When user add rule to autogrouped flow table, we search
for flow group with the same match criteria, if we don't
find such group then we create new flow group with the
required match criteria and insert the rule to this group.

We divide the flow table into required_groups + 1,
in order to reserve a part of the flow table for rules
which don't match any existing group.
Signed-off-by: NMaor Gottlieb <maorg@mellanox.com>
Signed-off-by: NMoni Shoua <monis@mellanox.com>
Signed-off-by: NMatan Barak <matanb@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Daniel Borkmann says:

====================
BPF update

This set adds IPv6 support for bpf_skb_{set,get}_tunnel_key() helper.
It also exports flags to user space that are being used in helpers and
weren't exported thus far. For more details, please see the individual
patches.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

After IPv6 support has recently been added to metadata dst and related
encaps, add support for populating/reading it from an eBPF program.

Commit d3aa45ce ("bpf: add helpers to access tunnel metadata") started
with initial IPv4-only support back then (due to IPv6 metadata support
not being available yet).

To stay compatible with older programs, we need to test for the passed
structure size. Also TOS and TTL support from the ip_tunnel_info key has
been added. Tested with vxlan devs in collect meta data mode with IPv4,
IPv6 and in compat mode over different network namespaces.
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>
Acked-by: NAlexei Starovoitov <ast@kernel.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Export flags used by eBPF helper functions through UAPI, so they can be
used by programs (instead of them redefining all flags each time or just
using the hard-coded values). It also gives a better overview what flags
are used where and we can further get rid of the extra macros defined in
filter.c. Moreover, reject invalid flags.
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>
Acked-by: NAlexei Starovoitov <ast@kernel.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The spew in /proc/net/bonding/bond0 uses netif_carrier_ok() to determine
mii_status, while /sys/class/net/bond0/bonding/mii_status looks at
curr_active_slave, which doesn't actually seem to be set sometimes when
the bond actually is up. A mode 4 bond configured via ifcfg-foo files on a
Red Hat Enterprise Linux system, after boot, comes up clean and
functional, but the sysfs node shows mii_status of down, while proc shows
up. A simple enough fix here seems to be to use the same method for
determining up or down in both places, and I'd opt for the one that seems
to match reality.

CC: Jay Vosburgh <j.vosburgh@gmail.com>
CC: Veaceslav Falico <vfalico@gmail.com>
CC: Andy Gospodarek <gospo@cumulusnetworks.com>
CC: netdev@vger.kernel.org
Signed-off-by: NJarod Wilson <jarod@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The sz_idx variable is defined in the rtnetlink_rcv_msg(), but
not used anywhere. Let's remove it.
Signed-off-by: NAlexander Kuleshov <kuleshovmail@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Use phy_find_first() to find the first phy device instead of
open-coding it.

Cc: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Jean Sacren says:

====================
Trivial fix-ups for openvswitch

This series does trivial fix-ups for openvswitch as follows:

1) Clean up the leftover of the unused function.

2) Fix up the twisted struct geneve_port member name.

3) Update the kernel doc to reflect the changes in struct vport.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

commit be4ace6e ("openvswitch: Move dev pointer into vport itself")

The commit above added @dev and moved @rcu to the bottom of struct
vport, but the change was not reflected in the kernel doc. So let's
update the kernel doc as well.
Signed-off-by: NJean Sacren <sakiwit@gmail.com>
Cc: Thomas Graf <tgraf@suug.ch>
Acked-by: NThomas Graf <tgraf@suug.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

commit 6b001e68 ("openvswitch: Use Geneve device.")

The commit above introduced 'port_no' as the name for the member of
struct geneve_port. The correct name should be 'dst_port' as described
in the kernel doc. Let's fix that member name and all the pertinent
instances so that both doc and code would be consistent.
Signed-off-by: NJean Sacren <sakiwit@gmail.com>
Acked-by: NThomas Graf <tgraf@suug.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

commit 6b001e68 ("openvswitch: Use Geneve device.")

The commit above deleted the only call site of ovs_tunnel_route_lookup()
and now that function is not used any more. So let's delete the function
definition as well.
Signed-off-by: NJean Sacren <sakiwit@gmail.com>
Acked-by: NThomas Graf <tgraf@suug.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Commit 7f854420 ("phy: Add API for {un}registering an mdio device to
a bus") introduces an API to access mii_bus structures, but missed to
update the TI cpamc driver. This results in the following error message.

drivers/net/ethernet/ti/cpmac.c: In function 'cpmac_probe':
drivers/net/ethernet/ti/cpmac.c:1119:18: error:
	'struct mii_bus' has no member named 'phy_map'

Fixes: 7f854420 ("phy: Add API for {un}registering an mdio device to a bus")
Cc: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Acked-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Commit e7f4dc35 ("mdio: Move allocation of interrupts into core")
removes some code from tc_mii_init(), but does not remove a now unused
variable. This results in the following build warning.

