In hmm_mirror_unregister(), mm->hmm is set to NULL and then
mmu_notifier_unregister_no_release() is called.  That creates a small
window where mmu_notifier can call mmu_notifier_ops with mm->hmm equal to
NULL.  Fix this by first unregistering mmu notifier callbacks and then
setting mm->hmm to NULL.

Similarly in hmm_register(), set mm->hmm before registering mmu_notifier
callbacks so callback functions always see mm->hmm set.

Link: http://lkml.kernel.org/r/20181019160442.18723-4-jglisse@redhat.comSigned-off-by: Ralph Campbell <rcampbell@nvidia.com>
Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
Reviewed-by: John Hubbard <jhubbard@nvidia.com>
Reviewed-by: Jérôme Glisse <jglisse@redhat.com>
Reviewed-by: Balbir Singh <bsingharora@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Private ZONE_DEVICE pages use a special pte entry and thus are not
present.  Properly handle this case in map_pte(), it is already handled in
check_pte(), the map_pte() part was lost in some rebase most probably.

Without this patch the slow migration path can not migrate back to any
private ZONE_DEVICE memory to regular memory.  This was found after stress
testing migration back to system memory.  This ultimatly can lead to the
CPU constantly page fault looping on the special swap entry.

Link: http://lkml.kernel.org/r/20181019160442.18723-3-jglisse@redhat.comSigned-off-by: Ralph Campbell <rcampbell@nvidia.com>
Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
Reviewed-by: Balbir Singh <bsingharora@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Patch series "HMM updates, improvements and fixes", v2

Few fixes that only affect HMM users.  Improve the synchronization call
back so that we match was other mmu_notifier listener do and add proper
support to the new blockable flags in the process.

For curious folks here are branches to leverage HMM in various existing
device drivers:

https://cgit.freedesktop.org/~glisse/linux/log/?h=hmm-nouveau-v01
https://cgit.freedesktop.org/~glisse/linux/log/?h=hmm-radeon-v00
https://cgit.freedesktop.org/~glisse/linux/log/?h=hmm-intel-v00

More to come (amd gpu, Mellanox, ...)

I expect more of the preparatory work for nouveau will be merge in 4.20
(like we have been doing since 4.16) and i will wait until this patchset
is upstream before pushing the patches that actualy make use of HMM (to
avoid complex tree inter-dependency).

This patch (of 6):

Somehow utf=8 must have been broken.

Link: http://lkml.kernel.org/r/20181019160442.18723-2-jglisse@redhat.comSigned-off-by: Jérôme Glisse <jglisse@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Pull tracing updates from Steven Rostedt:
 "The biggest change here is the updates to kprobes

  Back in January I posted patches to create function based events.
  These were the events that you suggested I make to allow developers to
  easily create events in code where no trace event exists. After
  posting those changes for review, it was suggested that we implement
  this instead with kprobes.

  The problem with kprobes is that the interface is too complex and
  needs to be simplified. Masami Hiramatsu posted patches in March and
  I've been playing with them a bit. There's been a bit of clean up in
  the kprobe code that was inspired by the function based event patches,
  and a couple of enhancements to the kprobe event interface.

   - If the arch supports it (we added support for x86), you can place a
     kprobe event at the start of a function and use $arg1, $arg2, etc
     to reference the arguments of a function. (Before you needed to
     know what register or where on the stack the argument was).

   - The second is a way to see array of events. For example, if you
     reference a mac address, you can add:

	echo 'p:mac ip_rcv perm_addr=+574($arg2):x8[6]' > kprobe_events

     And this will produce:

	mac: (ip_rcv+0x0/0x140) perm_addr={0x52,0x54,0x0,0xc0,0x76,0xec}

  Other changes include

   - Exporting trace_dump_stack to modules

   - Have the stack tracer trace the entire stack (stop trying to remove
     tracing itself, as we keep removing too much).

   - Added support for SDT in uprobes"

[ SDT - "Statically Defined Tracing" are userspace markers for tracing.
  Let's not use random TLA's in explanations unless they are fairly
  well-established as generic (at least for kernel people) - Linus ]

* tag 'trace-v4.20' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace: (24 commits)
  tracing: Have stack tracer trace full stack
  tracing: Export trace_dump_stack to modules
  tracing: probeevent: Fix uninitialized used of offset in parse args
  tracing/kprobes: Allow kprobe-events to record module symbol
  tracing/kprobes: Check the probe on unloaded module correctly
  tracing/uprobes: Fix to return -EFAULT if copy_from_user failed
  tracing: probeevent: Add $argN for accessing function args
  x86: ptrace: Add function argument access API
  tracing: probeevent: Add array type support
  tracing: probeevent: Add symbol type
  tracing: probeevent: Unify fetch_insn processing common part
  tracing: probeevent: Append traceprobe_ for exported function
  tracing: probeevent: Return consumed bytes of dynamic area
  tracing: probeevent: Unify fetch type tables
  tracing: probeevent: Introduce new argument fetching code
  tracing: probeevent: Remove NOKPROBE_SYMBOL from print functions
  tracing: probeevent: Cleanup argument field definition
  tracing: probeevent: Cleanup print argument functions
  trace_uprobe: support reference counter in fd-based uprobe
  perf probe: Support SDT markers having reference counter (semaphore)
  ...

Pull tracing fixes from Steven Rostedt:
 "Masami had a couple more fixes to the synthetic events. One was a
  proper error return value, and the other is for the self tests"

* tag 'trace-v4.19-rc8-3' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
  selftests/ftrace: Fix synthetic event test to delete event correctly
  tracing: Return -ENOENT if there is no target synthetic event

Pull xen fixes from Juergen Gross:
 "Only several small fixes and cleanups this time"

* tag 'for-linus-4.20a-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
  xen: drop writing error messages to xenstore
  xen/pvh: don't try to unplug emulated devices
  add myself as reviewer for Xen support in Linux
  xen: remove redundant 'default n' from Kconfig
  xen/balloon: Support xend-based toolstack
  xen/pvh: increase early stack size
  xen: make xen_qlock_wait() nestable
  xen: fix race in xen_qlock_wait()
  xen/balloon: Grammar s/Is it/It is/
  xen: Make XEN_BACKEND selectable by DomU

Pull more ACPI updates from Rafael Wysocki:
 "Rework the handling of the P-unit semaphore on Intel Baytrail and
  Cherrytrail systems to avoid race conditions and excessive overhead
  related to it (Hans de Goede)"

* tag 'acpi-4.20-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  ACPI / PMIC: xpower: Add depends on IOSF_MBI to Kconfig entry
  i2c: designware: Cleanup bus lock handling
  ACPI / PMIC: xpower: Block P-Unit I2C access during read-modify-write
  x86: baytrail/cherrytrail: Rework and move P-Unit PMIC bus semaphore code

Pull more power management updates from Rafael Wysocki:
 "These remove a questionable heuristic from the menu cpuidle governor,
  fix a recent build regression in the intel_pstate driver, clean up ARM
  big-Little support in cpufreq and fix up hung task watchdog's
  interaction with system-wide power management transitions.

