Merge remote-tracking branch 'msm4.4/tmp-da9a92f' into msm-4.4

* origin/tmp-da9a92f:
  arm64: kaslr: increase randomization granularity
  arm64: relocatable: deal with physically misaligned kernel images
  arm64: don't map TEXT_OFFSET bytes below the kernel if we can avoid it
  arm64: kernel: replace early 64-bit literal loads with move-immediates
  arm64: introduce mov_q macro to move a constant into a 64-bit register
  arm64: kernel: perform relocation processing from ID map
  arm64: kernel: use literal for relocated address of __secondary_switched
  arm64: kernel: don't export local symbols from head.S
  arm64: simplify kernel segment mapping granularity
  arm64: cover the .head.text section in the .text segment mapping
  arm64: move early boot code to the .init segment
  arm64: use 'segment' rather than 'chunk' to describe mapped kernel regions
  arm64: mm: Mark .rodata as RO
  Linux 4.4.16
  ovl: verify upper dentry before unlink and rename
  drm/i915: Revert DisplayPort fast link training feature
  tmpfs: fix regression hang in fallocate undo
  tmpfs: don't undo fallocate past its last page
  crypto: qat - make qat_asym_algs.o depend on asn1 headers
  xen/acpi: allow xen-acpi-processor driver to load on Xen 4.7
  File names with trailing period or space need special case conversion
  cifs: dynamic allocation of ntlmssp blob
  Fix reconnect to not defer smb3 session reconnect long after socket reconnect
  53c700: fix BUG on untagged commands
  s390: fix test_fp_ctl inline assembly contraints
  scsi: fix race between simultaneous decrements of ->host_failed
  ovl: verify upper dentry in ovl_remove_and_whiteout()
  ovl: Copy up underlying inode's ->i_mode to overlay inode
  ARM: mvebu: fix HW I/O coherency related deadlocks
  ARM: dts: armada-38x: fix MBUS_ID for crypto SRAM on Armada 385 Linksys
  ARM: sunxi/dt: make the CHIP inherit from allwinner,sun5i-a13
  ALSA: hda: add AMD Stoney PCI ID with proper driver caps
  ALSA: hda - fix use-after-free after module unload
  ALSA: ctl: Stop notification after disconnection
  ALSA: pcm: Free chmap at PCM free callback, too
  ALSA: hda/realtek - add new pin definition in alc225 pin quirk table
  ALSA: hda - fix read before array start
  ALSA: hda - Add PCI ID for Kabylake-H
  ALSA: hda/realtek: Add Lenovo L460 to docking unit fixup
  ALSA: timer: Fix negative queue usage by racy accesses
  ALSA: echoaudio: Fix memory allocation
  ALSA: au88x0: Fix calculation in vortex_wtdma_bufshift()
  ALSA: hda / realtek - add two more Thinkpad IDs (5050,5053) for tpt460 fixup
  ALSA: hda - Fix the headset mic jack detection on Dell machine
  ALSA: dummy: Fix a use-after-free at closing
  hwmon: (dell-smm) Cache fan_type() calls and change fan detection
  hwmon: (dell-smm) Disallow fan_type() calls on broken machines
  hwmon: (dell-smm) Restrict fan control and serial number to CAP_SYS_ADMIN by default
  tty/vt/keyboard: fix OOB access in do_compute_shiftstate()
  tty: vt: Fix soft lockup in fbcon cursor blink timer.
  iio:ad7266: Fix probe deferral for vref
  iio:ad7266: Fix support for optional regulators
  iio:ad7266: Fix broken regulator error handling
  iio: accel: kxsd9: fix the usage of spi_w8r8()
  staging: iio: accel: fix error check
  iio: hudmidity: hdc100x: fix incorrect shifting and scaling
  iio: humidity: hdc100x: fix IIO_TEMP channel reporting
  iio: humidity: hdc100x: correct humidity integration time mask
  iio: proximity: as3935: fix buffer stack trashing
  iio: proximity: as3935: remove triggered buffer processing
  iio: proximity: as3935: correct IIO_CHAN_INFO_RAW output
  iio: light apds9960: Add the missing dev.parent
  iio:st_pressure: fix sampling gains (bring inline with ABI)
  iio: Fix error handling in iio_trigger_attach_poll_func
  xen/balloon: Fix declared-but-not-defined warning
  perf/x86: Fix undefined shift on 32-bit kernels
  memory: omap-gpmc: Fix omap gpmc EXTRADELAY timing
  drm/vmwgfx: Fix error paths when mapping framebuffer
  drm/vmwgfx: Delay pinning fbdev framebuffer until after mode set
  drm/vmwgfx: Check pin count before attempting to move a buffer
  drm/vmwgfx: Work around mode set failure in 2D VMs
  drm/vmwgfx: Add an option to change assumed FB bpp
  drm/ttm: Make ttm_bo_mem_compat available
  drm: atmel-hlcdc: actually disable scaling when no scaling is required
  drm: make drm_atomic_set_mode_prop_for_crtc() more reliable
  drm: add missing drm_mode_set_crtcinfo call
  drm/i915: Update CDCLK_FREQ register on BDW after changing cdclk frequency
  drm/i915: Update ifdeffery for mutex->owner
  drm/i915: Refresh cached DP port register value on resume
  drm/i915/ilk: Don't disable SSC source if it's in use
  drm/nouveau/disp/sor/gf119: select correct sor when poking training pattern
  drm/nouveau: fix for disabled fbdev emulation
  drm/nouveau/fbcon: fix out-of-bounds memory accesses
  drm/nouveau/gr/gf100-: update sm error decoding from gk20a nvgpu headers
  drm/nouveau/disp/sor/gf119: both links use the same training register
  virtio_balloon: fix PFN format for virtio-1
  drm/dp/mst: Always clear proposed vcpi table for port.
  drm/amdkfd: destroy dbgmgr in notifier release
  drm/amdkfd: unbind only existing processes
  ubi: Make recover_peb power cut aware
  drm/amdgpu/gfx7: fix broken condition check
  drm/radeon: fix asic initialization for virtualized environments
  btrfs: account for non-CoW'd blocks in btrfs_abort_transaction
  percpu: fix synchronization between synchronous map extension and chunk destruction
  percpu: fix synchronization between chunk->map_extend_work and chunk destruction
  af_unix: fix hard linked sockets on overlay
  vfs: add d_real_inode() helper
  arm64: Rework valid_user_regs
  ipmi: Remove smi_msg from waiting_rcv_msgs list before handle_one_recv_msg()
  drm/mgag200: Black screen fix for G200e rev 4
  iommu/amd: Fix unity mapping initialization race
  iommu/vt-d: Enable QI on all IOMMUs before setting root entry
  iommu/arm-smmu: Wire up map_sg for arm-smmu-v3
  base: make module_create_drivers_dir race-free
  tracing: Handle NULL formats in hold_module_trace_bprintk_format()
  HID: multitouch: enable palm rejection for Windows Precision Touchpad
  HID: hiddev: validate num_values for HIDIOCGUSAGES, HIDIOCSUSAGES commands
  HID: elo: kill not flush the work
  KVM: nVMX: VMX instructions: fix segment checks when L1 is in long mode.
  kvm: Fix irq route entries exceeding KVM_MAX_IRQ_ROUTES
  KEYS: potential uninitialized variable
  ARCv2: LLSC: software backoff is NOT needed starting HS2.1c
  ARCv2: Check for LL-SC livelock only if LLSC is enabled
  ipv6: Fix mem leak in rt6i_pcpu
  cdc_ncm: workaround for EM7455 "silent" data interface
  net_sched: fix mirrored packets checksum
  packet: Use symmetric hash for PACKET_FANOUT_HASH.
  sched/fair: Fix cfs_rq avg tracking underflow
  UBIFS: Implement ->migratepage()
  mm: Export migrate_page_move_mapping and migrate_page_copy
  MIPS: KVM: Fix modular KVM under QEMU
  ARM: 8579/1: mm: Fix definition of pmd_mknotpresent
  ARM: 8578/1: mm: ensure pmd_present only checks the valid bit
  ARM: imx6ul: Fix Micrel PHY mask
  NFS: Fix another OPEN_DOWNGRADE bug
  make nfs_atomic_open() call d_drop() on all ->open_context() errors.
  nfsd: check permissions when setting ACLs
  posix_acl: Add set_posix_acl
  nfsd: Extend the mutex holding region around in nfsd4_process_open2()
  nfsd: Always lock state exclusively.
