200000 40#define PCI_NOASSIGN_BARS 0x200000
41#define PCI_BIG_ROOT_WINDOW 0x400000
41 42
42extern unsigned int pci_probe; 43extern unsigned int pci_probe;
43extern unsigned long pirq_table_addr; 44extern unsigned long pirq_table_addr;
diff --git a/arch/x86/include/asm/processor-flags.h b/arch/x86/include/asm/processor-flags.h
index 6a60fea90b9d..625a52a5594f 100644
--- a/arch/x86/include/asm/processor-flags.h
+++ b/arch/x86/include/asm/processor-flags.h
@@ -40,7 +40,7 @@
40#define CR3_NOFLUSH BIT_ULL(63) 40#define CR3_NOFLUSH BIT_ULL(63)
41 41
42#ifdef CONFIG_PAGE_TABLE_ISOLATION 42#ifdef CONFIG_PAGE_TABLE_ISOLATION
43# define X86_CR3_PTI_SWITCH_BIT 11 43# define X86_CR3_PTI_PCID_USER_BIT 11
44#endif 44#endif
45 45
46#else 46#else
diff --git a/arch/x86/include/asm/tlbflush.h b/arch/x86/include/asm/tlbflush.h
index 4a08dd2ab32a..d33e4a26dc7e 100644
--- a/arch/x86/include/asm/tlbflush.h
+++ b/arch/x86/include/asm/tlbflush.h
@@ -81,13 +81,13 @@ static inline u16 kern_pcid(u16 asid)
81 * Make sure that the dynamic ASID space does not confict with the 81 * Make sure that the dynamic ASID space does not confict with the
82 * bit we are using to switch between user and kernel ASIDs. 82 * bit we are using to switch between user and kernel ASIDs.
83 */ 83 */
84 BUILD_BUG_ON(TLB_NR_DYN_ASIDS >= (1 << X86_CR3_PTI_SWITCH_BIT)); 84 BUILD_BUG_ON(TLB_NR_DYN_ASIDS >= (1 << X86_CR3_PTI_PCID_USER_BIT));
85 85
86 /* 86 /*
87 * The ASID being passed in here should have respected the 87 * The ASID being passed in here should have respected the
88 * MAX_ASID_AVAILABLE and thus never have the switch bit set. 88 * MAX_ASID_AVAILABLE and thus never have the switch bit set.
89 */ 89 */
90 VM_WARN_ON_ONCE(asid & (1 << X86_CR3_PTI_SWITCH_BIT)); 90 VM_WARN_ON_ONCE(asid & (1 << X86_CR3_PTI_PCID_USER_BIT));
91#endif 91#endif
92 /* 92 /*
93 * The dynamically-assigned ASIDs that get passed in are small 93 * The dynamically-assigned ASIDs that get passed in are small
@@ -112,7 +112,7 @@ static inline u16 user_pcid(u16 asid)
112{ 112{
113 u16 ret = kern_pcid(asid); 113 u16 ret = kern_pcid(asid);
114#ifdef CONFIG_PAGE_TABLE_ISOLATION 114#ifdef CONFIG_PAGE_TABLE_ISOLATION
115 ret |= 1 << X86_CR3_PTI_SWITCH_BIT; 115 ret |= 1 << X86_CR3_PTI_PCID_USER_BIT;
116#endif 116#endif
117 return ret; 117 return ret;
118} 118}
diff --git a/arch/x86/include/asm/xen/hypercall.h b/arch/x86/include/asm/xen/hypercall.h
index 7cb282e9e587..bfd882617613 100644
--- a/arch/x86/include/asm/xen/hypercall.h
+++ b/arch/x86/include/asm/xen/hypercall.h
@@ -44,6 +44,7 @@
44#include <asm/page.h> 44#include <asm/page.h>
45#include <asm/pgtable.h> 45#include <asm/pgtable.h>
46#include <asm/smap.h> 46#include <asm/smap.h>
47#include <asm/nospec-branch.h>
47 48
48#include <xen/interface/xen.h> 49#include <xen/interface/xen.h>
49#include <xen/interface/sched.h> 50#include <xen/interface/sched.h>
@@ -217,9 +218,9 @@ privcmd_call(unsigned call,
217 __HYPERCALL_5ARG(a1, a2, a3, a4, a5); 218 __HYPERCALL_5ARG(a1, a2, a3, a4, a5);
218 219
219 stac(); 220 stac();
220 asm volatile("call *%[call]" 221 asm volatile(CALL_NOSPEC
221 : __HYPERCALL_5PARAM 222 : __HYPERCALL_5PARAM
222 : [call] "a" (&hypercall_page[call]) 223 : [thunk_target] "a" (&hypercall_page[call])
223 : __HYPERCALL_CLOBBER5); 224 : __HYPERCALL_CLOBBER5);
224 clac(); 225 clac();
225 226
diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c
index dbaf14d69ebd..4817d743c263 100644
--- a/arch/x86/kernel/alternative.c
+++ b/arch/x86/kernel/alternative.c
@@ -344,9 +344,12 @@ done:
344static void __init_or_module noinline optimize_nops(struct alt_instr *a, u8 *instr) 344static void __init_or_module noinline optimize_nops(struct alt_instr *a, u8 *instr)
345{ 345{
346 unsigned long flags; 346 unsigned long flags;
347 int i;
347 348
348 if (instr[0] != 0x90) 349 for (i = 0; i < a->padlen; i++) {
349 return; 350 if (instr[i] != 0x90)
351 return;
352 }
350 353
351 local_irq_save(flags); 354 local_irq_save(flags);
352 add_nops(instr + (a->instrlen - a->padlen), a->padlen); 355 add_nops(instr + (a->instrlen - a->padlen), a->padlen);
diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
index bcb75dc97d44..ea831c858195 100644
--- a/arch/x86/kernel/cpu/amd.c
+++ b/arch/x86/kernel/cpu/amd.c
@@ -829,8 +829,32 @@ static void init_amd(struct cpuinfo_x86 *c)
829 set_cpu_cap(c, X86_FEATURE_K8); 829 set_cpu_cap(c, X86_FEATURE_K8);
830 830
831 if (cpu_has(c, X86_FEATURE_XMM2)) { 831 if (cpu_has(c, X86_FEATURE_XMM2)) {
832 /* MFENCE stops RDTSC speculation */ 832 unsigned long long val;
833 set_cpu_cap(c, X86_FEATURE_MFENCE_RDTSC); 833 int ret;
834
835 /*
836 * A serializing LFENCE has less overhead than MFENCE, so
837 * use it for execution serialization. On families which
838 * don't have that MSR, LFENCE is already serializing.
839 * msr_set_bit() uses the safe accessors, too, even if the MSR
840 * is not present.
841 */
842 msr_set_bit(MSR_F10H_DECFG,
843 MSR_F10H_DECFG_LFENCE_SERIALIZE_BIT);
844
845 /*
846 * Verify that the MSR write was successful (could be running
847 * under a hypervisor) and only then assume that LFENCE is
848 * serializing.
849 */
850 ret = rdmsrl_safe(MSR_F10H_DECFG, &val);
851 if (!ret && (val & MSR_F10H_DECFG_LFENCE_SERIALIZE)) {
852 /* A serializing LFENCE stops RDTSC speculation */
853 set_cpu_cap(c, X86_FEATURE_LFENCE_RDTSC);
854 } else {
855 /* MFENCE stops RDTSC speculation */
856 set_cpu_cap(c, X86_FEATURE_MFENCE_RDTSC);
857 }
834 } 858 }
835 859
836 /* 860 /*
diff --git a/arch/x86/kernel/cpu/bugs.c b/arch/x86/kernel/cpu/bugs.c
index ba0b2424c9b0..e4dc26185aa7 100644
--- a/arch/x86/kernel/cpu/bugs.c
+++ b/arch/x86/kernel/cpu/bugs.c
@@ -10,6 +10,10 @@
10 */ 10 */
11#include <linux/init.h> 11#include <linux/init.h>
12#include <linux/utsname.h> 12#include <linux/utsname.h>
13#include <linux/cpu.h>
14
15#include <asm/nospec-branch.h>
16#include <asm/cmdline.h>
13#include <asm/bugs.h> 17#include <asm/bugs.h>
14#include <asm/processor.h> 18#include <asm/processor.h>
15#include <asm/processor-flags.h> 19#include <asm/processor-flags.h>
@@ -20,6 +24,8 @@
20#include <asm/pgtable.h> 24#include <asm/pgtable.h>
21#include <asm/set_memory.h> 25#include <asm/set_memory.h>
22 26
27static void __init spectre_v2_select_mitigation(void);
28
23void __init check_bugs(void) 29void __init check_bugs(void)
24{ 30{
25 identify_boot_cpu(); 31 identify_boot_cpu();
@@ -29,6 +35,9 @@ void __init check_bugs(void)
29 print_cpu_info(&boot_cpu_data); 35 print_cpu_info(&boot_cpu_data);
30 } 36 }
31 37
38 /* Select the proper spectre mitigation before patching alternatives */
39 spectre_v2_select_mitigation();
40
32#ifdef CONFIG_X86_32 41#ifdef CONFIG_X86_32
33 /* 42 /*
34 * Check whether we are able to run this kernel safely on SMP. 43 * Check whether we are able to run this kernel safely on SMP.
@@ -60,3 +69,179 @@ void __init check_bugs(void)
60 set_memory_4k((unsigned long)__va(0), 1); 69 set_memory_4k((unsigned long)__va(0), 1);
61#endif 70#endif
62} 71}
72
73/* The kernel command line selection */
74enum spectre_v2_mitigation_cmd {
75 SPECTRE_V2_CMD_NONE,
76 SPECTRE_V2_CMD_AUTO,
77 SPECTRE_V2_CMD_FORCE,
78 SPECTRE_V2_CMD_RETPOLINE,
79 SPECTRE_V2_CMD_RETPOLINE_GENERIC,
80 SPECTRE_V2_CMD_RETPOLINE_AMD,
81};
82
83static const char *spectre_v2_strings[] = {
84 [SPECTRE_V2_NONE] = "Vulnerable",
85 [SPECTRE_V2_RETPOLINE_MINIMAL] = "Vulnerable: Minimal generic ASM retpoline",
86 [SPECTRE_V2_RETPOLINE_MINIMAL_AMD] = "Vulnerable: Minimal AMD ASM retpoline",
87 [SPECTRE_V2_RETPOLINE_GENERIC] = "Mitigation: Full generic retpoline",
88 [SPECTRE_V2_RETPOLINE_AMD] = "Mitigation: Full AMD retpoline",
89};
90
91#undef pr_fmt
92#define pr_fmt(fmt) "Spectre V2 mitigation: " fmt
93
94static enum spectre_v2_mitigation spectre_v2_enabled = SPECTRE_V2_NONE;
95
96static void __init spec2_print_if_insecure(const char *reason)
97{
98 if (boot_cpu_has_bug(X86_BUG_SPECTRE_V2))
99 pr_info("%s\n", reason);
100}
101
102static void __init spec2_print_if_secure(const char *reason)
103{
104 if (!boot_cpu_has_bug(X86_BUG_SPECTRE_V2))
105 pr_info("%s\n", reason);
106}
107
108static inline bool retp_compiler(void)
109{
110 return __is_defined(RETPOLINE);
111}
112
113static inline bool match_option(const char *arg, int arglen, const char *opt)
114{
115 int len = strlen(opt);
116
117 return len == arglen && !strncmp(arg, opt, len);
118}
119
120static enum spectre_v2_mitigation_cmd __init spectre_v2_parse_cmdline(void)
121{
122 char arg[20];
123 int ret;
124
125 ret = cmdline_find_option(boot_command_line, "spectre_v2", arg,
126 sizeof(arg));
127 if (ret > 0) {
128 if (match_option(arg, ret, "off")) {
129 goto disable;
130 } else if (match_option(arg, ret, "on")) {
131 spec2_print_if_secure("force enabled on command line.");
132 return SPECTRE_V2_CMD_FORCE;
133 } else if (match_option(arg, ret, "retpoline")) {
134 spec2_print_if_insecure("retpoline selected on command line.");
135 return SPECTRE_V2_CMD_RETPOLINE;
136 } else if (match_option(arg, ret, "retpoline,amd")) {
137 if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD) {
138 pr_err("retpoline,amd selected but CPU is not AMD. Switching to AUTO select\n");
139 return SPECTRE_V2_CMD_AUTO;
140 }
141 spec2_print_if_insecure("AMD retpoline selected on command line.");
142 return SPECTRE_V2_CMD_RETPOLINE_AMD;
143 } else if (match_option(arg, ret, "retpoline,generic")) {
144 spec2_print_if_insecure("generic retpoline selected on command line.");
145 return SPECTRE_V2_CMD_RETPOLINE_GENERIC;
146 } else if (match_option(arg, ret, "auto")) {
147 return SPECTRE_V2_CMD_AUTO;
148 }
149 }
150
151 if (!cmdline_find_option_bool(boot_command_line, "nospectre_v2"))
152 return SPECTRE_V2_CMD_AUTO;
153disable:
154 spec2_print_if_insecure("disabled on command line.");
155 return SPECTRE_V2_CMD_NONE;
156}
157
158static void __init spectre_v2_select_mitigation(void)
159{
160 enum spectre_v2_mitigation_cmd cmd = spectre_v2_parse_cmdline();
161 enum spectre_v2_mitigation mode = SPECTRE_V2_NONE;
162
163 /*
164 * If the CPU is not affected and the command line mode is NONE or AUTO
165 * then nothing to do.
166 */
167 if (!boot_cpu_has_bug(X86_BUG_SPECTRE_V2) &&
168 (cmd == SPECTRE_V2_CMD_NONE || cmd == SPECTRE_V2_CMD_AUTO))
169 return;
170
171 switch (cmd) {
172 case SPECTRE_V2_CMD_NONE:
173 return;
174
175 case SPECTRE_V2_CMD_FORCE:
176 /* FALLTRHU */
177 case SPECTRE_V2_CMD_AUTO:
178 goto retpoline_auto;
179
180 case SPECTRE_V2_CMD_RETPOLINE_AMD:
181 if (IS_ENABLED(CONFIG_RETPOLINE))
182 goto retpoline_amd;
183 break;
184 case SPECTRE_V2_CMD_RETPOLINE_GENERIC:
185 if (IS_ENABLED(CONFIG_RETPOLINE))
186 goto retpoline_generic;
187 break;
188 case SPECTRE_V2_CMD_RETPOLINE:
189 if (IS_ENABLED(CONFIG_RETPOLINE))
190 goto retpoline_auto;
191 break;
192 }
193 pr_err("kernel not compiled with retpoline; no mitigation available!");
194 return;
195
196retpoline_auto:
197 if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD) {
198 retpoline_amd:
199 if (!boot_cpu_has(X86_FEATURE_LFENCE_RDTSC)) {
200 pr_err("LFENCE not serializing. Switching to generic retpoline\n");
201 goto retpoline_generic;
202 }
203 mode = retp_compiler() ? SPECTRE_V2_RETPOLINE_AMD :
204 SPECTRE_V2_RETPOLINE_MINIMAL_AMD;
205 setup_force_cpu_cap(X86_FEATURE_RETPOLINE_AMD);
206 setup_force_cpu_cap(X86_FEATURE_RETPOLINE);
207 } else {
208 retpoline_generic:
209 mode = retp_compiler() ? SPECTRE_V2_RETPOLINE_GENERIC :
210 SPECTRE_V2_RETPOLINE_MINIMAL;
211 setup_force_cpu_cap(X86_FEATURE_RETPOLINE);
212 }
213
214 spectre_v2_enabled = mode;
215 pr_info("%s\n", spectre_v2_strings[mode]);
216}
217
218#undef pr_fmt
219
220#ifdef CONFIG_SYSFS
221ssize_t cpu_show_meltdown(struct device *dev,
222 struct device_attribute *attr, char *buf)
223{
224 if (!boot_cpu_has_bug(X86_BUG_CPU_MELTDOWN))
225 return sprintf(buf, "Not affected\n");
226 if (boot_cpu_has(X86_FEATURE_PTI))
227 return sprintf(buf, "Mitigation: PTI\n");
228 return sprintf(buf, "Vulnerable\n");
229}
230
231ssize_t cpu_show_spectre_v1(struct device *dev,
232 struct device_attribute *attr, char *buf)
233{
234 if (!boot_cpu_has_bug(X86_BUG_SPECTRE_V1))
235 return sprintf(buf, "Not affected\n");
236 return sprintf(buf, "Vulnerable\n");
237}
238
239ssize_t cpu_show_spectre_v2(struct device *dev,
240 struct device_attribute *attr, char *buf)
241{
242 if (!boot_cpu_has_bug(X86_BUG_SPECTRE_V2))
243 return sprintf(buf, "Not affected\n");
244
245 return sprintf(buf, "%s\n", spectre_v2_strings[spectre_v2_enabled]);
246}
247#endif
diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index 39d7ea865207..ef29ad001991 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -926,6 +926,9 @@ static void __init early_identify_cpu(struct cpuinfo_x86 *c)
926 if (c->x86_vendor != X86_VENDOR_AMD) 926 if (c->x86_vendor != X86_VENDOR_AMD)
927 setup_force_cpu_bug(X86_BUG_CPU_MELTDOWN); 927 setup_force_cpu_bug(X86_BUG_CPU_MELTDOWN);
928 928
929 setup_force_cpu_bug(X86_BUG_SPECTRE_V1);
930 setup_force_cpu_bug(X86_BUG_SPECTRE_V2);
931
929 fpu__init_system(c); 932 fpu__init_system(c);
930 933
931#ifdef CONFIG_X86_32 934#ifdef CONFIG_X86_32
diff --git a/arch/x86/kernel/cpu/microcode/intel.c b/arch/x86/kernel/cpu/microcode/intel.c
index 8ccdca6d3f9e..d9e460fc7a3b 100644
--- a/arch/x86/kernel/cpu/microcode/intel.c
+++ b/arch/x86/kernel/cpu/microcode/intel.c
@@ -910,8 +910,17 @@ static bool is_blacklisted(unsigned int cpu)
910{ 910{
911 struct cpuinfo_x86 *c = &cpu_data(cpu); 911 struct cpuinfo_x86 *c = &cpu_data(cpu);
912 912
913 if (c->x86 == 6 && c->x86_model == INTEL_FAM6_BROADWELL_X) { 913 /*
914 pr_err_once("late loading on model 79 is disabled.\n"); 914 * Late loading on model 79 with microcode revision less than 0x0b000021
915 * may result in a system hang. This behavior is documented in item
916 * BDF90, #334165 (Intel Xeon Processor E7-8800/4800 v4 Product Family).
917 */
918 if (c->x86 == 6 &&
919 c->x86_model == INTEL_FAM6_BROADWELL_X &&
920 c->x86_mask == 0x01 &&
921 c->microcode < 0x0b000021) {
922 pr_err_once("Erratum BDF90: late loading with revision < 0x0b000021 (0x%x) disabled.\n", c->microcode);
923 pr_err_once("Please consider either early loading through initrd/built-in or a potential BIOS update.\n");
915 return true; 924 return true;
916 } 925 }
917 926
diff --git a/arch/x86/kernel/ftrace_32.S b/arch/x86/kernel/ftrace_32.S
index b6c6468e10bc..4c8440de3355 100644
--- a/arch/x86/kernel/ftrace_32.S
+++ b/arch/x86/kernel/ftrace_32.S
@@ -8,6 +8,7 @@
8#include <asm/segment.h> 8#include <asm/segment.h>
9#include <asm/export.h> 9#include <asm/export.h>
10#include <asm/ftrace.h> 10#include <asm/ftrace.h>
11#include <asm/nospec-branch.h>
11 12
12#ifdef CC_USING_FENTRY 13#ifdef CC_USING_FENTRY
13# define function_hook __fentry__ 14# define function_hook __fentry__
@@ -197,7 +198,8 @@ ftrace_stub:
197 movl 0x4(%ebp), %edx 198 movl 0x4(%ebp), %edx
198 subl $MCOUNT_INSN_SIZE, %eax 199 subl $MCOUNT_INSN_SIZE, %eax
199 200
200 call *ftrace_trace_function 201 movl ftrace_trace_function, %ecx
202 CALL_NOSPEC %ecx
201 203
202 popl %edx 204 popl %edx
203 popl %ecx 205 popl %ecx
@@ -241,5 +243,5 @@ return_to_handler:
241 movl %eax, %ecx 243 movl %eax, %ecx
242 popl %edx 244 popl %edx
243 popl %eax 245 popl %eax
244 jmp *%ecx 246 JMP_NOSPEC %ecx
245#endif 247#endif
diff --git a/arch/x86/kernel/ftrace_64.S b/arch/x86/kernel/ftrace_64.S
index c832291d948a..7cb8ba08beb9 100644
--- a/arch/x86/kernel/ftrace_64.S
+++ b/arch/x86/kernel/ftrace_64.S
@@ -7,7 +7,7 @@
7#include <asm/ptrace.h> 7#include <asm/ptrace.h>
8#include <asm/ftrace.h> 8#include <asm/ftrace.h>
9#include <asm/export.h> 9#include <asm/export.h>
10 10#include <asm/nospec-branch.h>
11 11
12 .code64 12 .code64
13 .section .entry.text, "ax" 13 .section .entry.text, "ax"
@@ -286,8 +286,8 @@ trace:
286 * ip and parent ip are used and the list function is called when 286 * ip and parent ip are used and the list function is called when
287 * function tracing is enabled. 287 * function tracing is enabled.
288 */ 288 */
289 call *ftrace_trace_function 289 movq ftrace_trace_function, %r8
290 290 CALL_NOSPEC %r8
291 restore_mcount_regs 291 restore_mcount_regs
292 292
293 jmp fgraph_trace 293 jmp fgraph_trace
@@ -329,5 +329,5 @@ GLOBAL(return_to_handler)
329 movq 8(%rsp), %rdx 329 movq 8(%rsp), %rdx
330 movq (%rsp), %rax 330 movq (%rsp), %rax
331 addq $24, %rsp 331 addq $24, %rsp
332 jmp *%rdi 332 JMP_NOSPEC %rdi
333#endif 333#endif
diff --git a/arch/x86/kernel/irq_32.c b/arch/x86/kernel/irq_32.c
index a83b3346a0e1..c1bdbd3d3232 100644
--- a/arch/x86/kernel/irq_32.c
+++ b/arch/x86/kernel/irq_32.c
@@ -20,6 +20,7 @@
20#include <linux/mm.h> 20#include <linux/mm.h>
21 21
22#include <asm/apic.h> 22#include <asm/apic.h>
23#include <asm/nospec-branch.h>
23 24
24#ifdef CONFIG_DEBUG_STACKOVERFLOW 25#ifdef CONFIG_DEBUG_STACKOVERFLOW
25 26
@@ -55,11 +56,11 @@ DEFINE_PER_CPU(struct irq_stack *, softirq_stack);
55static void call_on_stack(void *func, void *stack) 56static void call_on_stack(void *func, void *stack)
56{ 57{
57 asm volatile("xchgl %%ebx,%%esp \n" 58 asm volatile("xchgl %%ebx,%%esp \n"
58 "call *%%edi \n" 59 CALL_NOSPEC
59 "movl %%ebx,%%esp \n" 60 "movl %%ebx,%%esp \n"
60 : "=b" (stack) 61 : "=b" (stack)
61 : "0" (stack), 62 : "0" (stack),
62 "D"(func) 63 [thunk_target] "D"(func)
63 : "memory", "cc", "edx", "ecx", "eax"); 64 : "memory", "cc", "edx", "ecx", "eax");
64} 65}
65 66
@@ -95,11 +96,11 @@ static inline int execute_on_irq_stack(int overflow, struct irq_desc *desc)
95 call_on_stack(print_stack_overflow, isp); 96 call_on_stack(print_stack_overflow, isp);
96 97
97 asm volatile("xchgl %%ebx,%%esp \n" 98 asm volatile("xchgl %%ebx,%%esp \n"
98 "call *%%edi \n" 99 CALL_NOSPEC
99 "movl %%ebx,%%esp \n" 100 "movl %%ebx,%%esp \n"
100 : "=a" (arg1), "=b" (isp) 101 : "=a" (arg1), "=b" (isp)
101 : "0" (desc), "1" (isp), 102 : "0" (desc), "1" (isp),
102 "D" (desc->handle_irq) 103 [thunk_target] "D" (desc->handle_irq)
103 : "memory", "cc", "ecx"); 104 : "memory", "cc", "ecx");
104 return 1; 105 return 1;
105} 106}
diff --git a/arch/x86/kernel/tboot.c b/arch/x86/kernel/tboot.c
index a4eb27918ceb..a2486f444073 100644
--- a/arch/x86/kernel/tboot.c
+++ b/arch/x86/kernel/tboot.c
@@ -138,6 +138,17 @@ static int map_tboot_page(unsigned long vaddr, unsigned long pfn,
138 return -1; 138 return -1;
139 set_pte_at(&tboot_mm, vaddr, pte, pfn_pte(pfn, prot)); 139 set_pte_at(&tboot_mm, vaddr, pte, pfn_pte(pfn, prot));
140 pte_unmap(pte); 140 pte_unmap(pte);
141
142 /*
143 * PTI poisons low addresses in the kernel page tables in the
144 * name of making them unusable for userspace. To execute
145 * code at such a low address, the poison must be cleared.
146 *
147 * Note: 'pgd' actually gets set in p4d_alloc() _or_
148 * pud_alloc() depending on 4/5-level paging.
149 */
150 pgd->pgd &= ~_PAGE_NX;
151
141 return 0; 152 return 0;
142} 153}
143 154
diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index c4deb1f34faa..2b8eb4da4d08 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -3781,7 +3781,8 @@ static int kvm_arch_setup_async_pf(struct kvm_vcpu *vcpu, gva_t gva, gfn_t gfn)
3781bool kvm_can_do_async_pf(struct kvm_vcpu *vcpu) 3781bool kvm_can_do_async_pf(struct kvm_vcpu *vcpu)
3782{ 3782{
3783 if (unlikely(!lapic_in_kernel(vcpu) || 3783 if (unlikely(!lapic_in_kernel(vcpu) ||
3784 kvm_event_needs_reinjection(vcpu))) 3784 kvm_event_needs_reinjection(vcpu) ||
3785 vcpu->arch.exception.pending))
3785 return false; 3786 return false;
3786 3787
3787 if (!vcpu->arch.apf.delivery_as_pf_vmexit && is_guest_mode(vcpu)) 3788 if (!vcpu->arch.apf.delivery_as_pf_vmexit && is_guest_mode(vcpu))
@@ -5465,30 +5466,34 @@ static void mmu_destroy_caches(void)
5465 5466
5466int kvm_mmu_module_init(void) 5467int kvm_mmu_module_init(void)
5467{ 5468{
5469 int ret = -ENOMEM;
5470
5468 kvm_mmu_clear_all_pte_masks(); 5471 kvm_mmu_clear_all_pte_masks();
5469 5472
5470 pte_list_desc_cache = kmem_cache_create("pte_list_desc", 5473 pte_list_desc_cache = kmem_cache_create("pte_list_desc",
5471 sizeof(struct pte_list_desc), 5474 sizeof(struct pte_list_desc),
5472 0, SLAB_ACCOUNT, NULL); 5475 0, SLAB_ACCOUNT, NULL);
5473 if (!pte_list_desc_cache) 5476 if (!pte_list_desc_cache)
5474 goto nomem; 5477 goto out;
5475 5478
5476 mmu_page_header_cache = kmem_cache_create("kvm_mmu_page_header", 5479 mmu_page_header_cache = kmem_cache_create("kvm_mmu_page_header",
5477 sizeof(struct kvm_mmu_page), 5480 sizeof(struct kvm_mmu_page),
5478 0, SLAB_ACCOUNT, NULL); 5481 0, SLAB_ACCOUNT, NULL);
5479 if (!mmu_page_header_cache) 5482 if (!mmu_page_header_cache)
5480 goto nomem; 5483 goto out;
5481 5484
5482 if (percpu_counter_init(&kvm_total_used_mmu_pages, 0, GFP_KERNEL)) 5485 if (percpu_counter_init(&kvm_total_used_mmu_pages, 0, GFP_KERNEL))
5483 goto nomem; 5486 goto out;
5484 5487
5485 register_shrinker(&mmu_shrinker); 5488 ret = register_shrinker(&mmu_shrinker);
5489 if (ret)
5490 goto out;
5486 5491
5487 return 0; 5492 return 0;
5488 5493
5489nomem: 5494out:
5490 mmu_destroy_caches(); 5495 mmu_destroy_caches();
5491 return -ENOMEM; 5496 return ret;
5492} 5497}
5493 5498
5494/* 5499/*
diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
index bb31c801f1fc..f40d0da1f1d3 100644
--- a/arch/x86/kvm/svm.c
+++ b/arch/x86/kvm/svm.c
@@ -45,6 +45,7 @@
45#include <asm/debugreg.h> 45#include <asm/debugreg.h>
46#include <asm/kvm_para.h> 46#include <asm/kvm_para.h>
47#include <asm/irq_remapping.h> 47#include <asm/irq_remapping.h>
48#include <asm/nospec-branch.h>
48 49
49#include <asm/virtext.h> 50#include <asm/virtext.h>
50#include "trace.h" 51#include "trace.h"
@@ -361,7 +362,6 @@ static void recalc_intercepts(struct vcpu_svm *svm)
361{ 362{
362 struct vmcb_control_area *c, *h; 363 struct vmcb_control_area *c, *h;
363 struct nested_state *g; 364 struct nested_state *g;
364 u32 h_intercept_exceptions;
365 365
366 mark_dirty(svm->vmcb, VMCB_INTERCEPTS); 366 mark_dirty(svm->vmcb, VMCB_INTERCEPTS);
367 367
@@ -372,14 +372,9 @@ static void recalc_intercepts(struct vcpu_svm *svm)
372 h = &svm->nested.hsave->control; 372 h = &svm->nested.hsave->control;
373 g = &svm->nested; 373 g = &svm->nested;
374 374
375 /* No need to intercept #UD if L1 doesn't intercept it */
376 h_intercept_exceptions =
377 h->intercept_exceptions & ~(1U << UD_VECTOR);
378
379 c->intercept_cr = h->intercept_cr | g->intercept_cr; 375 c->intercept_cr = h->intercept_cr | g->intercept_cr;
380 c->intercept_dr = h->intercept_dr | g->intercept_dr; 376 c->intercept_dr = h->intercept_dr | g->intercept_dr;
381 c->intercept_exceptions = 377 c->intercept_exceptions = h->intercept_exceptions | g->intercept_exceptions;
382 h_intercept_exceptions | g->intercept_exceptions;
383 c->intercept = h->intercept | g->intercept; 378 c->intercept = h->intercept | g->intercept;
384} 379}
385 380
@@ -2202,7 +2197,6 @@ static int ud_interception(struct vcpu_svm *svm)
2202{ 2197{
2203 int er; 2198 int er;
2204 2199
2205 WARN_ON_ONCE(is_guest_mode(&svm->vcpu));
2206 er = emulate_instruction(&svm->vcpu, EMULTYPE_TRAP_UD); 2200 er = emulate_instruction(&svm->vcpu, EMULTYPE_TRAP_UD);
2207 if (er == EMULATE_USER_EXIT) 2201 if (er == EMULATE_USER_EXIT)
2208 return 0; 2202 return 0;
@@ -5034,6 +5028,9 @@ static void svm_vcpu_run(struct kvm_vcpu *vcpu)
5034#endif 5028#endif
5035 ); 5029 );
5036 5030
5031 /* Eliminate branch target predictions from guest mode */
5032 vmexit_fill_RSB();
5033
5037#ifdef CONFIG_X86_64 5034#ifdef CONFIG_X86_64
5038 wrmsrl(MSR_GS_BASE, svm->host.gs_base); 5035 wrmsrl(MSR_GS_BASE, svm->host.gs_base);
5039#else 5036#else
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 5c14d65f676a..c829d89e2e63 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -50,6 +50,7 @@
50#include <asm/apic.h> 50#include <asm/apic.h>
51#include <asm/irq_remapping.h> 51#include <asm/irq_remapping.h>
52#include <asm/mmu_context.h> 52#include <asm/mmu_context.h>
53#include <asm/nospec-branch.h>
53 54
54#include "trace.h" 55#include "trace.h"
55#include "pmu.h" 56#include "pmu.h"
@@ -899,8 +900,16 @@ static inline short vmcs_field_to_offset(unsigned long field)
899{ 900{
900 BUILD_BUG_ON(ARRAY_SIZE(vmcs_field_to_offset_table) > SHRT_MAX); 901 BUILD_BUG_ON(ARRAY_SIZE(vmcs_field_to_offset_table) > SHRT_MAX);
901 902
902 if (field >= ARRAY_SIZE(vmcs_field_to_offset_table) || 903 if (field >= ARRAY_SIZE(vmcs_field_to_offset_table))
903 vmcs_field_to_offset_table[field] == 0) 904 return -ENOENT;
905
906 /*
907 * FIXME: Mitigation for CVE-2017-5753. To be replaced with a
908 * generic mechanism.
909 */
910 asm("lfence");
911
912 if (vmcs_field_to_offset_table[field] == 0)
904 return -ENOENT; 913 return -ENOENT;
905 914
906 return vmcs_field_to_offset_table[field]; 915 return vmcs_field_to_offset_table[field];
@@ -1887,7 +1896,7 @@ static void update_exception_bitmap(struct kvm_vcpu *vcpu)
1887{ 1896{
1888 u32 eb; 1897 u32 eb;
1889 1898
1890 eb = (1u << PF_VECTOR) | (1u << MC_VECTOR) | 1899 eb = (1u << PF_VECTOR) | (1u << UD_VECTOR) | (1u << MC_VECTOR) |
1891 (1u << DB_VECTOR) | (1u << AC_VECTOR); 1900 (1u << DB_VECTOR) | (1u << AC_VECTOR);
1892 if ((vcpu->guest_debug & 1901 if ((vcpu->guest_debug &
1893 (KVM_GUESTDBG_ENABLE | KVM_GUESTDBG_USE_SW_BP)) == 1902 (KVM_GUESTDBG_ENABLE | KVM_GUESTDBG_USE_SW_BP)) ==
@@ -1905,8 +1914,6 @@ static void update_exception_bitmap(struct kvm_vcpu *vcpu)
1905 */ 1914 */
1906 if (is_guest_mode(vcpu)) 1915 if (is_guest_mode(vcpu))
1907 eb |= get_vmcs12(vcpu)->exception_bitmap; 1916 eb |= get_vmcs12(vcpu)->exception_bitmap;
1908 else
1909 eb |= 1u << UD_VECTOR;
1910 1917
1911 vmcs_write32(EXCEPTION_BITMAP, eb); 1918 vmcs_write32(EXCEPTION_BITMAP, eb);
1912} 1919}
@@ -5917,7 +5924,6 @@ static int handle_exception(struct kvm_vcpu *vcpu)
5917 return 1; /* already handled by vmx_vcpu_run() */ 5924 return 1; /* already handled by vmx_vcpu_run() */
5918 5925
5919 if (is_invalid_opcode(intr_info)) { 5926 if (is_invalid_opcode(intr_info)) {
5920 WARN_ON_ONCE(is_guest_mode(vcpu));
5921 er = emulate_instruction(vcpu, EMULTYPE_TRAP_UD); 5927 er = emulate_instruction(vcpu, EMULTYPE_TRAP_UD);
5922 if (er == EMULATE_USER_EXIT) 5928 if (er == EMULATE_USER_EXIT)
5923 return 0; 5929 return 0;
@@ -9485,6 +9491,9 @@ static void __noclone vmx_vcpu_run(struct kvm_vcpu *vcpu)
9485#endif 9491#endif
9486 ); 9492 );
9487 9493
9494 /* Eliminate branch target predictions from guest mode */
9495 vmexit_fill_RSB();
9496
9488 /* MSR_IA32_DEBUGCTLMSR is zeroed on vmexit. Restore it if needed */ 9497 /* MSR_IA32_DEBUGCTLMSR is zeroed on vmexit. Restore it if needed */
9489 if (debugctlmsr) 9498 if (debugctlmsr)
9490 update_debugctlmsr(debugctlmsr); 9499 update_debugctlmsr(debugctlmsr);
diff --git a/arch/x86/lib/Makefile b/arch/x86/lib/Makefile
index 7b181b61170e..f23934bbaf4e 100644
--- a/arch/x86/lib/Makefile
+++ b/arch/x86/lib/Makefile
@@ -26,6 +26,7 @@ lib-y += memcpy_$(BITS).o
26lib-$(CONFIG_RWSEM_XCHGADD_ALGORITHM) += rwsem.o 26lib-$(CONFIG_RWSEM_XCHGADD_ALGORITHM) += rwsem.o
27lib-$(CONFIG_INSTRUCTION_DECODER) += insn.o inat.o insn-eval.o 27lib-$(CONFIG_INSTRUCTION_DECODER) += insn.o inat.o insn-eval.o
28lib-$(CONFIG_RANDOMIZE_BASE) += kaslr.o 28lib-$(CONFIG_RANDOMIZE_BASE) += kaslr.o
29lib-$(CONFIG_RETPOLINE) += retpoline.o
29 30
30obj-y += msr.o msr-reg.o msr-reg-export.o hweight.o 31obj-y += msr.o msr-reg.o msr-reg-export.o hweight.o
31 32
diff --git a/arch/x86/lib/checksum_32.S b/arch/x86/lib/checksum_32.S
index 4d34bb548b41..46e71a74e612 100644
--- a/arch/x86/lib/checksum_32.S
+++ b/arch/x86/lib/checksum_32.S
@@ -29,7 +29,8 @@
29#include <asm/errno.h> 29#include <asm/errno.h>
30#include <asm/asm.h> 30#include <asm/asm.h>
31#include <asm/export.h> 31#include <asm/export.h>
32 32#include <asm/nospec-branch.h>
33
33/* 34/*
34 * computes a partial checksum, e.g. for TCP/UDP fragments 35 * computes a partial checksum, e.g. for TCP/UDP fragments
35 */ 36 */
@@ -156,7 +157,7 @@ ENTRY(csum_partial)
156 negl %ebx 157 negl %ebx
157 lea 45f(%ebx,%ebx,2), %ebx 158 lea 45f(%ebx,%ebx,2), %ebx
158 testl %esi, %esi 159 testl %esi, %esi
159 jmp *%ebx 160 JMP_NOSPEC %ebx
160 161
161 # Handle 2-byte-aligned regions 162 # Handle 2-byte-aligned regions
16220: addw (%esi), %ax 16320: addw (%esi), %ax
@@ -439,7 +440,7 @@ ENTRY(csum_partial_copy_generic)
439 andl $-32,%edx 440 andl $-32,%edx
440 lea 3f(%ebx,%ebx), %ebx 441 lea 3f(%ebx,%ebx), %ebx
441 testl %esi, %esi 442 testl %esi, %esi
442 jmp *%ebx 443 JMP_NOSPEC %ebx
4431: addl $64,%esi 4441: addl $64,%esi
444 addl $64,%edi 445 addl $64,%edi
445 SRC(movb -32(%edx),%bl) ; SRC(movb (%edx),%bl) 446 SRC(movb -32(%edx),%bl) ; SRC(movb (%edx),%bl)
diff --git a/arch/x86/lib/retpoline.S b/arch/x86/lib/retpoline.S
new file mode 100644
index 000000000000..cb45c6cb465f
--- /dev/null
+++ b/arch/x86/lib/retpoline.S
@@ -0,0 +1,48 @@
1/* SPDX-License-Identifier: GPL-2.0 */
2
3#include <linux/stringify.h>
4#include <linux/linkage.h>
5#include <asm/dwarf2.h>
6#include <asm/cpufeatures.h>
7#include <asm/alternative-asm.h>
8#include <asm/export.h>
9#include <asm/nospec-branch.h>
10
11.macro THUNK reg
12 .section .text.__x86.indirect_thunk.\reg
13
14ENTRY(__x86_indirect_thunk_\reg)
15 CFI_STARTPROC
16 JMP_NOSPEC %\reg
17 CFI_ENDPROC
18ENDPROC(__x86_indirect_thunk_\reg)
19.endm
20
21/*
22 * Despite being an assembler file we can't just use .irp here
23 * because __KSYM_DEPS__ only uses the C preprocessor and would
24 * only see one instance of "__x86_indirect_thunk_\reg" rather
25 * than one per register with the correct names. So we do it
26 * the simple and nasty way...
27 */
28#define EXPORT_THUNK(reg) EXPORT_SYMBOL(__x86_indirect_thunk_ ## reg)
29#define GENERATE_THUNK(reg) THUNK reg ; EXPORT_THUNK(reg)
30
31GENERATE_THUNK(_ASM_AX)
32GENERATE_THUNK(_ASM_BX)
33GENERATE_THUNK(_ASM_CX)
34GENERATE_THUNK(_ASM_DX)
35GENERATE_THUNK(_ASM_SI)
36GENERATE_THUNK(_ASM_DI)
37GENERATE_THUNK(_ASM_BP)
38GENERATE_THUNK(_ASM_SP)
39#ifdef CONFIG_64BIT
40GENERATE_THUNK(r8)
41GENERATE_THUNK(r9)
42GENERATE_THUNK(r10)
43GENERATE_THUNK(r11)
44GENERATE_THUNK(r12)
45GENERATE_THUNK(r13)
46GENERATE_THUNK(r14)
47GENERATE_THUNK(r15)
48#endif
diff --git a/arch/x86/mm/pti.c b/arch/x86/mm/pti.c
index 43d4a4a29037..ce38f165489b 100644
--- a/arch/x86/mm/pti.c
+++ b/arch/x86/mm/pti.c
@@ -149,7 +149,7 @@ pgd_t __pti_set_user_pgd(pgd_t *pgdp, pgd_t pgd)
149 * 149 *
150 * Returns a pointer to a P4D on success, or NULL on failure. 150 * Returns a pointer to a P4D on success, or NULL on failure.
151 */ 151 */
152static p4d_t *pti_user_pagetable_walk_p4d(unsigned long address) 152static __init p4d_t *pti_user_pagetable_walk_p4d(unsigned long address)
153{ 153{
154 pgd_t *pgd = kernel_to_user_pgdp(pgd_offset_k(address)); 154 pgd_t *pgd = kernel_to_user_pgdp(pgd_offset_k(address));
155 gfp_t gfp = (GFP_KERNEL | __GFP_NOTRACK | __GFP_ZERO); 155 gfp_t gfp = (GFP_KERNEL | __GFP_NOTRACK | __GFP_ZERO);
@@ -164,12 +164,7 @@ static p4d_t *pti_user_pagetable_walk_p4d(unsigned long address)
164 if (!new_p4d_page) 164 if (!new_p4d_page)
165 return NULL; 165 return NULL;
166 166
167 if (pgd_none(*pgd)) { 167 set_pgd(pgd, __pgd(_KERNPG_TABLE | __pa(new_p4d_page)));
168 set_pgd(pgd, __pgd(_KERNPG_TABLE | __pa(new_p4d_page)));
169 new_p4d_page = 0;
170 }
171 if (new_p4d_page)
172 free_page(new_p4d_page);
173 } 168 }
174 BUILD_BUG_ON(pgd_large(*pgd) != 0); 169 BUILD_BUG_ON(pgd_large(*pgd) != 0);
175 170
@@ -182,7 +177,7 @@ static p4d_t *pti_user_pagetable_walk_p4d(unsigned long address)
182 * 177 *
183 * Returns a pointer to a PMD on success, or NULL on failure. 178 * Returns a pointer to a PMD on success, or NULL on failure.
184 */ 179 */
185static pmd_t *pti_user_pagetable_walk_pmd(unsigned long address) 180static __init pmd_t *pti_user_pagetable_walk_pmd(unsigned long address)
186{ 181{
187 gfp_t gfp = (GFP_KERNEL | __GFP_NOTRACK | __GFP_ZERO); 182 gfp_t gfp = (GFP_KERNEL | __GFP_NOTRACK | __GFP_ZERO);
188 p4d_t *p4d = pti_user_pagetable_walk_p4d(address); 183 p4d_t *p4d = pti_user_pagetable_walk_p4d(address);
@@ -194,12 +189,7 @@ static pmd_t *pti_user_pagetable_walk_pmd(unsigned long address)
194 if (!new_pud_page) 189 if (!new_pud_page)
195 return NULL; 190 return NULL;
196 191
197 if (p4d_none(*p4d)) { 192 set_p4d(p4d, __p4d(_KERNPG_TABLE | __pa(new_pud_page)));
198 set_p4d(p4d, __p4d(_KERNPG_TABLE | __pa(new_pud_page)));
199 new_pud_page = 0;
200 }
201 if (new_pud_page)
202 free_page(new_pud_page);
203 } 193 }
204 194
205 pud = pud_offset(p4d, address); 195 pud = pud_offset(p4d, address);
@@ -213,12 +203,7 @@ static pmd_t *pti_user_pagetable_walk_pmd(unsigned long address)
213 if (!new_pmd_page) 203 if (!new_pmd_page)
214 return NULL; 204 return NULL;
215 205
216 if (pud_none(*pud)) { 206 set_pud(pud, __pud(_KERNPG_TABLE | __pa(new_pmd_page)));
217 set_pud(pud, __pud(_KERNPG_TABLE | __pa(new_pmd_page)));
218 new_pmd_page = 0;
219 }
220 if (new_pmd_page)
221 free_page(new_pmd_page);
222 } 207 }
223 208
224 return pmd_offset(pud, address); 209 return pmd_offset(pud, address);
@@ -251,12 +236,7 @@ static __init pte_t *pti_user_pagetable_walk_pte(unsigned long address)
251 if (!new_pte_page) 236 if (!new_pte_page)
252 return NULL; 237 return NULL;
253 238
254 if (pmd_none(*pmd)) { 239 set_pmd(pmd, __pmd(_KERNPG_TABLE | __pa(new_pte_page)));
255 set_pmd(pmd, __pmd(_KERNPG_TABLE | __pa(new_pte_page)));
256 new_pte_page = 0;
257 }
258 if (new_pte_page)
259 free_page(new_pte_page);
260 } 240 }
261 241
262 pte = pte_offset_kernel(pmd, address); 242 pte = pte_offset_kernel(pmd, address);
diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c
index 7a5350d08cef..563049c483a1 100644
--- a/arch/x86/pci/common.c
+++ b/arch/x86/pci/common.c
@@ -594,6 +594,11 @@ char *__init pcibios_setup(char *str)
594 } else if (!strcmp(str, "nocrs")) { 594 } else if (!strcmp(str, "nocrs")) {
595 pci_probe |= PCI_ROOT_NO_CRS; 595 pci_probe |= PCI_ROOT_NO_CRS;
596 return NULL; 596 return NULL;
597#ifdef CONFIG_PHYS_ADDR_T_64BIT
598 } else if (!strcmp(str, "big_root_window")) {
599 pci_probe |= PCI_BIG_ROOT_WINDOW;
600 return NULL;
601#endif
597 } else if (!strcmp(str, "earlydump")) { 602 } else if (!strcmp(str, "earlydump")) {
598 pci_early_dump_regs = 1; 603 pci_early_dump_regs = 1;
599 return NULL; 604 return NULL;
diff --git a/arch/x86/pci/fixup.c b/arch/x86/pci/fixup.c
index e663d6bf1328..f6a26e3cb476 100644
--- a/arch/x86/pci/fixup.c
+++ b/arch/x86/pci/fixup.c
@@ -662,10 +662,14 @@ DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x2033, quirk_no_aersid);
662 */ 662 */
663static void pci_amd_enable_64bit_bar(struct pci_dev *dev) 663static void pci_amd_enable_64bit_bar(struct pci_dev *dev)
664{ 664{
665 unsigned i;
666 u32 base, limit, high; 665 u32 base, limit, high;
667 struct resource *res, *conflict;
668 struct pci_dev *other; 666 struct pci_dev *other;
667 struct resource *res;
668 unsigned i;
669 int r;
670
671 if (!(pci_probe & PCI_BIG_ROOT_WINDOW))
672 return;
669 673
670 /* Check that we are the only device of that type */ 674 /* Check that we are the only device of that type */
671 other = pci_get_device(dev->vendor, dev->device, NULL); 675 other = pci_get_device(dev->vendor, dev->device, NULL);
@@ -699,22 +703,25 @@ static void pci_amd_enable_64bit_bar(struct pci_dev *dev)
699 if (!res) 703 if (!res)
700 return; 704 return;
701 705
706 /*
707 * Allocate a 256GB window directly below the 0xfd00000000 hardware
708 * limit (see AMD Family 15h Models 30h-3Fh BKDG, sec 2.4.6).
709 */
702 res->name = "PCI Bus 0000:00"; 710 res->name = "PCI Bus 0000:00";
703 res->flags = IORESOURCE_PREFETCH | IORESOURCE_MEM | 711 res->flags = IORESOURCE_PREFETCH | IORESOURCE_MEM |
704 IORESOURCE_MEM_64 | IORESOURCE_WINDOW; 712 IORESOURCE_MEM_64 | IORESOURCE_WINDOW;
705 res->start = 0x100000000ull; 713 res->start = 0xbd00000000ull;
706 res->end = 0xfd00000000ull - 1; 714 res->end = 0xfd00000000ull - 1;
707 715
708 /* Just grab the free area behind system memory for this */ 716 r = request_resource(&iomem_resource, res);
709 while ((conflict = request_resource_conflict(&iomem_resource, res))) { 717 if (r) {
710 if (conflict->end >= res->end) { 718 kfree(res);
711 kfree(res); 719 return;
712 return;
713 }
714 res->start = conflict->end + 1;
715 } 720 }
716 721
717 dev_info(&dev->dev, "adding root bus resource %pR\n", res); 722 dev_info(&dev->dev, "adding root bus resource %pR (tainting kernel)\n",
723 res);
724 add_taint(TAINT_FIRMWARE_WORKAROUND, LOCKDEP_STILL_OK);
718 725
719 base = ((res->start >> 8) & AMD_141b_MMIO_BASE_MMIOBASE_MASK) | 726 base = ((res->start >> 8) & AMD_141b_MMIO_BASE_MMIOBASE_MASK) |
720 AMD_141b_MMIO_BASE_RE_MASK | AMD_141b_MMIO_BASE_WE_MASK; 727 AMD_141b_MMIO_BASE_RE_MASK | AMD_141b_MMIO_BASE_WE_MASK;
diff --git a/arch/x86/platform/efi/efi_64.c b/arch/x86/platform/efi/efi_64.c
index d87ac96e37ed..2dd15e967c3f 100644
--- a/arch/x86/platform/efi/efi_64.c
+++ b/arch/x86/platform/efi/efi_64.c
@@ -135,7 +135,9 @@ pgd_t * __init efi_call_phys_prolog(void)
135 pud[j] = *pud_offset(p4d_k, vaddr); 135 pud[j] = *pud_offset(p4d_k, vaddr);
136 } 136 }
137 } 137 }
138 pgd_offset_k(pgd * PGDIR_SIZE)->pgd &= ~_PAGE_NX;
138 } 139 }
140
139out: 141out:
140 __flush_tlb_all(); 142 __flush_tlb_all();
141 143
diff --git a/arch/x86/platform/intel-mid/device_libs/platform_bt.c b/arch/x86/platform/intel-mid/device_libs/platform_bt.c
index dc036e511f48..5a0483e7bf66 100644
--- a/arch/x86/platform/intel-mid/device_libs/platform_bt.c
+++ b/arch/x86/platform/intel-mid/device_libs/platform_bt.c
@@ -60,7 +60,7 @@ static int __init tng_bt_sfi_setup(struct bt_sfi_data *ddata)
60 return 0; 60 return 0;
61} 61}
62 62
63static const struct bt_sfi_data tng_bt_sfi_data __initdata = { 63static struct bt_sfi_data tng_bt_sfi_data __initdata = {
64 .setup = tng_bt_sfi_setup, 64 .setup = tng_bt_sfi_setup,
65}; 65};
66 66
diff --git a/arch/x86/xen/mmu_pv.c b/arch/x86/xen/mmu_pv.c
index 4d62c071b166..d85076223a69 100644
--- a/arch/x86/xen/mmu_pv.c
+++ b/arch/x86/xen/mmu_pv.c
@@ -1325,20 +1325,18 @@ static void xen_flush_tlb_others(const struct cpumask *cpus,
1325{ 1325{
1326 struct { 1326 struct {
1327 struct mmuext_op op; 1327 struct mmuext_op op;
1328#ifdef CONFIG_SMP
1329 DECLARE_BITMAP(mask, num_processors);
1330#else
1331 DECLARE_BITMAP(mask, NR_CPUS); 1328 DECLARE_BITMAP(mask, NR_CPUS);
1332#endif
1333 } *args; 1329 } *args;
1334 struct multicall_space mcs; 1330 struct multicall_space mcs;
1331 const size_t mc_entry_size = sizeof(args->op) +
1332 sizeof(args->mask[0]) * BITS_TO_LONGS(num_possible_cpus());
1335 1333
1336 trace_xen_mmu_flush_tlb_others(cpus, info->mm, info->start, info->end); 1334 trace_xen_mmu_flush_tlb_others(cpus, info->mm, info->start, info->end);
1337 1335
1338 if (cpumask_empty(cpus)) 1336 if (cpumask_empty(cpus))
1339 return; /* nothing to do */ 1337 return; /* nothing to do */
1340 1338
1341 mcs = xen_mc_entry(sizeof(*args)); 1339 mcs = xen_mc_entry(mc_entry_size);
1342 args = mcs.args; 1340 args = mcs.args;
1343 args->op.arg2.vcpumask = to_cpumask(args->mask); 1341 args->op.arg2.vcpumask = to_cpumask(args->mask);
1344 1342
diff --git a/arch/x86/xen/xen-ops.h b/arch/x86/xen/xen-ops.h
index 75011b80660f..3b34745d0a52 100644
--- a/arch/x86/xen/xen-ops.h
+++ b/arch/x86/xen/xen-ops.h
@@ -72,7 +72,7 @@ u64 xen_clocksource_read(void);
72void xen_setup_cpu_clockevents(void); 72void xen_setup_cpu_clockevents(void);
73void xen_save_time_memory_area(void); 73void xen_save_time_memory_area(void);
74void xen_restore_time_memory_area(void); 74void xen_restore_time_memory_area(void);
75void __init xen_init_time_ops(void); 75void __ref xen_init_time_ops(void);
76void __init xen_hvm_init_time_ops(void); 76void __init xen_hvm_init_time_ops(void);
77 77
78irqreturn_t xen_debug_interrupt(int irq, void *dev_id); 78irqreturn_t xen_debug_interrupt(int irq, void *dev_id);
diff --git a/block/blk-core.c b/block/blk-core.c
index b8881750a3ac..3ba4326a63b5 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -562,6 +562,13 @@ static void __blk_drain_queue(struct request_queue *q, bool drain_all)
562 } 562 }
563} 563}
564 564
565void blk_drain_queue(struct request_queue *q)
566{
567 spin_lock_irq(q->queue_lock);
568 __blk_drain_queue(q, true);
569 spin_unlock_irq(q->queue_lock);
570}
571
565/** 572/**
566 * blk_queue_bypass_start - enter queue bypass mode 573 * blk_queue_bypass_start - enter queue bypass mode
567 * @q: queue of interest 574 * @q: queue of interest
@@ -689,8 +696,6 @@ void blk_cleanup_queue(struct request_queue *q)
689 */ 696 */
690 blk_freeze_queue(q); 697 blk_freeze_queue(q);
691 spin_lock_irq(lock); 698 spin_lock_irq(lock);
692 if (!q->mq_ops)
693 __blk_drain_queue(q, true);
694 queue_flag_set(QUEUE_FLAG_DEAD, q); 699 queue_flag_set(QUEUE_FLAG_DEAD, q);
695 spin_unlock_irq(lock); 700 spin_unlock_irq(lock);
696 701
diff --git a/block/blk-mq.c b/block/blk-mq.c
index 11097477eeab..3d3797327491 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -161,6 +161,8 @@ void blk_freeze_queue(struct request_queue *q)
161 * exported to drivers as the only user for unfreeze is blk_mq. 161 * exported to drivers as the only user for unfreeze is blk_mq.