  Specifics:

   - Fix build regression in the intel_pstate driver that doesn't build
     without CONFIG_ACPI after recent changes (Dominik Brodowski).

   - One of the heuristics in the menu cpuidle governor is based on a
     function returning 0 most of the time, so drop it and clean up the
     scheduler code related to it (Daniel Lezcano).

   - Prevent the arm_big_little cpufreq driver from being used on ARM64
     which is not suitable for it and drop the arm_big_little_dt driver
     that is not used any more (Sudeep Holla).

   - Prevent the hung task watchdog from triggering during resume from
     system-wide sleep states by disabling it before freezing tasks and
     enabling it again after they have been thawed (Vitaly Kuznetsov)"

* tag 'pm-4.20-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  kernel: hung_task.c: disable on suspend
  cpufreq: remove unused arm_big_little_dt driver
  cpufreq: drop ARM_BIG_LITTLE_CPUFREQ support for ARM64
  cpufreq: intel_pstate: Fix compilation for !CONFIG_ACPI
  cpuidle: menu: Remove get_loadavg() from the performance multiplier
  sched: Factor out nr_iowait and nr_iowait_cpu

Pull more btrfs updates from David Sterba:
 "This contains a few minor updates and fixes that were under testing or
  arrived shortly after the merge window freeze, mostly stable material"

* tag 'for-4.20-part2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  Btrfs: fix use-after-free when dumping free space
  Btrfs: fix use-after-free during inode eviction
  btrfs: move the dio_sem higher up the callchain
  btrfs: don't run delayed_iputs in commit
  btrfs: fix insert_reserved error handling
  btrfs: only free reserved extent if we didn't insert it
  btrfs: don't use ctl->free_space for max_extent_size
  btrfs: set max_extent_size properly
  btrfs: reset max_extent_size properly
  MAINTAINERS: update my email address for btrfs
  btrfs: delayed-ref: extract find_first_ref_head from find_ref_head
  Btrfs: fix deadlock when writing out free space caches
  Btrfs: fix assertion on fsync of regular file when using no-holes feature
  Btrfs: fix null pointer dereference on compressed write path error

* pm-cpuidle:
  cpuidle: menu: Remove get_loadavg() from the performance multiplier
  sched: Factor out nr_iowait and nr_iowait_cpu

* pm-cpufreq:
  cpufreq: remove unused arm_big_little_dt driver
  cpufreq: drop ARM_BIG_LITTLE_CPUFREQ support for ARM64
  cpufreq: intel_pstate: Fix compilation for !CONFIG_ACPI

Pull rpmsg updates from Bjorn Andersson:
 "This migrates rpmsg_char to use read/write_iter to allow being
  operated using aio, removes the message size alignment requirements
  from glink, closes a potential memory leak in SMD and switches to
  %pOFn for printing device_node names"

* tag 'rpmsg-v4.20' of git://github.com/andersson/remoteproc:
  rpmsg: glink: smem: Support rx peak for size less than 4 bytes
  rpmsg: smd: fix memory leak on channel create
  rpmsg: glink: Remove chunk size word align warning
  rpmsg: Convert to using %pOFn instead of device_node.name
  rpmsg: char: Migrate to iter versions of read and write

Pull remoteproc updates from Bjorn Andersson:
 "This contains a series of patches that reworks the memory carveout
  handling in remoteproc, in order to allow this to be reused for
  statically allocated memory regions to be used for e.g. firmware.

  It adds support for audio DSP (both TZ-assisted and non-TZ assisted)
  and compute DSP on Qualcomm SDM845, TZ-assisted audio DSP, compute DSP
  and WiFi processor on Qualcomm QCS404 and through some renaming of the
  drivers cleans up the naming situation.

  Finally support for custom coreudmp segment handlers is added and is
  used in the Qualcomm modem remoteproc driver to gather memory dumps of
  the firmware"

* tag 'rproc-v4.20' of git://github.com/andersson/remoteproc: (36 commits)
  remoteproc: qcom: q6v5-mss: Register segments/dumpfn for coredump
  remoteproc: qcom: q6v5-mss: Add custom dump function for modem
  remoteproc: qcom: q6v5-mss: Refactor mba load/unload sequence
  remoteproc: Add mechanism for custom dump function assignment
  remoteproc: Introduce custom dump function for each remoteproc segment
  remoteproc: modify vring allocation to rely on centralized carveout allocator
  remoteproc: qcom: q6v5: shore up resource probe handling
  remoteproc: qcom: qcom_q6v5_adsp: Fix some return value check
  remoteproc: modify rproc_handle_carveout to support pre-registered region
  remoteproc: add helper function to check carveout device address
  remoteproc: add helper function to allocate rproc_mem_entry from reserved memory
  remoteproc: add alloc ops in rproc_mem_entry struct
  remoteproc: introduce rproc_find_carveout_by_name function
  remoteproc: introduce rproc_add_carveout function
  remoteproc: add helper function to allocate and init rproc_mem_entry struct
  remoteproc: add name in rproc_mem_entry struct
  remoteproc: add release ops in rproc_mem_entry struct
  remoteproc: add rproc_va_to_pa function
  remoteproc: configure IOMMU only if device address requested
  remoteproc: qcom: q6v5-mss: add SCM probe dependency
  ...