  nfsd4/rpc: move backchannel create logic into rpc code
  writeback: use higher precision calculation in domain_dirty_limits()
  thermal: cpu_cooling: fix improper order during initialization
  uvc: Forward compat ioctls to their handlers directly
  Revert "gpiolib: Split GPIO flags parsing and GPIO configuration"
  x86/amd_nb: Fix boot crash on non-AMD systems
  kprobes/x86: Clear TF bit in fault on single-stepping
  x86, build: copy ldlinux.c32 to image.iso
  locking/static_key: Fix concurrent static_key_slow_inc()
  locking/qspinlock: Fix spin_unlock_wait() some more
  locking/ww_mutex: Report recursive ww_mutex locking early
  of: irq: fix of_irq_get[_byname]() kernel-doc
  of: fix autoloading due to broken modalias with no 'compatible'
  mnt: If fs_fully_visible fails call put_filesystem.
  mnt: Account for MS_RDONLY in fs_fully_visible
  mnt: fs_fully_visible test the proper mount for MNT_LOCKED
  usb: common: otg-fsm: add license to usb-otg-fsm
  USB: EHCI: declare hostpc register as zero-length array
  usb: dwc2: fix regression on big-endian PowerPC/ARM systems
  powerpc/tm: Always reclaim in start_thread() for exec() class syscalls
  powerpc/pseries: Fix IBM_ARCH_VEC_NRCORES_OFFSET since POWER8NVL was added
  powerpc/pseries: Fix PCI config address for DDW
  powerpc/iommu: Remove the dependency on EEH struct in DDW mechanism
  IB/mlx4: Properly initialize GRH TClass and FlowLabel in AHs
  IB/cm: Fix a recently introduced locking bug
  EDAC, sb_edac: Fix rank lookup on Broadwell
  mac80211: Fix mesh estab_plinks counting in STA removal case
  mac80211_hwsim: Add missing check for HWSIM_ATTR_SIGNAL
  mac80211: mesh: flush mesh paths unconditionally
  mac80211: fix fast_tx header alignment
  Linux 4.4.15
  usb: dwc3: exynos: Fix deferred probing storm.
  usb: host: ehci-tegra: Grab the correct UTMI pads reset
  usb: gadget: fix spinlock dead lock in gadgetfs
  USB: mos7720: delete parport
  xhci: Fix handling timeouted commands on hosts in weird states.
  USB: xhci: Add broken streams quirk for Frescologic device id 1009
  usb: xhci-plat: properly handle probe deferral for devm_clk_get()
  xhci: Cleanup only when releasing primary hcd
  usb: musb: host: correct cppi dma channel for isoch transfer
  usb: musb: Ensure rx reinit occurs for shared_fifo endpoints
  usb: musb: Stop bulk endpoint while queue is rotated
  usb: musb: only restore devctl when session was set in backup
  usb: quirks: Add no-lpm quirk for Acer C120 LED Projector
  usb: quirks: Fix sorting
  USB: uas: Fix slave queue_depth not being set
  crypto: user - re-add size check for CRYPTO_MSG_GETALG
  crypto: ux500 - memmove the right size
  crypto: vmx - Increase priority of aes-cbc cipher
  AX.25: Close socket connection on session completion
  bpf: try harder on clones when writing into skb
  net: alx: Work around the DMA RX overflow issue
  net: macb: fix default configuration for GMAC on AT91
  neigh: Explicitly declare RCU-bh read side critical section in neigh_xmit()
  bpf, perf: delay release of BPF prog after grace period
  sock_diag: do not broadcast raw socket destruction
  Bridge: Fix ipv6 mc snooping if bridge has no ipv6 address
  ipmr/ip6mr: Initialize the last assert time of mfc entries.
  netem: fix a use after free
  esp: Fix ESN generation under UDP encapsulation
  sit: correct IP protocol used in ipip6_err
  net: Don't forget pr_fmt on net_dbg_ratelimited for CONFIG_DYNAMIC_DEBUG
  net_sched: fix pfifo_head_drop behavior vs backlog
  sdcardfs: Truncate packages_gid.list on overflow
  UPSTREAM: cdc_ncm: do not call usbnet_link_change from cdc_ncm_bind
  BACKPORT: proc: add /proc/<pid>/timerslack_ns interface
  BACKPORT: timer: convert timer_slack_ns from unsigned long to u64
  netfilter: xt_quota2: make quota2_log work well
  Revert "usb: gadget: prevent change of Host MAC address of 'usb0' interface"
  BACKPORT: PM / sleep: Go direct_complete if driver has no callbacks
  ANDROID: base-cfg: enable UID_CPUTIME
  UPSTREAM: USB: usbfs: fix potential infoleak in devio
  UPSTREAM: ALSA: timer: Fix leak in events via snd_timer_user_ccallback
  UPSTREAM: ALSA: timer: Fix leak in events via snd_timer_user_tinterrupt
  UPSTREAM: ALSA: timer: Fix leak in SNDRV_TIMER_IOCTL_PARAMS
  ANDROID: configs: remove unused configs
  ANDROID: cpu: send KOBJ_ONLINE event when enabling cpus
  ANDROID: dm verity fec: initialize recursion level
  ANDROID: dm verity fec: fix RS block calculation
  Linux 4.4.14
  netfilter: x_tables: introduce and use xt_copy_counters_from_user
  netfilter: x_tables: do compat validation via translate_table
  netfilter: x_tables: xt_compat_match_from_user doesn't need a retval
  netfilter: ip6_tables: simplify translate_compat_table args
  netfilter: ip_tables: simplify translate_compat_table args
  netfilter: arp_tables: simplify translate_compat_table args
  netfilter: x_tables: don't reject valid target size on some architectures
  netfilter: x_tables: validate all offsets and sizes in a rule
  netfilter: x_tables: check for bogus target offset
  netfilter: x_tables: check standard target size too
  netfilter: x_tables: add compat version of xt_check_entry_offsets
  netfilter: x_tables: assert minimum target size
  netfilter: x_tables: kill check_entry helper
  netfilter: x_tables: add and use xt_check_entry_offsets
  netfilter: x_tables: validate targets of jumps
  netfilter: x_tables: don't move to non-existent next rule
  drm/core: Do not preserve framebuffer on rmfb, v4.
  crypto: qat - fix adf_ctl_drv.c:undefined reference to adf_init_pf_wq
  netfilter: x_tables: fix unconditional helper
  netfilter: x_tables: make sure e->next_offset covers remaining blob size
  netfilter: x_tables: validate e->target_offset early
  MIPS: Fix 64k page support for 32 bit kernels.
  sparc64: Fix return from trap window fill crashes.
  sparc: Harden signal return frame checks.
  sparc64: Take ctx_alloc_lock properly in hugetlb_setup().
  sparc64: Reduce TLB flushes during hugepte changes
  sparc/PCI: Fix for panic while enabling SR-IOV
  sparc64: Fix sparc64_set_context stack handling.
  sparc64: Fix numa node distance initialization
  sparc64: Fix bootup regressions on some Kconfig combinations.
  sparc: Fix system call tracing register handling.
  fix d_walk()/non-delayed __d_free() race
  sched: panic on corrupted stack end
  proc: prevent stacking filesystems on top
  x86/entry/traps: Don't force in_interrupt() to return true in IST handlers
  wext: Fix 32 bit iwpriv compatibility issue with 64 bit Kernel
  ecryptfs: forbid opening files without mmap handler
  memcg: add RCU locking around css_for_each_descendant_pre() in memcg_offline_kmem()
  parisc: Fix pagefault crash in unaligned __get_user() call
  pinctrl: mediatek: fix dual-edge code defect
  powerpc/pseries: Add POWER8NVL support to ibm,client-architecture-support call
  powerpc: Use privileged SPR number for MMCR2
  powerpc: Fix definition of SIAR and SDAR registers
  powerpc/pseries/eeh: Handle RTAS delay requests in configure_bridge
  arm64: mm: always take dirty state from new pte in ptep_set_access_flags
  arm64: Provide "model name" in /proc/cpuinfo for PER_LINUX32 tasks
  crypto: ccp - Fix AES XTS error for request sizes above 4096
  crypto: public_key: select CRYPTO_AKCIPHER
  irqchip/gic-v3: Fix ICC_SGI1R_EL1.INTID decoding mask
  s390/bpf: reduce maximum program size to 64 KB