162 */ 162 */
163 blk_freeze_queue_start(q); 163 blk_freeze_queue_start(q);
164 if (!q->mq_ops)
165 blk_drain_queue(q);
164 blk_mq_freeze_queue_wait(q); 166 blk_mq_freeze_queue_wait(q);
165} 167}
166 168
diff --git a/block/blk.h b/block/blk.h
index 3f1446937aec..442098aa9463 100644
--- a/block/blk.h
+++ b/block/blk.h
@@ -330,4 +330,6 @@ static inline void blk_queue_bounce(struct request_queue *q, struct bio **bio)
330} 330}
331#endif /* CONFIG_BOUNCE */ 331#endif /* CONFIG_BOUNCE */
332 332
333extern void blk_drain_queue(struct request_queue *q);
334
333#endif /* BLK_INTERNAL_H */ 335#endif /* BLK_INTERNAL_H */
diff --git a/crypto/algapi.c b/crypto/algapi.c
index 60d7366ed343..9a636f961572 100644
--- a/crypto/algapi.c
+++ b/crypto/algapi.c
@@ -167,6 +167,18 @@ void crypto_remove_spawns(struct crypto_alg *alg, struct list_head *list,
167 167
168 spawn->alg = NULL; 168 spawn->alg = NULL;
169 spawns = &inst->alg.cra_users; 169 spawns = &inst->alg.cra_users;
170
171 /*
172 * We may encounter an unregistered instance here, since
173 * an instance's spawns are set up prior to the instance
174 * being registered. An unregistered instance will have
175 * NULL ->cra_users.next, since ->cra_users isn't
176 * properly initialized until registration. But an
177 * unregistered instance cannot have any users, so treat
178 * it the same as ->cra_users being empty.
179 */
180 if (spawns->next == NULL)
181 break;
170 } 182 }
171 } while ((spawns = crypto_more_spawns(alg, &stack, &top, 183 } while ((spawns = crypto_more_spawns(alg, &stack, &top,
172 &secondary_spawns))); 184 &secondary_spawns)));
diff --git a/drivers/acpi/utils.c b/drivers/acpi/utils.c
index 9d49a1acebe3..78db97687f26 100644
--- a/drivers/acpi/utils.c
+++ b/drivers/acpi/utils.c
@@ -737,16 +737,17 @@ bool acpi_dev_found(const char *hid)
737} 737}
738EXPORT_SYMBOL(acpi_dev_found); 738EXPORT_SYMBOL(acpi_dev_found);
739 739
740struct acpi_dev_present_info { 740struct acpi_dev_match_info {
741 const char *dev_name;
741 struct acpi_device_id hid[2]; 742 struct acpi_device_id hid[2];
742 const char *uid; 743 const char *uid;
743 s64 hrv; 744 s64 hrv;
744}; 745};
745 746
746static int acpi_dev_present_cb(struct device *dev, void *data) 747static int acpi_dev_match_cb(struct device *dev, void *data)
747{ 748{
748 struct acpi_device *adev = to_acpi_device(dev); 749 struct acpi_device *adev = to_acpi_device(dev);
749 struct acpi_dev_present_info *match = data; 750 struct acpi_dev_match_info *match = data;
750 unsigned long long hrv; 751 unsigned long long hrv;
751 acpi_status status; 752 acpi_status status;
752 753
@@ -757,6 +758,8 @@ static int acpi_dev_present_cb(struct device *dev, void *data)
757 strcmp(adev->pnp.unique_id, match->uid))) 758 strcmp(adev->pnp.unique_id, match->uid)))
758 return 0; 759 return 0;
759 760
761 match->dev_name = acpi_dev_name(adev);
762
760 if (match->hrv == -1) 763 if (match->hrv == -1)
761 return 1; 764 return 1;
762 765
@@ -789,20 +792,44 @@ static int acpi_dev_present_cb(struct device *dev, void *data)
789 */ 792 */
790bool acpi_dev_present(const char *hid, const char *uid, s64 hrv) 793bool acpi_dev_present(const char *hid, const char *uid, s64 hrv)
791{ 794{
792 struct acpi_dev_present_info match = {}; 795 struct acpi_dev_match_info match = {};
793 struct device *dev; 796 struct device *dev;
794 797
795 strlcpy(match.hid[0].id, hid, sizeof(match.hid[0].id)); 798 strlcpy(match.hid[0].id, hid, sizeof(match.hid[0].id));
796 match.uid = uid; 799 match.uid = uid;
797 match.hrv = hrv; 800 match.hrv = hrv;
798 801
799 dev = bus_find_device(&acpi_bus_type, NULL, &match, 802 dev = bus_find_device(&acpi_bus_type, NULL, &match, acpi_dev_match_cb);
800 acpi_dev_present_cb);
801
802 return !!dev; 803 return !!dev;
803} 804}
804EXPORT_SYMBOL(acpi_dev_present); 805EXPORT_SYMBOL(acpi_dev_present);
805 806
807/**
808 * acpi_dev_get_first_match_name - Return name of first match of ACPI device
809 * @hid: Hardware ID of the device.
810 * @uid: Unique ID of the device, pass NULL to not check _UID
811 * @hrv: Hardware Revision of the device, pass -1 to not check _HRV
812 *
813 * Return device name if a matching device was present
814 * at the moment of invocation, or NULL otherwise.
815 *
816 * See additional information in acpi_dev_present() as well.
817 */
818const char *
819acpi_dev_get_first_match_name(const char *hid, const char *uid, s64 hrv)
820{
821 struct acpi_dev_match_info match = {};
822 struct device *dev;
823
824 strlcpy(match.hid[0].id, hid, sizeof(match.hid[0].id));
825 match.uid = uid;
826 match.hrv = hrv;
827
828 dev = bus_find_device(&acpi_bus_type, NULL, &match, acpi_dev_match_cb);
829 return dev ? match.dev_name : NULL;
830}
831EXPORT_SYMBOL(acpi_dev_get_first_match_name);
832
806/* 833/*
807 * acpi_backlight= handling, this is done here rather then in video_detect.c 834 * acpi_backlight= handling, this is done here rather then in video_detect.c
808 * because __setup cannot be used in modules. 835 * because __setup cannot be used in modules.
diff --git a/drivers/base/Kconfig b/drivers/base/Kconfig
index bdc87907d6a1..2415ad9f6dd4 100644
--- a/drivers/base/Kconfig
+++ b/drivers/base/Kconfig
@@ -236,6 +236,9 @@ config GENERIC_CPU_DEVICES
236config GENERIC_CPU_AUTOPROBE 236config GENERIC_CPU_AUTOPROBE
237 bool 237 bool
238 238
239config GENERIC_CPU_VULNERABILITIES
240 bool
241
239config SOC_BUS 242config SOC_BUS
240 bool 243 bool
241 select GLOB 244 select GLOB
diff --git a/drivers/base/cpu.c b/drivers/base/cpu.c
index 58a9b608d821..d99038487a0d 100644
--- a/drivers/base/cpu.c
+++ b/drivers/base/cpu.c
@@ -511,10 +511,58 @@ static void __init cpu_dev_register_generic(void)
511#endif 511#endif
512} 512}
513 513
514#ifdef CONFIG_GENERIC_CPU_VULNERABILITIES
515
516ssize_t __weak cpu_show_meltdown(struct device *dev,
517 struct device_attribute *attr, char *buf)
518{
519 return sprintf(buf, "Not affected\n");
520}
521
522ssize_t __weak cpu_show_spectre_v1(struct device *dev,
523 struct device_attribute *attr, char *buf)
524{
525 return sprintf(buf, "Not affected\n");
526}
527
528ssize_t __weak cpu_show_spectre_v2(struct device *dev,
529 struct device_attribute *attr, char *buf)
530{
531 return sprintf(buf, "Not affected\n");
532}
533
534static DEVICE_ATTR(meltdown, 0444, cpu_show_meltdown, NULL);
535static DEVICE_ATTR(spectre_v1, 0444, cpu_show_spectre_v1, NULL);
536static DEVICE_ATTR(spectre_v2, 0444, cpu_show_spectre_v2, NULL);
537
538static struct attribute *cpu_root_vulnerabilities_attrs[] = {
539 &dev_attr_meltdown.attr,
540 &dev_attr_spectre_v1.attr,
541 &dev_attr_spectre_v2.attr,
542 NULL
543};
544
545static const struct attribute_group cpu_root_vulnerabilities_group = {
546 .name = "vulnerabilities",
547 .attrs = cpu_root_vulnerabilities_attrs,
548};
549
550static void __init cpu_register_vulnerabilities(void)
551{
552 if (sysfs_create_group(&cpu_subsys.dev_root->kobj,
553 &cpu_root_vulnerabilities_group))
554 pr_err("Unable to register CPU vulnerabilities\n");
555}
556
557#else
558static inline void cpu_register_vulnerabilities(void) { }
559#endif
560
514void __init cpu_dev_init(void) 561void __init cpu_dev_init(void)
515{ 562{
516 if (subsys_system_register(&cpu_subsys, cpu_root_attr_groups)) 563 if (subsys_system_register(&cpu_subsys, cpu_root_attr_groups))
517 panic("Failed to register CPU subsystem"); 564 panic("Failed to register CPU subsystem");
518 565
519 cpu_dev_register_generic(); 566 cpu_dev_register_generic();
567 cpu_register_vulnerabilities();
520} 568}
diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index bc8e61506968..d5fe720cf149 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -1581,9 +1581,8 @@ out:
1581 return err; 1581 return err;
1582} 1582}
1583 1583
1584static void lo_release(struct gendisk *disk, fmode_t mode) 1584static void __lo_release(struct loop_device *lo)
1585{ 1585{
1586 struct loop_device *lo = disk->private_data;
1587 int err; 1586 int err;
1588 1587
1589 if (atomic_dec_return(&lo->lo_refcnt)) 1588 if (atomic_dec_return(&lo->lo_refcnt))
@@ -1610,6 +1609,13 @@ static void lo_release(struct gendisk *disk, fmode_t mode)
1610 mutex_unlock(&lo->lo_ctl_mutex); 1609 mutex_unlock(&lo->lo_ctl_mutex);
1611} 1610}
1612 1611
1612static void lo_release(struct gendisk *disk, fmode_t mode)
1613{
1614 mutex_lock(&loop_index_mutex);
1615 __lo_release(disk->private_data);
1616 mutex_unlock(&loop_index_mutex);
1617}
1618
1613static const struct block_device_operations lo_fops = { 1619static const struct block_device_operations lo_fops = {
1614 .owner = THIS_MODULE, 1620 .owner = THIS_MODULE,
1615 .open = lo_open, 1621 .open = lo_open,
diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c
index 38fc5f397fde..cc93522a6d41 100644
--- a/drivers/block/rbd.c
+++ b/drivers/block/rbd.c
@@ -3047,13 +3047,21 @@ static void format_lock_cookie(struct rbd_device *rbd_dev, char *buf)
3047 mutex_unlock(&rbd_dev->watch_mutex); 3047 mutex_unlock(&rbd_dev->watch_mutex);
3048} 3048}
3049 3049
3050static void __rbd_lock(struct rbd_device *rbd_dev, const char *cookie)
3051{
3052 struct rbd_client_id cid = rbd_get_cid(rbd_dev);
3053
3054 strcpy(rbd_dev->lock_cookie, cookie);
3055 rbd_set_owner_cid(rbd_dev, &cid);
3056 queue_work(rbd_dev->task_wq, &rbd_dev->acquired_lock_work);
3057}
3058
3050/* 3059/*
3051 * lock_rwsem must be held for write 3060 * lock_rwsem must be held for write
3052 */ 3061 */
3053static int rbd_lock(struct rbd_device *rbd_dev) 3062static int rbd_lock(struct rbd_device *rbd_dev)
3054{ 3063{
3055 struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc; 3064 struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
3056 struct rbd_client_id cid = rbd_get_cid(rbd_dev);
3057 char cookie[32]; 3065 char cookie[32];
3058 int ret; 3066 int ret;
3059 3067
@@ -3068,9 +3076,7 @@ static int rbd_lock(struct rbd_device *rbd_dev)
3068 return ret; 3076 return ret;
3069 3077
3070 rbd_dev->lock_state = RBD_LOCK_STATE_LOCKED; 3078 rbd_dev->lock_state = RBD_LOCK_STATE_LOCKED;
3071 strcpy(rbd_dev->lock_cookie, cookie); 3079 __rbd_lock(rbd_dev, cookie);
3072 rbd_set_owner_cid(rbd_dev, &cid);
3073 queue_work(rbd_dev->task_wq, &rbd_dev->acquired_lock_work);
3074 return 0; 3080 return 0;
3075} 3081}
3076 3082
@@ -3856,7 +3862,7 @@ static void rbd_reacquire_lock(struct rbd_device *rbd_dev)
3856 queue_delayed_work(rbd_dev->task_wq, 3862 queue_delayed_work(rbd_dev->task_wq,
3857 &rbd_dev->lock_dwork, 0); 3863 &rbd_dev->lock_dwork, 0);
3858 } else { 3864 } else {
3859 strcpy(rbd_dev->lock_cookie, cookie); 3865 __rbd_lock(rbd_dev, cookie);
3860 } 3866 }
3861} 3867}
3862 3868
@@ -4381,7 +4387,7 @@ static int rbd_init_disk(struct rbd_device *rbd_dev)
4381 segment_size = rbd_obj_bytes(&rbd_dev->header); 4387 segment_size = rbd_obj_bytes(&rbd_dev->header);
4382 blk_queue_max_hw_sectors(q, segment_size / SECTOR_SIZE); 4388 blk_queue_max_hw_sectors(q, segment_size / SECTOR_SIZE);
4383 q->limits.max_sectors = queue_max_hw_sectors(q); 4389 q->limits.max_sectors = queue_max_hw_sectors(q);
4384 blk_queue_max_segments(q, segment_size / SECTOR_SIZE); 4390 blk_queue_max_segments(q, USHRT_MAX);
4385 blk_queue_max_segment_size(q, segment_size); 4391 blk_queue_max_segment_size(q, segment_size);
4386 blk_queue_io_min(q, segment_size); 4392 blk_queue_io_min(q, segment_size);
4387 blk_queue_io_opt(q, segment_size); 4393 blk_queue_io_opt(q, segment_size);
diff --git a/drivers/gpio/gpio-merrifield.c b/drivers/gpio/gpio-merrifield.c
index dd67a31ac337..c38624ea0251 100644
--- a/drivers/gpio/gpio-merrifield.c
+++ b/drivers/gpio/gpio-merrifield.c
@@ -9,6 +9,7 @@
9 * published by the Free Software Foundation. 9 * published by the Free Software Foundation.
10 */ 10 */
11 11
12#include <linux/acpi.h>
12#include <linux/bitops.h> 13#include <linux/bitops.h>
13#include <linux/gpio/driver.h> 14#include <linux/gpio/driver.h>
14#include <linux/init.h> 15#include <linux/init.h>
@@ -380,9 +381,16 @@ static void mrfld_irq_init_hw(struct mrfld_gpio *priv)
380 } 381 }
381} 382}
382 383
384static const char *mrfld_gpio_get_pinctrl_dev_name(void)
385{
386 const char *dev_name = acpi_dev_get_first_match_name("INTC1002", NULL, -1);
387 return dev_name ? dev_name : "pinctrl-merrifield";
388}
389
383static int mrfld_gpio_probe(struct pci_dev *pdev, const struct pci_device_id *id) 390static int mrfld_gpio_probe(struct pci_dev *pdev, const struct pci_device_id *id)
384{ 391{
385 const struct mrfld_gpio_pinrange *range; 392 const struct mrfld_gpio_pinrange *range;
393 const char *pinctrl_dev_name;
386 struct mrfld_gpio *priv; 394 struct mrfld_gpio *priv;
387 u32 gpio_base, irq_base; 395 u32 gpio_base, irq_base;
388 void __iomem *base; 396 void __iomem *base;
@@ -439,10 +447,11 @@ static int mrfld_gpio_probe(struct pci_dev *pdev, const struct pci_device_id *id
439 return retval; 447 return retval;
440 } 448 }
441 449
450 pinctrl_dev_name = mrfld_gpio_get_pinctrl_dev_name();
442 for (i = 0; i < ARRAY_SIZE(mrfld_gpio_ranges); i++) { 451 for (i = 0; i < ARRAY_SIZE(mrfld_gpio_ranges); i++) {
443 range = &mrfld_gpio_ranges[i]; 452 range = &mrfld_gpio_ranges[i];
444 retval = gpiochip_add_pin_range(&priv->chip, 453 retval = gpiochip_add_pin_range(&priv->chip,
445 "pinctrl-merrifield", 454 pinctrl_dev_name,
446 range->gpio_base, 455 range->gpio_base,
447 range->pin_base, 456 range->pin_base,
448 range->npins); 457 range->npins);
diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
index 44332b793718..14532d9576e4 100644
--- a/drivers/gpio/gpiolib.c
+++ b/drivers/gpio/gpiolib.c
@@ -2893,6 +2893,27 @@ void gpiod_set_raw_value(struct gpio_desc *desc, int value)
2893EXPORT_SYMBOL_GPL(gpiod_set_raw_value); 2893EXPORT_SYMBOL_GPL(gpiod_set_raw_value);
2894 2894
2895/** 2895/**
2896 * gpiod_set_value_nocheck() - set a GPIO line value without checking
2897 * @desc: the descriptor to set the value on
2898 * @value: value to set
2899 *
2900 * This sets the value of a GPIO line backing a descriptor, applying
2901 * different semantic quirks like active low and open drain/source
2902 * handling.
2903 */
2904static void gpiod_set_value_nocheck(struct gpio_desc *desc, int value)
2905{
2906 if (test_bit(FLAG_ACTIVE_LOW, &desc->flags))
2907 value = !value;
2908 if (test_bit(FLAG_OPEN_DRAIN, &desc->flags))
2909 gpio_set_open_drain_value_commit(desc, value);
2910 else if (test_bit(FLAG_OPEN_SOURCE, &desc->flags))
2911 gpio_set_open_source_value_commit(desc, value);
2912 else
2913 gpiod_set_raw_value_commit(desc, value);
2914}
2915
2916/**
2896 * gpiod_set_value() - assign a gpio's value 2917 * gpiod_set_value() - assign a gpio's value
2897 * @desc: gpio whose value will be assigned 2918 * @desc: gpio whose value will be assigned
2898 * @value: value to assign 2919 * @value: value to assign
@@ -2906,16 +2927,8 @@ EXPORT_SYMBOL_GPL(gpiod_set_raw_value);
2906void gpiod_set_value(struct gpio_desc *desc, int value) 2927void gpiod_set_value(struct gpio_desc *desc, int value)
2907{ 2928{
2908 VALIDATE_DESC_VOID(desc); 2929 VALIDATE_DESC_VOID(desc);
2909 /* Should be using gpiod_set_value_cansleep() */
2910 WARN_ON(desc->gdev->chip->can_sleep); 2930 WARN_ON(desc->gdev->chip->can_sleep);
2911 if (test_bit(FLAG_ACTIVE_LOW, &desc->flags)) 2931 gpiod_set_value_nocheck(desc, value);
2912 value = !value;
2913 if (test_bit(FLAG_OPEN_DRAIN, &desc->flags))
2914 gpio_set_open_drain_value_commit(desc, value);
2915 else if (test_bit(FLAG_OPEN_SOURCE, &desc->flags))
2916 gpio_set_open_source_value_commit(desc, value);
2917 else
2918 gpiod_set_raw_value_commit(desc, value);
2919} 2932}
2920EXPORT_SYMBOL_GPL(gpiod_set_value); 2933EXPORT_SYMBOL_GPL(gpiod_set_value);
2921 2934
@@ -3243,9 +3256,7 @@ void gpiod_set_value_cansleep(struct gpio_desc *desc, int value)
3243{ 3256{
3244 might_sleep_if(extra_checks); 3257 might_sleep_if(extra_checks);
3245 VALIDATE_DESC_VOID(desc); 3258 VALIDATE_DESC_VOID(desc);
3246 if (test_bit(FLAG_ACTIVE_LOW, &desc->flags)) 3259 gpiod_set_value_nocheck(desc, value);
3247 value = !value;
3248 gpiod_set_raw_value_commit(desc, value);
3249} 3260}
3250EXPORT_SYMBOL_GPL(gpiod_set_value_cansleep); 3261EXPORT_SYMBOL_GPL(gpiod_set_value_cansleep);
3251 3262
diff --git a/drivers/gpu/drm/i915/gvt/cmd_parser.c b/drivers/gpu/drm/i915/gvt/cmd_parser.c
index 85d4c57870fb..49af94627c8a 100644
--- a/drivers/gpu/drm/i915/gvt/cmd_parser.c
+++ b/drivers/gpu/drm/i915/gvt/cmd_parser.c
@@ -2777,12 +2777,12 @@ int intel_gvt_scan_and_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx)
2777} 2777}
2778 2778
2779static struct cmd_info *find_cmd_entry_any_ring(struct intel_gvt *gvt, 2779static struct cmd_info *find_cmd_entry_any_ring(struct intel_gvt *gvt,
2780 unsigned int opcode, int rings) 2780 unsigned int opcode, unsigned long rings)
2781{ 2781{
2782 struct cmd_info *info = NULL; 2782 struct cmd_info *info = NULL;
2783 unsigned int ring; 2783 unsigned int ring;
2784 2784
2785 for_each_set_bit(ring, (unsigned long *)&rings, I915_NUM_ENGINES) { 2785 for_each_set_bit(ring, &rings, I915_NUM_ENGINES) {
2786 info = find_cmd_entry(gvt, opcode, ring); 2786 info = find_cmd_entry(gvt, opcode, ring);
2787 if (info) 2787 if (info)
2788 break; 2788 break;
diff --git a/drivers/gpu/drm/i915/gvt/gtt.c b/drivers/gpu/drm/i915/gvt/gtt.c
index 8e331142badb..64d67ff9bf08 100644
--- a/drivers/gpu/drm/i915/gvt/gtt.c
+++ b/drivers/gpu/drm/i915/gvt/gtt.c
@@ -1359,12 +1359,15 @@ static int ppgtt_handle_guest_write_page_table_bytes(void *gp,
1359 return ret; 1359 return ret;
1360 } else { 1360 } else {
1361 if (!test_bit(index, spt->post_shadow_bitmap)) { 1361 if (!test_bit(index, spt->post_shadow_bitmap)) {
1362 int type = spt->shadow_page.type;
1363
1362 ppgtt_get_shadow_entry(spt, &se, index); 1364 ppgtt_get_shadow_entry(spt, &se, index);
1363 ret = ppgtt_handle_guest_entry_removal(gpt, &se, index); 1365 ret = ppgtt_handle_guest_entry_removal(gpt, &se, index);
1364 if (ret) 1366 if (ret)
1365 return ret; 1367 return ret;
1368 ops->set_pfn(&se, vgpu->gtt.scratch_pt[type].page_mfn);
1369 ppgtt_set_shadow_entry(spt, &se, index);
1366 } 1370 }
1367
1368 ppgtt_set_post_shadow(spt, index); 1371 ppgtt_set_post_shadow(spt, index);
1369 } 1372 }
1370 1373
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 18de6569d04a..5cfba89ed586 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -467,7 +467,7 @@ static void __fence_set_priority(struct dma_fence *fence, int prio)
467 struct drm_i915_gem_request *rq; 467 struct drm_i915_gem_request *rq;
468 struct intel_engine_cs *engine; 468 struct intel_engine_cs *engine;
469 469
470 if (!dma_fence_is_i915(fence)) 470 if (dma_fence_is_signaled(fence) || !dma_fence_is_i915(fence))
471 return; 471 return;
472 472
473 rq = to_request(fence); 473 rq = to_request(fence);
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 333f40bc03bb..7923dfd9963c 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -7027,6 +7027,8 @@ enum {
7027#define GEN9_SLICE_COMMON_ECO_CHICKEN0 _MMIO(0x7308) 7027#define GEN9_SLICE_COMMON_ECO_CHICKEN0 _MMIO(0x7308)
7028#define DISABLE_PIXEL_MASK_CAMMING (1<<14) 7028#define DISABLE_PIXEL_MASK_CAMMING (1<<14)
7029 7029
7030#define GEN9_SLICE_COMMON_ECO_CHICKEN1 _MMIO(0x731c)
7031
7030#define GEN7_L3SQCREG1 _MMIO(0xB010) 7032#define GEN7_L3SQCREG1 _MMIO(0xB010)
7031#define VLV_B0_WA_L3SQCREG1_VALUE 0x00D30000 7033#define VLV_B0_WA_L3SQCREG1_VALUE 0x00D30000
7032 7034
diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c
index ab5bf4e2e28e..6074e04dc99f 100644
--- a/drivers/gpu/drm/i915/intel_engine_cs.c
+++ b/drivers/gpu/drm/i915/intel_engine_cs.c
@@ -1390,6 +1390,11 @@ static int glk_init_workarounds(struct intel_engine_cs *engine)
1390 if (ret) 1390 if (ret)
1391 return ret; 1391 return ret;
1392 1392
1393 /* WA #0862: Userspace has to set "Barrier Mode" to avoid hangs. */
1394 ret = wa_ring_whitelist_reg(engine, GEN9_SLICE_COMMON_ECO_CHICKEN1);
1395 if (ret)
1396 return ret;
1397
1393 /* WaToEnableHwFixForPushConstHWBug:glk */ 1398 /* WaToEnableHwFixForPushConstHWBug:glk */
1394 WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, 1399 WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2,
1395 GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION); 1400 GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
index d36e25607435..e71a8cd50498 100644
--- a/drivers/gpu/drm/i915/intel_lrc.c
+++ b/drivers/gpu/drm/i915/intel_lrc.c
@@ -974,6 +974,9 @@ static void execlists_schedule(struct drm_i915_gem_request *request, int prio)
974 974
975 GEM_BUG_ON(prio == I915_PRIORITY_INVALID); 975 GEM_BUG_ON(prio == I915_PRIORITY_INVALID);
976 976
977 if (i915_gem_request_completed(request))
978 return;
979
977 if (prio <= READ_ONCE(request->priotree.priority)) 980 if (prio <= READ_ONCE(request->priotree.priority))
978 return; 981 return;
979 982
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgf119.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgf119.c
index a2978a37b4f3..700fc754f28a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgf119.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgf119.c
@@ -174,6 +174,7 @@ gf119_sor = {
174 .links = gf119_sor_dp_links, 174 .links = gf119_sor_dp_links,
175 .power = g94_sor_dp_power, 175 .power = g94_sor_dp_power,
176 .pattern = gf119_sor_dp_pattern, 176 .pattern = gf119_sor_dp_pattern,
177 .drive = gf119_sor_dp_drive,
177 .vcpi = gf119_sor_dp_vcpi, 178 .vcpi = gf119_sor_dp_vcpi,
178 .audio = gf119_sor_dp_audio, 179 .audio = gf119_sor_dp_audio,
179 .audio_sym = gf119_sor_dp_audio_sym, 180 .audio_sym = gf119_sor_dp_audio_sym,
diff --git a/drivers/gpu/drm/tegra/sor.c b/drivers/gpu/drm/tegra/sor.c
index b0a1dedac802..476079f1255f 100644
--- a/drivers/gpu/drm/tegra/sor.c
+++ b/drivers/gpu/drm/tegra/sor.c
@@ -2656,6 +2656,9 @@ static int tegra_sor_probe(struct platform_device *pdev)
2656 name, err); 2656 name, err);
2657 goto remove; 2657 goto remove;
2658 } 2658 }
2659 } else {
2660 /* fall back to the module clock on SOR0 (eDP/LVDS only) */
2661 sor->clk_out = sor->clk;
2659 } 2662 }
2660 2663
2661 sor->clk_parent = devm_clk_get(&pdev->dev, "parent"); 2664 sor->clk_parent = devm_clk_get(&pdev->dev, "parent");
diff --git a/drivers/gpu/drm/vc4/vc4_irq.c b/drivers/gpu/drm/vc4/vc4_irq.c
index 26eddbb62893..3dd62d75f531 100644
--- a/drivers/gpu/drm/vc4/vc4_irq.c
+++ b/drivers/gpu/drm/vc4/vc4_irq.c
@@ -209,9 +209,6 @@ vc4_irq_postinstall(struct drm_device *dev)
209{ 209{
210 struct vc4_dev *vc4 = to_vc4_dev(dev); 210 struct vc4_dev *vc4 = to_vc4_dev(dev);
211 211
212 /* Undo the effects of a previous vc4_irq_uninstall. */
213 enable_irq(dev->irq);
214
215 /* Enable both the render done and out of memory interrupts. */ 212 /* Enable both the render done and out of memory interrupts. */
216 V3D_WRITE(V3D_INTENA, V3D_DRIVER_IRQS); 213 V3D_WRITE(V3D_INTENA, V3D_DRIVER_IRQS);
217 214
diff --git a/drivers/gpu/drm/vc4/vc4_v3d.c b/drivers/gpu/drm/vc4/vc4_v3d.c
index 622cd43840b8..493f392b3a0a 100644
--- a/drivers/gpu/drm/vc4/vc4_v3d.c
+++ b/drivers/gpu/drm/vc4/vc4_v3d.c
@@ -327,6 +327,9 @@ static int vc4_v3d_runtime_resume(struct device *dev)
327 return ret; 327 return ret;
328 328
329 vc4_v3d_init_hw(vc4->dev); 329 vc4_v3d_init_hw(vc4->dev);
330
331 /* We disabled the IRQ as part of vc4_irq_uninstall in suspend. */
332 enable_irq(vc4->dev->irq);
330 vc4_irq_postinstall(vc4->dev); 333 vc4_irq_postinstall(vc4->dev);
331 334
332 return 0; 335 return 0;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
index 21c62a34e558..87e8af5776a3 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
@@ -2731,6 +2731,8 @@ static int vmw_cmd_dx_view_define(struct vmw_private *dev_priv,
2731 } 2731 }
2732 2732
2733 view_type = vmw_view_cmd_to_type(header->id); 2733 view_type = vmw_view_cmd_to_type(header->id);
2734 if (view_type == vmw_view_max)
2735 return -EINVAL;
2734 cmd = container_of(header, typeof(*cmd), header); 2736 cmd = container_of(header, typeof(*cmd), header);
2735 ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_surface, 2737 ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_surface,
2736 user_surface_converter, 2738 user_surface_converter,
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
index 0545740b3724..641294aef165 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
@@ -697,7 +697,6 @@ vmw_du_plane_duplicate_state(struct drm_plane *plane)
697 vps->pinned = 0; 697 vps->pinned = 0;
698 698
699 /* Mapping is managed by prepare_fb/cleanup_fb */ 699 /* Mapping is managed by prepare_fb/cleanup_fb */
700 memset(&vps->guest_map, 0, sizeof(vps->guest_map));
701 memset(&vps->host_map, 0, sizeof(vps->host_map)); 700 memset(&vps->host_map, 0, sizeof(vps->host_map));
702 vps->cpp = 0; 701 vps->cpp = 0;
703 702
@@ -760,11 +759,6 @@ vmw_du_plane_destroy_state(struct drm_plane *plane,
760 759
761 760
762 /* Should have been freed by cleanup_fb */ 761 /* Should have been freed by cleanup_fb */
763 if (vps->guest_map.virtual) {
764 DRM_ERROR("Guest mapping not freed\n");
765 ttm_bo_kunmap(&vps->guest_map);
766 }
767
768 if (vps->host_map.virtual) { 762 if (vps->host_map.virtual) {
769 DRM_ERROR("Host mapping not freed\n"); 763 DRM_ERROR("Host mapping not freed\n");
770 ttm_bo_kunmap(&vps->host_map); 764 ttm_bo_kunmap(&vps->host_map);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h
index ff9c8389ff21..cd9da2dd79af 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h
@@ -175,7 +175,7 @@ struct vmw_plane_state {
175 int pinned; 175 int pinned;
176 176
177 /* For CPU Blit */ 177 /* For CPU Blit */
178 struct ttm_bo_kmap_obj host_map, guest_map; 178 struct ttm_bo_kmap_obj host_map;
179 unsigned int cpp; 179 unsigned int cpp;
180}; 180};
181 181
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c b/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c
index 90b5437fd787..b68d74888ab1 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c
@@ -114,7 +114,7 @@ struct vmw_screen_target_display_unit {
114 bool defined; 114 bool defined;
115 115
116 /* For CPU Blit */ 116 /* For CPU Blit */
117 struct ttm_bo_kmap_obj host_map, guest_map; 117 struct ttm_bo_kmap_obj host_map;
118 unsigned int cpp; 118 unsigned int cpp;
119}; 119};
120 120
@@ -695,7 +695,8 @@ static void vmw_stdu_dmabuf_cpu_commit(struct vmw_kms_dirty *dirty)
695 s32 src_pitch, dst_pitch; 695 s32 src_pitch, dst_pitch;
696 u8 *src, *dst; 696 u8 *src, *dst;
697 bool not_used; 697 bool not_used;
698 698 struct ttm_bo_kmap_obj guest_map;
699 int ret;
699 700
700 if (!dirty->num_hits) 701 if (!dirty->num_hits)
701 return; 702 return;
@@ -706,6 +707,13 @@ static void vmw_stdu_dmabuf_cpu_commit(struct vmw_kms_dirty *dirty)
706 if (width == 0 || height == 0) 707 if (width == 0 || height == 0)
707 return; 708 return;
708 709
710 ret = ttm_bo_kmap(&ddirty->buf->base, 0, ddirty->buf->base.num_pages,
711 &guest_map);
712 if (ret) {
713 DRM_ERROR("Failed mapping framebuffer for blit: %d\n",
714 ret);
715 goto out_cleanup;
716 }
709 717
710 /* Assume we are blitting from Host (display_srf) to Guest (dmabuf) */ 718 /* Assume we are blitting from Host (display_srf) to Guest (dmabuf) */
711 src_pitch = stdu->display_srf->base_size.width * stdu->cpp; 719 src_pitch = stdu->display_srf->base_size.width * stdu->cpp;
@@ -713,7 +721,7 @@ static void vmw_stdu_dmabuf_cpu_commit(struct vmw_kms_dirty *dirty)
713 src += ddirty->top * src_pitch + ddirty->left * stdu->cpp; 721 src += ddirty->top * src_pitch + ddirty->left * stdu->cpp;
714 722
715 dst_pitch = ddirty->pitch; 723 dst_pitch = ddirty->pitch;
716 dst = ttm_kmap_obj_virtual(&stdu->guest_map, &not_used); 724 dst = ttm_kmap_obj_virtual(&guest_map, &not_used);
717 dst += ddirty->fb_top * dst_pitch + ddirty->fb_left * stdu->cpp; 725 dst += ddirty->fb_top * dst_pitch + ddirty->fb_left * stdu->cpp;
718 726
719 727
@@ -772,6 +780,7 @@ static void vmw_stdu_dmabuf_cpu_commit(struct vmw_kms_dirty *dirty)
772 vmw_fifo_commit(dev_priv, sizeof(*cmd)); 780 vmw_fifo_commit(dev_priv, sizeof(*cmd));
773 } 781 }
774 782
783 ttm_bo_kunmap(&guest_map);
775out_cleanup: 784out_cleanup:
776 ddirty->left = ddirty->top = ddirty->fb_left = ddirty->fb_top = S32_MAX; 785 ddirty->left = ddirty->top = ddirty->fb_left = ddirty->fb_top = S32_MAX;
777 ddirty->right = ddirty->bottom = S32_MIN; 786 ddirty->right = ddirty->bottom = S32_MIN;
@@ -1109,9 +1118,6 @@ vmw_stdu_primary_plane_cleanup_fb(struct drm_plane *plane,
1109{ 1118{
1110 struct vmw_plane_state *vps = vmw_plane_state_to_vps(old_state); 1119 struct vmw_plane_state *vps = vmw_plane_state_to_vps(old_state);
1111 1120
1112 if (vps->guest_map.virtual)
1113 ttm_bo_kunmap(&vps->guest_map);
1114
1115 if (vps->host_map.virtual) 1121 if (vps->host_map.virtual)
1116 ttm_bo_kunmap(&vps->host_map); 1122 ttm_bo_kunmap(&vps->host_map);
1117 1123
@@ -1277,33 +1283,11 @@ vmw_stdu_primary_plane_prepare_fb(struct drm_plane *plane,
1277 */ 1283 */
1278 if (vps->content_fb_type == SEPARATE_DMA && 1284 if (vps->content_fb_type == SEPARATE_DMA &&
1279 !(dev_priv->capabilities & SVGA_CAP_3D)) { 1285 !(dev_priv->capabilities & SVGA_CAP_3D)) {
1280
1281 struct vmw_framebuffer_dmabuf *new_vfbd;
1282
1283 new_vfbd = vmw_framebuffer_to_vfbd(new_fb);
1284
1285 ret = ttm_bo_reserve(&new_vfbd->buffer->base, false, false,
1286 NULL);
1287 if (ret)
1288 goto out_srf_unpin;
1289
1290 ret = ttm_bo_kmap(&new_vfbd->buffer->base, 0,
1291 new_vfbd->buffer->base.num_pages,
1292 &vps->guest_map);
1293
1294 ttm_bo_unreserve(&new_vfbd->buffer->base);
1295
1296 if (ret) {
1297 DRM_ERROR("Failed to map content buffer to CPU\n");
1298 goto out_srf_unpin;
1299 }
1300
1301 ret = ttm_bo_kmap(&vps->surf->res.backup->base, 0, 1286 ret = ttm_bo_kmap(&vps->surf->res.backup->base, 0,
1302 vps->surf->res.backup->base.num_pages, 1287 vps->surf->res.backup->base.num_pages,
1303 &vps->host_map); 1288 &vps->host_map);
1304 if (ret) { 1289 if (ret) {
1305 DRM_ERROR("Failed to map display buffer to CPU\n"); 1290 DRM_ERROR("Failed to map display buffer to CPU\n");
1306 ttm_bo_kunmap(&vps->guest_map);
1307 goto out_srf_unpin; 1291 goto out_srf_unpin;
1308 } 1292 }
1309 1293
@@ -1350,7 +1334,6 @@ vmw_stdu_primary_plane_atomic_update(struct drm_plane *plane,
1350 stdu->display_srf = vps->surf; 1334 stdu->display_srf = vps->surf;
1351 stdu->content_fb_type = vps->content_fb_type; 1335 stdu->content_fb_type = vps->content_fb_type;
1352 stdu->cpp = vps->cpp; 1336 stdu->cpp = vps->cpp;
1353 memcpy(&stdu->guest_map, &vps->guest_map, sizeof(vps->guest_map));
1354 memcpy(&stdu->host_map, &vps->host_map, sizeof(vps->host_map)); 1337 memcpy(&stdu->host_map, &vps->host_map, sizeof(vps->host_map));
1355 1338
1356 if (!stdu->defined) 1339 if (!stdu->defined)
diff --git a/drivers/iio/adc/stm32-dfsdm-adc.c b/drivers/iio/adc/stm32-dfsdm-adc.c
index 5e871404f565..daa026d6a94f 100644
--- a/drivers/iio/adc/stm32-dfsdm-adc.c
+++ b/drivers/iio/adc/stm32-dfsdm-adc.c
@@ -1087,18 +1087,11 @@ static int stm32_dfsdm_adc_probe(struct platform_device *pdev)
1087 struct device_node *np = dev->of_node; 1087 struct device_node *np = dev->of_node;
1088 const struct stm32_dfsdm_dev_data *dev_data; 1088 const struct stm32_dfsdm_dev_data *dev_data;
1089 struct iio_dev *iio; 1089 struct iio_dev *iio;
1090 const struct of_device_id *of_id;
1091 char *name; 1090 char *name;
1092 int ret, irq, val; 1091 int ret, irq, val;
1093 1092
1094 of_id = of_match_node(stm32_dfsdm_adc_match, np);
1095 if (!of_id->data) {
1096 dev_err(&pdev->dev, "Data associated to device is missing\n");
1097 return -EINVAL;
1098 }
1099
1100 dev_data = (const struct stm32_dfsdm_dev_data *)of_id->data;
1101 1093
1094 dev_data = of_device_get_match_data(dev);
1102 iio = devm_iio_device_alloc(dev, sizeof(*adc)); 1095 iio = devm_iio_device_alloc(dev, sizeof(*adc));
1103 if (!iio) { 1096 if (!iio) {
1104 dev_err(dev, "%s: Failed to allocate IIO\n", __func__); 1097 dev_err(dev, "%s: Failed to allocate IIO\n", __func__);
@@ -1106,10 +1099,6 @@ static int stm32_dfsdm_adc_probe(struct platform_device *pdev)
1106 } 1099 }
1107 1100
1108 adc = iio_priv(iio); 1101 adc = iio_priv(iio);
1109 if (IS_ERR(adc)) {
1110 dev_err(dev, "%s: Failed to allocate ADC\n", __func__);
1111 return PTR_ERR(adc);
1112 }
1113 adc->dfsdm = dev_get_drvdata(dev->parent); 1102 adc->dfsdm = dev_get_drvdata(dev->parent);
1114 1103
1115 iio->dev.parent = dev; 1104 iio->dev.parent = dev;
diff --git a/drivers/iio/adc/stm32-dfsdm-core.c b/drivers/iio/adc/stm32-dfsdm-core.c
index 6cd655f8239b..6290332cfd3f 100644
--- a/drivers/iio/adc/stm32-dfsdm-core.c
+++ b/drivers/iio/adc/stm32-dfsdm-core.c
@@ -242,7 +242,6 @@ MODULE_DEVICE_TABLE(of, stm32_dfsdm_of_match);
242static int stm32_dfsdm_probe(struct platform_device *pdev) 242static int stm32_dfsdm_probe(struct platform_device *pdev)
243{ 243{
244 struct dfsdm_priv *priv; 244 struct dfsdm_priv *priv;
245 const struct of_device_id *of_id;
246 const struct stm32_dfsdm_dev_data *dev_data; 245 const struct stm32_dfsdm_dev_data *dev_data;
247 struct stm32_dfsdm *dfsdm; 246 struct stm32_dfsdm *dfsdm;
248 int ret; 247 int ret;
@@ -253,13 +252,8 @@ static int stm32_dfsdm_probe(struct platform_device *pdev)
253 252
254 priv->pdev = pdev; 253 priv->pdev = pdev;
255 254
256 of_id = of_match_node(stm32_dfsdm_of_match, pdev->dev.of_node); 255 dev_data = of_device_get_match_data(&pdev->dev);
257 if (!of_id->data) {
258 dev_err(&pdev->dev, "Data associated to device is missing\n");
259 return -EINVAL;
260 }
261 256
262 dev_data = (const struct stm32_dfsdm_dev_data *)of_id->data;
263 dfsdm = &priv->dfsdm; 257 dfsdm = &priv->dfsdm;
264 dfsdm->fl_list = devm_kcalloc(&pdev->dev, dev_data->num_filters, 258 dfsdm->fl_list = devm_kcalloc(&pdev->dev, dev_data->num_filters,
265 sizeof(*dfsdm->fl_list), GFP_KERNEL); 259 sizeof(*dfsdm->fl_list), GFP_KERNEL);
diff --git a/drivers/infiniband/core/core_priv.h b/drivers/infiniband/core/core_priv.h
index a1d687a664f8..66f0268f37a6 100644
--- a/drivers/infiniband/core/core_priv.h
+++ b/drivers/infiniband/core/core_priv.h
@@ -314,7 +314,7 @@ static inline int ib_mad_enforce_security(struct ib_mad_agent_private *map,
314} 314}
315#endif 315#endif
316 316
317struct ib_device *__ib_device_get_by_index(u32 ifindex); 317struct ib_device *ib_device_get_by_index(u32 ifindex);
318/* RDMA device netlink */ 318/* RDMA device netlink */
319void nldev_init(void); 319void nldev_init(void);
320void nldev_exit(void); 320void nldev_exit(void);
diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c
index 30914f3baa5f..465520627e4b 100644
--- a/drivers/infiniband/core/device.c
+++ b/drivers/infiniband/core/device.c
@@ -134,7 +134,7 @@ static int ib_device_check_mandatory(struct ib_device *device)
134 return 0; 134 return 0;
135} 135}
136 136
137struct ib_device *__ib_device_get_by_index(u32 index) 137static struct ib_device *__ib_device_get_by_index(u32 index)
138{ 138{
139 struct ib_device *device; 139 struct ib_device *device;
140 140
@@ -145,6 +145,22 @@ struct ib_device *__ib_device_get_by_index(u32 index)
145 return NULL; 145 return NULL;
146} 146}
147 147
148/*
149 * Caller is responsible to return refrerence count by calling put_device()
150 */
151struct ib_device *ib_device_get_by_index(u32 index)
152{
153 struct ib_device *device;
154
155 down_read(&lists_rwsem);
156 device = __ib_device_get_by_index(index);
157 if (device)
158 get_device(&device->dev);
159
160 up_read(&lists_rwsem);
161 return device;
162}
163
148static struct ib_device *__ib_device_get_by_name(const char *name) 164static struct ib_device *__ib_device_get_by_name(const char *name)
149{ 165{
150 struct ib_device *device; 166 struct ib_device *device;
diff --git a/drivers/infiniband/core/nldev.c b/drivers/infiniband/core/nldev.c
index 9a05245a1acf..0dcd1aa6f683 100644
--- a/drivers/infiniband/core/nldev.c
+++ b/drivers/infiniband/core/nldev.c
@@ -142,27 +142,34 @@ static int nldev_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh,
142 142
143 index = nla_get_u32(tb[RDMA_NLDEV_ATTR_DEV_INDEX]); 143 index = nla_get_u32(tb[RDMA_NLDEV_ATTR_DEV_INDEX]);
144 144
145 device = __ib_device_get_by_index(index); 145 device = ib_device_get_by_index(index);
146 if (!device) 146 if (!device)
147 return -EINVAL; 147 return -EINVAL;
148 148
149 msg = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL); 149 msg = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL);
150 if (!msg) 150 if (!msg) {
151 return -ENOMEM; 151 err = -ENOMEM;
152 goto err;
153 }
152 154
153 nlh = nlmsg_put(msg, NETLINK_CB(skb).portid, nlh->nlmsg_seq, 155 nlh = nlmsg_put(msg, NETLINK_CB(skb).portid, nlh->nlmsg_seq,
154 RDMA_NL_GET_TYPE(RDMA_NL_NLDEV, RDMA_NLDEV_CMD_GET), 156 RDMA_NL_GET_TYPE(RDMA_NL_NLDEV, RDMA_NLDEV_CMD_GET),
155 0, 0); 157 0, 0);
156 158
157 err = fill_dev_info(msg, device); 159 err = fill_dev_info(msg, device);
158 if (err) { 160 if (err)
159 nlmsg_free(msg); 161 goto err_free;
160 return err;
161 }
162 162
163 nlmsg_end(msg, nlh); 163 nlmsg_end(msg, nlh);
164 164
165 put_device(&device->dev);
165 return rdma_nl_unicast(msg, NETLINK_CB(skb).portid); 166 return rdma_nl_unicast(msg, NETLINK_CB(skb).portid);
167
168err_free:
169 nlmsg_free(msg);
170err:
171 put_device(&device->dev);
172 return err;
166} 173}
167 174
168static int _nldev_get_dumpit(struct ib_device *device, 175static int _nldev_get_dumpit(struct ib_device *device,
@@ -220,31 +227,40 @@ static int nldev_port_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh,
220 return -EINVAL; 227 return -EINVAL;
221 228
222 index = nla_get_u32(tb[RDMA_NLDEV_ATTR_DEV_INDEX]); 229 index = nla_get_u32(tb[RDMA_NLDEV_ATTR_DEV_INDEX]);
223 device = __ib_device_get_by_index(index); 230 device = ib_device_get_by_index(index);
224 if (!device) 231 if (!device)
225 return -EINVAL; 232 return -EINVAL;
226 233
227 port = nla_get_u32(tb[RDMA_NLDEV_ATTR_PORT_INDEX]); 234 port = nla_get_u32(tb[RDMA_NLDEV_ATTR_PORT_INDEX]);
228 if (!rdma_is_port_valid(device, port)) 235 if (!rdma_is_port_valid(device, port)) {
229 return -EINVAL; 236 err = -EINVAL;
237 goto err;
238 }
230 239
231 msg = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL); 240 msg = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL);
232 if (!msg) 241 if (!msg) {
233 return -ENOMEM; 242 err = -ENOMEM;
243 goto err;
244 }
234 245
235 nlh = nlmsg_put(msg, NETLINK_CB(skb).portid, nlh->nlmsg_seq, 246 nlh = nlmsg_put(msg, NETLINK_CB(skb).portid, nlh->nlmsg_seq,
236 RDMA_NL_GET_TYPE(RDMA_NL_NLDEV, RDMA_NLDEV_CMD_GET), 247 RDMA_NL_GET_TYPE(RDMA_NL_NLDEV, RDMA_NLDEV_CMD_GET),
237 0, 0); 248 0, 0);
238 249
239 err = fill_port_info(msg, device, port); 250 err = fill_port_info(msg, device, port);
240 if (err) { 251 if (err)
241 nlmsg_free(msg); 252 goto err_free;
242 return err;
243 }
244 253
245 nlmsg_end(msg, nlh); 254 nlmsg_end(msg, nlh);
255 put_device(&device->dev);
246 256
247 return rdma_nl_unicast(msg, NETLINK_CB(skb).portid); 257 return rdma_nl_unicast(msg, NETLINK_CB(skb).portid);
258
259err_free:
260 nlmsg_free(msg);
261err:
262 put_device(&device->dev);
263 return err;
248} 264}
249 265
250static int nldev_port_get_dumpit(struct sk_buff *skb, 266static int nldev_port_get_dumpit(struct sk_buff *skb,
@@ -265,7 +281,7 @@ static int nldev_port_get_dumpit(struct sk_buff *skb,
265 return -EINVAL; 281 return -EINVAL;
266 282
267 ifindex = nla_get_u32(tb[RDMA_NLDEV_ATTR_DEV_INDEX]); 283 ifindex = nla_get_u32(tb[RDMA_NLDEV_ATTR_DEV_INDEX]);
268 device = __ib_device_get_by_index(ifindex); 284 device = ib_device_get_by_index(ifindex);
269 if (!device) 285 if (!device)
270 return -EINVAL; 286 return -EINVAL;
271 287
@@ -299,7 +315,9 @@ static int nldev_port_get_dumpit(struct sk_buff *skb,
299 nlmsg_end(skb, nlh); 315 nlmsg_end(skb, nlh);
300 } 316 }
301 317
302out: cb->args[0] = idx; 318out:
319 put_device(&device->dev);
320 cb->args[0] = idx;
303 return skb->len; 321 return skb->len;
304} 322}
305 323
diff --git a/drivers/infiniband/hw/mlx4/mr.c b/drivers/infiniband/hw/mlx4/mr.c
index 313bfb9ccb71..4975f3e6596e 100644
--- a/drivers/infiniband/hw/mlx4/mr.c
+++ b/drivers/infiniband/hw/mlx4/mr.c
@@ -642,7 +642,6 @@ struct ib_mr *mlx4_ib_alloc_mr(struct ib_pd *pd,
642 goto err_free_mr; 642 goto err_free_mr;
643 643
644 mr->max_pages = max_num_sg; 644 mr->max_pages = max_num_sg;
645
646 err = mlx4_mr_enable(dev->dev, &mr->mmr); 645 err = mlx4_mr_enable(dev->dev, &mr->mmr);
647 if (err) 646 if (err)
648 goto err_free_pl; 647 goto err_free_pl;
@@ -653,6 +652,7 @@ struct ib_mr *mlx4_ib_alloc_mr(struct ib_pd *pd,
653 return &mr->ibmr; 652 return &mr->ibmr;
654 653
655err_free_pl: 654err_free_pl:
655 mr->ibmr.device = pd->device;
656 mlx4_free_priv_pages(mr); 656 mlx4_free_priv_pages(mr);
657err_free_mr: 657err_free_mr:
658 (void) mlx4_mr_free(dev->dev, &mr->mmr); 658 (void) mlx4_mr_free(dev->dev, &mr->mmr);
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c
index 12b7f911f0e5..8880351df179 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_main.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c
@@ -902,8 +902,8 @@ static int path_rec_start(struct net_device *dev,
902 return 0; 902 return 0;
903} 903}
904 904
905static void neigh_add_path(struct sk_buff *skb, u8 *daddr, 905static struct ipoib_neigh *neigh_add_path(struct sk_buff *skb, u8 *daddr,
906 struct net_device *dev) 906 struct net_device *dev)
907{ 907{
908 struct ipoib_dev_priv *priv = ipoib_priv(dev); 908 struct ipoib_dev_priv *priv = ipoib_priv(dev);
909 struct rdma_netdev *rn = netdev_priv(dev); 909 struct rdma_netdev *rn = netdev_priv(dev);
@@ -917,7 +917,15 @@ static void neigh_add_path(struct sk_buff *skb, u8 *daddr,
917 spin_unlock_irqrestore(&priv->lock, flags); 917 spin_unlock_irqrestore(&priv->lock, flags);
918 ++dev->stats.tx_dropped; 918 ++dev->stats.tx_dropped;
919 dev_kfree_skb_any(skb); 919 dev_kfree_skb_any(skb);
920 return; 920 return NULL;
921 }
922
923 /* To avoid race condition, make sure that the
924 * neigh will be added only once.