Pull ARM SoC platform updates from Arnd Bergmann:
 "A couple of platforms change hands in the MAINTAINERS file:

   - Linus Walleij lists himself for the ARM Reference platforms:
     versatile, vexpress, integrator and realview. He has been the main
     contributor for these for a while, and makes it official now.

   - Vladimir Zapolskiy takes over the LPC18xx platform from Joachim
     Eastwood

   - Manivannan Sadhasivam becomes a secondary maintainer for the
     Actions Semi machines

   - Nicolas Ferre lists updates the MAINTAINER listing for the AT91
     platform: Ludovic Desroches is now a co-maintainer for the
     platform, and several other people (Claudiu Beznea, Cristian
     Birsan, Eugen Hristev, Codrin Ciubotariu) take over individual
     device drivers.

  Thanks everyone for working on this, and welcome to the new
  maintainers!

  The "virt" platform on qemy or kvm can now be used in big-endian mode
  without additional tricks, thanks to Jason Donenfeld.

  Once again, we gain support for another NXP i.MX6 variant, this time
  it's the i.MX 6ULZ 32-bit single-core version.

  On arm64, we add support for two SoCs from Renesas: RZ/G2E (r8a774c0)
  and RZ/G2M (r8a774a1). These are described as microcontrollers on the
  manufacturer website, but appear to be rather powerful. The RZ/G2M is
  used on the reference board for the CIP Super Long Term Support (SLTS)
  Linux Kernels"

* tag 'armsoc-soc' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (54 commits)
  MAINTAINERS: Assign myself as a maintainer of ARM/LPC18XX architecture
  arm64: exynos: Enable generic power domain support
  MAINTAINERS: remove non-exsiting email address of Baoyou
  MAINTAINERS: fix pattern in ARM/Synaptics berlin SoC section
  MAINTAINERS: Drop dt-bindings/genpd/k2g.h
  ARM: samsung: Limit SAMSUNG_PM_CHECK config option to non-Exynos platforms
  arm64: actions: Enable PINCTRL in platforms Kconfig
  MAINTAINERS: Add entry for Actions Semi Owl SoCs DMA driver
  MAINTAINERS: Add entry for Actions Semiconductor Owl I2C driver
  MAINTAINERS: Update clock binding entry for Actions Semi Owl SoCs
  ARM: imx: add i.mx6ulz msl support
  ARM: Assume maintainership of ARM reference designs
  ARM: support big-endian for the virt architecture
  MAINTAINERS: sdhci: move the Microchip entry to proper location
  MAINTAINERS: move former ATMEL entries to proper MICROCHIP location
  MAINTAINERS: remove the / ATMEL string from MICROCHIP entries
  MAINTAINERS: iio: add co-maintainer to SAMA5D2-compatible ADC driver
  MAINTAINERS: pwm: add entry for Microchip pwm driver
  MAINTAINERS: dmaengine: add files to Microchip dma entry
  MAINTAINERS: USB: change maintainer for Microchip USBA gadget driver
  ...

Pull ARM SoC driver updates from Arnd Bergmann:
 "The most noteworthy SoC driver changes this time include:

   - The TEE subsystem gains an in-kernel interface to access the TEE
     from device drivers.

   - The reset controller subsystem gains a driver for the Qualcomm
     Snapdragon 845 Power Domain Controller.

   - The Xilinx Zynq platform now has a firmware interface for its
     platform management unit. This contains a firmware "ioctl"
     interface that was a little controversial at first, but the version
     we merged solved that by not exposing arbitrary firmware calls to
     user space.

   - The Amlogic Meson platform gains a "canvas" driver that is used for
     video processing and shared between different high-level drivers.

  The rest is more of the usual, mostly related to SoC specific power
  management support and core drivers in drivers/soc:

   - Several Renesas SoCs (RZ/G1N, RZ/G2M, R-Car V3M, RZ/A2M) gain new
     features related to power and reset control.

   - The Mediatek mt8183 and mt6765 SoC platforms gain support for their
     respective power management chips.

   - A new driver for NXP i.MX8, which need a firmware interface for
     power management.

   - The SCPI firmware interface now contains support estimating power
     usage of performance states

   - The NVIDIA Tegra "pmc" driver gains a few new features, in
     particular a pinctrl interface for configuring the pads.

   - Lots of small changes for Qualcomm, in particular the "smem" device
     driver.

   - Some cleanups for the TI OMAP series related to their sysc
     controller.

  Additional cleanups and bugfixes in SoC specific drivers include the
  Meson, Keystone, NXP, AT91, Sunxi, Actions, and Tegra platforms"

* tag 'armsoc-drivers' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (129 commits)
  firmware: tegra: bpmp: Implement suspend/resume support
  drivers: clk: Add ZynqMP clock driver
  dt-bindings: clock: Add bindings for ZynqMP clock driver
  firmware: xilinx: Add zynqmp IOCTL API for device control
  Documentation: xilinx: Add documentation for eemi APIs
  MAINTAINERS: imx: include drivers/firmware/imx path
  firmware: imx: add misc svc support
  firmware: imx: add SCU firmware driver support
  reset: Fix potential use-after-free in __of_reset_control_get()
  dt-bindings: arm: fsl: add scu binding doc
  soc: fsl: qbman: add interrupt coalesce changing APIs
  soc: fsl: bman_portals: defer probe after bman's probe
  soc: fsl: qbman: Use last response to determine valid bit
  soc: fsl: qbman: Add 64 bit DMA addressing requirement to QBMan
  soc: fsl: qbman: replace CPU 0 with any online CPU in hotplug handlers
  soc: fsl: qbman: Check if CPU is offline when initializing portals
  reset: qcom: PDC Global (Power Domain Controller) reset controller
  dt-bindings: reset: Add PDC Global binding for SDM845 SoCs
  reset: Grammar s/more then once/more than once/
  bus: ti-sysc: Just use SET_NOIRQ_SYSTEM_SLEEP_PM_OPS
  ...

Pull ARM SoC defconfig updates from Arnd Bergmann:
 "The defconfig changes are split out from the rest again. This time we
  have a number of changes for NXP i.MX and Renesas, including a cleanup
  of old options.