  s390/bpf: fix recache skb->data/hlen for skb_vlan_push/pop
  gpio: bcm-kona: fix bcm_kona_gpio_reset() warnings
  ARM: fix PTRACE_SETVFPREGS on SMP systems
  ALSA: hda/realtek: Add T560 docking unit fixup
  ALSA: hda/realtek - Add support for new codecs ALC700/ALC701/ALC703
  ALSA: hda/realtek - ALC256 speaker noise issue
  ALSA: hda - Fix headset mic detection problem for Dell machine
  ALSA: hda - Add PCI ID for Kabylake
  KVM: irqfd: fix NULL pointer dereference in kvm_irq_map_gsi
  KVM: x86: fix OOPS after invalid KVM_SET_DEBUGREGS
  vxlan, gre, geneve: Set a large MTU on ovs-created tunnel devices
  geneve: Relax MTU constraints
  vxlan: Relax MTU constraints
  ipv6: Skip XFRM lookup if dst_entry in socket cache is valid
  l2tp: fix configuration passed to setup_udp_tunnel_sock()
  bridge: Don't insert unnecessary local fdb entry on changing mac address
  tcp: record TLP and ER timer stats in v6 stats
  vxlan: Accept user specified MTU value when create new vxlan link
  team: don't call netdev_change_features under team->lock
  sfc: on MC reset, clear PIO buffer linkage in TXQs
  bpf, inode: disallow userns mounts
  uapi glibc compat: fix compilation when !__USE_MISC in glibc
  udp: prevent skbs lingering in tunnel socket queues
  bpf: Use mount_nodev not mount_ns to mount the bpf filesystem
  tuntap: correctly wake up process during uninit
  switchdev: pass pointer to fib_info instead of copy
  tipc: fix nametable publication field in nl compat
  netlink: Fix dump skb leak/double free
  tipc: check nl sock before parsing nested attributes
  scsi: Add QEMU CD-ROM to VPD Inquiry Blacklist
  scsi_lib: correctly retry failed zero length REQ_TYPE_FS commands
  cs-etm: associating output packet with CPU they executed on
  cs-etm: removing unecessary structure field
  cs-etm: account for each trace buffer in the queue
  cs-etm: avoid casting variable
  perf tools: fixing Makefile problems
  perf tools: new naming convention for openCSD
  perf scripts: Add python scripts for CoreSight traces
  perf tools: decoding capailitity for CoreSight traces
  perf symbols: Check before overwriting build_id
  perf tools: pushing driver configuration down to the kernel
  perf tools: add infrastructure for PMU specific configuration
  coresight: etm-perf: incorporating sink definition from the cmd line
  coresight: adding sink parameter to function coresight_build_path()
  perf: passing struct perf_event to function setup_aux()
  perf/core: adding PMU driver specific configuration
  perf tools: adding coresight etm PMU record capabilities
  perf tools: making coresight PMU listable
  coresight: tmc: implementing TMC-ETR AUX space API
  coresight: Add support for Juno platform
  coresight: Handle build path error
  coresight: Fix erroneous memset in tmc_read_unprepare_etr
  coresight: Fix tmc_read_unprepare_etr
  coresight: Fix NULL pointer dereference in _coresight_build_path
  ANDROID: dm verity fec: add missing release from fec_ktype
  ANDROID: dm verity fec: limit error correction recursion
  ANDROID: restrict access to perf events
  FROMLIST: security,perf: Allow further restriction of perf_event_open
  BACKPORT: perf tools: Document the perf sysctls
  Revert "armv6 dcc tty driver"
  Revert "arm: dcc_tty: fix armv6 dcc tty build failure"
  ARM64: Ignore Image-dtb from git point of view
  arm64: add option to build Image-dtb
  ANDROID: usb: gadget: f_midi: set fi->f to NULL when free f_midi function
  Linux 4.4.13
  xfs: handle dquot buffer readahead in log recovery correctly
  xfs: print name of verifier if it fails
  xfs: skip stale inodes in xfs_iflush_cluster
  xfs: fix inode validity check in xfs_iflush_cluster
  xfs: xfs_iflush_cluster fails to abort on error
  xfs: Don't wrap growfs AGFL indexes
  xfs: disallow rw remount on fs with unknown ro-compat features
  gcov: disable tree-loop-im to reduce stack usage
  scripts/package/Makefile: rpmbuild add support of RPMOPTS
  dma-debug: avoid spinlock recursion when disabling dma-debug
  PM / sleep: Handle failures in device_suspend_late() consistently
  ext4: silence UBSAN in ext4_mb_init()
  ext4: address UBSAN warning in mb_find_order_for_block()
  ext4: fix oops on corrupted filesystem
  ext4: clean up error handling when orphan list is corrupted
  ext4: fix hang when processing corrupted orphaned inode list
  drm/imx: Match imx-ipuv3-crtc components using device node in platform data
  drm/i915: Don't leave old junk in ilk active watermarks on readout
  drm/atomic: Verify connector->funcs != NULL when clearing states
  drm/fb_helper: Fix references to dev->mode_config.num_connector
  drm/i915/fbdev: Fix num_connector references in intel_fb_initial_config()
  drm/amdgpu: Fix hdmi deep color support.
  drm/amdgpu: use drm_mode_vrefresh() rather than mode->vrefresh
  drm/vmwgfx: Fix order of operation
  drm/vmwgfx: use vmw_cmd_dx_cid_check for query commands.
  drm/vmwgfx: Enable SVGA_3D_CMD_DX_SET_PREDICATION
  drm/gma500: Fix possible out of bounds read
  sunrpc: fix stripping of padded MIC tokens
  xen: use same main loop for counting and remapping pages
  xen/events: Don't move disabled irqs
  powerpc/eeh: Restore initial state in eeh_pe_reset_and_recover()
  Revert "powerpc/eeh: Fix crash in eeh_add_device_early() on Cell"
  powerpc/eeh: Don't report error in eeh_pe_reset_and_recover()
  powerpc/book3s64: Fix branching to OOL handlers in relocatable kernel
  pipe: limit the per-user amount of pages allocated in pipes
  QE-UART: add "fsl,t1040-ucc-uart" to of_device_id
  wait/ptrace: assume __WALL if the child is traced
  mm: use phys_addr_t for reserve_bootmem_region() arguments
  media: v4l2-compat-ioctl32: fix missing reserved field copy in put_v4l2_create32
  PCI: Disable all BAR sizing for devices with non-compliant BARs
  pinctrl: exynos5440: Use off-stack memory for pinctrl_gpio_range
  clk: bcm2835: divider value has to be 1 or more
  clk: bcm2835: pll_off should only update CM_PLL_ANARST
  clk: at91: fix check of clk_register() returned value
  clk: bcm2835: Fix PLL poweron
  cpuidle: Fix cpuidle_state_is_coupled() argument in cpuidle_enter()
  cpuidle: Indicate when a device has been unregistered
  PM / Runtime: Fix error path in pm_runtime_force_resume()
  mfd: intel_soc_pmic_core: Terminate panel control GPIO lookup table correctly
  mfd: intel-lpss: Save register context on suspend
  hwmon: (ads7828) Enable internal reference
  aacraid: Fix for KDUMP driver hang
  aacraid: Fix for aac_command_thread hang
  aacraid: Relinquish CPU during timeout wait
  rtlwifi: pci: use dev_kfree_skb_irq instead of kfree_skb in rtl_pci_reset_trx_ring
  rtlwifi: Fix logic error in enter/exit power-save mode
  rtlwifi: btcoexist: Implement antenna selection
  rtlwifi: rtl8723be: Add antenna select module parameter
  hwrng: exynos - Fix unbalanced PM runtime put on timeout error path
  ath5k: Change led pin configuration for compaq c700 laptop
  ath10k: fix kernel panic, move arvifs list head init before htt init
  ath10k: fix rx_channel during hw reconfigure
  ath10k: fix firmware assert in monitor mode
  ath10k: fix debugfs pktlog_filter write
  ath9k: Fix LED polarity for some Mini PCI AR9220 MB92 cards.
  ath9k: Add a module parameter to invert LED polarity.
  ARM: dts: imx35: restore existing used clock enumeration
  ARM: dts: exynos: Add interrupt line to MAX8997 PMIC on exynos4210-trats
  ARM: dts: at91: fix typo in sama5d2 PIN_PD24 description
