bpf: offload: ignore namespace moves

We are currently destroying the device offload state when device
moves to another net namespace.  This doesn't break with current
NFP code, because offload state is not used on program removal,
but it's not correct behaviour.

Ignore the device unregister notifications on namespace move.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

bpf: turn bpf_prog_get_type() into a wrapper

bpf_prog_get_type() is identical to bpf_prog_get_type_dev(),
with false passed as attach_drv.  Instead of keeping it as
an exported symbol turn it into static inline wrapper.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Quentin Monnet <quentin.monnet@netronome.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

net: xdp: don't allow device-bound programs in driver mode

Currently device-bound programs are not able to run on the host
to save resources (host JIT is not invoked).  Don't allow XDP
programs to be attached without the HW_MODE flag.  In theory
if program is already translated for device offload the driver
should choose to offload it instead of loading it in the driver.
However, offloading translated program may still fail resulting
in device-bound program being run on the host.

Prevent this by refusing to attach device bound programs if
XDP_FLAGS_HW_MODE is not set.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Quentin Monnet <quentin.monnet@netronome.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

bpf: offload: move offload device validation out to the drivers

With TC shared block changes we can't depend on correct netdev
pointer being available in cls_bpf.  Move the device validation
to the driver.  Core will only make sure that offloaded programs
are always attached in the driver (or in HW by the driver).  We
trust that drivers which implement offload callbacks will perform
necessary checks.

Moving the checks to the driver is generally a useful thing,
in practice the check should be against a switchdev instance,
not a netdev, given that most ASICs will probably allow using
the same program on many ports.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Quentin Monnet <quentin.monnet@netronome.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

bpf: offload: rename the ifindex field

bpf_target_prog seems long and clunky, rename it to prog_ifindex.
We don't want to call this field just ifindex, because maps
may need a similar field in the future and bpf_attr members for
programs and maps are unnamed.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Quentin Monnet <quentin.monnet@netronome.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

bpf: offload: limit offload to cls_bpf and xdp programs only

We are currently only allowing attachment of device-bound
cls_bpf and XDP programs.  Make this restriction explicit in
the BPF offload code.  This way we can potentially reuse the
ifindex field in the future.

Since XDP and cls_bpf programs can only be loaded by admin,
we can drop the explicit capability check from offload code.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Quentin Monnet <quentin.monnet@netronome.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

bpf: offload: add comment warning developers about double destroy

Offload state may get destroyed either because the device for which
it was constructed is going away, or because the refcount of bpf
program itself has reached 0.  In both of those cases we will call
__bpf_prog_offload_destroy() to unlink the offload from the device.
We may in fact call it twice, which works just fine, but we should
make clear this is intended and caution others trying to extend the
function.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Quentin Monnet <quentin.monnet@netronome.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

net: vxge: Fix some indentation issues

Some statements are not enough or too much indented.
Fix it to improve readalbility.

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ena: fix race condition between device reset and link up setup

In rare cases, ena driver would reset and re-start the device,
for example, in case of misbehaving application that causes
transmit timeout

The first step in the reset procedure is to stop the Tx traffic by
calling ena_carrier_off().

After the driver have just started the device reset procedure, device
happens to send an asynchronous notification (via AENQ) to the driver
than there was a link change (to link-up state).
This link change is mapped to a call to netif_carrier_on() which
re-activates the Tx queues, violating the assumption of no tx traffic
until device reset is completed, as the reset task might still be in
the process of queues initialization, leading to an access to
uninitialized memory.

Signed-off-by: Netanel Belgazal <netanel@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

r8169: use same RTL8111EVL green settings as in vendor driver

Adjust the code to use the same green settings as in the latest
vendor driver.

Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

r8169: fix RTL8111EVL EEE and green settings

Name of functions rtl_w0w1_eri and rtl_w0w1_phy is somewhat misleading
regarding order of arguments. One could assume that w0w1 means
argument with bits to be reset comes before argument with bits to set.
However this is not the case.
So fix the order of arguments in several statements.

In addition fix EEE advertisement. The current code resets the bits
for 100BaseT and 1000BaseT EEE advertisement what is not what we want.

I have a little of a hard time to find a proper "Fixes" line as the
issue seems to have been there forever (at least it existed already
when the driver was moved to the current place in 2011).

The patch was tested on a Zotac Mini-PC with a RTL8111E-VL chip.
Before the patch EEE was disabled, now it's properly advertised and
works fine.

Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

tun: fix rcu_read_lock imbalance in tun_build_skb

rcu_read_lock in tun_build_skb is used to rcu_dereference tun->xdp_prog
safely, rcu_read_unlock should be done in every return path.

Now I could see one place missing it, where it returns NULL in switch-case
XDP_REDIRECT,  another palce using rcu_read_lock wrongly, where it returns
NULL in if (xdp_xmit) chunk.

So fix both in this patch.

Fixes: 761876c857cb ("tap: XDP support")
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

tcp: when scheduling TLP, time of RTO should account for current ACK

Fix the TLP scheduling logic so that when scheduling a TLP probe, we
ensure that the estimated time at which an RTO would fire accounts for
the fact that ACKs indicating forward progress should push back RTO
times.

After the following fix:

df92c8394e6e ("tcp: fix xmit timer to only be reset if data ACKed/SACKed")

we had an unintentional behavior change in the following kind of
scenario: suppose the RTT variance has been very low recently. Then
suppose we send out a flight of N packets and our RTT is 100ms:

t=0: send a flight of N packets
t=100ms: receive an ACK for N-1 packets

The response before df92c8394e6e that was:
  -> schedule a TLP for now + RTO_interval

The response after df92c8394e6e is:
  -> schedule a TLP for t=0 + RTO_interval

Since RTO_interval = srtt + RTT_variance, this means that we have
scheduled a TLP timer at a point in the future that only accounts for
RTT_variance. If the RTT_variance term is small, this means that the
timer fires soon.

Before df92c8394e6e this would not happen, because in that code, when
we receive an ACK for a prefix of flight, we did:

    1) Near the top of tcp_ack(), switch from TLP timer to RTO
       at write_queue_head->paket_tx_time + RTO_interval:
            if (icsk->icsk_pending == ICSK_TIME_LOSS_PROBE)
                   tcp_rearm_rto(sk);

    2) In tcp_clean_rtx_queue(), update the RTO to now + RTO_interval:
            if (flag & FLAG_ACKED) {
                   tcp_rearm_rto(sk);

    3) In tcp_ack() after tcp_fastretrans_alert() switch from RTO
       to TLP at now + RTO_interval:
            if (icsk->icsk_pending == ICSK_TIME_RETRANS)
                   tcp_schedule_loss_probe(sk);

In df92c8394e6e we removed that 3-phase dance, and instead directly
set the TLP timer once: we set the TLP timer in cases like this to
write_queue_head->packet_tx_time + RTO_interval. So if the RTT
variance is small, then this means that this is setting the TLP timer
to fire quite soon. This means if the ACK for the tail of the flight
takes longer than an RTT to arrive (often due to delayed ACKs), then
the TLP timer fires too quickly.

Fixes: df92c8394e6e ("tcp: fix xmit timer to only be reset if data ACKed/SACKed")
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

usbnet: ipheth: fix potential null pointer dereference in ipheth_carrier_set

_dev_ is being dereferenced before it is null checked, hence there
is a potential null pointer dereference.

Fix this by moving the pointer dereference after _dev_ has been null
checked.

Addresses-Coverity-ID: 1462020
Fixes: bb1b40c7cb86 ("usbnet: ipheth: prevent TX queue timeouts when device not ready")
Signed-off-by: Gustavo A. R. Silva <garsilva@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

gre6: use log_ecn_error module parameter in ip6_tnl_rcv()

After commit 308edfdf1563 ("gre6: Cleanup GREv6 receive path, call
common GRE functions") it's not used anywhere in the module, but
previously was used in ip6gre_rcv().

Fixes: 308edfdf1563 ("gre6: Cleanup GREv6 receive path, call common GRE functions")
Signed-off-by: Alexey Kodanev <alexey.kodanev@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge git://git./linux/kernel/git/davem/sparc

Pull sparc updates from David Miller:

 1) Add missing cmpxchg64() for 32-bit sparc.

 2) Timer conversions from Allen Pais and Kees Cook.

 3) vDSO support, from Nagarathnam Muthusamy.

 4) Fix sparc64 huge page table walks based upon bug report by Al Viro,
    from Nitin Gupta.

 5) Optimized fls() for T4 and above, from Vijay Kumar.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc:
  sparc64: Fix page table walk for PUD hugepages
  sparc64: Convert timers to user timer_setup()
  sparc64: convert mdesc_handle.refcnt from atomic_t to refcount_t
  sparc/led: Convert timers to use timer_setup()
  sparc64: Use sparc optimized fls and __fls for T4 and above
  sparc64: SPARC optimized __fls function
  sparc64: SPARC optimized fls function
  sparc64: Define SPARC default __fls function
  sparc64: Define SPARC default fls function
  vDSO for sparc
  sparc32: Add cmpxchg64().
  sbus: char: Move D7S_MINOR to include/linux/miscdevice.h
  sparc: time: Remove unneeded linux/miscdevice.h include
  sparc64: mmu_context: Add missing include files

Merge git://git./linux/kernel/git/davem/net

Pull networking fixes from David Miller:

 1) Revert regression inducing change to the IPSEC template resolver,
    from Steffen Klassert.

 2) Peeloffs can cause the wrong sk to be waken up in SCTP, fix from Xin
    Long.

 3) Min packet MTU size is wrong in cpsw driver, from Grygorii Strashko.

 4) Fix build failure in netfilter ctnetlink, from Arnd Bergmann.

 5) ISDN hisax driver checks pnp_irq() for errors incorrectly, from
    Arvind Yadav.

 6) Fix fealnx driver build failure on MIPS, from Huacai Chen.

 7) Fix into leak in SCTP, the scope_id of socket addresses is not
    always filled in. From Eric W. Biederman.

 8) MTU inheritance between physical function and representor fix in nfp
    driver, from Dirk van der Merwe.

 9) Fix memory leak in rsi driver, from Colin Ian King.