  Some smaller changes are for Socionext Uniphier, Allwinner, Qualcomm,
  Rockchip, Renesas, AT91, Hisilicon, and STM32. All of these just
  enable platform specific device drivers"

* tag 'armsoc-defconfig' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (40 commits)
  arm64: defconfig: Enable SERIAL_8250_OMAP
  arm64: defconfig: Enable TI_SCI related configs
  ARM: imx_v6_v7_defconfig: Select CONFIG_TMPFS_POSIX_ACL
  ARM: imx_v4_v5_defconfig: Remove unneeded options
  ARM: imx_v4_v5_defconfig: Re-sync defconfig
  ARM: mxs_defconfig: Remove unneeded options
  ARM: mxs_defconfig: Re-sync defconfig
  ARM: imx_v6_v7_defconfig: Remove unneeded options
  ARM: imx_v6_v7_defconfig: Re-sync defconfig
  ARM: imx_v6_v7_defconfig: select CONFIG_ARM_CPUIDLE by default
  ARM: imx_v6_v7_defconfig: Make usbnet drivers builtin for boot
  ARM: imx_v6_v7_defconfig: add CONFIG_FW_LOADER_USER_HELPER
  ARM: imx_v6_v7_defconfig: Select CONFIG_SENSORS_MC13783_ADC
  ARM: multi_v7_defconfig: enable CONFIG_MMC_UNIPHIER
  arm64: defconfig: enable CONFIG_MMC_UNIPHIER
  ARM: multi_v7_defconfig: Enable USB phys for UniPhier SoCs
  arm64: defconfig: Enable USB phys for UniPhier SoCs
  arm64: defconfig: enable Rockchip Innosilicon hdmiphy
  arm64: defconfig: Enable PCIEPORTBUS
  arm64: defconfig: enable HiSilicon HNS3 driver
  ...

Pull ARM SoC device tree updates from Arnd Bergmann:
 "There are close to 800 indivudal changesets in this branch again,
  which feels like a lot. There are particularly many changes for the
  NVIDIA Tegra platform this time, in fact more than it has seen in the
  two years since the v4.9 merge window. Aside from this, it's been
  fairly normal, with lots of changes going into Renesas R-CAR, NXP
  i.MX, Allwinner Sunxi, Samsung Exynos, and TI OMAP.

  Most of the changes are for adding new features into existing boards,
  for brevity I'm only mentioning completely new machines and SoCs here.
  For the first time I think we have (slightly) more new 64-bit hardware
  than 32-bit:

  Two boards get added for TI OMAP: Moxa UC-2101 is an industrial
  computer, see https://www.moxa.com/product/UC-2100.htm; GTA04A5 is a
  minor variation of the motherboards of the GTA04 phone, see
  https://shop.goldelico.com/wiki.php?page=GTA04A5

  Clearfog is a nice little board for quad-core Marvell Armada 8040
  network processor, see
  https://www.solid-run.com/marvell-armada-family/clearfog-gt-8k/

  Two additional server boards come with the Aspeed baseboard management
  controllers: Stardragon4800 is an arm64 reference platform made by HXT
  (based on Qualcomm's server chips), and TiogaPass is an Open Compute
  mainboard with x86 CPUs. Both use the ARM11 based AST2500 chips in the
  BMC.

  NXP i.MX usually sees a lot of new boards each release. This time
  there we only add one minor variant: ConnectCore 6UL SBC Pro uses the
  same SoM design as the ConnectCore 6UL SBC Express added later.
  However, there is a new chip, the i.MX6ULZ, which is an even smaller
  variant of the i.MX6ULL, with features removed. There is also support
  for the reference board design, the i.MX6ULZ 14x14 EVK.

  A new Raspberry Pi variant gets added, this one is the CM3 compute
  module based on bcm2837, it was launched in early 2017 but only now
  added to the kernel, both as 32-bit and as 64-bit files, as we tend to
  do for Raspberry Pi.

  On the Allwinner side, everything is again about cheap development
  boards, usually of the "Fruit Pi" variety. The new ones this time are:
   - Orange Pi Zero Plus2: http://www.orangepi.org/OrangePiZeroPlus2/
   - Orange Pi One Plus: http://www.orangepi.org/OrangePiOneplus/
   - Pine64 LTS: https://www.pine64.org/?product=pine-a64-lts
   - Banana Pi M2+ H5: http://www.banana-pi.org/m2plus.html
  The last one of these is now a 64-bit version of the earlier Banana Pi
  M2+ H3, with the same board layout.

  Similarly, for Rockchips, get get another variant of the 32-bit Asus
  Tinker board, the model 'S' based on rk3288, and three now boards
  based on the popular RK3399 chip:
   - ROC-RK3399-PC: https://libre.computer/products/boards/roc-rk3399-pc/
   - Rock960: https://www.96boards.org/product/rock960/
   - RockPro64: https://www.pine64.org/?page_id=61454
  These are all quite powerful boards with lots of RAM and I/O, and the
  RK3399 is the same chip used in several Chromebooks. Finally, we get
  support for the PX30 (aka rk3326) chip, which is based on the low-end
  64-bit Cortex-A35 CPU core. So far, only the evaluation board is
  supported.

  One more Banana Pi is added with a Mediatek chip: Banana Pi R64 is
  based on the MT7622 WiFi router platform, and the first product I've
  seen with a 64-bit Mediatek chip in that market:
  http://www.banana-pi.org/r64.html

  For HiSilicon, we gain support for the Hi3670 SoC and HiKey 370
  development board, which are similar to the Hi3660 and Hikey 360
  respectively, but add support for an NPU.

  Amlogic gets initial support for the Meson-G12A chip (S905D2), another
  quad-core Cortex-A53 SoC, and its evaluation platform. On the 32-bit
  side, we gain support for an actual end-user product, the Endless
  Computers Endless Mini based on Meson8b (S805), see
  https://endlessos.com/computers/

  Qualcomm adds support for their MSM8998 SoC and evaluation platform.
  This chip is commonly known as the Snapdragon 835, and is used in
  high-end phones as well as low-end laptops.

  For Renesas, a very bare support for the r8a774a1 (RZ/G2M) is added,
  but no boards for this one. However, we do add boards for the
  previously added r8a77965 (R-Car M3-N): the M3NULCB Kingfisher and the
  M3NULCB Starter Kit Pro.

  While we have lots of DT changes for NVIDIA to update the existing
  files, the only board that gets added is the Toradex Colibri T20 on
  Colibri Evaluation Board for the old Tegra2.