925 */
926 if (unlikely(!list_empty(&neigh->list))) {
927 spin_unlock_irqrestore(&priv->lock, flags);
928 return neigh;
921 } 929 }
922 930
923 path = __path_find(dev, daddr + 4); 931 path = __path_find(dev, daddr + 4);
@@ -956,7 +964,7 @@ static void neigh_add_path(struct sk_buff *skb, u8 *daddr,
956 path->ah->last_send = rn->send(dev, skb, path->ah->ah, 964 path->ah->last_send = rn->send(dev, skb, path->ah->ah,
957 IPOIB_QPN(daddr)); 965 IPOIB_QPN(daddr));
958 ipoib_neigh_put(neigh); 966 ipoib_neigh_put(neigh);
959 return; 967 return NULL;
960 } 968 }
961 } else { 969 } else {
962 neigh->ah = NULL; 970 neigh->ah = NULL;
@@ -973,7 +981,7 @@ static void neigh_add_path(struct sk_buff *skb, u8 *daddr,
973 981
974 spin_unlock_irqrestore(&priv->lock, flags); 982 spin_unlock_irqrestore(&priv->lock, flags);
975 ipoib_neigh_put(neigh); 983 ipoib_neigh_put(neigh);
976 return; 984 return NULL;
977 985
978err_path: 986err_path:
979 ipoib_neigh_free(neigh); 987 ipoib_neigh_free(neigh);
@@ -983,6 +991,8 @@ err_drop:
983 991
984 spin_unlock_irqrestore(&priv->lock, flags); 992 spin_unlock_irqrestore(&priv->lock, flags);
985 ipoib_neigh_put(neigh); 993 ipoib_neigh_put(neigh);
994
995 return NULL;
986} 996}
987 997
988static void unicast_arp_send(struct sk_buff *skb, struct net_device *dev, 998static void unicast_arp_send(struct sk_buff *skb, struct net_device *dev,
@@ -1091,8 +1101,9 @@ static int ipoib_start_xmit(struct sk_buff *skb, struct net_device *dev)
1091 case htons(ETH_P_TIPC): 1101 case htons(ETH_P_TIPC):
1092 neigh = ipoib_neigh_get(dev, phdr->hwaddr); 1102 neigh = ipoib_neigh_get(dev, phdr->hwaddr);
1093 if (unlikely(!neigh)) { 1103 if (unlikely(!neigh)) {
1094 neigh_add_path(skb, phdr->hwaddr, dev); 1104 neigh = neigh_add_path(skb, phdr->hwaddr, dev);
1095 return NETDEV_TX_OK; 1105 if (likely(!neigh))
1106 return NETDEV_TX_OK;
1096 } 1107 }
1097 break; 1108 break;
1098 case htons(ETH_P_ARP): 1109 case htons(ETH_P_ARP):
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
index 93e149efc1f5..9b3f47ae2016 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
@@ -816,7 +816,10 @@ void ipoib_mcast_send(struct net_device *dev, u8 *daddr, struct sk_buff *skb)
816 spin_lock_irqsave(&priv->lock, flags); 816 spin_lock_irqsave(&priv->lock, flags);
817 if (!neigh) { 817 if (!neigh) {
818 neigh = ipoib_neigh_alloc(daddr, dev); 818 neigh = ipoib_neigh_alloc(daddr, dev);
819 if (neigh) { 819 /* Make sure that the neigh will be added only
820 * once to mcast list.
821 */
822 if (neigh && list_empty(&neigh->list)) {
820 kref_get(&mcast->ah->ref); 823 kref_get(&mcast->ah->ref);
821 neigh->ah = mcast->ah; 824 neigh->ah = mcast->ah;
822 list_add_tail(&neigh->list, &mcast->neigh_list); 825 list_add_tail(&neigh->list, &mcast->neigh_list);
diff --git a/drivers/infiniband/ulp/srpt/ib_srpt.c b/drivers/infiniband/ulp/srpt/ib_srpt.c
index 8a1bd354b1cc..bfa576aa9f03 100644
--- a/drivers/infiniband/ulp/srpt/ib_srpt.c
+++ b/drivers/infiniband/ulp/srpt/ib_srpt.c
@@ -1013,8 +1013,7 @@ static int srpt_init_ch_qp(struct srpt_rdma_ch *ch, struct ib_qp *qp)
1013 return -ENOMEM; 1013 return -ENOMEM;
1014 1014
1015 attr->qp_state = IB_QPS_INIT; 1015 attr->qp_state = IB_QPS_INIT;
1016 attr->qp_access_flags = IB_ACCESS_LOCAL_WRITE | IB_ACCESS_REMOTE_READ | 1016 attr->qp_access_flags = IB_ACCESS_LOCAL_WRITE;
1017 IB_ACCESS_REMOTE_WRITE;
1018 attr->port_num = ch->sport->port; 1017 attr->port_num = ch->sport->port;
1019 attr->pkey_index = 0; 1018 attr->pkey_index = 0;
1020 1019
@@ -2078,7 +2077,7 @@ static int srpt_cm_req_recv(struct ib_cm_id *cm_id,
2078 goto destroy_ib; 2077 goto destroy_ib;
2079 } 2078 }
2080 2079
2081 guid = (__be16 *)&param->primary_path->sgid.global.interface_id; 2080 guid = (__be16 *)&param->primary_path->dgid.global.interface_id;
2082 snprintf(ch->ini_guid, sizeof(ch->ini_guid), "%04x:%04x:%04x:%04x", 2081 snprintf(ch->ini_guid, sizeof(ch->ini_guid), "%04x:%04x:%04x:%04x",
2083 be16_to_cpu(guid[0]), be16_to_cpu(guid[1]), 2082 be16_to_cpu(guid[0]), be16_to_cpu(guid[1]),
2084 be16_to_cpu(guid[2]), be16_to_cpu(guid[3])); 2083 be16_to_cpu(guid[2]), be16_to_cpu(guid[3]));
diff --git a/drivers/mmc/host/renesas_sdhi_core.c b/drivers/mmc/host/renesas_sdhi_core.c
index fcf7235d5742..157e1d9e7725 100644
--- a/drivers/mmc/host/renesas_sdhi_core.c
+++ b/drivers/mmc/host/renesas_sdhi_core.c
@@ -24,6 +24,7 @@
24#include <linux/kernel.h> 24#include <linux/kernel.h>
25#include <linux/clk.h> 25#include <linux/clk.h>
26#include <linux/slab.h> 26#include <linux/slab.h>
27#include <linux/module.h>
27#include <linux/of_device.h> 28#include <linux/of_device.h>
28#include <linux/platform_device.h> 29#include <linux/platform_device.h>
29#include <linux/mmc/host.h> 30#include <linux/mmc/host.h>
@@ -667,3 +668,5 @@ int renesas_sdhi_remove(struct platform_device *pdev)
667 return 0; 668 return 0;
668} 669}
669EXPORT_SYMBOL_GPL(renesas_sdhi_remove); 670EXPORT_SYMBOL_GPL(renesas_sdhi_remove);
671
672MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/s3cmci.c b/drivers/mmc/host/s3cmci.c
index f7f157a62a4a..555c7f133eb8 100644
--- a/drivers/mmc/host/s3cmci.c
+++ b/drivers/mmc/host/s3cmci.c
@@ -1424,7 +1424,9 @@ static const struct file_operations s3cmci_fops_state = {
1424struct s3cmci_reg { 1424struct s3cmci_reg {
1425 unsigned short addr; 1425 unsigned short addr;
1426 unsigned char *name; 1426 unsigned char *name;
1427} debug_regs[] = { 1427};
1428
1429static const struct s3cmci_reg debug_regs[] = {
1428 DBG_REG(CON), 1430 DBG_REG(CON),
1429 DBG_REG(PRE), 1431 DBG_REG(PRE),
1430 DBG_REG(CMDARG), 1432 DBG_REG(CMDARG),
@@ -1446,7 +1448,7 @@ struct s3cmci_reg {
1446static int s3cmci_regs_show(struct seq_file *seq, void *v) 1448static int s3cmci_regs_show(struct seq_file *seq, void *v)
1447{ 1449{
1448 struct s3cmci_host *host = seq->private; 1450 struct s3cmci_host *host = seq->private;
1449 struct s3cmci_reg *rptr = debug_regs; 1451 const struct s3cmci_reg *rptr = debug_regs;
1450 1452
1451 for (; rptr->name; rptr++) 1453 for (; rptr->name; rptr++)
1452 seq_printf(seq, "SDI%s\t=0x%08x\n", rptr->name, 1454 seq_printf(seq, "SDI%s\t=0x%08x\n", rptr->name,
diff --git a/drivers/mux/core.c b/drivers/mux/core.c
index 2260063b0ea8..6e5cf9d9cd99 100644
--- a/drivers/mux/core.c
+++ b/drivers/mux/core.c
@@ -413,6 +413,7 @@ static int of_dev_node_match(struct device *dev, const void *data)
413 return dev->of_node == data; 413 return dev->of_node == data;
414} 414}
415 415
416/* Note this function returns a reference to the mux_chip dev. */
416static struct mux_chip *of_find_mux_chip_by_node(struct device_node *np) 417static struct mux_chip *of_find_mux_chip_by_node(struct device_node *np)
417{ 418{
418 struct device *dev; 419 struct device *dev;
@@ -466,6 +467,7 @@ struct mux_control *mux_control_get(struct device *dev, const char *mux_name)
466 (!args.args_count && (mux_chip->controllers > 1))) { 467 (!args.args_count && (mux_chip->controllers > 1))) {
467 dev_err(dev, "%pOF: wrong #mux-control-cells for %pOF\n", 468 dev_err(dev, "%pOF: wrong #mux-control-cells for %pOF\n",
468 np, args.np); 469 np, args.np);
470 put_device(&mux_chip->dev);
469 return ERR_PTR(-EINVAL); 471 return ERR_PTR(-EINVAL);
470 } 472 }
471 473
@@ -476,10 +478,10 @@ struct mux_control *mux_control_get(struct device *dev, const char *mux_name)
476 if (controller >= mux_chip->controllers) { 478 if (controller >= mux_chip->controllers) {
477 dev_err(dev, "%pOF: bad mux controller %u specified in %pOF\n", 479 dev_err(dev, "%pOF: bad mux controller %u specified in %pOF\n",
478 np, controller, args.np); 480 np, controller, args.np);
481 put_device(&mux_chip->dev);
479 return ERR_PTR(-EINVAL); 482 return ERR_PTR(-EINVAL);
480 } 483 }
481 484
482 get_device(&mux_chip->dev);
483 return &mux_chip->mux[controller]; 485 return &mux_chip->mux[controller];
484} 486}
485EXPORT_SYMBOL_GPL(mux_control_get); 487EXPORT_SYMBOL_GPL(mux_control_get);
diff --git a/drivers/net/can/flexcan.c b/drivers/net/can/flexcan.c
index 0626dcfd1f3d..760d2c07e3a2 100644
--- a/drivers/net/can/flexcan.c
+++ b/drivers/net/can/flexcan.c
@@ -526,7 +526,7 @@ static int flexcan_start_xmit(struct sk_buff *skb, struct net_device *dev)
526 data = be32_to_cpup((__be32 *)&cf->data[0]); 526 data = be32_to_cpup((__be32 *)&cf->data[0]);
527 flexcan_write(data, &priv->tx_mb->data[0]); 527 flexcan_write(data, &priv->tx_mb->data[0]);
528 } 528 }
529 if (cf->can_dlc > 3) { 529 if (cf->can_dlc > 4) {
530 data = be32_to_cpup((__be32 *)&cf->data[4]); 530 data = be32_to_cpup((__be32 *)&cf->data[4]);
531 flexcan_write(data, &priv->tx_mb->data[1]); 531 flexcan_write(data, &priv->tx_mb->data[1]);
532 } 532 }
diff --git a/drivers/net/can/usb/ems_usb.c b/drivers/net/can/usb/ems_usb.c
index b00358297424..12ff0020ecd6 100644
--- a/drivers/net/can/usb/ems_usb.c
+++ b/drivers/net/can/usb/ems_usb.c
@@ -395,6 +395,7 @@ static void ems_usb_rx_err(struct ems_usb *dev, struct ems_cpc_msg *msg)
395 395
396 if (dev->can.state == CAN_STATE_ERROR_WARNING || 396 if (dev->can.state == CAN_STATE_ERROR_WARNING ||
397 dev->can.state == CAN_STATE_ERROR_PASSIVE) { 397 dev->can.state == CAN_STATE_ERROR_PASSIVE) {
398 cf->can_id |= CAN_ERR_CRTL;
398 cf->data[1] = (txerr > rxerr) ? 399 cf->data[1] = (txerr > rxerr) ?
399 CAN_ERR_CRTL_TX_PASSIVE : CAN_ERR_CRTL_RX_PASSIVE; 400 CAN_ERR_CRTL_TX_PASSIVE : CAN_ERR_CRTL_RX_PASSIVE;
400 } 401 }
diff --git a/drivers/net/can/usb/gs_usb.c b/drivers/net/can/usb/gs_usb.c
index 68ac3e88a8ce..8bf80ad9dc44 100644
--- a/drivers/net/can/usb/gs_usb.c
+++ b/drivers/net/can/usb/gs_usb.c
@@ -449,7 +449,7 @@ static int gs_usb_set_bittiming(struct net_device *netdev)
449 dev_err(netdev->dev.parent, "Couldn't set bittimings (err=%d)", 449 dev_err(netdev->dev.parent, "Couldn't set bittimings (err=%d)",
450 rc); 450 rc);
451 451
452 return rc; 452 return (rc > 0) ? 0 : rc;
453} 453}
454 454
455static void gs_usb_xmit_callback(struct urb *urb) 455static void gs_usb_xmit_callback(struct urb *urb)
diff --git a/drivers/net/can/vxcan.c b/drivers/net/can/vxcan.c
index 8404e8852a0f..b4c4a2c76437 100644
--- a/drivers/net/can/vxcan.c
+++ b/drivers/net/can/vxcan.c
@@ -194,7 +194,7 @@ static int vxcan_newlink(struct net *net, struct net_device *dev,
194 tbp = peer_tb; 194 tbp = peer_tb;
195 } 195 }
196 196
197 if (tbp[IFLA_IFNAME]) { 197 if (ifmp && tbp[IFLA_IFNAME]) {
198 nla_strlcpy(ifname, tbp[IFLA_IFNAME], IFNAMSIZ); 198 nla_strlcpy(ifname, tbp[IFLA_IFNAME], IFNAMSIZ);
199 name_assign_type = NET_NAME_USER; 199 name_assign_type = NET_NAME_USER;
200 } else { 200 } else {
diff --git a/drivers/net/dsa/b53/b53_common.c b/drivers/net/dsa/b53/b53_common.c
index f5a8dd96fd75..4498ab897d94 100644
--- a/drivers/net/dsa/b53/b53_common.c
+++ b/drivers/net/dsa/b53/b53_common.c
@@ -1500,10 +1500,13 @@ static enum dsa_tag_protocol b53_get_tag_protocol(struct dsa_switch *ds,
1500{ 1500{
1501 struct b53_device *dev = ds->priv; 1501 struct b53_device *dev = ds->priv;
1502 1502
1503 /* Older models support a different tag format that we do not 1503 /* Older models (5325, 5365) support a different tag format that we do
1504 * support in net/dsa/tag_brcm.c yet. 1504 * not support in net/dsa/tag_brcm.c yet. 539x and 531x5 require managed
1505 * mode to be turned on which means we need to specifically manage ARL
1506 * misses on multicast addresses (TBD).
1505 */ 1507 */
1506 if (is5325(dev) || is5365(dev) || !b53_can_enable_brcm_tags(ds, port)) 1508 if (is5325(dev) || is5365(dev) || is539x(dev) || is531x5(dev) ||
1509 !b53_can_enable_brcm_tags(ds, port))
1507 return DSA_TAG_PROTO_NONE; 1510 return DSA_TAG_PROTO_NONE;
1508 1511
1509 /* Broadcom BCM58xx chips have a flow accelerator on Port 8 1512 /* Broadcom BCM58xx chips have a flow accelerator on Port 8
diff --git a/drivers/net/ethernet/3com/3c59x.c b/drivers/net/ethernet/3com/3c59x.c
index f4e13a7014bd..36c8950dbd2d 100644
--- a/drivers/net/ethernet/3com/3c59x.c
+++ b/drivers/net/ethernet/3com/3c59x.c
@@ -602,7 +602,7 @@ struct vortex_private {
602 struct sk_buff* rx_skbuff[RX_RING_SIZE]; 602 struct sk_buff* rx_skbuff[RX_RING_SIZE];
603 struct sk_buff* tx_skbuff[TX_RING_SIZE]; 603 struct sk_buff* tx_skbuff[TX_RING_SIZE];
604 unsigned int cur_rx, cur_tx; /* The next free ring entry */ 604 unsigned int cur_rx, cur_tx; /* The next free ring entry */
605 unsigned int dirty_rx, dirty_tx; /* The ring entries to be free()ed. */ 605 unsigned int dirty_tx; /* The ring entries to be free()ed. */
606 struct vortex_extra_stats xstats; /* NIC-specific extra stats */ 606 struct vortex_extra_stats xstats; /* NIC-specific extra stats */
607 struct sk_buff *tx_skb; /* Packet being eaten by bus master ctrl. */ 607 struct sk_buff *tx_skb; /* Packet being eaten by bus master ctrl. */
608 dma_addr_t tx_skb_dma; /* Allocated DMA address for bus master ctrl DMA. */ 608 dma_addr_t tx_skb_dma; /* Allocated DMA address for bus master ctrl DMA. */
@@ -618,7 +618,6 @@ struct vortex_private {
618 618
619 /* The remainder are related to chip state, mostly media selection. */ 619 /* The remainder are related to chip state, mostly media selection. */
620 struct timer_list timer; /* Media selection timer. */ 620 struct timer_list timer; /* Media selection timer. */
621 struct timer_list rx_oom_timer; /* Rx skb allocation retry timer */
622 int options; /* User-settable misc. driver options. */ 621 int options; /* User-settable misc. driver options. */
623 unsigned int media_override:4, /* Passed-in media type. */ 622 unsigned int media_override:4, /* Passed-in media type. */
624 default_media:4, /* Read from the EEPROM/Wn3_Config. */ 623 default_media:4, /* Read from the EEPROM/Wn3_Config. */
@@ -760,7 +759,6 @@ static void mdio_sync(struct vortex_private *vp, int bits);
760static int mdio_read(struct net_device *dev, int phy_id, int location); 759static int mdio_read(struct net_device *dev, int phy_id, int location);
761static void mdio_write(struct net_device *vp, int phy_id, int location, int value); 760static void mdio_write(struct net_device *vp, int phy_id, int location, int value);
762static void vortex_timer(struct timer_list *t); 761static void vortex_timer(struct timer_list *t);
763static void rx_oom_timer(struct timer_list *t);
764static netdev_tx_t vortex_start_xmit(struct sk_buff *skb, 762static netdev_tx_t vortex_start_xmit(struct sk_buff *skb,
765 struct net_device *dev); 763 struct net_device *dev);
766static netdev_tx_t boomerang_start_xmit(struct sk_buff *skb, 764static netdev_tx_t boomerang_start_xmit(struct sk_buff *skb,
@@ -1601,7 +1599,6 @@ vortex_up(struct net_device *dev)
1601 1599
1602 timer_setup(&vp->timer, vortex_timer, 0); 1600 timer_setup(&vp->timer, vortex_timer, 0);
1603 mod_timer(&vp->timer, RUN_AT(media_tbl[dev->if_port].wait)); 1601 mod_timer(&vp->timer, RUN_AT(media_tbl[dev->if_port].wait));
1604 timer_setup(&vp->rx_oom_timer, rx_oom_timer, 0);
1605 1602
1606 if (vortex_debug > 1) 1603 if (vortex_debug > 1)
1607 pr_debug("%s: Initial media type %s.\n", 1604 pr_debug("%s: Initial media type %s.\n",
@@ -1676,7 +1673,7 @@ vortex_up(struct net_device *dev)
1676 window_write16(vp, 0x0040, 4, Wn4_NetDiag); 1673 window_write16(vp, 0x0040, 4, Wn4_NetDiag);
1677 1674
1678 if (vp->full_bus_master_rx) { /* Boomerang bus master. */ 1675 if (vp->full_bus_master_rx) { /* Boomerang bus master. */
1679 vp->cur_rx = vp->dirty_rx = 0; 1676 vp->cur_rx = 0;
1680 /* Initialize the RxEarly register as recommended. */ 1677 /* Initialize the RxEarly register as recommended. */
1681 iowrite16(SetRxThreshold + (1536>>2), ioaddr + EL3_CMD); 1678 iowrite16(SetRxThreshold + (1536>>2), ioaddr + EL3_CMD);
1682 iowrite32(0x0020, ioaddr + PktStatus); 1679 iowrite32(0x0020, ioaddr + PktStatus);
@@ -1729,6 +1726,7 @@ vortex_open(struct net_device *dev)
1729 struct vortex_private *vp = netdev_priv(dev); 1726 struct vortex_private *vp = netdev_priv(dev);
1730 int i; 1727 int i;
1731 int retval; 1728 int retval;
1729 dma_addr_t dma;
1732 1730
1733 /* Use the now-standard shared IRQ implementation. */ 1731 /* Use the now-standard shared IRQ implementation. */
1734 if ((retval = request_irq(dev->irq, vp->full_bus_master_rx ? 1732 if ((retval = request_irq(dev->irq, vp->full_bus_master_rx ?
@@ -1753,7 +1751,11 @@ vortex_open(struct net_device *dev)
1753 break; /* Bad news! */ 1751 break; /* Bad news! */
1754 1752
1755 skb_reserve(skb, NET_IP_ALIGN); /* Align IP on 16 byte boundaries */ 1753 skb_reserve(skb, NET_IP_ALIGN); /* Align IP on 16 byte boundaries */
1756 vp->rx_ring[i].addr = cpu_to_le32(pci_map_single(VORTEX_PCI(vp), skb->data, PKT_BUF_SZ, PCI_DMA_FROMDEVICE)); 1754 dma = pci_map_single(VORTEX_PCI(vp), skb->data,
1755 PKT_BUF_SZ, PCI_DMA_FROMDEVICE);
1756 if (dma_mapping_error(&VORTEX_PCI(vp)->dev, dma))
1757 break;
1758 vp->rx_ring[i].addr = cpu_to_le32(dma);
1757 } 1759 }
1758 if (i != RX_RING_SIZE) { 1760 if (i != RX_RING_SIZE) {
1759 pr_emerg("%s: no memory for rx ring\n", dev->name); 1761 pr_emerg("%s: no memory for rx ring\n", dev->name);
@@ -2067,6 +2069,12 @@ vortex_start_xmit(struct sk_buff *skb, struct net_device *dev)
2067 int len = (skb->len + 3) & ~3; 2069 int len = (skb->len + 3) & ~3;
2068 vp->tx_skb_dma = pci_map_single(VORTEX_PCI(vp), skb->data, len, 2070 vp->tx_skb_dma = pci_map_single(VORTEX_PCI(vp), skb->data, len,
2069 PCI_DMA_TODEVICE); 2071 PCI_DMA_TODEVICE);
2072 if (dma_mapping_error(&VORTEX_PCI(vp)->dev, vp->tx_skb_dma)) {
2073 dev_kfree_skb_any(skb);
2074 dev->stats.tx_dropped++;
2075 return NETDEV_TX_OK;
2076 }
2077
2070 spin_lock_irq(&vp->window_lock); 2078 spin_lock_irq(&vp->window_lock);
2071 window_set(vp, 7); 2079 window_set(vp, 7);
2072 iowrite32(vp->tx_skb_dma, ioaddr + Wn7_MasterAddr); 2080 iowrite32(vp->tx_skb_dma, ioaddr + Wn7_MasterAddr);
@@ -2593,7 +2601,7 @@ boomerang_rx(struct net_device *dev)
2593 int entry = vp->cur_rx % RX_RING_SIZE; 2601 int entry = vp->cur_rx % RX_RING_SIZE;
2594 void __iomem *ioaddr = vp->ioaddr; 2602 void __iomem *ioaddr = vp->ioaddr;
2595 int rx_status; 2603 int rx_status;
2596 int rx_work_limit = vp->dirty_rx + RX_RING_SIZE - vp->cur_rx; 2604 int rx_work_limit = RX_RING_SIZE;
2597 2605
2598 if (vortex_debug > 5) 2606 if (vortex_debug > 5)
2599 pr_debug("boomerang_rx(): status %4.4x\n", ioread16(ioaddr+EL3_STATUS)); 2607 pr_debug("boomerang_rx(): status %4.4x\n", ioread16(ioaddr+EL3_STATUS));
@@ -2614,7 +2622,8 @@ boomerang_rx(struct net_device *dev)
2614 } else { 2622 } else {
2615 /* The packet length: up to 4.5K!. */ 2623 /* The packet length: up to 4.5K!. */
2616 int pkt_len = rx_status & 0x1fff; 2624 int pkt_len = rx_status & 0x1fff;
2617 struct sk_buff *skb; 2625 struct sk_buff *skb, *newskb;
2626 dma_addr_t newdma;
2618 dma_addr_t dma = le32_to_cpu(vp->rx_ring[entry].addr); 2627 dma_addr_t dma = le32_to_cpu(vp->rx_ring[entry].addr);
2619 2628
2620 if (vortex_debug > 4) 2629 if (vortex_debug > 4)
@@ -2633,9 +2642,27 @@ boomerang_rx(struct net_device *dev)
2633 pci_dma_sync_single_for_device(VORTEX_PCI(vp), dma, PKT_BUF_SZ, PCI_DMA_FROMDEVICE); 2642 pci_dma_sync_single_for_device(VORTEX_PCI(vp), dma, PKT_BUF_SZ, PCI_DMA_FROMDEVICE);
2634 vp->rx_copy++; 2643 vp->rx_copy++;
2635 } else { 2644 } else {
2645 /* Pre-allocate the replacement skb. If it or its
2646 * mapping fails then recycle the buffer thats already
2647 * in place
2648 */
2649 newskb = netdev_alloc_skb_ip_align(dev, PKT_BUF_SZ);
2650 if (!newskb) {
2651 dev->stats.rx_dropped++;
2652 goto clear_complete;
2653 }
2654 newdma = pci_map_single(VORTEX_PCI(vp), newskb->data,
2655 PKT_BUF_SZ, PCI_DMA_FROMDEVICE);
2656 if (dma_mapping_error(&VORTEX_PCI(vp)->dev, newdma)) {
2657 dev->stats.rx_dropped++;
2658 consume_skb(newskb);
2659 goto clear_complete;
2660 }
2661
2636 /* Pass up the skbuff already on the Rx ring. */ 2662 /* Pass up the skbuff already on the Rx ring. */
2637 skb = vp->rx_skbuff[entry]; 2663 skb = vp->rx_skbuff[entry];
2638 vp->rx_skbuff[entry] = NULL; 2664 vp->rx_skbuff[entry] = newskb;
2665 vp->rx_ring[entry].addr = cpu_to_le32(newdma);
2639 skb_put(skb, pkt_len); 2666 skb_put(skb, pkt_len);
2640 pci_unmap_single(VORTEX_PCI(vp), dma, PKT_BUF_SZ, PCI_DMA_FROMDEVICE); 2667 pci_unmap_single(VORTEX_PCI(vp), dma, PKT_BUF_SZ, PCI_DMA_FROMDEVICE);
2641 vp->rx_nocopy++; 2668 vp->rx_nocopy++;
@@ -2653,55 +2680,15 @@ boomerang_rx(struct net_device *dev)
2653 netif_rx(skb); 2680 netif_rx(skb);
2654 dev->stats.rx_packets++; 2681 dev->stats.rx_packets++;
2655 } 2682 }
2656 entry = (++vp->cur_rx) % RX_RING_SIZE;
2657 }
2658 /* Refill the Rx ring buffers. */
2659 for (; vp->cur_rx - vp->dirty_rx > 0; vp->dirty_rx++) {
2660 struct sk_buff *skb;
2661 entry = vp->dirty_rx % RX_RING_SIZE;
2662 if (vp->rx_skbuff[entry] == NULL) {
2663 skb = netdev_alloc_skb_ip_align(dev, PKT_BUF_SZ);
2664 if (skb == NULL) {
2665 static unsigned long last_jif;
2666 if (time_after(jiffies, last_jif + 10 * HZ)) {
2667 pr_warn("%s: memory shortage\n",
2668 dev->name);
2669 last_jif = jiffies;
2670 }
2671 if ((vp->cur_rx - vp->dirty_rx) == RX_RING_SIZE)
2672 mod_timer(&vp->rx_oom_timer, RUN_AT(HZ * 1));
2673 break; /* Bad news! */
2674 }
2675 2683
2676 vp->rx_ring[entry].addr = cpu_to_le32(pci_map_single(VORTEX_PCI(vp), skb->data, PKT_BUF_SZ, PCI_DMA_FROMDEVICE)); 2684clear_complete:
2677 vp->rx_skbuff[entry] = skb;
2678 }
2679 vp->rx_ring[entry].status = 0; /* Clear complete bit. */ 2685 vp->rx_ring[entry].status = 0; /* Clear complete bit. */
2680 iowrite16(UpUnstall, ioaddr + EL3_CMD); 2686 iowrite16(UpUnstall, ioaddr + EL3_CMD);
2687 entry = (++vp->cur_rx) % RX_RING_SIZE;
2681 } 2688 }
2682 return 0; 2689 return 0;
2683} 2690}
2684 2691
2685/*
2686 * If we've hit a total OOM refilling the Rx ring we poll once a second
2687 * for some memory. Otherwise there is no way to restart the rx process.
2688 */
2689static void
2690rx_oom_timer(struct timer_list *t)
2691{
2692 struct vortex_private *vp = from_timer(vp, t, rx_oom_timer);
2693 struct net_device *dev = vp->mii.dev;
2694
2695 spin_lock_irq(&vp->lock);
2696 if ((vp->cur_rx - vp->dirty_rx) == RX_RING_SIZE) /* This test is redundant, but makes me feel good */
2697 boomerang_rx(dev);
2698 if (vortex_debug > 1) {
2699 pr_debug("%s: rx_oom_timer %s\n", dev->name,
2700 ((vp->cur_rx - vp->dirty_rx) != RX_RING_SIZE) ? "succeeded" : "retrying");
2701 }
2702 spin_unlock_irq(&vp->lock);
2703}
2704
2705static void 2692static void
2706vortex_down(struct net_device *dev, int final_down) 2693vortex_down(struct net_device *dev, int final_down)
2707{ 2694{
@@ -2711,7 +2698,6 @@ vortex_down(struct net_device *dev, int final_down)
2711 netdev_reset_queue(dev); 2698 netdev_reset_queue(dev);
2712 netif_stop_queue(dev); 2699 netif_stop_queue(dev);
2713 2700
2714 del_timer_sync(&vp->rx_oom_timer);
2715 del_timer_sync(&vp->timer); 2701 del_timer_sync(&vp->timer);
2716 2702
2717 /* Turn off statistics ASAP. We update dev->stats below. */ 2703 /* Turn off statistics ASAP. We update dev->stats below. */
diff --git a/drivers/net/ethernet/amazon/ena/ena_netdev.c b/drivers/net/ethernet/amazon/ena/ena_netdev.c
index 97c5a89a9cf7..fbe21a817bd8 100644
--- a/drivers/net/ethernet/amazon/ena/ena_netdev.c
+++ b/drivers/net/ethernet/amazon/ena/ena_netdev.c
@@ -75,6 +75,9 @@ static struct workqueue_struct *ena_wq;
75MODULE_DEVICE_TABLE(pci, ena_pci_tbl); 75MODULE_DEVICE_TABLE(pci, ena_pci_tbl);
76 76
77static int ena_rss_init_default(struct ena_adapter *adapter); 77static int ena_rss_init_default(struct ena_adapter *adapter);
78static void check_for_admin_com_state(struct ena_adapter *adapter);
79static void ena_destroy_device(struct ena_adapter *adapter);
80static int ena_restore_device(struct ena_adapter *adapter);
78 81
79static void ena_tx_timeout(struct net_device *dev) 82static void ena_tx_timeout(struct net_device *dev)
80{ 83{
@@ -1565,7 +1568,7 @@ static int ena_rss_configure(struct ena_adapter *adapter)
1565 1568
1566static int ena_up_complete(struct ena_adapter *adapter) 1569static int ena_up_complete(struct ena_adapter *adapter)
1567{ 1570{
1568 int rc, i; 1571 int rc;
1569 1572
1570 rc = ena_rss_configure(adapter); 1573 rc = ena_rss_configure(adapter);
1571 if (rc) 1574 if (rc)
@@ -1584,17 +1587,6 @@ static int ena_up_complete(struct ena_adapter *adapter)
1584 1587
1585 ena_napi_enable_all(adapter); 1588 ena_napi_enable_all(adapter);
1586 1589
1587 /* Enable completion queues interrupt */
1588 for (i = 0; i < adapter->num_queues; i++)
1589 ena_unmask_interrupt(&adapter->tx_ring[i],
1590 &adapter->rx_ring[i]);
1591
1592 /* schedule napi in case we had pending packets
1593 * from the last time we disable napi
1594 */
1595 for (i = 0; i < adapter->num_queues; i++)
1596 napi_schedule(&adapter->ena_napi[i].napi);
1597
1598 return 0; 1590 return 0;
1599} 1591}
1600 1592
@@ -1731,7 +1723,7 @@ create_err:
1731 1723
1732static int ena_up(struct ena_adapter *adapter) 1724static int ena_up(struct ena_adapter *adapter)
1733{ 1725{
1734 int rc; 1726 int rc, i;
1735 1727
1736 netdev_dbg(adapter->netdev, "%s\n", __func__); 1728 netdev_dbg(adapter->netdev, "%s\n", __func__);
1737 1729
@@ -1774,6 +1766,17 @@ static int ena_up(struct ena_adapter *adapter)
1774 1766
1775 set_bit(ENA_FLAG_DEV_UP, &adapter->flags); 1767 set_bit(ENA_FLAG_DEV_UP, &adapter->flags);
1776 1768
1769 /* Enable completion queues interrupt */
1770 for (i = 0; i < adapter->num_queues; i++)
1771 ena_unmask_interrupt(&adapter->tx_ring[i],
1772 &adapter->rx_ring[i]);
1773
1774 /* schedule napi in case we had pending packets
1775 * from the last time we disable napi
1776 */
1777 for (i = 0; i < adapter->num_queues; i++)
1778 napi_schedule(&adapter->ena_napi[i].napi);
1779
1777 return rc; 1780 return rc;
1778 1781
1779err_up: 1782err_up:
@@ -1884,6 +1887,17 @@ static int ena_close(struct net_device *netdev)
1884 if (test_bit(ENA_FLAG_DEV_UP, &adapter->flags)) 1887 if (test_bit(ENA_FLAG_DEV_UP, &adapter->flags))
1885 ena_down(adapter); 1888 ena_down(adapter);
1886 1889
1890 /* Check for device status and issue reset if needed*/
1891 check_for_admin_com_state(adapter);
1892 if (unlikely(test_bit(ENA_FLAG_TRIGGER_RESET, &adapter->flags))) {
1893 netif_err(adapter, ifdown, adapter->netdev,
1894 "Destroy failure, restarting device\n");
1895 ena_dump_stats_to_dmesg(adapter);
1896 /* rtnl lock already obtained in dev_ioctl() layer */
1897 ena_destroy_device(adapter);
1898 ena_restore_device(adapter);
1899 }
1900
1887 return 0; 1901 return 0;
1888} 1902}
1889 1903
@@ -2544,11 +2558,12 @@ static void ena_destroy_device(struct ena_adapter *adapter)
2544 2558
2545 ena_com_set_admin_running_state(ena_dev, false); 2559 ena_com_set_admin_running_state(ena_dev, false);
2546 2560
2547 ena_close(netdev); 2561 if (test_bit(ENA_FLAG_DEV_UP, &adapter->flags))
2562 ena_down(adapter);
2548 2563
2549 /* Before releasing the ENA resources, a device reset is required. 2564 /* Before releasing the ENA resources, a device reset is required.
2550 * (to prevent the device from accessing them). 2565 * (to prevent the device from accessing them).
2551 * In case the reset flag is set and the device is up, ena_close 2566 * In case the reset flag is set and the device is up, ena_down()
2552 * already perform the reset, so it can be skipped. 2567 * already perform the reset, so it can be skipped.
2553 */ 2568 */
2554 if (!(test_bit(ENA_FLAG_TRIGGER_RESET, &adapter->flags) && dev_up)) 2569 if (!(test_bit(ENA_FLAG_TRIGGER_RESET, &adapter->flags) && dev_up))
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c
index 5ee18660bc33..c9617675f934 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c
@@ -70,7 +70,7 @@ static int bnxt_vf_ndo_prep(struct bnxt *bp, int vf_id)
70 netdev_err(bp->dev, "vf ndo called though sriov is disabled\n"); 70 netdev_err(bp->dev, "vf ndo called though sriov is disabled\n");
71 return -EINVAL; 71 return -EINVAL;
72 } 72 }
73 if (vf_id >= bp->pf.max_vfs) { 73 if (vf_id >= bp->pf.active_vfs) {
74 netdev_err(bp->dev, "Invalid VF id %d\n", vf_id); 74 netdev_err(bp->dev, "Invalid VF id %d\n", vf_id);
75 return -EINVAL; 75 return -EINVAL;
76 } 76 }
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_tc.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_tc.c
index 3d201d7324bd..d8fee26cd45e 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_tc.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_tc.c
@@ -421,7 +421,7 @@ static int bnxt_hwrm_cfa_flow_alloc(struct bnxt *bp, struct bnxt_tc_flow *flow,
421 } 421 }
422 422
423 /* If all IP and L4 fields are wildcarded then this is an L2 flow */ 423 /* If all IP and L4 fields are wildcarded then this is an L2 flow */
424 if (is_wildcard(&l3_mask, sizeof(l3_mask)) && 424 if (is_wildcard(l3_mask, sizeof(*l3_mask)) &&
425 is_wildcard(&flow->l4_mask, sizeof(flow->l4_mask))) { 425 is_wildcard(&flow->l4_mask, sizeof(flow->l4_mask))) {
426 flow_flags |= CFA_FLOW_ALLOC_REQ_FLAGS_FLOWTYPE_L2; 426 flow_flags |= CFA_FLOW_ALLOC_REQ_FLAGS_FLOWTYPE_L2;
427 } else { 427 } else {
diff --git a/drivers/net/ethernet/chelsio/cxgb4/cxgb4.h b/drivers/net/ethernet/chelsio/cxgb4/cxgb4.h
index 6f9fa6e3c42a..d8424ed16c33 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/cxgb4.h
+++ b/drivers/net/ethernet/chelsio/cxgb4/cxgb4.h
@@ -344,7 +344,6 @@ struct adapter_params {
344 344
345 unsigned int sf_size; /* serial flash size in bytes */ 345 unsigned int sf_size; /* serial flash size in bytes */
346 unsigned int sf_nsec; /* # of flash sectors */ 346 unsigned int sf_nsec; /* # of flash sectors */
347 unsigned int sf_fw_start; /* start of FW image in flash */
348 347
349 unsigned int fw_vers; /* firmware version */ 348 unsigned int fw_vers; /* firmware version */
350 unsigned int bs_vers; /* bootstrap version */ 349 unsigned int bs_vers; /* bootstrap version */
diff --git a/drivers/net/ethernet/chelsio/cxgb4/t4_hw.c b/drivers/net/ethernet/chelsio/cxgb4/t4_hw.c
index f63210f15579..375ef86a84da 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/t4_hw.c
+++ b/drivers/net/ethernet/chelsio/cxgb4/t4_hw.c
@@ -2844,8 +2844,6 @@ enum {
2844 SF_RD_DATA_FAST = 0xb, /* read flash */ 2844 SF_RD_DATA_FAST = 0xb, /* read flash */
2845 SF_RD_ID = 0x9f, /* read ID */ 2845 SF_RD_ID = 0x9f, /* read ID */
2846 SF_ERASE_SECTOR = 0xd8, /* erase sector */ 2846 SF_ERASE_SECTOR = 0xd8, /* erase sector */
2847
2848 FW_MAX_SIZE = 16 * SF_SEC_SIZE,
2849}; 2847};
2850 2848
2851/** 2849/**
@@ -3558,8 +3556,9 @@ int t4_load_fw(struct adapter *adap, const u8 *fw_data, unsigned int size)
3558 const __be32 *p = (const __be32 *)fw_data; 3556 const __be32 *p = (const __be32 *)fw_data;
3559 const struct fw_hdr *hdr = (const struct fw_hdr *)fw_data; 3557 const struct fw_hdr *hdr = (const struct fw_hdr *)fw_data;
3560 unsigned int sf_sec_size = adap->params.sf_size / adap->params.sf_nsec; 3558 unsigned int sf_sec_size = adap->params.sf_size / adap->params.sf_nsec;
3561 unsigned int fw_img_start = adap->params.sf_fw_start; 3559 unsigned int fw_start_sec = FLASH_FW_START_SEC;
3562 unsigned int fw_start_sec = fw_img_start / sf_sec_size; 3560 unsigned int fw_size = FLASH_FW_MAX_SIZE;
3561 unsigned int fw_start = FLASH_FW_START;
3563 3562
3564 if (!size) { 3563 if (!size) {
3565 dev_err(adap->pdev_dev, "FW image has no data\n"); 3564 dev_err(adap->pdev_dev, "FW image has no data\n");
@@ -3575,9 +3574,9 @@ int t4_load_fw(struct adapter *adap, const u8 *fw_data, unsigned int size)
3575 "FW image size differs from size in FW header\n"); 3574 "FW image size differs from size in FW header\n");
3576 return -EINVAL; 3575 return -EINVAL;
3577 } 3576 }
3578 if (size > FW_MAX_SIZE) { 3577 if (size > fw_size) {
3579 dev_err(adap->pdev_dev, "FW image too large, max is %u bytes\n", 3578 dev_err(adap->pdev_dev, "FW image too large, max is %u bytes\n",
3580 FW_MAX_SIZE); 3579 fw_size);
3581 return -EFBIG; 3580 return -EFBIG;
3582 } 3581 }
3583 if (!t4_fw_matches_chip(adap, hdr)) 3582 if (!t4_fw_matches_chip(adap, hdr))
@@ -3604,11 +3603,11 @@ int t4_load_fw(struct adapter *adap, const u8 *fw_data, unsigned int size)
3604 */ 3603 */
3605 memcpy(first_page, fw_data, SF_PAGE_SIZE); 3604 memcpy(first_page, fw_data, SF_PAGE_SIZE);
3606 ((struct fw_hdr *)first_page)->fw_ver = cpu_to_be32(0xffffffff); 3605 ((struct fw_hdr *)first_page)->fw_ver = cpu_to_be32(0xffffffff);
3607 ret = t4_write_flash(adap, fw_img_start, SF_PAGE_SIZE, first_page); 3606 ret = t4_write_flash(adap, fw_start, SF_PAGE_SIZE, first_page);
3608 if (ret) 3607 if (ret)
3609 goto out; 3608 goto out;
3610 3609
3611 addr = fw_img_start; 3610 addr = fw_start;
3612 for (size -= SF_PAGE_SIZE; size; size -= SF_PAGE_SIZE) { 3611 for (size -= SF_PAGE_SIZE; size; size -= SF_PAGE_SIZE) {
3613 addr += SF_PAGE_SIZE; 3612 addr += SF_PAGE_SIZE;
3614 fw_data += SF_PAGE_SIZE; 3613 fw_data += SF_PAGE_SIZE;
@@ -3618,7 +3617,7 @@ int t4_load_fw(struct adapter *adap, const u8 *fw_data, unsigned int size)
3618 } 3617 }
3619 3618
3620 ret = t4_write_flash(adap, 3619 ret = t4_write_flash(adap,
3621 fw_img_start + offsetof(struct fw_hdr, fw_ver), 3620 fw_start + offsetof(struct fw_hdr, fw_ver),
3622 sizeof(hdr->fw_ver), (const u8 *)&hdr->fw_ver); 3621 sizeof(hdr->fw_ver), (const u8 *)&hdr->fw_ver);
3623out: 3622out:
3624 if (ret) 3623 if (ret)
diff --git a/drivers/net/ethernet/freescale/fec_main.c b/drivers/net/ethernet/freescale/fec_main.c
index 8184d2fca9be..a74300a4459c 100644
--- a/drivers/net/ethernet/freescale/fec_main.c
+++ b/drivers/net/ethernet/freescale/fec_main.c
@@ -3469,6 +3469,10 @@ fec_probe(struct platform_device *pdev)
3469 goto failed_regulator; 3469 goto failed_regulator;
3470 } 3470 }
3471 } else { 3471 } else {
3472 if (PTR_ERR(fep->reg_phy) == -EPROBE_DEFER) {
3473 ret = -EPROBE_DEFER;
3474 goto failed_regulator;
3475 }
3472 fep->reg_phy = NULL; 3476 fep->reg_phy = NULL;
3473 } 3477 }
3474 3478
@@ -3552,8 +3556,9 @@ failed_clk_ipg:
3552failed_clk: 3556failed_clk:
3553 if (of_phy_is_fixed_link(np)) 3557 if (of_phy_is_fixed_link(np))
3554 of_phy_deregister_fixed_link(np); 3558 of_phy_deregister_fixed_link(np);
3555failed_phy:
3556 of_node_put(phy_node); 3559 of_node_put(phy_node);
3560failed_phy:
3561 dev_id--;
3557failed_ioremap: 3562failed_ioremap:
3558 free_netdev(ndev); 3563 free_netdev(ndev);
3559 3564
diff --git a/drivers/net/ethernet/freescale/gianfar_ptp.c b/drivers/net/ethernet/freescale/gianfar_ptp.c
index 544114281ea7..9f8d4f8e57e3 100644
--- a/drivers/net/ethernet/freescale/gianfar_ptp.c
+++ b/drivers/net/ethernet/freescale/gianfar_ptp.c
@@ -319,11 +319,10 @@ static int ptp_gianfar_adjtime(struct ptp_clock_info *ptp, s64 delta)
319 now = tmr_cnt_read(etsects); 319 now = tmr_cnt_read(etsects);
320 now += delta; 320 now += delta;
321 tmr_cnt_write(etsects, now); 321 tmr_cnt_write(etsects, now);
322 set_fipers(etsects);
322 323
323 spin_unlock_irqrestore(&etsects->lock, flags); 324 spin_unlock_irqrestore(&etsects->lock, flags);
324 325
325 set_fipers(etsects);
326
327 return 0; 326 return 0;
328} 327}
329 328
diff --git a/drivers/net/ethernet/intel/e1000/e1000.h b/drivers/net/ethernet/intel/e1000/e1000.h
index d7bdea79e9fa..8fd2458060a0 100644
--- a/drivers/net/ethernet/intel/e1000/e1000.h
+++ b/drivers/net/ethernet/intel/e1000/e1000.h
@@ -331,7 +331,8 @@ struct e1000_adapter {
331enum e1000_state_t { 331enum e1000_state_t {
332 __E1000_TESTING, 332 __E1000_TESTING,
333 __E1000_RESETTING, 333 __E1000_RESETTING,
334 __E1000_DOWN 334 __E1000_DOWN,
335 __E1000_DISABLED
335}; 336};
336 337
337#undef pr_fmt 338#undef pr_fmt
diff --git a/drivers/net/ethernet/intel/e1000/e1000_main.c b/drivers/net/ethernet/intel/e1000/e1000_main.c
index 1982f7917a8d..3dd4aeb2706d 100644
--- a/drivers/net/ethernet/intel/e1000/e1000_main.c
+++ b/drivers/net/ethernet/intel/e1000/e1000_main.c
@@ -945,7 +945,7 @@ static int e1000_init_hw_struct(struct e1000_adapter *adapter,
945static int e1000_probe(struct pci_dev *pdev, const struct pci_device_id *ent) 945static int e1000_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
946{ 946{
947 struct net_device *netdev; 947 struct net_device *netdev;
948 struct e1000_adapter *adapter; 948 struct e1000_adapter *adapter = NULL;
949 struct e1000_hw *hw; 949 struct e1000_hw *hw;
950 950
951 static int cards_found; 951 static int cards_found;
@@ -955,6 +955,7 @@ static int e1000_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
955 u16 tmp = 0; 955 u16 tmp = 0;
956 u16 eeprom_apme_mask = E1000_EEPROM_APME; 956 u16 eeprom_apme_mask = E1000_EEPROM_APME;
957 int bars, need_ioport; 957 int bars, need_ioport;
958 bool disable_dev = false;
958 959
959 /* do not allocate ioport bars when not needed */ 960 /* do not allocate ioport bars when not needed */
960 need_ioport = e1000_is_need_ioport(pdev); 961 need_ioport = e1000_is_need_ioport(pdev);
@@ -1259,11 +1260,13 @@ err_mdio_ioremap:
1259 iounmap(hw->ce4100_gbe_mdio_base_virt); 1260 iounmap(hw->ce4100_gbe_mdio_base_virt);
1260 iounmap(hw->hw_addr); 1261 iounmap(hw->hw_addr);
1261err_ioremap: 1262err_ioremap:
1263 disable_dev = !test_and_set_bit(__E1000_DISABLED, &adapter->flags);
1262 free_netdev(netdev); 1264 free_netdev(netdev);
1263err_alloc_etherdev: 1265err_alloc_etherdev:
1264 pci_release_selected_regions(pdev, bars); 1266 pci_release_selected_regions(pdev, bars);
1265err_pci_reg: 1267err_pci_reg:
1266 pci_disable_device(pdev); 1268 if (!adapter || disable_dev)
1269 pci_disable_device(pdev);
1267 return err; 1270 return err;
1268} 1271}
1269 1272
@@ -1281,6 +1284,7 @@ static void e1000_remove(struct pci_dev *pdev)
1281 struct net_device *netdev = pci_get_drvdata(pdev); 1284 struct net_device *netdev = pci_get_drvdata(pdev);
1282 struct e1000_adapter *adapter = netdev_priv(netdev); 1285 struct e1000_adapter *adapter = netdev_priv(netdev);
1283 struct e1000_hw *hw = &adapter->hw; 1286 struct e1000_hw *hw = &adapter->hw;
1287 bool disable_dev;
1284 1288
1285 e1000_down_and_stop(adapter); 1289 e1000_down_and_stop(adapter);
1286 e1000_release_manageability(adapter); 1290 e1000_release_manageability(adapter);
@@ -1299,9 +1303,11 @@ static void e1000_remove(struct pci_dev *pdev)
1299 iounmap(hw->flash_address); 1303 iounmap(hw->flash_address);
1300 pci_release_selected_regions(pdev, adapter->bars); 1304 pci_release_selected_regions(pdev, adapter->bars);
1301 1305
1306 disable_dev = !test_and_set_bit(__E1000_DISABLED, &adapter->flags);
1302 free_netdev(netdev); 1307 free_netdev(netdev);
1303 1308
1304 pci_disable_device(pdev); 1309 if (disable_dev)
1310 pci_disable_device(pdev);
1305} 1311}
1306 1312
1307/** 1313/**
@@ -5156,7 +5162,8 @@ static int __e1000_shutdown(struct pci_dev *pdev, bool *enable_wake)
5156 if (netif_running(netdev)) 5162 if (netif_running(netdev))
5157 e1000_free_irq(adapter); 5163 e1000_free_irq(adapter);
5158 5164
5159 pci_disable_device(pdev); 5165 if (!test_and_set_bit(__E1000_DISABLED, &adapter->flags))
5166 pci_disable_device(pdev);
5160 5167
5161 return 0; 5168 return 0;
5162} 5169}
@@ -5200,6 +5207,10 @@ static int e1000_resume(struct pci_dev *pdev)
5200 pr_err("Cannot enable PCI device from suspend\n"); 5207 pr_err("Cannot enable PCI device from suspend\n");
5201 return err; 5208 return err;
5202 } 5209 }
5210
5211 /* flush memory to make sure state is correct */
5212 smp_mb__before_atomic();
5213 clear_bit(__E1000_DISABLED, &adapter->flags);
5203 pci_set_master(pdev); 5214 pci_set_master(pdev);
5204 5215
5205 pci_enable_wake(pdev, PCI_D3hot, 0); 5216 pci_enable_wake(pdev, PCI_D3hot, 0);
@@ -5274,7 +5285,9 @@ static pci_ers_result_t e1000_io_error_detected(struct pci_dev *pdev,
5274 5285
5275 if (netif_running(netdev)) 5286 if (netif_running(netdev))
5276 e1000_down(adapter); 5287 e1000_down(adapter);
5277 pci_disable_device(pdev); 5288
5289 if (!test_and_set_bit(__E1000_DISABLED, &adapter->flags))
5290 pci_disable_device(pdev);
5278 5291
5279 /* Request a slot slot reset. */ 5292 /* Request a slot slot reset. */
5280 return PCI_ERS_RESULT_NEED_RESET; 5293 return PCI_ERS_RESULT_NEED_RESET;
@@ -5302,6 +5315,10 @@ static pci_ers_result_t e1000_io_slot_reset(struct pci_dev *pdev)
5302 pr_err("Cannot re-enable PCI device after reset.\n"); 5315 pr_err("Cannot re-enable PCI device after reset.\n");
5303 return PCI_ERS_RESULT_DISCONNECT; 5316 return PCI_ERS_RESULT_DISCONNECT;
5304 } 5317 }
5318
5319 /* flush memory to make sure state is correct */
5320 smp_mb__before_atomic();
5321 clear_bit(__E1000_DISABLED, &adapter->flags);
5305 pci_set_master(pdev); 5322 pci_set_master(pdev);
5306 5323
5307 pci_enable_wake(pdev, PCI_D3hot, 0); 5324 pci_enable_wake(pdev, PCI_D3hot, 0);
diff --git a/drivers/net/ethernet/intel/e1000e/ich8lan.c b/drivers/net/ethernet/intel/e1000e/ich8lan.c
index d6d4ed7acf03..31277d3bb7dc 100644
--- a/drivers/net/ethernet/intel/e1000e/ich8lan.c
+++ b/drivers/net/ethernet/intel/e1000e/ich8lan.c
@@ -1367,6 +1367,9 @@ out:
1367 * Checks to see of the link status of the hardware has changed. If a 1367 * Checks to see of the link status of the hardware has changed. If a
1368 * change in link status has been detected, then we read the PHY registers 1368 * change in link status has been detected, then we read the PHY registers
1369 * to get the current speed/duplex if link exists. 1369 * to get the current speed/duplex if link exists.
1370 *
1371 * Returns a negative error code (-E1000_ERR_*) or 0 (link down) or 1 (link
1372 * up).
1370 **/ 1373 **/
1371static s32 e1000_check_for_copper_link_ich8lan(struct e1000_hw *hw) 1374static s32 e1000_check_for_copper_link_ich8lan(struct e1000_hw *hw)
1372{ 1375{
@@ -1382,7 +1385,7 @@ static s32 e1000_check_for_copper_link_ich8lan(struct e1000_hw *hw)
1382 * Change or Rx Sequence Error interrupt. 1385 * Change or Rx Sequence Error interrupt.
1383 */ 1386 */
1384 if (!mac->get_link_status) 1387 if (!mac->get_link_status)
1385 return 0; 1388 return 1;
1386 1389
1387 /* First we want to see if the MII Status Register reports 1390 /* First we want to see if the MII Status Register reports
1388 * link. If so, then we want to get the current speed/duplex 1391 * link. If so, then we want to get the current speed/duplex
@@ -1613,10 +1616,12 @@ static s32 e1000_check_for_copper_link_ich8lan(struct e1000_hw *hw)
1613 * different link partner. 1616 * different link partner.
