When xfrm is applied to TSO/GSO packets, it follows this path:

    xfrm_output() -> xfrm_output_gso() -> skb_gso_segment()

where skb_gso_segment() relies on skb->protocol to function properly.

This patch sets skb->protocol to ETH_P_IP before dst_output() is called,
fixing a bug where GSO packets sent through a sit tunnel are dropped
when xfrm is involved.

Cc: stable@vger.kernel.org
Signed-off-by: Eli Cooper <elicooper@gmx.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

When packet_set_ring creates a ring buffer it will initialize a
struct timer_list if the packet version is TPACKET_V3. This value
can then be raced by a different thread calling setsockopt to
set the version to TPACKET_V1 before packet_set_ring has finished.

This leads to a use-after-free on a function pointer in the
struct timer_list when the socket is closed as the previously
initialized timer will not be deleted.

The bug is fixed by taking lock_sock(sk) in packet_setsockopt when
changing the packet version while also taking the lock at the start
of packet_set_ring.

Fixes: f6fb8f10 ("af-packet: TPACKET_V3 flexible buffer implementation.")
Signed-off-by: Philip Pettersson <philip.pettersson@gmail.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The driver already uses its private lock for synchronization between xmit
and xmit completion handler making the additional use of the xmit_lock
unnecessary.
Furthermore the driver does not set NETIF_F_LLTX resulting in xmit to be
called with the xmit_lock held and then taking the private lock while xmit
completion handler does the reverse, first take the private lock, then the
xmit_lock.
Fix these issues by not taking the xmit_lock in the tx completion handler.
Signed-off-by: Lino Sanfilippo <LinoSanfilippo@gmx.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

An explicit dma sync for device directly after mapping as well as an
explicit dma sync for cpu directly before unmapping is unnecessary and
costly on the hotpath. So remove these calls.
Signed-off-by: Lino Sanfilippo <LinoSanfilippo@gmx.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

Johan Hovold says:

====================
net: stmmac: fix probe error handling and phydev leaks

This series fixes a number of issues with the stmmac-driver probe error
handling, which for example left clocks enabled after probe failures.

The final patch fixes a failure to deregister and free any fixed-link
PHYs that were registered during probe on probe errors and on driver
unbind. It also fixes a related of-node leak on late probe errors.

This series depends on the of_phy_deregister_fixed_link() helper that
was just merged to net.

As mentioned earlier, one staging driver also suffers from a similar
leak and can be fixed up once the above mentioned helper hits mainline.

Note that these patches have only been compile tested.
====================
Acked-by: Giuseppe Cavallaro <peppe.cavallaro@st.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Make sure to deregister and free any fixed-link phy registered during
probe on probe errors and on driver unbind by adding a new glue helper
function.

Drop the of-node reference taken in the same path also on late probe
errors (and not just on driver unbind) by moving the put from
stmmac_dvr_remove() to the new helper.

Fixes: 27732381 ("stmmac: add fixed-link device-tree support")
Fixes: 4613b279 ("ethernet: stmicro: stmmac: add missing of_node_put
after calling of_parse_phandle")
Signed-off-by: Johan Hovold <johan@kernel.org>
Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Fix the OF-helper function header to reflect that the function no longer
has a platform-data parameter.

Fixes: b0003ead ("stmmac: make stmmac_probe_config_dt return the
platform data struct")
Signed-off-by: Johan Hovold <johan@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Make sure to disable clocks before returning on late probe errors.

Fixes: 566e8251 ("net: stmmac: add a glue driver for the Amlogic
Meson 8b / GXBB DWMAC")
Signed-off-by: Johan Hovold <johan@kernel.org>
Acked-by: Kevin Hilman <khilman@baylibre.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Make sure to call any exit() callback to undo the effect of init()
before returning on late probe errors.

Fixes: cf3f047b ("stmmac: move hw init in the probe (v2)")
Signed-off-by: Johan Hovold <johan@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Make sure to disable runtime PM, power down the PHY, and disable clocks
before returning on late probe errors.

