Pull NTB updates from Jon Mason:

 - reorg and clean-up of the Intel NTB driver

 - a trivial comment change

 - change GFP_ATOMIC to GFP_KERNEL where appropriate

* tag 'ntb-4.18' of git://github.com/jonmason/ntb:
  ntb: ntb_transport: Replace GFP_ATOMIC with GFP_KERNEL in ntb_transport_create_queue
  ntb: ntb_transport: Replace GFP_ATOMIC with GFP_KERNEL in ntb_transport_setup_qp_mw
  NTB: ntb_hw_idt: fix typo 'can by' to 'can be'
  ntb: intel: change references of skx to gen3
  ntb: intel: split out the gen3 code
  ntb: intel: header definitions refactor

Pull VFIO updates from Alex Williamson:

 - Bind type1 task tracking to group_leader to facilitate vCPU hotplug
   in QEMU (Alex Williamson)

 - Sample mdev display drivers, including region-based host and guest
   Linux drivers and bochs compatible dmabuf device
   (Gerd Hoffmann)

 - Fix vfio-platform reset module leak (Geert Uytterhoeven)

 - vfio-platform error message consistency (Geert Uytterhoeven)

 - Global checking for mdev uuid collisions rather than per parent
   device (Alex Williamson)

 - Use match_string() helper (Yisheng Xie)

 - vfio-platform PM domain fixes (Geert Uytterhoeven)

 - Fix sample mbochs driver build dependency (Arnd Bergmann)

* tag 'vfio-v4.18-rc1' of git://github.com/awilliam/linux-vfio:
  samples: mbochs: add DMA_SHARED_BUFFER dependency
  vfio: platform: Fix using devices in PM Domains
  vfio: use match_string() helper
  vfio/mdev: Re-order sysfs attribute creation
  vfio/mdev: Check globally for duplicate devices
  vfio: platform: Make printed error messages more consistent
  vfio: platform: Fix reset module leak in error path
  sample: vfio bochs vbe display (host device for bochs-drm)
  sample: vfio mdev display - guest driver
  sample: vfio mdev display - host device
  vfio/type1: Fix task tracking for QEMU vCPU hotplug

Pull MIPS updates from James Hogan:
 "These are the main MIPS changes for 4.18.

  Rough overview:

   - MAINTAINERS: Add Paul Burton as MIPS co-maintainer

   - Misc: Generic compiler intrinsics, Y2038 improvements, Perf+MT fixes

   - Platform support: Netgear WNR1000 V3, Microsemi Ocelot integrated
     switch, Ingenic watchdog cleanups

  More detailed  summary:

  Maintainers:

   - Add Paul Burton as MIPS co-maintainer, as I soon won't have access
     to much MIPS hardware, nor enough time to properly maintain MIPS on
     my own.

  Miscellaneous:

   - Use generic GCC library routines from lib/
     - Add notrace to generic ucmpdi2 implementation
     - Rename compiler intrinsic selects to GENERIC_LIB_*
     - vmlinuz: Use generic ashldi3

   - y2038: Convert update/read_persistent_clock() to *_clock64()
     - sni: Remove read_persistent_clock()

   - perf: Fix perf with MT counting other threads
     - Probe for per-TC perf counters in cpu-probe.c
     - Use correct VPE ID for VPE tracing

  Minor cleanups:

   - Avoid unneeded built-in.a in DTS dirs

   - sc-debugfs: Re-use kstrtobool_from_user

   - memset.S: Reinstate delay slot indentation

   - VPE: Fix spelling "uneeded" -> "Unneeded"

  Platform support:

  BCM47xx:

   - Add support for Netgear WNR1000 V3

   - firmware: Support small NVRAM partitions

   - Use __initdata for LEDs platform data

  Ingenic:

   - Watchdog driver & platform code improvements:
     - Disable clock after stopping counter
     - Use devm_* functions
     - Drop module remove function
     - Move platform reset code to restart handler in driver
     - JZ4740: Convert watchdog instantiation to DT
     - JZ4780: Fix watchdog DT node
     - qi_lb60_defconfig: Enable watchdog driver

  Microsemi:

   - Ocelot: Add support for integrated switch
     - pcb123: Connect phys to ports"

* tag 'mips_4.18' of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux: (30 commits)
  MAINTAINERS: Add Paul Burton as MIPS co-maintainer
  MIPS: ptrace: Make FPU context layout comments match reality
  MIPS: memset.S: Reinstate delay slot indentation
  MIPS: perf: Fix perf with MT counting other threads
  MIPS: perf: Use correct VPE ID when setting up VPE tracing
  MIPS: perf: More robustly probe for the presence of per-tc counters
  MIPS: Probe for MIPS MT perf counters per TC
  MIPS: mscc: Connect phys to ports on ocelot_pcb123
  MIPS: mscc: Add switch to ocelot
  MIPS: JZ4740: Drop old platform reset code
  MIPS: qi_lb60: Enable the jz4740-wdt driver
  MIPS: JZ4780: dts: Fix watchdog node
  MIPS: JZ4740: dts: Add bindings for the jz4740-wdt driver
  watchdog: JZ4740: Drop module remove function
  watchdog: JZ4740: Register a restart handler
  watchdog: JZ4740: Use devm_* functions
  watchdog: JZ4740: Disable clock after stopping counter
  MIPS: VPE: Fix spelling mistake: "uneeded" -> "unneeded"
  MIPS: Re-use kstrtobool_from_user()
  MIPS: Convert update_persistent_clock() to update_persistent_clock64()
  ...