10) Fix expiration and generation ID handling of cached ipv4 redirect
    routes, from Xin Long.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (40 commits)
  net: usb: hso.c: remove unneeded DRIVER_LICENSE #define
  ibmvnic: fix dma_mapping_error call
  ipvlan: NULL pointer dereference panic in ipvlan_port_destroy
  route: also update fnhe_genid when updating a route cache
  route: update fnhe_expires for redirect when the fnhe exists
  sctp: set frag_point in sctp_setsockopt_maxseg correctly
  rsi: fix memory leak on buf and usb_reg_buf
  net/netlabel: Add list_next_rcu() in rcu_dereference().
  nfp: remove false positive offloads in flower vxlan
  nfp: register flower reprs for egress dev offload
  nfp: inherit the max_mtu from the PF netdev
  nfp: fix vlan receive MAC statistics typo
  nfp: fix flower offload metadata flag usage
  virto_net: remove empty file 'virtio_net.'
  net/sctp: Always set scope_id in sctp_inet6_skb_msgname
  fealnx: Fix building error on MIPS
  isdn: hisax: Fix pnp_irq's error checking for setup_teles3
  isdn: hisax: Fix pnp_irq's error checking for setup_sedlbauer_isapnp
  isdn: hisax: Fix pnp_irq's error checking for setup_niccy
  isdn: hisax: Fix pnp_irq's error checking for setup_ix1micro
  ...

Merge tag 'hwlock-v4.15' of git://github.com/andersson/remoteproc

Pull hwspinlock update from Bjorn Andersson:
 "This changes the HWSPINLOCK core Kconfig option to bool, to aid when
  other core code depends on it"

* tag 'hwlock-v4.15' of git://github.com/andersson/remoteproc:
  hwspinlock: Change hwspinlock to a bool

Merge tag 'rproc-v4.15' of git://github.com/andersson/remoteproc

Pull remoteproc updates from Bjorn Andersson:
 "This adds an interface for configuring Qualcomm's "secure SMMU" and
  adds support for booting the modem Hexagon on MSM8996.

  Two new debugfs entries are added in the remoteproc core to introspect
  the list of memory carveouts and the loaded resource table"

* tag 'rproc-v4.15' of git://github.com/andersson/remoteproc:
  remoteproc: qcom: Fix error handling paths in order to avoid memory leaks
  remoteproc: qcom: Drop pr_err in q6v5_xfer_mem_ownership()
  remoteproc: debug: add carveouts list dump feature
  remoteproc: debug: add resource table dump feature
  remoteproc: qcom: Add support for mss remoteproc on msm8996
  remoteproc: qcom: Make secure world call for mem ownership switch
  remoteproc: qcom: refactor mss fw image loading sequence
  firmware: scm: Add new SCM call API for switching memory ownership

Merge tag 'rpmsg-v4.15' of git://github.com/andersson/remoteproc

Pull rpmsg updates from Bjorn Andersson:

 - turn RPMSG_VIRTIO into a user selectable config

 - fix few bugs in GLINK

 - provide the support for specifying initial buffer sizes for GLINK
   channels.

* tag 'rpmsg-v4.15' of git://github.com/andersson/remoteproc:
  rpmsg: glink: The mbox client knows_txdone
  rpmsg: glink: Add missing MODULE_LICENSE
  rpmsg: glink: Use best fit intent during tx
  rpmsg: glink: Add support to preallocate intents
  dt-bindings: soc: qcom: Support GLINK intents
  rpmsg: glink: Initialize the "intent_req_comp" completion variable
  rpmsg: Allow RPMSG_VIRTIO to be enabled via menuconfig or defconfig

Merge tag 'hwmon-for-linus-v4.15-take2' of git://git./linux/kernel/git/groeck/linux-staging

Pull more hwmon updates/fixes from Guenter Roeck:

 - minor bug fix in k10temp driver

 - take advantage of added NULL check in i2c_unregister_device()

* tag 'hwmon-for-linus-v4.15-take2' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging:
  hwmon: (w83793) Remove duplicate NULL check
  hwmon: (w83792d) Remove duplicate NULL check
  hwmon: (w83791d) Remove duplicate NULL check
  hwmon: (w83781d) Remove duplicate NULL check
  hwmon: (k10temp) Correct model name for Ryzen 1600X

Merge tag 'clk-for-linus' of git://git./linux/kernel/git/clk/linux

Pull clk updates from Stephen Boyd:
 "We have two changes to the core framework this time around.

  The first being a large change that introduces runtime PM support to
  the clk framework. Now we properly call runtime PM operations on the
  device providing a clk when the clk is in use. This helps on SoCs
  where the clks provided by a device need something to be powered on
  before using the clks, like power domains or regulators. It also helps
  power those things down when clks aren't in use.

  The other core change is a devm API addition for clk providers so we
  can get rid of a bunch of clk driver remove functions that are just
  doing of_clk_del_provider().

  Outside of the core, we have the usual addition of clk drivers and
  smattering of non-critical fixes to existing drivers. The biggest diff
  is support for Mediatek MT2712 and MT7622 SoCs, but those patches
  really just add a bunch of data.

  By the way, we're trying something new here where we build the tree up
  with topic branches. We plan to work this into our workflow so that we
  don't step on each other's toes, and so the fixes branch can be merged
  on an as-needed basis.

  Summary:

  Core:
   - runtime PM support for clk providers
   - devm API for of_clk_add_hw_provider()

  New Drivers:
   - Mediatek MT2712 and MT7622
   - Renesas R-Car V3M SoC

  Updates:
   - runtime PM support for Samsung exynos5433/exynos4412 providers
   - removal of clkdev aliases on Samsung SoCs
   - convert clk-gpio to use gpio descriptors
   - various driver cleanups to match kernel coding style
   - Amlogic Video Processing Unit VPU and VAPB clks
   - sigma-delta modulation for Allwinner audio PLLs
   - Allwinner A83t Display clks
   - support for the second display unit clock on Renesas RZ/G1E
   - suspend/resume support for Renesas R-Car Gen3 CPG/MSSR
   - new clock ids for Rockchip rk3188 and rk3368 SoCs
   - various 'const' markings on clk_ops structures
   - RPM clk support on Qualcomm MSM8996/MSM8660 SoCs"

* tag 'clk-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux: (137 commits)
  clk: stm32h7: fix test of clock config
  clk: pxa: fix building on older compilers
  clk: sunxi-ng: a83t: Fix i2c buses bits
  clk: ti: dra7-atl-clock: fix child-node lookups
  clk: qcom: common: fix legacy board-clock registration
  clk: uniphier: fix DAPLL2 clock rate of Pro5
  clk: uniphier: fix parent of miodmac clock data
  clk: hi3798cv200: correct parent mux clock for 'clk_sdio0_ciu'
  clk: hisilicon: Delete an error message for a failed memory allocation in hisi_register_clkgate_sep()
  clk: hi3660: fix incorrect uart3 clock freqency
  clk: kona-setup: Delete error messages for failed memory allocations
  ARC: clk: fix spelling mistake: "configurarion" -> "configuration"
  clk: cdce925: remove redundant check for non-null parent_name
  clk: versatile: Improve sizeof() usage
  clk: versatile: Delete error messages for failed memory allocations
  clk: ux500: Improve sizeof() usage
  clk: ux500: Delete error messages for failed memory allocations
  clk: spear: Delete error messages for failed memory allocations
  clk: ti: Delete error messages for failed memory allocations
  clk: mmp: Adjust checks for NULL pointers
  ...

Merge tag 'kbuild-misc-v4.15' of git://git./linux/kernel/git/masahiroy/linux-kbuild

Pull Kbuild misc updates from Masahiro Yamada:

 - Clean up and fix RPM package build

 - Fix a warning in DEB package build

 - Improve coccicheck script

 - Improve some semantic patches

* tag 'kbuild-misc-v4.15' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
  docs: dev-tools: coccinelle: delete out of date wiki reference
  coccinelle: orplus: reorganize to improve performance
  coccinelle: use exists to improve efficiency
  builddeb: Pass the kernel:debarch substvar to dpkg-genchanges
  Coccinelle: use false positive annotation
  coccinelle: fix verbose message about .cocci file being run
  coccinelle: grep Options and Requires fields more precisely
  Coccinelle: make DEBUG_FILE option more useful
  coccinelle: api: detect identical chip data arrays
  coccinelle: Improve setup_timer.cocci matching
  Coccinelle: setup_timer: improve messages from setup_timer
  kbuild: rpm-pkg: do not force -jN in submake
  kbuild: rpm-pkg: keep spec file until make mrproper
  kbuild: rpm-pkg: fix jobserver unavailable warning
  kbuild: rpm-pkg: replace $RPM_BUILD_ROOT with %{buildroot}
  kbuild: rpm-pkg: fix build error when CONFIG_MODULES is disabled
  kbuild: rpm-pkg: refactor mkspec with here doc
  kbuild: rpm-pkg: clean up mkspec
  kbuild: rpm-pkg: install vmlinux.bz2 unconditionally
  kbuild: rpm-pkg: remove ppc64 specific image handling

Merge tag 'kbuild-v4.15' of git://git./linux/kernel/git/masahiroy/linux-kbuild

Pull Kbuild updates from Masahiro Yamada:
 "One of the most remarkable improvements in this cycle is, Kbuild is
  now able to cache the result of shell commands. Some variables are
  expensive to compute, for example, $(call cc-option,...) invokes the
  compiler. It is not efficient to redo this computation every time,
  even when we are not actually building anything. Kbuild creates a
  hidden file ".cache.mk" that contains invoked shell commands and their
  results. The speed-up should be noticeable.

  Summary:

   - Fix arch build issues (hexagon, sh)

   - Clean up various Makefiles and scripts

   - Fix wrong usage of {CFLAGS,LDFLAGS}_MODULE in arch Makefiles

   - Cache variables that are expensive to compute

   - Improve cc-ldopton and ld-option for Clang

   - Optimize output directory creation"

* tag 'kbuild-v4.15' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild: (30 commits)
  kbuild: move coccicheck help from scripts/Makefile.help to top Makefile
  sh: decompressor: add shipped files to .gitignore
  frv: .gitignore: ignore vmlinux.lds
  selinux: remove unnecessary assignment to subdir-
  kbuild: specify FORCE in Makefile.headersinst as .PHONY target
  kbuild: remove redundant mkdir from ./Kbuild
  kbuild: optimize object directory creation for incremental build
  kbuild: create object directories simpler and faster
  kbuild: filter-out PHONY targets from "targets"
  kbuild: remove redundant $(wildcard ...) for cmd_files calculation
  kbuild: create directory for make cache only when necessary
  sh: select KBUILD_DEFCONFIG depending on ARCH
  kbuild: fix linker feature test macros when cross compiling with Clang
  kbuild: shrink .cache.mk when it exceeds 1000 lines
  kbuild: do not call cc-option before KBUILD_CFLAGS initialization
  kbuild: Cache a few more calls to the compiler
  kbuild: Add a cache for generated variables
  kbuild: add forward declaration of default target to Makefile.asm-generic
  kbuild: remove KBUILD_SUBDIR_ASFLAGS and KBUILD_SUBDIR_CCFLAGS
  hexagon/kbuild: replace CFLAGS_MODULE with KBUILD_CFLAGS_MODULE
  ...

net: usb: hso.c: remove unneeded DRIVER_LICENSE #define

There is no need to #define the license of the driver, just put it in
the MODULE_LICENSE() line directly as a text string.