  Synaptics add support for their AS370 SoC, which is part of the
  (formerly Marvell) Berlin line of set-top-box chips used e.g. in the
  various Google Chromecast. Only the .dtsi gets added at this point, no
  actual machines"

* tag 'armsoc-dt' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (721 commits)
  ARM: dts: socfgpa: remove ethernet aliases from dtsi
  arm64: dts: stratix10: add ethernet aliases
  dt-bindings: mediatek: Add bindig for MT7623 IOMMU and SMI
  dt-bindings: mediatek: Add JPEG Decoder binding for MT7623
  dt-bindings: iommu: mediatek: Add binding for MT7623
  dt-bindings: clock: mediatek: add support for MT7623
  ARM: dts: mvebu: armada-385-db-88f6820-amc: auto-detect nand ECC properites
  ARM: dts: da850-lego-ev3: slow down A/DC as much as possible
  ARM: dts: da850-evm: Enable tca6416 on baseboard
  arm64: dts: uniphier: Add USB2 PHY nodes
  arm64: dts: uniphier: Add USB3 controller nodes
  ARM: dts: uniphier: Add USB2 PHY nodes
  ARM: dts: uniphier: Add USB3 controller nodes
  arm64: dts: meson-axg: s400: disable emmc
  arm64: dts: meson-axg: s400: add missing emmc pwrseq
  arm64: dts: clearfog-gt-8k: add PCIe slot description
  ARM: dts: at91: sama5d4_xplained: even nand memory partitions
  ARM: dts: at91: sama5d3_xplained: even nand memory partitions
  ARM: dts: at91: at91sam9x5cm: even nand memory partitions
  ARM: dts: at91: sama5d2_ptc_ek: fix bootloader env offsets
  ...

Pull parisc updates from Helge Deller:
 "Three small patches:

   - A boot fix for A500 machines, crash was caused by the new
     alternative patching code from this merge window (Dave)

   - Change __kernel_suseconds_t to match glibc on 64-bit parisc (Arnd)

   - Use constants instead of hard-coded numbers (me)"

* 'parisc-4.20-2' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux:
  parisc: Fix A500 boot crash
  parisc: Use LINUX_GATEWAY_SPACE constant in entry.S
  parisc64: change __kernel_suseconds_t to match glibc

Pull i2c updates from Wolfram Sang:
 "I2C has not so much stuff this time. Mostly driver enablement for new
  SoCs, some driver bugfixes, and some cleanups"

* 'i2c/for-4.20' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux: (35 commits)
  MAINTAINERS: add maintainer for Renesas RIIC driver
  i2c: sh_mobile: Remove dummy runtime PM callbacks
  i2c: uniphier-f: fix race condition when IRQ is cleared
  i2c: uniphier-f: fix occasional timeout error
  i2c: uniphier-f: make driver robust against concurrency
  i2c: i2c-qcom-geni: Simplify irq handler
  i2c: i2c-qcom-geni: Simplify tx/rx functions
  i2c: designware: Set IRQF_NO_SUSPEND flag for all BYT and CHT controllers
  i2c: mux: mlxcpld: simplify code to reach the adapter
  i2c: mux: ltc4306: simplify code to reach the adapter
  i2c: mux: pca954x: simplify code to reach the adapter
  i2c: core: remove level of indentation in i2c_transfer
  i2c: core: remove outdated DEBUG output
  i2c: zx2967: use core to detect 'no zero length' quirk
  i2c: tegra: use core to detect 'no zero length' quirk
  i2c: qup: use core to detect 'no zero length' quirk
  i2c: omap: use core to detect 'no zero length' quirk
  i2c: Convert to using %pOFn instead of device_node.name
  i2c: brcmstb: Allow enabling the driver on DSL SoCs
  eeprom: at24: fix unexpected timeout under high load
  ...

Pull media updates from Mauro Carvalho Chehab:

 - new dvb frontend driver: lnbh29

 - new sensor drivers: imx319 and imx 355

 - some old soc_camera driver renames to avoid conflict with new
   drivers

 - new i.MX Pixel Pipeline (PXP) mem-to-mem platform driver

 - a new V4L2 frontend for the FWHT codec

 - several other improvements, bug fixes, code cleanups, etc

* tag 'media/v4.20-1' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media: (289 commits)
  media: rename soc_camera I2C drivers
  media: cec: forgot to cancel delayed work
  media: vivid: Support 480p for webcam capture
  media: v4l2-tpg: fix kernel oops when enabling HFLIP and OSD
  media: vivid: Add 16-bit bayer to format list
  media: v4l2-tpg-core: Add 16-bit bayer
  media: pvrusb2: replace `printk` with `pr_*`
  media: venus: vdec: fix decoded data size
  media: cx231xx: fix potential sign-extension overflow on large shift
  media: dt-bindings: media: rcar_vin: add device tree support for r8a7744
  media: isif: fix a NULL pointer dereference bug
  media: exynos4-is: make const array config_ids static
  media: cx23885: make const array addr_list static
  media: ivtv: make const array addr_list static
  media: bttv-input: make const array addr_list static
  media: cx18: Don't check for address of video_dev
  media: dw9807-vcm: Fix probe error handling
  media: dw9714: Remove useless error message
  media: dw9714: Fix error handling in probe function
  media: cec: name for RC passthrough device does not need 'RC for'
  ...

The RIIC I2C controller is used in Renesas RZ/A SoCs.
Signed-off-by: Chris Brandt <chris.brandt@renesas.com>
[wsa: added documentation file]
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Platform drivers don't need dummy runtime PM callbacks that just return
success and non-NULL pm pointer in their struct device_driver in order
to have runtime PM happening. This has changed since following commits:

05aa55dd ("PM / Runtime: Lenient generic runtime pm callbacks")
543f2503 ("PM / platform_bus: Allow runtime PM by default")
8b313a38 ("PM / Platform: Use generic runtime PM callbacks directly")
Signed-off-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Reviewed-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

The current IRQ handler clears all the IRQ status bits when it bails
out. This is dangerous because it might clear away the status bits
that have just been set while processing the current handler. If this
happens, the IRQ event for the latest transfer is lost forever.