Fixes: a2225d93 ("autofs: remove left-over autofs4 stubs")
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Pull KVM updates from Paolo Bonzini:
 "Small update for KVM:

  ARM:
   - lazy context-switching of FPSIMD registers on arm64
   - "split" regions for vGIC redistributor

  s390:
   - cleanups for nested
   - clock handling
   - crypto
   - storage keys
   - control register bits

  x86:
   - many bugfixes
   - implement more Hyper-V super powers
   - implement lapic_timer_advance_ns even when the LAPIC timer is
     emulated using the processor's VMX preemption timer.
   - two security-related bugfixes at the top of the branch"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (79 commits)
  kvm: fix typo in flag name
  kvm: x86: use correct privilege level for sgdt/sidt/fxsave/fxrstor access
  KVM: x86: pass kvm_vcpu to kvm_read_guest_virt and kvm_write_guest_virt_system
  KVM: x86: introduce linear_{read,write}_system
  kvm: nVMX: Enforce cpl=0 for VMX instructions
  kvm: nVMX: Add support for "VMWRITE to any supported field"
  kvm: nVMX: Restrict VMX capability MSR changes
  KVM: VMX: Optimize tscdeadline timer latency
  KVM: docs: nVMX: Remove known limitations as they do not exist now
  KVM: docs: mmu: KVM support exposing SLAT to guests
  kvm: no need to check return value of debugfs_create functions
  kvm: Make VM ioctl do valloc for some archs
  kvm: Change return type to vm_fault_t
  KVM: docs: mmu: Fix link to NPT presentation from KVM Forum 2008
  kvm: x86: Amend the KVM_GET_SUPPORTED_CPUID API documentation
  KVM: x86: hyperv: declare KVM_CAP_HYPERV_TLBFLUSH capability
  KVM: x86: hyperv: simplistic HVCALL_FLUSH_VIRTUAL_ADDRESS_{LIST,SPACE}_EX implementation
  KVM: x86: hyperv: simplistic HVCALL_FLUSH_VIRTUAL_ADDRESS_{LIST,SPACE} implementation
  KVM: introduce kvm_make_vcpus_request_mask() API
  KVM: x86: hyperv: do rep check for each hypercall separately
  ...

Pull NFS client updates from Trond Myklebust:
 "Highlights include:

  Stable fixes:

   - Fix a 1-byte stack overflow in nfs_idmap_read_and_verify_message

   - Fix a hang due to incorrect error returns in rpcrdma_convert_iovs()

   - Revert an incorrect change to the NFSv4.1 callback channel

   - Fix a bug in the NFSv4.1 sequence error handling

  Features and optimisations:

   - Support for piggybacking a LAYOUTGET operation to the OPEN compound

   - RDMA performance enhancements to deal with transport congestion

   - Add proper SPDX tags for NetApp-contributed RDMA source

   - Do not request delegated file attributes (size+change) from the
     server

   - Optimise away a GETATTR in the lookup revalidate code when doing
     NFSv4 OPEN

   - Optimise away unnecessary lookups for rename targets

   - Misc performance improvements when freeing NFSv4 delegations

  Bugfixes and cleanups:

   - Try to fail quickly if proto=rdma

   - Clean up RDMA receive trace points

   - Fix sillyrename to return the delegation when appropriate

   - Misc attribute revalidation fixes

   - Immediately clear the pNFS layout on a file when the server returns
     ESTALE

   - Return NFS4ERR_DELAY when delegation/layout recalls fail due to
     igrab()

   - Fix the client behaviour on NFS4ERR_SEQ_FALSE_RETRY"

* tag 'nfs-for-4.18-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs: (80 commits)
  skip LAYOUTRETURN if layout is invalid
  NFSv4.1: Fix the client behaviour on NFS4ERR_SEQ_FALSE_RETRY
  NFSv4: Fix a typo in nfs41_sequence_process
  NFSv4: Revert commit 5f83d86c ("NFSv4.x: Fix wraparound issues..")
  NFSv4: Return NFS4ERR_DELAY when a layout recall fails due to igrab()
  NFSv4: Return NFS4ERR_DELAY when a delegation recall fails due to igrab()
  NFSv4.0: Remove transport protocol name from non-UCS client ID
  NFSv4.0: Remove cl_ipaddr from non-UCS client ID
  NFSv4: Fix a compiler warning when CONFIG_NFS_V4_1 is undefined
  NFS: Filter cache invalidation when holding a delegation
  NFS: Ignore NFS_INO_REVAL_FORCED in nfs_check_inode_attributes()
  NFS: Improve caching while holding a delegation
  NFS: Fix attribute revalidation
  NFS: fix up nfs_setattr_update_inode
  NFSv4: Ensure the inode is clean when we set a delegation
  NFSv4: Ignore NFS_INO_REVAL_FORCED in nfs4_proc_access
  NFSv4: Don't ask for delegated attributes when adding a hard link
  NFSv4: Don't ask for delegated attributes when revalidating the inode
  NFS: Pass the inode down to the getattr() callback
  NFSv4: Don't request size+change attribute if they are delegated to us
  ...

Pull nfsd updates from Bruce Fields:
 "A relatively quiet cycle for nfsd.

  The largest piece is an RDMA update from Chuck Lever with new trace
  points, miscellaneous cleanups, and streamlining of the send and
  receive paths.

  Other than that, some miscellaneous bugfixes"

* tag 'nfsd-4.18' of git://linux-nfs.org/~bfields/linux: (26 commits)
  nfsd: fix error handling in nfs4_set_delegation()
  nfsd: fix potential use-after-free in nfsd4_decode_getdeviceinfo
  Fix 16-byte memory leak in gssp_accept_sec_context_upcall
  svcrdma: Fix incorrect return value/type in svc_rdma_post_recvs
  svcrdma: Remove unused svc_rdma_op_ctxt
  svcrdma: Persistently allocate and DMA-map Send buffers
  svcrdma: Simplify svc_rdma_send()
  svcrdma: Remove post_send_wr
  svcrdma: Don't overrun the SGE array in svc_rdma_send_ctxt
  svcrdma: Introduce svc_rdma_send_ctxt
  svcrdma: Clean up Send SGE accounting
  svcrdma: Refactor svc_rdma_dma_map_buf
  svcrdma: Allocate recv_ctxt's on CPU handling Receives
  svcrdma: Persistently allocate and DMA-map Receive buffers
  svcrdma: Preserve Receive buffer until svc_rdma_sendto
  svcrdma: Simplify svc_rdma_recv_ctxt_put
  svcrdma: Remove sc_rq_depth
  svcrdma: Introduce svc_rdma_recv_ctxt
  svcrdma: Trace key RDMA API events
  svcrdma: Trace key RPC/RDMA protocol events
  ...