This allows tools that check that the module license matches the source
code license to work properly, as there is no need to unwind the
unneeded dereference.

Cc: "David S. Miller" <davem@davemloft.net>
Cc: Andreas Kemnade <andreas@kemnade.info>
Cc: Johan Hovold <johan@kernel.org>
Reported-by: Philippe Ombredanne <pombredanne@nexb.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: Philippe Ombredanne <pombredanne@nexb.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ibmvnic: fix dma_mapping_error call

This patch fixes the dma_mapping_error call to use the correct dma_addr
which is inside the ibmvnic_vpd struct. Moreover, it fixes an uninitialized
warning regarding a local dma_addr variable which is not used anymore.

Fixes: 4e6759be28e4 ("ibmvnic: Feature implementation of VPD for the ibmvnic driver")
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Desnes A. Nunes do Rosario <desnesn@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipvlan: NULL pointer dereference panic in ipvlan_port_destroy

When call to register_netdevice() (called from ipvlan_link_new()) fails,
we call ipvlan_uninit() (through ndo_uninit()) to destroy the ipvlan
port. After returning unsuccessfully from register_netdevice() we go
ahead and call ipvlan_port_destroy() again which causes NULL pointer
dereference panic. Fix the issue by making ipvlan_init() and
ipvlan_uninit() call symmetric.

The ipvlan port will now be created inside ipvlan_init() and will be
destroyed in ipvlan_uninit().

Fixes: 2ad7bf363841 (ipvlan: Initial check-in of the IPVLAN driver)
Signed-off-by: Girish Moodalbail <girish.moodalbail@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

route: also update fnhe_genid when updating a route cache

Now when ip route flush cache and it turn out all fnhe_genid != genid.
If a redirect/pmtu icmp packet comes and the old fnhe is found and all
it's members but fnhe_genid will be updated.

Then next time when it looks up route and tries to rebind this fnhe to
the new dst, the fnhe will be flushed due to fnhe_genid != genid. It
causes this redirect/pmtu icmp packet acutally not to be applied.

This patch is to also reset fnhe_genid when updating a route cache.

Fixes: 5aad1de5ea2c ("ipv4: use separate genid for next hop exceptions")
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

route: update fnhe_expires for redirect when the fnhe exists

Now when creating fnhe for redirect, it sets fnhe_expires for this
new route cache. But when updating the exist one, it doesn't do it.
It will cause this fnhe never to be expired.

Paolo already noticed it before, in Jianlin's test case, it became
even worse:

When ip route flush cache, the old fnhe is not to be removed, but
only clean it's members. When redirect comes again, this fnhe will
be found and updated, but never be expired due to fnhe_expires not
being set.

So fix it by simply updating fnhe_expires even it's for redirect.

Fixes: aee06da6726d ("ipv4: use seqlock for nh_exceptions")
Reported-by: Jianlin Shi <jishi@redhat.com>
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

sctp: set frag_point in sctp_setsockopt_maxseg correctly

Now in sctp_setsockopt_maxseg user_frag or frag_point can be set with
val >= 8 and val <= SCTP_MAX_CHUNK_LEN. But both checks are incorrect.

val >= 8 means frag_point can even be less than SCTP_DEFAULT_MINSEGMENT.
Then in sctp_datamsg_from_user(), when it's value is greater than cookie
echo len and trying to bundle with cookie echo chunk, the first_len will
overflow.

The worse case is when it's value is equal as cookie echo len, first_len
becomes 0, it will go into a dead loop for fragment later on. In Hangbin
syzkaller testing env, oom was even triggered due to consecutive memory
allocation in that loop.

Besides, SCTP_MAX_CHUNK_LEN is the max size of the whole chunk, it should
deduct the data header for frag_point or user_frag check.

This patch does a proper check with SCTP_DEFAULT_MINSEGMENT subtracting
the sctphdr and datahdr, SCTP_MAX_CHUNK_LEN subtracting datahdr when
setting frag_point via sockopt. It also improves sctp_setsockopt_maxseg
codes.

Suggested-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Reported-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

rsi: fix memory leak on buf and usb_reg_buf

In the cases where len is too long, the error return path fails to
kfree allocated buffers buf and usb_reg_buf.  The simplest fix is to
perform the sanity check on len before the allocations to avoid having
to do the kfree'ing in the first place.

Detected by CoverityScan, CID#1452258,1452259 ("Resource Leak")

Fixes: 59f73e2ae185 ("rsi: check length before USB read/write register")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/netlabel: Add list_next_rcu() in rcu_dereference().

Add list_next_rcu() for fetching next list in rcu_deference safely.

Found with sparse in linux-next tree on tag next-20171116.

Signed-off-by: Tim Hansen <devtimhansen@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'akpm' (patches from Andrew)

Merge more updates from Andrew Morton:

 - a bit more MM

 - procfs updates

 - dynamic-debug fixes

 - lib/ updates

 - checkpatch

 - epoll

 - nilfs2

 - signals

 - rapidio

 - PID management cleanup and optimization

 - kcov updates

 - sysvipc updates

 - quite a few misc things all over the place

* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (94 commits)
  EXPERT Kconfig menu: fix broken EXPERT menu
  include/asm-generic/topology.h: remove unused parent_node() macro
  arch/tile/include/asm/topology.h: remove unused parent_node() macro
  arch/sparc/include/asm/topology_64.h: remove unused parent_node() macro
  arch/sh/include/asm/topology.h: remove unused parent_node() macro
  arch/ia64/include/asm/topology.h: remove unused parent_node() macro
  drivers/pcmcia/sa1111_badge4.c: avoid unused function warning
  mm: add infrastructure for get_user_pages_fast() benchmarking
  sysvipc: make get_maxid O(1) again
  sysvipc: properly name ipc_addid() limit parameter
  sysvipc: duplicate lock comments wrt ipc_addid()
  sysvipc: unteach ids->next_id for !CHECKPOINT_RESTORE
  initramfs: use time64_t timestamps
  drivers/watchdog: make use of devm_register_reboot_notifier()
  kernel/reboot.c: add devm_register_reboot_notifier()
  kcov: update documentation
  Makefile: support flag -fsanitizer-coverage=trace-cmp
  kcov: support comparison operands collection
  kcov: remove pointless current != NULL check
  kernel/panic.c: add TAINT_AUX
  ...

EXPERT Kconfig menu: fix broken EXPERT menu

Clean up the EXPERT menu (yet again).

Move FHANDLE and CHECKPOINT_RESTORE into the primary EXPERT menu since
they already depend on EXPERT.

Move BPF_SYSCALL and USERFAULTFD out of the EXPERT Kconfig symbols menu
list since they do not depend on EXPERT and were breaking the continuity
of that menu list.

Move all of the KALLSYMS Kconfig symbols to the end of the EXPERT menu.
This separates the kernel services from the build options.

This patch depends on [PATCH] pci: move PCI_QUIRKS to the PCI bus menu
(https://lkml.org/lkml/2017/11/2/907).

Link: http://lkml.kernel.org/r/72e4465a-a5ff-cb3c-1a90-11aa4861b161@infradead.org
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Acked-by: Daniel Borkmann <daniel@iogearbox.net> [BPF]
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

include/asm-generic/topology.h: remove unused parent_node() macro

Commit a7be6e5a7f8d ("mm: drop useless local parameters of
__register_one_node()") removed the last user of parent_node().

The parent_node() macro in generic situation is unnecessary.

Remove it for cleanup.

Link: http://lkml.kernel.org/r/1504234599-29533-8-git-send-email-douly.fnst@cn.fujitsu.com
Signed-off-by: Dou Liyang <douly.fnst@cn.fujitsu.com>
Reported-by: Michael Ellerman <mpe@ellerman.id.au>
Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

arch/tile/include/asm/topology.h: remove unused parent_node() macro

Commit a7be6e5a7f8d ("mm: drop useless local parameters of
__register_one_node()") removed the last user of parent_node().

The parent_node() macro in tile platform is unnecessary.

Remove it for cleanup.

Link: http://lkml.kernel.org/r/1504234599-29533-7-git-send-email-douly.fnst@cn.fujitsu.com
Signed-off-by: Dou Liyang <douly.fnst@cn.fujitsu.com>
Reported-by: Michael Ellerman <mpe@ellerman.id.au>
Acked-by: Chris Metcalf <cmetcalf@mellanox.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

arch/sparc/include/asm/topology_64.h: remove unused parent_node() macro

Commit a7be6e5a7f8d ("mm: drop useless local parameters of
__register_one_node()") removed the last user of parent_node().

The parent_node() macro in SPARC64 platform is unnecessary.

Remove it for cleanup.

Link: http://lkml.kernel.org/r/1504234599-29533-6-git-send-email-douly.fnst@cn.fujitsu.com
Signed-off-by: Dou Liyang <douly.fnst@cn.fujitsu.com>
Reported-by: Michael Ellerman <mpe@ellerman.id.au>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

arch/sh/include/asm/topology.h: remove unused parent_node() macro

Commit a7be6e5a7f8d ("mm: drop useless local parameters of
__register_one_node()") removed the last user of parent_node().

The parent_node() macro in SUPERH platform is unnecessary.

Remove it for cleanup.