  ARM: mvebu: fix GPIO config on the Linksys boards
  Input: uinput - handle compat ioctl for UI_SET_PHYS
  ASoC: ak4642: Enable cache usage to fix crashes on resume
  affs: fix remount failure when there are no options changed
  MIPS: VDSO: Build with `-fno-strict-aliasing'
  MIPS: lib: Mark intrinsics notrace
  MIPS: Build microMIPS VDSO for microMIPS kernels
  MIPS: Fix sigreturn via VDSO on microMIPS kernel
  MIPS: ptrace: Prevent writes to read-only FCSR bits
  MIPS: ptrace: Fix FP context restoration FCSR regression
  MIPS: Disable preemption during prctl(PR_SET_FP_MODE, ...)
  MIPS: Prevent "restoration" of MSA context in non-MSA kernels
  MIPS: Fix MSA ld_*/st_* asm macros to use PTR_ADDU
  MIPS: Use copy_s.fmt rather than copy_u.fmt
  MIPS: Loongson-3: Reserve 32MB for RS780E integrated GPU
  MIPS: Reserve nosave data for hibernation
  MIPS: ath79: make bootconsole wait for both THRE and TEMT
  MIPS: Sync icache & dcache in set_pte_at
  MIPS: Handle highmem pages in __update_cache
  MIPS: Flush highmem pages in __flush_dcache_page
  MIPS: Fix watchpoint restoration
  MIPS: Fix uapi include in exported asm/siginfo.h
  MIPS: Fix siginfo.h to use strict posix types
  MIPS: Avoid using unwind_stack() with usermode
  MIPS: Don't unwind to user mode with EVA
  MIPS: MSA: Fix a link error on `_init_msa_upper' with older GCC
  MIPS: math-emu: Fix jalr emulation when rd == $0
  MIPS64: R6: R2 emulation bugfix
  coresight: etb10: adjust read pointer only when needed
  coresight: configuring ETF in FIFO mode when acting as link
  coresight: tmc: implementing TMC-ETF AUX space API
  coresight: moving struct cs_buffers to header file
  coresight: tmc: keep track of memory width
  coresight: tmc: make sysFS and Perf mode mutually exclusive
  coresight: tmc: dump system memory content only when needed
  coresight: tmc: adding mode of operation for link/sinks
  coresight: tmc: getting rid of multiple read access
  coresight: tmc: allocating memory when needed
  coresight: tmc: making prepare/unprepare functions generic
  coresight: tmc: splitting driver in ETB/ETF and ETR components
  coresight: tmc: cleaning up header file
  coresight: tmc: introducing new header file
  coresight: tmc: clearly define number of transfers per burst
  coresight: tmc: re-implementing tmc_read_prepare/unprepare() functions
  coresight: tmc: waiting for TMCReady bit before programming
  coresight: tmc: modifying naming convention
  coresight: tmc: adding sysFS management entries
  coresight: etm4x: add tracer ID for A72 Maia processor.
  coresight: etb10: fixing the right amount of words to read
  coresight: stm: adding driver for CoreSight STM component
  coresight: adding path for STM device
  coresight: etm4x: modify q_support type
  coresight: no need to do the forced type conversion
  coresight: removing gratuitous boot time log messages
  coresight: etb10: splitting sysFS "status" entry
  coresight: moving coresight_simple_func() to header file
  coresight: etm4x: implementing the perf PMU API
  coresight: etm4x: implementing user/kernel mode tracing
  coresight: etm4x: moving etm_drvdata::enable to atomic field
  coresight: etm4x: unlocking tracers in default arch init
  coresight: etm4x: splitting etmv4 default configuration
  coresight: etm4x: splitting struct etmv4_drvdata
  coresight: etm4x: adding config and traceid registers
  coresight: etm4x: moving sysFS entries to a dedicated file
  stm class: Support devices that override software assigned masters
  stm class: Remove unnecessary pointer increment
  stm class: Fix stm device initialization order
  stm class: Do not leak the chrdev in error path
  stm class: Remove a pointless line
  stm class: stm_heartbeat: Make nr_devs parameter read-only
  stm class: dummy_stm: Make nr_dummies parameter read-only
  MAINTAINERS: Add a git tree for the stm class
  perf/ring_buffer: Document AUX API usage
  perf/core: Free AUX pages in unmap path
  perf/ring_buffer: Refuse to begin AUX transaction after rb->aux_mmap_count drops
  perf auxtrace: Add perf_evlist pointer to *info_priv_size()
  perf session: Simplify tool stubs
  perf inject: Hit all DSOs for AUX data in JIT and other cases
  perf tools: tracepoint_error() can receive e=NULL, robustify it
  perf evlist: Make perf_evlist__open() open evsels with their cpus and threads (like perf record does)
  perf evsel: Introduce disable() method
  perf cpumap: Auto initialize cpu__max_{node,cpu}
  drivers/hwtracing: make coresight-etm-perf.c explicitly non-modular
  drivers/hwtracing: make coresight-* explicitly non-modular
  coresight: introducing a global trace ID function
  coresight: etm-perf: new PMU driver for ETM tracers
  coresight: etb10: implementing AUX API
  coresight: etb10: adding operation mode for sink->enable()
  coresight: etb10: moving to local atomic operations
  coresight: etm3x: implementing perf_enable/disable() API
  coresight: etm3x: implementing user/kernel mode tracing
  coresight: etm3x: consolidating initial config
  coresight: etm3x: changing default trace configuration
  coresight: etm3x: set progbit to stop trace collection
  coresight: etm3x: adding operation mode for etm_enable()
  coresight: etm3x: splitting struct etm_drvdata
  coresight: etm3x: unlocking tracers in default arch init
  coresight: etm3x: moving sysFS entries to dedicated file
  coresight: etm3x: moving etm_readl/writel to header file
  coresight: moving PM runtime operations to core framework
  coresight: add API to get sink from path
  coresight: associating path with session rather than tracer
  coresight: etm4x: Check every parameter used by dma_xx_coherent.
  coresight: "DEVICE_ATTR_RO" should defined as static.
  coresight: implementing 'cpu_id()' API
  coresight: removing bind/unbind options from sysfs
  coresight: remove csdev's link from topology
  coresight: release reference taken by 'bus_find_device()'
  coresight: coresight_unregister() function cleanup
  coresight: fixing lockdep error
  coresight: fixing indentation problem
  coresight: Fix a typo in Kconfig
  coresight: checking for NULL string in coresight_name_match()
  perf/core: Disable the event on a truncated AUX record
  perf/core: Don't leak event in the syscall error path
  perf/core: Fix perf_sched_count derailment
  stm class: dummy_stm: Add link callback for fault injection
  stm class: Plug stm device's unlink callback
  stm class: Fix a race in unlinking
  stm class: Fix unbalanced module/device refcounting
  stm class: Guard output assignment against concurrency
  stm class: Fix unlocking braino in the error path
  stm class: Add heartbeat stm source device
  stm class: dummy_stm: Create multiple devices
  stm class: Support devices with multiple instances
  stm class: Use driver's packet callback return value
  stm class: Prevent user-controllable allocations
  stm class: Fix link list locking
  stm class: Fix locking in unbinding policy path
  stm class: Select CONFIG_SRCU
  stm class: Hide STM-specific options if STM is disabled
  perf: Synchronously free aux pages in case of allocation failure
  Linux 4.4.12
  kbuild: move -Wunused-const-variable to W=1 warning level
  Revert "scsi: fix soft lockup in scsi_remove_target() on module removal"
  scsi: Add intermediate STARGET_REMOVE state to scsi_target_state
  hpfs: implement the show_options method
  hpfs: fix remount failure when there are no options changed