Fixes: 27ffefd2 ("stmmac: dwmac-rk: create a new probe function")
Signed-off-by: Johan Hovold <johan@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Make sure to disable clocks before returning on late probe errors.

Fixes: 8387ee21 ("stmmac: dwmac-sti: turn setup callback into a
probe function")
Signed-off-by: Johan Hovold <johan@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Make sure to call stmmac_dvr_remove() before returning on late probe
errors so that memory is freed, clocks are disabled, and the netdev is
deregistered before its resources go away.

Fixes: 3c201b5a ("net: stmmac: socfpga: Remove re-registration of
reset controller")
Signed-off-by: Johan Hovold <johan@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Use the correct attribute constant names IFLA_GSO_MAX_{SEGS,SIZE}
instead of IFLA_MAX_GSO_{SEGS,SIZE} for the comments int nlmsg_size().

Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

In the case of IPIP and SIT tunnel frames the outer transport header
offset is actually set to the same offset as the inner transport header.
This results in the lco_csum call not doing any checksum computation over
the inner IPv4/v6 header data.

In order to account for that I am updating the code so that we determine
the location to start the checksum ourselves based on the location of the
IPv4 header and the length.

Fixes: b83e3010 ("ixgbe/ixgbevf: Add support for GSO partial")
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

In the case of IPIP and SIT tunnel frames the outer transport header
offset is actually set to the same offset as the inner transport header.
This results in the lco_csum call not doing any checksum computation over
the inner IPv4/v6 header data.

In order to account for that I am updating the code so that we determine
the location to start the checksum ourselves based on the location of the
IPv4 header and the length.

Fixes: e10715d3 ("igb/igbvf: Add support for GSO partial")
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The change fixes AX88772_suspend() USB vendor commands failure issues.
Signed-off-by: Allan Chou <allan@asix.com.tw>
Tested-by: Allan Chou <allan@asix.com.tw>
Tested-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Steffen Klassert says:

====================
pull request (net): ipsec 2016-12-01

1) Change the error value when someone tries to run 32bit
   userspace on a 64bit host from -ENOTSUPP to the userspace
   exported -EOPNOTSUPP. Fix from Yi Zhao.

2) On inbound, ESN sequence numbers are already in network
   byte order. So don't try to convert it again, this fixes
   integrity verification for ESN. Fixes from Tobias Brunner.

Please pull or let me know if there are problems.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Pablo Neira Ayuso says:

====================
Netfilter fixes for net

This is a large batch of Netfilter fixes for net, they are:

1) Three patches to fix NAT conversion to rhashtable: Switch to rhlist
   structure that allows to have several objects with the same key.
   Moreover, fix wrong comparison logic in nf_nat_bysource_cmp() as this is
   expecting a return value similar to memcmp(). Change location of
   the nat_bysource field in the nf_conn structure to avoid zeroing
   this as it breaks interaction with SLAB_DESTROY_BY_RCU and lead us
   to crashes. From Florian Westphal.

2) Don't allow malformed fragments go through in IPv6, drop them,
   otherwise we hit GPF, patch from Florian Westphal.

3) Fix crash if attributes are missing in nft_range, from Liping Zhang.

4) Fix arptables 32-bits userspace 64-bits kernel compat, from Hongxu Jia.

5) Two patches from David Ahern to fix netfilter interaction with vrf.
   From David Ahern.

6) Fix element timeout calculation in nf_tables, we take milliseconds
   from userspace, but we use jiffies from kernelspace. Patch from
   Anders K.  Pedersen.

7) Missing validation length netlink attribute for nft_hash, from
   Laura Garcia.

8) Fix nf_conntrack_helper documentation, we don't default to off
   anymore for a bit of time so let's get this in sync with the code.