1614 */ 1617 */
1615 ret_val = e1000e_config_fc_after_link_up(hw); 1618 ret_val = e1000e_config_fc_after_link_up(hw);
1616 if (ret_val) 1619 if (ret_val) {
1617 e_dbg("Error configuring flow control\n"); 1620 e_dbg("Error configuring flow control\n");
1621 return ret_val;
1622 }
1618 1623
1619 return ret_val; 1624 return 1;
1620} 1625}
1621 1626
1622static s32 e1000_get_variants_ich8lan(struct e1000_adapter *adapter) 1627static s32 e1000_get_variants_ich8lan(struct e1000_adapter *adapter)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 321d8be80871..42dcaefc4c19 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -1573,11 +1573,18 @@ static int i40e_set_mac(struct net_device *netdev, void *p)
1573 else 1573 else
1574 netdev_info(netdev, "set new mac address %pM\n", addr->sa_data); 1574 netdev_info(netdev, "set new mac address %pM\n", addr->sa_data);
1575 1575
1576 /* Copy the address first, so that we avoid a possible race with
1577 * .set_rx_mode(). If we copy after changing the address in the filter
1578 * list, we might open ourselves to a narrow race window where
1579 * .set_rx_mode could delete our dev_addr filter and prevent traffic
1580 * from passing.
1581 */
1582 ether_addr_copy(netdev->dev_addr, addr->sa_data);
1583
1576 spin_lock_bh(&vsi->mac_filter_hash_lock); 1584 spin_lock_bh(&vsi->mac_filter_hash_lock);
1577 i40e_del_mac_filter(vsi, netdev->dev_addr); 1585 i40e_del_mac_filter(vsi, netdev->dev_addr);
1578 i40e_add_mac_filter(vsi, addr->sa_data); 1586 i40e_add_mac_filter(vsi, addr->sa_data);
1579 spin_unlock_bh(&vsi->mac_filter_hash_lock); 1587 spin_unlock_bh(&vsi->mac_filter_hash_lock);
1580 ether_addr_copy(netdev->dev_addr, addr->sa_data);
1581 if (vsi->type == I40E_VSI_MAIN) { 1588 if (vsi->type == I40E_VSI_MAIN) {
1582 i40e_status ret; 1589 i40e_status ret;
1583 1590
@@ -1923,6 +1930,14 @@ static int i40e_addr_unsync(struct net_device *netdev, const u8 *addr)
1923 struct i40e_netdev_priv *np = netdev_priv(netdev); 1930 struct i40e_netdev_priv *np = netdev_priv(netdev);
1924 struct i40e_vsi *vsi = np->vsi; 1931 struct i40e_vsi *vsi = np->vsi;
1925 1932
1933 /* Under some circumstances, we might receive a request to delete
1934 * our own device address from our uc list. Because we store the
1935 * device address in the VSI's MAC/VLAN filter list, we need to ignore
1936 * such requests and not delete our device address from this list.
1937 */
1938 if (ether_addr_equal(addr, netdev->dev_addr))
1939 return 0;
1940
1926 i40e_del_mac_filter(vsi, addr); 1941 i40e_del_mac_filter(vsi, addr);
1927 1942
1928 return 0; 1943 return 0;
@@ -6038,8 +6053,8 @@ static int i40e_validate_and_set_switch_mode(struct i40e_vsi *vsi)
6038 /* Set Bit 7 to be valid */ 6053 /* Set Bit 7 to be valid */
6039 mode = I40E_AQ_SET_SWITCH_BIT7_VALID; 6054 mode = I40E_AQ_SET_SWITCH_BIT7_VALID;
6040 6055
6041 /* Set L4type to both TCP and UDP support */ 6056 /* Set L4type for TCP support */
6042 mode |= I40E_AQ_SET_SWITCH_L4_TYPE_BOTH; 6057 mode |= I40E_AQ_SET_SWITCH_L4_TYPE_TCP;
6043 6058
6044 /* Set cloud filter mode */ 6059 /* Set cloud filter mode */
6045 mode |= I40E_AQ_SET_SWITCH_MODE_NON_TUNNEL; 6060 mode |= I40E_AQ_SET_SWITCH_MODE_NON_TUNNEL;
@@ -6969,18 +6984,18 @@ static int i40e_add_del_cloud_filter_big_buf(struct i40e_vsi *vsi,
6969 is_valid_ether_addr(filter->src_mac)) || 6984 is_valid_ether_addr(filter->src_mac)) ||
6970 (is_multicast_ether_addr(filter->dst_mac) && 6985 (is_multicast_ether_addr(filter->dst_mac) &&
6971 is_multicast_ether_addr(filter->src_mac))) 6986 is_multicast_ether_addr(filter->src_mac)))
6972 return -EINVAL; 6987 return -EOPNOTSUPP;
6973 6988
6974 /* Make sure port is specified, otherwise bail out, for channel 6989 /* Big buffer cloud filter needs 'L4 port' to be non-zero. Also, UDP
6975 * specific cloud filter needs 'L4 port' to be non-zero 6990 * ports are not supported via big buffer now.
6976 */ 6991 */
6977 if (!filter->dst_port) 6992 if (!filter->dst_port || filter->ip_proto == IPPROTO_UDP)
6978 return -EINVAL; 6993 return -EOPNOTSUPP;
6979 6994
6980 /* adding filter using src_port/src_ip is not supported at this stage */ 6995 /* adding filter using src_port/src_ip is not supported at this stage */
6981 if (filter->src_port || filter->src_ipv4 || 6996 if (filter->src_port || filter->src_ipv4 ||
6982 !ipv6_addr_any(&filter->ip.v6.src_ip6)) 6997 !ipv6_addr_any(&filter->ip.v6.src_ip6))
6983 return -EINVAL; 6998 return -EOPNOTSUPP;
6984 6999
6985 /* copy element needed to add cloud filter from filter */ 7000 /* copy element needed to add cloud filter from filter */
6986 i40e_set_cld_element(filter, &cld_filter.element); 7001 i40e_set_cld_element(filter, &cld_filter.element);
@@ -6991,7 +7006,7 @@ static int i40e_add_del_cloud_filter_big_buf(struct i40e_vsi *vsi,
6991 is_multicast_ether_addr(filter->src_mac)) { 7006 is_multicast_ether_addr(filter->src_mac)) {
6992 /* MAC + IP : unsupported mode */ 7007 /* MAC + IP : unsupported mode */
6993 if (filter->dst_ipv4) 7008 if (filter->dst_ipv4)
6994 return -EINVAL; 7009 return -EOPNOTSUPP;
6995 7010
6996 /* since we validated that L4 port must be valid before 7011 /* since we validated that L4 port must be valid before
6997 * we get here, start with respective "flags" value 7012 * we get here, start with respective "flags" value
@@ -7356,7 +7371,7 @@ static int i40e_configure_clsflower(struct i40e_vsi *vsi,
7356 7371
7357 if (tc < 0) { 7372 if (tc < 0) {
7358 dev_err(&vsi->back->pdev->dev, "Invalid traffic class\n"); 7373 dev_err(&vsi->back->pdev->dev, "Invalid traffic class\n");
7359 return -EINVAL; 7374 return -EOPNOTSUPP;
7360 } 7375 }
7361 7376
7362 if (test_bit(__I40E_RESET_RECOVERY_PENDING, pf->state) || 7377 if (test_bit(__I40E_RESET_RECOVERY_PENDING, pf->state) ||
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
index 4566d66ffc7c..5bc2748ac468 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
@@ -3047,10 +3047,30 @@ bool __i40e_chk_linearize(struct sk_buff *skb)
3047 /* Walk through fragments adding latest fragment, testing it, and 3047 /* Walk through fragments adding latest fragment, testing it, and
3048 * then removing stale fragments from the sum. 3048 * then removing stale fragments from the sum.
3049 */ 3049 */
3050 stale = &skb_shinfo(skb)->frags[0]; 3050 for (stale = &skb_shinfo(skb)->frags[0];; stale++) {
3051 for (;;) { 3051 int stale_size = skb_frag_size(stale);
3052
3052 sum += skb_frag_size(frag++); 3053 sum += skb_frag_size(frag++);
3053 3054
3055 /* The stale fragment may present us with a smaller
3056 * descriptor than the actual fragment size. To account
3057 * for that we need to remove all the data on the front and
3058 * figure out what the remainder would be in the last
3059 * descriptor associated with the fragment.
3060 */
3061 if (stale_size > I40E_MAX_DATA_PER_TXD) {
3062 int align_pad = -(stale->page_offset) &
3063 (I40E_MAX_READ_REQ_SIZE - 1);
3064
3065 sum -= align_pad;
3066 stale_size -= align_pad;
3067
3068 do {
3069 sum -= I40E_MAX_DATA_PER_TXD_ALIGNED;
3070 stale_size -= I40E_MAX_DATA_PER_TXD_ALIGNED;
3071 } while (stale_size > I40E_MAX_DATA_PER_TXD);
3072 }
3073
3054 /* if sum is negative we failed to make sufficient progress */ 3074 /* if sum is negative we failed to make sufficient progress */
3055 if (sum < 0) 3075 if (sum < 0)
3056 return true; 3076 return true;
@@ -3058,7 +3078,7 @@ bool __i40e_chk_linearize(struct sk_buff *skb)
3058 if (!nr_frags--) 3078 if (!nr_frags--)
3059 break; 3079 break;
3060 3080
3061 sum -= skb_frag_size(stale++); 3081 sum -= stale_size;
3062 } 3082 }
3063 3083
3064 return false; 3084 return false;
diff --git a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
index 50864f99446d..1ba29bb85b67 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
@@ -2012,10 +2012,30 @@ bool __i40evf_chk_linearize(struct sk_buff *skb)
2012 /* Walk through fragments adding latest fragment, testing it, and 2012 /* Walk through fragments adding latest fragment, testing it, and
2013 * then removing stale fragments from the sum. 2013 * then removing stale fragments from the sum.
2014 */ 2014 */
2015 stale = &skb_shinfo(skb)->frags[0]; 2015 for (stale = &skb_shinfo(skb)->frags[0];; stale++) {
2016 for (;;) { 2016 int stale_size = skb_frag_size(stale);
2017
2017 sum += skb_frag_size(frag++); 2018 sum += skb_frag_size(frag++);
2018 2019
2020 /* The stale fragment may present us with a smaller
2021 * descriptor than the actual fragment size. To account
2022 * for that we need to remove all the data on the front and
2023 * figure out what the remainder would be in the last
2024 * descriptor associated with the fragment.
2025 */
2026 if (stale_size > I40E_MAX_DATA_PER_TXD) {
2027 int align_pad = -(stale->page_offset) &
2028 (I40E_MAX_READ_REQ_SIZE - 1);
2029
2030 sum -= align_pad;
2031 stale_size -= align_pad;
2032
2033 do {
2034 sum -= I40E_MAX_DATA_PER_TXD_ALIGNED;
2035 stale_size -= I40E_MAX_DATA_PER_TXD_ALIGNED;
2036 } while (stale_size > I40E_MAX_DATA_PER_TXD);
2037 }
2038
2019 /* if sum is negative we failed to make sufficient progress */ 2039 /* if sum is negative we failed to make sufficient progress */
2020 if (sum < 0) 2040 if (sum < 0)
2021 return true; 2041 return true;
@@ -2023,7 +2043,7 @@ bool __i40evf_chk_linearize(struct sk_buff *skb)
2023 if (!nr_frags--) 2043 if (!nr_frags--)
2024 break; 2044 break;
2025 2045
2026 sum -= skb_frag_size(stale++); 2046 sum -= stale_size;
2027 } 2047 }
2028 2048
2029 return false; 2049 return false;
diff --git a/drivers/net/ethernet/mellanox/mlxsw/pci.c b/drivers/net/ethernet/mellanox/mlxsw/pci.c
index 23f7d828cf67..6ef20e5cc77d 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/pci.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/pci.c
@@ -1643,7 +1643,12 @@ static int mlxsw_pci_sw_reset(struct mlxsw_pci *mlxsw_pci,
1643 return 0; 1643 return 0;
1644 } 1644 }
1645 1645
1646 wmb(); /* reset needs to be written before we read control register */ 1646 /* Reset needs to be written before we read control register, and
1647 * we must wait for the HW to become responsive once again
1648 */
1649 wmb();
1650 msleep(MLXSW_PCI_SW_RESET_WAIT_MSECS);
1651
1647 end = jiffies + msecs_to_jiffies(MLXSW_PCI_SW_RESET_TIMEOUT_MSECS); 1652 end = jiffies + msecs_to_jiffies(MLXSW_PCI_SW_RESET_TIMEOUT_MSECS);
1648 do { 1653 do {
1649 u32 val = mlxsw_pci_read32(mlxsw_pci, FW_READY); 1654 u32 val = mlxsw_pci_read32(mlxsw_pci, FW_READY);
diff --git a/drivers/net/ethernet/mellanox/mlxsw/pci_hw.h b/drivers/net/ethernet/mellanox/mlxsw/pci_hw.h
index a6441208e9d9..fb082ad21b00 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/pci_hw.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/pci_hw.h
@@ -59,6 +59,7 @@
59#define MLXSW_PCI_SW_RESET 0xF0010 59#define MLXSW_PCI_SW_RESET 0xF0010
60#define MLXSW_PCI_SW_RESET_RST_BIT BIT(0) 60#define MLXSW_PCI_SW_RESET_RST_BIT BIT(0)
61#define MLXSW_PCI_SW_RESET_TIMEOUT_MSECS 5000 61#define MLXSW_PCI_SW_RESET_TIMEOUT_MSECS 5000
62#define MLXSW_PCI_SW_RESET_WAIT_MSECS 100
62#define MLXSW_PCI_FW_READY 0xA1844 63#define MLXSW_PCI_FW_READY 0xA1844
63#define MLXSW_PCI_FW_READY_MASK 0xFFFF 64#define MLXSW_PCI_FW_READY_MASK 0xFFFF
64#define MLXSW_PCI_FW_READY_MAGIC 0x5E 65#define MLXSW_PCI_FW_READY_MAGIC 0x5E
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
index 9bd8d28de152..c3837ca7a705 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
@@ -4376,7 +4376,10 @@ static int mlxsw_sp_netdevice_port_upper_event(struct net_device *lower_dev,
4376 } 4376 }
4377 if (!info->linking) 4377 if (!info->linking)
4378 break; 4378 break;
4379 if (netdev_has_any_upper_dev(upper_dev)) { 4379 if (netdev_has_any_upper_dev(upper_dev) &&
4380 (!netif_is_bridge_master(upper_dev) ||
4381 !mlxsw_sp_bridge_device_is_offloaded(mlxsw_sp,
4382 upper_dev))) {
4380 NL_SET_ERR_MSG(extack, 4383 NL_SET_ERR_MSG(extack,
4381 "spectrum: Enslaving a port to a device that already has an upper device is not supported"); 4384 "spectrum: Enslaving a port to a device that already has an upper device is not supported");
4382 return -EINVAL; 4385 return -EINVAL;
@@ -4504,6 +4507,7 @@ static int mlxsw_sp_netdevice_port_vlan_event(struct net_device *vlan_dev,
4504 u16 vid) 4507 u16 vid)
4505{ 4508{
4506 struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev); 4509 struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
4510 struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
4507 struct netdev_notifier_changeupper_info *info = ptr; 4511 struct netdev_notifier_changeupper_info *info = ptr;
4508 struct netlink_ext_ack *extack; 4512 struct netlink_ext_ack *extack;
4509 struct net_device *upper_dev; 4513 struct net_device *upper_dev;
@@ -4520,7 +4524,10 @@ static int mlxsw_sp_netdevice_port_vlan_event(struct net_device *vlan_dev,
4520 } 4524 }
4521 if (!info->linking) 4525 if (!info->linking)
4522 break; 4526 break;
4523 if (netdev_has_any_upper_dev(upper_dev)) { 4527 if (netdev_has_any_upper_dev(upper_dev) &&
4528 (!netif_is_bridge_master(upper_dev) ||
4529 !mlxsw_sp_bridge_device_is_offloaded(mlxsw_sp,
4530 upper_dev))) {
4524 NL_SET_ERR_MSG(extack, "spectrum: Enslaving a port to a device that already has an upper device is not supported"); 4531 NL_SET_ERR_MSG(extack, "spectrum: Enslaving a port to a device that already has an upper device is not supported");
4525 return -EINVAL; 4532 return -EINVAL;
4526 } 4533 }
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
index 432ab9b12b7f..05ce1befd9b3 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
@@ -365,6 +365,8 @@ int mlxsw_sp_port_bridge_join(struct mlxsw_sp_port *mlxsw_sp_port,
365void mlxsw_sp_port_bridge_leave(struct mlxsw_sp_port *mlxsw_sp_port, 365void mlxsw_sp_port_bridge_leave(struct mlxsw_sp_port *mlxsw_sp_port,
366 struct net_device *brport_dev, 366 struct net_device *brport_dev,
367 struct net_device *br_dev); 367 struct net_device *br_dev);
368bool mlxsw_sp_bridge_device_is_offloaded(const struct mlxsw_sp *mlxsw_sp,
369 const struct net_device *br_dev);
368 370
369/* spectrum.c */ 371/* spectrum.c */
370int mlxsw_sp_port_ets_set(struct mlxsw_sp_port *mlxsw_sp_port, 372int mlxsw_sp_port_ets_set(struct mlxsw_sp_port *mlxsw_sp_port,
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_qdisc.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_qdisc.c
index c33beac5def0..b5397da94d7f 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_qdisc.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_qdisc.c
@@ -46,7 +46,8 @@ mlxsw_sp_tclass_congestion_enable(struct mlxsw_sp_port *mlxsw_sp_port,
46 int tclass_num, u32 min, u32 max, 46 int tclass_num, u32 min, u32 max,
47 u32 probability, bool is_ecn) 47 u32 probability, bool is_ecn)
48{ 48{
49 char cwtp_cmd[max_t(u8, MLXSW_REG_CWTP_LEN, MLXSW_REG_CWTPM_LEN)]; 49 char cwtpm_cmd[MLXSW_REG_CWTPM_LEN];
50 char cwtp_cmd[MLXSW_REG_CWTP_LEN];
50 struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp; 51 struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
51 int err; 52 int err;
52 53
@@ -60,10 +61,10 @@ mlxsw_sp_tclass_congestion_enable(struct mlxsw_sp_port *mlxsw_sp_port,
60 if (err) 61 if (err)
61 return err; 62 return err;
62 63
63 mlxsw_reg_cwtpm_pack(cwtp_cmd, mlxsw_sp_port->local_port, tclass_num, 64 mlxsw_reg_cwtpm_pack(cwtpm_cmd, mlxsw_sp_port->local_port, tclass_num,
64 MLXSW_REG_CWTP_DEFAULT_PROFILE, true, is_ecn); 65 MLXSW_REG_CWTP_DEFAULT_PROFILE, true, is_ecn);
65 66
66 return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(cwtpm), cwtp_cmd); 67 return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(cwtpm), cwtpm_cmd);
67} 68}
68 69
69static int 70static int
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index be657b8533f0..434b3922b34f 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -3228,7 +3228,7 @@ static void __mlxsw_sp_nexthop_neigh_update(struct mlxsw_sp_nexthop *nh,
3228{ 3228{
3229 if (!removing) 3229 if (!removing)
3230 nh->should_offload = 1; 3230 nh->should_offload = 1;
3231 else if (nh->offloaded) 3231 else
3232 nh->should_offload = 0; 3232 nh->should_offload = 0;
3233 nh->update = 1; 3233 nh->update = 1;
3234} 3234}
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
index 7b8548e25ae7..593ad31be749 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
@@ -152,6 +152,12 @@ mlxsw_sp_bridge_device_find(const struct mlxsw_sp_bridge *bridge,
152 return NULL; 152 return NULL;
153} 153}
154 154
155bool mlxsw_sp_bridge_device_is_offloaded(const struct mlxsw_sp *mlxsw_sp,
156 const struct net_device *br_dev)
157{
158 return !!mlxsw_sp_bridge_device_find(mlxsw_sp->bridge, br_dev);
159}
160
155static struct mlxsw_sp_bridge_device * 161static struct mlxsw_sp_bridge_device *
156mlxsw_sp_bridge_device_create(struct mlxsw_sp_bridge *bridge, 162mlxsw_sp_bridge_device_create(struct mlxsw_sp_bridge *bridge,
157 struct net_device *br_dev) 163 struct net_device *br_dev)
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
index 1a603fdd9e80..99b0487b6d82 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
@@ -568,6 +568,7 @@ nfp_net_aux_irq_request(struct nfp_net *nn, u32 ctrl_offset,
568 return err; 568 return err;
569 } 569 }
570 nn_writeb(nn, ctrl_offset, entry->entry); 570 nn_writeb(nn, ctrl_offset, entry->entry);
571 nfp_net_irq_unmask(nn, entry->entry);
571 572
572 return 0; 573 return 0;
573} 574}
@@ -582,6 +583,7 @@ static void nfp_net_aux_irq_free(struct nfp_net *nn, u32 ctrl_offset,
582 unsigned int vector_idx) 583 unsigned int vector_idx)
583{ 584{
584 nn_writeb(nn, ctrl_offset, 0xff); 585 nn_writeb(nn, ctrl_offset, 0xff);
586 nn_pci_flush(nn);
585 free_irq(nn->irq_entries[vector_idx].vector, nn); 587 free_irq(nn->irq_entries[vector_idx].vector, nn);
586} 588}
587 589
diff --git a/drivers/net/ethernet/renesas/sh_eth.c b/drivers/net/ethernet/renesas/sh_eth.c
index 75323000c364..b9e2846589f8 100644
--- a/drivers/net/ethernet/renesas/sh_eth.c
+++ b/drivers/net/ethernet/renesas/sh_eth.c
@@ -147,7 +147,7 @@ static const u16 sh_eth_offset_gigabit[SH_ETH_MAX_REGISTER_OFFSET] = {
147 [FWNLCR0] = 0x0090, 147 [FWNLCR0] = 0x0090,
148 [FWALCR0] = 0x0094, 148 [FWALCR0] = 0x0094,
149 [TXNLCR1] = 0x00a0, 149 [TXNLCR1] = 0x00a0,
150 [TXALCR1] = 0x00a0, 150 [TXALCR1] = 0x00a4,
151 [RXNLCR1] = 0x00a8, 151 [RXNLCR1] = 0x00a8,
152 [RXALCR1] = 0x00ac, 152 [RXALCR1] = 0x00ac,
153 [FWNLCR1] = 0x00b0, 153 [FWNLCR1] = 0x00b0,
@@ -399,7 +399,7 @@ static const u16 sh_eth_offset_fast_sh3_sh2[SH_ETH_MAX_REGISTER_OFFSET] = {
399 [FWNLCR0] = 0x0090, 399 [FWNLCR0] = 0x0090,
400 [FWALCR0] = 0x0094, 400 [FWALCR0] = 0x0094,
401 [TXNLCR1] = 0x00a0, 401 [TXNLCR1] = 0x00a0,
402 [TXALCR1] = 0x00a0, 402 [TXALCR1] = 0x00a4,
403 [RXNLCR1] = 0x00a8, 403 [RXNLCR1] = 0x00a8,
404 [RXALCR1] = 0x00ac, 404 [RXALCR1] = 0x00ac,
405 [FWNLCR1] = 0x00b0, 405 [FWNLCR1] = 0x00b0,
@@ -3225,18 +3225,37 @@ static int sh_eth_drv_probe(struct platform_device *pdev)
3225 /* ioremap the TSU registers */ 3225 /* ioremap the TSU registers */
3226 if (mdp->cd->tsu) { 3226 if (mdp->cd->tsu) {
3227 struct resource *rtsu; 3227 struct resource *rtsu;
3228
3228 rtsu = platform_get_resource(pdev, IORESOURCE_MEM, 1); 3229 rtsu = platform_get_resource(pdev, IORESOURCE_MEM, 1);
3229 mdp->tsu_addr = devm_ioremap_resource(&pdev->dev, rtsu); 3230 if (!rtsu) {
3230 if (IS_ERR(mdp->tsu_addr)) { 3231 dev_err(&pdev->dev, "no TSU resource\n");
3231 ret = PTR_ERR(mdp->tsu_addr); 3232 ret = -ENODEV;
3233 goto out_release;
3234 }
3235 /* We can only request the TSU region for the first port
3236 * of the two sharing this TSU for the probe to succeed...
3237 */
3238 if (devno % 2 == 0 &&
3239 !devm_request_mem_region(&pdev->dev, rtsu->start,
3240 resource_size(rtsu),
3241 dev_name(&pdev->dev))) {
3242 dev_err(&pdev->dev, "can't request TSU resource.\n");
3243 ret = -EBUSY;
3244 goto out_release;
3245 }
3246 mdp->tsu_addr = devm_ioremap(&pdev->dev, rtsu->start,
3247 resource_size(rtsu));
3248 if (!mdp->tsu_addr) {
3249 dev_err(&pdev->dev, "TSU region ioremap() failed.\n");
3250 ret = -ENOMEM;
3232 goto out_release; 3251 goto out_release;
3233 } 3252 }
3234 mdp->port = devno % 2; 3253 mdp->port = devno % 2;
3235 ndev->features = NETIF_F_HW_VLAN_CTAG_FILTER; 3254 ndev->features = NETIF_F_HW_VLAN_CTAG_FILTER;
3236 } 3255 }
3237 3256
3238 /* initialize first or needed device */ 3257 /* Need to init only the first port of the two sharing a TSU */
3239 if (!devno || pd->needs_init) { 3258 if (devno % 2 == 0) {
3240 if (mdp->cd->chip_reset) 3259 if (mdp->cd->chip_reset)
3241 mdp->cd->chip_reset(ndev); 3260 mdp->cd->chip_reset(ndev);
3242 3261
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
index 337d53d12e94..c0af0bc4e714 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
@@ -364,9 +364,15 @@ static void stmmac_eee_ctrl_timer(struct timer_list *t)
364bool stmmac_eee_init(struct stmmac_priv *priv) 364bool stmmac_eee_init(struct stmmac_priv *priv)
365{ 365{
366 struct net_device *ndev = priv->dev; 366 struct net_device *ndev = priv->dev;
367 int interface = priv->plat->interface;
367 unsigned long flags; 368 unsigned long flags;
368 bool ret = false; 369 bool ret = false;
369 370
371 if ((interface != PHY_INTERFACE_MODE_MII) &&
372 (interface != PHY_INTERFACE_MODE_GMII) &&
373 !phy_interface_mode_is_rgmii(interface))
374 goto out;
375
370 /* Using PCS we cannot dial with the phy registers at this stage 376 /* Using PCS we cannot dial with the phy registers at this stage
371 * so we do not support extra feature like EEE. 377 * so we do not support extra feature like EEE.
372 */ 378 */
diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
index b718a02a6bb6..0a48b3073d3d 100644
--- a/drivers/net/geneve.c
+++ b/drivers/net/geneve.c
@@ -825,6 +825,13 @@ static int geneve_xmit_skb(struct sk_buff *skb, struct net_device *dev,
825 if (IS_ERR(rt)) 825 if (IS_ERR(rt))
826 return PTR_ERR(rt); 826 return PTR_ERR(rt);
827 827
828 if (skb_dst(skb)) {
829 int mtu = dst_mtu(&rt->dst) - sizeof(struct iphdr) -
830 GENEVE_BASE_HLEN - info->options_len - 14;
831
832 skb_dst(skb)->ops->update_pmtu(skb_dst(skb), NULL, skb, mtu);
833 }
834
828 sport = udp_flow_src_port(geneve->net, skb, 1, USHRT_MAX, true); 835 sport = udp_flow_src_port(geneve->net, skb, 1, USHRT_MAX, true);
829 if (geneve->collect_md) { 836 if (geneve->collect_md) {
830 tos = ip_tunnel_ecn_encap(key->tos, ip_hdr(skb), skb); 837 tos = ip_tunnel_ecn_encap(key->tos, ip_hdr(skb), skb);
@@ -864,6 +871,13 @@ static int geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
864 if (IS_ERR(dst)) 871 if (IS_ERR(dst))
865 return PTR_ERR(dst); 872 return PTR_ERR(dst);
866 873
874 if (skb_dst(skb)) {
875 int mtu = dst_mtu(dst) - sizeof(struct ipv6hdr) -
876 GENEVE_BASE_HLEN - info->options_len - 14;
877
878 skb_dst(skb)->ops->update_pmtu(skb_dst(skb), NULL, skb, mtu);
879 }
880
867 sport = udp_flow_src_port(geneve->net, skb, 1, USHRT_MAX, true); 881 sport = udp_flow_src_port(geneve->net, skb, 1, USHRT_MAX, true);
868 if (geneve->collect_md) { 882 if (geneve->collect_md) {
869 prio = ip_tunnel_ecn_encap(key->tos, ip_hdr(skb), skb); 883 prio = ip_tunnel_ecn_encap(key->tos, ip_hdr(skb), skb);
diff --git a/drivers/net/macvlan.c b/drivers/net/macvlan.c
index a178c5efd33e..a0f2be81d52e 100644
--- a/drivers/net/macvlan.c
+++ b/drivers/net/macvlan.c
@@ -1444,9 +1444,14 @@ int macvlan_common_newlink(struct net *src_net, struct net_device *dev,
1444 return 0; 1444 return 0;
1445 1445
1446unregister_netdev: 1446unregister_netdev:
1447 /* macvlan_uninit would free the macvlan port */
1447 unregister_netdevice(dev); 1448 unregister_netdevice(dev);
1449 return err;
1448destroy_macvlan_port: 1450destroy_macvlan_port:
1449 if (create) 1451 /* the macvlan port may be freed by macvlan_uninit when fail to register.
1452 * so we destroy the macvlan port only when it's valid.
1453 */
1454 if (create && macvlan_port_get_rtnl(dev))
1450 macvlan_port_destroy(port->dev); 1455 macvlan_port_destroy(port->dev);
1451 return err; 1456 return err;
1452} 1457}
diff --git a/drivers/net/phy/mdio-sun4i.c b/drivers/net/phy/mdio-sun4i.c
index 135296508a7e..6425ce04d3f9 100644
--- a/drivers/net/phy/mdio-sun4i.c
+++ b/drivers/net/phy/mdio-sun4i.c
@@ -118,8 +118,10 @@ static int sun4i_mdio_probe(struct platform_device *pdev)
118 118
119 data->regulator = devm_regulator_get(&pdev->dev, "phy"); 119 data->regulator = devm_regulator_get(&pdev->dev, "phy");
120 if (IS_ERR(data->regulator)) { 120 if (IS_ERR(data->regulator)) {
121 if (PTR_ERR(data->regulator) == -EPROBE_DEFER) 121 if (PTR_ERR(data->regulator) == -EPROBE_DEFER) {
122 return -EPROBE_DEFER; 122 ret = -EPROBE_DEFER;
123 goto err_out_free_mdiobus;
124 }
123 125
124 dev_info(&pdev->dev, "no regulator found\n"); 126 dev_info(&pdev->dev, "no regulator found\n");
125 data->regulator = NULL; 127 data->regulator = NULL;
diff --git a/drivers/net/phy/phylink.c b/drivers/net/phy/phylink.c
index 827f3f92560e..249ce5cbea22 100644
--- a/drivers/net/phy/phylink.c
+++ b/drivers/net/phy/phylink.c
@@ -1296,6 +1296,7 @@ int phylink_mii_ioctl(struct phylink *pl, struct ifreq *ifr, int cmd)
1296 switch (cmd) { 1296 switch (cmd) {
1297 case SIOCGMIIPHY: 1297 case SIOCGMIIPHY:
1298 mii->phy_id = pl->phydev->mdio.addr; 1298 mii->phy_id = pl->phydev->mdio.addr;
1299 /* fall through */
1299 1300
1300 case SIOCGMIIREG: 1301 case SIOCGMIIREG:
1301 ret = phylink_phy_read(pl, mii->phy_id, mii->reg_num); 1302 ret = phylink_phy_read(pl, mii->phy_id, mii->reg_num);
@@ -1318,6 +1319,7 @@ int phylink_mii_ioctl(struct phylink *pl, struct ifreq *ifr, int cmd)
1318 switch (cmd) { 1319 switch (cmd) {
1319 case SIOCGMIIPHY: 1320 case SIOCGMIIPHY:
1320 mii->phy_id = 0; 1321 mii->phy_id = 0;
1322 /* fall through */
1321 1323
1322 case SIOCGMIIREG: 1324 case SIOCGMIIREG:
1323 ret = phylink_mii_read(pl, mii->phy_id, mii->reg_num); 1325 ret = phylink_mii_read(pl, mii->phy_id, mii->reg_num);
@@ -1429,9 +1431,8 @@ static void phylink_sfp_link_down(void *upstream)
1429 WARN_ON(!lockdep_rtnl_is_held()); 1431 WARN_ON(!lockdep_rtnl_is_held());
1430 1432
1431 set_bit(PHYLINK_DISABLE_LINK, &pl->phylink_disable_state); 1433 set_bit(PHYLINK_DISABLE_LINK, &pl->phylink_disable_state);
1434 queue_work(system_power_efficient_wq, &pl->resolve);
1432 flush_work(&pl->resolve); 1435 flush_work(&pl->resolve);
1433
1434 netif_carrier_off(pl->netdev);
1435} 1436}
1436 1437
1437static void phylink_sfp_link_up(void *upstream) 1438static void phylink_sfp_link_up(void *upstream)
diff --git a/drivers/net/phy/sfp-bus.c b/drivers/net/phy/sfp-bus.c
index 8a1b1f4c1b7c..ab64a142b832 100644
--- a/drivers/net/phy/sfp-bus.c
+++ b/drivers/net/phy/sfp-bus.c
@@ -356,7 +356,8 @@ EXPORT_SYMBOL_GPL(sfp_register_upstream);
356void sfp_unregister_upstream(struct sfp_bus *bus) 356void sfp_unregister_upstream(struct sfp_bus *bus)
357{ 357{
358 rtnl_lock(); 358 rtnl_lock();
359 sfp_unregister_bus(bus); 359 if (bus->sfp)
360 sfp_unregister_bus(bus);
360 bus->upstream = NULL; 361 bus->upstream = NULL;
361 bus->netdev = NULL; 362 bus->netdev = NULL;
362 rtnl_unlock(); 363 rtnl_unlock();
@@ -459,7 +460,8 @@ EXPORT_SYMBOL_GPL(sfp_register_socket);
459void sfp_unregister_socket(struct sfp_bus *bus) 460void sfp_unregister_socket(struct sfp_bus *bus)
460{ 461{
461 rtnl_lock(); 462 rtnl_lock();
462 sfp_unregister_bus(bus); 463 if (bus->netdev)
464 sfp_unregister_bus(bus);
463 bus->sfp_dev = NULL; 465 bus->sfp_dev = NULL;
464 bus->sfp = NULL; 466 bus->sfp = NULL;
465 bus->socket_ops = NULL; 467 bus->socket_ops = NULL;
diff --git a/drivers/net/usb/qmi_wwan.c b/drivers/net/usb/qmi_wwan.c
index 3000ddd1c7e2..728819feab44 100644
--- a/drivers/net/usb/qmi_wwan.c
+++ b/drivers/net/usb/qmi_wwan.c
@@ -1100,6 +1100,7 @@ static const struct usb_device_id products[] = {
1100 {QMI_FIXED_INTF(0x05c6, 0x9084, 4)}, 1100 {QMI_FIXED_INTF(0x05c6, 0x9084, 4)},
1101 {QMI_FIXED_INTF(0x05c6, 0x920d, 0)}, 1101 {QMI_FIXED_INTF(0x05c6, 0x920d, 0)},
1102 {QMI_FIXED_INTF(0x05c6, 0x920d, 5)}, 1102 {QMI_FIXED_INTF(0x05c6, 0x920d, 5)},
1103 {QMI_QUIRK_SET_DTR(0x05c6, 0x9625, 4)}, /* YUGA CLM920-NC5 */
1103 {QMI_FIXED_INTF(0x0846, 0x68a2, 8)}, 1104 {QMI_FIXED_INTF(0x0846, 0x68a2, 8)},
1104 {QMI_FIXED_INTF(0x12d1, 0x140c, 1)}, /* Huawei E173 */ 1105 {QMI_FIXED_INTF(0x12d1, 0x140c, 1)}, /* Huawei E173 */
1105 {QMI_FIXED_INTF(0x12d1, 0x14ac, 1)}, /* Huawei E1820 */ 1106 {QMI_FIXED_INTF(0x12d1, 0x14ac, 1)}, /* Huawei E1820 */
diff --git a/drivers/net/wireless/ath/wcn36xx/main.c b/drivers/net/wireless/ath/wcn36xx/main.c
index f7d228b5ba93..987f1252a3cf 100644
--- a/drivers/net/wireless/ath/wcn36xx/main.c
+++ b/drivers/net/wireless/ath/wcn36xx/main.c
@@ -384,6 +384,18 @@ static int wcn36xx_config(struct ieee80211_hw *hw, u32 changed)
384 } 384 }
385 } 385 }
386 386
387 if (changed & IEEE80211_CONF_CHANGE_PS) {
388 list_for_each_entry(tmp, &wcn->vif_list, list) {
389 vif = wcn36xx_priv_to_vif(tmp);
390 if (hw->conf.flags & IEEE80211_CONF_PS) {
391 if (vif->bss_conf.ps) /* ps allowed ? */
392 wcn36xx_pmc_enter_bmps_state(wcn, vif);
393 } else {
394 wcn36xx_pmc_exit_bmps_state(wcn, vif);
395 }
396 }
397 }
398
387 mutex_unlock(&wcn->conf_mutex); 399 mutex_unlock(&wcn->conf_mutex);
388 400
389 return 0; 401 return 0;
@@ -747,17 +759,6 @@ static void wcn36xx_bss_info_changed(struct ieee80211_hw *hw,
747 vif_priv->dtim_period = bss_conf->dtim_period; 759 vif_priv->dtim_period = bss_conf->dtim_period;
748 } 760 }
749 761
750 if (changed & BSS_CHANGED_PS) {
751 wcn36xx_dbg(WCN36XX_DBG_MAC,
752 "mac bss PS set %d\n",
753 bss_conf->ps);
754 if (bss_conf->ps) {
755 wcn36xx_pmc_enter_bmps_state(wcn, vif);
756 } else {
757 wcn36xx_pmc_exit_bmps_state(wcn, vif);
758 }
759 }
760
761 if (changed & BSS_CHANGED_BSSID) { 762 if (changed & BSS_CHANGED_BSSID) {
762 wcn36xx_dbg(WCN36XX_DBG_MAC, "mac bss changed_bssid %pM\n", 763 wcn36xx_dbg(WCN36XX_DBG_MAC, "mac bss changed_bssid %pM\n",
763 bss_conf->bssid); 764 bss_conf->bssid);
diff --git a/drivers/net/wireless/ath/wcn36xx/pmc.c b/drivers/net/wireless/ath/wcn36xx/pmc.c
index 589fe5f70971..1976b80c235f 100644
--- a/drivers/net/wireless/ath/wcn36xx/pmc.c
+++ b/drivers/net/wireless/ath/wcn36xx/pmc.c
@@ -45,8 +45,10 @@ int wcn36xx_pmc_exit_bmps_state(struct wcn36xx *wcn,
45 struct wcn36xx_vif *vif_priv = wcn36xx_vif_to_priv(vif); 45 struct wcn36xx_vif *vif_priv = wcn36xx_vif_to_priv(vif);
46 46
47 if (WCN36XX_BMPS != vif_priv->pw_state) { 47 if (WCN36XX_BMPS != vif_priv->pw_state) {
48 wcn36xx_err("Not in BMPS mode, no need to exit from BMPS mode!\n"); 48 /* Unbalanced call or last BMPS enter failed */
49 return -EINVAL; 49 wcn36xx_dbg(WCN36XX_DBG_PMC,
50 "Not in BMPS mode, no need to exit\n");
51 return -EALREADY;
50 } 52 }
51 wcn36xx_smd_exit_bmps(wcn, vif); 53 wcn36xx_smd_exit_bmps(wcn, vif);
52 vif_priv->pw_state = WCN36XX_FULL_POWER; 54 vif_priv->pw_state = WCN36XX_FULL_POWER;
diff --git a/drivers/net/wireless/intel/iwlwifi/pcie/internal.h b/drivers/net/wireless/intel/iwlwifi/pcie/internal.h
index d749abeca3ae..403e65c309d0 100644
--- a/drivers/net/wireless/intel/iwlwifi/pcie/internal.h
+++ b/drivers/net/wireless/intel/iwlwifi/pcie/internal.h
@@ -670,11 +670,15 @@ static inline u8 iwl_pcie_get_cmd_index(struct iwl_txq *q, u32 index)
670 return index & (q->n_window - 1); 670 return index & (q->n_window - 1);
671} 671}
672 672
673static inline void *iwl_pcie_get_tfd(struct iwl_trans_pcie *trans_pcie, 673static inline void *iwl_pcie_get_tfd(struct iwl_trans *trans,
674 struct iwl_txq *txq, int idx) 674 struct iwl_txq *txq, int idx)
675{ 675{
676 return txq->tfds + trans_pcie->tfd_size * iwl_pcie_get_cmd_index(txq, 676 struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
677 idx); 677
678 if (trans->cfg->use_tfh)
679 idx = iwl_pcie_get_cmd_index(txq, idx);
680
681 return txq->tfds + trans_pcie->tfd_size * idx;
678} 682}
679 683
680static inline void iwl_enable_rfkill_int(struct iwl_trans *trans) 684static inline void iwl_enable_rfkill_int(struct iwl_trans *trans)
diff --git a/drivers/net/wireless/intel/iwlwifi/pcie/tx-gen2.c b/drivers/net/wireless/intel/iwlwifi/pcie/tx-gen2.c
index 16b345f54ff0..6d0a907d5ba5 100644
--- a/drivers/net/wireless/intel/iwlwifi/pcie/tx-gen2.c
+++ b/drivers/net/wireless/intel/iwlwifi/pcie/tx-gen2.c
@@ -171,8 +171,6 @@ static void iwl_pcie_gen2_tfd_unmap(struct iwl_trans *trans,
171 171
172static void iwl_pcie_gen2_free_tfd(struct iwl_trans *trans, struct iwl_txq *txq) 172static void iwl_pcie_gen2_free_tfd(struct iwl_trans *trans, struct iwl_txq *txq)
173{ 173{
174 struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
175
176 /* rd_ptr is bounded by TFD_QUEUE_SIZE_MAX and 174 /* rd_ptr is bounded by TFD_QUEUE_SIZE_MAX and
177 * idx is bounded by n_window 175 * idx is bounded by n_window
178 */ 176 */
@@ -181,7 +179,7 @@ static void iwl_pcie_gen2_free_tfd(struct iwl_trans *trans, struct iwl_txq *txq)
181 lockdep_assert_held(&txq->lock); 179 lockdep_assert_held(&txq->lock);
182 180
183 iwl_pcie_gen2_tfd_unmap(trans, &txq->entries[idx].meta, 181 iwl_pcie_gen2_tfd_unmap(trans, &txq->entries[idx].meta,
184 iwl_pcie_get_tfd(trans_pcie, txq, idx)); 182 iwl_pcie_get_tfd(trans, txq, idx));
185 183
186 /* free SKB */ 184 /* free SKB */
187 if (txq->entries) { 185 if (txq->entries) {
@@ -364,11 +362,9 @@ struct iwl_tfh_tfd *iwl_pcie_gen2_build_tfd(struct iwl_trans *trans,
364 struct sk_buff *skb, 362 struct sk_buff *skb,
365 struct iwl_cmd_meta *out_meta) 363 struct iwl_cmd_meta *out_meta)
366{ 364{
367 struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
368 struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)skb->data; 365 struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)skb->data;
369 int idx = iwl_pcie_get_cmd_index(txq, txq->write_ptr); 366 int idx = iwl_pcie_get_cmd_index(txq, txq->write_ptr);
370 struct iwl_tfh_tfd *tfd = 367 struct iwl_tfh_tfd *tfd = iwl_pcie_get_tfd(trans, txq, idx);
371 iwl_pcie_get_tfd(trans_pcie, txq, idx);
372 dma_addr_t tb_phys; 368 dma_addr_t tb_phys;
373 bool amsdu; 369 bool amsdu;
374 int i, len, tb1_len, tb2_len, hdr_len; 370 int i, len, tb1_len, tb2_len, hdr_len;
@@ -565,8 +561,7 @@ static int iwl_pcie_gen2_enqueue_hcmd(struct iwl_trans *trans,
565 u8 group_id = iwl_cmd_groupid(cmd->id); 561 u8 group_id = iwl_cmd_groupid(cmd->id);
566 const u8 *cmddata[IWL_MAX_CMD_TBS_PER_TFD]; 562 const u8 *cmddata[IWL_MAX_CMD_TBS_PER_TFD];
567 u16 cmdlen[IWL_MAX_CMD_TBS_PER_TFD]; 563 u16 cmdlen[IWL_MAX_CMD_TBS_PER_TFD];
568 struct iwl_tfh_tfd *tfd = 564 struct iwl_tfh_tfd *tfd = iwl_pcie_get_tfd(trans, txq, txq->write_ptr);
569 iwl_pcie_get_tfd(trans_pcie, txq, txq->write_ptr);
570 565
571 memset(tfd, 0, sizeof(*tfd)); 566 memset(tfd, 0, sizeof(*tfd));
572 567
diff --git a/drivers/net/wireless/intel/iwlwifi/pcie/tx.c b/drivers/net/wireless/intel/iwlwifi/pcie/tx.c
index fed6d842a5e1..3f85713c41dc 100644
--- a/drivers/net/wireless/intel/iwlwifi/pcie/tx.c
+++ b/drivers/net/wireless/intel/iwlwifi/pcie/tx.c
@@ -373,7 +373,7 @@ static void iwl_pcie_tfd_unmap(struct iwl_trans *trans,
373{ 373{
374 struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans); 374 struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
375 int i, num_tbs; 375 int i, num_tbs;
376 void *tfd = iwl_pcie_get_tfd(trans_pcie, txq, index); 376 void *tfd = iwl_pcie_get_tfd(trans, txq, index);
377 377
378 /* Sanity check on number of chunks */ 378 /* Sanity check on number of chunks */
379 num_tbs = iwl_pcie_tfd_get_num_tbs(trans, tfd); 379 num_tbs = iwl_pcie_tfd_get_num_tbs(trans, tfd);
@@ -2018,7 +2018,7 @@ static int iwl_fill_data_tbs(struct iwl_trans *trans, struct sk_buff *skb,
2018 } 2018 }
2019 2019
2020 trace_iwlwifi_dev_tx(trans->dev, skb, 2020 trace_iwlwifi_dev_tx(trans->dev, skb,
2021 iwl_pcie_get_tfd(trans_pcie, txq, txq->write_ptr), 2021 iwl_pcie_get_tfd(trans, txq, txq->write_ptr),
2022 trans_pcie->tfd_size, 2022 trans_pcie->tfd_size,
2023 &dev_cmd->hdr, IWL_FIRST_TB_SIZE + tb1_len, 2023 &dev_cmd->hdr, IWL_FIRST_TB_SIZE + tb1_len,
2024 hdr_len); 2024 hdr_len);
@@ -2092,7 +2092,7 @@ static int iwl_fill_data_tbs_amsdu(struct iwl_trans *trans, struct sk_buff *skb,
2092 IEEE80211_CCMP_HDR_LEN : 0; 2092 IEEE80211_CCMP_HDR_LEN : 0;
2093 2093
2094 trace_iwlwifi_dev_tx(trans->dev, skb, 2094 trace_iwlwifi_dev_tx(trans->dev, skb,
2095 iwl_pcie_get_tfd(trans_pcie, txq, txq->write_ptr), 2095 iwl_pcie_get_tfd(trans, txq, txq->write_ptr),
2096 trans_pcie->tfd_size, 2096 trans_pcie->tfd_size,
2097 &dev_cmd->hdr, IWL_FIRST_TB_SIZE + tb1_len, 0); 2097 &dev_cmd->hdr, IWL_FIRST_TB_SIZE + tb1_len, 0);
2098 2098
@@ -2425,7 +2425,7 @@ int iwl_trans_pcie_tx(struct iwl_trans *trans, struct sk_buff *skb,
2425 memcpy(&txq->first_tb_bufs[txq->write_ptr], &dev_cmd->hdr, 2425 memcpy(&txq->first_tb_bufs[txq->write_ptr], &dev_cmd->hdr,
2426 IWL_FIRST_TB_SIZE); 2426 IWL_FIRST_TB_SIZE);
2427 2427
2428 tfd = iwl_pcie_get_tfd(trans_pcie, txq, txq->write_ptr); 2428 tfd = iwl_pcie_get_tfd(trans, txq, txq->write_ptr);
2429 /* Set up entry for this TFD in Tx byte-count array */ 2429 /* Set up entry for this TFD in Tx byte-count array */
2430 iwl_pcie_txq_update_byte_cnt_tbl(trans, txq, le16_to_cpu(tx_cmd->len), 2430 iwl_pcie_txq_update_byte_cnt_tbl(trans, txq, le16_to_cpu(tx_cmd->len),
2431 iwl_pcie_tfd_get_num_tbs(trans, tfd)); 2431 iwl_pcie_tfd_get_num_tbs(trans, tfd));
diff --git a/drivers/net/xen-netfront.c b/drivers/net/xen-netfront.c
index c5a34671abda..9bd7ddeeb6a5 100644
--- a/drivers/net/xen-netfront.c
+++ b/drivers/net/xen-netfront.c
@@ -1326,6 +1326,7 @@ static struct net_device *xennet_create_dev(struct xenbus_device *dev)
1326 1326
1327 netif_carrier_off(netdev); 1327 netif_carrier_off(netdev);
1328 1328
1329 xenbus_switch_state(dev, XenbusStateInitialising);
1329 return netdev; 1330 return netdev;
1330 1331
1331 exit: 1332 exit:
diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index 1e46e60b8f10..839650e0926a 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -1335,6 +1335,7 @@ static void nvme_update_disk_info(struct gendisk *disk,
1335 struct nvme_ns *ns, struct nvme_id_ns *id) 1335 struct nvme_ns *ns, struct nvme_id_ns *id)
1336{ 1336{
1337 sector_t capacity = le64_to_cpup(&id->nsze) << (ns->lba_shift - 9); 1337 sector_t capacity = le64_to_cpup(&id->nsze) << (ns->lba_shift - 9);
1338 unsigned short bs = 1 << ns->lba_shift;
1338 unsigned stream_alignment = 0; 1339 unsigned stream_alignment = 0;
1339 1340
1340 if (ns->ctrl->nr_streams && ns->sws && ns->sgs) 1341 if (ns->ctrl->nr_streams && ns->sws && ns->sgs)
@@ -1343,7 +1344,10 @@ static void nvme_update_disk_info(struct gendisk *disk,
1343 blk_mq_freeze_queue(disk->queue); 1344 blk_mq_freeze_queue(disk->queue);
1344 blk_integrity_unregister(disk); 1345 blk_integrity_unregister(disk);
1345 1346
1346 blk_queue_logical_block_size(disk->queue, 1 << ns->lba_shift); 1347 blk_queue_logical_block_size(disk->queue, bs);
1348 blk_queue_physical_block_size(disk->queue, bs);
1349 blk_queue_io_min(disk->queue, bs);
1350
1347 if (ns->ms && !ns->ext && 1351 if (ns->ms && !ns->ext &&
1348 (ns->ctrl->ops->flags & NVME_F_METADATA_SUPPORTED)) 1352 (ns->ctrl->ops->flags & NVME_F_METADATA_SUPPORTED))
1349 nvme_init_integrity(disk, ns->ms, ns->pi_type); 1353 nvme_init_integrity(disk, ns->ms, ns->pi_type);
@@ -2987,6 +2991,7 @@ static void nvme_ns_remove(struct nvme_ns *ns)
2987 mutex_unlock(&ns->ctrl->namespaces_mutex); 2991 mutex_unlock(&ns->ctrl->namespaces_mutex);
2988 2992
2989 synchronize_srcu(&ns->head->srcu); 2993 synchronize_srcu(&ns->head->srcu);
2994 nvme_mpath_check_last_path(ns);
2990 nvme_put_ns(ns); 2995 nvme_put_ns(ns);
2991} 2996}
2992 2997
diff --git a/drivers/nvme/host/fabrics.c b/drivers/nvme/host/fabrics.c
index 76b4fe6816a0..894c2ccb3891 100644
--- a/drivers/nvme/host/fabrics.c
+++ b/drivers/nvme/host/fabrics.c
@@ -74,6 +74,7 @@ static struct nvmf_host *nvmf_host_default(void)
74 return NULL; 74 return NULL;
75 75
76 kref_init(&host->ref); 76 kref_init(&host->ref);
77 uuid_gen(&host->id);
77 snprintf(host->nqn, NVMF_NQN_SIZE, 78 snprintf(host->nqn, NVMF_NQN_SIZE,
78 "nqn.2014-08.org.nvmexpress:uuid:%pUb", &host->id); 79 "nqn.2014-08.org.nvmexpress:uuid:%pUb", &host->id);
79 80
diff --git a/drivers/nvme/host/nvme.h b/drivers/nvme/host/nvme.h
index ea1aa5283e8e..a00eabd06427 100644
--- a/drivers/nvme/host/nvme.h
+++ b/drivers/nvme/host/nvme.h
@@ -417,6 +417,15 @@ static inline void nvme_mpath_clear_current_path(struct nvme_ns *ns)
417 rcu_assign_pointer(head->current_path, NULL); 417 rcu_assign_pointer(head->current_path, NULL);
418} 418}
419struct nvme_ns *nvme_find_path(struct nvme_ns_head *head); 419struct nvme_ns *nvme_find_path(struct nvme_ns_head *head);
420
421static inline void nvme_mpath_check_last_path(struct nvme_ns *ns)
422{
423 struct nvme_ns_head *head = ns->head;
424
425 if (head->disk && list_empty(&head->list))
426 kblockd_schedule_work(&head->requeue_work);
427}
428
420#else 429#else
421static inline void nvme_failover_req(struct request *req) 430static inline void nvme_failover_req(struct request *req)
422{ 431{
@@ -448,6 +457,9 @@ static inline void nvme_mpath_remove_disk_links(struct nvme_ns *ns)
448static inline void nvme_mpath_clear_current_path(struct nvme_ns *ns) 457static inline void nvme_mpath_clear_current_path(struct nvme_ns *ns)
449{ 458{
450} 459}
460static inline void nvme_mpath_check_last_path(struct nvme_ns *ns)
461{
462}
451#endif /* CONFIG_NVME_MULTIPATH */ 463#endif /* CONFIG_NVME_MULTIPATH */
452 464
453#ifdef CONFIG_NVM 465#ifdef CONFIG_NVM
diff --git a/drivers/nvme/host/pci.c b/drivers/nvme/host/pci.c
index f5800c3c9082..d53550e612bc 100644
--- a/drivers/nvme/host/pci.c
+++ b/drivers/nvme/host/pci.c
@@ -448,12 +448,31 @@ static void **nvme_pci_iod_list(struct request *req)
448 return (void **)(iod->sg + blk_rq_nr_phys_segments(req)); 448 return (void **)(iod->sg + blk_rq_nr_phys_segments(req));
449} 449}
450 450
451static inline bool nvme_pci_use_sgls(struct nvme_dev *dev, struct request *req)
452{
453 struct nvme_iod *iod = blk_mq_rq_to_pdu(req);
454 unsigned int avg_seg_size;
455
456 avg_seg_size = DIV_ROUND_UP(blk_rq_payload_bytes(req),
457 blk_rq_nr_phys_segments(req));
458
459 if (!(dev->ctrl.sgls & ((1 << 0) | (1 << 1))))
460 return false;
461 if (!iod->nvmeq->qid)
462 return false;
463 if (!sgl_threshold || avg_seg_size < sgl_threshold)
464 return false;
465 return true;
466}
467
451static blk_status_t nvme_init_iod(struct request *rq, struct nvme_dev *dev) 468static blk_status_t nvme_init_iod(struct request *rq, struct nvme_dev *dev)
452{ 469{
453 struct nvme_iod *iod = blk_mq_rq_to_pdu(rq); 470 struct nvme_iod *iod = blk_mq_rq_to_pdu(rq);