KVM_X86_DISABLE_EXITS_HTL really refers to exit on halt.
Obviously a typo: should be named KVM_X86_DISABLE_EXITS_HLT.

Fixes: caa057a2 ("KVM: X86: Provide a capability to disable HLT intercepts")
Cc: stable@vger.kernel.org
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

The functions that were used in the emulation of fxrstor, fxsave, sgdt and
sidt were originally meant for task switching, and as such they did not
check privilege levels.  This is very bad when the same functions are used
in the emulation of unprivileged instructions.  This is CVE-2018-10853.

The obvious fix is to add a new argument to ops->read_std and ops->write_std,
which decides whether the access is a "system" access or should use the
processor's CPL.

Fixes: 129a72a0 ("KVM: x86: Introduce segmented_write_std", 2017-01-12)
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Int the next patch the emulator's .read_std and .write_std callbacks will
grow another argument, which is not needed in kvm_read_guest_virt and
kvm_write_guest_virt_system's callers.  Since we have to make separate
functions, let's give the currently existing names a nicer interface, too.

Fixes: 129a72a0 ("KVM: x86: Introduce segmented_write_std", 2017-01-12)
Cc: stable@vger.kernel.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Wrap the common invocation of ctxt->ops->read_std and ctxt->ops->write_std, so
as to have a smaller patch when the functions grow another argument.

Fixes: 129a72a0 ("KVM: x86: Introduce segmented_write_std", 2017-01-12)
Cc: stable@vger.kernel.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

VMX instructions executed inside a L1 VM will always trigger a VM exit
even when executed with cpl 3. This means we must perform the
privilege check in software.

Fixes: 70f3aac9("kvm: nVMX: Remove superfluous VMX instruction fault checks")
Cc: stable@vger.kernel.org
Signed-off-by: Felix Wilhelm <fwilhelm@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Currently, when IO to DS fails, client returns the layout and
retries against the MDS. However, then on umounting (inode eviction)
it returns the layout again.

This is because pnfs_return_layout() was changed in
commit d78471d3 ("pnfs/blocklayout: set PNFS_LAYOUTRETURN_ON_ERROR")
to always set NFS_LAYOUT_RETURN_REQUESTED so even if we returned
the layout, it will be returned again. Instead, let's also check
if we have already marked the layout invalid.
Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

Pull ARM SoC late updates from Olof Johansson:
 "This is a branch with a few merge requests that either came in late,
  or took a while longer for us to review and merge than usual and thus
  cut it a bit close to the merge window. We stage them in a separate
  branch and if things look good, we still send them up -- and that's
  the case here.

  This is mostly DT additions for Renesas platforms, adding IP block
  descriptions for existing and new SoCs.

  There are also some driver updates for Qualcomm platforms for SMEM/QMI
  and GENI, which is their generalized serial protocol interface"

* tag 'armsoc-late' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (186 commits)
  soc: qcom: smem: introduce qcom_smem_virt_to_phys()
  soc: qcom: qmi: fix a buffer sizing bug
  MAINTAINERS: Update pattern for qcom_scm
  soc: Unconditionally include qcom Makefile
  soc: qcom: smem: check sooner in qcom_smem_set_global_partition()
  soc: qcom: smem: fix qcom_smem_set_global_partition()
  soc: qcom: smem: fix off-by-one error in qcom_smem_alloc_private()
  soc: qcom: smem: byte swap values properly
  soc: qcom: smem: return proper type for cached entry functions
  soc: qcom: smem: fix first cache entry calculation
  soc: qcom: cmd-db: Make endian-agnostic
  drivers: qcom: add command DB driver
  arm64: dts: renesas: salvator-common: Add ADV7482 support
  ARM: dts: r8a7740: Add CEU1
  ARM: dts: r8a7740: Add CEU0
  arm64: dts: renesas: salvator-common: enable VIN
  arm64: dts: renesas: r8a77970: add VIN and CSI-2 nodes
  arm64: dts: renesas: r8a77965: add VIN and CSI-2 nodes
  arm64: dts: renesas: r8a7796: add VIN and CSI-2 nodes
  arm64: dts: renesas: r8a7795-es1: add CSI-2 node
  ...

Pull ARM SoC driver updates from Olof Johansson:
 "This contains platform-related driver updates for ARM and ARM64.

  Highlights:

   - ARM SCMI (System Control & Management Interface) driver cleanups

   - Hisilicon support for LPC bus w/ ACPI

   - Reset driver updates for several platforms: Uniphier,

   - Rockchip power domain bindings and hardware descriptions for
     several SoCs.

   - Tegra memory controller reset improvements"

* tag 'armsoc-drivers' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (59 commits)
  ARM: tegra: fix compile-testing PCI host driver
  soc: rockchip: power-domain: add power domain support for px30
  dt-bindings: power: add binding for px30 power domains
  dt-bindings: power: add PX30 SoCs header for power-domain
  soc: rockchip: power-domain: add power domain support for rk3228
  dt-bindings: power: add binding for rk3228 power domains
  dt-bindings: power: add RK3228 SoCs header for power-domain
  soc: rockchip: power-domain: add power domain support for rk3128
  dt-bindings: power: add binding for rk3128 power domains
  dt-bindings: power: add RK3128 SoCs header for power-domain
  soc: rockchip: power-domain: add power domain support for rk3036
  dt-bindings: power: add binding for rk3036 power domains
  dt-bindings: power: add RK3036 SoCs header for power-domain
  dt-bindings: memory: tegra: Remove Tegra114 SATA and AFI reset definitions
  memory: tegra: Remove Tegra114 SATA and AFI reset definitions
  memory: tegra: Register SMMU after MC driver became ready
  soc: mediatek: remove unneeded semicolon
  soc: mediatek: add a fixed wait for SRAM stable
  soc: mediatek: introduce a CAPS flag for scp_domain_data
  soc: mediatek: reuse regmap_read_poll_timeout helpers
  ...