Link: http://lkml.kernel.org/r/1504234599-29533-5-git-send-email-douly.fnst@cn.fujitsu.com
Signed-off-by: Dou Liyang <douly.fnst@cn.fujitsu.com>
Reported-by: Michael Ellerman <mpe@ellerman.id.au>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Rich Felker <dalias@libc.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

arch/ia64/include/asm/topology.h: remove unused parent_node() macro

Commit a7be6e5a7f8d ("mm: drop useless local parameters of
__register_one_node()") removed the last user of parent_node().

The parent_node() macro in IA64(Itanium) platform is unnecessary.

Remove it for cleanup.

Link: http://lkml.kernel.org/r/1504234599-29533-2-git-send-email-douly.fnst@cn.fujitsu.com
Signed-off-by: Dou Liyang <douly.fnst@cn.fujitsu.com>
Reported-by: Michael Ellerman <mpe@ellerman.id.au>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

drivers/pcmcia/sa1111_badge4.c: avoid unused function warning

pcmv_setup() is only used when the badge4 driver is built-in, but not
when it is a loadable module:

  drivers/pcmcia/sa1111_badge4.c:153:122: error: 'pcmv_setup' defined but not used [-Werror=unused-function]

This adds an #ifdef to avoid the definition of the unused function in
the modular case.

Link: http://lkml.kernel.org/r/20170911201133.3421636-1-arnd@arndb.de
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Russell King <rmk@armlinux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm: add infrastructure for get_user_pages_fast() benchmarking

Performance of get_user_pages_fast() is critical for some workloads, but
it's tricky to test it directly.

This patch provides /sys/kernel/debug/gup_benchmark that helps with
testing performance of it.

See tools/testing/selftests/vm/gup_benchmark.c for userspace
counterpart.

Link: http://lkml.kernel.org/r/20170908215603.9189-2-kirill.shutemov@linux.intel.com
Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thorsten Leemhuis <regressions@leemhuis.info>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Huang Ying <ying.huang@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

sysvipc: make get_maxid O(1) again

For a custom microbenchmark on a 3.30GHz Xeon SandyBridge, which calls
IPC_STAT over and over, it was calculated that, on avg the cost of
ipc_get_maxid() for increasing amounts of keys was:

 10 keys: ~900 cycles
 100 keys: ~15000 cycles
 1000 keys: ~150000 cycles
 10000 keys: ~2100000 cycles

This is unsurprising as maxid is currently O(n).

By having the max_id available in O(1) we save all those cycles for each
semctl(_STAT) command, the idr_find can be expensive -- which some real
(customer) workloads actually poll on.

Note that this used to be the case, until commit 7ca7e564e04 ("ipc:
store ipcs into IDRs").  The cost is the extra idr_find when doing
RMIDs, but we simply go backwards, and should not take too many
iterations to find the new value.

[akpm@linux-foundation.org: coding-style fixes]
Link: http://lkml.kernel.org/r/20170831172049.14576-5-dave@stgolabs.net
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
Cc: Manfred Spraul <manfred@colorfullife.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

sysvipc: properly name ipc_addid() limit parameter

This is better understood as a limit, instead of size; exactly like the
function comment indicates.  Rename it.

Link: http://lkml.kernel.org/r/20170831172049.14576-4-dave@stgolabs.net
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
Cc: Manfred Spraul <manfred@colorfullife.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

sysvipc: duplicate lock comments wrt ipc_addid()

The comment in msgqueues when using ipc_addid() is quite useful imo.
Duplicate it for shm and semaphores.

Link: http://lkml.kernel.org/r/20170831172049.14576-3-dave@stgolabs.net
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
Cc: Manfred Spraul <manfred@colorfullife.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

sysvipc: unteach ids->next_id for !CHECKPOINT_RESTORE

Patch series "sysvipc: ipc-key management improvements".

Here are a few improvements I spotted while eyeballing Guillaume's
rhashtable implementation for ipc keys.  The first and fourth patches
are the interesting ones, the middle two are trivial.

This patch (of 4):

The next_id object-allocation functionality was introduced in commit
03f595668017 ("ipc: add sysctl to specify desired next object id").

Given that these new entries are _only_ exported under the
CONFIG_CHECKPOINT_RESTORE option, there is no point for the common case
to even know about ->next_id.  As such rewrite ipc_buildid() such that
it can do away with the field as well as unnecessary branches when
adding a new identifier.  The end result also better differentiates both
cases, so the code ends up being cleaner; albeit the small duplications
regarding the default case.

[akpm@linux-foundation.org: coding-style fixes]
Link: http://lkml.kernel.org/r/20170831172049.14576-2-dave@stgolabs.net
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
Cc: Manfred Spraul <manfred@colorfullife.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

initramfs: use time64_t timestamps

The cpio format uses a 32-bit number to encode file timestamps, which
breaks initramfs support in 2038.  This reinterprets the timestamp as
unsigned, to give us another 68 years and avoids breaking until 2106.

Link: http://lkml.kernel.org/r/20171019095536.801199-1-arnd@arndb.de
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Deepa Dinamani <deepa.kernel@gmail.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Daniel Thompson <daniel.thompson@linaro.org>
Cc: Lokesh Vutla <lokeshvutla@ti.com>
Cc: Stafford Horne <shorne@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

drivers/watchdog: make use of devm_register_reboot_notifier()

Save a bit of cleanup code by leveraging newly added
devm_register_reboot_notifier().

[akpm@linux-foundation.org: small cleanup: avoid 80-col tricks]
Link: http://lkml.kernel.org/r/20170411160615.9784-1-andrew.smirnov@gmail.com
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Acked-by: Guenter Roeck <linux@roeck-us.net>
Cc: Chris Healy <cphealy@gmail.com>
Cc: Wim Van Sebroeck <wim@iguana.be>
Cc: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kernel/reboot.c: add devm_register_reboot_notifier()

Add devm_* wrapper around register_reboot_notifier to simplify device
specific reboot notifier registration/unregistration.

[akpm@linux-foundation.org: move `struct device' forward decl to top-of-file]
Link: http://lkml.kernel.org/r/20170320171753.1705-1-andrew.smirnov@gmail.com
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kcov: update documentation

The updated documentation describes new KCOV mode for collecting
comparison operands.

Link: http://lkml.kernel.org/r/20171011095459.70721-3-glider@google.com
Signed-off-by: Victor Chibotaru <tchibo@google.com>
Signed-off-by: Alexander Potapenko <glider@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Andrey Konovalov <andreyknvl@google.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Alexander Popov <alex.popov@linux.com>
Cc: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Vegard Nossum <vegard.nossum@oracle.com>
Cc: Quentin Casasnovas <quentin.casasnovas@oracle.com>
Cc: <syzkaller@googlegroups.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Makefile: support flag -fsanitizer-coverage=trace-cmp

The flag enables Clang instrumentation of comparison operations
(currently not supported by GCC).  This instrumentation is needed by the
new KCOV device to collect comparison operands.

Link: http://lkml.kernel.org/r/20171011095459.70721-2-glider@google.com
Signed-off-by: Victor Chibotaru <tchibo@google.com>
Signed-off-by: Alexander Potapenko <glider@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Andrey Konovalov <andreyknvl@google.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Alexander Popov <alex.popov@linux.com>
Cc: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Vegard Nossum <vegard.nossum@oracle.com>
Cc: Quentin Casasnovas <quentin.casasnovas@oracle.com>
Cc: <syzkaller@googlegroups.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kcov: support comparison operands collection

Enables kcov to collect comparison operands from instrumented code.
This is done by using Clang's -fsanitize=trace-cmp instrumentation
(currently not available for GCC).

The comparison operands help a lot in fuzz testing.  E.g.  they are used
in Syzkaller to cover the interiors of conditional statements with way
less attempts and thus make previously unreachable code reachable.

To allow separate collection of coverage and comparison operands two
different work modes are implemented.  Mode selection is now done via a
KCOV_ENABLE ioctl call with corresponding argument value.

Link: http://lkml.kernel.org/r/20171011095459.70721-1-glider@google.com
Signed-off-by: Victor Chibotaru <tchibo@google.com>
Signed-off-by: Alexander Potapenko <glider@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Andrey Konovalov <andreyknvl@google.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Alexander Popov <alex.popov@linux.com>
Cc: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Vegard Nossum <vegard.nossum@oracle.com>
Cc: Quentin Casasnovas <quentin.casasnovas@oracle.com>
Cc: <syzkaller@googlegroups.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kcov: remove pointless current != NULL check

__sanitizer_cov_trace_pc() is a hot code, so it's worth to remove
pointless '!current' check.  Current is never NULL.

Link: http://lkml.kernel.org/r/20170929162221.32500-1-aryabinin@virtuozzo.com
Signed-off-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Acked-by: Dmitry Vyukov <dvyukov@google.com>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Cc: Andrey Konovalov <andreyknvl@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kernel/panic.c: add TAINT_AUX

This is the gist of a patch which we've been forward-porting in our
kernels for a long time now and it probably would make a good sense to
have such TAINT_AUX flag upstream which can be used by each distro etc,
how they see fit.  This way, we won't need to forward-port a distro-only
version indefinitely.

Add an auxiliary taint flag to be used by distros and others.  This
obviates the need to forward-port whatever internal solutions people
have in favor of a single flag which they can map arbitrarily to a
definition of their pleasing.

The "X" mnemonic could also mean eXternal, which would be taint from a
distro or something else but not the upstream kernel.  We will use it to
mark modules for which we don't provide support.  I.e., a really
eXternal module.

Link: http://lkml.kernel.org/r/20170911134533.dp5mtyku5bongx4c@pd.tnic
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: Kees Cook <keescook@chromium.org>
Cc: Jessica Yu <jeyu@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Jiri Slaby <jslaby@suse.cz>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Michal Marek <mmarek@suse.cz>
Cc: Jiri Kosina <jkosina@suse.cz>
Cc: Takashi Iwai <tiwai@suse.de>
Cc: Petr Mladek <pmladek@suse.com>
Cc: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

pid: remove pidhash

pidhash is no longer required as all the information can be looked up
from idr tree.  nr_hashed represented the number of pids that had been
hashed.  Since, nr_hashed and PIDNS_HASH_ADDING are no longer relevant,
it has been renamed to pid_allocated and PIDNS_ADDING respectively.