454 int nseg = blk_rq_nr_phys_segments(rq); 471 int nseg = blk_rq_nr_phys_segments(rq);
455 unsigned int size = blk_rq_payload_bytes(rq); 472 unsigned int size = blk_rq_payload_bytes(rq);
456 473
474 iod->use_sgl = nvme_pci_use_sgls(dev, rq);
475
457 if (nseg > NVME_INT_PAGES || size > NVME_INT_BYTES(dev)) { 476 if (nseg > NVME_INT_PAGES || size > NVME_INT_BYTES(dev)) {
458 size_t alloc_size = nvme_pci_iod_alloc_size(dev, size, nseg, 477 size_t alloc_size = nvme_pci_iod_alloc_size(dev, size, nseg,
459 iod->use_sgl); 478 iod->use_sgl);
@@ -604,8 +623,6 @@ static blk_status_t nvme_pci_setup_prps(struct nvme_dev *dev,
604 dma_addr_t prp_dma; 623 dma_addr_t prp_dma;
605 int nprps, i; 624 int nprps, i;
606 625
607 iod->use_sgl = false;
608
609 length -= (page_size - offset); 626 length -= (page_size - offset);
610 if (length <= 0) { 627 if (length <= 0) {
611 iod->first_dma = 0; 628 iod->first_dma = 0;
@@ -715,8 +732,6 @@ static blk_status_t nvme_pci_setup_sgls(struct nvme_dev *dev,
715 int entries = iod->nents, i = 0; 732 int entries = iod->nents, i = 0;
716 dma_addr_t sgl_dma; 733 dma_addr_t sgl_dma;
717 734
718 iod->use_sgl = true;
719
720 /* setting the transfer type as SGL */ 735 /* setting the transfer type as SGL */
721 cmd->flags = NVME_CMD_SGL_METABUF; 736 cmd->flags = NVME_CMD_SGL_METABUF;
722 737
@@ -770,23 +785,6 @@ static blk_status_t nvme_pci_setup_sgls(struct nvme_dev *dev,
770 return BLK_STS_OK; 785 return BLK_STS_OK;
771} 786}
772 787
773static inline bool nvme_pci_use_sgls(struct nvme_dev *dev, struct request *req)
774{
775 struct nvme_iod *iod = blk_mq_rq_to_pdu(req);
776 unsigned int avg_seg_size;
777
778 avg_seg_size = DIV_ROUND_UP(blk_rq_payload_bytes(req),
779 blk_rq_nr_phys_segments(req));
780
781 if (!(dev->ctrl.sgls & ((1 << 0) | (1 << 1))))
782 return false;
783 if (!iod->nvmeq->qid)
784 return false;
785 if (!sgl_threshold || avg_seg_size < sgl_threshold)
786 return false;
787 return true;
788}
789
790static blk_status_t nvme_map_data(struct nvme_dev *dev, struct request *req, 788static blk_status_t nvme_map_data(struct nvme_dev *dev, struct request *req,
791 struct nvme_command *cmnd) 789 struct nvme_command *cmnd)
792{ 790{
@@ -806,7 +804,7 @@ static blk_status_t nvme_map_data(struct nvme_dev *dev, struct request *req,
806 DMA_ATTR_NO_WARN)) 804 DMA_ATTR_NO_WARN))
807 goto out; 805 goto out;
808 806
809 if (nvme_pci_use_sgls(dev, req)) 807 if (iod->use_sgl)
810 ret = nvme_pci_setup_sgls(dev, req, &cmnd->rw); 808 ret = nvme_pci_setup_sgls(dev, req, &cmnd->rw);
811 else 809 else
812 ret = nvme_pci_setup_prps(dev, req, &cmnd->rw); 810 ret = nvme_pci_setup_prps(dev, req, &cmnd->rw);
diff --git a/drivers/nvme/host/rdma.c b/drivers/nvme/host/rdma.c
index 37af56596be6..2a0bba7f50cf 100644
--- a/drivers/nvme/host/rdma.c
+++ b/drivers/nvme/host/rdma.c
@@ -974,12 +974,18 @@ static void nvme_rdma_error_recovery_work(struct work_struct *work)
974 blk_mq_unquiesce_queue(ctrl->ctrl.admin_q); 974 blk_mq_unquiesce_queue(ctrl->ctrl.admin_q);
975 nvme_start_queues(&ctrl->ctrl); 975 nvme_start_queues(&ctrl->ctrl);
976 976
977 if (!nvme_change_ctrl_state(&ctrl->ctrl, NVME_CTRL_RECONNECTING)) {
978 /* state change failure should never happen */
979 WARN_ON_ONCE(1);
980 return;
981 }
982
977 nvme_rdma_reconnect_or_remove(ctrl); 983 nvme_rdma_reconnect_or_remove(ctrl);
978} 984}
979 985
980static void nvme_rdma_error_recovery(struct nvme_rdma_ctrl *ctrl) 986static void nvme_rdma_error_recovery(struct nvme_rdma_ctrl *ctrl)
981{ 987{
982 if (!nvme_change_ctrl_state(&ctrl->ctrl, NVME_CTRL_RECONNECTING)) 988 if (!nvme_change_ctrl_state(&ctrl->ctrl, NVME_CTRL_RESETTING))
983 return; 989 return;
984 990
985 queue_work(nvme_wq, &ctrl->err_work); 991 queue_work(nvme_wq, &ctrl->err_work);
@@ -1753,6 +1759,12 @@ static void nvme_rdma_reset_ctrl_work(struct work_struct *work)
1753 nvme_stop_ctrl(&ctrl->ctrl); 1759 nvme_stop_ctrl(&ctrl->ctrl);
1754 nvme_rdma_shutdown_ctrl(ctrl, false); 1760 nvme_rdma_shutdown_ctrl(ctrl, false);
1755 1761
1762 if (!nvme_change_ctrl_state(&ctrl->ctrl, NVME_CTRL_RECONNECTING)) {
1763 /* state change failure should never happen */
1764 WARN_ON_ONCE(1);
1765 return;
1766 }
1767
1756 ret = nvme_rdma_configure_admin_queue(ctrl, false); 1768 ret = nvme_rdma_configure_admin_queue(ctrl, false);
1757 if (ret) 1769 if (ret)
1758 goto out_fail; 1770 goto out_fail;
diff --git a/drivers/nvme/target/fcloop.c b/drivers/nvme/target/fcloop.c
index 7b75d9de55ab..6a018a0bd6ce 100644
--- a/drivers/nvme/target/fcloop.c
+++ b/drivers/nvme/target/fcloop.c
@@ -1085,7 +1085,7 @@ fcloop_delete_target_port(struct device *dev, struct device_attribute *attr,
1085 const char *buf, size_t count) 1085 const char *buf, size_t count)
1086{ 1086{
1087 struct fcloop_nport *nport = NULL, *tmpport; 1087 struct fcloop_nport *nport = NULL, *tmpport;
1088 struct fcloop_tport *tport; 1088 struct fcloop_tport *tport = NULL;
1089 u64 nodename, portname; 1089 u64 nodename, portname;
1090 unsigned long flags; 1090 unsigned long flags;
1091 int ret; 1091 int ret;
diff --git a/drivers/of/of_mdio.c b/drivers/of/of_mdio.c
index 3481e69738b5..a327be1d264b 100644
--- a/drivers/of/of_mdio.c
+++ b/drivers/of/of_mdio.c
@@ -231,7 +231,12 @@ int of_mdiobus_register(struct mii_bus *mdio, struct device_node *np)
231 rc = of_mdiobus_register_phy(mdio, child, addr); 231 rc = of_mdiobus_register_phy(mdio, child, addr);
232 else 232 else
233 rc = of_mdiobus_register_device(mdio, child, addr); 233 rc = of_mdiobus_register_device(mdio, child, addr);
234 if (rc) 234
235 if (rc == -ENODEV)
236 dev_err(&mdio->dev,
237 "MDIO device at address %d is missing.\n",
238 addr);
239 else if (rc)
235 goto unregister; 240 goto unregister;
236 } 241 }
237 242
@@ -255,7 +260,7 @@ int of_mdiobus_register(struct mii_bus *mdio, struct device_node *np)
255 260
256 if (of_mdiobus_child_is_phy(child)) { 261 if (of_mdiobus_child_is_phy(child)) {
257 rc = of_mdiobus_register_phy(mdio, child, addr); 262 rc = of_mdiobus_register_phy(mdio, child, addr);
258 if (rc) 263 if (rc && rc != -ENODEV)
259 goto unregister; 264 goto unregister;
260 } 265 }
261 } 266 }
diff --git a/drivers/platform/x86/wmi.c b/drivers/platform/x86/wmi.c
index 791449a2370f..daa68acbc900 100644
--- a/drivers/platform/x86/wmi.c
+++ b/drivers/platform/x86/wmi.c
@@ -1458,5 +1458,5 @@ static void __exit acpi_wmi_exit(void)
1458 class_unregister(&wmi_bus_class); 1458 class_unregister(&wmi_bus_class);
1459} 1459}
1460 1460
1461subsys_initcall(acpi_wmi_init); 1461subsys_initcall_sync(acpi_wmi_init);
1462module_exit(acpi_wmi_exit); 1462module_exit(acpi_wmi_exit);
diff --git a/drivers/staging/android/ashmem.c b/drivers/staging/android/ashmem.c
index 0f695df14c9d..372ce9913e6d 100644
--- a/drivers/staging/android/ashmem.c
+++ b/drivers/staging/android/ashmem.c
@@ -765,10 +765,12 @@ static long ashmem_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
765 break; 765 break;
766 case ASHMEM_SET_SIZE: 766 case ASHMEM_SET_SIZE:
767 ret = -EINVAL; 767 ret = -EINVAL;
768 mutex_lock(&ashmem_mutex);
768 if (!asma->file) { 769 if (!asma->file) {
769 ret = 0; 770 ret = 0;
770 asma->size = (size_t)arg; 771 asma->size = (size_t)arg;
771 } 772 }
773 mutex_unlock(&ashmem_mutex);
772 break; 774 break;
773 case ASHMEM_GET_SIZE: 775 case ASHMEM_GET_SIZE:
774 ret = asma->size; 776 ret = asma->size;
diff --git a/drivers/usb/gadget/udc/core.c b/drivers/usb/gadget/udc/core.c
index 93eff7dec2f5..1b3efb14aec7 100644
--- a/drivers/usb/gadget/udc/core.c
+++ b/drivers/usb/gadget/udc/core.c
@@ -1147,11 +1147,7 @@ int usb_add_gadget_udc_release(struct device *parent, struct usb_gadget *gadget,
1147 1147
1148 udc = kzalloc(sizeof(*udc), GFP_KERNEL); 1148 udc = kzalloc(sizeof(*udc), GFP_KERNEL);
1149 if (!udc) 1149 if (!udc)
1150 goto err1; 1150 goto err_put_gadget;
1151
1152 ret = device_add(&gadget->dev);
1153 if (ret)
1154 goto err2;
1155 1151
1156 device_initialize(&udc->dev); 1152 device_initialize(&udc->dev);
1157 udc->dev.release = usb_udc_release; 1153 udc->dev.release = usb_udc_release;
@@ -1160,7 +1156,11 @@ int usb_add_gadget_udc_release(struct device *parent, struct usb_gadget *gadget,
1160 udc->dev.parent = parent; 1156 udc->dev.parent = parent;
1161 ret = dev_set_name(&udc->dev, "%s", kobject_name(&parent->kobj)); 1157 ret = dev_set_name(&udc->dev, "%s", kobject_name(&parent->kobj));
1162 if (ret) 1158 if (ret)
1163 goto err3; 1159 goto err_put_udc;
1160
1161 ret = device_add(&gadget->dev);
1162 if (ret)
1163 goto err_put_udc;
1164 1164
1165 udc->gadget = gadget; 1165 udc->gadget = gadget;
1166 gadget->udc = udc; 1166 gadget->udc = udc;
@@ -1170,7 +1170,7 @@ int usb_add_gadget_udc_release(struct device *parent, struct usb_gadget *gadget,
1170 1170
1171 ret = device_add(&udc->dev); 1171 ret = device_add(&udc->dev);
1172 if (ret) 1172 if (ret)
1173 goto err4; 1173 goto err_unlist_udc;
1174 1174
1175 usb_gadget_set_state(gadget, USB_STATE_NOTATTACHED); 1175 usb_gadget_set_state(gadget, USB_STATE_NOTATTACHED);
1176 udc->vbus = true; 1176 udc->vbus = true;
@@ -1178,27 +1178,25 @@ int usb_add_gadget_udc_release(struct device *parent, struct usb_gadget *gadget,
1178 /* pick up one of pending gadget drivers */ 1178 /* pick up one of pending gadget drivers */
1179 ret = check_pending_gadget_drivers(udc); 1179 ret = check_pending_gadget_drivers(udc);
1180 if (ret) 1180 if (ret)
1181 goto err5; 1181 goto err_del_udc;
1182 1182
1183 mutex_unlock(&udc_lock); 1183 mutex_unlock(&udc_lock);
1184 1184
1185 return 0; 1185 return 0;
1186 1186
1187err5: 1187 err_del_udc:
1188 device_del(&udc->dev); 1188 device_del(&udc->dev);
1189 1189
1190err4: 1190 err_unlist_udc:
1191 list_del(&udc->list); 1191 list_del(&udc->list);
1192 mutex_unlock(&udc_lock); 1192 mutex_unlock(&udc_lock);
1193 1193
1194err3:
1195 put_device(&udc->dev);
1196 device_del(&gadget->dev); 1194 device_del(&gadget->dev);
1197 1195
1198err2: 1196 err_put_udc:
1199 kfree(udc); 1197 put_device(&udc->dev);
1200 1198
1201err1: 1199 err_put_gadget:
1202 put_device(&gadget->dev); 1200 put_device(&gadget->dev);
1203 return ret; 1201 return ret;
1204} 1202}
diff --git a/drivers/usb/misc/usb3503.c b/drivers/usb/misc/usb3503.c
index 465dbf68b463..f723f7b8c9ac 100644
--- a/drivers/usb/misc/usb3503.c
+++ b/drivers/usb/misc/usb3503.c
@@ -279,6 +279,8 @@ static int usb3503_probe(struct usb3503 *hub)
279 if (gpio_is_valid(hub->gpio_reset)) { 279 if (gpio_is_valid(hub->gpio_reset)) {
280 err = devm_gpio_request_one(dev, hub->gpio_reset, 280 err = devm_gpio_request_one(dev, hub->gpio_reset,
281 GPIOF_OUT_INIT_LOW, "usb3503 reset"); 281 GPIOF_OUT_INIT_LOW, "usb3503 reset");
282 /* Datasheet defines a hardware reset to be at least 100us */
283 usleep_range(100, 10000);
282 if (err) { 284 if (err) {
283 dev_err(dev, 285 dev_err(dev,
284 "unable to request GPIO %d as reset pin (%d)\n", 286 "unable to request GPIO %d as reset pin (%d)\n",
diff --git a/drivers/usb/mon/mon_bin.c b/drivers/usb/mon/mon_bin.c
index f6ae753ab99b..f932f40302df 100644
--- a/drivers/usb/mon/mon_bin.c
+++ b/drivers/usb/mon/mon_bin.c
@@ -1004,7 +1004,9 @@ static long mon_bin_ioctl(struct file *file, unsigned int cmd, unsigned long arg
1004 break; 1004 break;
1005 1005
1006 case MON_IOCQ_RING_SIZE: 1006 case MON_IOCQ_RING_SIZE:
1007 mutex_lock(&rp->fetch_lock);
1007 ret = rp->b_size; 1008 ret = rp->b_size;
1009 mutex_unlock(&rp->fetch_lock);
1008 break; 1010 break;
1009 1011
1010 case MON_IOCT_RING_SIZE: 1012 case MON_IOCT_RING_SIZE:
@@ -1231,12 +1233,16 @@ static int mon_bin_vma_fault(struct vm_fault *vmf)
1231 unsigned long offset, chunk_idx; 1233 unsigned long offset, chunk_idx;
1232 struct page *pageptr; 1234 struct page *pageptr;
1233 1235
1236 mutex_lock(&rp->fetch_lock);
1234 offset = vmf->pgoff << PAGE_SHIFT; 1237 offset = vmf->pgoff << PAGE_SHIFT;
1235 if (offset >= rp->b_size) 1238 if (offset >= rp->b_size) {
1239 mutex_unlock(&rp->fetch_lock);
1236 return VM_FAULT_SIGBUS; 1240 return VM_FAULT_SIGBUS;
1241 }
1237 chunk_idx = offset / CHUNK_SIZE; 1242 chunk_idx = offset / CHUNK_SIZE;
1238 pageptr = rp->b_vec[chunk_idx].pg; 1243 pageptr = rp->b_vec[chunk_idx].pg;
1239 get_page(pageptr); 1244 get_page(pageptr);
1245 mutex_unlock(&rp->fetch_lock);
1240 vmf->page = pageptr; 1246 vmf->page = pageptr;
1241 return 0; 1247 return 0;
1242} 1248}
diff --git a/drivers/usb/serial/cp210x.c b/drivers/usb/serial/cp210x.c
index 7c6273bf5beb..06d502b3e913 100644
--- a/drivers/usb/serial/cp210x.c
+++ b/drivers/usb/serial/cp210x.c
@@ -124,6 +124,7 @@ static const struct usb_device_id id_table[] = {
124 { USB_DEVICE(0x10C4, 0x8470) }, /* Juniper Networks BX Series System Console */ 124 { USB_DEVICE(0x10C4, 0x8470) }, /* Juniper Networks BX Series System Console */
125 { USB_DEVICE(0x10C4, 0x8477) }, /* Balluff RFID */ 125 { USB_DEVICE(0x10C4, 0x8477) }, /* Balluff RFID */
126 { USB_DEVICE(0x10C4, 0x84B6) }, /* Starizona Hyperion */ 126 { USB_DEVICE(0x10C4, 0x84B6) }, /* Starizona Hyperion */
127 { USB_DEVICE(0x10C4, 0x85A7) }, /* LifeScan OneTouch Verio IQ */
127 { USB_DEVICE(0x10C4, 0x85EA) }, /* AC-Services IBUS-IF */ 128 { USB_DEVICE(0x10C4, 0x85EA) }, /* AC-Services IBUS-IF */
128 { USB_DEVICE(0x10C4, 0x85EB) }, /* AC-Services CIS-IBUS */ 129 { USB_DEVICE(0x10C4, 0x85EB) }, /* AC-Services CIS-IBUS */
129 { USB_DEVICE(0x10C4, 0x85F8) }, /* Virtenio Preon32 */ 130 { USB_DEVICE(0x10C4, 0x85F8) }, /* Virtenio Preon32 */
@@ -174,6 +175,7 @@ static const struct usb_device_id id_table[] = {
174 { USB_DEVICE(0x1843, 0x0200) }, /* Vaisala USB Instrument Cable */ 175 { USB_DEVICE(0x1843, 0x0200) }, /* Vaisala USB Instrument Cable */
175 { USB_DEVICE(0x18EF, 0xE00F) }, /* ELV USB-I2C-Interface */ 176 { USB_DEVICE(0x18EF, 0xE00F) }, /* ELV USB-I2C-Interface */
176 { USB_DEVICE(0x18EF, 0xE025) }, /* ELV Marble Sound Board 1 */ 177 { USB_DEVICE(0x18EF, 0xE025) }, /* ELV Marble Sound Board 1 */
178 { USB_DEVICE(0x18EF, 0xE030) }, /* ELV ALC 8xxx Battery Charger */
177 { USB_DEVICE(0x18EF, 0xE032) }, /* ELV TFD500 Data Logger */ 179 { USB_DEVICE(0x18EF, 0xE032) }, /* ELV TFD500 Data Logger */
178 { USB_DEVICE(0x1901, 0x0190) }, /* GE B850 CP2105 Recorder interface */ 180 { USB_DEVICE(0x1901, 0x0190) }, /* GE B850 CP2105 Recorder interface */
179 { USB_DEVICE(0x1901, 0x0193) }, /* GE B650 CP2104 PMC interface */ 181 { USB_DEVICE(0x1901, 0x0193) }, /* GE B650 CP2104 PMC interface */
diff --git a/drivers/usb/storage/unusual_uas.h b/drivers/usb/storage/unusual_uas.h
index e6127fb21c12..a7d08ae0adad 100644
--- a/drivers/usb/storage/unusual_uas.h
+++ b/drivers/usb/storage/unusual_uas.h
@@ -143,6 +143,13 @@ UNUSUAL_DEV(0x2109, 0x0711, 0x0000, 0x9999,
143 USB_SC_DEVICE, USB_PR_DEVICE, NULL, 143 USB_SC_DEVICE, USB_PR_DEVICE, NULL,
144 US_FL_NO_ATA_1X), 144 US_FL_NO_ATA_1X),
145 145
146/* Reported-by: Icenowy Zheng <icenowy@aosc.io> */
147UNUSUAL_DEV(0x2537, 0x1068, 0x0000, 0x9999,
148 "Norelsys",
149 "NS1068X",
150 USB_SC_DEVICE, USB_PR_DEVICE, NULL,
151 US_FL_IGNORE_UAS),
152
146/* Reported-by: Takeo Nakayama <javhera@gmx.com> */ 153/* Reported-by: Takeo Nakayama <javhera@gmx.com> */
147UNUSUAL_DEV(0x357d, 0x7788, 0x0000, 0x9999, 154UNUSUAL_DEV(0x357d, 0x7788, 0x0000, 0x9999,
148 "JMicron", 155 "JMicron",
diff --git a/drivers/usb/usbip/usbip_common.c b/drivers/usb/usbip/usbip_common.c
index 7b219d9109b4..ee2bbce24584 100644
--- a/drivers/usb/usbip/usbip_common.c
+++ b/drivers/usb/usbip/usbip_common.c
@@ -91,7 +91,7 @@ static void usbip_dump_usb_device(struct usb_device *udev)
91 dev_dbg(dev, " devnum(%d) devpath(%s) usb speed(%s)", 91 dev_dbg(dev, " devnum(%d) devpath(%s) usb speed(%s)",
92 udev->devnum, udev->devpath, usb_speed_string(udev->speed)); 92 udev->devnum, udev->devpath, usb_speed_string(udev->speed));
93 93
94 pr_debug("tt %p, ttport %d\n", udev->tt, udev->ttport); 94 pr_debug("tt hub ttport %d\n", udev->ttport);
95 95
96 dev_dbg(dev, " "); 96 dev_dbg(dev, " ");
97 for (i = 0; i < 16; i++) 97 for (i = 0; i < 16; i++)
@@ -124,12 +124,8 @@ static void usbip_dump_usb_device(struct usb_device *udev)
124 } 124 }
125 pr_debug("\n"); 125 pr_debug("\n");
126 126
127 dev_dbg(dev, "parent %p, bus %p\n", udev->parent, udev->bus); 127 dev_dbg(dev, "parent %s, bus %s\n", dev_name(&udev->parent->dev),
128 128 udev->bus->bus_name);
129 dev_dbg(dev,
130 "descriptor %p, config %p, actconfig %p, rawdescriptors %p\n",
131 &udev->descriptor, udev->config,
132 udev->actconfig, udev->rawdescriptors);
133 129
134 dev_dbg(dev, "have_langid %d, string_langid %d\n", 130 dev_dbg(dev, "have_langid %d, string_langid %d\n",
135 udev->have_langid, udev->string_langid); 131 udev->have_langid, udev->string_langid);
@@ -237,9 +233,6 @@ void usbip_dump_urb(struct urb *urb)
237 233
238 dev = &urb->dev->dev; 234 dev = &urb->dev->dev;
239 235
240 dev_dbg(dev, " urb :%p\n", urb);
241 dev_dbg(dev, " dev :%p\n", urb->dev);
242
243 usbip_dump_usb_device(urb->dev); 236 usbip_dump_usb_device(urb->dev);
244 237
245 dev_dbg(dev, " pipe :%08x ", urb->pipe); 238 dev_dbg(dev, " pipe :%08x ", urb->pipe);
@@ -248,11 +241,9 @@ void usbip_dump_urb(struct urb *urb)
248 241
249 dev_dbg(dev, " status :%d\n", urb->status); 242 dev_dbg(dev, " status :%d\n", urb->status);
250 dev_dbg(dev, " transfer_flags :%08X\n", urb->transfer_flags); 243 dev_dbg(dev, " transfer_flags :%08X\n", urb->transfer_flags);
251 dev_dbg(dev, " transfer_buffer :%p\n", urb->transfer_buffer);
252 dev_dbg(dev, " transfer_buffer_length:%d\n", 244 dev_dbg(dev, " transfer_buffer_length:%d\n",
253 urb->transfer_buffer_length); 245 urb->transfer_buffer_length);
254 dev_dbg(dev, " actual_length :%d\n", urb->actual_length); 246 dev_dbg(dev, " actual_length :%d\n", urb->actual_length);
255 dev_dbg(dev, " setup_packet :%p\n", urb->setup_packet);
256 247
257 if (urb->setup_packet && usb_pipetype(urb->pipe) == PIPE_CONTROL) 248 if (urb->setup_packet && usb_pipetype(urb->pipe) == PIPE_CONTROL)
258 usbip_dump_usb_ctrlrequest( 249 usbip_dump_usb_ctrlrequest(
@@ -262,8 +253,6 @@ void usbip_dump_urb(struct urb *urb)
262 dev_dbg(dev, " number_of_packets :%d\n", urb->number_of_packets); 253 dev_dbg(dev, " number_of_packets :%d\n", urb->number_of_packets);
263 dev_dbg(dev, " interval :%d\n", urb->interval); 254 dev_dbg(dev, " interval :%d\n", urb->interval);
264 dev_dbg(dev, " error_count :%d\n", urb->error_count); 255 dev_dbg(dev, " error_count :%d\n", urb->error_count);
265 dev_dbg(dev, " context :%p\n", urb->context);
266 dev_dbg(dev, " complete :%p\n", urb->complete);
267} 256}
268EXPORT_SYMBOL_GPL(usbip_dump_urb); 257EXPORT_SYMBOL_GPL(usbip_dump_urb);
269 258
diff --git a/drivers/usb/usbip/vudc_rx.c b/drivers/usb/usbip/vudc_rx.c
index df1e30989148..1e8a23d92cb4 100644
--- a/drivers/usb/usbip/vudc_rx.c
+++ b/drivers/usb/usbip/vudc_rx.c
@@ -120,6 +120,25 @@ static int v_recv_cmd_submit(struct vudc *udc,
120 urb_p->new = 1; 120 urb_p->new = 1;
121 urb_p->seqnum = pdu->base.seqnum; 121 urb_p->seqnum = pdu->base.seqnum;
122 122
123 if (urb_p->ep->type == USB_ENDPOINT_XFER_ISOC) {
124 /* validate packet size and number of packets */
125 unsigned int maxp, packets, bytes;
126
127 maxp = usb_endpoint_maxp(urb_p->ep->desc);
128 maxp *= usb_endpoint_maxp_mult(urb_p->ep->desc);
129 bytes = pdu->u.cmd_submit.transfer_buffer_length;
130 packets = DIV_ROUND_UP(bytes, maxp);
131
132 if (pdu->u.cmd_submit.number_of_packets < 0 ||
133 pdu->u.cmd_submit.number_of_packets > packets) {
134 dev_err(&udc->gadget.dev,
135 "CMD_SUBMIT: isoc invalid num packets %d\n",
136 pdu->u.cmd_submit.number_of_packets);
137 ret = -EMSGSIZE;
138 goto free_urbp;
139 }
140 }
141
123 ret = alloc_urb_from_cmd(&urb_p->urb, pdu, urb_p->ep->type); 142 ret = alloc_urb_from_cmd(&urb_p->urb, pdu, urb_p->ep->type);
124 if (ret) { 143 if (ret) {
125 usbip_event_add(&udc->ud, VUDC_EVENT_ERROR_MALLOC); 144 usbip_event_add(&udc->ud, VUDC_EVENT_ERROR_MALLOC);
diff --git a/drivers/usb/usbip/vudc_tx.c b/drivers/usb/usbip/vudc_tx.c
index 1440ae0919ec..3ccb17c3e840 100644
--- a/drivers/usb/usbip/vudc_tx.c
+++ b/drivers/usb/usbip/vudc_tx.c
@@ -85,6 +85,13 @@ static int v_send_ret_submit(struct vudc *udc, struct urbp *urb_p)
85 memset(&pdu_header, 0, sizeof(pdu_header)); 85 memset(&pdu_header, 0, sizeof(pdu_header));
86 memset(&msg, 0, sizeof(msg)); 86 memset(&msg, 0, sizeof(msg));
87 87
88 if (urb->actual_length > 0 && !urb->transfer_buffer) {
89 dev_err(&udc->gadget.dev,
90 "urb: actual_length %d transfer_buffer null\n",
91 urb->actual_length);
92 return -1;
93 }
94
88 if (urb_p->type == USB_ENDPOINT_XFER_ISOC) 95 if (urb_p->type == USB_ENDPOINT_XFER_ISOC)
89 iovnum = 2 + urb->number_of_packets; 96 iovnum = 2 + urb->number_of_packets;
90 else 97 else
@@ -100,8 +107,8 @@ static int v_send_ret_submit(struct vudc *udc, struct urbp *urb_p)
100 107
101 /* 1. setup usbip_header */ 108 /* 1. setup usbip_header */
102 setup_ret_submit_pdu(&pdu_header, urb_p); 109 setup_ret_submit_pdu(&pdu_header, urb_p);
103 usbip_dbg_stub_tx("setup txdata seqnum: %d urb: %p\n", 110 usbip_dbg_stub_tx("setup txdata seqnum: %d\n",
104 pdu_header.base.seqnum, urb); 111 pdu_header.base.seqnum);
105 usbip_header_correct_endian(&pdu_header, 1); 112 usbip_header_correct_endian(&pdu_header, 1);
106 113
107 iov[iovnum].iov_base = &pdu_header; 114 iov[iovnum].iov_base = &pdu_header;
diff --git a/drivers/xen/gntdev.c b/drivers/xen/gntdev.c
index 57efbd3b053b..bd56653b9bbc 100644
--- a/drivers/xen/gntdev.c
+++ b/drivers/xen/gntdev.c
@@ -380,10 +380,8 @@ static int unmap_grant_pages(struct grant_map *map, int offset, int pages)
380 } 380 }
381 range = 0; 381 range = 0;
382 while (range < pages) { 382 while (range < pages) {
383 if (map->unmap_ops[offset+range].handle == -1) { 383 if (map->unmap_ops[offset+range].handle == -1)
384 range--;
385 break; 384 break;
386 }
387 range++; 385 range++;
388 } 386 }
389 err = __unmap_grant_pages(map, offset, range); 387 err = __unmap_grant_pages(map, offset, range);
@@ -1073,8 +1071,10 @@ unlock_out:
1073out_unlock_put: 1071out_unlock_put:
1074 mutex_unlock(&priv->lock); 1072 mutex_unlock(&priv->lock);
1075out_put_map: 1073out_put_map:
1076 if (use_ptemod) 1074 if (use_ptemod) {
1077 map->vma = NULL; 1075 map->vma = NULL;
1076 unmap_grant_pages(map, 0, map->count);
1077 }
1078 gntdev_put_map(priv, map); 1078 gntdev_put_map(priv, map);
1079 return err; 1079 return err;
1080} 1080}
diff --git a/include/acpi/acpi_bus.h b/include/acpi/acpi_bus.h
index 79287629c888..c9608b0b80c6 100644
--- a/include/acpi/acpi_bus.h
+++ b/include/acpi/acpi_bus.h
@@ -91,6 +91,9 @@ acpi_evaluate_dsm_typed(acpi_handle handle, const guid_t *guid, u64 rev,
91bool acpi_dev_found(const char *hid); 91bool acpi_dev_found(const char *hid);
92bool acpi_dev_present(const char *hid, const char *uid, s64 hrv); 92bool acpi_dev_present(const char *hid, const char *uid, s64 hrv);
93 93
94const char *
95acpi_dev_get_first_match_name(const char *hid, const char *uid, s64 hrv);
96
94#ifdef CONFIG_ACPI 97#ifdef CONFIG_ACPI
95 98
96#include <linux/proc_fs.h> 99#include <linux/proc_fs.h>
diff --git a/include/linux/acpi.h b/include/linux/acpi.h
index dc1ebfeeb5ec..d918f1ea84e6 100644
--- a/include/linux/acpi.h
+++ b/include/linux/acpi.h
@@ -640,6 +640,12 @@ static inline bool acpi_dev_present(const char *hid, const char *uid, s64 hrv)
640 return false; 640 return false;
641} 641}
642 642
643static inline const char *
644acpi_dev_get_first_match_name(const char *hid, const char *uid, s64 hrv)
645{
646 return NULL;
647}
648
643static inline bool is_acpi_node(struct fwnode_handle *fwnode) 649static inline bool is_acpi_node(struct fwnode_handle *fwnode)
644{ 650{
645 return false; 651 return false;
diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index b63a592ad29d..0b25cf87b6d6 100644
--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -43,7 +43,14 @@ struct bpf_map_ops {
43}; 43};
44 44
45struct bpf_map { 45struct bpf_map {
46 atomic_t refcnt; 46 /* 1st cacheline with read-mostly members of which some
47 * are also accessed in fast-path (e.g. ops, max_entries).
48 */
49 const struct bpf_map_ops *ops ____cacheline_aligned;
50 struct bpf_map *inner_map_meta;
51#ifdef CONFIG_SECURITY
52 void *security;
53#endif
47 enum bpf_map_type map_type; 54 enum bpf_map_type map_type;
48 u32 key_size; 55 u32 key_size;
49 u32 value_size; 56 u32 value_size;
@@ -52,15 +59,17 @@ struct bpf_map {
52 u32 pages; 59 u32 pages;
53 u32 id; 60 u32 id;
54 int numa_node; 61 int numa_node;
55 struct user_struct *user; 62 bool unpriv_array;
56 const struct bpf_map_ops *ops; 63 /* 7 bytes hole */
57 struct work_struct work; 64
65 /* 2nd cacheline with misc members to avoid false sharing
66 * particularly with refcounting.
67 */
68 struct user_struct *user ____cacheline_aligned;
69 atomic_t refcnt;
58 atomic_t usercnt; 70 atomic_t usercnt;
59 struct bpf_map *inner_map_meta; 71 struct work_struct work;
60 char name[BPF_OBJ_NAME_LEN]; 72 char name[BPF_OBJ_NAME_LEN];
61#ifdef CONFIG_SECURITY
62 void *security;
63#endif
64}; 73};
65 74
66/* function argument constraints */ 75/* function argument constraints */
@@ -221,6 +230,7 @@ struct bpf_prog_aux {
221struct bpf_array { 230struct bpf_array {
222 struct bpf_map map; 231 struct bpf_map map;
223 u32 elem_size; 232 u32 elem_size;
233 u32 index_mask;
224 /* 'ownership' of prog_array is claimed by the first program that 234 /* 'ownership' of prog_array is claimed by the first program that
225 * is going to use this map or by the first program which FD is stored 235 * is going to use this map or by the first program which FD is stored
226 * in the map to make sure that all callers and callees have the same 236 * in the map to make sure that all callers and callees have the same
diff --git a/include/linux/completion.h b/include/linux/completion.h
index 94a59ba7d422..519e94915d18 100644
--- a/include/linux/completion.h
+++ b/include/linux/completion.h
@@ -32,7 +32,6 @@ struct completion {
32#define init_completion(x) __init_completion(x) 32#define init_completion(x) __init_completion(x)
33static inline void complete_acquire(struct completion *x) {} 33static inline void complete_acquire(struct completion *x) {}
34static inline void complete_release(struct completion *x) {} 34static inline void complete_release(struct completion *x) {}
35static inline void complete_release_commit(struct completion *x) {}
36 35
37#define COMPLETION_INITIALIZER(work) \ 36#define COMPLETION_INITIALIZER(work) \
38 { 0, __WAIT_QUEUE_HEAD_INITIALIZER((work).wait) } 37 { 0, __WAIT_QUEUE_HEAD_INITIALIZER((work).wait) }
diff --git a/include/linux/cpu.h b/include/linux/cpu.h
index a04ef7c15c6a..7b01bc11c692 100644
--- a/include/linux/cpu.h
+++ b/include/linux/cpu.h
@@ -47,6 +47,13 @@ extern void cpu_remove_dev_attr(struct device_attribute *attr);
47extern int cpu_add_dev_attr_group(struct attribute_group *attrs); 47extern int cpu_add_dev_attr_group(struct attribute_group *attrs);
48extern void cpu_remove_dev_attr_group(struct attribute_group *attrs); 48extern void cpu_remove_dev_attr_group(struct attribute_group *attrs);
49 49
50extern ssize_t cpu_show_meltdown(struct device *dev,
51 struct device_attribute *attr, char *buf);
52extern ssize_t cpu_show_spectre_v1(struct device *dev,
53 struct device_attribute *attr, char *buf);
54extern ssize_t cpu_show_spectre_v2(struct device *dev,
55 struct device_attribute *attr, char *buf);
56
50extern __printf(4, 5) 57extern __printf(4, 5)
51struct device *cpu_device_create(struct device *parent, void *drvdata, 58struct device *cpu_device_create(struct device *parent, void *drvdata,
52 const struct attribute_group **groups, 59 const struct attribute_group **groups,
diff --git a/include/linux/crash_core.h b/include/linux/crash_core.h
index 06097ef30449..b511f6d24b42 100644
--- a/include/linux/crash_core.h
+++ b/include/linux/crash_core.h
@@ -42,6 +42,8 @@ phys_addr_t paddr_vmcoreinfo_note(void);
42 vmcoreinfo_append_str("PAGESIZE=%ld\n", value) 42 vmcoreinfo_append_str("PAGESIZE=%ld\n", value)
43#define VMCOREINFO_SYMBOL(name) \ 43#define VMCOREINFO_SYMBOL(name) \
44 vmcoreinfo_append_str("SYMBOL(%s)=%lx\n", #name, (unsigned long)&name) 44 vmcoreinfo_append_str("SYMBOL(%s)=%lx\n", #name, (unsigned long)&name)
45#define VMCOREINFO_SYMBOL_ARRAY(name) \
46 vmcoreinfo_append_str("SYMBOL(%s)=%lx\n", #name, (unsigned long)name)
45#define VMCOREINFO_SIZE(name) \ 47#define VMCOREINFO_SIZE(name) \
46 vmcoreinfo_append_str("SIZE(%s)=%lu\n", #name, \ 48 vmcoreinfo_append_str("SIZE(%s)=%lu\n", #name, \
47 (unsigned long)sizeof(name)) 49 (unsigned long)sizeof(name))
diff --git a/include/linux/irqflags.h b/include/linux/irqflags.h
index 46cb57d5eb13..1b3996ff3f16 100644
--- a/include/linux/irqflags.h
+++ b/include/linux/irqflags.h
@@ -27,22 +27,18 @@
27# define trace_hardirq_enter() \ 27# define trace_hardirq_enter() \
28do { \ 28do { \
29 current->hardirq_context++; \ 29 current->hardirq_context++; \
30 crossrelease_hist_start(XHLOCK_HARD); \
31} while (0) 30} while (0)
32# define trace_hardirq_exit() \ 31# define trace_hardirq_exit() \
33do { \ 32do { \
34 current->hardirq_context--; \ 33 current->hardirq_context--; \
35 crossrelease_hist_end(XHLOCK_HARD); \
36} while (0) 34} while (0)
37# define lockdep_softirq_enter() \ 35# define lockdep_softirq_enter() \
38do { \ 36do { \
39 current->softirq_context++; \ 37 current->softirq_context++; \
40 crossrelease_hist_start(XHLOCK_SOFT); \
41} while (0) 38} while (0)
42# define lockdep_softirq_exit() \ 39# define lockdep_softirq_exit() \
43do { \ 40do { \
44 current->softirq_context--; \ 41 current->softirq_context--; \
45 crossrelease_hist_end(XHLOCK_SOFT); \
46} while (0) 42} while (0)
47# define INIT_TRACE_IRQFLAGS .softirqs_enabled = 1, 43# define INIT_TRACE_IRQFLAGS .softirqs_enabled = 1,
48#else 44#else
diff --git a/include/linux/lockdep.h b/include/linux/lockdep.h
index 2e75dc34bff5..3251d9c0d313 100644
--- a/include/linux/lockdep.h
+++ b/include/linux/lockdep.h
@@ -475,8 +475,6 @@ enum xhlock_context_t {
475#define STATIC_LOCKDEP_MAP_INIT(_name, _key) \ 475#define STATIC_LOCKDEP_MAP_INIT(_name, _key) \
476 { .name = (_name), .key = (void *)(_key), } 476 { .name = (_name), .key = (void *)(_key), }
477 477
478static inline void crossrelease_hist_start(enum xhlock_context_t c) {}
479static inline void crossrelease_hist_end(enum xhlock_context_t c) {}
480static inline void lockdep_invariant_state(bool force) {} 478static inline void lockdep_invariant_state(bool force) {}
481static inline void lockdep_init_task(struct task_struct *task) {} 479static inline void lockdep_init_task(struct task_struct *task) {}
482static inline void lockdep_free_task(struct task_struct *task) {} 480static inline void lockdep_free_task(struct task_struct *task) {}
diff --git a/include/linux/sh_eth.h b/include/linux/sh_eth.h
index ff3642d267f7..94081e9a5010 100644
--- a/include/linux/sh_eth.h
+++ b/include/linux/sh_eth.h
@@ -17,7 +17,6 @@ struct sh_eth_plat_data {
17 unsigned char mac_addr[ETH_ALEN]; 17 unsigned char mac_addr[ETH_ALEN];
18 unsigned no_ether_link:1; 18 unsigned no_ether_link:1;
19 unsigned ether_link_active_low:1; 19 unsigned ether_link_active_low:1;
20 unsigned needs_init:1;
21}; 20};
22 21
23#endif 22#endif
diff --git a/include/net/sctp/structs.h b/include/net/sctp/structs.h
index 2f8f93da5dc2..9a5ccf03a59b 100644
--- a/include/net/sctp/structs.h
+++ b/include/net/sctp/structs.h
@@ -966,7 +966,7 @@ void sctp_transport_burst_limited(struct sctp_transport *);
966void sctp_transport_burst_reset(struct sctp_transport *); 966void sctp_transport_burst_reset(struct sctp_transport *);
967unsigned long sctp_transport_timeout(struct sctp_transport *); 967unsigned long sctp_transport_timeout(struct sctp_transport *);
968void sctp_transport_reset(struct sctp_transport *t); 968void sctp_transport_reset(struct sctp_transport *t);
969void sctp_transport_update_pmtu(struct sctp_transport *t, u32 pmtu); 969bool sctp_transport_update_pmtu(struct sctp_transport *t, u32 pmtu);
970void sctp_transport_immediate_rtx(struct sctp_transport *); 970void sctp_transport_immediate_rtx(struct sctp_transport *);
971void sctp_transport_dst_release(struct sctp_transport *t); 971void sctp_transport_dst_release(struct sctp_transport *t);
972void sctp_transport_dst_confirm(struct sctp_transport *t); 972void sctp_transport_dst_confirm(struct sctp_transport *t);
diff --git a/include/net/vxlan.h b/include/net/vxlan.h
index 13223396dc64..f96391e84a8a 100644
--- a/include/net/vxlan.h
+++ b/include/net/vxlan.h
@@ -146,7 +146,7 @@ struct vxlanhdr_gpe {
146 np_applied:1, 146 np_applied:1,
147 instance_applied:1, 147 instance_applied:1,
148 version:2, 148 version:2,
149reserved_flags2:2; 149 reserved_flags2:2;
150#elif defined(__BIG_ENDIAN_BITFIELD) 150#elif defined(__BIG_ENDIAN_BITFIELD)
151 u8 reserved_flags2:2, 151 u8 reserved_flags2:2,
152 version:2, 152 version:2,
diff --git a/include/sound/soc-acpi.h b/include/sound/soc-acpi.h
index d1aaf876cd26..082224275f52 100644
--- a/include/sound/soc-acpi.h
+++ b/include/sound/soc-acpi.h
@@ -27,17 +27,13 @@ struct snd_soc_acpi_package_context {
27 bool data_valid; 27 bool data_valid;
28}; 28};
29 29
30/* codec name is used in DAIs is i2c-<HID>:00 with HID being 8 chars */
31#define SND_ACPI_I2C_ID_LEN (4 + ACPI_ID_LEN + 3 + 1)
32
30#if IS_ENABLED(CONFIG_ACPI) 33#if IS_ENABLED(CONFIG_ACPI)
31/* translation fron HID to I2C name, needed for DAI codec_name */
32const char *snd_soc_acpi_find_name_from_hid(const u8 hid[ACPI_ID_LEN]);
33bool snd_soc_acpi_find_package_from_hid(const u8 hid[ACPI_ID_LEN], 34bool snd_soc_acpi_find_package_from_hid(const u8 hid[ACPI_ID_LEN],
34 struct snd_soc_acpi_package_context *ctx); 35 struct snd_soc_acpi_package_context *ctx);
35#else 36#else
36static inline const char *
37snd_soc_acpi_find_name_from_hid(const u8 hid[ACPI_ID_LEN])
38{
39 return NULL;
40}
41static inline bool 37static inline bool
42snd_soc_acpi_find_package_from_hid(const u8 hid[ACPI_ID_LEN], 38snd_soc_acpi_find_package_from_hid(const u8 hid[ACPI_ID_LEN],
43 struct snd_soc_acpi_package_context *ctx) 39 struct snd_soc_acpi_package_context *ctx)
diff --git a/include/sound/soc.h b/include/sound/soc.h
index a34aa200e113..b655d987fbe7 100644
--- a/include/sound/soc.h
+++ b/include/sound/soc.h
@@ -804,6 +804,9 @@ struct snd_soc_component_driver {
804 int (*suspend)(struct snd_soc_component *); 804 int (*suspend)(struct snd_soc_component *);
805 int (*resume)(struct snd_soc_component *); 805 int (*resume)(struct snd_soc_component *);
806 806
807 unsigned int (*read)(struct snd_soc_component *, unsigned int);
808 int (*write)(struct snd_soc_component *, unsigned int, unsigned int);
809
807 /* pcm creation and destruction */ 810 /* pcm creation and destruction */
808 int (*pcm_new)(struct snd_soc_pcm_runtime *); 811 int (*pcm_new)(struct snd_soc_pcm_runtime *);
809 void (*pcm_free)(struct snd_pcm *); 812 void (*pcm_free)(struct snd_pcm *);
diff --git a/include/uapi/linux/if_ether.h b/include/uapi/linux/if_ether.h
index 3ee3bf7c8526..144de4d2f385 100644
--- a/include/uapi/linux/if_ether.h
+++ b/include/uapi/linux/if_ether.h
@@ -23,6 +23,7 @@
23#define _UAPI_LINUX_IF_ETHER_H 23#define _UAPI_LINUX_IF_ETHER_H
24 24
25#include <linux/types.h> 25#include <linux/types.h>
26#include <linux/libc-compat.h>
26 27
27/* 28/*
28 * IEEE 802.3 Ethernet magic constants. The frame sizes omit the preamble 29 * IEEE 802.3 Ethernet magic constants. The frame sizes omit the preamble
@@ -149,11 +150,13 @@
149 * This is an Ethernet frame header. 150 * This is an Ethernet frame header.
150 */ 151 */
151 152
153#if __UAPI_DEF_ETHHDR
152struct ethhdr { 154struct ethhdr {
153 unsigned char h_dest[ETH_ALEN]; /* destination eth addr */ 155 unsigned char h_dest[ETH_ALEN]; /* destination eth addr */
154 unsigned char h_source[ETH_ALEN]; /* source ether addr */ 156 unsigned char h_source[ETH_ALEN]; /* source ether addr */
155 __be16 h_proto; /* packet type ID field */ 157 __be16 h_proto; /* packet type ID field */
156} __attribute__((packed)); 158} __attribute__((packed));
159#endif
157 160
158 161
159#endif /* _UAPI_LINUX_IF_ETHER_H */ 162#endif /* _UAPI_LINUX_IF_ETHER_H */
diff --git a/include/uapi/linux/libc-compat.h b/include/uapi/linux/libc-compat.h
index 282875cf8056..fc29efaa918c 100644
--- a/include/uapi/linux/libc-compat.h
+++ b/include/uapi/linux/libc-compat.h
@@ -168,47 +168,106 @@
168 168
169/* If we did not see any headers from any supported C libraries, 169/* If we did not see any headers from any supported C libraries,
170 * or we are being included in the kernel, then define everything 170 * or we are being included in the kernel, then define everything
171 * that we need. */ 171 * that we need. Check for previous __UAPI_* definitions to give
172 * unsupported C libraries a way to opt out of any kernel definition. */
172#else /* !defined(__GLIBC__) */ 173#else /* !defined(__GLIBC__) */
173 174
174/* Definitions for if.h */ 175/* Definitions for if.h */
176#ifndef __UAPI_DEF_IF_IFCONF
175#define __UAPI_DEF_IF_IFCONF 1 177#define __UAPI_DEF_IF_IFCONF 1
178#endif
179#ifndef __UAPI_DEF_IF_IFMAP
176#define __UAPI_DEF_IF_IFMAP 1 180#define __UAPI_DEF_IF_IFMAP 1
181#endif
182#ifndef __UAPI_DEF_IF_IFNAMSIZ
177#define __UAPI_DEF_IF_IFNAMSIZ 1 183#define __UAPI_DEF_IF_IFNAMSIZ 1
184#endif
185#ifndef __UAPI_DEF_IF_IFREQ
178#define __UAPI_DEF_IF_IFREQ 1 186#define __UAPI_DEF_IF_IFREQ 1
187#endif
179/* Everything up to IFF_DYNAMIC, matches net/if.h until glibc 2.23 */ 188/* Everything up to IFF_DYNAMIC, matches net/if.h until glibc 2.23 */
189#ifndef __UAPI_DEF_IF_NET_DEVICE_FLAGS
180#define __UAPI_DEF_IF_NET_DEVICE_FLAGS 1 190#define __UAPI_DEF_IF_NET_DEVICE_FLAGS 1
191#endif
181/* For the future if glibc adds IFF_LOWER_UP, IFF_DORMANT and IFF_ECHO */ 192/* For the future if glibc adds IFF_LOWER_UP, IFF_DORMANT and IFF_ECHO */
193#ifndef __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO
182#define __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO 1 194#define __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO 1
195#endif
183 196
184/* Definitions for in.h */ 197/* Definitions for in.h */
198#ifndef __UAPI_DEF_IN_ADDR
185#define __UAPI_DEF_IN_ADDR 1 199#define __UAPI_DEF_IN_ADDR 1
200#endif
201#ifndef __UAPI_DEF_IN_IPPROTO
186#define __UAPI_DEF_IN_IPPROTO 1 202#define __UAPI_DEF_IN_IPPROTO 1
203#endif
204#ifndef __UAPI_DEF_IN_PKTINFO
187#define __UAPI_DEF_IN_PKTINFO 1 205#define __UAPI_DEF_IN_PKTINFO 1
206#endif
207#ifndef __UAPI_DEF_IP_MREQ
188#define __UAPI_DEF_IP_MREQ 1 208#define __UAPI_DEF_IP_MREQ 1
209#endif
210#ifndef __UAPI_DEF_SOCKADDR_IN
189#define __UAPI_DEF_SOCKADDR_IN 1 211#define __UAPI_DEF_SOCKADDR_IN 1
212#endif
213#ifndef __UAPI_DEF_IN_CLASS
190#define __UAPI_DEF_IN_CLASS 1 214#define __UAPI_DEF_IN_CLASS 1
215#endif
191 216
192/* Definitions for in6.h */ 217/* Definitions for in6.h */
218#ifndef __UAPI_DEF_IN6_ADDR
193#define __UAPI_DEF_IN6_ADDR 1 219#define __UAPI_DEF_IN6_ADDR 1
220#endif
221#ifndef __UAPI_DEF_IN6_ADDR_ALT
194#define __UAPI_DEF_IN6_ADDR_ALT 1 222#define __UAPI_DEF_IN6_ADDR_ALT 1
223#endif
224#ifndef __UAPI_DEF_SOCKADDR_IN6
195#define __UAPI_DEF_SOCKADDR_IN6 1 225#define __UAPI_DEF_SOCKADDR_IN6 1
226#endif
227#ifndef __UAPI_DEF_IPV6_MREQ
196#define __UAPI_DEF_IPV6_MREQ 1 228#define __UAPI_DEF_IPV6_MREQ 1
229#endif
230#ifndef __UAPI_DEF_IPPROTO_V6
197#define __UAPI_DEF_IPPROTO_V6 1 231#define __UAPI_DEF_IPPROTO_V6 1
232#endif
233#ifndef __UAPI_DEF_IPV6_OPTIONS
198#define __UAPI_DEF_IPV6_OPTIONS 1 234#define __UAPI_DEF_IPV6_OPTIONS 1
235#endif
236#ifndef __UAPI_DEF_IN6_PKTINFO
199#define __UAPI_DEF_IN6_PKTINFO 1 237#define __UAPI_DEF_IN6_PKTINFO 1
238#endif
239#ifndef __UAPI_DEF_IP6_MTUINFO
200#define __UAPI_DEF_IP6_MTUINFO 1 240#define __UAPI_DEF_IP6_MTUINFO 1
241#endif
201 242
202/* Definitions for ipx.h */ 243/* Definitions for ipx.h */
244#ifndef __UAPI_DEF_SOCKADDR_IPX
203#define __UAPI_DEF_SOCKADDR_IPX 1 245#define __UAPI_DEF_SOCKADDR_IPX 1
246#endif
247#ifndef __UAPI_DEF_IPX_ROUTE_DEFINITION
204#define __UAPI_DEF_IPX_ROUTE_DEFINITION 1 248#define __UAPI_DEF_IPX_ROUTE_DEFINITION 1
249#endif
250#ifndef __UAPI_DEF_IPX_INTERFACE_DEFINITION
205#define __UAPI_DEF_IPX_INTERFACE_DEFINITION 1 251#define __UAPI_DEF_IPX_INTERFACE_DEFINITION 1
252#endif
253#ifndef __UAPI_DEF_IPX_CONFIG_DATA
206#define __UAPI_DEF_IPX_CONFIG_DATA 1 254#define __UAPI_DEF_IPX_CONFIG_DATA 1
255#endif
256#ifndef __UAPI_DEF_IPX_ROUTE_DEF
207#define __UAPI_DEF_IPX_ROUTE_DEF 1 257#define __UAPI_DEF_IPX_ROUTE_DEF 1
258#endif
208 259
209/* Definitions for xattr.h */ 260/* Definitions for xattr.h */
261#ifndef __UAPI_DEF_XATTR
210#define __UAPI_DEF_XATTR 1 262#define __UAPI_DEF_XATTR 1
263#endif
211 264
212#endif /* __GLIBC__ */ 265#endif /* __GLIBC__ */
213 266
267/* Definitions for if_ether.h */
268/* allow libcs like musl to deactivate this, glibc does not implement this. */
269#ifndef __UAPI_DEF_ETHHDR
270#define __UAPI_DEF_ETHHDR 1
271#endif
272
214#endif /* _UAPI_LIBC_COMPAT_H */ 273#endif /* _UAPI_LIBC_COMPAT_H */
diff --git a/include/uapi/linux/netfilter/nf_conntrack_common.h b/include/uapi/linux/netfilter/nf_conntrack_common.h
index 3fea7709a441..57ccfb32e87f 100644
--- a/include/uapi/linux/netfilter/nf_conntrack_common.h
+++ b/include/uapi/linux/netfilter/nf_conntrack_common.h
@@ -36,7 +36,7 @@ enum ip_conntrack_info {
36 36
37#define NF_CT_STATE_INVALID_BIT (1 << 0) 37#define NF_CT_STATE_INVALID_BIT (1 << 0)
38#define NF_CT_STATE_BIT(ctinfo) (1 << ((ctinfo) % IP_CT_IS_REPLY + 1)) 38#define NF_CT_STATE_BIT(ctinfo) (1 << ((ctinfo) % IP_CT_IS_REPLY + 1))
39#define NF_CT_STATE_UNTRACKED_BIT (1 << (IP_CT_UNTRACKED + 1)) 39#define NF_CT_STATE_UNTRACKED_BIT (1 << 6)
40 40
41/* Bitset representing status of connection. */ 41/* Bitset representing status of connection. */
42enum ip_conntrack_status { 42enum ip_conntrack_status {
diff --git a/init/Kconfig b/init/Kconfig
index 690a381adee0..a9a2e2c86671 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -461,6 +461,7 @@ endmenu # "CPU/Task time and stats accounting"
461 461
462config CPU_ISOLATION 462config CPU_ISOLATION
463 bool "CPU isolation" 463 bool "CPU isolation"
464 depends on SMP || COMPILE_TEST
464 default y 465 default y
465 help 466 help
466 Make sure that CPUs running critical tasks are not disturbed by 467 Make sure that CPUs running critical tasks are not disturbed by
@@ -1396,6 +1397,13 @@ config BPF_SYSCALL
1396 Enable the bpf() system call that allows to manipulate eBPF 1397 Enable the bpf() system call that allows to manipulate eBPF
1397 programs and maps via file descriptors. 1398 programs and maps via file descriptors.