Pull ARM SoC device tree updates from Olof Johansson:
 "As always, a large number of DT updates. Too many to enumerate them
  all, but at a glance:

  New SoCs introduced in this release:

   - Amlogic:
      + Meson 8M2 SoC, a.k.a. S812. A quad Cortex-A9 SoC used in some
        set top boxes and other products.

   - Mediatek:
      + MT7623A, which is a flavor of the MT7623 family with other
        on-chip ethernet options.

   - Qualcomm:
      + SDM845, a.k.a Snapdragon 845, an 4+4-core Kryo 385/845
        (Cortex-A75/A55 derivative) SoC that's one of the current
        high-end mobile SoCs.

        It's great to see mainline support for it. So far, you can't do
        much with it, since a lot of peripherals are not yet in the DTs
        but driver support for USB, GPU and other pieces are starting to
        trickle in. This might end up being a well-supported SoC
        upstream if the momentum keeps up.

   - Renesas:
      + R8A77990, a.k.a R-Car E3, a new automotive
        entertainment-targeted SoC. Currently only one Cortex-A53 CPU is
        enabled, we are eagerly awaiting more. So far, basic drivers
        such as serial, gpios, PMU and ethernet are enabled.
      + R8A77470, a.k.a. RZ/G1C, a new dual Cortex-A7 SoC with PowerVR
        GPU. Same here, basic set of drivers such as serial, gpios and
        ethernet enabled, and SMP support is also forthcoming.

   - STMicroelectronics:
      + STM32F469, very similar tih STM32F429 but with display support

  Enhancements to SoCs/platforms (DTS contents, some driver portions
  might not be in yet):
   - Allwinner sun8i (h3/a33/a83t) SMP, DVFS tweaks, misc
   - Amlogic Meson: I2C, UFS, TDM, GPIO external interrupts, MMC resets
   - Hisilicon hi3660: Thermal cooling, CPU frequency scaling, mailbox interfaces
   - Marvell Berlin2CD: SMP support, thermal sensors
   - Mediatek MT7623: Highspeed DMA, audio support
   - Qualcomm IPQ8074 PCIe support, MSM8996 UFS support
   - Renesas: Watchdog and PMU support across many platforms
   - Rockchip RK3399: USB3 OTG support
   - Samsung Exynos: Audio-over-HDMI on Odroid X/X2/U3
   - STMicro STM32: Lots of peripherals added to STM32MP175C
   - Uniphier: Ethernet support

  New boards:
   - Allwinner A20: Olimex A20-SOM-EVB-eMMC variant
   - Allwinner H2+: Libre Computer ALL-H3-CC (h2+ version)
   - Allwinner A33: Nintendo NES/SuperNES Classic Edition
   - Aspeed: S2600WF, Inventec Lanyang BMC, Portwell Neptune
   - Berlin2CD: Valve Steam Link
   - Broadcom BCM5301X: Luxul XAP-1610 and XWR-3150 V1
   - Broadcom: Raspberry Pi 3 B+
   - Mediatek MT7623N and MT7623A: reference boards
   - Meson 8M2: Tronsmart MXIII Plus
   - NXP i.MX: Engicam i.CoreM6, DHCOM iMX6 SOM, BTicino i.MX6DL Mamoj
   - Qualcomm MSM8974: Sony Xperia Z1 Compact support
   - Qualcomm SDM845: MTP development board
   - Renesas: Ebisu R8A77990 board
   - Renesas RZ/G1C: iwg23s: iWave G235-SDB
   - TI am335x: Pocketbeagle support"

* tag 'armsoc-dt' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (448 commits)
  ARM: dts: aspeed: Fix hwrng register address
  arm64: dts: sprd: whale2: Add the rtc enable clock for watchdog
  arm64: dts: sprd: Add GPIO and GPIO keys device nodes
  arm64: dts: sprd: fix typo in 'remote-endpoint'
  arm64: dts: apq8096-db820c: Removed bt-en-1-8v regulator
  arm64: dts: fix regulator property name for wlan pcie endpoint
  arm64: dts: qcom: msm8996: Use UFS_GDSC for UFS
  ARM: dts: pxa3xx: fix MMC clocks
  ARM: pxa: dts: add pin definitions for extended GPIOs
  ARM: pxa: dts: add gpio-ranges to gpio controller
  ARM: dts: ipq8074: Enable few peripherals for hk01 board
  ARM: dts: ipq8074: Add pcie nodes
  ARM: dts: ipq8074: Add peripheral nodes
  ARM: dts: ipq4019: Add qcom-ipq4019-ap.dk07.1-c2 board file
  ARM: dts: ipq4019: Add qcom-ipq4019-ap.dk07.1-c1 board file
  ARM: dts: ipq4019: Add ipq4019-ap.dk07.1 common data
  ARM: dts: ipq4019: Add qcom-ipq4019-ap.dk04.1-c3 board file
  ARM: dts: ipq4019: Add ipq4019-ap.dk04.1-c1 board file
  ARM: dts: ipq4019: Add ipq4019-ap.dk04.dtsi
  ARM: dts: ipq4019: Change the max opp frequency
  ...

Pull ARM SoC platform updates from Olof Johansson:
 "Here are the main updates for SoC support (besides DT additions) for
  ARM 32- and 64-bit platforms. The branch also contains defconfig
  updates to turn on drivers and options as needed on the various
  platforms.

  The largest parts of the delta are from cleanups moving platform data
  and board file setup of TI platforms to ti-sysc bus drivers. There are
  also some sweeping changes of eeprom and nand setup on Davinci, i.MX
  and other platforms.

  Samsung is removing support for Exynos5440, which was an oddball SoC
  that hasn't been seen much use in designs.

  Renesas is adding support for new SoCs (R-Car E3, RZ/G1C and RZ/N1D).

