virtio_net XDP support expects receive buffers to be contiguous.
If this is not the case we enable a slowpath to allow connectivity
to continue but at a significan performance overhead associated with
linearizing data. To make it painfully aware to users that XDP is
running in a degraded mode we throw an xdp buffer error.

To linearize packets we allocate a page and copy the segments of
the data, including the header, into it. After this the page can be
handled by XDP code flow as normal.

Then depending on the return code the page is either freed or sent
to the XDP xmit path. There is no attempt to optimize this path.

This case is being handled simple as a precaution in case some
unknown backend were to generate packets in this form. To test this
I had to hack qemu and force it to generate these packets. I do not
expect this case to be generated by "real" backends.
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This adds support for the XDP_TX action to virtio_net. When an XDP
program is run and returns the XDP_TX action the virtio_net XDP
implementation will transmit the packet on a TX queue that aligns
with the current CPU that the XDP packet was processed on.

Before sending the packet the header is zeroed.  Also XDP is expected
to handle checksum correctly so no checksum offload  support is
provided.
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

XDP requires using isolated transmit queues to avoid interference
with normal networking stack (BQL, NETDEV_TX_BUSY, etc). This patch
adds a XDP queue per cpu when a XDP program is loaded and does not
expose the queues to the OS via the normal API call to
netif_set_real_num_tx_queues(). This way the stack will never push
an skb to these queues.

However virtio/vhost/qemu implementation only allows for creating
TX/RX queue pairs at this time so creating only TX queues was not
possible. And because the associated RX queues are being created I
went ahead and exposed these to the stack and let the backend use
them. This creates more RX queues visible to the network stack than
TX queues which is worth mentioning but does not cause any issues as
far as I can tell.
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This adds XDP support to virtio_net. Some requirements must be
met for XDP to be enabled depending on the mode. First it will
only be supported with LRO disabled so that data is not pushed
across multiple buffers. Second the MTU must be less than a page
size to avoid having to handle XDP across multiple pages.

If mergeable receive is enabled this patch only supports the case
where header and data are in the same buf which we can check when
a packet is received by looking at num_buf. If the num_buf is
greater than 1 and a XDP program is loaded the packet is dropped
and a warning is thrown. When any_header_sg is set this does not
happen and both header and data is put in a single buffer as expected
so we check this when XDP programs are loaded.  Subsequent patches
will process the packet in a degraded mode to ensure connectivity
and correctness is not lost even if backend pushes packets into
multiple buffers.

If big packets mode is enabled and MTU/LRO conditions above are
met then XDP is allowed.

This patch was tested with qemu with vhost=on and vhost=off where
mergeable and big_packet modes were forced via hard coding feature
negotiation. Multiple buffers per packet was forced via a small
test patch to vhost.c in the vhost=on qemu mode.
Suggested-by: Shrijeet Mukherjee <shrijeet@gmail.com>
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This adds a warning for drivers to use when encountering an invalid
buffer for XDP. For normal cases this should not happen but to catch
this in virtual/qemu setups that I may not have expected from the
emulation layer having a standard warning is useful.
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Prior to this patch, sctp_transport_lookup_process didn't rcu_read_unlock
when it failed to find a transport by sctp_addrs_lookup_transport.

This patch is to fix it by moving up rcu_read_unlock right before checking
transport and also to remove the out path.

Fixes: 1cceda78 ("sctp: fix the issue sctp_diag uses lock_sock in rcu_read_lock")
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>

Since commit 7fda702f ("sctp: use new rhlist interface on sctp transport
rhashtable"), sctp has changed to use rhlist_lookup to look up transport, but
rhlist_lookup doesn't call rcu_read_lock inside, unlike rhashtable_lookup_fast.

It is called in sctp_epaddr_lookup_transport and sctp_addrs_lookup_transport.
sctp_addrs_lookup_transport is always in the protection of rcu_read_lock(),
as __sctp_lookup_association is called in rx path or sctp_lookup_association
which are in the protection of rcu_read_lock() already.