1398 1399
1400config BPF_JIT_ALWAYS_ON
1401 bool "Permanently enable BPF JIT and remove BPF interpreter"
1402 depends on BPF_SYSCALL && HAVE_EBPF_JIT && BPF_JIT
1403 help
1404 Enables BPF JIT and removes BPF interpreter to avoid
1405 speculative execution of BPF instructions by the interpreter
1406
1399config USERFAULTFD 1407config USERFAULTFD
1400 bool "Enable userfaultfd() system call" 1408 bool "Enable userfaultfd() system call"
1401 select ANON_INODES 1409 select ANON_INODES
diff --git a/kernel/bpf/arraymap.c b/kernel/bpf/arraymap.c
index 7c25426d3cf5..aaa319848e7d 100644
--- a/kernel/bpf/arraymap.c
+++ b/kernel/bpf/arraymap.c
@@ -53,9 +53,10 @@ static struct bpf_map *array_map_alloc(union bpf_attr *attr)
53{ 53{
54 bool percpu = attr->map_type == BPF_MAP_TYPE_PERCPU_ARRAY; 54 bool percpu = attr->map_type == BPF_MAP_TYPE_PERCPU_ARRAY;
55 int numa_node = bpf_map_attr_numa_node(attr); 55 int numa_node = bpf_map_attr_numa_node(attr);
56 u32 elem_size, index_mask, max_entries;
57 bool unpriv = !capable(CAP_SYS_ADMIN);
56 struct bpf_array *array; 58 struct bpf_array *array;
57 u64 array_size; 59 u64 array_size;
58 u32 elem_size;
59 60
60 /* check sanity of attributes */ 61 /* check sanity of attributes */
61 if (attr->max_entries == 0 || attr->key_size != 4 || 62 if (attr->max_entries == 0 || attr->key_size != 4 ||
@@ -72,11 +73,20 @@ static struct bpf_map *array_map_alloc(union bpf_attr *attr)
72 73
73 elem_size = round_up(attr->value_size, 8); 74 elem_size = round_up(attr->value_size, 8);
74 75
76 max_entries = attr->max_entries;
77 index_mask = roundup_pow_of_two(max_entries) - 1;
78
79 if (unpriv)
80 /* round up array size to nearest power of 2,
81 * since cpu will speculate within index_mask limits
82 */
83 max_entries = index_mask + 1;
84
75 array_size = sizeof(*array); 85 array_size = sizeof(*array);
76 if (percpu) 86 if (percpu)
77 array_size += (u64) attr->max_entries * sizeof(void *); 87 array_size += (u64) max_entries * sizeof(void *);
78 else 88 else
79 array_size += (u64) attr->max_entries * elem_size; 89 array_size += (u64) max_entries * elem_size;
80 90
81 /* make sure there is no u32 overflow later in round_up() */ 91 /* make sure there is no u32 overflow later in round_up() */
82 if (array_size >= U32_MAX - PAGE_SIZE) 92 if (array_size >= U32_MAX - PAGE_SIZE)
@@ -86,6 +96,8 @@ static struct bpf_map *array_map_alloc(union bpf_attr *attr)
86 array = bpf_map_area_alloc(array_size, numa_node); 96 array = bpf_map_area_alloc(array_size, numa_node);
87 if (!array) 97 if (!array)
88 return ERR_PTR(-ENOMEM); 98 return ERR_PTR(-ENOMEM);
99 array->index_mask = index_mask;
100 array->map.unpriv_array = unpriv;
89 101
90 /* copy mandatory map attributes */ 102 /* copy mandatory map attributes */
91 array->map.map_type = attr->map_type; 103 array->map.map_type = attr->map_type;
@@ -121,12 +133,13 @@ static void *array_map_lookup_elem(struct bpf_map *map, void *key)
121 if (unlikely(index >= array->map.max_entries)) 133 if (unlikely(index >= array->map.max_entries))
122 return NULL; 134 return NULL;
123 135
124 return array->value + array->elem_size * index; 136 return array->value + array->elem_size * (index & array->index_mask);
125} 137}
126 138
127/* emit BPF instructions equivalent to C code of array_map_lookup_elem() */ 139/* emit BPF instructions equivalent to C code of array_map_lookup_elem() */
128static u32 array_map_gen_lookup(struct bpf_map *map, struct bpf_insn *insn_buf) 140static u32 array_map_gen_lookup(struct bpf_map *map, struct bpf_insn *insn_buf)
129{ 141{
142 struct bpf_array *array = container_of(map, struct bpf_array, map);
130 struct bpf_insn *insn = insn_buf; 143 struct bpf_insn *insn = insn_buf;
131 u32 elem_size = round_up(map->value_size, 8); 144 u32 elem_size = round_up(map->value_size, 8);
132 const int ret = BPF_REG_0; 145 const int ret = BPF_REG_0;
@@ -135,7 +148,12 @@ static u32 array_map_gen_lookup(struct bpf_map *map, struct bpf_insn *insn_buf)
135 148
136 *insn++ = BPF_ALU64_IMM(BPF_ADD, map_ptr, offsetof(struct bpf_array, value)); 149 *insn++ = BPF_ALU64_IMM(BPF_ADD, map_ptr, offsetof(struct bpf_array, value));
137 *insn++ = BPF_LDX_MEM(BPF_W, ret, index, 0); 150 *insn++ = BPF_LDX_MEM(BPF_W, ret, index, 0);
138 *insn++ = BPF_JMP_IMM(BPF_JGE, ret, map->max_entries, 3); 151 if (map->unpriv_array) {
152 *insn++ = BPF_JMP_IMM(BPF_JGE, ret, map->max_entries, 4);
153 *insn++ = BPF_ALU32_IMM(BPF_AND, ret, array->index_mask);
154 } else {
155 *insn++ = BPF_JMP_IMM(BPF_JGE, ret, map->max_entries, 3);
156 }
139 157
140 if (is_power_of_2(elem_size)) { 158 if (is_power_of_2(elem_size)) {
141 *insn++ = BPF_ALU64_IMM(BPF_LSH, ret, ilog2(elem_size)); 159 *insn++ = BPF_ALU64_IMM(BPF_LSH, ret, ilog2(elem_size));
@@ -157,7 +175,7 @@ static void *percpu_array_map_lookup_elem(struct bpf_map *map, void *key)
157 if (unlikely(index >= array->map.max_entries)) 175 if (unlikely(index >= array->map.max_entries))
158 return NULL; 176 return NULL;
159 177
160 return this_cpu_ptr(array->pptrs[index]); 178 return this_cpu_ptr(array->pptrs[index & array->index_mask]);
161} 179}
162 180
163int bpf_percpu_array_copy(struct bpf_map *map, void *key, void *value) 181int bpf_percpu_array_copy(struct bpf_map *map, void *key, void *value)
@@ -177,7 +195,7 @@ int bpf_percpu_array_copy(struct bpf_map *map, void *key, void *value)
177 */ 195 */
178 size = round_up(map->value_size, 8); 196 size = round_up(map->value_size, 8);
179 rcu_read_lock(); 197 rcu_read_lock();
180 pptr = array->pptrs[index]; 198 pptr = array->pptrs[index & array->index_mask];
181 for_each_possible_cpu(cpu) { 199 for_each_possible_cpu(cpu) {
182 bpf_long_memcpy(value + off, per_cpu_ptr(pptr, cpu), size); 200 bpf_long_memcpy(value + off, per_cpu_ptr(pptr, cpu), size);
183 off += size; 201 off += size;
@@ -225,10 +243,11 @@ static int array_map_update_elem(struct bpf_map *map, void *key, void *value,
225 return -EEXIST; 243 return -EEXIST;
226 244
227 if (array->map.map_type == BPF_MAP_TYPE_PERCPU_ARRAY) 245 if (array->map.map_type == BPF_MAP_TYPE_PERCPU_ARRAY)
228 memcpy(this_cpu_ptr(array->pptrs[index]), 246 memcpy(this_cpu_ptr(array->pptrs[index & array->index_mask]),
229 value, map->value_size); 247 value, map->value_size);
230 else 248 else
231 memcpy(array->value + array->elem_size * index, 249 memcpy(array->value +
250 array->elem_size * (index & array->index_mask),
232 value, map->value_size); 251 value, map->value_size);
233 return 0; 252 return 0;
234} 253}
@@ -262,7 +281,7 @@ int bpf_percpu_array_update(struct bpf_map *map, void *key, void *value,
262 */ 281 */
263 size = round_up(map->value_size, 8); 282 size = round_up(map->value_size, 8);
264 rcu_read_lock(); 283 rcu_read_lock();
265 pptr = array->pptrs[index]; 284 pptr = array->pptrs[index & array->index_mask];
266 for_each_possible_cpu(cpu) { 285 for_each_possible_cpu(cpu) {
267 bpf_long_memcpy(per_cpu_ptr(pptr, cpu), value + off, size); 286 bpf_long_memcpy(per_cpu_ptr(pptr, cpu), value + off, size);
268 off += size; 287 off += size;
@@ -613,6 +632,7 @@ static void *array_of_map_lookup_elem(struct bpf_map *map, void *key)
613static u32 array_of_map_gen_lookup(struct bpf_map *map, 632static u32 array_of_map_gen_lookup(struct bpf_map *map,
614 struct bpf_insn *insn_buf) 633 struct bpf_insn *insn_buf)
615{ 634{
635 struct bpf_array *array = container_of(map, struct bpf_array, map);
616 u32 elem_size = round_up(map->value_size, 8); 636 u32 elem_size = round_up(map->value_size, 8);
617 struct bpf_insn *insn = insn_buf; 637 struct bpf_insn *insn = insn_buf;
618 const int ret = BPF_REG_0; 638 const int ret = BPF_REG_0;
@@ -621,7 +641,12 @@ static u32 array_of_map_gen_lookup(struct bpf_map *map,
621 641
622 *insn++ = BPF_ALU64_IMM(BPF_ADD, map_ptr, offsetof(struct bpf_array, value)); 642 *insn++ = BPF_ALU64_IMM(BPF_ADD, map_ptr, offsetof(struct bpf_array, value));
623 *insn++ = BPF_LDX_MEM(BPF_W, ret, index, 0); 643 *insn++ = BPF_LDX_MEM(BPF_W, ret, index, 0);
624 *insn++ = BPF_JMP_IMM(BPF_JGE, ret, map->max_entries, 5); 644 if (map->unpriv_array) {
645 *insn++ = BPF_JMP_IMM(BPF_JGE, ret, map->max_entries, 6);
646 *insn++ = BPF_ALU32_IMM(BPF_AND, ret, array->index_mask);
647 } else {
648 *insn++ = BPF_JMP_IMM(BPF_JGE, ret, map->max_entries, 5);
649 }
625 if (is_power_of_2(elem_size)) 650 if (is_power_of_2(elem_size))
626 *insn++ = BPF_ALU64_IMM(BPF_LSH, ret, ilog2(elem_size)); 651 *insn++ = BPF_ALU64_IMM(BPF_LSH, ret, ilog2(elem_size));
627 else 652 else
diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
index 86b50aa26ee8..51ec2dda7f08 100644
--- a/kernel/bpf/core.c
+++ b/kernel/bpf/core.c
@@ -767,6 +767,7 @@ noinline u64 __bpf_call_base(u64 r1, u64 r2, u64 r3, u64 r4, u64 r5)
767} 767}
768EXPORT_SYMBOL_GPL(__bpf_call_base); 768EXPORT_SYMBOL_GPL(__bpf_call_base);
769 769
770#ifndef CONFIG_BPF_JIT_ALWAYS_ON
770/** 771/**
771 * __bpf_prog_run - run eBPF program on a given context 772 * __bpf_prog_run - run eBPF program on a given context
772 * @ctx: is the data we are operating on 773 * @ctx: is the data we are operating on
@@ -1317,6 +1318,14 @@ EVAL6(PROG_NAME_LIST, 224, 256, 288, 320, 352, 384)
1317EVAL4(PROG_NAME_LIST, 416, 448, 480, 512) 1318EVAL4(PROG_NAME_LIST, 416, 448, 480, 512)
1318}; 1319};
1319 1320
1321#else
1322static unsigned int __bpf_prog_ret0(const void *ctx,
1323 const struct bpf_insn *insn)
1324{
1325 return 0;
1326}
1327#endif
1328
1320bool bpf_prog_array_compatible(struct bpf_array *array, 1329bool bpf_prog_array_compatible(struct bpf_array *array,
1321 const struct bpf_prog *fp) 1330 const struct bpf_prog *fp)
1322{ 1331{
@@ -1364,9 +1373,13 @@ static int bpf_check_tail_call(const struct bpf_prog *fp)
1364 */ 1373 */
1365struct bpf_prog *bpf_prog_select_runtime(struct bpf_prog *fp, int *err) 1374struct bpf_prog *bpf_prog_select_runtime(struct bpf_prog *fp, int *err)
1366{ 1375{
1376#ifndef CONFIG_BPF_JIT_ALWAYS_ON
1367 u32 stack_depth = max_t(u32, fp->aux->stack_depth, 1); 1377 u32 stack_depth = max_t(u32, fp->aux->stack_depth, 1);
1368 1378
1369 fp->bpf_func = interpreters[(round_up(stack_depth, 32) / 32) - 1]; 1379 fp->bpf_func = interpreters[(round_up(stack_depth, 32) / 32) - 1];
1380#else
1381 fp->bpf_func = __bpf_prog_ret0;
1382#endif
1370 1383
1371 /* eBPF JITs can rewrite the program in case constant 1384 /* eBPF JITs can rewrite the program in case constant
1372 * blinding is active. However, in case of error during 1385 * blinding is active. However, in case of error during
@@ -1376,6 +1389,12 @@ struct bpf_prog *bpf_prog_select_runtime(struct bpf_prog *fp, int *err)
1376 */ 1389 */
1377 if (!bpf_prog_is_dev_bound(fp->aux)) { 1390 if (!bpf_prog_is_dev_bound(fp->aux)) {
1378 fp = bpf_int_jit_compile(fp); 1391 fp = bpf_int_jit_compile(fp);
1392#ifdef CONFIG_BPF_JIT_ALWAYS_ON
1393 if (!fp->jited) {
1394 *err = -ENOTSUPP;
1395 return fp;
1396 }
1397#endif
1379 } else { 1398 } else {
1380 *err = bpf_prog_offload_compile(fp); 1399 *err = bpf_prog_offload_compile(fp);
1381 if (*err) 1400 if (*err)
diff --git a/kernel/bpf/sockmap.c b/kernel/bpf/sockmap.c
index 5ee2e41893d9..1712d319c2d8 100644
--- a/kernel/bpf/sockmap.c
+++ b/kernel/bpf/sockmap.c
@@ -591,8 +591,15 @@ static void sock_map_free(struct bpf_map *map)
591 591
592 write_lock_bh(&sock->sk_callback_lock); 592 write_lock_bh(&sock->sk_callback_lock);
593 psock = smap_psock_sk(sock); 593 psock = smap_psock_sk(sock);
594 smap_list_remove(psock, &stab->sock_map[i]); 594 /* This check handles a racing sock event that can get the
595 smap_release_sock(psock, sock); 595 * sk_callback_lock before this case but after xchg happens
596 * causing the refcnt to hit zero and sock user data (psock)
597 * to be null and queued for garbage collection.
598 */
599 if (likely(psock)) {
600 smap_list_remove(psock, &stab->sock_map[i]);
601 smap_release_sock(psock, sock);
602 }
596 write_unlock_bh(&sock->sk_callback_lock); 603 write_unlock_bh(&sock->sk_callback_lock);
597 } 604 }
598 rcu_read_unlock(); 605 rcu_read_unlock();
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index 04b24876cd23..b414d6b2d470 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -1729,6 +1729,13 @@ static int check_call(struct bpf_verifier_env *env, int func_id, int insn_idx)
1729 err = check_func_arg(env, BPF_REG_2, fn->arg2_type, &meta); 1729 err = check_func_arg(env, BPF_REG_2, fn->arg2_type, &meta);
1730 if (err) 1730 if (err)
1731 return err; 1731 return err;
1732 if (func_id == BPF_FUNC_tail_call) {
1733 if (meta.map_ptr == NULL) {
1734 verbose(env, "verifier bug\n");
1735 return -EINVAL;
1736 }
1737 env->insn_aux_data[insn_idx].map_ptr = meta.map_ptr;
1738 }
1732 err = check_func_arg(env, BPF_REG_3, fn->arg3_type, &meta); 1739 err = check_func_arg(env, BPF_REG_3, fn->arg3_type, &meta);
1733 if (err) 1740 if (err)
1734 return err; 1741 return err;
@@ -4456,6 +4463,35 @@ static int fixup_bpf_calls(struct bpf_verifier_env *env)
4456 */ 4463 */
4457 insn->imm = 0; 4464 insn->imm = 0;
4458 insn->code = BPF_JMP | BPF_TAIL_CALL; 4465 insn->code = BPF_JMP | BPF_TAIL_CALL;
4466
4467 /* instead of changing every JIT dealing with tail_call
4468 * emit two extra insns:
4469 * if (index >= max_entries) goto out;
4470 * index &= array->index_mask;
4471 * to avoid out-of-bounds cpu speculation
4472 */
4473 map_ptr = env->insn_aux_data[i + delta].map_ptr;
4474 if (map_ptr == BPF_MAP_PTR_POISON) {
4475 verbose(env, "tail_call obusing map_ptr\n");
4476 return -EINVAL;
4477 }
4478 if (!map_ptr->unpriv_array)
4479 continue;
4480 insn_buf[0] = BPF_JMP_IMM(BPF_JGE, BPF_REG_3,
4481 map_ptr->max_entries, 2);
4482 insn_buf[1] = BPF_ALU32_IMM(BPF_AND, BPF_REG_3,
4483 container_of(map_ptr,
4484 struct bpf_array,
4485 map)->index_mask);
4486 insn_buf[2] = *insn;
4487 cnt = 3;
4488 new_prog = bpf_patch_insn_data(env, i + delta, insn_buf, cnt);
4489 if (!new_prog)
4490 return -ENOMEM;
4491
4492 delta += cnt - 1;
4493 env->prog = prog = new_prog;
4494 insn = new_prog->insnsi + i + delta;
4459 continue; 4495 continue;
4460 } 4496 }
4461 4497
diff --git a/kernel/cgroup/cgroup-v1.c b/kernel/cgroup/cgroup-v1.c
index 024085daab1a..a2c05d2476ac 100644
--- a/kernel/cgroup/cgroup-v1.c
+++ b/kernel/cgroup/cgroup-v1.c
@@ -123,7 +123,11 @@ int cgroup_transfer_tasks(struct cgroup *to, struct cgroup *from)
123 */ 123 */
124 do { 124 do {
125 css_task_iter_start(&from->self, 0, &it); 125 css_task_iter_start(&from->self, 0, &it);
126 task = css_task_iter_next(&it); 126
127 do {
128 task = css_task_iter_next(&it);
129 } while (task && (task->flags & PF_EXITING));
130
127 if (task) 131 if (task)
128 get_task_struct(task); 132 get_task_struct(task);
129 css_task_iter_end(&it); 133 css_task_iter_end(&it);
diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c
index 0b1ffe147f24..2cf06c274e4c 100644
--- a/kernel/cgroup/cgroup.c
+++ b/kernel/cgroup/cgroup.c
@@ -1397,7 +1397,7 @@ static char *cgroup_file_name(struct cgroup *cgrp, const struct cftype *cft,
1397 cgroup_on_dfl(cgrp) ? ss->name : ss->legacy_name, 1397 cgroup_on_dfl(cgrp) ? ss->name : ss->legacy_name,
1398 cft->name); 1398 cft->name);
1399 else 1399 else
1400 strncpy(buf, cft->name, CGROUP_FILE_NAME_MAX); 1400 strlcpy(buf, cft->name, CGROUP_FILE_NAME_MAX);
1401 return buf; 1401 return buf;
1402} 1402}
1403 1403
@@ -1864,9 +1864,9 @@ void init_cgroup_root(struct cgroup_root *root, struct cgroup_sb_opts *opts)
1864 1864
1865 root->flags = opts->flags; 1865 root->flags = opts->flags;
1866 if (opts->release_agent) 1866 if (opts->release_agent)
1867 strcpy(root->release_agent_path, opts->release_agent); 1867 strlcpy(root->release_agent_path, opts->release_agent, PATH_MAX);
1868 if (opts->name) 1868 if (opts->name)
1869 strcpy(root->name, opts->name); 1869 strlcpy(root->name, opts->name, MAX_CGROUP_ROOT_NAMELEN);
1870 if (opts->cpuset_clone_children) 1870 if (opts->cpuset_clone_children)
1871 set_bit(CGRP_CPUSET_CLONE_CHILDREN, &root->cgrp.flags); 1871 set_bit(CGRP_CPUSET_CLONE_CHILDREN, &root->cgrp.flags);
1872} 1872}
@@ -4125,26 +4125,24 @@ static void css_task_iter_advance_css_set(struct css_task_iter *it)
4125 4125
4126static void css_task_iter_advance(struct css_task_iter *it) 4126static void css_task_iter_advance(struct css_task_iter *it)
4127{ 4127{
4128 struct list_head *l = it->task_pos; 4128 struct list_head *next;
4129 4129
4130 lockdep_assert_held(&css_set_lock); 4130 lockdep_assert_held(&css_set_lock);
4131 WARN_ON_ONCE(!l);
4132
4133repeat: 4131repeat:
4134 /* 4132 /*
4135 * Advance iterator to find next entry. cset->tasks is consumed 4133 * Advance iterator to find next entry. cset->tasks is consumed
4136 * first and then ->mg_tasks. After ->mg_tasks, we move onto the 4134 * first and then ->mg_tasks. After ->mg_tasks, we move onto the
4137 * next cset. 4135 * next cset.
4138 */ 4136 */
4139 l = l->next; 4137 next = it->task_pos->next;
4140 4138
4141 if (l == it->tasks_head) 4139 if (next == it->tasks_head)
4142 l = it->mg_tasks_head->next; 4140 next = it->mg_tasks_head->next;
4143 4141
4144 if (l == it->mg_tasks_head) 4142 if (next == it->mg_tasks_head)
4145 css_task_iter_advance_css_set(it); 4143 css_task_iter_advance_css_set(it);
4146 else 4144 else
4147 it->task_pos = l; 4145 it->task_pos = next;
4148 4146
4149 /* if PROCS, skip over tasks which aren't group leaders */ 4147 /* if PROCS, skip over tasks which aren't group leaders */
4150 if ((it->flags & CSS_TASK_ITER_PROCS) && it->task_pos && 4148 if ((it->flags & CSS_TASK_ITER_PROCS) && it->task_pos &&
diff --git a/kernel/crash_core.c b/kernel/crash_core.c
index b3663896278e..4f63597c824d 100644
--- a/kernel/crash_core.c
+++ b/kernel/crash_core.c
@@ -410,7 +410,7 @@ static int __init crash_save_vmcoreinfo_init(void)
410 VMCOREINFO_SYMBOL(contig_page_data); 410 VMCOREINFO_SYMBOL(contig_page_data);
411#endif 411#endif
412#ifdef CONFIG_SPARSEMEM 412#ifdef CONFIG_SPARSEMEM
413 VMCOREINFO_SYMBOL(mem_section); 413 VMCOREINFO_SYMBOL_ARRAY(mem_section);
414 VMCOREINFO_LENGTH(mem_section, NR_SECTION_ROOTS); 414 VMCOREINFO_LENGTH(mem_section, NR_SECTION_ROOTS);
415 VMCOREINFO_STRUCT_SIZE(mem_section); 415 VMCOREINFO_STRUCT_SIZE(mem_section);
416 VMCOREINFO_OFFSET(mem_section, section_mem_map); 416 VMCOREINFO_OFFSET(mem_section, section_mem_map);
diff --git a/kernel/sched/completion.c b/kernel/sched/completion.c
index 2ddaec40956f..0926aef10dad 100644
--- a/kernel/sched/completion.c
+++ b/kernel/sched/completion.c
@@ -34,11 +34,6 @@ void complete(struct completion *x)
34 34
35 spin_lock_irqsave(&x->wait.lock, flags); 35 spin_lock_irqsave(&x->wait.lock, flags);
36 36
37 /*
38 * Perform commit of crossrelease here.
39 */
40 complete_release_commit(x);
41
42 if (x->done != UINT_MAX) 37 if (x->done != UINT_MAX)
43 x->done++; 38 x->done++;
44 __wake_up_locked(&x->wait, TASK_NORMAL, 1); 39 __wake_up_locked(&x->wait, TASK_NORMAL, 1);
diff --git a/kernel/sched/membarrier.c b/kernel/sched/membarrier.c
index dd7908743dab..9bcbacba82a8 100644
--- a/kernel/sched/membarrier.c
+++ b/kernel/sched/membarrier.c
@@ -89,7 +89,9 @@ static int membarrier_private_expedited(void)
89 rcu_read_unlock(); 89 rcu_read_unlock();
90 } 90 }
91 if (!fallback) { 91 if (!fallback) {
92 preempt_disable();
92 smp_call_function_many(tmpmask, ipi_mb, NULL, 1); 93 smp_call_function_many(tmpmask, ipi_mb, NULL, 1);
94 preempt_enable();
93 free_cpumask_var(tmpmask); 95 free_cpumask_var(tmpmask);
94 } 96 }
95 cpus_read_unlock(); 97 cpus_read_unlock();
diff --git a/lib/test_bpf.c b/lib/test_bpf.c
index 9e9748089270..f369889e521d 100644
--- a/lib/test_bpf.c
+++ b/lib/test_bpf.c
@@ -6250,9 +6250,8 @@ static struct bpf_prog *generate_filter(int which, int *err)
6250 return NULL; 6250 return NULL;
6251 } 6251 }
6252 } 6252 }
6253 /* We don't expect to fail. */
6254 if (*err) { 6253 if (*err) {
6255 pr_cont("FAIL to attach err=%d len=%d\n", 6254 pr_cont("FAIL to prog_create err=%d len=%d\n",
6256 *err, fprog.len); 6255 *err, fprog.len);
6257 return NULL; 6256 return NULL;
6258 } 6257 }
@@ -6276,6 +6275,10 @@ static struct bpf_prog *generate_filter(int which, int *err)
6276 * checks. 6275 * checks.
6277 */ 6276 */
6278 fp = bpf_prog_select_runtime(fp, err); 6277 fp = bpf_prog_select_runtime(fp, err);
6278 if (*err) {
6279 pr_cont("FAIL to select_runtime err=%d\n", *err);
6280 return NULL;
6281 }
6279 break; 6282 break;
6280 } 6283 }
6281 6284
@@ -6461,8 +6464,8 @@ static __init int test_bpf(void)
6461 pass_cnt++; 6464 pass_cnt++;
6462 continue; 6465 continue;
6463 } 6466 }
6464 6467 err_cnt++;
6465 return err; 6468 continue;
6466 } 6469 }
6467 6470
6468 pr_cont("jited:%u ", fp->jited); 6471 pr_cont("jited:%u ", fp->jited);
diff --git a/mm/kmemleak.c b/mm/kmemleak.c
index d73c14294f3a..f656ca27f6c2 100644
--- a/mm/kmemleak.c
+++ b/mm/kmemleak.c
@@ -127,7 +127,7 @@
127/* GFP bitmask for kmemleak internal allocations */ 127/* GFP bitmask for kmemleak internal allocations */
128#define gfp_kmemleak_mask(gfp) (((gfp) & (GFP_KERNEL | GFP_ATOMIC)) | \ 128#define gfp_kmemleak_mask(gfp) (((gfp) & (GFP_KERNEL | GFP_ATOMIC)) | \
129 __GFP_NORETRY | __GFP_NOMEMALLOC | \ 129 __GFP_NORETRY | __GFP_NOMEMALLOC | \
130 __GFP_NOWARN) 130 __GFP_NOWARN | __GFP_NOFAIL)
131 131
132/* scanning area inside a memory block */ 132/* scanning area inside a memory block */
133struct kmemleak_scan_area { 133struct kmemleak_scan_area {
diff --git a/net/8021q/vlan.c b/net/8021q/vlan.c
index 8dfdd94e430f..bad01b14a4ad 100644
--- a/net/8021q/vlan.c
+++ b/net/8021q/vlan.c
@@ -111,12 +111,7 @@ void unregister_vlan_dev(struct net_device *dev, struct list_head *head)
111 vlan_gvrp_uninit_applicant(real_dev); 111 vlan_gvrp_uninit_applicant(real_dev);
112 } 112 }
113 113
114 /* Take it out of our own structures, but be sure to interlock with 114 vlan_vid_del(real_dev, vlan->vlan_proto, vlan_id);
115 * HW accelerating devices or SW vlan input packet processing if
116 * VLAN is not 0 (leave it there for 802.1p).
117 */
118 if (vlan_id)
119 vlan_vid_del(real_dev, vlan->vlan_proto, vlan_id);
120 115
121 /* Get rid of the vlan's reference to real_dev */ 116 /* Get rid of the vlan's reference to real_dev */
122 dev_put(real_dev); 117 dev_put(real_dev);
diff --git a/net/bluetooth/l2cap_core.c b/net/bluetooth/l2cap_core.c
index 43ba91c440bc..fc6615d59165 100644
--- a/net/bluetooth/l2cap_core.c
+++ b/net/bluetooth/l2cap_core.c
@@ -3363,9 +3363,10 @@ static int l2cap_parse_conf_req(struct l2cap_chan *chan, void *data, size_t data
3363 break; 3363 break;
3364 3364
3365 case L2CAP_CONF_EFS: 3365 case L2CAP_CONF_EFS:
3366 remote_efs = 1; 3366 if (olen == sizeof(efs)) {
3367 if (olen == sizeof(efs)) 3367 remote_efs = 1;
3368 memcpy(&efs, (void *) val, olen); 3368 memcpy(&efs, (void *) val, olen);
3369 }
3369 break; 3370 break;
3370 3371
3371 case L2CAP_CONF_EWS: 3372 case L2CAP_CONF_EWS:
@@ -3584,16 +3585,17 @@ static int l2cap_parse_conf_rsp(struct l2cap_chan *chan, void *rsp, int len,
3584 break; 3585 break;
3585 3586
3586 case L2CAP_CONF_EFS: 3587 case L2CAP_CONF_EFS:
3587 if (olen == sizeof(efs)) 3588 if (olen == sizeof(efs)) {
3588 memcpy(&efs, (void *)val, olen); 3589 memcpy(&efs, (void *)val, olen);
3589 3590
3590 if (chan->local_stype != L2CAP_SERV_NOTRAFIC && 3591 if (chan->local_stype != L2CAP_SERV_NOTRAFIC &&
3591 efs.stype != L2CAP_SERV_NOTRAFIC && 3592 efs.stype != L2CAP_SERV_NOTRAFIC &&
3592 efs.stype != chan->local_stype) 3593 efs.stype != chan->local_stype)
3593 return -ECONNREFUSED; 3594 return -ECONNREFUSED;
3594 3595
3595 l2cap_add_conf_opt(&ptr, L2CAP_CONF_EFS, sizeof(efs), 3596 l2cap_add_conf_opt(&ptr, L2CAP_CONF_EFS, sizeof(efs),
3596 (unsigned long) &efs, endptr - ptr); 3597 (unsigned long) &efs, endptr - ptr);
3598 }
3597 break; 3599 break;
3598 3600
3599 case L2CAP_CONF_FCS: 3601 case L2CAP_CONF_FCS:
diff --git a/net/caif/caif_dev.c b/net/caif/caif_dev.c
index 2d38b6e34203..e0adcd123f48 100644
--- a/net/caif/caif_dev.c
+++ b/net/caif/caif_dev.c
@@ -334,9 +334,8 @@ void caif_enroll_dev(struct net_device *dev, struct caif_dev_common *caifdev,
334 mutex_lock(&caifdevs->lock); 334 mutex_lock(&caifdevs->lock);
335 list_add_rcu(&caifd->list, &caifdevs->list); 335 list_add_rcu(&caifd->list, &caifdevs->list);
336 336
337 strncpy(caifd->layer.name, dev->name, 337 strlcpy(caifd->layer.name, dev->name,
338 sizeof(caifd->layer.name) - 1); 338 sizeof(caifd->layer.name));
339 caifd->layer.name[sizeof(caifd->layer.name) - 1] = 0;
340 caifd->layer.transmit = transmit; 339 caifd->layer.transmit = transmit;
341 cfcnfg_add_phy_layer(cfg, 340 cfcnfg_add_phy_layer(cfg,
342 dev, 341 dev,
diff --git a/net/caif/caif_usb.c b/net/caif/caif_usb.c
index 5cd44f001f64..1a082a946045 100644
--- a/net/caif/caif_usb.c
+++ b/net/caif/caif_usb.c
@@ -176,9 +176,7 @@ static int cfusbl_device_notify(struct notifier_block *me, unsigned long what,
176 dev_add_pack(&caif_usb_type); 176 dev_add_pack(&caif_usb_type);
177 pack_added = true; 177 pack_added = true;
178 178
179 strncpy(layer->name, dev->name, 179 strlcpy(layer->name, dev->name, sizeof(layer->name));
180 sizeof(layer->name) - 1);
181 layer->name[sizeof(layer->name) - 1] = 0;
182 180
183 return 0; 181 return 0;
184} 182}
diff --git a/net/caif/cfcnfg.c b/net/caif/cfcnfg.c
index 273cb07f57d8..8f00bea093b9 100644
--- a/net/caif/cfcnfg.c
+++ b/net/caif/cfcnfg.c
@@ -268,17 +268,15 @@ static int caif_connect_req_to_link_param(struct cfcnfg *cnfg,
268 case CAIFPROTO_RFM: 268 case CAIFPROTO_RFM:
269 l->linktype = CFCTRL_SRV_RFM; 269 l->linktype = CFCTRL_SRV_RFM;
270 l->u.datagram.connid = s->sockaddr.u.rfm.connection_id; 270 l->u.datagram.connid = s->sockaddr.u.rfm.connection_id;
271 strncpy(l->u.rfm.volume, s->sockaddr.u.rfm.volume, 271 strlcpy(l->u.rfm.volume, s->sockaddr.u.rfm.volume,
272 sizeof(l->u.rfm.volume)-1); 272 sizeof(l->u.rfm.volume));
273 l->u.rfm.volume[sizeof(l->u.rfm.volume)-1] = 0;
274 break; 273 break;
275 case CAIFPROTO_UTIL: 274 case CAIFPROTO_UTIL:
276 l->linktype = CFCTRL_SRV_UTIL; 275 l->linktype = CFCTRL_SRV_UTIL;
277 l->endpoint = 0x00; 276 l->endpoint = 0x00;
278 l->chtype = 0x00; 277 l->chtype = 0x00;
279 strncpy(l->u.utility.name, s->sockaddr.u.util.service, 278 strlcpy(l->u.utility.name, s->sockaddr.u.util.service,
280 sizeof(l->u.utility.name)-1); 279 sizeof(l->u.utility.name));
281 l->u.utility.name[sizeof(l->u.utility.name)-1] = 0;
282 caif_assert(sizeof(l->u.utility.name) > 10); 280 caif_assert(sizeof(l->u.utility.name) > 10);
283 l->u.utility.paramlen = s->param.size; 281 l->u.utility.paramlen = s->param.size;
284 if (l->u.utility.paramlen > sizeof(l->u.utility.params)) 282 if (l->u.utility.paramlen > sizeof(l->u.utility.params))
diff --git a/net/caif/cfctrl.c b/net/caif/cfctrl.c
index f5afda1abc76..655ed7032150 100644
--- a/net/caif/cfctrl.c
+++ b/net/caif/cfctrl.c
@@ -258,8 +258,8 @@ int cfctrl_linkup_request(struct cflayer *layer,
258 tmp16 = cpu_to_le16(param->u.utility.fifosize_bufs); 258 tmp16 = cpu_to_le16(param->u.utility.fifosize_bufs);
259 cfpkt_add_body(pkt, &tmp16, 2); 259 cfpkt_add_body(pkt, &tmp16, 2);
260 memset(utility_name, 0, sizeof(utility_name)); 260 memset(utility_name, 0, sizeof(utility_name));
261 strncpy(utility_name, param->u.utility.name, 261 strlcpy(utility_name, param->u.utility.name,
262 UTILITY_NAME_LENGTH - 1); 262 UTILITY_NAME_LENGTH);
263 cfpkt_add_body(pkt, utility_name, UTILITY_NAME_LENGTH); 263 cfpkt_add_body(pkt, utility_name, UTILITY_NAME_LENGTH);
264 tmp8 = param->u.utility.paramlen; 264 tmp8 = param->u.utility.paramlen;
265 cfpkt_add_body(pkt, &tmp8, 1); 265 cfpkt_add_body(pkt, &tmp8, 1);
diff --git a/net/core/dev.c b/net/core/dev.c
index 01ee854454a8..0e0ba36eeac9 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -1146,7 +1146,19 @@ EXPORT_SYMBOL(dev_alloc_name);
1146int dev_get_valid_name(struct net *net, struct net_device *dev, 1146int dev_get_valid_name(struct net *net, struct net_device *dev,
1147 const char *name) 1147 const char *name)
1148{ 1148{
1149 return dev_alloc_name_ns(net, dev, name); 1149 BUG_ON(!net);
1150
1151 if (!dev_valid_name(name))
1152 return -EINVAL;
1153
1154 if (strchr(name, '%'))
1155 return dev_alloc_name_ns(net, dev, name);
1156 else if (__dev_get_by_name(net, name))
1157 return -EEXIST;
1158 else if (dev->name != name)
1159 strlcpy(dev->name, name, IFNAMSIZ);
1160
1161 return 0;
1150} 1162}
1151EXPORT_SYMBOL(dev_get_valid_name); 1163EXPORT_SYMBOL(dev_get_valid_name);
1152 1164
diff --git a/net/core/ethtool.c b/net/core/ethtool.c
index f8fcf450a36e..8225416911ae 100644
--- a/net/core/ethtool.c
+++ b/net/core/ethtool.c
@@ -770,15 +770,6 @@ static int ethtool_set_link_ksettings(struct net_device *dev,
770 return dev->ethtool_ops->set_link_ksettings(dev, &link_ksettings); 770 return dev->ethtool_ops->set_link_ksettings(dev, &link_ksettings);
771} 771}
772 772
773static void
774warn_incomplete_ethtool_legacy_settings_conversion(const char *details)
775{
776 char name[sizeof(current->comm)];
777
778 pr_info_once("warning: `%s' uses legacy ethtool link settings API, %s\n",
779 get_task_comm(name, current), details);
780}
781
782/* Query device for its ethtool_cmd settings. 773/* Query device for its ethtool_cmd settings.
783 * 774 *
784 * Backward compatibility note: for compatibility with legacy ethtool, 775 * Backward compatibility note: for compatibility with legacy ethtool,
@@ -805,10 +796,8 @@ static int ethtool_get_settings(struct net_device *dev, void __user *useraddr)
805 &link_ksettings); 796 &link_ksettings);
806 if (err < 0) 797 if (err < 0)
807 return err; 798 return err;
808 if (!convert_link_ksettings_to_legacy_settings(&cmd, 799 convert_link_ksettings_to_legacy_settings(&cmd,
809 &link_ksettings)) 800 &link_ksettings);
810 warn_incomplete_ethtool_legacy_settings_conversion(
811 "link modes are only partially reported");
812 801
813 /* send a sensible cmd tag back to user */ 802 /* send a sensible cmd tag back to user */
814 cmd.cmd = ETHTOOL_GSET; 803 cmd.cmd = ETHTOOL_GSET;
diff --git a/net/core/filter.c b/net/core/filter.c
index 6a85e67fafce..d339ef170df6 100644
--- a/net/core/filter.c
+++ b/net/core/filter.c
@@ -1054,11 +1054,9 @@ static struct bpf_prog *bpf_migrate_filter(struct bpf_prog *fp)
1054 */ 1054 */
1055 goto out_err_free; 1055 goto out_err_free;
1056 1056
1057 /* We are guaranteed to never error here with cBPF to eBPF
1058 * transitions, since there's no issue with type compatibility
1059 * checks on program arrays.
1060 */
1061 fp = bpf_prog_select_runtime(fp, &err); 1057 fp = bpf_prog_select_runtime(fp, &err);
1058 if (err)
1059 goto out_err_free;
1062 1060
1063 kfree(old_prog); 1061 kfree(old_prog);
1064 return fp; 1062 return fp;
diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
index dabba2a91fc8..778d7f03404a 100644
--- a/net/core/rtnetlink.c
+++ b/net/core/rtnetlink.c
@@ -1681,18 +1681,18 @@ static bool link_dump_filtered(struct net_device *dev,
1681 return false; 1681 return false;
1682} 1682}
1683 1683
1684static struct net *get_target_net(struct sk_buff *skb, int netnsid) 1684static struct net *get_target_net(struct sock *sk, int netnsid)
1685{ 1685{
1686 struct net *net; 1686 struct net *net;
1687 1687
1688 net = get_net_ns_by_id(sock_net(skb->sk), netnsid); 1688 net = get_net_ns_by_id(sock_net(sk), netnsid);
1689 if (!net) 1689 if (!net)
1690 return ERR_PTR(-EINVAL); 1690 return ERR_PTR(-EINVAL);
1691 1691
1692 /* For now, the caller is required to have CAP_NET_ADMIN in 1692 /* For now, the caller is required to have CAP_NET_ADMIN in
1693 * the user namespace owning the target net ns. 1693 * the user namespace owning the target net ns.
1694 */ 1694 */
1695 if (!netlink_ns_capable(skb, net->user_ns, CAP_NET_ADMIN)) { 1695 if (!sk_ns_capable(sk, net->user_ns, CAP_NET_ADMIN)) {
1696 put_net(net); 1696 put_net(net);
1697 return ERR_PTR(-EACCES); 1697 return ERR_PTR(-EACCES);
1698 } 1698 }
@@ -1733,7 +1733,7 @@ static int rtnl_dump_ifinfo(struct sk_buff *skb, struct netlink_callback *cb)
1733 ifla_policy, NULL) >= 0) { 1733 ifla_policy, NULL) >= 0) {
1734 if (tb[IFLA_IF_NETNSID]) { 1734 if (tb[IFLA_IF_NETNSID]) {
1735 netnsid = nla_get_s32(tb[IFLA_IF_NETNSID]); 1735 netnsid = nla_get_s32(tb[IFLA_IF_NETNSID]);
1736 tgt_net = get_target_net(skb, netnsid); 1736 tgt_net = get_target_net(skb->sk, netnsid);
1737 if (IS_ERR(tgt_net)) { 1737 if (IS_ERR(tgt_net)) {
1738 tgt_net = net; 1738 tgt_net = net;
1739 netnsid = -1; 1739 netnsid = -1;
@@ -2883,7 +2883,7 @@ static int rtnl_getlink(struct sk_buff *skb, struct nlmsghdr *nlh,
2883 2883
2884 if (tb[IFLA_IF_NETNSID]) { 2884 if (tb[IFLA_IF_NETNSID]) {
2885 netnsid = nla_get_s32(tb[IFLA_IF_NETNSID]); 2885 netnsid = nla_get_s32(tb[IFLA_IF_NETNSID]);
2886 tgt_net = get_target_net(skb, netnsid); 2886 tgt_net = get_target_net(NETLINK_CB(skb).sk, netnsid);
2887 if (IS_ERR(tgt_net)) 2887 if (IS_ERR(tgt_net))
2888 return PTR_ERR(tgt_net); 2888 return PTR_ERR(tgt_net);
2889 } 2889 }
diff --git a/net/core/sock_diag.c b/net/core/sock_diag.c
index 217f4e3b82f6..146b50e30659 100644
--- a/net/core/sock_diag.c
+++ b/net/core/sock_diag.c
@@ -288,7 +288,7 @@ static int sock_diag_bind(struct net *net, int group)
288 case SKNLGRP_INET6_UDP_DESTROY: 288 case SKNLGRP_INET6_UDP_DESTROY:
289 if (!sock_diag_handlers[AF_INET6]) 289 if (!sock_diag_handlers[AF_INET6])
290 request_module("net-pf-%d-proto-%d-type-%d", PF_NETLINK, 290 request_module("net-pf-%d-proto-%d-type-%d", PF_NETLINK,
291 NETLINK_SOCK_DIAG, AF_INET); 291 NETLINK_SOCK_DIAG, AF_INET6);
292 break; 292 break;
293 } 293 }
294 return 0; 294 return 0;
diff --git a/net/core/sysctl_net_core.c b/net/core/sysctl_net_core.c
index cbc3dde4cfcc..a47ad6cd41c0 100644
--- a/net/core/sysctl_net_core.c
+++ b/net/core/sysctl_net_core.c
@@ -325,7 +325,13 @@ static struct ctl_table net_core_table[] = {
325 .data = &bpf_jit_enable, 325 .data = &bpf_jit_enable,
326 .maxlen = sizeof(int), 326 .maxlen = sizeof(int),
327 .mode = 0644, 327 .mode = 0644,
328#ifndef CONFIG_BPF_JIT_ALWAYS_ON
328 .proc_handler = proc_dointvec 329 .proc_handler = proc_dointvec
330#else
331 .proc_handler = proc_dointvec_minmax,
332 .extra1 = &one,
333 .extra2 = &one,
334#endif
329 }, 335 },
330# ifdef CONFIG_HAVE_EBPF_JIT 336# ifdef CONFIG_HAVE_EBPF_JIT
331 { 337 {
diff --git a/net/ipv4/raw.c b/net/ipv4/raw.c
index 125c1eab3eaa..5e570aa9e43b 100644
--- a/net/ipv4/raw.c
+++ b/net/ipv4/raw.c
@@ -520,9 +520,11 @@ static int raw_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
520 goto out; 520 goto out;
521 521
522 /* hdrincl should be READ_ONCE(inet->hdrincl) 522 /* hdrincl should be READ_ONCE(inet->hdrincl)
523 * but READ_ONCE() doesn't work with bit fields 523 * but READ_ONCE() doesn't work with bit fields.
524 * Doing this indirectly yields the same result.
524 */ 525 */
525 hdrincl = inet->hdrincl; 526 hdrincl = inet->hdrincl;
527 hdrincl = READ_ONCE(hdrincl);
526 /* 528 /*
527 * Check the flags. 529 * Check the flags.
528 */ 530 */
diff --git a/net/ipv6/exthdrs.c b/net/ipv6/exthdrs.c
index 83bd75713535..bc68eb661970 100644
--- a/net/ipv6/exthdrs.c
+++ b/net/ipv6/exthdrs.c
@@ -925,6 +925,15 @@ static void ipv6_push_rthdr4(struct sk_buff *skb, u8 *proto,
925 sr_phdr->segments[0] = **addr_p; 925 sr_phdr->segments[0] = **addr_p;
926 *addr_p = &sr_ihdr->segments[sr_ihdr->segments_left]; 926 *addr_p = &sr_ihdr->segments[sr_ihdr->segments_left];
927 927
928 if (sr_ihdr->hdrlen > hops * 2) {
929 int tlvs_offset, tlvs_length;
930
931 tlvs_offset = (1 + hops * 2) << 3;
932 tlvs_length = (sr_ihdr->hdrlen - hops * 2) << 3;
933 memcpy((char *)sr_phdr + tlvs_offset,
934 (char *)sr_ihdr + tlvs_offset, tlvs_length);
935 }
936
928#ifdef CONFIG_IPV6_SEG6_HMAC 937#ifdef CONFIG_IPV6_SEG6_HMAC
929 if (sr_has_hmac(sr_phdr)) { 938 if (sr_has_hmac(sr_phdr)) {
930 struct net *net = NULL; 939 struct net *net = NULL;
diff --git a/net/ipv6/ip6_fib.c b/net/ipv6/ip6_fib.c
index f5285f4e1d08..9dcc3924a975 100644
--- a/net/ipv6/ip6_fib.c
+++ b/net/ipv6/ip6_fib.c
@@ -640,6 +640,11 @@ static struct fib6_node *fib6_add_1(struct net *net,
640 if (!(fn->fn_flags & RTN_RTINFO)) { 640 if (!(fn->fn_flags & RTN_RTINFO)) {
641 RCU_INIT_POINTER(fn->leaf, NULL); 641 RCU_INIT_POINTER(fn->leaf, NULL);
642 rt6_release(leaf); 642 rt6_release(leaf);
643 /* remove null_entry in the root node */
644 } else if (fn->fn_flags & RTN_TL_ROOT &&
645 rcu_access_pointer(fn->leaf) ==
646 net->ipv6.ip6_null_entry) {
647 RCU_INIT_POINTER(fn->leaf, NULL);
643 } 648 }
644 649
645 return fn; 650 return fn;
@@ -1241,23 +1246,28 @@ out:
1241 * If fib6_add_1 has cleared the old leaf pointer in the 1246 * If fib6_add_1 has cleared the old leaf pointer in the
1242 * super-tree leaf node we have to find a new one for it. 1247 * super-tree leaf node we have to find a new one for it.
1243 */ 1248 */
1244 struct rt6_info *pn_leaf = rcu_dereference_protected(pn->leaf, 1249 if (pn != fn) {
1245 lockdep_is_held(&table->tb6_lock)); 1250 struct rt6_info *pn_leaf =
1246 if (pn != fn && pn_leaf == rt) { 1251 rcu_dereference_protected(pn->leaf,
1247 pn_leaf = NULL; 1252 lockdep_is_held(&table->tb6_lock));
1248 RCU_INIT_POINTER(pn->leaf, NULL); 1253 if (pn_leaf == rt) {
1249 atomic_dec(&rt->rt6i_ref); 1254 pn_leaf = NULL;
1250 } 1255 RCU_INIT_POINTER(pn->leaf, NULL);
1251 if (pn != fn && !pn_leaf && !(pn->fn_flags & RTN_RTINFO)) { 1256 atomic_dec(&rt->rt6i_ref);
1252 pn_leaf = fib6_find_prefix(info->nl_net, table, pn);
1253#if RT6_DEBUG >= 2
1254 if (!pn_leaf) {
1255 WARN_ON(!pn_leaf);
1256 pn_leaf = info->nl_net->ipv6.ip6_null_entry;
1257 } 1257 }
1258 if (!pn_leaf && !(pn->fn_flags & RTN_RTINFO)) {
1259 pn_leaf = fib6_find_prefix(info->nl_net, table,
1260 pn);
1261#if RT6_DEBUG >= 2
1262 if (!pn_leaf) {
1263 WARN_ON(!pn_leaf);
1264 pn_leaf =
1265 info->nl_net->ipv6.ip6_null_entry;
1266 }
1258#endif 1267#endif
1259 atomic_inc(&pn_leaf->rt6i_ref); 1268 atomic_inc(&pn_leaf->rt6i_ref);
1260 rcu_assign_pointer(pn->leaf, pn_leaf); 1269 rcu_assign_pointer(pn->leaf, pn_leaf);
1270 }
1261 } 1271 }
1262#endif 1272#endif
1263 goto failure; 1273 goto failure;
@@ -1265,13 +1275,17 @@ out:
1265 return err; 1275 return err;
1266 1276
1267failure: 1277failure:
1268 /* fn->leaf could be NULL if fn is an intermediate node and we 1278 /* fn->leaf could be NULL and fib6_repair_tree() needs to be called if:
1269 * failed to add the new route to it in both subtree creation 1279 * 1. fn is an intermediate node and we failed to add the new
1270 * failure and fib6_add_rt2node() failure case. 1280 * route to it in both subtree creation failure and fib6_add_rt2node()
1271 * In both cases, fib6_repair_tree() should be called to fix 1281 * failure case.