[gs051095@gmail.com: v6]
Link: http://lkml.kernel.org/r/1507760379-21662-3-git-send-email-gs051095@gmail.com
Link: http://lkml.kernel.org/r/1507583624-22146-3-git-send-email-gs051095@gmail.com
Signed-off-by: Gargi Sharma <gs051095@gmail.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Tested-by: Tony Luck <tony.luck@intel.com> [ia64]
Cc: Julia Lawall <julia.lawall@lip6.fr>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Pavel Tatashin <pasha.tatashin@oracle.com>
Cc: Kirill Tkhai <ktkhai@virtuozzo.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Eric W. Biederman <ebiederm@xmission.com>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

pid: replace pid bitmap implementation with IDR API

Patch series "Replacing PID bitmap implementation with IDR API", v4.

This series replaces kernel bitmap implementation of PID allocation with
IDR API.  These patches are written to simplify the kernel by replacing
custom code with calls to generic code.

The following are the stats for pid and pid_namespace object files
before and after the replacement.  There is a noteworthy change between
the IDR and bitmap implementation.

Before
   text       data        bss        dec        hex    filename
   8447       3894         64      12405       3075    kernel/pid.o
After
   text       data        bss        dec        hex    filename
   3397        304          0       3701        e75    kernel/pid.o

Before
   text       data        bss        dec        hex    filename
   5692       1842        192       7726       1e2e    kernel/pid_namespace.o
After
   text       data        bss        dec        hex    filename
   2854        216         16       3086        c0e    kernel/pid_namespace.o

The following are the stats for ps, pstree and calling readdir on /proc
for 10,000 processes.

ps:
        With IDR API    With bitmap
real    0m1.479s        0m2.319s
user    0m0.070s        0m0.060s
sys     0m0.289s        0m0.516s

pstree:
        With IDR API    With bitmap
real    0m1.024s        0m1.794s
user    0m0.348s        0m0.612s
sys     0m0.184s        0m0.264s

proc:
        With IDR API    With bitmap
real    0m0.059s        0m0.074s
user    0m0.000s        0m0.004s
sys     0m0.016s        0m0.016s

This patch (of 2):

Replace the current bitmap implementation for Process ID allocation.
Functions that are no longer required, for example, free_pidmap(),
alloc_pidmap(), etc.  are removed.  The rest of the functions are
modified to use the IDR API.  The change was made to make the PID
allocation less complex by replacing custom code with calls to generic
API.

[gs051095@gmail.com: v6]
Link: http://lkml.kernel.org/r/1507760379-21662-2-git-send-email-gs051095@gmail.com
[avagin@openvz.org: restore the old behaviour of the ns_last_pid sysctl]
Link: http://lkml.kernel.org/r/20171106183144.16368-1-avagin@openvz.org
Link: http://lkml.kernel.org/r/1507583624-22146-2-git-send-email-gs051095@gmail.com
Signed-off-by: Gargi Sharma <gs051095@gmail.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Cc: Julia Lawall <julia.lawall@lip6.fr>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Pavel Tatashin <pasha.tatashin@oracle.com>
Cc: Kirill Tkhai <ktkhai@virtuozzo.com>
Cc: Eric W. Biederman <ebiederm@xmission.com>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kernel/sysctl.c: code cleanups

Remove unnecessary else block, remove redundant return and call to kfree
in if block.

Link: http://lkml.kernel.org/r/1510238435-1655-1-git-send-email-mail@okal.no
Signed-off-by: Ola N. Kaldestad <mail@okal.no>
Acked-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Documentation/sysctl/vm.txt: fix typo

Link: http://lkml.kernel.org/r/CAKW4uUyCi=PnKf3epgFVz8z=1tMtHSOHNm+fdNxrNw3-THvRCA@mail.gmail.com
Signed-off-by: Kangmin Park <l4stpr0gr4m@gmail.com>
Cc: Jiri Kosina <trivial@kernel.org>
Cc: Alan Cox <alan@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

drivers/rapidio/devices/rio_mport_cdev.c: fix error handling in 'rio_dma_transfer()'

In case of error, 'dma_map_sg()' returns 0, not a negative value.  There
is BUG_ON() in 'dma_map_sg_attrs()' which makes sure of that.

Link: http://lkml.kernel.org/r/d4235bd2b9274e99f6c86ea71b1fa1c7bd8d0c08.1505687047.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Reviewed-by: Logan Gunthorpe <logang@deltatee.com>
Cc: Matt Porter <mporter@kernel.crashing.org>
Cc: Alexandre Bounine <alexandre.bounine@idt.com>
Cc: Lorenzo Stoakes <lstoakes@gmail.com>
Cc: Jesper Nilsson <jesper.nilsson@axis.com>
Cc: Christian K_nig <christian.koenig@amd.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

drivers/rapidio/devices/rio_mport_cdev.c: fix resource leak in error handling path in 'rio_dma_transfer()'

If 'dma_map_sg()', we should branch to the existing error handling path
to free some resources before returning.

Link: http://lkml.kernel.org/r/61292a4f369229eee03394247385e955027283f8.1505687047.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Reviewed-by: Logan Gunthorpe <logang@deltatee.com>
Cc: Matt Porter <mporter@kernel.crashing.org>
Cc: Alexandre Bounine <alexandre.bounine@idt.com>
Cc: Lorenzo Stoakes <lstoakes@gmail.com>
Cc: Jesper Nilsson <jesper.nilsson@axis.com>
Cc: Christian K_nig <christian.koenig@amd.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

rapidio: constify rio_device_id

rio_device_id are not supposed to change at runtime.  rio driver is
working with const 'id_table'.  So mark the non-const rio_device_id
structs as const.

Link: http://lkml.kernel.org/r/1503734627-6058-2-git-send-email-arvind.yadav.cs@gmail.com
Link: http://lkml.kernel.org/r/1503734627-6058-3-git-send-email-arvind.yadav.cs@gmail.com
Link: http://lkml.kernel.org/r/1503734627-6058-4-git-send-email-arvind.yadav.cs@gmail.com
Link: http://lkml.kernel.org/r/1503734627-6058-5-git-send-email-arvind.yadav.cs@gmail.com
Link: http://lkml.kernel.org/r/1503734627-6058-6-git-send-email-arvind.yadav.cs@gmail.com
Signed-off-by: Arvind Yadav <arvind.yadav.cs@gmail.com>
Acked-by: Alexandre Bounine <alexandre.bounine@idt.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kdump: print a message in case parse_crashkernel_mem resulted in zero bytes

parse_crashkernel_mem() silently returns if we get zero bytes in the
parsing function.  It is useful for debugging to add a message,
especially if the kernel cannot boot correctly.

Add a pr_info instead of pr_warn because it is expected behavior for
size = 0, eg.  crashkernel=2G-4G:128M, size will be 0 in case system
memory is less than 2G.

Link: http://lkml.kernel.org/r/20171114080129.GA6115@dhcp-128-65.nay.redhat.com
Signed-off-by: Dave Young <dyoung@redhat.com>
Cc: Baoquan He <bhe@redhat.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Bhupesh Sharma <bhsharma@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kernel/signal.c: remove the no longer needed SIGNAL_UNKILLABLE check in complete_signal()

complete_signal() checks SIGNAL_UNKILLABLE before it starts to destroy
the thread group, today this is wrong in many ways.

If nothing else, fatal_signal_pending() should always imply that the
whole thread group (except ->group_exit_task if it is not NULL) is
killed, this check breaks the rule.

After the previous changes we can rely on sig_task_ignored();
sig_fatal(sig) && SIGNAL_UNKILLABLE can only be true if we actually want
to kill this task and sig == SIGKILL OR it is traced and debugger can
intercept the signal.

This should hopefully fix the problem reported by Dmitry.  This
test-case

static int init(void *arg)
{
for (;;)
pause();
}

int main(void)
{
char stack[16 * 1024];

for (;;) {
int pid = clone(init, stack + sizeof(stack)/2,
CLONE_NEWPID | SIGCHLD, NULL);
assert(pid > 0);

assert(ptrace(PTRACE_ATTACH, pid, 0, 0) == 0);
assert(waitpid(-1, NULL, WSTOPPED) == pid);

assert(ptrace(PTRACE_DETACH, pid, 0, SIGSTOP) == 0);
assert(syscall(__NR_tkill, pid, SIGKILL) == 0);
assert(pid == wait(NULL));
}
}

triggers the WARN_ON_ONCE(!(task->jobctl & JOBCTL_STOP_PENDING)) in
task_participate_group_stop().  do_signal_stop()->signal_group_exit()
checks SIGNAL_GROUP_EXIT and return false, but task_set_jobctl_pending()
checks fatal_signal_pending() and does not set JOBCTL_STOP_PENDING.

And his should fix the minor security problem reported by Kyle,
SECCOMP_RET_TRACE can miss fatal_signal_pending() the same way if the
task is the root of a pid namespace.

Link: http://lkml.kernel.org/r/20171103184246.GD21036@redhat.com
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Reported-by: Dmitry Vyukov <dvyukov@google.com>
Reported-by: Kyle Huey <me@kylehuey.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Tested-by: Kyle Huey <me@kylehuey.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kernel/signal.c: protect the SIGNAL_UNKILLABLE tasks from !sig_kernel_only() signals

Change sig_task_ignored() to drop the SIG_DFL && !sig_kernel_only()
signals even if force == T.  This simplifies the next change and this
matches the same check in get_signal() which will drop these signals
anyway.

Link: http://lkml.kernel.org/r/20171103184227.GC21036@redhat.com
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Tested-by: Kyle Huey <me@kylehuey.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kernel/signal.c: protect the traced SIGNAL_UNKILLABLE tasks from SIGKILL

The comment in sig_ignored() says "Tracers may want to know about even
ignored signals" but SIGKILL can not be reported to debugger and it is
just wrong to return 0 in this case: SIGKILL should only kill the
SIGNAL_UNKILLABLE task if it comes from the parent ns.

Change sig_ignored() to ignore ->ptrace if sig == SIGKILL and rely on
sig_task_ignored().

SISGTOP coming from within the namespace is not really right too but at
least debugger can intercept it, and we can't drop it here because this
will break "gdb -p 1": ptrace_attach() won't work.  Perhaps we will add
another ->ptrace check later, we will see.

Link: http://lkml.kernel.org/r/20171103184206.GB21036@redhat.com
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Tested-by: Kyle Huey <me@kylehuey.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

fat: remove redundant assignment of 0 to slots

The variable slots is being assigned a value of zero that is never read,
slots is being updated again a few lines later.  Remove this redundant
assignment.