But sctp_epaddr_lookup_transport is called by sctp_endpoint_lookup_assoc, it
doesn't call rcu_read_lock, which may cause "suspicious rcu_dereference_check
usage' in __rhashtable_lookup.

This patch is to fix it by adding rcu_read_lock in sctp_endpoint_lookup_assoc
before calling sctp_epaddr_lookup_transport.

Fixes: 7fda702f ("sctp: use new rhlist interface on sctp transport rhashtable")
Reported-by: Dmitry Vyukov <dvyukov@google.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>

Madalin Bucur says:

====================
dpaa_eth: a couple of fixes

This patch set introduces big endian accessors in the dpaa_eth driver
making sure accesses to the QBMan HW are correct on little endian
platforms. Removing a redundant Kconfig dependency on FSL_SOC.
Adding myself as maintainer of the dpaa_eth driver.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Add record for Freescale QORIQ DPAA Ethernet driver adding myself as
maintainer.
Signed-off-by: Madalin Bucur <madalin.bucur@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Madalin Bucur <madalin.bucur@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Ensure correct access to the big endian QMan HW through proper
accessors.
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Signed-off-by: Madalin Bucur <madalin.bucur@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Since the struct miscdevice have many members, it is dangerous to init
it without members name relying only on member order.

This patch add member name to the init declaration.
Signed-off-by: Corentin Labbe <clabbe.montjoie@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The IRNET_MAJOR define is not used, so this patch remove it.
Signed-off-by: Corentin Labbe <clabbe.montjoie@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch move the define for IRNET_MINOR to include/linux/miscdevice.h
It is better that all minor number definitions are in the same place.
Signed-off-by: Corentin Labbe <clabbe.montjoie@gmail.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

The only use of miscdevice is irda_ppp so no need to include
linux/miscdevice.h for all irda files.
This patch move the linux/miscdevice.h include to irnet_ppp.h
Signed-off-by: Corentin Labbe <clabbe.montjoie@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

irproc.c does not use any miscdevice so this patch remove this
unnecessary inclusion.
Signed-off-by: Corentin Labbe <clabbe.montjoie@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The member 'effective' in 'struct cgroup_bpf' is protected by RCU.
Annotate it accordingly to squelch a sparse warning.
Signed-off-by: Daniel Mack <daniel@zonque.org>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Tom Herbert says:

====================
inet: Fixes for inet_csk_get_port and soreusport

This patch set fixes a couple of issues I noticed while debugging our
softlockup issue in inet_csk_get_port.

- Don't allow jump into port scan in inet_csk_get_port if function
  was called with non-zero port number (looking up explicit port
  number).
- When inet_csk_get_port is called with zero port number (ie. perform
  scan) an reuseport is set on the socket, don't match sockets that
  also have reuseport set. The intent from the user should be
  to get a new port number and then explictly bind other
  sockets to that number using soreuseport.

Tested:

Ran first patch on production workload with no ill effect.

For second patch, ran a little listener application and first
demonstrated that unbound sockets with soreuseport can indeed
be bound to unrelated soreuseport sockets.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

A user may call listen with binding an explicit port with the intent
that the kernel will assign an available port to the socket. In this
case inet_csk_get_port does a port scan. For such sockets, the user may
also set soreuseport with the intent a creating more sockets for the
port that is selected. The problem is that the initial socket being
opened could inadvertently choose an existing and unreleated port
number that was already created with soreuseport.