  Linus Walleij is also removing support for ux500 (Sony Ericsson)
  U8540/9540 SoCs that never made it to significant mass production and
  products"

* tag 'armsoc-soc' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (133 commits)
  MAINTAINERS: add NXP linux team maillist as i.MX reviewer
  ARM: stm32: Don't select DMA unconditionally on STM32MP157C
  arm64: defconfig: Enable PCIe on msm8996 and db820c
  ARM: pxa3xx: enable external wakeup pins
  ARM: pxa: stargate2: use device properties for at24 eeprom
  arm64: defconfig: Enable HISILICON_LPC
  arm64: defconfig: enable drivers for Poplar support
  arm64: defconfig: Enable UFS on msm8996
  ARM: berlin: switch to SPDX license identifier
  arm: berlin: remove non-necessary flush_cache_all()
  ARM: berlin: extend BG2CD Kconfig entry
  OMAP: CLK: CLKSRC: Add suspend resume hooks
  ARM: AM43XX: Add functions to save/restore am43xx control registers
  ASoC: ams_delta: use GPIO lookup table
  ARM: OMAP1: ams-delta: add GPIO lookup tables
  bus: ti-sysc: Fix optional clocks array access
  ARM: OMAP2+: Make sure LOGICRETSTATE bits are not cleared
  ARM: OMAP2+: prm44xx: Inroduce cpu_pm notifiers for context save/restore
  ARM: OMAP2+: prm44xx: Introduce context save/restore for am43 PRCM IO
  ARM: OMAP2+: powerdomain: Introduce cpu_pm notifiers for context save/restore
  ...

Merging in defconfig updates. It's useful to keep them separate during
development but little point in merging them upstream on their own.

* next/defconfig: (40 commits)
  arm64: defconfig: Enable PCIe on msm8996 and db820c
  arm64: defconfig: Enable HISILICON_LPC
  arm64: defconfig: enable drivers for Poplar support
  arm64: defconfig: Enable UFS on msm8996
  arm64: defconfig: enable the Armada thermal driver
  ARM: multi_v7_defconfig: Enable RENESAS_WDT
  ARM: shmobile: defconfig: Enable RENESAS_WDT_GEN
  arm64: defconfig: enable R8A77990 SoC
  ARM: multi_v7_defconfig: Add STM32F7 I2C & STM32 USBPHYC support
  arm64: defconfig: Increase CMA size for VC4
  arm64: defconfig: enable rockchip efuse
  ARM: imx_v6_v7_defconfig: Select CONFIG_GPIO_MAX732X by default
  ARM: imx_v6_v7_defconfig: enable imx6sll by default
  arm64: defconfig: Enable CONFIG_SND_AUDIO_GRAPH_CARD
  arm64: defconfig: makes SND_SIMPLE_CARD to module
  ARM: multi_v7_defconfig: enable the Meson PWM controller
  arm: multi_v7_defconfig: enable the Amlogic Meson I2C driver
  arm64: defconfig: enable CONFIG_SPI_ARMADA_3700
  arm64: defconfig: Enable CONFIG_PINCTRL_MT7622 by default
  arm64: defconfig: Enable bluetooth USB support
  ...
Signed-off-by: Olof Johansson <olof@lixom.net>

Pull watchdog updates from Wim Van Sebroeck:
 "Core:
   - simplify getting .drvdata

  renesas_wdt:
   - add support for the R8A77965 WDT and R-Car Gen2
   - remove R-Car M2-W ES2.x from blacklist

  sp805:
   - add restart handler

  hpwdt:
   - claim NMIs generated by iLO5

  mena21_wdt:
   - drop unnecessary mutex lock

  da9062 & da9063:
   - fixes and general cleanups"

* tag 'linux-watchdog-4.18-rc1' of git://www.linux-watchdog.org/linux-watchdog:
  watchdog: da9063: remove duplicated timeout_to_sel calls
  watchdog: da9063: rename helper function to avoid misunderstandings
  watchdog: da9062: remove unused code
  watchdog: da9063: Fix timeout handling during probe
  watchdog: da9063: Fix updating timeout value
  watchdog: da9063: Fix setting/changing timeout
  dt-bindings: watchdog: renesas-wdt: Add R-Car Gen2 support
  watchdog: mena21_wdt: Drop unnecessary mutex lock
  watchdog: renesas-wdt: Add support for the R8A77965 WDT
  watchdog: hpwdt: Claim NMIs generated by iLO5
  watchdog: sp805: add restart handler
  watchdog: renesas-wdt: Remove R-Car M2-W ES2.x from blacklist
  watchdog: simplify getting .drvdata

This reverts commit b468620f.

It turns out that this broke drm on AMD platforms. Quoting Gabriel C:
 "I can confirm reverting b468620f fixes
  that issue for me.

  The GPU is working fine with SME enabled.

  Now with working GPU :) I can also confirm performance is back to
  normal without doing any other workarounds"

Christan König analyzed it partially:
 "As far as I analyzed it we now get an -ENOMEM from dma_alloc_attrs()
  in drivers/gpu/drm/ttm/ttm_page_alloc_dma.c when IOMMU is enabled"

and Christoph Hellwig responded:
 "I think the prime issue is that dma_direct_alloc respects the dma
  mask. Which we don't need if actually using the iommu. This would be
  mostly harmless exept for the the SEV bit high in the address that
  makes the checks fail.

  For now I'd say revert this commit for 4.17/4.18-rc and I'll look into
  addressing these issues properly"
Reported-and-bisected-by: Gabriel C <nix.or.die@gmail.com>
Acked-by: Christoph Hellwig <hch@lst.de>
Cc: Christian König <christian.koenig@amd.com>
Cc: Michel Dänzer <michel.daenzer@amd.com>
Cc: Joerg Roedel <jroedel@suse.de>
Cc: Tom Lendacky <thomas.lendacky@amd.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: stable@kernel.org		# v4.17
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

ntb_transport_create_queue() is never called in atomic context.

ntb_transport_create_queue() is only called by ntb_netdev_probe(),
which is set as ".probe" in struct ntb_transport_client.