Cleans clang warning: Value stored to 'slots' is never read

Link: http://lkml.kernel.org/r/20171017140258.22536-1-colin.king@canonical.com
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Acked-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

hfs/hfsplus: clean up unused variables in bnode.c

Delete variables 'tree' and 'sb', which are set but never used.

Link: http://lkml.kernel.org/r/1507977146-15875-1-git-send-email-chris.gekas@gmail.com
Signed-off-by: Christos Gkekas <chris.gekas@gmail.com>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

nilfs2: remove inode->i_version initialization

It's never used in nilfs2.

Link: http://lkml.kernel.org/r/1510064486-1728-2-git-send-email-konishi.ryusuke@lab.ntt.co.jp
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

nilfs2: use octal for unreadable permission macro

Replace S_IRWXUGO with 0777 because symbolic permissions are considered
harmful:

 https://lwn.net/Articles/696229/

Link: http://lkml.kernel.org/r/1509367935-3086-5-git-send-email-konishi.ryusuke@lab.ntt.co.jp
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

nilfs2: align block comments of nilfs_sufile_truncate_range() at *

Fix the following checkpatch warning:

 WARNING: Block comments should align the * on each line
 #633: FILE: sufile.c:633:
 +/**
 +  * nilfs_sufile_truncate_range - truncate range of segment array

Link: http://lkml.kernel.org/r/1509367935-3086-4-git-send-email-konishi.ryusuke@lab.ntt.co.jp
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

fs, nilfs: convert nilfs_root.count from atomic_t to refcount_t

atomic_t variables are currently used to implement reference counters
with the following properties:

 - counter is initialized to 1 using atomic_set()
 - a resource is freed upon counter reaching zero
 - once counter reaches zero, its further
   increments aren't allowed
 - counter schema uses basic atomic operations
   (set, inc, inc_not_zero, dec_and_test, etc.)

Such atomic variables should be converted to a newly provided refcount_t
type and API that prevents accidental counter overflows and underflows.
This is important since overflows and underflows can lead to
use-after-free situation and be exploitable.

The variable nilfs_root.count is used as pure reference counter.
Convert it to refcount_t and fix up the operations.

Link: http://lkml.kernel.org/r/1509367935-3086-3-git-send-email-konishi.ryusuke@lab.ntt.co.jp
Signed-off-by: Elena Reshetova <elena.reshetova@intel.com>
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
Suggested-by: Kees Cook <keescook@chromium.org>
Reviewed-by: David Windsor <dwindsor@gmail.com>
Reviewed-by: Hans Liljestrand <ishkamiel@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

nilfs2: fix race condition that causes file system corruption

There is a race condition between nilfs_dirty_inode() and
nilfs_set_file_dirty().

When a file is opened, nilfs_dirty_inode() is called to update the
access timestamp in the inode.  It calls __nilfs_mark_inode_dirty() in a
separate transaction.  __nilfs_mark_inode_dirty() caches the ifile
buffer_head in the i_bh field of the inode info structure and marks it
as dirty.

After some data was written to the file in another transaction, the
function nilfs_set_file_dirty() is called, which adds the inode to the
ns_dirty_files list.

Then the segment construction calls nilfs_segctor_collect_dirty_files(),
which goes through the ns_dirty_files list and checks the i_bh field.
If there is a cached buffer_head in i_bh it is not marked as dirty
again.

Since nilfs_dirty_inode() and nilfs_set_file_dirty() use separate
transactions, it is possible that a segment construction that writes out
the ifile occurs in-between the two.  If this happens the inode is not
on the ns_dirty_files list, but its ifile block is still marked as dirty
and written out.

In the next segment construction, the data for the file is written out
and nilfs_bmap_propagate() updates the b-tree.  Eventually the bmap root
is written into the i_bh block, which is not dirty, because it was
written out in another segment construction.

As a result the bmap update can be lost, which leads to file system
corruption.  Either the virtual block address points to an unallocated
DAT block, or the DAT entry will be reused for something different.

The error can remain undetected for a long time.  A typical error
message would be one of the "bad btree" errors or a warning that a DAT
entry could not be found.

This bug can be reproduced reliably by a simple benchmark that creates
and overwrites millions of 4k files.

Link: http://lkml.kernel.org/r/1509367935-3086-2-git-send-email-konishi.ryusuke@lab.ntt.co.jp
Signed-off-by: Andreas Rohner <andreas.rohner@gmx.net>
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
Tested-by: Andreas Rohner <andreas.rohner@gmx.net>
Tested-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

fs/nilfs2: convert timers to use timer_setup()

In preparation for unconditionally passing the struct timer_list pointer
to all timer callbacks, switch to using the new timer_setup() and
from_timer() to pass the timer pointer explicitly.  This requires adding
a pointer to hold the timer's target task, as the lifetime of sc_task
doesn't appear to match the timer's task.

Link: http://lkml.kernel.org/r/20171016235900.GA102729@beast
Signed-off-by: Kees Cook <keescook@chromium.org>
Acked-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

sysctl: check for UINT_MAX before unsigned int min/max

Mikulas noticed in the existing do_proc_douintvec_minmax_conv() and
do_proc_dopipe_max_size_conv() introduced in this patchset, that they
inconsistently handle overflow and min/max range inputs:

For example:

  0 ... param->min - 1 ---> ERANGE
  param->min ... param->max ---> the value is accepted
  param->max + 1 ... 0x100000000L + param->min - 1 ---> ERANGE
  0x100000000L + param->min ... 0x100000000L + param->max ---> EINVAL
  0x100000000L + param->max + 1, 0x200000000L + param->min - 1 ---> ERANGE
  0x200000000L + param->min ... 0x200000000L + param->max ---> EINVAL
  0x200000000L + param->max + 1, 0x300000000L + param->min - 1 ---> ERANGE

In do_proc_do*() routines which store values into unsigned int variables
(4 bytes wide for 64-bit builds), first validate that the input unsigned
long value (8 bytes wide for 64-bit builds) will fit inside the smaller
unsigned int variable.  Then check that the unsigned int value falls
inside the specified parameter min, max range.  Otherwise the unsigned
long -> unsigned int conversion drops leading bits from the input value,
leading to the inconsistent pattern Mikulas documented above.

Link: http://lkml.kernel.org/r/1507658689-11669-5-git-send-email-joe.lawrence@redhat.com
Signed-off-by: Joe Lawrence <joe.lawrence@redhat.com>
Reported-by: Mikulas Patocka <mpatocka@redhat.com>
Reviewed-by: Mikulas Patocka <mpatocka@redhat.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

pipe: add proc_dopipe_max_size() to safely assign pipe_max_size

pipe_max_size is assigned directly via procfs sysctl:

  static struct ctl_table fs_table[] = {
          ...
          {
                  .procname       = "pipe-max-size",
                  .data           = &pipe_max_size,
                  .maxlen         = sizeof(int),
                  .mode           = 0644,
                  .proc_handler   = &pipe_proc_fn,
                  .extra1         = &pipe_min_size,
          },
          ...

  int pipe_proc_fn(struct ctl_table *table, int write, void __user *buf,
                   size_t *lenp, loff_t *ppos)
  {
          ...
          ret = proc_dointvec_minmax(table, write, buf, lenp, ppos)
          ...

and then later rounded in-place a few statements later:

          ...
          pipe_max_size = round_pipe_size(pipe_max_size);
          ...

This leaves a window of time between initial assignment and rounding
that may be visible to other threads.  (For example, one thread sets a
non-rounded value to pipe_max_size while another reads its value.)

Similar reads of pipe_max_size are potentially racy:

  pipe.c :: alloc_pipe_info()
  pipe.c :: pipe_set_size()

Add a new proc_dopipe_max_size() that consolidates reading the new value
from the user buffer, verifying bounds, and calling round_pipe_size()
with a single assignment to pipe_max_size.

Link: http://lkml.kernel.org/r/1507658689-11669-4-git-send-email-joe.lawrence@redhat.com
Signed-off-by: Joe Lawrence <joe.lawrence@redhat.com>
Reported-by: Mikulas Patocka <mpatocka@redhat.com>
Reviewed-by: Mikulas Patocka <mpatocka@redhat.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

pipe: avoid round_pipe_size() nr_pages overflow on 32-bit

round_pipe_size() contains a right-bit-shift expression which may
overflow, which would cause undefined results in a subsequent
roundup_pow_of_two() call.

  static inline unsigned int round_pipe_size(unsigned int size)
  {
          unsigned long nr_pages;

          nr_pages = (size + PAGE_SIZE - 1) >> PAGE_SHIFT;
          return roundup_pow_of_two(nr_pages) << PAGE_SHIFT;
  }

PAGE_SIZE is defined as (1UL << PAGE_SHIFT), so:
  - 4 bytes wide on 32-bit (0 to 0xffffffff)
  - 8 bytes wide on 64-bit (0 to 0xffffffffffffffff)

That means that 32-bit round_pipe_size(), nr_pages may overflow to 0:

  size=0x00000000    nr_pages=0x0
  size=0x00000001    nr_pages=0x1
  size=0xfffff000    nr_pages=0xfffff
  size=0xfffff001    nr_pages=0x0         << !
  size=0xffffffff    nr_pages=0x0         << !

This is bad because roundup_pow_of_two(n) is undefined when n == 0!

64-bit is not a problem as the unsigned int size is 4 bytes wide
(similar to 32-bit) and the larger, 8 byte wide unsigned long, is
sufficient to handle the largest value of the bit shift expression:

  size=0xffffffff    nr_pages=100000

Modify round_pipe_size() to return 0 if n == 0 and updates its callers to
handle accordingly.

Link: http://lkml.kernel.org/r/1507658689-11669-3-git-send-email-joe.lawrence@redhat.com
Signed-off-by: Joe Lawrence <joe.lawrence@redhat.com>
Reported-by: Mikulas Patocka <mpatocka@redhat.com>
Reviewed-by: Mikulas Patocka <mpatocka@redhat.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

pipe: match pipe_max_size data type with procfs

Patch series "A few round_pipe_size() and pipe-max-size fixups", v3.

While backporting Michael's "pipe: fix limit handling" patchset to a
distro-kernel, Mikulas noticed that current upstream pipe limit handling
contains a few problems:

  1 - procfs signed wrap: echo'ing a large number into
      /proc/sys/fs/pipe-max-size and then cat'ing it back out shows a
      negative value.