The IRQ status bits must be cleared *before* the next transfer is
kicked.

Fixes: 6a62974b ("i2c: uniphier_f: add UniPhier FIFO-builtin I2C driver")
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Currently, a timeout error could happen at a repeated START condition.

For a (non-repeated) START condition, the controller starts sending
data when the UNIPHIER_FI2C_CR_STA bit is set. However, for a repeated
START condition, the hardware starts running when the slave address is
written to the TX FIFO - the write to the UNIPHIER_FI2C_CR register is
actually unneeded.

Because the hardware is already running before the IRQ is enabled for
a repeated START, the driver may miss the IRQ event. In most cases,
this problem does not show up since modern CPUs are much faster than
the I2C transfer. However, it is still possible that a context switch
happens after the controller starts, but before the IRQ register is
set up.

To fix this,

 - Do not write UNIPHIER_FI2C_CR for repeated START conditions.

 - Enable IRQ *before* writing the slave address to the TX FIFO.

 - Disable IRQ for the current CPU while queuing up the TX FIFO;
   If the CPU is interrupted by some task, the interrupt handler
   might be invoked due to the empty TX FIFO before completing the
   setup.

Fixes: 6a62974b ("i2c: uniphier_f: add UniPhier FIFO-builtin I2C driver")
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

This is unlikely to happen, but it is possible for a CPU to enter
the interrupt handler just after wait_for_completion_timeout() has
expired. If this happens, the hardware is accessed from multiple
contexts concurrently.

Disable the IRQ after wait_for_completion_timeout(), and do nothing
from the handler when the IRQ is disabled.

Fixes: 6a62974b ("i2c: uniphier_f: add UniPhier FIFO-builtin I2C driver")
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Commit 1ff2e1a4 ("HID: input: Create a utility class for counting
scroll events") created the helper function

    hid_scroll_counter_handle_scroll()

to handle high-res scroll events and also expose them as regular wheel
events.

But the resulting algorithm was unstable, and causes scrolling to be
very unreliable.  When you hit the half-way mark of the highres
multiplier, small highres movements will incorrectly translate into big
traditional wheel movements, causing odd jitters.

Simplify the code and make the output stable.

NOTE! I'm pretty sure this will need further tweaking.  But this at
least turns a unusable mouse wheel on my Logitech MX Anywhere 2S into
a usable one.

Cc: Jiri Kosina <jikos@kernel.org>
Cc: Harry Cutts <hcutts@chromium.org>
Cc: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Cc: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Pull tty/serial updates from Greg KH:
 "Here is the big tty and serial pull request for 4.20-rc1

  Lots of little things here, including a merge from the SPI tree in
  order to keep things simpler for everyone to sync around for one
  platform.

  Major stuff is:

   - tty buffer clearing after use

   - atmel_serial fixes and additions

   - xilinx uart driver updates

  and of course, lots of tiny fixes and additions to individual serial
  drivers.

  All of these have been in linux-next with no reported issues for a
  while"

* tag 'tty-4.20-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty: (66 commits)
  of: base: Change logic in of_alias_get_alias_list()
  of: base: Fix english spelling in of_alias_get_alias_list()
  serial: sh-sci: do not warn if DMA transfers are not supported
  serial: uartps: Do not allow use aliases >= MAX_UART_INSTANCES
  tty: check name length in tty_find_polling_driver()
  serial: sh-sci: Add r8a77990 support
  tty: wipe buffer if not echoing data
  tty: wipe buffer.
  serial: fsl_lpuart: Remove the alias node dependence
  TTY: sn_console: Replace spin_is_locked() with spin_trylock()
  Revert "serial:serial_core: Allow use of CTS for PPS line discipline"
  serial: 8250_uniphier: add auto-flow-control support
  serial: 8250_uniphier: flatten probe function
  serial: 8250_uniphier: remove unused "fifo-size" property
  dt-bindings: serial: sh-sci: Document r8a7744 bindings
  serial: uartps: Fix missing unlock on error in cdns_get_id()
  tty/serial: atmel: add ISO7816 support
  tty/serial_core: add ISO7816 infrastructure
  serial:serial_core: Allow use of CTS for PPS line discipline
  serial: docs: Fix filename for serial reference implementation
  ...

Pull staging/IIO driver updates from Greg KH:
 "Here is the big staging and IIO driver pull request for 4.20-rc1.

  There are lots of things here, we ended up adding more lines than
  removing, thanks to a large influx of Comedi National Instrument
  device support. Someday soon we need to get comedi out of staging...

  Other than the comedi drivers, the "big" things here are:

   - new iio drivers

   - delete dgnc driver (no one used it and no one had the hardware
     anymore)

   - vbox driver updates and fixes

   - erofs fixes

   - tons and tons of tiny checkpatch fixes for almost all staging
     drivers

  All of these have been in linux-next, with the last few happening a
  bit "late" due to them getting stuck on my laptop during travel to the
  Mantainers summit"

* tag 'staging-4.20-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging: (690 commits)
  staging: gasket: Fix sparse "incorrect type in assignment" warnings.
  staging: gasket: remove debug logs for callback invocation
  staging: gasket: remove debug logs in page table mapping calls
  staging: rtl8188eu: core: Use sizeof(*p) instead of sizeof(struct P) for memory allocation
  staging: ks7010: Remove extra blank line
  staging: gasket: Remove extra blank line
  staging: media: davinci_vpfe: Fix spelling mistake in enum
  staging: speakup: Add a pair of braces
  staging: wlan-ng: Replace long int with long
  staging: MAINTAINERS: remove obsolete IPX staging directory
  staging: MAINTAINERS: remove NCP filesystem entry
  staging: rtl8188eu: cleanup comparsions to false
  staging: gasket: Update device virtual address comment
  staging: gasket: sysfs: fix attribute release comment
  staging: gasket: apex: fix sysfs_show
  staging: gasket: page_table: simplify gasket_components_to_dev_address
  staging: gasket: page_table: fix comment in components_to_dev_address
  staging: gasket: page table: fixup error path allocating coherent mem
  staging: gasket: page_table: rearrange gasket_page_table_entry
  staging: gasket: page_table: remove unnecessary PTE status set to free
  ...