I know is late but I think these are important, specifically the NAT
bits, as they are mostly addressing fallout from recent changes. I also
read there are chances to have -rc8, if that is the case, that would
also give us a bit more time to test this.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

We trigger uarg->callback() immediately after we decide do datacopy
even if caller want to do zerocopy. This will cause the callback
(vhost_net_zerocopy_callback) decrease the refcount. But when we meet
an error afterwards, the error handling in vhost handle_tx() will try
to decrease it again. This is wrong and fix this by delay the
uarg->callback() until we're sure there's no errors.
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>

We trigger uarg->callback() immediately after we decide do datacopy
even if caller want to do zerocopy. This will cause the callback
(vhost_net_zerocopy_callback) decrease the refcount. But when we meet
an error afterwards, the error handling in vhost handle_tx() will try
to decrease it again. This is wrong and fix this by delay the
uarg->callback() until we're sure there's no errors.
Reported-by: wangyunjian <wangyunjian@huawei.com>
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>

netif_set_real_num_tx/rx_queues() are required to be called with rtnl_lock
taken, otherwise ASSERT_RTNL() warning will be triggered - which happens
now during System resume from suspend:
cpsw_resume()
|- cpsw_ndo_open()
  |- netif_set_real_num_tx/rx_queues()
     |- ASSERT_RTNL();

Hence, fix it by surrounding cpsw_ndo_open() by rtnl_lock/unlock() calls.

Cc: Dave Gerlach <d-gerlach@ti.com>
Cc: Ivan Khoronzhuk <ivan.khoronzhuk@linaro.org>
Fixes: commit e05107e6 ("net: ethernet: ti: cpsw: add multi queue support")
Signed-off-by: Grygorii Strashko <grygorii.strashko@ti.com>
Reviewed-by: Ivan Khoronzhuk <ivan.khoronzhuk@linaro.org>
Tested-by: Dave Gerlach <d-gerlach@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

If we have a branch that looks something like this

int foo = map->value;
if (condition) {
  foo += blah;
} else {
  foo = bar;
}
map->array[foo] = baz;

We will incorrectly assume that the !condition branch is equal to the condition
branch as the register for foo will be UNKNOWN_VALUE in both cases.  We need to
adjust this logic to only do this if we didn't do a varlen access after we
processed the !condition branch, otherwise we have different ranges and need to
check the other branch as well.

Fixes: 48461135 ("bpf: allow access into map value arrays")
Reported-by: Jann Horn <jannh@google.com>
Signed-off-by: Josef Bacik <jbacik@fb.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

Since 09d96860 ("netfilter: x_tables: do compat validation via
translate_table"), it used compatr structure to assign newinfo
structure.  In translate_compat_table of ip_tables.c and ip6_tables.c,
it used compatr->hook_entry to replace info->hook_entry and
compatr->underflow to replace info->underflow, but not do the same
replacement in arp_tables.c.

It caused invoking 32-bit "arptbale -P INPUT ACCEPT" failed in 64bit
kernel.
--------------------------------------
root@qemux86-64:~# arptables -P INPUT ACCEPT
root@qemux86-64:~# arptables -P INPUT ACCEPT
ERROR: Policy for `INPUT' offset 448 != underflow 0
arptables: Incompatible with this kernel
--------------------------------------

Fixes: 09d96860 ("netfilter: x_tables: do compat validation via translate_table")
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Acked-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>

Merge tag 'wireless-drivers-for-davem-2016-11-29' of git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers

Kalle Valo says:

====================
wireless-drivers fixes for 4.9

mwifiex

* properly terminate SSIDs so that uninitalised memory is not printed
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Guillaume Nault says:

====================
l2tp: fixes for l2tp_ip and l2tp_ip6 socket handling

This series addresses problems found while working on commit 32c23116
("l2tp: fix racy SOCK_ZAPPED flag check in l2tp_ip{,6}_bind()").

The first three patches fix races in socket's connect, recv and bind
operations. The last two ones fix scenarios where l2tp fails to
correctly lookup its userspace sockets.