Despite never getting called from atomic context,
ntb_transport_create_queue() calls kzalloc_node() with GFP_ATOMIC,
which does not sleep for allocation.
GFP_ATOMIC is not necessary and can be replaced with GFP_KERNEL,
which can sleep and improve the possibility of sucessful allocation.

This is found by a static analysis tool named DCNS written by myself.
And I also manually check it
Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>

ntb_transport_setup_qp_mw() is never called in atomic context.

ntb_transport_setup_qp_mw() is only called by ntb_transport_link_work(),
which is set as a parameter of INIT_DELAYED_WORK()
in ntb_transport_probe().

Despite never getting called from atomic context,
ntb_transport_setup_qp_mw() calls kzalloc_node() with GFP_ATOMIC,
which does not sleep for allocation.
GFP_ATOMIC is not necessary and can be replaced with GFP_KERNEL,
which can sleep and improve the possibility of sucessful allocation.

This is found by a static analysis tool named DCNS written by myself.
And I also manually check it.
Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Acked-by: Serge Semin <fancer.lancer@gmail.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>

Change all references to skx to gen3 NTB.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>

Move the Intel hw gen3 code to its own source file. The ntb_hw_intel.c was
getting too large and makes it hard to maintain with future hardware
changes.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>

Break out the generation specific definitions to different headers
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>

Pull rpmsg updates from Bjorn Andersson:
 "This migrates rpmsg to use SPDX license headers and fixes a
  use-after-free in SMD"

* tag 'rpmsg-v4.18' of git://github.com/andersson/remoteproc:
  rpmsg: smd: do not use mananged resources for endpoints and channels
  rpmsg: char: Switch to SPDX license identifier
  rpmsg: glink: Switch to SPDX license identifier
  rpmsg: smd: Switch to SPDX license identifier
  rpmsg: virtio_rpmsg_bus: Switch to SPDX license identifier
  rpmsg: Switch to SPDX license identifier
  rpmsg: qcom_smd: Access APCS through mailbox framework
  rpmsg: Add driver_override device attribute for rpmsg_device

Pull remoteproc updates from Bjorn Andersson:
 "This brings a few minor fixes to the Davinci driver, drops a orphan
  include file from the StE cleanup done ealier and introduces support
  for booting the modem on Qualcomm's SDM845 platform"

* tag 'rproc-v4.18' of git://github.com/andersson/remoteproc:
  remoteproc: q6v5: Allow defining GLINK edge for mss remoteproc
  remoteproc: q6v5: Add support for mss remoteproc on SDM845
  remoteproc: q6v5: Introduce reset assert/deassert helper functions
  dt-bindings: remoteproc: Add Q6v5 Modem PIL binding for SDM845
  remoteproc: q6v5: Move proxy unvote to handover irq handler
  remoteproc: q6v5: Return irq from q6v5_request_irq()
  remoteproc/ste: remove abandoned include file
  remoteproc/davinci: use octal permissions for module_param()
  remoteproc/davinci: prepare and unprepare the clock where needed
  remoteproc/davinci: add the missing retval check for clk_enable()
  remoteproc: Remove depends on HAS_DMA in case of platform dependency
  remoteproc: Prevent incorrect rproc state on xfer mem ownership failure

Pull hwspinlock updates from Bjorn Andersson:
 "In addition to migrating the files to use SPDX license headers this
  introduces the ability for clients to operate a hwlock without the
  framework taking any additional locks"

* tag 'hwlock-v4.18' of git://github.com/andersson/remoteproc:
  hwspinlock/u8500: Switch to SPDX license identifier
  hwspinlock: sprd: Switch to SPDX license identifier
  hwspinlock/sirf: Switch to SPDX license identifier
  hwspinlock: qcom: Switch to SPDX license identifier
  hwspinlock/omap: Switch to SPDX license identifier
  hwspinlock/core: Switch to SPDX license identifier
  hwspinlock: Introduce one new mode for hwspinlock
  hwspinlock: Convert to use 'switch' statement

Al noted that the semantics of the ia64 perfmon mmap() is really wrong,
and can cause serious VM problems.  Before we look at whether it needs
to be fixed, let's see if anybody even uses it, and mark it broken.  It
may be that we can just remove the code entirely.
Reported-by: Alexander Viro <viro@zeniv.linux.org.uk>
Acked-by: Matthew Wilcox <willy@infradead.org>
Acked-by: Stephane Eranian <eranian@google.com>
Cc: Tony Luck <tony.luck@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Pull f2fs updates from Jaegeuk Kim:
 "In this round, we've mainly focused on discard, aka unmap, control
  along with fstrim for Android-specific usage model. In addition, we've
  fixed writepage flow which returned EAGAIN previously resulting in EIO
  of fsync(2) due to mapping's error state. In order to avoid old MM bug
  [1], we decided not to use __GFP_ZERO for the mapping for node and
  meta page caches. As always, we've cleaned up many places for future
  fsverity and symbol conflicts.

  Enhancements:
   - do discard/fstrim in lower priority considering fs utilization
   - split large discard commands into smaller ones for better responsiveness
   - add more sanity checks to address syzbot reports
   - add a mount option, fsync_mode=nobarrier, which can reduce # of cache flushes
   - clean up symbol namespace with modified function names
   - be strict on block allocation and IO control in corner cases

  Bug fixes:
   - don't use __GFP_ZERO for mappings
   - fix error reports in writepage to avoid fsync() failure
   - avoid selinux denial on CAP_RESOURCE on resgid/resuid
   - fix some subtle race conditions in GC/atomic writes/shutdown
   - fix overflow bugs in sanity_check_raw_super
   - fix missing bits on get_flags

  Clean-ups:
   - prepare the generic flow for future fsverity integration
   - fix some broken coding standard"