Pull mailbox updates from Jassi Brar:

 - convert print users to use the %pOFn format specifier

 - enable ti-msgmr driver for the K3 platform as well

 - add QCS404 to compatible list of QCOM's APCS IPC driver

 - minor spelling fixes toogle -> toggle

 - kzalloc failure catch in Mediatek driver

* tag 'mailbox-v4.20' of git://git.linaro.org/landing-teams/working/fujitsu/integration:
  mailbox: mediatek: Add check for possible failure of kzalloc
  mailbox: bcm-flexrm-mailbox: fix spelling mistake "toogle" -> "toggle"
  mailbox: qcom: Add QCS404 APPS Global compatible
  drivers: mailbox: Make ti-msgmr driver depend on ARCH_K3
  mailbox: Convert to using %pOFn instead of device_node.name

Merge tag 'filesystems_for_v4.20-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs

Pull ext2 and udf updates from Jan Kara:
 "Small ext2 cleanups and a couple of udf fixes"

* tag 'filesystems_for_v4.20-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
  ext2: remove redundant building macro check
  udf: Drop pack pragma from udf_sb.h
  udf: Drop freed bitmap / table support
  udf: Fix crash during mount
  udf: Prevent write-unsupported filesystem to be remounted read-write
  ext2: cache NULL when both default_acl and acl are NULL
  udf: remove unused variables group_start and nr_groups

Pull fsnotify updates from Jan Kara:
 "Amir's patches to implement superblock fanotify watches, Xiaoming's
  patch to enable reporting of thread IDs in fanotify events instead of
  TGIDs (sadly the patch got mis-attributed to Amir and I've noticed
  only now), and a fix of possible oops on umount caused by fsnotify
  infrastructure"

* tag 'for_v4.20-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
  fsnotify: Fix busy inodes during unmount
  fs: group frequently accessed fields of struct super_block together
  fanotify: support reporting thread id instead of process id
  fanotify: add BUILD_BUG_ON() to count the bits of fanotify constants
  fsnotify: convert runtime BUG_ON() to BUILD_BUG_ON()
  fanotify: deprecate uapi FAN_ALL_* constants
  fanotify: simplify handling of FAN_ONDIR
  fsnotify: generalize handling of extra event flags
  fanotify: fix collision of internal and uapi mark flags
  fanotify: store fanotify_init() flags in group's fanotify_data
  fanotify: add API to attach/detach super block mark
  fsnotify: send path type events to group with super block marks
  fsnotify: add super block object type

Pull 9p updates from Dominique Martinet:
 "Highlights this time around are the end of Matthew's work to remove
  the custom 9p request cache and use a slab directly for requests, with
  some extra patches on my end to not degrade performance, but it's a
  very good cleanup.

  Tomas and I fixed a few more syzkaller bugs (refcount is the big one),
  and I had a go at the coverity bugs and at some of the bugzilla
  reports we had open for a while.

  I'm a bit disappointed that I couldn't get much reviews for a few of
  my own patches, but the big ones got some and it's all been soaking in
  linux-next for quite a while so I think it should be OK.

  Summary:

   - Finish removing the custom 9p request cache mechanism

   - Embed part of the fcall in the request to have better slab
     performance (msize usually is power of two aligned)

   - syzkaller fixes:
      * add a refcount to 9p requests to avoid use after free
      * a few double free issues

   - A few coverity fixes

   - Some old patches that were in the bugzilla:
      * do not trust pdu content for size header
      * mount option for lock retry interval"

* tag '9p-for-4.20' of git://github.com/martinetd/linux: (21 commits)
  9p/trans_fd: put worker reqs on destroy
  9p/trans_fd: abort p9_read_work if req status changed
  9p: potential NULL dereference
  9p locks: fix glock.client_id leak in do_lock
  9p: p9dirent_read: check network-provided name length
  9p/rdma: remove useless check in cm_event_handler
  9p: acl: fix uninitialized iattr access
  9p locks: add mount option for lock retry interval
  9p: do not trust pdu content for stat item size
  9p: Rename req to rreq in trans_fd
  9p: fix spelling mistake in fall-through annotation
  9p/rdma: do not disconnect on down_interruptible EAGAIN
  9p: Add refcount to p9_req_t
  9p: rename p9_free_req() function
  9p: add a per-client fcall kmem_cache
  9p: embed fcall in req to round down buffer allocs
  9p: Remove p9_idpool
  9p: Use a slab for allocating requests
  9p: clear dangling pointers in p9stat_free
  v9fs_dir_readdir: fix double-free on p9stat_read error
  ...

Pull m68k nommu fix from Greg Ungerer:
 "Only a single change to fix an out of bounds array access when parsing
  boot command line"

* 'for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/gerg/m68knommu:
  m68k: fix command-line parsing when passed from u-boot

Pull m68k updates from Geert Uytterhoeven:
 "Just two small cleanups"

* tag 'm68k-for-v4.20-tag1' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k:
  m68k/sun3: Remove is_medusa and m68k_pgtable_cachemode
  m68k/atari: ARAnyM - Remove reference to long-deprecated MODULE_PARM

Pull C-SKY architecture port from Guo Ren:
 "This contains the Linux port for C-SKY(csky) based on linux-4.19
  Release, which has been through 10 rounds of review on mailing list.

  More information:

    http://en.c-sky.com

  The development repo:

    https://github.com/c-sky/csky-linux

  ABI Documentation:

    https://github.com/c-sky/csky-doc

  Here is the pre-built cross compiler for fast test from our CI:

    https://gitlab.com/c-sky/buildroot/-/jobs/101608095/artifacts/file/output/images/csky_toolchain_qemu_csky_ck807f_4.18_glibc_defconfig_482b221e52908be1c9b2ccb444255e1562bb7025.tar.xz

  We use buildroot as our CI-test enviornment. "LTP, Lmbench ..." will
  be tested for every commit. See here for more details:

    https://gitlab.com/c-sky/buildroot/pipelines

  We'll continouslly improve csky subsystem in future"

Arnd acks, and adds the following notes:
 "I did a thorough review of the ABI, which as usual mainly consists of
  spotting any files that don't use the asm-generic ABI itself, and
  having it changed to it matches exactly what we do on other new
  architectures.