  UBI: Fix static volume checks when Fastmap is used
  SIGNAL: Move generic copy_siginfo() to signal.h
  thunderbolt: Fix double free of drom buffer
  IB/srp: Fix a debug kernel crash
  ALSA: hda - Fix headset mic detection problem for one Dell machine
  ALSA: hda/realtek - Add support for ALC295/ALC3254
  ALSA: hda - Fix headphone noise on Dell XPS 13 9360
  ALSA: hda/realtek - New codecs support for ALC234/ALC274/ALC294
  mcb: Fixed bar number assignment for the gdd
  clk: bcm2835: add locking to pll*_on/off methods
  locking,qspinlock: Fix spin_is_locked() and spin_unlock_wait()
  serial: samsung: Reorder the sequence of clock control when call s3c24xx_serial_set_termios()
  serial: 8250_mid: recognize interrupt source in handler
  serial: 8250_mid: use proper bar for DNV platform
  serial: 8250_pci: fix divide error bug if baud rate is 0
  Fix OpenSSH pty regression on close
  tty/serial: atmel: fix hardware handshake selection
  TTY: n_gsm, fix false positive WARN_ON
  tty: vt, return error when con_startup fails
  xen/x86: actually allocate legacy interrupts on PV guests
  KVM: x86: mask CPUID(0xD,0x1).EAX against host value
  MIPS: KVM: Fix timer IRQ race when writing CP0_Compare
  MIPS: KVM: Fix timer IRQ race when freezing timer
  KVM: x86: fix ordering of cr0 initialization code in vmx_cpu_reset
  KVM: MTRR: remove MSR 0x2f8
  staging: comedi: das1800: fix possible NULL dereference
  usb: gadget: udc: core: Fix argument of dev_err() in usb_gadget_map_request()
  USB: leave LPM alone if possible when binding/unbinding interface drivers
  usb: misc: usbtest: fix pattern tests for scatterlists.
  usb: f_mass_storage: test whether thread is running before starting another
  usb: gadget: f_fs: Fix EFAULT generation for async read operations
  USB: serial: option: add even more ZTE device ids
  USB: serial: option: add more ZTE device ids
  USB: serial: option: add support for Cinterion PH8 and AHxx
  USB: serial: io_edgeport: fix memory leaks in probe error path
  USB: serial: io_edgeport: fix memory leaks in attach error path
  USB: serial: quatech2: fix use-after-free in probe error path
  USB: serial: keyspan: fix use-after-free in probe error path
  USB: serial: mxuport: fix use-after-free in probe error path
  mei: bus: call mei_cl_read_start under device lock
  mei: amthif: discard not read messages
  mei: fix NULL dereferencing during FW initiated disconnection
  Bluetooth: vhci: Fix race at creating hci device
  Bluetooth: vhci: purge unhandled skbs
  Bluetooth: vhci: fix open_timeout vs. hdev race
  mmc: sdhci-pci: Remove MMC_CAP_BUS_WIDTH_TEST for Intel controllers
  mmc: longer timeout for long read time quirk
  dell-rbtn: Ignore ACPI notifications if device is suspended
  ACPI / osi: Fix an issue that acpi_osi=!* cannot disable ACPICA internal strings
  mmc: sdhci-acpi: Remove MMC_CAP_BUS_WIDTH_TEST for Intel controllers
  mmc: mmc: Fix partition switch timeout for some eMMCs
  can: fix handling of unmodifiable configuration options
  irqchip/gic-v3: Configure all interrupts as non-secure Group-1
  irqchip/gic: Ensure ordering between read of INTACK and shared data
  Input: pwm-beeper - fix - scheduling while atomic
  mfd: omap-usb-tll: Fix scheduling while atomic BUG
  sched/loadavg: Fix loadavg artifacts on fully idle and on fully loaded systems
  clk: qcom: msm8916: Fix crypto clock flags
  crypto: sun4i-ss - Replace spinlock_bh by spin_lock_irq{save|restore}
  crypto: talitos - fix ahash algorithms registration
  crypto: caam - fix caam_jr_alloc() ret code
  ring-buffer: Prevent overflow of size in ring_buffer_resize()
  ring-buffer: Use long for nr_pages to avoid overflow failures
  asix: Fix offset calculation in asix_rx_fixup() causing slow transmissions
  fs/cifs: correctly to anonymous authentication for the NTLM(v2) authentication
  fs/cifs: correctly to anonymous authentication for the NTLM(v1) authentication
  fs/cifs: correctly to anonymous authentication for the LANMAN authentication
  fs/cifs: correctly to anonymous authentication via NTLMSSP
  remove directory incorrectly tries to set delete on close on non-empty directories
  kvm: arm64: Fix EC field in inject_abt64
  arm/arm64: KVM: Enforce Break-Before-Make on Stage-2 page tables
  arm64: cpuinfo: Missing NULL terminator in compat_hwcap_str
  arm64: Implement pmdp_set_access_flags() for hardware AF/DBM
  arm64: Implement ptep_set_access_flags() for hardware AF/DBM
  arm64: Ensure pmd_present() returns false after pmd_mknotpresent()
  arm64: Fix typo in the pmdp_huge_get_and_clear() definition
  ext4: iterate over buffer heads correctly in move_extent_per_page()
  perf test: Fix build of BPF and LLVM on older glibc libraries
  perf/core: Fix perf_event_open() vs. execve() race
  perf/x86/intel/pt: Generate PMI in the STOP region as well
  Btrfs: don't use src fd for printk
  UPSTREAM: mac80211: fix "warning: ‘target_metric’ may be used uninitialized"
  Revert "drivers: power: use 'current' instead of 'get_current()'"
  cpufreq: interactive: drop cpufreq_{get,put}_global_kobject func calls
  Revert "cpufreq: interactive: build fixes for 4.4"
  xt_qtaguid: Fix panic caused by processing non-full socket.
  fiq_debugger: Add fiq_debugger.disable option
  UPSTREAM: procfs: fixes pthread cross-thread naming if !PR_DUMPABLE
  FROMLIST: wlcore: Disable filtering in AP role
  Revert "drivers: power: Add watchdog timer to catch drivers which lockup during suspend."
  fiq_debugger: Add option to apply uart overlay by FIQ_DEBUGGER_UART_OVERLAY
  Revert "Recreate asm/mach/mmc.h include file"
  Revert "ARM: Add 'card_present' state to mmc_platfrom_data"
  usb: dual-role: make stub functions inline
  Revert "mmc: Add status IRQ and status callback function to mmc platform data"
  quick selinux support for tracefs
  Revert "hid-multitouch: Filter collections by application usage."
  Revert "HID: steelseries: validate output report details"
  xt_qtaguid: Fix panic caused by synack processing
  Revert "mm: vmscan: Add a debug file for shrinkers"
  Revert "SELinux: Enable setting security contexts on rootfs inodes."
  Revert "SELinux: build fix for 4.1"
  fuse: Add support for d_canonical_path
  vfs: change d_canonical_path to take two paths
  android: recommended.cfg: remove CONFIG_UID_STAT
  netfilter: xt_qtaguid: seq_printf fixes
  Revert "misc: uidstat: Adding uid stat driver to collect network statistics."
  Revert "net: activity_stats: Add statistics for network transmission activity"
  Revert "net: activity_stats: Stop using obsolete create_proc_read_entry api"
  Revert "misc: uidstat: avoid create_stat() race and blockage."
  Revert "misc: uidstat: Remove use of obsolete create_proc_read_entry api"
  Revert "misc seq_printf fixes for 4.4"
  Revert "misc: uid_stat: Include linux/atomic.h instead of asm/atomic.h"
  Revert "net: socket ioctl to reset connections matching local address"
  Revert "net: fix iterating over hashtable in tcp_nuke_addr()"
  Revert "net: fix crash in tcp_nuke_addr()"
  Revert "Don't kill IPv4 sockets when killing IPv6 sockets was requested."
  Revert "tcp: Fix IPV6 module build errors"
  android: base-cfg: remove CONFIG_SWITCH
  Revert "switch: switch class and GPIO drivers."
  Revert "drivers: switch: remove S_IWUSR from dev_attr"
  ANDROID: base-cfg: enable CONFIG_IP_NF_NAT
  BACKPORT: selinux: restrict kernel module loading
  android: base-cfg: enable CONFIG_QUOTA