[1] https://lkml.org/lkml/2018/4/8/661

* tag 'f2fs-for-4.18' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs: (79 commits)
  f2fs: fix to clear FI_VOLATILE_FILE correctly
  f2fs: let sync node IO interrupt async one
  f2fs: don't change wbc->sync_mode
  f2fs: fix to update mtime correctly
  fs: f2fs: insert space around that ':' and ', '
  fs: f2fs: add missing blank lines after declarations
  fs: f2fs: changed variable type of offset "unsigned" to "loff_t"
  f2fs: clean up symbol namespace
  f2fs: make set_de_type() static
  f2fs: make __f2fs_write_data_pages() static
  f2fs: fix to avoid accessing cross the boundary
  f2fs: fix to let caller retry allocating block address
  disable loading f2fs module on PAGE_SIZE > 4KB
  f2fs: fix error path of move_data_page
  f2fs: don't drop dentry pages after fs shutdown
  f2fs: fix to avoid race during access gc_thread pointer
  f2fs: clean up with clear_radix_tree_dirty_tag
  f2fs: fix to don't trigger writeback during recovery
  f2fs: clear discard_wake earlier
  f2fs: let discard thread wait a little longer if dev is busy
  ...

There's no need to retain the fs/autofs4 directory for backward
compatibility.

Adding an AUTOFS4_FS fragment to the autofs Kconfig and a module alias
for autofs4 is sufficient for almost all cases. Not keeping fs/autofs4
remnants will prevent "insmod <path>/autofs4/autofs4.ko" from working
but this shouldn't be used in automation scripts rather than
modprobe(8).

There were some comments about things to look out for with the module
rename in the fs/autofs4/Kconfig that is removed by this patch, see the
commit patch if you are interested.

One potential problem with this change is that when the
fs/autofs/Kconfig fragment for AUTOFS4_FS is removed any AUTOFS4_FS
entries will be removed from the kernel config, resulting in no autofs
file system being built if there is no AUTOFS_FS entry also.

This would have also happened if the fs/autofs4 remnants had remained
and is most likely to be a problem with automated builds.

Please check your build configurations before the removal which will
occur after the next couple of kernel releases.
Acked-by: Ian Kent <raven@themaw.net>
[ With edits and commit message from Ian Kent ]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Pull backlight updates from Lee Jones:
 "Core Frameworks
   - Provide helpers to enable/disable backlight
   - Provide standard and devres versions OF find helpers

  New Drivers:
   - Add support for the Zodiac Inflight Innovations RAVE Supervisory
     Processor

  New Functionality:
   - Allow pwm-on/pwm-off delay to be specified via DT

  Bug Fixes:
   - Fix ordering of the power {en,dis}able and PWM {en,dis}able
     signals
   - Fix Device Tree node look-up"

* tag 'backlight-next-4.18' of git://git.kernel.org/pub/scm/linux/kernel/git/lee/backlight:
  backlight: as3711_bl: Fix Device Tree node leaks
  backlight: tps65217_bl: Fix Device Tree node lookup
  backlight: max8925_bl: Fix Device Tree node lookup
  backlight: as3711_bl: Fix Device Tree node lookup
  MAINTAINERS: Add dri-devel for backlight subsystem patches
  backlight: Nuke BL_CORE_DRIVER1
  staging: fbtft: Stop using BL_CORE_DRIVER1
  backlight: pandora: Stop using BL_CORE_DRIVER1
  backlight: generic-bl: Remove DRIVER1 state
  backlight: Nuke unused backlight.props.state states
  backlight: otm3225a: Add support for ORISE OTM3225A LCD SoC
  backlight: pwm_bl: Don't use GPIOF_* with gpiod_get_direction
  pwm-backlight: Add support for PWM delays proprieties.
  dt-bindings: pwm-backlight: Add PWM delay proprieties.
  pwm-backlight: Enable/disable the PWM before/after LCD enable toggle.
  dt-bindings: backlight: Add binding for RAVE SP backlight driver
  backlight: Add RAVE SP backlight driver

Pull MFD updates from Lee Jones:
 "New Device Support:
   - Add support for AXP813 ADC to AXP20x
   - Add support for PM8005, PM8998 and PMI8998

  New Functionality:
   - Add support for Battery Power Supply to AXP813
   - Add support for SYSCON to SPARD SC27XX SPI
   - Add support for RTC to ChromeOS Embedded-Controller

  Fix-ups:
   - Remove unused code; exynos{4,5}-pmu, cros_ec, cros_ec_acpi_gpe
   - Remove duplicate error messages (-ENOMEM, etc); htc-i2cpld,
        janz-cmodio, max8997, rc5t583, sm501, smsc-ece1099, abx500-core,
        si476x-i2c, ti_am335x_tscadc, tps65090, tps6586x, tps65910,
        tps80031, twl6030-irq, viperboard
   - Succinctly use ptr to struct in sizeof(); rc5t583, abx500-core,
        sm501, smsc-ece1099
   - Simplify syntax for NULL ptr checking; abx500-core, sm501
   - No not unnecessarily initialise variables; tps65910, tps65910
   - Reorganise and simplify driver data; omap-usb-tll
   - Move to SPDX license statement; tps68470
   - Probe ADCs via DT; axp20x
   - Use new GPIOD API; arizona-core
   - Constify things; axp20x
   - Reduce code-size (use MACROS, etc); axp20x, omap-usb-host
   - Add DT support/docs; motorola-cpcap
   - Remove VLAs; rave-sp
   - Use devm_* managed resources; cros_ec
   - Interrogate HW for firmware version; rave-sp
   - Provide ACPI support for ChromeOS Embedded-Controller

  Bug Fixes:
   - Reorder ordered (enum) device list; tps65218
   - Only accept valid data from the offset; rave-sp
   - Refrain from copying junk from failed SPI read; cros_ec_dev
   - Fix potential memory leaks; pcf50633-core
   - Fix clock initialisation; twl-core
   - Fix build-issue; tps65911
   - Fix off-by-one error; tps65911
   - Fix code ordering issues; intel-lpss
   - Fix COMPILE_TEST related issues; pwm-stm32
   - Fix broken MMC card detection; asic3
   - Fix clocking related issues; intel-lpss-pci"