Apart from the last patch, which is l2tp_ip6 specific, every patch
fixes the same problem in the L2TP IPv4 and IPv6 code.

All problems fixed by this series exist since the creation of the
l2tp_ip and l2tp_ip6 modules.

Changes since v1:
  * Patch #3: fix possible uninitialised use of 'ret' in l2tp_ip_bind().
====================
Acked-by: James Chapman <jchapman@katalix.com>

The '!(addr && ipv6_addr_equal(addr, laddr))' part of the conditional
matches if addr is NULL or if addr != laddr.
But the intend of __l2tp_ip6_bind_lookup() is to find a sockets with
the same address, so the ipv6_addr_equal() condition needs to be
inverted.

For better clarity and consistency with the rest of the expression, the
(!X || X == Y) notation is used instead of !(X && X != Y).
Signed-off-by: Guillaume Nault <g.nault@alphalink.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>

When looking up an l2tp socket, we must consider a null netdevice id as
wild card. There are currently two problems caused by
__l2tp_ip_bind_lookup() not considering 'dif' as wild card when set to 0:

  * A socket bound to a device (i.e. with sk->sk_bound_dev_if != 0)
    never receives any packet. Since __l2tp_ip_bind_lookup() is called
    with dif == 0 in l2tp_ip_recv(), sk->sk_bound_dev_if is always
    different from 'dif' so the socket doesn't match.

  * Two sockets, one bound to a device but not the other, can be bound
    to the same address. If the first socket binding to the address is
    the one that is also bound to a device, the second socket can bind
    to the same address without __l2tp_ip_bind_lookup() noticing the
    overlap.

To fix this issue, we need to consider that any null device index, be
it 'sk->sk_bound_dev_if' or 'dif', matches with any other value.
We also need to pass the input device index to __l2tp_ip_bind_lookup()
on reception so that sockets bound to a device never receive packets
from other devices.

This patch fixes l2tp_ip6 in the same way.
Signed-off-by: Guillaume Nault <g.nault@alphalink.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>

It's not enough to check for sockets bound to same address at the
beginning of l2tp_ip{,6}_bind(): even if no socket is found at that
time, a socket with the same address could be bound before we take
the l2tp lock again.

This patch moves the lookup right before inserting the new socket, so
that no change can ever happen to the list between address lookup and
socket insertion.

Care is taken to avoid side effects on the socket in case of failure.
That is, modifications of the socket are done after the lookup, when
binding is guaranteed to succeed, and before releasing the l2tp lock,
so that concurrent lookups will always see fully initialised sockets.

For l2tp_ip, 'ret' is set to -EINVAL before checking the SOCK_ZAPPED
bit. Error code was mistakenly set to -EADDRINUSE on error by commit
32c23116 ("l2tp: fix racy SOCK_ZAPPED flag check in l2tp_ip{,6}_bind()").
Using -EINVAL restores original behaviour.

For l2tp_ip6, the lookup is now always done with the correct bound
device. Before this patch, when binding to a link-local address, the
lookup was done with the original sk->sk_bound_dev_if, which was later
overwritten with addr->l2tp_scope_id. Lookup is now performed with the
final sk->sk_bound_dev_if value.

Finally, the (addr_len >= sizeof(struct sockaddr_in6)) check has been
dropped: addr is a sockaddr_l2tpip6 not sockaddr_in6 and addr_len has
already been checked at this point (this part of the code seems to have
been copy-pasted from net/ipv6/raw.c).
Signed-off-by: Guillaume Nault <g.nault@alphalink.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>

Socket must be held while under the protection of the l2tp lock; there
is no guarantee that sk remains valid after the read_unlock_bh() call.

Same issue for l2tp_ip and l2tp_ip6.
Signed-off-by: Guillaume Nault <g.nault@alphalink.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>

Socket flags aren't updated atomically, so the socket must be locked
while reading the SOCK_ZAPPED flag.