Conflicts:
	Documentation/sysctl/kernel.txt
	drivers/cpufreq/cpufreq_interactive.c
	drivers/hwtracing/coresight/Kconfig
	drivers/hwtracing/coresight/Makefile
	drivers/hwtracing/coresight/coresight-etm4x.c
	drivers/hwtracing/coresight/coresight-etm4x.h
	drivers/hwtracing/coresight/coresight-priv.h
	drivers/hwtracing/coresight/coresight-stm.c
	drivers/hwtracing/coresight/coresight-tmc.c
	drivers/mmc/core/core.c
	include/linux/coresight-stm.h
	include/linux/coresight.h
	include/linux/msm_mdp.h
	include/uapi/linux/coresight-stm.h
	kernel/events/core.c
	kernel/sched/fair.c
	net/Makefile
	net/ipv4/netfilter/arp_tables.c
	net/ipv4/netfilter/ip_tables.c
	net/ipv4/tcp.c
	net/ipv6/netfilter/ip6_tables.c
	net/netfilter/xt_quota2.c
	sound/core/pcm.c

Change-Id: I17aa0002815014e9bddc47e67769a53c15768a99
Signed-off-by: Runmin Wang <runminw@codeaurora.org>

7 files changed, 67 insertions, 85 deletions

diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index d1f6dc5a715d..08806bb1f070 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -3661,6 +3661,7 @@ static void memcg_deactivate_kmem(struct mem_cgroup *memcg)
 	 * ordering is imposed by list_lru_node->lock taken by
 	 * memcg_drain_all_list_lrus().
 	 */
+	rcu_read_lock(); /* can be called from css_free w/o cgroup_mutex */
 	css_for_each_descendant_pre(css, &memcg->css) {
 		child = mem_cgroup_from_css(css);
 		BUG_ON(child->kmemcg_id != kmemcg_id);
@@ -3668,6 +3669,8 @@ static void memcg_deactivate_kmem(struct mem_cgroup *memcg)
 		if (!memcg->use_hierarchy)
 			break;
 	}
+	rcu_read_unlock();
+
 	memcg_drain_all_list_lrus(kmemcg_id, parent->kmemcg_id);
 
 	memcg_free_cache_id(kmemcg_id);
diff --git a/mm/migrate.c b/mm/migrate.c
index 3db1b0277eb4..a9ce3783361a 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -429,6 +429,7 @@ int migrate_page_move_mapping(struct address_space *mapping,
 
 	return MIGRATEPAGE_SUCCESS;
 }
+EXPORT_SYMBOL(migrate_page_move_mapping);
 
 /*
  * The expected number of remaining references is the same as that
@@ -579,6 +580,7 @@ void migrate_page_copy(struct page *newpage, struct page *page)
 	if (PageWriteback(newpage))
 		end_page_writeback(newpage);
 }
+EXPORT_SYMBOL(migrate_page_copy);
 