1272 * fn->leaf. 1282 * 2. fn is the root node in the table and we fail to add the first
1283 * default route to it.
1273 */ 1284 */
1274 if (fn && !(fn->fn_flags & (RTN_RTINFO|RTN_ROOT))) 1285 if (fn &&
1286 (!(fn->fn_flags & (RTN_RTINFO|RTN_ROOT)) ||
1287 (fn->fn_flags & RTN_TL_ROOT &&
1288 !rcu_access_pointer(fn->leaf))))
1275 fib6_repair_tree(info->nl_net, table, fn); 1289 fib6_repair_tree(info->nl_net, table, fn);
1276 /* Always release dst as dst->__refcnt is guaranteed 1290 /* Always release dst as dst->__refcnt is guaranteed
1277 * to be taken before entering this function 1291 * to be taken before entering this function
@@ -1526,6 +1540,12 @@ static struct fib6_node *fib6_repair_tree(struct net *net,
1526 struct fib6_walker *w; 1540 struct fib6_walker *w;
1527 int iter = 0; 1541 int iter = 0;
1528 1542
1543 /* Set fn->leaf to null_entry for root node. */
1544 if (fn->fn_flags & RTN_TL_ROOT) {
1545 rcu_assign_pointer(fn->leaf, net->ipv6.ip6_null_entry);
1546 return fn;
1547 }
1548
1529 for (;;) { 1549 for (;;) {
1530 struct fib6_node *fn_r = rcu_dereference_protected(fn->right, 1550 struct fib6_node *fn_r = rcu_dereference_protected(fn->right,
1531 lockdep_is_held(&table->tb6_lock)); 1551 lockdep_is_held(&table->tb6_lock));
@@ -1680,10 +1700,15 @@ static void fib6_del_route(struct fib6_table *table, struct fib6_node *fn,
1680 } 1700 }
1681 read_unlock(&net->ipv6.fib6_walker_lock); 1701 read_unlock(&net->ipv6.fib6_walker_lock);
1682 1702
1683 /* If it was last route, expunge its radix tree node */ 1703 /* If it was last route, call fib6_repair_tree() to:
1704 * 1. For root node, put back null_entry as how the table was created.
1705 * 2. For other nodes, expunge its radix tree node.
1706 */
1684 if (!rcu_access_pointer(fn->leaf)) { 1707 if (!rcu_access_pointer(fn->leaf)) {
1685 fn->fn_flags &= ~RTN_RTINFO; 1708 if (!(fn->fn_flags & RTN_TL_ROOT)) {
1686 net->ipv6.rt6_stats->fib_route_nodes--; 1709 fn->fn_flags &= ~RTN_RTINFO;
1710 net->ipv6.rt6_stats->fib_route_nodes--;
1711 }
1687 fn = fib6_repair_tree(net, table, fn); 1712 fn = fib6_repair_tree(net, table, fn);
1688 } 1713 }
1689 1714
diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
index f7dd51c42314..688ba5f7516b 100644
--- a/net/ipv6/ip6_output.c
+++ b/net/ipv6/ip6_output.c
@@ -1735,9 +1735,10 @@ struct sk_buff *ip6_make_skb(struct sock *sk,
1735 cork.base.opt = NULL; 1735 cork.base.opt = NULL;
1736 v6_cork.opt = NULL; 1736 v6_cork.opt = NULL;
1737 err = ip6_setup_cork(sk, &cork, &v6_cork, ipc6, rt, fl6); 1737 err = ip6_setup_cork(sk, &cork, &v6_cork, ipc6, rt, fl6);
1738 if (err) 1738 if (err) {
1739 ip6_cork_release(&cork, &v6_cork);
1739 return ERR_PTR(err); 1740 return ERR_PTR(err);
1740 1741 }
1741 if (ipc6->dontfrag < 0) 1742 if (ipc6->dontfrag < 0)
1742 ipc6->dontfrag = inet6_sk(sk)->dontfrag; 1743 ipc6->dontfrag = inet6_sk(sk)->dontfrag;
1743 1744
diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c
index 931c38f6ff4a..9a7cf355bc8c 100644
--- a/net/ipv6/ip6_tunnel.c
+++ b/net/ipv6/ip6_tunnel.c
@@ -1074,10 +1074,11 @@ int ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev, __u8 dsfield,
1074 memcpy(&fl6->daddr, addr6, sizeof(fl6->daddr)); 1074 memcpy(&fl6->daddr, addr6, sizeof(fl6->daddr));
1075 neigh_release(neigh); 1075 neigh_release(neigh);
1076 } 1076 }
1077 } else if (!(t->parms.flags & 1077 } else if (t->parms.proto != 0 && !(t->parms.flags &
1078 (IP6_TNL_F_USE_ORIG_TCLASS | IP6_TNL_F_USE_ORIG_FWMARK))) { 1078 (IP6_TNL_F_USE_ORIG_TCLASS |
1079 /* enable the cache only only if the routing decision does 1079 IP6_TNL_F_USE_ORIG_FWMARK))) {
1080 * not depend on the current inner header value 1080 /* enable the cache only if neither the outer protocol nor the
1081 * routing decision depends on the current inner header value
1081 */ 1082 */
1082 use_cache = true; 1083 use_cache = true;
1083 } 1084 }
@@ -1676,11 +1677,11 @@ int ip6_tnl_change_mtu(struct net_device *dev, int new_mtu)
1676{ 1677{
1677 struct ip6_tnl *tnl = netdev_priv(dev); 1678 struct ip6_tnl *tnl = netdev_priv(dev);
1678 1679
1679 if (tnl->parms.proto == IPPROTO_IPIP) { 1680 if (tnl->parms.proto == IPPROTO_IPV6) {
1680 if (new_mtu < ETH_MIN_MTU) 1681 if (new_mtu < IPV6_MIN_MTU)
1681 return -EINVAL; 1682 return -EINVAL;
1682 } else { 1683 } else {
1683 if (new_mtu < IPV6_MIN_MTU) 1684 if (new_mtu < ETH_MIN_MTU)
1684 return -EINVAL; 1685 return -EINVAL;
1685 } 1686 }
1686 if (new_mtu > 0xFFF8 - dev->hard_header_len) 1687 if (new_mtu > 0xFFF8 - dev->hard_header_len)
diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c
index 70e9d2ca8bbe..4daafb07602f 100644
--- a/net/mac80211/rx.c
+++ b/net/mac80211/rx.c
@@ -3632,6 +3632,8 @@ static bool ieee80211_accept_frame(struct ieee80211_rx_data *rx)
3632 } 3632 }
3633 return true; 3633 return true;
3634 case NL80211_IFTYPE_MESH_POINT: 3634 case NL80211_IFTYPE_MESH_POINT:
3635 if (ether_addr_equal(sdata->vif.addr, hdr->addr2))
3636 return false;
3635 if (multicast) 3637 if (multicast)
3636 return true; 3638 return true;
3637 return ether_addr_equal(sdata->vif.addr, hdr->addr1); 3639 return ether_addr_equal(sdata->vif.addr, hdr->addr1);
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 10798b357481..07bd4138c84e 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -2072,7 +2072,7 @@ static int nf_tables_dump_rules(struct sk_buff *skb,
2072 continue; 2072 continue;
2073 2073
2074 list_for_each_entry_rcu(chain, &table->chains, list) { 2074 list_for_each_entry_rcu(chain, &table->chains, list) {
2075 if (ctx && ctx->chain[0] && 2075 if (ctx && ctx->chain &&
2076 strcmp(ctx->chain, chain->name) != 0) 2076 strcmp(ctx->chain, chain->name) != 0)
2077 continue; 2077 continue;
2078 2078
@@ -4665,8 +4665,10 @@ static int nf_tables_dump_obj_done(struct netlink_callback *cb)
4665{ 4665{
4666 struct nft_obj_filter *filter = cb->data; 4666 struct nft_obj_filter *filter = cb->data;
4667 4667
4668 kfree(filter->table); 4668 if (filter) {
4669 kfree(filter); 4669 kfree(filter->table);
4670 kfree(filter);
4671 }
4670 4672
4671 return 0; 4673 return 0;
4672} 4674}
diff --git a/net/rds/rdma.c b/net/rds/rdma.c
index bc2f1e0977d6..634cfcb7bba6 100644
--- a/net/rds/rdma.c
+++ b/net/rds/rdma.c
@@ -525,6 +525,9 @@ int rds_rdma_extra_size(struct rds_rdma_args *args)
525 525
526 local_vec = (struct rds_iovec __user *)(unsigned long) args->local_vec_addr; 526 local_vec = (struct rds_iovec __user *)(unsigned long) args->local_vec_addr;
527 527
528 if (args->nr_local == 0)
529 return -EINVAL;
530
528 /* figure out the number of pages in the vector */ 531 /* figure out the number of pages in the vector */
529 for (i = 0; i < args->nr_local; i++) { 532 for (i = 0; i < args->nr_local; i++) {
530 if (copy_from_user(&vec, &local_vec[i], 533 if (copy_from_user(&vec, &local_vec[i],
@@ -874,6 +877,7 @@ int rds_cmsg_atomic(struct rds_sock *rs, struct rds_message *rm,
874err: 877err:
875 if (page) 878 if (page)
876 put_page(page); 879 put_page(page);
880 rm->atomic.op_active = 0;
877 kfree(rm->atomic.op_notifier); 881 kfree(rm->atomic.op_notifier);
878 882
879 return ret; 883 return ret;
diff --git a/net/sched/act_gact.c b/net/sched/act_gact.c
index e29a48ef7fc3..a0ac42b3ed06 100644
--- a/net/sched/act_gact.c
+++ b/net/sched/act_gact.c
@@ -159,7 +159,7 @@ static void tcf_gact_stats_update(struct tc_action *a, u64 bytes, u32 packets,
159 if (action == TC_ACT_SHOT) 159 if (action == TC_ACT_SHOT)
160 this_cpu_ptr(gact->common.cpu_qstats)->drops += packets; 160 this_cpu_ptr(gact->common.cpu_qstats)->drops += packets;
161 161
162 tm->lastuse = lastuse; 162 tm->lastuse = max_t(u64, tm->lastuse, lastuse);
163} 163}
164 164
165static int tcf_gact_dump(struct sk_buff *skb, struct tc_action *a, 165static int tcf_gact_dump(struct sk_buff *skb, struct tc_action *a,
diff --git a/net/sched/act_mirred.c b/net/sched/act_mirred.c
index 8b3e59388480..08b61849c2a2 100644
--- a/net/sched/act_mirred.c
+++ b/net/sched/act_mirred.c
@@ -239,7 +239,7 @@ static void tcf_stats_update(struct tc_action *a, u64 bytes, u32 packets,
239 struct tcf_t *tm = &m->tcf_tm; 239 struct tcf_t *tm = &m->tcf_tm;
240 240
241 _bstats_cpu_update(this_cpu_ptr(a->cpu_bstats), bytes, packets); 241 _bstats_cpu_update(this_cpu_ptr(a->cpu_bstats), bytes, packets);
242 tm->lastuse = lastuse; 242 tm->lastuse = max_t(u64, tm->lastuse, lastuse);
243} 243}
244 244
245static int tcf_mirred_dump(struct sk_buff *skb, struct tc_action *a, int bind, 245static int tcf_mirred_dump(struct sk_buff *skb, struct tc_action *a, int bind,
diff --git a/net/sctp/input.c b/net/sctp/input.c
index 621b5ca3fd1c..141c9c466ec1 100644
--- a/net/sctp/input.c
+++ b/net/sctp/input.c
@@ -399,20 +399,24 @@ void sctp_icmp_frag_needed(struct sock *sk, struct sctp_association *asoc,
399 return; 399 return;
400 } 400 }
401 401
402 if (t->param_flags & SPP_PMTUD_ENABLE) { 402 if (!(t->param_flags & SPP_PMTUD_ENABLE))
403 /* Update transports view of the MTU */ 403 /* We can't allow retransmitting in such case, as the
404 sctp_transport_update_pmtu(t, pmtu); 404 * retransmission would be sized just as before, and thus we
405 405 * would get another icmp, and retransmit again.
406 /* Update association pmtu. */ 406 */
407 sctp_assoc_sync_pmtu(asoc); 407 return;
408 }
409 408
410 /* Retransmit with the new pmtu setting. 409 /* Update transports view of the MTU. Return if no update was needed.
411 * Normally, if PMTU discovery is disabled, an ICMP Fragmentation 410 * If an update wasn't needed/possible, it also doesn't make sense to
412 * Needed will never be sent, but if a message was sent before 411 * try to retransmit now.
413 * PMTU discovery was disabled that was larger than the PMTU, it
414 * would not be fragmented, so it must be re-transmitted fragmented.
415 */ 412 */
413 if (!sctp_transport_update_pmtu(t, pmtu))
414 return;
415
416 /* Update association pmtu. */
417 sctp_assoc_sync_pmtu(asoc);
418
419 /* Retransmit with the new pmtu setting. */
416 sctp_retransmit(&asoc->outqueue, t, SCTP_RTXR_PMTUD); 420 sctp_retransmit(&asoc->outqueue, t, SCTP_RTXR_PMTUD);
417} 421}
418 422
diff --git a/net/sctp/socket.c b/net/sctp/socket.c
index b4fb6e4886d2..9b01e994f661 100644
--- a/net/sctp/socket.c
+++ b/net/sctp/socket.c
@@ -2277,7 +2277,7 @@ static int sctp_setsockopt_events(struct sock *sk, char __user *optval,
2277 2277
2278 if (asoc && sctp_outq_is_empty(&asoc->outqueue)) { 2278 if (asoc && sctp_outq_is_empty(&asoc->outqueue)) {
2279 event = sctp_ulpevent_make_sender_dry_event(asoc, 2279 event = sctp_ulpevent_make_sender_dry_event(asoc,
2280 GFP_ATOMIC); 2280 GFP_USER | __GFP_NOWARN);
2281 if (!event) 2281 if (!event)
2282 return -ENOMEM; 2282 return -ENOMEM;
2283 2283
@@ -3498,6 +3498,8 @@ static int sctp_setsockopt_hmac_ident(struct sock *sk,
3498 3498
3499 if (optlen < sizeof(struct sctp_hmacalgo)) 3499 if (optlen < sizeof(struct sctp_hmacalgo))
3500 return -EINVAL; 3500 return -EINVAL;
3501 optlen = min_t(unsigned int, optlen, sizeof(struct sctp_hmacalgo) +
3502 SCTP_AUTH_NUM_HMACS * sizeof(u16));
3501 3503
3502 hmacs = memdup_user(optval, optlen); 3504 hmacs = memdup_user(optval, optlen);
3503 if (IS_ERR(hmacs)) 3505 if (IS_ERR(hmacs))
@@ -3536,6 +3538,11 @@ static int sctp_setsockopt_auth_key(struct sock *sk,
3536 3538
3537 if (optlen <= sizeof(struct sctp_authkey)) 3539 if (optlen <= sizeof(struct sctp_authkey))
3538 return -EINVAL; 3540 return -EINVAL;
3541 /* authkey->sca_keylength is u16, so optlen can't be bigger than
3542 * this.
3543 */
3544 optlen = min_t(unsigned int, optlen, USHRT_MAX +
3545 sizeof(struct sctp_authkey));
3539 3546
3540 authkey = memdup_user(optval, optlen); 3547 authkey = memdup_user(optval, optlen);
3541 if (IS_ERR(authkey)) 3548 if (IS_ERR(authkey))
@@ -3893,6 +3900,9 @@ static int sctp_setsockopt_reset_streams(struct sock *sk,
3893 3900
3894 if (optlen < sizeof(*params)) 3901 if (optlen < sizeof(*params))
3895 return -EINVAL; 3902 return -EINVAL;
3903 /* srs_number_streams is u16, so optlen can't be bigger than this. */
3904 optlen = min_t(unsigned int, optlen, USHRT_MAX +
3905 sizeof(__u16) * sizeof(*params));
3896 3906
3897 params = memdup_user(optval, optlen); 3907 params = memdup_user(optval, optlen);
3898 if (IS_ERR(params)) 3908 if (IS_ERR(params))
@@ -5015,7 +5025,7 @@ static int sctp_getsockopt_autoclose(struct sock *sk, int len, char __user *optv
5015 len = sizeof(int); 5025 len = sizeof(int);
5016 if (put_user(len, optlen)) 5026 if (put_user(len, optlen))
5017 return -EFAULT; 5027 return -EFAULT;
5018 if (copy_to_user(optval, &sctp_sk(sk)->autoclose, sizeof(int))) 5028 if (copy_to_user(optval, &sctp_sk(sk)->autoclose, len))
5019 return -EFAULT; 5029 return -EFAULT;
5020 return 0; 5030 return 0;
5021} 5031}
@@ -5645,6 +5655,9 @@ copy_getaddrs:
5645 err = -EFAULT; 5655 err = -EFAULT;
5646 goto out; 5656 goto out;
5647 } 5657 }
5658 /* XXX: We should have accounted for sizeof(struct sctp_getaddrs) too,
5659 * but we can't change it anymore.
5660 */
5648 if (put_user(bytes_copied, optlen)) 5661 if (put_user(bytes_copied, optlen))
5649 err = -EFAULT; 5662 err = -EFAULT;
5650out: 5663out:
@@ -6081,7 +6094,7 @@ static int sctp_getsockopt_maxseg(struct sock *sk, int len,
6081 params.assoc_id = 0; 6094 params.assoc_id = 0;
6082 } else if (len >= sizeof(struct sctp_assoc_value)) { 6095 } else if (len >= sizeof(struct sctp_assoc_value)) {
6083 len = sizeof(struct sctp_assoc_value); 6096 len = sizeof(struct sctp_assoc_value);
6084 if (copy_from_user(&params, optval, sizeof(params))) 6097 if (copy_from_user(&params, optval, len))
6085 return -EFAULT; 6098 return -EFAULT;
6086 } else 6099 } else
6087 return -EINVAL; 6100 return -EINVAL;
@@ -6251,7 +6264,9 @@ static int sctp_getsockopt_active_key(struct sock *sk, int len,
6251 6264
6252 if (len < sizeof(struct sctp_authkeyid)) 6265 if (len < sizeof(struct sctp_authkeyid))
6253 return -EINVAL; 6266 return -EINVAL;
6254 if (copy_from_user(&val, optval, sizeof(struct sctp_authkeyid))) 6267
6268 len = sizeof(struct sctp_authkeyid);
6269 if (copy_from_user(&val, optval, len))
6255 return -EFAULT; 6270 return -EFAULT;
6256 6271
6257 asoc = sctp_id2assoc(sk, val.scact_assoc_id); 6272 asoc = sctp_id2assoc(sk, val.scact_assoc_id);
@@ -6263,7 +6278,6 @@ static int sctp_getsockopt_active_key(struct sock *sk, int len,
6263 else 6278 else
6264 val.scact_keynumber = ep->active_key_id; 6279 val.scact_keynumber = ep->active_key_id;
6265 6280
6266 len = sizeof(struct sctp_authkeyid);
6267 if (put_user(len, optlen)) 6281 if (put_user(len, optlen))
6268 return -EFAULT; 6282 return -EFAULT;
6269 if (copy_to_user(optval, &val, len)) 6283 if (copy_to_user(optval, &val, len))
@@ -6289,7 +6303,7 @@ static int sctp_getsockopt_peer_auth_chunks(struct sock *sk, int len,
6289 if (len < sizeof(struct sctp_authchunks)) 6303 if (len < sizeof(struct sctp_authchunks))
6290 return -EINVAL; 6304 return -EINVAL;
6291 6305
6292 if (copy_from_user(&val, optval, sizeof(struct sctp_authchunks))) 6306 if (copy_from_user(&val, optval, sizeof(val)))
6293 return -EFAULT; 6307 return -EFAULT;
6294 6308
6295 to = p->gauth_chunks; 6309 to = p->gauth_chunks;
@@ -6334,7 +6348,7 @@ static int sctp_getsockopt_local_auth_chunks(struct sock *sk, int len,
6334 if (len < sizeof(struct sctp_authchunks)) 6348 if (len < sizeof(struct sctp_authchunks))
6335 return -EINVAL; 6349 return -EINVAL;
6336 6350
6337 if (copy_from_user(&val, optval, sizeof(struct sctp_authchunks))) 6351 if (copy_from_user(&val, optval, sizeof(val)))
6338 return -EFAULT; 6352 return -EFAULT;
6339 6353
6340 to = p->gauth_chunks; 6354 to = p->gauth_chunks;
diff --git a/net/sctp/stream.c b/net/sctp/stream.c
index 76ea66be0bbe..524dfeb94c41 100644
--- a/net/sctp/stream.c
+++ b/net/sctp/stream.c
@@ -156,9 +156,9 @@ int sctp_stream_init(struct sctp_stream *stream, __u16 outcnt, __u16 incnt,
156 sctp_stream_outq_migrate(stream, NULL, outcnt); 156 sctp_stream_outq_migrate(stream, NULL, outcnt);
157 sched->sched_all(stream); 157 sched->sched_all(stream);
158 158
159 i = sctp_stream_alloc_out(stream, outcnt, gfp); 159 ret = sctp_stream_alloc_out(stream, outcnt, gfp);
160 if (i) 160 if (ret)
161 return i; 161 goto out;
162 162
163 stream->outcnt = outcnt; 163 stream->outcnt = outcnt;
164 for (i = 0; i < stream->outcnt; i++) 164 for (i = 0; i < stream->outcnt; i++)
@@ -170,19 +170,17 @@ in:
170 if (!incnt) 170 if (!incnt)
171 goto out; 171 goto out;
172 172
173 i = sctp_stream_alloc_in(stream, incnt, gfp); 173 ret = sctp_stream_alloc_in(stream, incnt, gfp);
174 if (i) { 174 if (ret) {
175 ret = -ENOMEM; 175 sched->free(stream);
176 goto free; 176 kfree(stream->out);
177 stream->out = NULL;
178 stream->outcnt = 0;
179 goto out;
177 } 180 }
178 181
179 stream->incnt = incnt; 182 stream->incnt = incnt;
180 goto out;
181 183
182free:
183 sched->free(stream);
184 kfree(stream->out);
185 stream->out = NULL;
186out: 184out:
187 return ret; 185 return ret;
188} 186}
diff --git a/net/sctp/transport.c b/net/sctp/transport.c
index 1e5a22430cf5..47f82bd794d9 100644
--- a/net/sctp/transport.c
+++ b/net/sctp/transport.c
@@ -248,28 +248,37 @@ void sctp_transport_pmtu(struct sctp_transport *transport, struct sock *sk)
248 transport->pathmtu = SCTP_DEFAULT_MAXSEGMENT; 248 transport->pathmtu = SCTP_DEFAULT_MAXSEGMENT;
249} 249}
250 250
251void sctp_transport_update_pmtu(struct sctp_transport *t, u32 pmtu) 251bool sctp_transport_update_pmtu(struct sctp_transport *t, u32 pmtu)
252{ 252{
253 struct dst_entry *dst = sctp_transport_dst_check(t); 253 struct dst_entry *dst = sctp_transport_dst_check(t);
254 bool change = true;
254 255
255 if (unlikely(pmtu < SCTP_DEFAULT_MINSEGMENT)) { 256 if (unlikely(pmtu < SCTP_DEFAULT_MINSEGMENT)) {
256 pr_warn("%s: Reported pmtu %d too low, using default minimum of %d\n", 257 pr_warn_ratelimited("%s: Reported pmtu %d too low, using default minimum of %d\n",
257 __func__, pmtu, SCTP_DEFAULT_MINSEGMENT); 258 __func__, pmtu, SCTP_DEFAULT_MINSEGMENT);
258 /* Use default minimum segment size and disable 259 /* Use default minimum segment instead */
259 * pmtu discovery on this transport. 260 pmtu = SCTP_DEFAULT_MINSEGMENT;
260 */
261 t->pathmtu = SCTP_DEFAULT_MINSEGMENT;
262 } else {
263 t->pathmtu = pmtu;
264 } 261 }
262 pmtu = SCTP_TRUNC4(pmtu);
265 263
266 if (dst) { 264 if (dst) {
267 dst->ops->update_pmtu(dst, t->asoc->base.sk, NULL, pmtu); 265 dst->ops->update_pmtu(dst, t->asoc->base.sk, NULL, pmtu);
268 dst = sctp_transport_dst_check(t); 266 dst = sctp_transport_dst_check(t);
269 } 267 }
270 268
271 if (!dst) 269 if (!dst) {
272 t->af_specific->get_dst(t, &t->saddr, &t->fl, t->asoc->base.sk); 270 t->af_specific->get_dst(t, &t->saddr, &t->fl, t->asoc->base.sk);
271 dst = t->dst;
272 }
273
274 if (dst) {
275 /* Re-fetch, as under layers may have a higher minimum size */
276 pmtu = SCTP_TRUNC4(dst_mtu(dst));
277 change = t->pathmtu != pmtu;
278 }
279 t->pathmtu = pmtu;
280
281 return change;
273} 282}
274 283
275/* Caches the dst entry and source address for a transport's destination 284/* Caches the dst entry and source address for a transport's destination
diff --git a/net/socket.c b/net/socket.c
index 05f361faec45..6f05d5c4bf30 100644
--- a/net/socket.c
+++ b/net/socket.c
@@ -436,8 +436,10 @@ static int sock_map_fd(struct socket *sock, int flags)
436{ 436{
437 struct file *newfile; 437 struct file *newfile;
438 int fd = get_unused_fd_flags(flags); 438 int fd = get_unused_fd_flags(flags);
439 if (unlikely(fd < 0)) 439 if (unlikely(fd < 0)) {
440 sock_release(sock);
440 return fd; 441 return fd;
442 }
441 443
442 newfile = sock_alloc_file(sock, flags, NULL); 444 newfile = sock_alloc_file(sock, flags, NULL);
443 if (likely(!IS_ERR(newfile))) { 445 if (likely(!IS_ERR(newfile))) {
@@ -2619,6 +2621,15 @@ out_fs:
2619 2621
2620core_initcall(sock_init); /* early initcall */ 2622core_initcall(sock_init); /* early initcall */
2621 2623
2624static int __init jit_init(void)
2625{
2626#ifdef CONFIG_BPF_JIT_ALWAYS_ON
2627 bpf_jit_enable = 1;
2628#endif
2629 return 0;
2630}
2631pure_initcall(jit_init);
2632
2622#ifdef CONFIG_PROC_FS 2633#ifdef CONFIG_PROC_FS
2623void socket_seq_show(struct seq_file *seq) 2634void socket_seq_show(struct seq_file *seq)
2624{ 2635{
diff --git a/net/tipc/group.c b/net/tipc/group.c
index 8e12ab55346b..5f4ffae807ee 100644
--- a/net/tipc/group.c
+++ b/net/tipc/group.c
@@ -109,7 +109,8 @@ static void tipc_group_proto_xmit(struct tipc_group *grp, struct tipc_member *m,
109static void tipc_group_decr_active(struct tipc_group *grp, 109static void tipc_group_decr_active(struct tipc_group *grp,
110 struct tipc_member *m) 110 struct tipc_member *m)
111{ 111{
112 if (m->state == MBR_ACTIVE || m->state == MBR_RECLAIMING) 112 if (m->state == MBR_ACTIVE || m->state == MBR_RECLAIMING ||
113 m->state == MBR_REMITTED)
113 grp->active_cnt--; 114 grp->active_cnt--;
114} 115}
115 116
@@ -562,7 +563,7 @@ void tipc_group_update_rcv_win(struct tipc_group *grp, int blks, u32 node,
562 int max_active = grp->max_active; 563 int max_active = grp->max_active;
563 int reclaim_limit = max_active * 3 / 4; 564 int reclaim_limit = max_active * 3 / 4;
564 int active_cnt = grp->active_cnt; 565 int active_cnt = grp->active_cnt;
565 struct tipc_member *m, *rm; 566 struct tipc_member *m, *rm, *pm;
566 567
567 m = tipc_group_find_member(grp, node, port); 568 m = tipc_group_find_member(grp, node, port);
568 if (!m) 569 if (!m)
@@ -605,6 +606,17 @@ void tipc_group_update_rcv_win(struct tipc_group *grp, int blks, u32 node,
605 pr_warn_ratelimited("Rcv unexpected msg after REMIT\n"); 606 pr_warn_ratelimited("Rcv unexpected msg after REMIT\n");
606 tipc_group_proto_xmit(grp, m, GRP_ADV_MSG, xmitq); 607 tipc_group_proto_xmit(grp, m, GRP_ADV_MSG, xmitq);
607 } 608 }
609 grp->active_cnt--;
610 list_del_init(&m->list);
611 if (list_empty(&grp->pending))
612 return;
613
614 /* Set oldest pending member to active and advertise */
615 pm = list_first_entry(&grp->pending, struct tipc_member, list);
616 pm->state = MBR_ACTIVE;
617 list_move_tail(&pm->list, &grp->active);
618 grp->active_cnt++;
619 tipc_group_proto_xmit(grp, pm, GRP_ADV_MSG, xmitq);
608 break; 620 break;
609 case MBR_RECLAIMING: 621 case MBR_RECLAIMING:
610 case MBR_DISCOVERED: 622 case MBR_DISCOVERED:
@@ -742,14 +754,14 @@ void tipc_group_proto_rcv(struct tipc_group *grp, bool *usr_wakeup,
742 if (!m || m->state != MBR_RECLAIMING) 754 if (!m || m->state != MBR_RECLAIMING)
743 return; 755 return;
744 756
745 list_del_init(&m->list);
746 grp->active_cnt--;
747 remitted = msg_grp_remitted(hdr); 757 remitted = msg_grp_remitted(hdr);
748 758
749 /* Messages preceding the REMIT still in receive queue */ 759 /* Messages preceding the REMIT still in receive queue */
750 if (m->advertised > remitted) { 760 if (m->advertised > remitted) {
751 m->state = MBR_REMITTED; 761 m->state = MBR_REMITTED;
752 in_flight = m->advertised - remitted; 762 in_flight = m->advertised - remitted;
763 m->advertised = ADV_IDLE + in_flight;
764 return;
753 } 765 }
754 /* All messages preceding the REMIT have been read */ 766 /* All messages preceding the REMIT have been read */
755 if (m->advertised <= remitted) { 767 if (m->advertised <= remitted) {
@@ -761,6 +773,8 @@ void tipc_group_proto_rcv(struct tipc_group *grp, bool *usr_wakeup,
761 tipc_group_proto_xmit(grp, m, GRP_ADV_MSG, xmitq); 773 tipc_group_proto_xmit(grp, m, GRP_ADV_MSG, xmitq);
762 774
763 m->advertised = ADV_IDLE + in_flight; 775 m->advertised = ADV_IDLE + in_flight;
776 grp->active_cnt--;
777 list_del_init(&m->list);
764 778
765 /* Set oldest pending member to active and advertise */ 779 /* Set oldest pending member to active and advertise */
766 if (list_empty(&grp->pending)) 780 if (list_empty(&grp->pending))
diff --git a/net/wireless/nl80211.c b/net/wireless/nl80211.c
index 213d0c498c97..2b3dbcd40e46 100644
--- a/net/wireless/nl80211.c
+++ b/net/wireless/nl80211.c
@@ -11361,7 +11361,8 @@ static int nl80211_nan_add_func(struct sk_buff *skb,
11361 break; 11361 break;
11362 case NL80211_NAN_FUNC_FOLLOW_UP: 11362 case NL80211_NAN_FUNC_FOLLOW_UP:
11363 if (!tb[NL80211_NAN_FUNC_FOLLOW_UP_ID] || 11363 if (!tb[NL80211_NAN_FUNC_FOLLOW_UP_ID] ||
11364 !tb[NL80211_NAN_FUNC_FOLLOW_UP_REQ_ID]) { 11364 !tb[NL80211_NAN_FUNC_FOLLOW_UP_REQ_ID] ||
11365 !tb[NL80211_NAN_FUNC_FOLLOW_UP_DEST]) {
11365 err = -EINVAL; 11366 err = -EINVAL;
11366 goto out; 11367 goto out;
11367 } 11368 }
diff --git a/scripts/genksyms/.gitignore b/scripts/genksyms/.gitignore
index 86dc07a01b43..e7836b47f060 100644
--- a/scripts/genksyms/.gitignore
+++ b/scripts/genksyms/.gitignore
@@ -1,4 +1,3 @@
1*.hash.c
2*.lex.c 1*.lex.c
3*.tab.c 2*.tab.c
4*.tab.h 3*.tab.h
diff --git a/scripts/kconfig/expr.c b/scripts/kconfig/expr.c
index cbf4996dd9c1..8cee597d33a5 100644
--- a/scripts/kconfig/expr.c
+++ b/scripts/kconfig/expr.c
@@ -893,7 +893,10 @@ static enum string_value_kind expr_parse_string(const char *str,
893 switch (type) { 893 switch (type) {
894 case S_BOOLEAN: 894 case S_BOOLEAN:
895 case S_TRISTATE: 895 case S_TRISTATE:
896 return k_string; 896 val->s = !strcmp(str, "n") ? 0 :
897 !strcmp(str, "m") ? 1 :
898 !strcmp(str, "y") ? 2 : -1;
899 return k_signed;
897 case S_INT: 900 case S_INT:
898 val->s = strtoll(str, &tail, 10); 901 val->s = strtoll(str, &tail, 10);
899 kind = k_signed; 902 kind = k_signed;
diff --git a/security/Kconfig b/security/Kconfig
index 3d4debd0257e..b0cb9a5f9448 100644
--- a/security/Kconfig
+++ b/security/Kconfig
@@ -63,7 +63,7 @@ config PAGE_TABLE_ISOLATION
63 ensuring that the majority of kernel addresses are not mapped 63 ensuring that the majority of kernel addresses are not mapped
64 into userspace. 64 into userspace.
65 65
66 See Documentation/x86/pagetable-isolation.txt for more details. 66 See Documentation/x86/pti.txt for more details.
67 67
68config SECURITY_INFINIBAND 68config SECURITY_INFINIBAND
69 bool "Infiniband Security Hooks" 69 bool "Infiniband Security Hooks"
diff --git a/security/apparmor/domain.c b/security/apparmor/domain.c
index 04ba9d0718ea..6a54d2ffa840 100644
--- a/security/apparmor/domain.c
+++ b/security/apparmor/domain.c
@@ -330,10 +330,7 @@ static struct aa_profile *__attach_match(const char *name,
330 continue; 330 continue;
331 331
332 if (profile->xmatch) { 332 if (profile->xmatch) {
333 if (profile->xmatch_len == len) { 333 if (profile->xmatch_len >= len) {
334 conflict = true;
335 continue;
336 } else if (profile->xmatch_len > len) {
337 unsigned int state; 334 unsigned int state;
338 u32 perm; 335 u32 perm;
339 336
@@ -342,6 +339,10 @@ static struct aa_profile *__attach_match(const char *name,
342 perm = dfa_user_allow(profile->xmatch, state); 339 perm = dfa_user_allow(profile->xmatch, state);
343 /* any accepting state means a valid match. */ 340 /* any accepting state means a valid match. */
344 if (perm & MAY_EXEC) { 341 if (perm & MAY_EXEC) {
342 if (profile->xmatch_len == len) {
343 conflict = true;
344 continue;
345 }
345 candidate = profile; 346 candidate = profile;
346 len = profile->xmatch_len; 347 len = profile->xmatch_len;
347 conflict = false; 348 conflict = false;
diff --git a/security/apparmor/include/perms.h b/security/apparmor/include/perms.h
index 2b27bb79aec4..d7b7e7115160 100644
--- a/security/apparmor/include/perms.h
+++ b/security/apparmor/include/perms.h
@@ -133,6 +133,9 @@ extern struct aa_perms allperms;
133#define xcheck_labels_profiles(L1, L2, FN, args...) \ 133#define xcheck_labels_profiles(L1, L2, FN, args...) \
134 xcheck_ns_labels((L1), (L2), xcheck_ns_profile_label, (FN), args) 134 xcheck_ns_labels((L1), (L2), xcheck_ns_profile_label, (FN), args)
135 135
136#define xcheck_labels(L1, L2, P, FN1, FN2) \
137 xcheck(fn_for_each((L1), (P), (FN1)), fn_for_each((L2), (P), (FN2)))
138
136 139
137void aa_perm_mask_to_str(char *str, const char *chrs, u32 mask); 140void aa_perm_mask_to_str(char *str, const char *chrs, u32 mask);
138void aa_audit_perm_names(struct audit_buffer *ab, const char **names, u32 mask); 141void aa_audit_perm_names(struct audit_buffer *ab, const char **names, u32 mask);
diff --git a/security/apparmor/ipc.c b/security/apparmor/ipc.c
index 7ca0032e7ba9..b40678f3c1d5 100644
--- a/security/apparmor/ipc.c
+++ b/security/apparmor/ipc.c
@@ -64,40 +64,48 @@ static void audit_ptrace_cb(struct audit_buffer *ab, void *va)
64 FLAGS_NONE, GFP_ATOMIC); 64 FLAGS_NONE, GFP_ATOMIC);
65} 65}
66 66
67/* assumes check for PROFILE_MEDIATES is already done */
67/* TODO: conditionals */ 68/* TODO: conditionals */
68static int profile_ptrace_perm(struct aa_profile *profile, 69static int profile_ptrace_perm(struct aa_profile *profile,
69 struct aa_profile *peer, u32 request, 70 struct aa_label *peer, u32 request,
70 struct common_audit_data *sa) 71 struct common_audit_data *sa)
71{ 72{
72 struct aa_perms perms = { }; 73 struct aa_perms perms = { };
73 74
74 /* need because of peer in cross check */ 75 aad(sa)->peer = peer;
75 if (profile_unconfined(profile) || 76 aa_profile_match_label(profile, peer, AA_CLASS_PTRACE, request,
76 !PROFILE_MEDIATES(profile, AA_CLASS_PTRACE))
77 return 0;
78
79 aad(sa)->peer = &peer->label;
80 aa_profile_match_label(profile, &peer->label, AA_CLASS_PTRACE, request,
81 &perms); 77 &perms);
82 aa_apply_modes_to_perms(profile, &perms); 78 aa_apply_modes_to_perms(profile, &perms);
83 return aa_check_perms(profile, &perms, request, sa, audit_ptrace_cb); 79 return aa_check_perms(profile, &perms, request, sa, audit_ptrace_cb);
84} 80}
85 81
86static int cross_ptrace_perm(struct aa_profile *tracer, 82static int profile_tracee_perm(struct aa_profile *tracee,
87 struct aa_profile *tracee, u32 request, 83 struct aa_label *tracer, u32 request,
88 struct common_audit_data *sa) 84 struct common_audit_data *sa)
89{ 85{
86 if (profile_unconfined(tracee) || unconfined(tracer) ||
87 !PROFILE_MEDIATES(tracee, AA_CLASS_PTRACE))
88 return 0;
89
90 return profile_ptrace_perm(tracee, tracer, request, sa);
91}
92
93static int profile_tracer_perm(struct aa_profile *tracer,
94 struct aa_label *tracee, u32 request,
95 struct common_audit_data *sa)
96{
97 if (profile_unconfined(tracer))
98 return 0;
99
90 if (PROFILE_MEDIATES(tracer, AA_CLASS_PTRACE)) 100 if (PROFILE_MEDIATES(tracer, AA_CLASS_PTRACE))
91 return xcheck(profile_ptrace_perm(tracer, tracee, request, sa), 101 return profile_ptrace_perm(tracer, tracee, request, sa);
92 profile_ptrace_perm(tracee, tracer, 102
93 request << PTRACE_PERM_SHIFT, 103 /* profile uses the old style capability check for ptrace */
94 sa)); 104 if (&tracer->label == tracee)
95 /* policy uses the old style capability check for ptrace */
96 if (profile_unconfined(tracer) || tracer == tracee)
97 return 0; 105 return 0;
98 106
99 aad(sa)->label = &tracer->label; 107 aad(sa)->label = &tracer->label;
100 aad(sa)->peer = &tracee->label; 108 aad(sa)->peer = tracee;
101 aad(sa)->request = 0; 109 aad(sa)->request = 0;
102 aad(sa)->error = aa_capable(&tracer->label, CAP_SYS_PTRACE, 1); 110 aad(sa)->error = aa_capable(&tracer->label, CAP_SYS_PTRACE, 1);
103 111
@@ -115,10 +123,13 @@ static int cross_ptrace_perm(struct aa_profile *tracer,
115int aa_may_ptrace(struct aa_label *tracer, struct aa_label *tracee, 123int aa_may_ptrace(struct aa_label *tracer, struct aa_label *tracee,
116 u32 request) 124 u32 request)
117{ 125{
126 struct aa_profile *profile;
127 u32 xrequest = request << PTRACE_PERM_SHIFT;
118 DEFINE_AUDIT_DATA(sa, LSM_AUDIT_DATA_NONE, OP_PTRACE); 128 DEFINE_AUDIT_DATA(sa, LSM_AUDIT_DATA_NONE, OP_PTRACE);
119 129
120 return xcheck_labels_profiles(tracer, tracee, cross_ptrace_perm, 130 return xcheck_labels(tracer, tracee, profile,
121 request, &sa); 131 profile_tracer_perm(profile, tracee, request, &sa),
132 profile_tracee_perm(profile, tracer, xrequest, &sa));
122} 133}
123 134
124 135
diff --git a/sound/soc/au1x/ac97c.c b/sound/soc/au1x/ac97c.c
index 29a97d52e8ad..66d6c52e7761 100644
--- a/sound/soc/au1x/ac97c.c
+++ b/sound/soc/au1x/ac97c.c
@@ -91,8 +91,8 @@ static unsigned short au1xac97c_ac97_read(struct snd_ac97 *ac97,
91 do { 91 do {
92 mutex_lock(&ctx->lock); 92 mutex_lock(&ctx->lock);
93 93
94 tmo = 5; 94 tmo = 6;
95 while ((RD(ctx, AC97_STATUS) & STAT_CP) && tmo--) 95 while ((RD(ctx, AC97_STATUS) & STAT_CP) && --tmo)
96 udelay(21); /* wait an ac97 frame time */ 96 udelay(21); /* wait an ac97 frame time */
97 if (!tmo) { 97 if (!tmo) {
98 pr_debug("ac97rd timeout #1\n"); 98 pr_debug("ac97rd timeout #1\n");
@@ -105,7 +105,7 @@ static unsigned short au1xac97c_ac97_read(struct snd_ac97 *ac97,
105 * poll, Forrest, poll... 105 * poll, Forrest, poll...