  I also looked at every other patch and commented on maybe half of them
  where I saw something that did not quite seem right. Others have
  reviewed specific patches in greater depth. I'm sure that one could
  fine more of the minor details, but as long as they are not ABI
  relevant, they can be fixed later.

  The only patch that is part of the ABI and that nobody reviewed is the
  signal handling. This is one of the areas I never worked on in much
  detail. I did not see anything wrong with it, but I also don't know
  what the problems with the other architectures are here, and we seem
  to be hitting issues occasionally, and we never managed to generalize
  this enough for new architectures to have a trivial implementation.

  I was originally hoping that we could have the 64-bit time_t
  interfaces ready in time to completely drop the 32-bit ones, but that
  did not happen. We might still remove them in the next merge window
  depending on whether the libc upstream people prefer to keep them or
  not.

  One more general comment: I think this may well be the last new CPU
  architecture we ever add to the kernel. Both nds32 and c-sky are made
  by companies that also work on risc-v, and generally speaking risc-v
  seems to be killing off any of the minor licensable instruction set
  projects, just like ARM has mostly killed off the custom
  vendor-specific instruction sets already.

  If we add another architecture in the future, it may instead be
  something like the LLVM bitcode or WebAssembly, who knows?"

To which Geert Uytterhoeven pipes in about another architecture still in
the pipeline: Kalray MPPA.

* tag 'csky-for-linus-4.20' of https://github.com/c-sky/csky-linux: (24 commits)
  dt-bindings: interrupt-controller: C-SKY APB intc
  irqchip: add C-SKY APB bus interrupt controller
  dt-bindings: interrupt-controller: C-SKY SMP intc
  irqchip: add C-SKY SMP interrupt controller
  MAINTAINERS: Add csky
  dt-bindings: Add vendor prefix for csky
  dt-bindings: csky CPU Bindings
  csky: Misc headers
  csky: SMP support
  csky: Debug and Ptrace GDB
  csky: User access
  csky: Library functions
  csky: ELF and module probe
  csky: Atomic operations
  csky: IRQ handling
  csky: VDSO and rt_sigreturn
  csky: Process management and Signal
  csky: MMU and page table management
  csky: Cache and TLB routines
  csky: System Call
  ...

Pull networking fixes from David Miller:

 1) GRO overflow entries are not unlinked properly, resulting in list
    poison pointers being dereferenced.

 2) Fix bridge build with ipv6 disabled, from Nikolay Aleksandrov.

 3) Direct packet access and other fixes in BPF from Daniel Borkmann.

 4) gred_change_table_def() gets passed the wrong pointer, a pointer to
    a set of unparsed attributes instead of the attribute itself. From
    Jakub Kicinski.

 5) Allow macsec device to be brought up even if it's lowerdev is down,
    from Sabrina Dubroca.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net:
  net: diag: document swapped src/dst in udp_dump_one.
  macsec: let the administrator set UP state even if lowerdev is down
  macsec: update operstate when lower device changes
  net: sched: gred: pass the right attribute to gred_change_table_def()
  ptp: drop redundant kasprintf() to create worker name
  net: bridge: remove ipv6 zero address check in mcast queries
  net: Properly unlink GRO packets on overflow.
  bpf: fix wrong helper enablement in cgroup local storage
  bpf: add bpf_jit_limit knob to restrict unpriv allocations
  bpf: make direct packet write unclone more robust
  bpf: fix leaking uninitialized memory on pop/peek helpers
  bpf: fix direct packet write into pop/peek helpers
  bpf: fix cg_skb types to hint access type in may_access_direct_pkt_data
  bpf: fix direct packet access for flow dissector progs
  bpf: disallow direct packet access for unpriv in cg_skb
  bpf: fix test suite to enable all unpriv program types
  bpf, btf: fix a missing check bug in btf_parse
  selftests/bpf: add config fragments BPF_STREAM_PARSER and XDP_SOCKETS
  bpf: devmap: fix wrong interface selection in notifier_call

Since its inception, udp_dump_one has had a bug where userspace
needs to swap src and dst addresses and ports in order to find
the socket it wants. This is because it passes the socket source
address to __udp[46]_lib_lookup's saddr argument, but those
functions are intended to find local sockets matching received
packets, so saddr is the remote address, not the local address.

This can no longer be fixed for backwards compatibility reasons,
so add a brief comment explaining that this is the case. This
will avoid confusion and help ensure SOCK_DIAG implementations
of new protocols don't have the same problem.

Fixes: a925aa00 ("udp_diag: Implement the get_exact dumping functionality")
Signed-off-by: Lorenzo Colitti <lorenzo@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Sabrina Dubroca says:

====================
macsec: linkstate fixes

This series fixes issues with handling administrative and operstate of
macsec devices.

Radu Rendec proposed another version of the first patch [0] but I'd
rather not follow the behavior of vlan devices, going with macvlan
does instead. Patrick Talbert also reported the same issue to me.

The second patch is a follow-up. The restriction on setting the device
up is a bit unreasonable, and operstate provides the information we
need in this case.

[0] https://patchwork.ozlabs.org/patch/971374/
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently, the kernel doesn't let the administrator set a macsec device
up unless its lower device is currently up. This is inconsistent, as a
macsec device that is up won't automatically go down when its lower
device goes down.

Now that linkstate propagation works, there's really no reason for this
limitation, so let's remove it.

Fixes: c09440f7 ("macsec: introduce IEEE 802.1AE driver")
Reported-by: Radu Rendec <radu.rendec@gmail.com>
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

Like all other virtual devices (macvlan, vlan), the operstate of a
macsec device should match the state of its lower device. This is done
by calling netif_stacked_transfer_operstate from its netdevice notifier.

We also need to call netif_stacked_transfer_operstate when a new macsec
device is created, so that its operstate is set properly. This is only
relevant when we try to bring the device up directly when we create it.

Radu Rendec proposed a similar patch, inspired from the 802.1q driver,
that included changing the administrative state of the macsec device,
instead of just the operstate. This version is similar to what the
macvlan driver does, and updates only the operstate.

Fixes: c09440f7 ("macsec: introduce IEEE 802.1AE driver")
Reported-by: Radu Rendec <radu.rendec@gmail.com>
Reported-by: Patrick Talbert <ptalbert@redhat.com>
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: David S. Miller <davem@davemloft.net>