 /************************************************************
  *                    Migration functions
diff --git a/mm/page-writeback.c b/mm/page-writeback.c
index 8bf8e06a56a6..6b4fa64a5c91 100644
--- a/mm/page-writeback.c
+++ b/mm/page-writeback.c
@@ -359,8 +359,9 @@ static void domain_dirty_limits(struct dirty_throttle_control *dtc)
 	struct dirty_throttle_control *gdtc = mdtc_gdtc(dtc);
 	unsigned long bytes = vm_dirty_bytes;
 	unsigned long bg_bytes = dirty_background_bytes;
-	unsigned long ratio = vm_dirty_ratio;
-	unsigned long bg_ratio = dirty_background_ratio;
+	/* convert ratios to per-PAGE_SIZE for higher precision */
+	unsigned long ratio = (vm_dirty_ratio * PAGE_SIZE) / 100;
+	unsigned long bg_ratio = (dirty_background_ratio * PAGE_SIZE) / 100;
 	unsigned long thresh;
 	unsigned long bg_thresh;
 	struct task_struct *tsk;
@@ -372,26 +373,28 @@ static void domain_dirty_limits(struct dirty_throttle_control *dtc)
 		/*
 		 * The byte settings can't be applied directly to memcg
 		 * domains.  Convert them to ratios by scaling against
-		 * globally available memory.
+		 * globally available memory.  As the ratios are in
+		 * per-PAGE_SIZE, they can be obtained by dividing bytes by
+		 * number of pages.
 		 */
 		if (bytes)
-			ratio = min(DIV_ROUND_UP(bytes, PAGE_SIZE) * 100 /
-				    global_avail, 100UL);
+			ratio = min(DIV_ROUND_UP(bytes, global_avail),
+				    PAGE_SIZE);
 		if (bg_bytes)
-			bg_ratio = min(DIV_ROUND_UP(bg_bytes, PAGE_SIZE) * 100 /
-				       global_avail, 100UL);
+			bg_ratio = min(DIV_ROUND_UP(bg_bytes, global_avail),
+				       PAGE_SIZE);
 		bytes = bg_bytes = 0;
 	}
 
 	if (bytes)
 		thresh = DIV_ROUND_UP(bytes, PAGE_SIZE);
 	else
-		thresh = (ratio * available_memory) / 100;
+		thresh = (ratio * available_memory) / PAGE_SIZE;
 
 	if (bg_bytes)
 		bg_thresh = DIV_ROUND_UP(bg_bytes, PAGE_SIZE);
 	else
-		bg_thresh = (bg_ratio * available_memory) / 100;
+		bg_thresh = (bg_ratio * available_memory) / PAGE_SIZE;
 
 	if (bg_thresh >= thresh)
 		bg_thresh = thresh / 2;
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index c8a31783c2d6..f12a0bde548b 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -967,7 +967,7 @@ static inline void init_reserved_page(unsigned long pfn)
  * marks the pages PageReserved. The remaining valid pages are later
  * sent to the buddy page allocator.
  */
-void __meminit reserve_bootmem_region(unsigned long start, unsigned long end)
+void __meminit reserve_bootmem_region(phys_addr_t start, phys_addr_t end)
 {
 	unsigned long start_pfn = PFN_DOWN(start);
 	unsigned long end_pfn = PFN_UP(end);
diff --git a/mm/percpu.c b/mm/percpu.c
index 8a943b97a053..1f376bce413c 100644
--- a/mm/percpu.c
+++ b/mm/percpu.c
@@ -110,7 +110,7 @@ struct pcpu_chunk {
 	int			map_used;	/* # of map entries used before the sentry */
 	int			map_alloc;	/* # of map entries allocated */
 	int			*map;		/* allocation map */
-	struct work_struct	map_extend_work;/* async ->map[] extension */
+	struct list_head	map_extend_list;/* on pcpu_map_extend_chunks */
 
 	void			*data;		/* chunk data */
 	int			first_free;	/* no free below this */
@@ -160,10 +160,13 @@ static struct pcpu_chunk *pcpu_reserved_chunk;
 static int pcpu_reserved_chunk_limit;
 
 static DEFINE_SPINLOCK(pcpu_lock);	/* all internal data structures */
-static DEFINE_MUTEX(pcpu_alloc_mutex);	/* chunk create/destroy, [de]pop */
+static DEFINE_MUTEX(pcpu_alloc_mutex);	/* chunk create/destroy, [de]pop, map ext */
 
 static struct list_head *pcpu_slot __read_mostly; /* chunk list slots */
 
+/* chunks which need their map areas extended, protected by pcpu_lock */
+static LIST_HEAD(pcpu_map_extend_chunks);
+
 /*
  * The number of empty populated pages, protected by pcpu_lock.  The
  * reserved chunk doesn't contribute to the count.
@@ -397,13 +400,19 @@ static int pcpu_need_to_extend(struct pcpu_chunk *chunk, bool is_atomic)
 {
 	int margin, new_alloc;
 
+	lockdep_assert_held(&pcpu_lock);
+
 	if (is_atomic) {
 		margin = 3;
 
 		if (chunk->map_alloc <
-		    chunk->map_used + PCPU_ATOMIC_MAP_MARGIN_LOW &&
-		    pcpu_async_enabled)
-			schedule_work(&chunk->map_extend_work);
+		    chunk->map_used + PCPU_ATOMIC_MAP_MARGIN_LOW) {
+			if (list_empty(&chunk->map_extend_list)) {
+				list_add_tail(&chunk->map_extend_list,
+					      &pcpu_map_extend_chunks);
+				pcpu_schedule_balance_work();
+			}
+		}
 	} else {
 		margin = PCPU_ATOMIC_MAP_MARGIN_HIGH;
 	}
@@ -437,6 +446,8 @@ static int pcpu_extend_area_map(struct pcpu_chunk *chunk, int new_alloc)
 	size_t old_size = 0, new_size = new_alloc * sizeof(new[0]);
 	unsigned long flags;
 
+	lockdep_assert_held(&pcpu_alloc_mutex);
+
 	new = pcpu_mem_zalloc(new_size);
 	if (!new)
 		return -ENOMEM;
@@ -469,20 +480,6 @@ out_unlock:
 	return 0;
 }
 
-static void pcpu_map_extend_workfn(struct work_struct *work)
-{
-	struct pcpu_chunk *chunk = container_of(work, struct pcpu_chunk,
-						map_extend_work);
-	int new_alloc;
-
-	spin_lock_irq(&pcpu_lock);
-	new_alloc = pcpu_need_to_extend(chunk, false);
-	spin_unlock_irq(&pcpu_lock);
-
-	if (new_alloc)
-		pcpu_extend_area_map(chunk, new_alloc);
-}
-
 /**
  * pcpu_fit_in_area - try to fit the requested allocation in a candidate area
  * @chunk: chunk the candidate area belongs to
@@ -742,7 +739,7 @@ static struct pcpu_chunk *pcpu_alloc_chunk(void)
 	chunk->map_used = 1;
 
 	INIT_LIST_HEAD(&chunk->list);
-	INIT_WORK(&chunk->map_extend_work, pcpu_map_extend_workfn);
+	INIT_LIST_HEAD(&chunk->map_extend_list);
 	chunk->free_size = pcpu_unit_size;
 	chunk->contig_hint = pcpu_unit_size;
 