106 */ 106 */
107 tmo = 0x10000; 107 tmo = 0x10000;
108 while ((RD(ctx, AC97_STATUS) & STAT_CP) && tmo--) 108 while ((RD(ctx, AC97_STATUS) & STAT_CP) && --tmo)
109 asm volatile ("nop"); 109 asm volatile ("nop");
110 data = RD(ctx, AC97_CMDRESP); 110 data = RD(ctx, AC97_CMDRESP);
111 111
diff --git a/sound/soc/bcm/bcm2835-i2s.c b/sound/soc/bcm/bcm2835-i2s.c
index 2e449d7173fc..d5f73a8ab893 100644
--- a/sound/soc/bcm/bcm2835-i2s.c
+++ b/sound/soc/bcm/bcm2835-i2s.c
@@ -130,6 +130,7 @@ struct bcm2835_i2s_dev {
130 struct regmap *i2s_regmap; 130 struct regmap *i2s_regmap;
131 struct clk *clk; 131 struct clk *clk;
132 bool clk_prepared; 132 bool clk_prepared;
133 int clk_rate;
133}; 134};
134 135
135static void bcm2835_i2s_start_clock(struct bcm2835_i2s_dev *dev) 136static void bcm2835_i2s_start_clock(struct bcm2835_i2s_dev *dev)
@@ -419,10 +420,19 @@ static int bcm2835_i2s_hw_params(struct snd_pcm_substream *substream,
419 } 420 }
420 421
421 /* Clock should only be set up here if CPU is clock master */ 422 /* Clock should only be set up here if CPU is clock master */
422 if (bit_clock_master) { 423 if (bit_clock_master &&
423 ret = clk_set_rate(dev->clk, bclk_rate); 424 (!dev->clk_prepared || dev->clk_rate != bclk_rate)) {
424 if (ret) 425 if (dev->clk_prepared)
425 return ret; 426 bcm2835_i2s_stop_clock(dev);
427
428 if (dev->clk_rate != bclk_rate) {
429 ret = clk_set_rate(dev->clk, bclk_rate);
430 if (ret)
431 return ret;
432 dev->clk_rate = bclk_rate;
433 }
434
435 bcm2835_i2s_start_clock(dev);
426 } 436 }
427 437
428 /* Setup the frame format */ 438 /* Setup the frame format */
@@ -618,8 +628,6 @@ static int bcm2835_i2s_prepare(struct snd_pcm_substream *substream,
618 struct bcm2835_i2s_dev *dev = snd_soc_dai_get_drvdata(dai); 628 struct bcm2835_i2s_dev *dev = snd_soc_dai_get_drvdata(dai);
619 uint32_t cs_reg; 629 uint32_t cs_reg;
620 630
621 bcm2835_i2s_start_clock(dev);
622
623 /* 631 /*
624 * Clear both FIFOs if the one that should be started 632 * Clear both FIFOs if the one that should be started
625 * is not empty at the moment. This should only happen 633 * is not empty at the moment. This should only happen
diff --git a/sound/soc/codecs/cs42l73.c b/sound/soc/codecs/cs42l73.c
index dde37e569ade..aebaa97490b6 100644
--- a/sound/soc/codecs/cs42l73.c
+++ b/sound/soc/codecs/cs42l73.c
@@ -1355,7 +1355,7 @@ static int cs42l73_i2c_probe(struct i2c_client *i2c_client,
1355 ret = regmap_read(cs42l73->regmap, CS42L73_REVID, &reg); 1355 ret = regmap_read(cs42l73->regmap, CS42L73_REVID, &reg);
1356 if (ret < 0) { 1356 if (ret < 0) {
1357 dev_err(&i2c_client->dev, "Get Revision ID failed\n"); 1357 dev_err(&i2c_client->dev, "Get Revision ID failed\n");
1358 return ret;; 1358 return ret;
1359 } 1359 }
1360 1360
1361 dev_info(&i2c_client->dev, 1361 dev_info(&i2c_client->dev,
diff --git a/sound/soc/codecs/cx20442.c b/sound/soc/codecs/cx20442.c
index 6b6f8e44369b..95bb10ba80dc 100644
--- a/sound/soc/codecs/cx20442.c
+++ b/sound/soc/codecs/cx20442.c
@@ -28,6 +28,7 @@
28struct cx20442_priv { 28struct cx20442_priv {
29 struct tty_struct *tty; 29 struct tty_struct *tty;
30 struct regulator *por; 30 struct regulator *por;
31 u8 reg_cache;
31}; 32};
32 33
33#define CX20442_PM 0x0 34#define CX20442_PM 0x0
@@ -88,6 +89,17 @@ static const struct snd_soc_dapm_route cx20442_audio_map[] = {
88 {"ADC", NULL, "Input Mixer"}, 89 {"ADC", NULL, "Input Mixer"},
89}; 90};
90 91
92static unsigned int cx20442_read_reg_cache(struct snd_soc_codec *codec,
93 unsigned int reg)
94{
95 struct cx20442_priv *cx20442 = snd_soc_codec_get_drvdata(codec);
96
97 if (reg >= 1)
98 return -EINVAL;
99
100 return cx20442->reg_cache;
101}
102
91enum v253_vls { 103enum v253_vls {
92 V253_VLS_NONE = 0, 104 V253_VLS_NONE = 0,
93 V253_VLS_T, 105 V253_VLS_T,
@@ -112,8 +124,6 @@ enum v253_vls {
112 V253_VLS_TEST, 124 V253_VLS_TEST,
113}; 125};
114 126
115#if 0
116/* FIXME : these function will be re-used */
117static int cx20442_pm_to_v253_vls(u8 value) 127static int cx20442_pm_to_v253_vls(u8 value)
118{ 128{
119 switch (value & ~(1 << CX20442_AGC)) { 129 switch (value & ~(1 << CX20442_AGC)) {
@@ -147,11 +157,10 @@ static int cx20442_write(struct snd_soc_codec *codec, unsigned int reg,
147 unsigned int value) 157 unsigned int value)
148{ 158{
149 struct cx20442_priv *cx20442 = snd_soc_codec_get_drvdata(codec); 159 struct cx20442_priv *cx20442 = snd_soc_codec_get_drvdata(codec);
150 u8 *reg_cache = codec->reg_cache;
151 int vls, vsp, old, len; 160 int vls, vsp, old, len;
152 char buf[18]; 161 char buf[18];
153 162
154 if (reg >= codec->driver->reg_cache_size) 163 if (reg >= 1)
155 return -EINVAL; 164 return -EINVAL;
156 165
157 /* tty and write pointers required for talking to the modem 166 /* tty and write pointers required for talking to the modem
@@ -159,8 +168,8 @@ static int cx20442_write(struct snd_soc_codec *codec, unsigned int reg,
159 if (!cx20442->tty || !cx20442->tty->ops->write) 168 if (!cx20442->tty || !cx20442->tty->ops->write)
160 return -EIO; 169 return -EIO;
161 170
162 old = reg_cache[reg]; 171 old = cx20442->reg_cache;
163 reg_cache[reg] = value; 172 cx20442->reg_cache = value;
164 173
165 vls = cx20442_pm_to_v253_vls(value); 174 vls = cx20442_pm_to_v253_vls(value);
166 if (vls < 0) 175 if (vls < 0)
@@ -190,7 +199,6 @@ static int cx20442_write(struct snd_soc_codec *codec, unsigned int reg,
190 199
191 return 0; 200 return 0;
192} 201}
193#endif
194 202
195/* 203/*
196 * Line discpline related code 204 * Line discpline related code
@@ -384,12 +392,12 @@ static int cx20442_codec_remove(struct snd_soc_codec *codec)
384 return 0; 392 return 0;
385} 393}
386 394
387static const u8 cx20442_reg;
388
389static const struct snd_soc_codec_driver cx20442_codec_dev = { 395static const struct snd_soc_codec_driver cx20442_codec_dev = {
390 .probe = cx20442_codec_probe, 396 .probe = cx20442_codec_probe,
391 .remove = cx20442_codec_remove, 397 .remove = cx20442_codec_remove,
392 .set_bias_level = cx20442_set_bias_level, 398 .set_bias_level = cx20442_set_bias_level,
399 .read = cx20442_read_reg_cache,
400 .write = cx20442_write,
393 401
394 .component_driver = { 402 .component_driver = {
395 .dapm_widgets = cx20442_dapm_widgets, 403 .dapm_widgets = cx20442_dapm_widgets,
diff --git a/sound/soc/codecs/sgtl5000.c b/sound/soc/codecs/sgtl5000.c
index f2bb4feba3b6..633cdcfc933d 100644
--- a/sound/soc/codecs/sgtl5000.c
+++ b/sound/soc/codecs/sgtl5000.c
@@ -1332,10 +1332,13 @@ static int sgtl5000_i2c_probe(struct i2c_client *client,
1332 sgtl5000->mclk = devm_clk_get(&client->dev, NULL); 1332 sgtl5000->mclk = devm_clk_get(&client->dev, NULL);
1333 if (IS_ERR(sgtl5000->mclk)) { 1333 if (IS_ERR(sgtl5000->mclk)) {
1334 ret = PTR_ERR(sgtl5000->mclk); 1334 ret = PTR_ERR(sgtl5000->mclk);
1335 dev_err(&client->dev, "Failed to get mclock: %d\n", ret);
1336 /* Defer the probe to see if the clk will be provided later */ 1335 /* Defer the probe to see if the clk will be provided later */
1337 if (ret == -ENOENT) 1336 if (ret == -ENOENT)
1338 ret = -EPROBE_DEFER; 1337 ret = -EPROBE_DEFER;
1338
1339 if (ret != -EPROBE_DEFER)
1340 dev_err(&client->dev, "Failed to get mclock: %d\n",
1341 ret);
1339 goto disable_regs; 1342 goto disable_regs;
1340 } 1343 }
1341 1344
diff --git a/sound/soc/codecs/tlv320dac33.c b/sound/soc/codecs/tlv320dac33.c
index 675f5b1b90a6..8c71d2f876ff 100644
--- a/sound/soc/codecs/tlv320dac33.c
+++ b/sound/soc/codecs/tlv320dac33.c
@@ -246,6 +246,19 @@ static int dac33_write(struct snd_soc_codec *codec, unsigned int reg,
246 return ret; 246 return ret;
247} 247}
248 248
249static int dac33_write_locked(struct snd_soc_codec *codec, unsigned int reg,
250 unsigned int value)
251{
252 struct tlv320dac33_priv *dac33 = snd_soc_codec_get_drvdata(codec);
253 int ret;
254
255 mutex_lock(&dac33->mutex);
256 ret = dac33_write(codec, reg, value);
257 mutex_unlock(&dac33->mutex);
258
259 return ret;
260}
261
249#define DAC33_I2C_ADDR_AUTOINC 0x80 262#define DAC33_I2C_ADDR_AUTOINC 0x80
250static int dac33_write16(struct snd_soc_codec *codec, unsigned int reg, 263static int dac33_write16(struct snd_soc_codec *codec, unsigned int reg,
251 unsigned int value) 264 unsigned int value)
@@ -1422,6 +1435,8 @@ static int dac33_soc_remove(struct snd_soc_codec *codec)
1422} 1435}
1423 1436
1424static const struct snd_soc_codec_driver soc_codec_dev_tlv320dac33 = { 1437static const struct snd_soc_codec_driver soc_codec_dev_tlv320dac33 = {
1438 .read = dac33_read_reg_cache,
1439 .write = dac33_write_locked,
1425 .set_bias_level = dac33_set_bias_level, 1440 .set_bias_level = dac33_set_bias_level,
1426 .idle_bias_off = true, 1441 .idle_bias_off = true,
1427 1442
diff --git a/sound/soc/codecs/uda1380.c b/sound/soc/codecs/uda1380.c
index 46a495b4da8d..c73e6a192224 100644
--- a/sound/soc/codecs/uda1380.c
+++ b/sound/soc/codecs/uda1380.c
@@ -726,6 +726,8 @@ static int uda1380_probe(struct snd_soc_codec *codec)
726 726
727static const struct snd_soc_codec_driver soc_codec_dev_uda1380 = { 727static const struct snd_soc_codec_driver soc_codec_dev_uda1380 = {
728 .probe = uda1380_probe, 728 .probe = uda1380_probe,
729 .read = uda1380_read_reg_cache,
730 .write = uda1380_write,
729 .set_bias_level = uda1380_set_bias_level, 731 .set_bias_level = uda1380_set_bias_level,
730 .suspend_bias_off = true, 732 .suspend_bias_off = true,
731 733
diff --git a/sound/soc/intel/Kconfig b/sound/soc/intel/Kconfig
index b0bd1938b71e..f2c9e8c5970a 100644
--- a/sound/soc/intel/Kconfig
+++ b/sound/soc/intel/Kconfig
@@ -77,7 +77,6 @@ config SND_SST_ATOM_HIFI2_PLATFORM_PCI
77 depends on X86 && PCI 77 depends on X86 && PCI
78 select SND_SST_IPC_PCI 78 select SND_SST_IPC_PCI
79 select SND_SOC_COMPRESS 79 select SND_SOC_COMPRESS
80 select SND_SOC_INTEL_COMMON
81 help 80 help
82 If you have a Intel Medfield or Merrifield/Edison platform, then 81 If you have a Intel Medfield or Merrifield/Edison platform, then
83 enable this option by saying Y or m. Distros will typically not 82 enable this option by saying Y or m. Distros will typically not
diff --git a/sound/soc/intel/boards/Kconfig b/sound/soc/intel/boards/Kconfig
index de598dcbef30..d4e103615f51 100644
--- a/sound/soc/intel/boards/Kconfig
+++ b/sound/soc/intel/boards/Kconfig
@@ -139,6 +139,7 @@ config SND_SOC_INTEL_BYT_CHT_DA7213_MACH
139config SND_SOC_INTEL_BYT_CHT_ES8316_MACH 139config SND_SOC_INTEL_BYT_CHT_ES8316_MACH
140 tristate "Baytrail & Cherrytrail with ES8316 codec" 140 tristate "Baytrail & Cherrytrail with ES8316 codec"
141 depends on X86_INTEL_LPSS && I2C && ACPI 141 depends on X86_INTEL_LPSS && I2C && ACPI
142 select SND_SOC_ACPI
142 select SND_SOC_ES8316 143 select SND_SOC_ES8316
143 help 144 help
144 This adds support for ASoC machine driver for Intel(R) Baytrail & 145 This adds support for ASoC machine driver for Intel(R) Baytrail &
diff --git a/sound/soc/intel/boards/bytcht_da7213.c b/sound/soc/intel/boards/bytcht_da7213.c
index c4d82ad41bd7..2179dedb28ad 100644
--- a/sound/soc/intel/boards/bytcht_da7213.c
+++ b/sound/soc/intel/boards/bytcht_da7213.c
@@ -219,7 +219,7 @@ static struct snd_soc_card bytcht_da7213_card = {
219 .num_dapm_routes = ARRAY_SIZE(audio_map), 219 .num_dapm_routes = ARRAY_SIZE(audio_map),
220}; 220};
221 221
222static char codec_name[16]; /* i2c-<HID>:00 with HID being 8 chars */ 222static char codec_name[SND_ACPI_I2C_ID_LEN];
223 223
224static int bytcht_da7213_probe(struct platform_device *pdev) 224static int bytcht_da7213_probe(struct platform_device *pdev)
225{ 225{
@@ -243,7 +243,7 @@ static int bytcht_da7213_probe(struct platform_device *pdev)
243 } 243 }
244 244
245 /* fixup codec name based on HID */ 245 /* fixup codec name based on HID */
246 i2c_name = snd_soc_acpi_find_name_from_hid(mach->id); 246 i2c_name = acpi_dev_get_first_match_name(mach->id, NULL, -1);
247 if (i2c_name) { 247 if (i2c_name) {
248 snprintf(codec_name, sizeof(codec_name), 248 snprintf(codec_name, sizeof(codec_name),
249 "%s%s", "i2c-", i2c_name); 249 "%s%s", "i2c-", i2c_name);
diff --git a/sound/soc/intel/boards/bytcht_es8316.c b/sound/soc/intel/boards/bytcht_es8316.c
index 8088396717e3..305e7f4fe55a 100644
--- a/sound/soc/intel/boards/bytcht_es8316.c
+++ b/sound/soc/intel/boards/bytcht_es8316.c
@@ -232,15 +232,39 @@ static struct snd_soc_card byt_cht_es8316_card = {
232 .fully_routed = true, 232 .fully_routed = true,
233}; 233};
234 234
235static char codec_name[SND_ACPI_I2C_ID_LEN];
236
235static int snd_byt_cht_es8316_mc_probe(struct platform_device *pdev) 237static int snd_byt_cht_es8316_mc_probe(struct platform_device *pdev)
236{ 238{
237 int ret = 0;
238 struct byt_cht_es8316_private *priv; 239 struct byt_cht_es8316_private *priv;
240 struct snd_soc_acpi_mach *mach;
241 const char *i2c_name = NULL;
242 int dai_index = 0;
243 int i;
244 int ret = 0;
239 245
240 priv = devm_kzalloc(&pdev->dev, sizeof(*priv), GFP_ATOMIC); 246 priv = devm_kzalloc(&pdev->dev, sizeof(*priv), GFP_ATOMIC);
241 if (!priv) 247 if (!priv)
242 return -ENOMEM; 248 return -ENOMEM;
243 249
250 mach = (&pdev->dev)->platform_data;
251 /* fix index of codec dai */
252 for (i = 0; i < ARRAY_SIZE(byt_cht_es8316_dais); i++) {
253 if (!strcmp(byt_cht_es8316_dais[i].codec_name,
254 "i2c-ESSX8316:00")) {
255 dai_index = i;
256 break;
257 }
258 }
259
260 /* fixup codec name based on HID */
261 i2c_name = acpi_dev_get_first_match_name(mach->id, NULL, -1);
262 if (i2c_name) {
263 snprintf(codec_name, sizeof(codec_name),
264 "%s%s", "i2c-", i2c_name);
265 byt_cht_es8316_dais[dai_index].codec_name = codec_name;
266 }
267
244 /* register the soc card */ 268 /* register the soc card */
245 byt_cht_es8316_card.dev = &pdev->dev; 269 byt_cht_es8316_card.dev = &pdev->dev;
246 snd_soc_card_set_drvdata(&byt_cht_es8316_card, priv); 270 snd_soc_card_set_drvdata(&byt_cht_es8316_card, priv);
diff --git a/sound/soc/intel/boards/bytcr_rt5640.c b/sound/soc/intel/boards/bytcr_rt5640.c
index f2c0fc415e52..b6a1cfeec830 100644
--- a/sound/soc/intel/boards/bytcr_rt5640.c
+++ b/sound/soc/intel/boards/bytcr_rt5640.c
@@ -713,7 +713,7 @@ static struct snd_soc_card byt_rt5640_card = {
713 .fully_routed = true, 713 .fully_routed = true,
714}; 714};
715 715
716static char byt_rt5640_codec_name[16]; /* i2c-<HID>:00 with HID being 8 chars */ 716static char byt_rt5640_codec_name[SND_ACPI_I2C_ID_LEN];
717static char byt_rt5640_codec_aif_name[12]; /* = "rt5640-aif[1|2]" */ 717static char byt_rt5640_codec_aif_name[12]; /* = "rt5640-aif[1|2]" */
718static char byt_rt5640_cpu_dai_name[10]; /* = "ssp[0|2]-port" */ 718static char byt_rt5640_cpu_dai_name[10]; /* = "ssp[0|2]-port" */
719 719
@@ -762,7 +762,7 @@ static int snd_byt_rt5640_mc_probe(struct platform_device *pdev)
762 } 762 }
763 763
764 /* fixup codec name based on HID */ 764 /* fixup codec name based on HID */
765 i2c_name = snd_soc_acpi_find_name_from_hid(mach->id); 765 i2c_name = acpi_dev_get_first_match_name(mach->id, NULL, -1);
766 if (i2c_name) { 766 if (i2c_name) {
767 snprintf(byt_rt5640_codec_name, sizeof(byt_rt5640_codec_name), 767 snprintf(byt_rt5640_codec_name, sizeof(byt_rt5640_codec_name),
768 "%s%s", "i2c-", i2c_name); 768 "%s%s", "i2c-", i2c_name);
diff --git a/sound/soc/intel/boards/bytcr_rt5651.c b/sound/soc/intel/boards/bytcr_rt5651.c
index 22c9cc5d135e..456526a93dd5 100644
--- a/sound/soc/intel/boards/bytcr_rt5651.c
+++ b/sound/soc/intel/boards/bytcr_rt5651.c
@@ -509,7 +509,7 @@ static struct snd_soc_card byt_rt5651_card = {
509 .fully_routed = true, 509 .fully_routed = true,
510}; 510};
511 511
512static char byt_rt5651_codec_name[16]; /* i2c-<HID>:00 with HID being 8 chars */ 512static char byt_rt5651_codec_name[SND_ACPI_I2C_ID_LEN];
513 513
514static int snd_byt_rt5651_mc_probe(struct platform_device *pdev) 514static int snd_byt_rt5651_mc_probe(struct platform_device *pdev)
515{ 515{
@@ -539,7 +539,7 @@ static int snd_byt_rt5651_mc_probe(struct platform_device *pdev)
539 } 539 }
540 540
541 /* fixup codec name based on HID */ 541 /* fixup codec name based on HID */
542 i2c_name = snd_soc_acpi_find_name_from_hid(mach->id); 542 i2c_name = acpi_dev_get_first_match_name(mach->id, NULL, -1);
543 if (i2c_name) { 543 if (i2c_name) {
544 snprintf(byt_rt5651_codec_name, sizeof(byt_rt5651_codec_name), 544 snprintf(byt_rt5651_codec_name, sizeof(byt_rt5651_codec_name),
545 "%s%s", "i2c-", i2c_name); 545 "%s%s", "i2c-", i2c_name);
diff --git a/sound/soc/intel/boards/cht_bsw_rt5645.c b/sound/soc/intel/boards/cht_bsw_rt5645.c
index f898ee140cdc..31641aab62cd 100644
--- a/sound/soc/intel/boards/cht_bsw_rt5645.c
+++ b/sound/soc/intel/boards/cht_bsw_rt5645.c
@@ -49,7 +49,7 @@ struct cht_acpi_card {
49struct cht_mc_private { 49struct cht_mc_private {
50 struct snd_soc_jack jack; 50 struct snd_soc_jack jack;
51 struct cht_acpi_card *acpi_card; 51 struct cht_acpi_card *acpi_card;
52 char codec_name[16]; 52 char codec_name[SND_ACPI_I2C_ID_LEN];
53 struct clk *mclk; 53 struct clk *mclk;
54}; 54};
55 55
@@ -506,7 +506,7 @@ static struct cht_acpi_card snd_soc_cards[] = {
506 {"10EC5650", CODEC_TYPE_RT5650, &snd_soc_card_chtrt5650}, 506 {"10EC5650", CODEC_TYPE_RT5650, &snd_soc_card_chtrt5650},
507}; 507};
508 508
509static char cht_rt5645_codec_name[16]; /* i2c-<HID>:00 with HID being 8 chars */ 509static char cht_rt5645_codec_name[SND_ACPI_I2C_ID_LEN];
510static char cht_rt5645_codec_aif_name[12]; /* = "rt5645-aif[1|2]" */ 510static char cht_rt5645_codec_aif_name[12]; /* = "rt5645-aif[1|2]" */
511static char cht_rt5645_cpu_dai_name[10]; /* = "ssp[0|2]-port" */ 511static char cht_rt5645_cpu_dai_name[10]; /* = "ssp[0|2]-port" */
512 512
@@ -573,7 +573,7 @@ static int snd_cht_mc_probe(struct platform_device *pdev)
573 } 573 }
574 574
575 /* fixup codec name based on HID */ 575 /* fixup codec name based on HID */
576 i2c_name = snd_soc_acpi_find_name_from_hid(mach->id); 576 i2c_name = acpi_dev_get_first_match_name(mach->id, NULL, -1);
577 if (i2c_name) { 577 if (i2c_name) {
578 snprintf(cht_rt5645_codec_name, sizeof(cht_rt5645_codec_name), 578 snprintf(cht_rt5645_codec_name, sizeof(cht_rt5645_codec_name),
579 "%s%s", "i2c-", i2c_name); 579 "%s%s", "i2c-", i2c_name);
diff --git a/sound/soc/intel/boards/cht_bsw_rt5672.c b/sound/soc/intel/boards/cht_bsw_rt5672.c
index f8f21eee9b2d..c14a52d2f714 100644
--- a/sound/soc/intel/boards/cht_bsw_rt5672.c
+++ b/sound/soc/intel/boards/cht_bsw_rt5672.c
@@ -35,7 +35,7 @@
35 35
36struct cht_mc_private { 36struct cht_mc_private {
37 struct snd_soc_jack headset; 37 struct snd_soc_jack headset;
38 char codec_name[16]; 38 char codec_name[SND_ACPI_I2C_ID_LEN];
39 struct clk *mclk; 39 struct clk *mclk;
40}; 40};
41 41
@@ -396,7 +396,7 @@ static int snd_cht_mc_probe(struct platform_device *pdev)
396 396
397 /* fixup codec name based on HID */ 397 /* fixup codec name based on HID */
398 if (mach) { 398 if (mach) {
399 i2c_name = snd_soc_acpi_find_name_from_hid(mach->id); 399 i2c_name = acpi_dev_get_first_match_name(mach->id, NULL, -1);
400 if (i2c_name) { 400 if (i2c_name) {
401 snprintf(drv->codec_name, sizeof(drv->codec_name), 401 snprintf(drv->codec_name, sizeof(drv->codec_name),
402 "i2c-%s", i2c_name); 402 "i2c-%s", i2c_name);
diff --git a/sound/soc/intel/skylake/skl-topology.c b/sound/soc/intel/skylake/skl-topology.c
index 28bc16a8e09a..73af6e19ebbd 100644
--- a/sound/soc/intel/skylake/skl-topology.c
+++ b/sound/soc/intel/skylake/skl-topology.c
@@ -190,7 +190,6 @@ skl_tplg_free_pipe_mcps(struct skl *skl, struct skl_module_cfg *mconfig)
190 u8 res_idx = mconfig->res_idx; 190 u8 res_idx = mconfig->res_idx;
191 struct skl_module_res *res = &mconfig->module->resources[res_idx]; 191 struct skl_module_res *res = &mconfig->module->resources[res_idx];
192 192
193 res = &mconfig->module->resources[res_idx];
194 skl->resource.mcps -= res->cps; 193 skl->resource.mcps -= res->cps;
195} 194}
196 195
diff --git a/sound/soc/mxs/mxs-sgtl5000.c b/sound/soc/mxs/mxs-sgtl5000.c
index 2ed3240cc682..2b3f2408301a 100644
--- a/sound/soc/mxs/mxs-sgtl5000.c
+++ b/sound/soc/mxs/mxs-sgtl5000.c
@@ -93,6 +93,14 @@ static struct snd_soc_dai_link mxs_sgtl5000_dai[] = {
93 }, 93 },
94}; 94};
95 95
96static const struct snd_soc_dapm_widget mxs_sgtl5000_dapm_widgets[] = {
97 SND_SOC_DAPM_MIC("Mic Jack", NULL),
98 SND_SOC_DAPM_LINE("Line In Jack", NULL),
99 SND_SOC_DAPM_HP("Headphone Jack", NULL),
100 SND_SOC_DAPM_SPK("Line Out Jack", NULL),
101 SND_SOC_DAPM_SPK("Ext Spk", NULL),
102};
103
96static struct snd_soc_card mxs_sgtl5000 = { 104static struct snd_soc_card mxs_sgtl5000 = {
97 .name = "mxs_sgtl5000", 105 .name = "mxs_sgtl5000",
98 .owner = THIS_MODULE, 106 .owner = THIS_MODULE,
@@ -141,10 +149,23 @@ static int mxs_sgtl5000_probe(struct platform_device *pdev)
141 149
142 card->dev = &pdev->dev; 150 card->dev = &pdev->dev;
143 151
152 if (of_find_property(np, "audio-routing", NULL)) {
153 card->dapm_widgets = mxs_sgtl5000_dapm_widgets;
154 card->num_dapm_widgets = ARRAY_SIZE(mxs_sgtl5000_dapm_widgets);
155
156 ret = snd_soc_of_parse_audio_routing(card, "audio-routing");
157 if (ret) {
158 dev_err(&pdev->dev, "failed to parse audio-routing (%d)\n",
159 ret);
160 return ret;
161 }
162 }
163
144 ret = devm_snd_soc_register_card(&pdev->dev, card); 164 ret = devm_snd_soc_register_card(&pdev->dev, card);
145 if (ret) { 165 if (ret) {
146 dev_err(&pdev->dev, "snd_soc_register_card failed (%d)\n", 166 if (ret != -EPROBE_DEFER)
147 ret); 167 dev_err(&pdev->dev, "snd_soc_register_card failed (%d)\n",
168 ret);
148 return ret; 169 return ret;
149 } 170 }
150 171
diff --git a/sound/soc/soc-acpi.c b/sound/soc/soc-acpi.c
index 7f43c9bf3d09..3d7e1ff79139 100644
--- a/sound/soc/soc-acpi.c
+++ b/sound/soc/soc-acpi.c
@@ -16,39 +16,6 @@
16 16
17#include <sound/soc-acpi.h> 17#include <sound/soc-acpi.h>
18 18
19static acpi_status snd_soc_acpi_find_name(acpi_handle handle, u32 level,
20 void *context, void **ret)
21{
22 struct acpi_device *adev;
23 const char *name = NULL;
24
25 if (acpi_bus_get_device(handle, &adev))
26 return AE_OK;
27
28 if (adev->status.present && adev->status.functional) {
29 name = acpi_dev_name(adev);
30 *(const char **)ret = name;
31 return AE_CTRL_TERMINATE;
32 }
33
34 return AE_OK;
35}
36
37const char *snd_soc_acpi_find_name_from_hid(const u8 hid[ACPI_ID_LEN])
38{
39 const char *name = NULL;
40 acpi_status status;
41
42 status = acpi_get_devices(hid, snd_soc_acpi_find_name, NULL,
43 (void **)&name);
44
45 if (ACPI_FAILURE(status) || name[0] == '\0')
46 return NULL;
47
48 return name;
49}
50EXPORT_SYMBOL_GPL(snd_soc_acpi_find_name_from_hid);
51
52struct snd_soc_acpi_mach * 19struct snd_soc_acpi_mach *
53snd_soc_acpi_find_machine(struct snd_soc_acpi_mach *machines) 20snd_soc_acpi_find_machine(struct snd_soc_acpi_mach *machines)
54{ 21{
diff --git a/sound/soc/soc-core.c b/sound/soc/soc-core.c
index d1f7e639d5b1..e91879569a0f 100644
--- a/sound/soc/soc-core.c
+++ b/sound/soc/soc-core.c
@@ -349,120 +349,84 @@ static void soc_init_codec_debugfs(struct snd_soc_component *component)
349 "ASoC: Failed to create codec register debugfs file\n"); 349 "ASoC: Failed to create codec register debugfs file\n");
350} 350}
351 351
352static ssize_t codec_list_read_file(struct file *file, char __user *user_buf, 352static int codec_list_seq_show(struct seq_file *m, void *v)
353 size_t count, loff_t *ppos)
354{ 353{
355 char *buf = kmalloc(PAGE_SIZE, GFP_KERNEL);
356 ssize_t len, ret = 0;
357 struct snd_soc_codec *codec; 354 struct snd_soc_codec *codec;
358 355
359 if (!buf)
360 return -ENOMEM;
361
362 mutex_lock(&client_mutex); 356 mutex_lock(&client_mutex);
363 357
364 list_for_each_entry(codec, &codec_list, list) { 358 list_for_each_entry(codec, &codec_list, list)
365 len = snprintf(buf + ret, PAGE_SIZE - ret, "%s\n", 359 seq_printf(m, "%s\n", codec->component.name);
366 codec->component.name);
367 if (len >= 0)
368 ret += len;
369 if (ret > PAGE_SIZE) {
370 ret = PAGE_SIZE;
371 break;
372 }
373 }
374 360
375 mutex_unlock(&client_mutex); 361 mutex_unlock(&client_mutex);
376 362
377 if (ret >= 0) 363 return 0;
378 ret = simple_read_from_buffer(user_buf, count, ppos, buf, ret); 364}
379
380 kfree(buf);
381 365
382 return ret; 366static int codec_list_seq_open(struct inode *inode, struct file *file)
367{
368 return single_open(file, codec_list_seq_show, NULL);
383} 369}
384 370
385static const struct file_operations codec_list_fops = { 371static const struct file_operations codec_list_fops = {
386 .read = codec_list_read_file, 372 .open = codec_list_seq_open,
387 .llseek = default_llseek,/* read accesses f_pos */ 373 .read = seq_read,
374 .llseek = seq_lseek,
375 .release = single_release,
388}; 376};
389 377
390static ssize_t dai_list_read_file(struct file *file, char __user *user_buf, 378static int dai_list_seq_show(struct seq_file *m, void *v)
391 size_t count, loff_t *ppos)
392{ 379{
393 char *buf = kmalloc(PAGE_SIZE, GFP_KERNEL);
394 ssize_t len, ret = 0;
395 struct snd_soc_component *component; 380 struct snd_soc_component *component;
396 struct snd_soc_dai *dai; 381 struct snd_soc_dai *dai;
397 382
398 if (!buf)
399 return -ENOMEM;
400
401 mutex_lock(&client_mutex); 383 mutex_lock(&client_mutex);
402 384
403 list_for_each_entry(component, &component_list, list) { 385 list_for_each_entry(component, &component_list, list)
404 list_for_each_entry(dai, &component->dai_list, list) { 386 list_for_each_entry(dai, &component->dai_list, list)
405 len = snprintf(buf + ret, PAGE_SIZE - ret, "%s\n", 387 seq_printf(m, "%s\n", dai->name);
406 dai->name);
407 if (len >= 0)
408 ret += len;
409 if (ret > PAGE_SIZE) {
410 ret = PAGE_SIZE;
411 break;
412 }
413 }
414 }
415 388
416 mutex_unlock(&client_mutex); 389 mutex_unlock(&client_mutex);
417 390
418 ret = simple_read_from_buffer(user_buf, count, ppos, buf, ret); 391 return 0;
419 392}
420 kfree(buf);
421 393
422 return ret; 394static int dai_list_seq_open(struct inode *inode, struct file *file)
395{
396 return single_open(file, dai_list_seq_show, NULL);
423} 397}
424 398
425static const struct file_operations dai_list_fops = { 399static const struct file_operations dai_list_fops = {
426 .read = dai_list_read_file, 400 .open = dai_list_seq_open,
427 .llseek = default_llseek,/* read accesses f_pos */ 401 .read = seq_read,
402 .llseek = seq_lseek,
403 .release = single_release,
428}; 404};
429 405
430static ssize_t platform_list_read_file(struct file *file, 406static int platform_list_seq_show(struct seq_file *m, void *v)
431 char __user *user_buf,
432 size_t count, loff_t *ppos)
433{ 407{
434 char *buf = kmalloc(PAGE_SIZE, GFP_KERNEL);
435 ssize_t len, ret = 0;
436 struct snd_soc_platform *platform; 408 struct snd_soc_platform *platform;
437 409
438 if (!buf)
439 return -ENOMEM;
440
441 mutex_lock(&client_mutex); 410 mutex_lock(&client_mutex);
442 411
443 list_for_each_entry(platform, &platform_list, list) { 412 list_for_each_entry(platform, &platform_list, list)
444 len = snprintf(buf + ret, PAGE_SIZE - ret, "%s\n", 413 seq_printf(m, "%s\n", platform->component.name);
445 platform->component.name);
446 if (len >= 0)
447 ret += len;
448 if (ret > PAGE_SIZE) {
449 ret = PAGE_SIZE;
450 break;
451 }
452 }
453 414
454 mutex_unlock(&client_mutex); 415 mutex_unlock(&client_mutex);
455 416
456 ret = simple_read_from_buffer(user_buf, count, ppos, buf, ret); 417 return 0;
457 418}
458 kfree(buf);
459 419
460 return ret; 420static int platform_list_seq_open(struct inode *inode, struct file *file)
421{
422 return single_open(file, platform_list_seq_show, NULL);
461} 423}
462 424
463static const struct file_operations platform_list_fops = { 425static const struct file_operations platform_list_fops = {
464 .read = platform_list_read_file, 426 .open = platform_list_seq_open,
465 .llseek = default_llseek,/* read accesses f_pos */ 427 .read = seq_read,
428 .llseek = seq_lseek,
429 .release = single_release,
466}; 430};
467 431
468static void soc_init_card_debugfs(struct snd_soc_card *card) 432static void soc_init_card_debugfs(struct snd_soc_card *card)
@@ -491,7 +455,6 @@ static void soc_cleanup_card_debugfs(struct snd_soc_card *card)
491 debugfs_remove_recursive(card->debugfs_card_root); 455 debugfs_remove_recursive(card->debugfs_card_root);
492} 456}
493 457
494
495static void snd_soc_debugfs_init(void) 458static void snd_soc_debugfs_init(void)
496{ 459{
497 snd_soc_debugfs_root = debugfs_create_dir("asoc", NULL); 460 snd_soc_debugfs_root = debugfs_create_dir("asoc", NULL);
@@ -598,6 +561,7 @@ struct snd_soc_component *snd_soc_rtdcom_lookup(struct snd_soc_pcm_runtime *rtd,
598 561
599 return NULL; 562 return NULL;
600} 563}
564EXPORT_SYMBOL_GPL(snd_soc_rtdcom_lookup);
601 565
602struct snd_pcm_substream *snd_soc_get_dai_substream(struct snd_soc_card *card, 566struct snd_pcm_substream *snd_soc_get_dai_substream(struct snd_soc_card *card,
603 const char *dai_link, int stream) 567 const char *dai_link, int stream)
diff --git a/sound/soc/soc-io.c b/sound/soc/soc-io.c
index 20340ade20a7..2bc1c4c17896 100644
--- a/sound/soc/soc-io.c
+++ b/sound/soc/soc-io.c
@@ -34,6 +34,10 @@ int snd_soc_component_read(struct snd_soc_component *component,
34 ret = regmap_read(component->regmap, reg, val); 34 ret = regmap_read(component->regmap, reg, val);
35 else if (component->read) 35 else if (component->read)
36 ret = component->read(component, reg, val); 36 ret = component->read(component, reg, val);
37 else if (component->driver->read) {
38 *val = component->driver->read(component, reg);
39 ret = 0;
40 }
37 else 41 else
38 ret = -EIO; 42 ret = -EIO;
39 43
@@ -70,6 +74,8 @@ int snd_soc_component_write(struct snd_soc_component *component,
70 return regmap_write(component->regmap, reg, val); 74 return regmap_write(component->regmap, reg, val);
71 else if (component->write) 75 else if (component->write)
72 return component->write(component, reg, val); 76 return component->write(component, reg, val);
77 else if (component->driver->write)
78 return component->driver->write(component, reg, val);
73 else 79 else
74 return -EIO; 80 return -EIO;
75} 81}
diff --git a/tools/objtool/Makefile b/tools/objtool/Makefile
index ae0272f9a091..e6acc281dd37 100644
--- a/tools/objtool/Makefile
+++ b/tools/objtool/Makefile
@@ -46,7 +46,7 @@ $(OBJTOOL_IN): fixdep FORCE
46 @$(MAKE) $(build)=objtool 46 @$(MAKE) $(build)=objtool
47 47
48$(OBJTOOL): $(LIBSUBCMD) $(OBJTOOL_IN) 48$(OBJTOOL): $(LIBSUBCMD) $(OBJTOOL_IN)
49 @./sync-check.sh 49 @$(CONFIG_SHELL) ./sync-check.sh
50 $(QUIET_LINK)$(CC) $(OBJTOOL_IN) $(LDFLAGS) -o $@ 50 $(QUIET_LINK)$(CC) $(OBJTOOL_IN) $(LDFLAGS) -o $@
51 51
52 52
diff --git a/tools/objtool/check.c b/tools/objtool/check.c
index 9b341584eb1b..f40d46e24bcc 100644
--- a/tools/objtool/check.c
+++ b/tools/objtool/check.c
@@ -428,6 +428,40 @@ static void add_ignores(struct objtool_file *file)
428} 428}
429 429
430/* 430/*
431 * FIXME: For now, just ignore any alternatives which add retpolines. This is
432 * a temporary hack, as it doesn't allow ORC to unwind from inside a retpoline.
433 * But it at least allows objtool to understand the control flow *around* the
434 * retpoline.
435 */
436static int add_nospec_ignores(struct objtool_file *file)
437{
438 struct section *sec;
439 struct rela *rela;
440 struct instruction *insn;
441
442 sec = find_section_by_name(file->elf, ".rela.discard.nospec");
443 if (!sec)
444 return 0;
445
446 list_for_each_entry(rela, &sec->rela_list, list) {
447 if (rela->sym->type != STT_SECTION) {
448 WARN("unexpected relocation symbol type in %s", sec->name);
449 return -1;
450 }
451
452 insn = find_insn(file, rela->sym->sec, rela->addend);
453 if (!insn) {
454 WARN("bad .discard.nospec entry");
455 return -1;
456 }
457
458 insn->ignore_alts = true;
459 }
460
461 return 0;
462}
463
464/*
431 * Find the destination instructions for all jumps. 465 * Find the destination instructions for all jumps.
432 */ 466 */
433static int add_jump_destinations(struct objtool_file *file) 467static int add_jump_destinations(struct objtool_file *file)
@@ -456,6 +490,13 @@ static int add_jump_destinations(struct objtool_file *file)
456 } else if (rela->sym->sec->idx) { 490 } else if (rela->sym->sec->idx) {
457 dest_sec = rela->sym->sec; 491 dest_sec = rela->sym->sec;
458 dest_off = rela->sym->sym.st_value + rela->addend + 4; 492 dest_off = rela->sym->sym.st_value + rela->addend + 4;
493 } else if (strstr(rela->sym->name, "_indirect_thunk_")) {
494 /*
495 * Retpoline jumps are really dynamic jumps in
496 * disguise, so convert them accordingly.
497 */
498 insn->type = INSN_JUMP_DYNAMIC;
499 continue;
459 } else { 500 } else {
460 /* sibling call */ 501 /* sibling call */
461 insn->jump_dest = 0; 502 insn->jump_dest = 0;
@@ -502,11 +543,18 @@ static int add_call_destinations(struct objtool_file *file)
502 dest_off = insn->offset + insn->len + insn->immediate; 543 dest_off = insn->offset + insn->len + insn->immediate;
503 insn->call_dest = find_symbol_by_offset(insn->sec, 544 insn->call_dest = find_symbol_by_offset(insn->sec,
504 dest_off); 545 dest_off);
546 /*
547 * FIXME: Thanks to retpolines, it's now considered
548 * normal for a function to call within itself. So
549 * disable this warning for now.
550 */
551#if 0
505 if (!insn->call_dest) { 552 if (!insn->call_dest) {
506 WARN_FUNC("can't find call dest symbol at offset 0x%lx", 553 WARN_FUNC("can't find call dest symbol at offset 0x%lx",
507 insn->sec, insn->offset, dest_off); 554 insn->sec, insn->offset, dest_off);
508 return -1; 555 return -1;
509 } 556 }
557#endif
510 } else if (rela->sym->type == STT_SECTION) { 558 } else if (rela->sym->type == STT_SECTION) {
511 insn->call_dest = find_symbol_by_offset(rela->sym->sec, 559 insn->call_dest = find_symbol_by_offset(rela->sym->sec,
512 rela->addend+4); 560 rela->addend+4);
@@ -671,12 +719,6 @@ static int add_special_section_alts(struct objtool_file *file)
671 return ret; 719 return ret;
672 720
673 list_for_each_entry_safe(special_alt, tmp, &special_alts, list) { 721 list_for_each_entry_safe(special_alt, tmp, &special_alts, list) {
674 alt = malloc(sizeof(*alt));
675 if (!alt) {
676 WARN("malloc failed");
677 ret = -1;
678 goto out;
679 }
680 722
681 orig_insn = find_insn(file, special_alt->orig_sec, 723 orig_insn = find_insn(file, special_alt->orig_sec,
682 special_alt->orig_off); 724 special_alt->orig_off);
@@ -687,6 +729,10 @@ static int add_special_section_alts(struct objtool_file *file)
687 goto out; 729 goto out;
688 } 730 }
689 731
732 /* Ignore retpoline alternatives. */
733 if (orig_insn->ignore_alts)
734 continue;
735
690 new_insn = NULL; 736 new_insn = NULL;
691 if (!special_alt->group || special_alt->new_len) { 737 if (!special_alt->group || special_alt->new_len) {
692 new_insn = find_insn(file, special_alt->new_sec, 738 new_insn = find_insn(file, special_alt->new_sec,
@@ -712,6 +758,13 @@ static int add_special_section_alts(struct objtool_file *file)
712 goto out; 758 goto out;
713 } 759 }
714 760
761 alt = malloc(sizeof(*alt));
762 if (!alt) {
763 WARN("malloc failed");
764 ret = -1;
765 goto out;
766 }
767
715 alt->insn = new_insn; 768 alt->insn = new_insn;
716 list_add_tail(&alt->list, &orig_insn->alts); 769 list_add_tail(&alt->list, &orig_insn->alts);
717 770
@@ -1028,6 +1081,10 @@ static int decode_sections(struct objtool_file *file)
1028 1081
1029 add_ignores(file); 1082 add_ignores(file);
1030 1083
1084 ret = add_nospec_ignores(file);
1085 if (ret)
1086 return ret;
1087
1031 ret = add_jump_destinations(file); 1088 ret = add_jump_destinations(file);
1032 if (ret) 1089 if (ret)
1033 return ret; 1090 return ret;
diff --git a/tools/objtool/check.h b/tools/objtool/check.h
index 47d9ea70a83d..dbadb304a410 100644
--- a/tools/objtool/check.h
+++ b/tools/objtool/check.h
@@ -44,7 +44,7 @@ struct instruction {
44 unsigned int len; 44 unsigned int len;
45 unsigned char type; 45 unsigned char type;
46 unsigned long immediate; 46 unsigned long immediate;
47 bool alt_group, visited, dead_end, ignore, hint, save, restore; 47 bool alt_group, visited, dead_end, ignore, hint, save, restore, ignore_alts;
48 struct symbol *call_dest; 48 struct symbol *call_dest;
49 struct instruction *jump_dest; 49 struct instruction *jump_dest;
50 struct list_head alts; 50 struct list_head alts;
diff --git a/tools/testing/selftests/bpf/test_align.c b/tools/testing/selftests/bpf/test_align.c
index 8591c89c0828..471bbbdb94db 100644
--- a/tools/testing/selftests/bpf/test_align.c
+++ b/tools/testing/selftests/bpf/test_align.c
@@ -474,27 +474,7 @@ static struct bpf_align_test tests[] = {
474 .result = REJECT, 474 .result = REJECT,
475 .matches = { 475 .matches = {
476 {4, "R5=pkt(id=0,off=0,r=0,imm=0)"}, 476 {4, "R5=pkt(id=0,off=0,r=0,imm=0)"},
477 /* ptr & 0x40 == either 0 or 0x40 */ 477 /* R5 bitwise operator &= on pointer prohibited */
478 {5, "R5=inv(id=0,umax_value=64,var_off=(0x0; 0x40))"},
479 /* ptr << 2 == unknown, (4n) */
480 {7, "R5=inv(id=0,smax_value=9223372036854775804,umax_value=18446744073709551612,var_off=(0x0; 0xfffffffffffffffc))"},
481 /* (4n) + 14 == (4n+2). We blow our bounds, because
482 * the add could overflow.
483 */
484 {8, "R5=inv(id=0,var_off=(0x2; 0xfffffffffffffffc))"},
485 /* Checked s>=0 */
486 {10, "R5=inv(id=0,umin_value=2,umax_value=9223372036854775806,var_off=(0x2; 0x7ffffffffffffffc))"},
487 /* packet pointer + nonnegative (4n+2) */
488 {12, "R6=pkt(id=1,off=0,r=0,umin_value=2,umax_value=9223372036854775806,var_off=(0x2; 0x7ffffffffffffffc))"},
489 {14, "R4=pkt(id=1,off=4,r=0,umin_value=2,umax_value=9223372036854775806,var_off=(0x2; 0x7ffffffffffffffc))"},
490 /* NET_IP_ALIGN + (4n+2) == (4n), alignment is fine.
491 * We checked the bounds, but it might have been able
492 * to overflow if the packet pointer started in the
493 * upper half of the address space.
494 * So we did not get a 'range' on R6, and the access
495 * attempt will fail.
496 */
497 {16, "R6=pkt(id=1,off=0,r=0,umin_value=2,umax_value=9223372036854775806,var_off=(0x2; 0x7ffffffffffffffc))"},
498 } 478 }
499 }, 479 },
500 { 480 {
diff --git a/tools/testing/selftests/x86/Makefile b/tools/testing/selftests/x86/Makefile
index 939a337128db..5d4f10ac2af2 100644
--- a/tools/testing/selftests/x86/Makefile
+++ b/tools/testing/selftests/x86/Makefile
@@ -7,7 +7,7 @@ include ../lib.mk
7 7
8TARGETS_C_BOTHBITS := single_step_syscall sysret_ss_attrs syscall_nt ptrace_syscall test_mremap_vdso \ 8TARGETS_C_BOTHBITS := single_step_syscall sysret_ss_attrs syscall_nt ptrace_syscall test_mremap_vdso \
9 check_initial_reg_state sigreturn ldt_gdt iopl mpx-mini-test ioperm \ 9 check_initial_reg_state sigreturn ldt_gdt iopl mpx-mini-test ioperm \
10 protection_keys test_vdso 10 protection_keys test_vdso test_vsyscall
11TARGETS_C_32BIT_ONLY := entry_from_vm86 syscall_arg_fault test_syscall_vdso unwind_vdso \ 11TARGETS_C_32BIT_ONLY := entry_from_vm86 syscall_arg_fault test_syscall_vdso unwind_vdso \
12 test_FCMOV test_FCOMI test_FISTTP \ 12 test_FCMOV test_FCOMI test_FISTTP \
13 vdso_restorer 13 vdso_restorer
diff --git a/tools/testing/selftests/x86/test_vsyscall.c b/tools/testing/selftests/x86/test_vsyscall.c
new file mode 100644
index 000000000000..7a744fa7b786
--- /dev/null
+++ b/tools/testing/selftests/x86/test_vsyscall.c
@@ -0,0 +1,500 @@
1/* SPDX-License-Identifier: GPL-2.0 */
2
3#define _GNU_SOURCE
4
5#include <stdio.h>
6#include <sys/time.h>
7#include <time.h>
8#include <stdlib.h>
9#include <sys/syscall.h>
10#include <unistd.h>
11#include <dlfcn.h>
12#include <string.h>
13#include <inttypes.h>
14#include <signal.h>
15#include <sys/ucontext.h>
16#include <errno.h>
17#include <err.h>
18#include <sched.h>
19#include <stdbool.h>
20#include <setjmp.h>
21
22#ifdef __x86_64__
23# define VSYS(x) (x)
24#else
25# define VSYS(x) 0
26#endif
27
28#ifndef SYS_getcpu
29# ifdef __x86_64__
30# define SYS_getcpu 309
31# else
32# define SYS_getcpu 318
33# endif
34#endif
35
36static void sethandler(int sig, void (*handler)(int, siginfo_t *, void *),
37 int flags)
38{
39 struct sigaction sa;
40 memset(&sa, 0, sizeof(sa));
41 sa.sa_sigaction = handler;
42 sa.sa_flags = SA_SIGINFO | flags;
43 sigemptyset(&sa.sa_mask);
44 if (sigaction(sig, &sa, 0))
45 err(1, "sigaction");
46}
47
48/* vsyscalls and vDSO */
49bool should_read_vsyscall = false;
50
51typedef long (*gtod_t)(struct timeval *tv, struct timezone *tz);
52gtod_t vgtod = (gtod_t)VSYS(0xffffffffff600000);
53gtod_t vdso_gtod;
54
55typedef int (*vgettime_t)(clockid_t, struct timespec *);
56vgettime_t vdso_gettime;
57
58typedef long (*time_func_t)(time_t *t);
59time_func_t vtime = (time_func_t)VSYS(0xffffffffff600400);
60time_func_t vdso_time;
61
62typedef long (*getcpu_t)(unsigned *, unsigned *, void *);
63getcpu_t vgetcpu = (getcpu_t)VSYS(0xffffffffff600800);
64getcpu_t vdso_getcpu;
65
66static void init_vdso(void)
67{
68 void *vdso = dlopen("linux-vdso.so.1", RTLD_LAZY | RTLD_LOCAL | RTLD_NOLOAD);
69 if (!vdso)
70 vdso = dlopen("linux-gate.so.1", RTLD_LAZY | RTLD_LOCAL | RTLD_NOLOAD);
71 if (!vdso) {
72 printf("[WARN]\tfailed to find vDSO\n");
73 return;
74 }
75
76 vdso_gtod = (gtod_t)dlsym(vdso, "__vdso_gettimeofday");
77 if (!vdso_gtod)
78 printf("[WARN]\tfailed to find gettimeofday in vDSO\n");
79
80 vdso_gettime = (vgettime_t)dlsym(vdso, "__vdso_clock_gettime");
81 if (!vdso_gettime)
82 printf("[WARN]\tfailed to find clock_gettime in vDSO\n");
83
84 vdso_time = (time_func_t)dlsym(vdso, "__vdso_time");
85 if (!vdso_time)
86 printf("[WARN]\tfailed to find time in vDSO\n");
87
88 vdso_getcpu = (getcpu_t)dlsym(vdso, "__vdso_getcpu");
89 if (!vdso_getcpu) {
90 /* getcpu() was never wired up in the 32-bit vDSO. */
91 printf("[%s]\tfailed to find getcpu in vDSO\n",
92 sizeof(long) == 8 ? "WARN" : "NOTE");
93 }
94}
95
96static int init_vsys(void)
97{
98#ifdef __x86_64__
99 int nerrs = 0;
100 FILE *maps;
101 char line[128];
102 bool found = false;
103
104 maps = fopen("/proc/self/maps", "r");
105 if (!maps) {
106 printf("[WARN]\tCould not open /proc/self/maps -- assuming vsyscall is r-x\n");
107 should_read_vsyscall = true;
108 return 0;
109 }
110
111 while (fgets(line, sizeof(line), maps)) {
112 char r, x;
113 void *start, *end;
114 char name[128];
115 if (sscanf(line, "%p-%p %c-%cp %*x %*x:%*x %*u %s",
116 &start, &end, &r, &x, name) != 5)
117 continue;
118
119 if (strcmp(name, "[vsyscall]"))
120 continue;
121
122 printf("\tvsyscall map: %s", line);
123
124 if (start != (void *)0xffffffffff600000 ||
125 end != (void *)0xffffffffff601000) {
126 printf("[FAIL]\taddress range is nonsense\n");
127 nerrs++;
128 }
129
130 printf("\tvsyscall permissions are %c-%c\n", r, x);
131 should_read_vsyscall = (r == 'r');
132 if (x != 'x') {
133 vgtod = NULL;
134 vtime = NULL;
135 vgetcpu = NULL;
136 }
137
138 found = true;
139 break;
140 }
141
142 fclose(maps);
143
144 if (!found) {
145 printf("\tno vsyscall map in /proc/self/maps\n");
146 should_read_vsyscall = false;
147 vgtod = NULL;
148 vtime = NULL;
149 vgetcpu = NULL;
150 }
151
152 return nerrs;
153#else
154 return 0;
155#endif
156}
157
158/* syscalls */
159static inline long sys_gtod(struct timeval *tv, struct timezone *tz)
160{
161 return syscall(SYS_gettimeofday, tv, tz);
162}
163
164static inline int sys_clock_gettime(clockid_t id, struct timespec *ts)
165{
166 return syscall(SYS_clock_gettime, id, ts);
167}
168
169static inline long sys_time(time_t *t)
170{
171 return syscall(SYS_time, t);
172}
173
174static inline long sys_getcpu(unsigned * cpu, unsigned * node,
175 void* cache)
176{
177 return syscall(SYS_getcpu, cpu, node, cache);
178}
179
180static jmp_buf jmpbuf;
181
182static void sigsegv(int sig, siginfo_t *info, void *ctx_void)
183{
184 siglongjmp(jmpbuf, 1);
185}
186
187static double tv_diff(const struct timeval *a, const struct timeval *b)
188{
189 return (double)(a->tv_sec - b->tv_sec) +
190 (double)((int)a->tv_usec - (int)b->tv_usec) * 1e-6;
191}
192
193static int check_gtod(const struct timeval *tv_sys1,
194 const struct timeval *tv_sys2,
195 const struct timezone *tz_sys,
196 const char *which,
197 const struct timeval *tv_other,
198 const struct timezone *tz_other)
199{
200 int nerrs = 0;
201 double d1, d2;
202
203 if (tz_other && (tz_sys->tz_minuteswest != tz_other->tz_minuteswest || tz_sys->tz_dsttime != tz_other->tz_dsttime)) {
204 printf("[FAIL] %s tz mismatch\n", which);
205 nerrs++;
206 }
207
208 d1 = tv_diff(tv_other, tv_sys1);
209 d2 = tv_diff(tv_sys2, tv_other);
210 printf("\t%s time offsets: %lf %lf\n", which, d1, d2);
211
212 if (d1 < 0 || d2 < 0) {
213 printf("[FAIL]\t%s time was inconsistent with the syscall\n", which);
214 nerrs++;
215 } else {
216 printf("[OK]\t%s gettimeofday()'s timeval was okay\n", which);
217 }
218
219 return nerrs;
220}
221
222static int test_gtod(void)
223{
224 struct timeval tv_sys1, tv_sys2, tv_vdso, tv_vsys;
225 struct timezone tz_sys, tz_vdso, tz_vsys;
226 long ret_vdso = -1;
227 long ret_vsys = -1;
228 int nerrs = 0;
229
230 printf("[RUN]\ttest gettimeofday()\n");
231
232 if (sys_gtod(&tv_sys1, &tz_sys) != 0)
233 err(1, "syscall gettimeofday");
234 if (vdso_gtod)
235 ret_vdso = vdso_gtod(&tv_vdso, &tz_vdso);
236 if (vgtod)
237 ret_vsys = vgtod(&tv_vsys, &tz_vsys);
238 if (sys_gtod(&tv_sys2, &tz_sys) != 0)
239 err(1, "syscall gettimeofday");
240
241 if (vdso_gtod) {
242 if (ret_vdso == 0) {
243 nerrs += check_gtod(&tv_sys1, &tv_sys2, &tz_sys, "vDSO", &tv_vdso, &tz_vdso);
244 } else {
245 printf("[FAIL]\tvDSO gettimeofday() failed: %ld\n", ret_vdso);
246 nerrs++;
247 }
248 }
249
250 if (vgtod) {
251 if (ret_vsys == 0) {
252 nerrs += check_gtod(&tv_sys1, &tv_sys2, &tz_sys, "vsyscall", &tv_vsys, &tz_vsys);
253 } else {
254 printf("[FAIL]\tvsys gettimeofday() failed: %ld\n", ret_vsys);
255 nerrs++;
256 }
257 }
258
259 return nerrs;
260}
261
262static int test_time(void) {
263 int nerrs = 0;
264
265 printf("[RUN]\ttest time()\n");
266 long t_sys1, t_sys2, t_vdso = 0, t_vsys = 0;
267 long t2_sys1 = -1, t2_sys2 = -1, t2_vdso = -1, t2_vsys = -1;
268 t_sys1 = sys_time(&t2_sys1);
269 if (vdso_time)
270 t_vdso = vdso_time(&t2_vdso);
271 if (vtime)
272 t_vsys = vtime(&t2_vsys);
273 t_sys2 = sys_time(&t2_sys2);
274 if (t_sys1 < 0 || t_sys1 != t2_sys1 || t_sys2 < 0 || t_sys2 != t2_sys2) {
275 printf("[FAIL]\tsyscall failed (ret1:%ld output1:%ld ret2:%ld output2:%ld)\n", t_sys1, t2_sys1, t_sys2, t2_sys2);
276 nerrs++;
277 return nerrs;
278 }
279
280 if (vdso_time) {
281 if (t_vdso < 0 || t_vdso != t2_vdso) {
282 printf("[FAIL]\tvDSO failed (ret:%ld output:%ld)\n", t_vdso, t2_vdso);
283 nerrs++;
284 } else if (t_vdso < t_sys1 || t_vdso > t_sys2) {
285 printf("[FAIL]\tvDSO returned the wrong time (%ld %ld %ld)\n", t_sys1, t_vdso, t_sys2);
286 nerrs++;
287 } else {
288 printf("[OK]\tvDSO time() is okay\n");
289 }
290 }
291
292 if (vtime) {
293 if (t_vsys < 0 || t_vsys != t2_vsys) {
294 printf("[FAIL]\tvsyscall failed (ret:%ld output:%ld)\n", t_vsys, t2_vsys);
295 nerrs++;
296 } else if (t_vsys < t_sys1 || t_vsys > t_sys2) {
297 printf("[FAIL]\tvsyscall returned the wrong time (%ld %ld %ld)\n", t_sys1, t_vsys, t_sys2);
298 nerrs++;
299 } else {
300 printf("[OK]\tvsyscall time() is okay\n");
301 }
302 }
303
304 return nerrs;
305}
306
307static int test_getcpu(int cpu)
308{
309 int nerrs = 0;
310 long ret_sys, ret_vdso = -1, ret_vsys = -1;
311
312 printf("[RUN]\tgetcpu() on CPU %d\n", cpu);
313
314 cpu_set_t cpuset;
315 CPU_ZERO(&cpuset);
316 CPU_SET(cpu, &cpuset);
317 if (sched_setaffinity(0, sizeof(cpuset), &cpuset) != 0) {
318 printf("[SKIP]\tfailed to force CPU %d\n", cpu);
319 return nerrs;
320 }
321
322 unsigned cpu_sys, cpu_vdso, cpu_vsys, node_sys, node_vdso, node_vsys;
323 unsigned node = 0;
324 bool have_node = false;
325 ret_sys = sys_getcpu(&cpu_sys, &node_sys, 0);
326 if (vdso_getcpu)
327 ret_vdso = vdso_getcpu(&cpu_vdso, &node_vdso, 0);
328 if (vgetcpu)
329 ret_vsys = vgetcpu(&cpu_vsys, &node_vsys, 0);
330
331 if (ret_sys == 0) {
332 if (cpu_sys != cpu) {
333 printf("[FAIL]\tsyscall reported CPU %hu but should be %d\n", cpu_sys, cpu);
334 nerrs++;
335 }
336
337 have_node = true;
338 node = node_sys;
339 }
340
341 if (vdso_getcpu) {
342 if (ret_vdso) {
343 printf("[FAIL]\tvDSO getcpu() failed\n");
344 nerrs++;
345 } else {
346 if (!have_node) {
347 have_node = true;
348 node = node_vdso;
349 }
350
351 if (cpu_vdso != cpu) {
352 printf("[FAIL]\tvDSO reported CPU %hu but should be %d\n", cpu_vdso, cpu);
353 nerrs++;
354 } else {
355 printf("[OK]\tvDSO reported correct CPU\n");
356 }
357
358 if (node_vdso != node) {
359 printf("[FAIL]\tvDSO reported node %hu but should be %hu\n", node_vdso, node);
360 nerrs++;
361 } else {
362 printf("[OK]\tvDSO reported correct node\n");
363 }
364 }
365 }
366
367 if (vgetcpu) {
368 if (ret_vsys) {
369 printf("[FAIL]\tvsyscall getcpu() failed\n");
370 nerrs++;
371 } else {
372 if (!have_node) {
373 have_node = true;
374 node = node_vsys;
375 }
376
377 if (cpu_vsys != cpu) {
378 printf("[FAIL]\tvsyscall reported CPU %hu but should be %d\n", cpu_vsys, cpu);
379 nerrs++;
380 } else {
381 printf("[OK]\tvsyscall reported correct CPU\n");
382 }
383
384 if (node_vsys != node) {
385 printf("[FAIL]\tvsyscall reported node %hu but should be %hu\n", node_vsys, node);
386 nerrs++;
387 } else {
388 printf("[OK]\tvsyscall reported correct node\n");
389 }
390 }
391 }
392
393 return nerrs;
394}
395
396static int test_vsys_r(void)
397{
398#ifdef __x86_64__
399 printf("[RUN]\tChecking read access to the vsyscall page\n");
400 bool can_read;
401 if (sigsetjmp(jmpbuf, 1) == 0) {
402 *(volatile int *)0xffffffffff600000;
403 can_read = true;
404 } else {
405 can_read = false;
406 }
407
408 if (can_read && !should_read_vsyscall) {
409 printf("[FAIL]\tWe have read access, but we shouldn't\n");
410 return 1;
411 } else if (!can_read && should_read_vsyscall) {
412 printf("[FAIL]\tWe don't have read access, but we should\n");
413 return 1;
414 } else {
415 printf("[OK]\tgot expected result\n");
416 }
417#endif
418
419 return 0;
420}
421
422
423#ifdef __x86_64__
424#define X86_EFLAGS_TF (1UL << 8)
425static volatile sig_atomic_t num_vsyscall_traps;
426
427static unsigned long get_eflags(void)
428{
429 unsigned long eflags;
430 asm volatile ("pushfq\n\tpopq %0" : "=rm" (eflags));
431 return eflags;
432}
433
434static void set_eflags(unsigned long eflags)
435{
436 asm volatile ("pushq %0\n\tpopfq" : : "rm" (eflags) : "flags");
437}
438
439static void sigtrap(int sig, siginfo_t *info, void *ctx_void)
440{
441 ucontext_t *ctx = (ucontext_t *)ctx_void;
442 unsigned long ip = ctx->uc_mcontext.gregs[REG_RIP];
443
444 if (((ip ^ 0xffffffffff600000UL) & ~0xfffUL) == 0)
445 num_vsyscall_traps++;
446}
447
448static int test_native_vsyscall(void)
449{
450 time_t tmp;
451 bool is_native;
452
453 if (!vtime)
454 return 0;
455
456 printf("[RUN]\tchecking for native vsyscall\n");
457 sethandler(SIGTRAP, sigtrap, 0);
458 set_eflags(get_eflags() | X86_EFLAGS_TF);
459 vtime(&tmp);
460 set_eflags(get_eflags() & ~X86_EFLAGS_TF);
461
462 /*
463 * If vsyscalls are emulated, we expect a single trap in the
464 * vsyscall page -- the call instruction will trap with RIP
465 * pointing to the entry point before emulation takes over.
466 * In native mode, we expect two traps, since whatever code
467 * the vsyscall page contains will be more than just a ret
468 * instruction.
469 */
470 is_native = (num_vsyscall_traps > 1);
471
472 printf("\tvsyscalls are %s (%d instructions in vsyscall page)\n",
473 (is_native ? "native" : "emulated"),
474 (int)num_vsyscall_traps);
475
476 return 0;
477}
478#endif
479
480int main(int argc, char **argv)
481{
482 int nerrs = 0;
483
484 init_vdso();
485 nerrs += init_vsys();
486
487 nerrs += test_gtod();
488 nerrs += test_time();
489 nerrs += test_getcpu(0);
490 nerrs += test_getcpu(1);
491
492 sethandler(SIGSEGV, sigsegv, 0);
493 nerrs += test_vsys_r();
494
495#ifdef __x86_64__
496 nerrs += test_native_vsyscall();
497#endif
498
499 return nerrs ? 1 : 0;
500}