This patch adds a boolean parameter to inet_bind_conflict that indicates
rather soreuseport is allowed for the check (in addition to
sk->sk_reuseport). In calls to inet_bind_conflict from inet_csk_get_port
the argument is set to true if an explicit port is being looked up (snum
argument is nonzero), and is false if port scan is done.
Signed-off-by: Tom Herbert <tom@herbertland.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

inet_csk_get_port is called with port number (snum argument) that may be
zero or nonzero. If it is zero, then the intent is to find an available
ephemeral port number to bind to. If snum is non-zero then the caller
is asking to allocate a specific port number. In the latter case we
never want to perform the scan in ephemeral port range. It is
conceivable that this can happen if the "goto again" in "tb_found:"
is done. This patch adds a check that snum is zero before doing
the "goto again".
Signed-off-by: Tom Herbert <tom@herbertland.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Running ./test_verifier as unprivileged lets 1 out of 98 tests fail:

  [...]
  #71 unpriv: check that printk is disallowed FAIL
  Unexpected error message!
  0: (7a) *(u64 *)(r10 -8) = 0
  1: (bf) r1 = r10
  2: (07) r1 += -8
  3: (b7) r2 = 8
  4: (bf) r3 = r1
  5: (85) call bpf_trace_printk#6
  unknown func bpf_trace_printk#6
  [...]

The test case is correct, just that the error outcome changed with
ebb676da ("bpf: Print function name in addition to function id").
Same as with e00c7b21 ("bpf: fix multiple issues in selftest suite
and samples") issue 2), so just fix up the function name.

Fixes: ebb676da ("bpf: Print function name in addition to function id")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Commit 57a09bf0 ("bpf: Detect identical PTR_TO_MAP_VALUE_OR_NULL
registers") introduced a regression where existing programs stopped
loading due to reaching the verifier's maximum complexity limit,
whereas prior to this commit they were loading just fine; the affected
program has roughly 2k instructions.

What was found is that state pruning couldn't be performed effectively
anymore due to mismatches of the verifier's register state, in particular
in the id tracking. It doesn't mean that 57a09bf0 is incorrect per
se, but rather that verifier needs to perform a lot more work for the
same program with regards to involved map lookups.

Since commit 57a09bf0 is only about tracking registers with type
PTR_TO_MAP_VALUE_OR_NULL, the id is only needed to follow registers
until they are promoted through pattern matching with a NULL check to
either PTR_TO_MAP_VALUE or UNKNOWN_VALUE type. After that point, the
id becomes irrelevant for the transitioned types.

For UNKNOWN_VALUE, id is already reset to 0 via mark_reg_unknown_value(),
but not so for PTR_TO_MAP_VALUE where id is becoming stale. It's even
transferred further into other types that don't make use of it. Among
others, one example is where UNKNOWN_VALUE is set on function call
return with RET_INTEGER return type.

states_equal() will then fall through the memcmp() on register state;
note that the second memcmp() uses offsetofend(), so the id is part of
that since d2a4dd37 ("bpf: fix state equivalence"). But the bisect
pointed already to 57a09bf0, where we really reach beyond complexity
limit. What I found was that states_equal() often failed in this
case due to id mismatches in spilled regs with registers in type
PTR_TO_MAP_VALUE. Unlike non-spilled regs, spilled regs just perform
a memcmp() on their reg state and don't have any other optimizations
in place, therefore also id was relevant in this case for making a
pruning decision.

We can safely reset id to 0 as well when converting to PTR_TO_MAP_VALUE.
For the affected program, it resulted in a ~17 fold reduction of
complexity and let the program load fine again. Selftest suite also
runs fine. The only other place where env->id_gen is used currently is
through direct packet access, but for these cases id is long living, thus
a different scenario.

Also, the current logic in mark_map_regs() is not fully correct when
marking NULL branch with UNKNOWN_VALUE. We need to cache the destination
reg's id in any case. Otherwise, once we marked that reg as UNKNOWN_VALUE,
it's id is reset and any subsequent registers that hold the original id
and are of type PTR_TO_MAP_VALUE_OR_NULL won't be marked UNKNOWN_VALUE
anymore, since mark_map_reg() reuses the uncached regs[regno].id that
was just overridden. Note, we don't need to cache it outside of
mark_map_regs(), since it's called once on this_branch and the other
time on other_branch, which are both two independent verifier states.
A test case for this is added here, too.