  2 - round_pipe_size() nr_pages overflow on 32bit:  this would
      subsequently try roundup_pow_of_two(0), which is undefined.

  3 - visible non-rounded pipe-max-size value: there is no mutual
      exclusion or protection between the time pipe_max_size is assigned
      a raw value from proc_dointvec_minmax() and when it is rounded.

  4 - unsigned long -> unsigned int conversion makes for potential odd
      return errors from do_proc_douintvec_minmax_conv() and
      do_proc_dopipe_max_size_conv().

This version underwent the same testing as v1:
https://marc.info/?l=linux-kernel&m=150643571406022&w=2

This patch (of 4):

pipe_max_size is defined as an unsigned int:

  unsigned int pipe_max_size = 1048576;

but its procfs/sysctl representation is an integer:

  static struct ctl_table fs_table[] = {
          ...
          {
                  .procname       = "pipe-max-size",
                  .data           = &pipe_max_size,
                  .maxlen         = sizeof(int),
                  .mode           = 0644,
                  .proc_handler   = &pipe_proc_fn,
                  .extra1         = &pipe_min_size,
          },
          ...

that is signed:

  int pipe_proc_fn(struct ctl_table *table, int write, void __user *buf,
                   size_t *lenp, loff_t *ppos)
  {
          ...
          ret = proc_dointvec_minmax(table, write, buf, lenp, ppos)

This leads to signed results via procfs for large values of pipe_max_size:

  % echo 2147483647 >/proc/sys/fs/pipe-max-size
  % cat /proc/sys/fs/pipe-max-size
  -2147483648

Use unsigned operations on this variable to avoid such negative values.

Link: http://lkml.kernel.org/r/1507658689-11669-2-git-send-email-joe.lawrence@redhat.com
Signed-off-by: Joe Lawrence <joe.lawrence@redhat.com>
Reported-by: Mikulas Patocka <mpatocka@redhat.com>
Reviewed-by: Mikulas Patocka <mpatocka@redhat.com>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

autofs: don't fail mount for transient error

Currently if the autofs kernel module gets an error when writing to the
pipe which links to the daemon, then it marks the whole moutpoint as
catatonic, and it will stop working.

It is possible that the error is transient.  This can happen if the
daemon is slow and more than 16 requests queue up.  If a subsequent
process tries to queue a request, and is then signalled, the write to
the pipe will return -ERESTARTSYS and autofs will take that as total
failure.

So change the code to assess -ERESTARTSYS and -ENOMEM as transient
failures which only abort the current request, not the whole mountpoint.

It isn't a crash or a data corruption, but having autofs mountpoints
suddenly stop working is rather inconvenient.

Ian said:

: And given the problems with a half dozen (or so) user space applications
: consuming large amounts of CPU under heavy mount and umount activity this
: could happen more easily than we expect.

Link: http://lkml.kernel.org/r/87y3norvgp.fsf@notabene.neil.brown.name
Signed-off-by: NeilBrown <neilb@suse.com>
Acked-by: Ian Kent <raven@themaw.net>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

init/version.c: include <linux/export.h> instead of <linux/module.h>

init/version.c has nothing to do with modules, so remove the
<linux/modude.h>.

Instead, include <linux/export.h> for EXPORT_SYMBOL_GPL.

This cuts off a lot of unnecessary header parsing.

Link: http://lkml.kernel.org/r/1505920984-8523-1-git-send-email-yamada.masahiro@socionext.com
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

epoll: remove ep_call_nested() from ep_eventpoll_poll()

The use of ep_call_nested() in ep_eventpoll_poll(), which is the .poll
routine for an epoll fd, is used to prevent excessively deep epoll
nesting, and to prevent circular paths.

However, we are already preventing these conditions during
EPOLL_CTL_ADD.  In terms of too deep epoll chains, we do in fact allow
deep nesting of the epoll fds themselves (deeper than EP_MAX_NESTS),
however we don't allow more than EP_MAX_NESTS when an epoll file
descriptor is actually connected to a wakeup source.  Thus, we do not
require the use of ep_call_nested(), since ep_eventpoll_poll(), which is
called via ep_scan_ready_list() only continues nesting if there are
events available.

Since ep_call_nested() is implemented using a global lock, applications
that make use of nested epoll can see large performance improvements
with this change.

Davidlohr said:

: Improvements are quite obscene actually, such as for the following
: epoll_wait() benchmark with 2 level nesting on a 80 core IvyBridge:
:
: ncpus  vanilla     dirty     delta
: 1      2447092     3028315   +23.75%
: 4      231265      2986954   +1191.57%
: 8      121631      2898796   +2283.27%
: 16     59749       2902056   +4757.07%
: 32     26837      2326314   +8568.30%
: 64     12926       1341281   +10276.61%
:
: (http://linux-scalability.org/epoll/epoll-test.c)

Link: http://lkml.kernel.org/r/1509430214-5599-1-git-send-email-jbaron@akamai.com
Signed-off-by: Jason Baron <jbaron@akamai.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Salman Qazi <sqazi@google.com>
Cc: Hou Tao <houtao1@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

epoll: avoid calling ep_call_nested() from ep_poll_safewake()

ep_poll_safewake() is used to wakeup potentially nested epoll file
descriptors.  The function uses ep_call_nested() to prevent entering the
same wake up queue more than once, and to prevent excessively deep
wakeup paths (deeper than EP_MAX_NESTS).  However, this is not necessary
since we are already preventing these conditions during EPOLL_CTL_ADD.
This saves extra function calls, and avoids taking a global lock during
the ep_call_nested() calls.

I have, however, left ep_call_nested() for the CONFIG_DEBUG_LOCK_ALLOC
case, since ep_call_nested() keeps track of the nesting level, and this
is required by the call to spin_lock_irqsave_nested().  It would be nice
to remove the ep_call_nested() calls for the CONFIG_DEBUG_LOCK_ALLOC
case as well, however its not clear how to simply pass the nesting level
through multiple wake_up() levels without more surgery.  In any case, I
don't think CONFIG_DEBUG_LOCK_ALLOC is generally used for production.
This patch, also apparently fixes a workload at Google that Salman Qazi
reported by completely removing the poll_safewake_ncalls->lock from
wakeup paths.

Link: http://lkml.kernel.org/r/1507920533-8812-1-git-send-email-jbaron@akamai.com
Signed-off-by: Jason Baron <jbaron@akamai.com>
Acked-by: Davidlohr Bueso <dbueso@suse.de>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Salman Qazi <sqazi@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

epoll: account epitem and eppoll_entry to kmemcg

A userspace application can directly trigger the allocations from
eventpoll_epi and eventpoll_pwq slabs.  A buggy or malicious application
can consume a significant amount of system memory by triggering such
allocations.  Indeed we have seen in production where a buggy
application was leaking the epoll references and causing a burst of
eventpoll_epi and eventpoll_pwq slab allocations.  This patch opt-in the
charging of eventpoll_epi and eventpoll_pwq slabs.

There is a per-user limit (~4% of total memory if no highmem) on these
caches.  I think it is too generous particularly in the scenario where
jobs of multiple users are running on the system and the administrator
is reducing cost by overcomitting the memory.  This is unaccounted
kernel memory and will not be considered by the oom-killer.  I think by
accounting it to kmemcg, for systems with kmem accounting enabled, we
can provide better isolation between jobs of different users.

Link: http://lkml.kernel.org/r/20171003021519.23907-1-shakeelb@google.com
Signed-off-by: Shakeel Butt <shakeelb@google.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Greg Thelen <gthelen@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

checkpatch: do not check missing blank line before builtin_*_driver

checkpatch.pl does not check missing blank line before module_*_driver.
I want it to behave likewise for builtin_*_driver.

Link: http://lkml.kernel.org/r/1505700081-12854-1-git-send-email-yamada.masahiro@socionext.com
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Acked-by: Joe Perches <joe@perches.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

checkpatch: add --strict test for lines ending in [ or (

Lines that end in an open bracket or open parenthesis are generally hard
to follow.  Lines following those ending with open parenthesis are also
rarely aligned to that open parenthesis.

Suggest not ending lines with '[' or '('

Link: http://lkml.kernel.org/r/8fd0b2b4a7482064254e37931eb9302a81d5aa2f.1508340786.git.joe@perches.com
Signed-off-by: Joe Perches <joe@perches.com>
Suggested-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

checkpatch: add TP_printk to list of logging functions

So the line length check can be bypassed by its callers.

Link: http://lkml.kernel.org/r/7de542c08a6e79f2ebe7c1416c9f403c23fdcc09.1508282823.git.joe@perches.com
Signed-off-by: Joe Perches <joe@perches.com>
Reported-by: Song Liu <songliubraving@fb.com>
Tested-by: Song Liu <songliubraving@fb.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

checkpatch: allow DEFINE_PER_CPU definitions to exceed line length

Some of the definitions are very long and can't be split into multiple
lines because ctags is limited.

Exempt these lines from the line length checks.

See commit 25528213fe9f ("tags: Fix DEFINE_PER_CPU expansions") for more
details.

Link: http://lkml.kernel.org/r/1508170320.6530.15.camel@perches.com
Signed-off-by: Joe Perches <joe@perches.com>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Cc: Marc Zyngier <marc.zyngier@arm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

checkpatch: printks always need a KERN_<LEVEL>

There was code in checkpatch that allowed continuation printks to be
used without KERN_CONT.  Remove the continuation check and always
require a KERN_<LEVEL>.

Link: http://lkml.kernel.org/r/61980ef41d5b9b6543da1c49055042e0ab74d308.1507047008.git.joe@perches.com
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

scripts/checkpatch.pl: avoid false warning missing break

void foo(int a)
switch (a) {
case 'h':
fun1();
exit(1);
default:
}

creates a warning "Possible switch case/default not preceded by break or
fallthrough comment".

exit( should be treated like return.

Link: http://lkml.kernel.org/r/20170910154618.25819-1-xypron.glpk@gmx.de
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Joe Perches <joe@perches.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

checkpatch: support function pointers for unnamed function definition arguments

Current unnamed function definition argument does not include function
pointer cases and it reports something like:

  WARNING: function definition argument 'void' should also have an identifier name
  +unsigned int (*dummy)(void);

Support function pointers for unnamed function arguments

Link: http://lkml.kernel.org/r/1505389925-31087-1-git-send-email-miles.chen@mediatek.com
Signed-off-by: Miles Chen <miles.chen@mediatek.com>
Acked-by: Joe Perches <joe@perches.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

lib: test module for find_*_bit() functions

find_bit functions are widely used in the kernel, including hot paths.
This module tests performance of those functions in 2 typical scenarios:
randomly filled bitmap with relatively equal distribution of set and
cleared bits, and sparse bitmap which has 1 set bit for 500 cleared
bits.