@@ -897,6 +894,9 @@ static void __percpu *pcpu_alloc(size_t size, size_t align, bool reserved,
 		return NULL;
 	}
 
+	if (!is_atomic)
+		mutex_lock(&pcpu_alloc_mutex);
+
 	spin_lock_irqsave(&pcpu_lock, flags);
 
 	/* serve reserved allocations from the reserved chunk if available */
@@ -969,12 +969,9 @@ restart:
 	if (is_atomic)
 		goto fail;
 
-	mutex_lock(&pcpu_alloc_mutex);
-
 	if (list_empty(&pcpu_slot[pcpu_nr_slots - 1])) {
 		chunk = pcpu_create_chunk();
 		if (!chunk) {
-			mutex_unlock(&pcpu_alloc_mutex);
 			err = "failed to allocate new chunk";
 			goto fail;
 		}
@@ -985,7 +982,6 @@ restart:
 		spin_lock_irqsave(&pcpu_lock, flags);
 	}
 
-	mutex_unlock(&pcpu_alloc_mutex);
 	goto restart;
 
 area_found:
@@ -995,8 +991,6 @@ area_found:
 	if (!is_atomic) {
 		int page_start, page_end, rs, re;
 
-		mutex_lock(&pcpu_alloc_mutex);
-
 		page_start = PFN_DOWN(off);
 		page_end = PFN_UP(off + size);
 
@@ -1007,7 +1001,6 @@ area_found:
 
 			spin_lock_irqsave(&pcpu_lock, flags);
 			if (ret) {
-				mutex_unlock(&pcpu_alloc_mutex);
 				pcpu_free_area(chunk, off, &occ_pages);
 				err = "failed to populate";
 				goto fail_unlock;
@@ -1047,6 +1040,8 @@ fail:
 		/* see the flag handling in pcpu_blance_workfn() */
 		pcpu_atomic_alloc_failed = true;
 		pcpu_schedule_balance_work();
+	} else {
+		mutex_unlock(&pcpu_alloc_mutex);
 	}
 	return NULL;
 }
@@ -1131,6 +1126,7 @@ static void pcpu_balance_workfn(struct work_struct *work)
 		if (chunk == list_first_entry(free_head, struct pcpu_chunk, list))
 			continue;
 
+		list_del_init(&chunk->map_extend_list);
 		list_move(&chunk->list, &to_free);
 	}
 
@@ -1148,6 +1144,25 @@ static void pcpu_balance_workfn(struct work_struct *work)
 		pcpu_destroy_chunk(chunk);
 	}
 
+	/* service chunks which requested async area map extension */
+	do {
+		int new_alloc = 0;
+
+		spin_lock_irq(&pcpu_lock);
+
+		chunk = list_first_entry_or_null(&pcpu_map_extend_chunks,
+					struct pcpu_chunk, map_extend_list);
+		if (chunk) {
+			list_del_init(&chunk->map_extend_list);
+			new_alloc = pcpu_need_to_extend(chunk, false);
+		}
+
+		spin_unlock_irq(&pcpu_lock);
+
+		if (new_alloc)
+			pcpu_extend_area_map(chunk, new_alloc);
+	} while (chunk);
+
 	/*
 	 * Ensure there are certain number of free populated pages for
 	 * atomic allocs.  Fill up from the most packed so that atomic
@@ -1646,7 +1661,7 @@ int __init pcpu_setup_first_chunk(const struct pcpu_alloc_info *ai,
 	 */
 	schunk = memblock_virt_alloc(pcpu_chunk_struct_size, 0);
 	INIT_LIST_HEAD(&schunk->list);
-	INIT_WORK(&schunk->map_extend_work, pcpu_map_extend_workfn);
+	INIT_LIST_HEAD(&schunk->map_extend_list);
 	schunk->base_addr = base_addr;
 	schunk->map = smap;
 	schunk->map_alloc = ARRAY_SIZE(smap);
@@ -1675,7 +1690,7 @@ int __init pcpu_setup_first_chunk(const struct pcpu_alloc_info *ai,
 	if (dyn_size) {
 		dchunk = memblock_virt_alloc(pcpu_chunk_struct_size, 0);
 		INIT_LIST_HEAD(&dchunk->list);
-		INIT_WORK(&dchunk->map_extend_work, pcpu_map_extend_workfn);
+		INIT_LIST_HEAD(&dchunk->map_extend_list);
 		dchunk->base_addr = base_addr;
 		dchunk->map = dmap;
 		dchunk->map_alloc = ARRAY_SIZE(dmap);
diff --git a/mm/shmem.c b/mm/shmem.c
index b82e56ebabdd..79997e8cf807 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -2153,9 +2153,11 @@ static long shmem_fallocate(struct file *file, int mode, loff_t offset,
 									NULL);
 		if (error) {
 			/* Remove the !PageUptodate pages we added */
-			shmem_undo_range(inode,
-				(loff_t)start << PAGE_CACHE_SHIFT,
-				(loff_t)index << PAGE_CACHE_SHIFT, true);
+			if (index > start) {
+				shmem_undo_range(inode,
+				 (loff_t)start << PAGE_CACHE_SHIFT,
+				 ((loff_t)index << PAGE_CACHE_SHIFT) - 1, true);
+			}
 			goto undone;
 		}
 
diff --git a/mm/vmscan.c b/mm/vmscan.c
index d5c3ef60a71e..ff408638fd95 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -46,7 +46,6 @@
 #include <linux/oom.h>
 #include <linux/prefetch.h>
 #include <linux/printk.h>
-#include <linux/debugfs.h>
 
 #include <asm/tlbflush.h>
 #include <asm/div64.h>
@@ -234,39 +233,6 @@ static unsigned long get_lru_size(struct lruvec *lruvec, enum lru_list lru)
 	return zone_page_state(lruvec_zone(lruvec), NR_LRU_BASE + lru);
 }
 
-struct dentry *debug_file;
-
-static int debug_shrinker_show(struct seq_file *s, void *unused)
-{
-	struct shrinker *shrinker;
-	struct shrink_control sc;
-
-	sc.gfp_mask = -1;
-	sc.nr_to_scan = 0;
-
-	down_read(&shrinker_rwsem);
-	list_for_each_entry(shrinker, &shrinker_list, list) {
-		int num_objs;
-
-		num_objs = shrinker->count_objects(shrinker, &sc);
-		seq_printf(s, "%pf %d\n", shrinker->scan_objects, num_objs);
-	}
-	up_read(&shrinker_rwsem);
-	return 0;
-}
-
-static int debug_shrinker_open(struct inode *inode, struct file *file)
-{
-        return single_open(file, debug_shrinker_show, inode->i_private);
-}
-
-static const struct file_operations debug_shrinker_fops = {
-        .open = debug_shrinker_open,
-        .read = seq_read,
-        .llseek = seq_lseek,
-        .release = single_release,
-};
-
 /*
  * Add a shrinker callback to be called from the vm.
  */
@@ -296,15 +262,6 @@ int register_shrinker(struct shrinker *shrinker)
 }
 EXPORT_SYMBOL(register_shrinker);
 
-static int __init add_shrinker_debug(void)
-{
-	debugfs_create_file("shrinker", 0644, NULL, NULL,
-			    &debug_shrinker_fops);
-	return 0;
-}
-
-late_initcall(add_shrinker_debug);
-
 /*
  * Remove one
  */