Fixes: 57a09bf0 ("bpf: Detect identical PTR_TO_MAP_VALUE_OR_NULL registers")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Thomas Graf <tgraf@suug.ch>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Locally originated traffic in a VRF fails in the presence of a POSTROUTING
rule. For example,

    $ iptables -t nat -A POSTROUTING -s 11.1.1.0/24  -j MASQUERADE
    $ ping -I red -c1 11.1.1.3
    ping: Warning: source address might be selected on device other than red.
    PING 11.1.1.3 (11.1.1.3) from 11.1.1.2 red: 56(84) bytes of data.
    ping: sendmsg: Operation not permitted

Worse, the above causes random corruption resulting in a panic in random
places (I have not seen a consistent backtrace).

Call nf_reset to drop the conntrack info following the pass through the
VRF device.  The nf_reset is needed on Tx but not Rx because of the order
in which NF_HOOK's are hit: on Rx the VRF device is after the real ingress
device and on Tx it is is before the real egress device. Connection
tracking should be tied to the real egress device and not the VRF device.

Fixes: 8f58336d ("net: Add ethernet header for pass through VRF device")
Fixes: 35402e31 ("net: Add IPv6 support to VRF device")
Signed-off-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Connection tracking with VRF is broken because the pass through the VRF
device drops the connection tracking info. Removing the call to nf_reset
allows DNAT and MASQUERADE to work across interfaces within a VRF.

Fixes: 73e20b76 ("net: vrf: Add support for PREROUTING rules on vrf device")
Signed-off-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Paul Blakey says:

====================
net/sched: cls_flower: Fix mask handling

The series fix how the mask is being handled.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Zero bits on the mask signify a "don't care" on the corresponding bits
in key. Some HWs require those bits on the key to be zero. Since these
bits are masked anyway, it's okay to provide the masked key to all
drivers.

Fixes: 5b33f488 ('net/flower: Introduce hardware offload support')
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

When addr_type is set, mask should also be set.

Fixes: 66530bdf ('sched,cls_flower: set key address type when present')
Fixes: bc3103f1 ('net/sched: cls_flower: Classify packet in ip tunnels')
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

There are hardware PCI implementations of Cadence GEM network
controller. This patch will allow to use such hardware with reuse of
existing Platform Driver.
Signed-off-by: Bartosz Folta <bfolta@cadence.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Include calculations to compute the number of segments
that comprise an aggregated large packet.
Signed-off-by: Thomas Falcon <tlfalcon@linux.vnet.ibm.com>
Reviewed-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Reviewed-by: Jonathan Maxwell <jmaxwell37@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

On ACPI systems, clocks are not available to drivers directly.  They are
handled exclusively by ACPI and/or firmware, so there is no clock driver.
Calls to clk_get() always fail, so we should not even attempt to claim
any clocks on ACPI systems.
Signed-off-by: Timur Tabi <timur@codeaurora.org>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Jeroen De Wachter <jeroen.de_wachter.ext@nokia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Before, encx24j600_rx_packets did not update encx24j600_priv's next_packet
member when an error occurred during packet handling (either because the
packet's RSV header indicates an error or because the encx24j600_receive_packet
method can't allocate an sk_buff).

If the next_packet member is not updated, the ERXTAIL register will be set to
the same value it had before, which means the bad packet remains in the
component's memory and its RSV header will be read again when a new packet
arrives. If the RSV header indicates a bad packet or if sk_buff allocation
continues to fail, new packets will be stored in the component's memory until
that memory is full, after which packets will be dropped.

The SETPKTDEC command is always executed though, so the encx24j600 hardware has
an incorrect count of the packets in its memory.