* tag 'mfd-next-4.18' of git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd: (84 commits)
  mfd: cros_ec: Remove unused __remove function
  mfd: wm97xx-core: Platform data can be NULL
  mfd: cros_ec_dev: Don't advertise junk features on failure
  mfd: cros_ec: Use devm_kzalloc for private data
  mfd: intel-lpss: Fix Intel Cannon Lake LPSS I2C input clock
  mfd: asic3: Fix broken MMC card detection
  mfd: timberdale: Fix spelling mistake "Uknown" -> "Unknown"
  mfd: omap-usb-host: Use match_string() helper
  mfd: stm32-timers: Fix pwm-stm32 linker issue with COMPILE_TEST
  pwm: stm32: Initialize raw local variables
  mfd: arizona: Update DT doc to support more standard Reset binding
  dt-bindings: mfd: Add bindings for DA9063L
  mfd: intel-lpss: Correct names of RESETS register bits
  mfd: qcom-spmi-pmic: Add support for pm8005, pm8998 and pmi8998
  mfd: intel-lpss: Program REMAP register in PIO mode
  mfd: cros_ec_i2c: Moving the system sleep pm ops to late
  mfd: cros_ec_i2c: Add ACPI module device table
  mfd: cros_ec_dev: Register shutdown function for debugfs
  mfd: cros_ec_dev: Register cros-ec-rtc driver as a subdevice
  mfd: cros_ec: Don't try to grab log when suspended
  ...

Pull drm msm updates and misc fixes from Dave Airlie:
 "I looked at Rob's msm tree, he kept it small due to being late, and it
  was in -next for a while before he was ill, so I think it should be
  fine.

  Otherwise this contains a set of i915 fixes and a v3d build fix, and
  vc4 leak fix"

* tag 'drm-next-2018-06-11' of git://anongit.freedesktop.org/drm/drm: (31 commits)
  drm/i915/icl: Don't update enabled dbuf slices struct until updated in hw
  drm/i915/icl: fix icl_unmap/map_plls_to_ports
  drm/i915: Remove bogus NV12 PLANE_COLOR_CTL setup
  drm/msm: Fix NULL deref on bind/probe deferral
  drm/msm: Switch to atomic_helper_commit()
  drm/msm: Remove msm_commit/worker, use atomic helper commit
  drm/msm: Issue queued events when disabling crtc
  drm/msm: Move implicit sync handling to prepare_fb
  drm/msm: Refactor complete_commit() to look more the helpers
  drm/msm: Don't subclass drm_atomic_state anymore
  drm/msm/mdp5: Use the new private_obj state
  drm/msm/mdp5: Add global state as a private atomic object
  drm/msm: use correct aspace pointer in msm_gem_put_iova()
  drm/msm: remove unbalanced mutex unlock
  drm/msm: don't deref error pointer in the msm_fbdev_create error path
  drm/msm/dsi: use correct enum in dsi_get_cmd_fmt
  drm/msm: Fix possible null dereference on failure of get_pages()
  drm/msm: Add modifier to mdp_get_format arguments
  drm/msm: Mark the crtc->state->event consumed
  drm/msm/dsi: implement auto PHY timing calculator for 10nm PHY
  ...

Two framebuffer device-node names were looked up during probe, but were
only used as flags to indicate the presence of two framebuffer device.

Drop the unused framebuffer name along with a likewise unused device
pointer from the driver data, and update the platform data to pass in
booleans instead of the framebuffer strings. This allows us do drop the
node references acquired during probe, which would otherwise leak.

Note that there are no other in-kernel users of the modified
platform-data fields.

Fixes: 59eb2b5e ("drivers/video/backlight/as3711_bl.c: add OF support")
Signed-off-by: Johan Hovold <johan@kernel.org>
Acked-by: Daniel Thompson <daniel.thompson@linaro.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>

Fix child-node lookup during probe, which ended up searching the whole
device tree depth-first starting at the parent rather than just matching
on its children.

This would only cause trouble if the child node is missing while there
is an unrelated node named "backlight" elsewhere in the tree.

Cc: stable <stable@vger.kernel.org>     # 3.7
Fixes: eebfdc17 ("backlight: Add TPS65217 WLED driver")
Signed-off-by: Johan Hovold <johan@kernel.org>
Acked-by: Daniel Thompson <daniel.thompson@linaro.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>

Fix child-node lookup during probe, which ended up searching the whole
device tree depth-first starting at the parent rather than just matching
on its children.

To make things worse, the parent mfd node was also prematurely freed,
while the child backlight node was leaked.

Cc: stable <stable@vger.kernel.org>     # 3.9
Fixes: 47ec340c ("mfd: max8925: Support dt for backlight")
Signed-off-by: Johan Hovold <johan@kernel.org>
Acked-by: Daniel Thompson <daniel.thompson@linaro.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>

Fix child-node lookup during probe, which ended up searching the whole
device tree depth-first starting at the parent rather than just matching
on its children.

To make things worse, the parent mfd node was also prematurely freed.

Cc: stable <stable@vger.kernel.org>     # 3.10
Fixes: 59eb2b5e ("drivers/video/backlight/as3711_bl.c: add OF support")
Signed-off-by: Johan Hovold <johan@kernel.org>
Acked-by: Daniel Thompson <daniel.thompson@linaro.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>

I soon won't have access to much MIPS hardware, nor enough time to
properly maintain MIPS on my own, so add Paul Burton as a co-maintainer.

Also add a link to a new shared git repository on kernel.org for
linux-next branches and pull request tags.
Signed-off-by: James Hogan <jhogan@kernel.org>
Acked-by: Paul Burton <paul.burton@mips.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Matt Redfearn <matt.redfearn@mips.com>
Cc: Maciej W. Rozycki <macro@mips.com>
Cc: Huacai Chen <chenhc@lemote.com>
Cc: Aaro Koskinen <aaro.koskinen@iki.fi>
Cc: John Crispin <john@phrozen.org>
Cc: Steven J. Hill <Steven.Hill@cavium.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/19473/