On ThunderX machine:

 Start testing find_bit() with random-filled bitmap
find_next_bit:          240043 cycles,  164062 iterations
find_next_zero_bit:     312848 cycles,  163619 iterations
find_last_bit:          193748 cycles,  164062 iterations
find_first_bit:      177720874 cycles,  164062 iterations

 Start testing find_bit() with sparse bitmap
find_next_bit:            3633 cycles,     656 iterations
find_next_zero_bit:     620399 cycles,  327025 iterations
find_last_bit:            3038 cycles,     656 iterations
find_first_bit:         691407 cycles,     656 iterations

[arnd@arndb.de: use correct format string for find-bit tests]
Link: http://lkml.kernel.org/r/20171113135605.3166307-1-arnd@arndb.de
Link: http://lkml.kernel.org/r/20171109140714.13168-1-ynorov@caviumnetworks.com
Signed-off-by: Yury Norov <ynorov@caviumnetworks.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Clement Courbet <courbet@google.com>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Matthew Wilcox <mawilcox@microsoft.com>
Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

lib/rbtree-test: lower default params

Fengguang reported soft lockups while running the rbtree and interval
tree test modules.  The logic for these tests all occur in init phase,
and we currently are pounding with the default values for number of
nodes and number of iterations of each test.  Reduce the latter by two
orders of magnitude.  This does not influence the value of the tests in
that one thousand times by default is enough to get the picture.

Link: http://lkml.kernel.org/r/20171109161715.xai2dtwqw2frhkcm@linux-n805
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

tools/lib/traceevent/parse-filter.c: clean up clang build warning

The uniform structure filter_arg sets its union based on the difference
of enum filter_arg_type, However, some functions use implicit type
conversion obviously.

warning: implicit conversion from enumeration type 'enum filter_exp_type'
 to different enumeration type 'enum filter_op_type'

warning: implicit conversion from enumeration type 'enum filter_cmp_type'
 to different enumeration type 'enum filter_exp_type'

Link: http://lkml.kernel.org/r/1509938415-113825-1-git-send-email-cj.chengjian@huawei.com
Signed-off-by: Cheng Jian <cj.chengjian@huawei.com>
Cc: Kate Stewart <kstewart@linuxfoundation.org>
Cc: Xie XiuQi <xiexiuqi@huawei.com>
Cc: Li Bin <huawei.libin@huawei.com>
Cc: Steven Rostedt (Red Hat) <rostedt@goodmis.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

lib/nmi_backtrace.c: fix kernel text address leak

Don't leak idle function address in NMI backtrace.

Link: http://lkml.kernel.org/r/20171106165648.GA95243@sofia
Signed-off-by: Liu Changcheng <changcheng.liu@intel.com>
Reviewed-by: Petr Mladek <pmladek@suse.com>
Reviewed-by: Josh Poimboeuf <jpoimboe@redhat.com>
Reviewed-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

lib/genalloc.c: make the avail variable an atomic_long_t

If the amount of resources allocated to a gen_pool exceeds 2^32 then the
avail atomic overflows and this causes problems when clients try and
borrow resources from the pool.  This is only expected to be an issue on
64 bit systems.

Add the <linux/atomic.h> header to pull in atomic_long* operations.  So
that 32 bit systems continue to use atomic32_t but 64 bit systems can
use atomic64_t.

Link: http://lkml.kernel.org/r/1509033843-25667-1-git-send-email-sbates@raithlin.com
Signed-off-by: Stephen Bates <sbates@raithlin.com>
Reviewed-by: Logan Gunthorpe <logang@deltatee.com>
Reviewed-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Reviewed-by: Daniel Mentz <danielmentz@google.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

lib/int_sqrt: adjust comments

Our current int_sqrt() is not rough nor any approximation; it calculates
the exact value of: floor(sqrt()).  Document this.

Link: http://lkml.kernel.org/r/20171020164645.001652117@infradead.org
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Anshul Garg <aksgarg1989@gmail.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: David Miller <davem@davemloft.net>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Joe Perches <joe@perches.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Matthew Wilcox <mawilcox@microsoft.com>
Cc: Michael Davidson <md@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

lib/int_sqrt: optimize initial value compute

The initial value (@m) compute is:

m = 1UL << (BITS_PER_LONG - 2);
while (m > x)
m >>= 2;

Which is a linear search for the highest even bit smaller or equal to @x
We can implement this using a binary search using __fls() (or better when
its hardware implemented).

m = 1UL << (__fls(x) & ~1UL);

Especially for small values of @x; which are the more common arguments
when doing a CDF on idle times; the linear search is near to worst case,
while the binary search of __fls() is a constant 6 (or 5 on 32bit)
branches.

      cycles:                 branches:              branch-misses:

PRE:

hot:   43.633557 +- 0.034373  45.333132 +- 0.002277  0.023529 +- 0.000681
cold: 207.438411 +- 0.125840  45.333132 +- 0.002277  6.976486 +- 0.004219

SOFTWARE FLS:

hot:   29.576176 +- 0.028850  26.666730 +- 0.004511  0.019463 +- 0.000663
cold: 165.947136 +- 0.188406  26.666746 +- 0.004511  6.133897 +- 0.004386

HARDWARE FLS:

hot:   24.720922 +- 0.025161  20.666784 +- 0.004509  0.020836 +- 0.000677
cold: 132.777197 +- 0.127471  20.666776 +- 0.004509  5.080285 +- 0.003874

Averages computed over all values <128k using a LFSR to generate order.
Cold numbers have a LFSR based branch trace buffer 'confuser' ran between
each int_sqrt() invocation.

Link: http://lkml.kernel.org/r/20171020164644.936577234@infradead.org
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Suggested-by: Joe Perches <joe@perches.com>
Acked-by: Will Deacon <will.deacon@arm.com>
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Anshul Garg <aksgarg1989@gmail.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: David Miller <davem@davemloft.net>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Matthew Wilcox <mawilcox@microsoft.com>
Cc: Michael Davidson <md@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

lib/int_sqrt: optimize small argument

The current int_sqrt() computation is sub-optimal for the case of small
@x.  Which is the interesting case when we're going to do cumulative
distribution functions on idle times, which we assume to be a random
variable, where the target residency of the deepest idle state gives an
upper bound on the variable (5e6ns on recent Intel chips).

In the case of small @x, the compute loop:

while (m != 0) {
b = y + m;
y >>= 1;

if (x >= b) {
x -= b;
y += m;
}
m >>= 2;
}

can be reduced to:

while (m > x)
m >>= 2;

Because y==0, b==m and until x>=m y will remain 0.

And while this is computationally equivalent, it runs much faster
because there's less code, in particular less branches.

      cycles:                 branches:              branch-misses:

OLD:

hot:   45.109444 +- 0.044117  44.333392 +- 0.002254  0.018723 +- 0.000593
cold: 187.737379 +- 0.156678  44.333407 +- 0.002254  6.272844 +- 0.004305

PRE:

hot:   67.937492 +- 0.064124  66.999535 +- 0.000488  0.066720 +- 0.001113
cold: 232.004379 +- 0.332811  66.999527 +- 0.000488  6.914634 +- 0.006568

POST:

hot:   43.633557 +- 0.034373  45.333132 +- 0.002277  0.023529 +- 0.000681
cold: 207.438411 +- 0.125840  45.333132 +- 0.002277  6.976486 +- 0.004219

Averages computed over all values <128k using a LFSR to generate order.
Cold numbers have a LFSR based branch trace buffer 'confuser' ran between
each int_sqrt() invocation.

Link: http://lkml.kernel.org/r/20171020164644.876503355@infradead.org
Fixes: 30493cc9dddb ("lib/int_sqrt.c: optimize square root algorithm")
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Suggested-by: Anshul Garg <aksgarg1989@gmail.com>
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Joe Perches <joe@perches.com>
Cc: David Miller <davem@davemloft.net>
Cc: Matthew Wilcox <mawilcox@microsoft.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Michael Davidson <md@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

lib/test: delete five error messages for failed memory allocations

Omit extra messages for a memory allocation failure in these functions.

This issue was detected by using the Coccinelle software.

Link: http://lkml.kernel.org/r/410a4c5a-4ee0-6fcc-969c-103d8e496b78@users.sourceforge.net
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Acked-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

lib: add module support to string tests

Extract the string test code into its own source file, to allow
compiling it either to a loadable module, or built into the kernel.

Fixes: 03270c13c5ffaa6a ("lib/string.c: add testcases for memset16/32/64")
Link: http://lkml.kernel.org/r/1505397744-3387-1-git-send-email-geert@linux-m68k.org
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Matthew Wilcox <mawilcox@microsoft.com>
Cc: Shuah Khan <shuah@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

include/linux/radix-tree.h: remove unneeded #include <linux/bug.h>

This include was added by commit 187f1882b5b0 ("BUG: headers with
BUG/BUG_ON etc. need linux/bug.h") because BUG_ON() was used in this
header at that time.

Some time later, commit 6d75f366b924 ("lib: radix-tree: check accounting
of existing slot replacement users") removed the use of BUG_ON() from
this header.

Since then, there is no reason to include <linux/bug.h>.

Link: http://lkml.kernel.org/r/1505660151-4383-1-git-send-email-yamada.masahiro@socionext.com
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Matthew Wilcox <mawilcox@microsoft.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Jan Kara <jack@suse.cz>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Chris Mi <chrism@mellanox.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

include/linux/bitfield.h: include <linux/build_bug.h> instead of <linux/bug.h>

Since commit bc6245e5efd7 ("bug: split BUILD_BUG stuff out into
<linux/build_bug.h>"), #include <linux/build_bug.h> is better to pull
minimal headers needed for BUILG_BUG() family.

Link: http://lkml.kernel.org/r/1505700775-19826-1-git-send-email-yamada.masahiro@socionext.com
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Cc: Dinan Gunawardena <dinan.gunawardena@netronome.com>
Cc: Kalle Valo <kvalo@codeaurora.org>
Cc: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>