This issue exists for both l2tp_ip and l2tp_ip6. For IPv6, this patch
also brings error handling for __ip6_datagram_connect() failures.
Signed-off-by: Guillaume Nault <g.nault@alphalink.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Execution 'ethtool -S' on fec device that is down causes OOPS on Vybrid
board:

Unhandled fault: external abort on non-linefetch (0x1008) at 0xe0898200
pgd = ddecc000
[e0898200] *pgd=9e406811, *pte=400d1653, *ppte=400d1453
Internal error: : 1008 [#1] SMP ARM
...

Reason of OOPS is that fec_enet_get_ethtool_stats() accesses fec
registers while IPG clock is stopped by PM.

Fix that by caching statistics in fec_enet_private. Cache is initialized
at device probe time, and updated at statistics request time if device
is up, and also just before turning device off on down path.

Additional locking is not needed, since cached statistics is accessed
either before device is registered, or under rtnl_lock().
Signed-off-by: Nikita Yushchenko <nikita.yoush@cogentembedded.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

vxlan_fdb_append may return error, so add the proper check,
otherwise it will cause memory leak.
Signed-off-by: Haishuang Yan <yanhaishuang@cmss.chinamobile.com>

Changes in v2:
  - Unnecessary to initialize rc to zero.
Acked-by: Jiri Benc <jbenc@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

If nf_ct_frag6_gather() returns an error other than -EINPROGRESS, it
means that we still have a reference to the skb.  We should free it
before returning from handle_fragments, as stated in the comment above.

Fixes: daaa7d64 ("netfilter: ipv6: avoid nf_iterate recursion")
CC: Florian Westphal <fw@strlen.de>
CC: Pravin B Shelar <pshelar@ovn.org>
CC: Joe Stringer <joe@ovn.org>
Signed-off-by: Daniele Di Proietto <diproiettod@ovn.org>
Acked-by: Pravin B Shelar <pshelar@ovn.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

When handling inbound packets, the two halves of the sequence number
stored on the skb are already in network order.

Fixes: 000ae7b2 ("esp6: Switch to new AEAD interface")
Signed-off-by: Tobias Brunner <tobias@strongswan.org>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

When handling inbound packets, the two halves of the sequence number
stored on the skb are already in network order.

Fixes: 7021b2e1 ("esp4: Switch to new AEAD interface")
Signed-off-by: Tobias Brunner <tobias@strongswan.org>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

It doesn't support to run 32bit 'ip' to set xfrm objdect on 64bit host.
But the return value is unknown for user program:

ip xfrm policy list
RTNETLINK answers: Unknown error 524

Replace ENOTSUPP with EOPNOTSUPP:

ip xfrm policy list
RTNETLINK answers: Operation not supported
Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Johan Hovold says:

====================
net: fix fixed-link phydev leaks

This series fixes failures to deregister and free fixed-link phydevs
that have been registered using the of_phy_register_fixed_link()
interface.

All but two drivers currently fail to do this and this series fixes most
of them with the exception of a staging driver and the stmmac drivers
which will be fixed by follow-on patches.

Included are also a couple of fixes for related of-node leaks.

Note that all patches except the of_mdio one have been compile-tested
only.

Also note that the series is against net due to dependencies not yet in
net-next.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Make sure to deregister and free any fixed-link PHY registered using
of_phy_register_fixed_link() on slave-setup errors and on slave destroy.

Fixes: 0d8bcdd3 ("net: dsa: allow for more complex PHY setups")
Signed-off-by: Johan Hovold <johan@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Make sure to deregister and free any fixed-link PHY registered using
of_phy_register_fixed_link() on probe errors and on driver unbind.

Also remember to put the of-node reference on probe errors.

Fixes: 1bb6aa56 ("net: davinci_emac: Add support for fixed-link
PHY")
Signed-off-by: Johan Hovold <johan@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>