To prevent this, the next_packet member should always be updated, allowing the
packet to be skipped (either because it's bad, as indicated in its RSV header,
or because allocating an sk_buff failed). In the allocation failure case, this
does mean dropping a valid packet, but dropping the oldest packet to keep as
much memory as possible available for new packets seems preferable to keeping
old (but valid) packets around while dropping new ones.
Signed-off-by: Jeroen De Wachter <jeroen.de_wachter.ext@nokia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Dongpo Li says:

====================
net: ethernet: hisilicon: set dev->dev.parent before PHY connect

This patch series builds atop:
ec988ad7 ("phy: Don't increment MDIO bus
refcount unless it's a different owner")

I have checked all the hisilicon ethernet driver and found only two drivers
need to be fixed to make sure set dev->dev.parent before PHY connect.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

The hip04 driver calls into PHYLIB which now checks for
net_device->dev.parent, so make sure we do set it before calling into
any MDIO/PHYLIB related function.

Fixes: ec988ad7 ("phy: Don't increment MDIO bus refcount unless it's a different owner")
Signed-off-by: Dongpo Li <lidongpo@hisilicon.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The hisi_femac driver calls into PHYLIB which now checks for
net_device->dev.parent, so make sure we do set it before calling into
any MDIO/PHYLIB related function.

Fixes: ec988ad7 ("phy: Don't increment MDIO bus refcount unless it's a different owner")
Signed-off-by: Dongpo Li <lidongpo@hisilicon.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

A port is not necessarily assigned to a netdev. And a port does not
need to be a member of a bridge. So when iterating over all ports,
check before using the netdev and bridge_dev for a port. Otherwise we
dereference a NULL pointer.

Fixes: da9c359e ("net: dsa: mv88e6xxx: check hardware VLAN in use")
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The timer handling in this driver is broken in several ways:

- corkscrew_open() initializes and arms a timer before requesting the
  device interrupt. If the request fails the timer stays armed.

  A second call to corkscrew_open will unconditionally reinitialize the
  quued timer and arm it again. Also a immediate device removal will leave
  the timer queued because close() is not called (open() failed) and
  therefore nothing issues del_timer().

  The reinitialization corrupts the link chain in the timer wheel hash
  bucket and causes a NULL pointer dereference when the timer wheel tries
  to operate on that hash bucket. Immediate device removal lets the link
  chain poke into freed and possibly reused memory.

  Solution: Arm the timer after the successful irq request.

- corkscrew_close() uses del_timer()

  On close the timer is disarmed with del_timer() which lets the following
  code race against a concurrent timer expiry function.

  Solution: Use del_timer_sync() instead

- corkscrew_close() calls del_timer() unconditionally

  del_timer() is invoked even if the timer was never initialized. This
  works by chance because the struct containing the timer is zeroed at
  allocation time.

  Solution: Move the setup of the timer into corkscrew_setup().
Reported-by: Matthew Whitehead <tedheadster@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: netdev@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>

Commit 4f7df337
"netlink: 2-clause nla_ok()" is BROKEN.

First clause tests if "->nla_len" could even be accessed at all,
it can not possibly be omitted.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Commit 44900010 ("virtio-net: enable
multiqueue by default") blindly set the affinity instead of queues
during probe which can cause a mismatch of #queues between guest and
host. This patch fixes it by setting queues.
Reported-by: Theodore Ts'o <tytso@mit.edu>
Tested-by: Theodore Ts'o <tytso@mit.edu>
Cc: Neil Horman <nhorman@tuxdriver.com>
Cc: Michael S. Tsirkin <mst@redhat.com>
Fixes: 49000102901 ("virtio-net: enable multiqueue by default")
Signed-off-by: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Pull CRIS updates from Jesper Nilsson:
 "Three patches for minor issues"

* tag 'cris-for-4.10' of git://git.kernel.org/pub/scm/linux/kernel/git/jesper/cris:
  cris: No need to append -O2 and $(LINUXINCLUDE)
  tty: serial: make crisv10 explicitly non-modular
  cris: Only build flash rescue image if CONFIG_ETRAX_AXISFLASHMAP is selected