drivers/net/ethernet/toshiba/tc35815.c: In function 'tc_mii_init':
drivers/net/ethernet/toshiba/tc35815.c:670:6: warning: unused variable 'i'

Fixes: e7f4dc35 ("mdio: Move allocation of interrupts into core")
Cc: Andrew Lunn <andrew@lunn.ch>
Acked-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Commit 7f854420 ("phy: Add API for {un}registering an mdio device to
a bus") introduces an API to access mii_bus structures, but missed to
update the tc35815 driver. This results in the following error message.

drivers/net/ethernet/toshiba/tc35815.c: In function 'tc_mii_probe':
drivers/net/ethernet/toshiba/tc35815.c:617:18: error:
	'struct mii_bus' has no member named 'phy_map'
drivers/net/ethernet/toshiba/tc35815.c:623:24: error:
	'struct mii_bus' has no member named 'phy_map'

Instead of looping over the list of phy addresses to find a phy chip,
use phy_find_first(). While the intent of the original code was to return
an error if more than one phy was specified, this code path was never
executed because the loop aborted after finding the first phy. The
original code is therefore semantically identical to phy_find_first(),
thus it is simpler and more straightforward to use phy_find_first()
directly.

Fixes: 7f854420 ("phy: Add API for {un}registering an mdio device to a bus")
Cc: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Acked-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

LAN8740 has a different phy_id than LAN8710/LAN8720.
Signed-off-by: NJoshua Henderson <joshua.henderson@microchip.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Merge tag 'wireless-drivers-next-for-davem-2016-01-09' of git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers-next

Kalle Valo says:

====================
brcmfmac

* query features through firmware command
* ARP offload through inet notifier
* force probe to succeed for debugging purposes
* random mac support for scheduled scan
* support wowl upon net detect

iwlwifi

* bug fixes and improvements for firmware debug system
* advertise support for Rx A-MSDU in A-MPDU
* support -20.ucode
* fix WoWLAN for iwldvm
* preparations towards multiple Rx queues
* platform power improvements for GO mode when no clients are associated
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Applications often have to reduce number of datagrams
they receive or send per system call to avoid starvation problems.

Really the kernel should take care of this by using cond_resched(),
so that applications can experiment bigger batch sizes.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The userspace needs to know why is the address being removed so that it can
perhaps obtain a new address.

Without the DADFAILED flag it's impossible to distinguish removal of a
temporary and tentative address due to DAD failure from other reasons (device
removed, manual address removal).
Signed-off-by: NLubomir Rintel <lkundrak@v3.sk>
Acked-by: NHannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Commit e7f4dc35 ("mdio: Move allocation of interrupts into core")
introduced the following build warnings:

drivers/net/ethernet/nxp/lpc_eth.c: In function 'lpc_mii_init':
drivers/net/ethernet/nxp/lpc_eth.c:865:1: warning: label 'err_out_1' defined but not used [-Wunused-label]
drivers/net/ethernet/nxp/lpc_eth.c:826:20: warning: unused variable 'i' [-Wunused-variable]

Remove the unused variables to fix them.
Reported-by: NOlof's autobuilder <build@lixom.net>
Signed-off-by: NFabio Estevam <fabio.estevam@nxp.com>
Acked-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

While building blackfin defconfig we were getting a build warning:
warning: label 'out_err_irq_alloc' defined but not used.

Commit e7f4dc35 ("mdio: Move allocation of interrupts into core")
removed the label out_err_mdiobus_register but then mistakenly jumped to
out_err_alloc. But it was actually supposed to jump to out_err_irq_alloc.

Fixes: e7f4dc35 ("mdio: Move allocation of interrupts into core")
Cc: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: NSudip Mukherjee <sudip@vectorindia.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The build of blackfin defconfig is failing with the error:
error: 'struct mii_bus' has no member named 'phy_map'

A new API mdiobus_get_phy() was introduced and phy_map was removed but
it was not changed here.

Fixes: 7f854420 ("phy: Add API for {un}registering an mdio device to a bus.")
Cc: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: NSudip Mukherjee <sudip@vectorindia.org>
Acked-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The commit 115b56af ("cxgb4: Update mps_tcam output to include T6
fields") from Dec 23, 2015, leads to the following static checker
warning:

        drivers/net/ethernet/chelsio/cxgb4/cxgb4_debugfs.c:1735
mps_tcam_show()
        warn: we tested 'lookup_type' before and it was 'true'

Fixing it.
Reported-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NHariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Xing Zheng says:

====================
Add support emac for the RK3036 SoC platform

  We have supported the emac for RK3066/RK3188, but the RK3036 have
some configuration different with them. We should let the driver of
emac_rockchip compatible with other Rockchip SoCs.

Changes in v2:
- Separate DTS from patch series.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The RK3036's GRFs offset are different with RK3066/RK3188, and need to set
mac TX/RX clock before probe emac.
Signed-off-by: NXing Zheng <zhengxing@rock-chips.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

On the RK3066/RK3188, there was fixed GRF offset configuration to set emac
and fixed DIV2 mac TX/RX clock. So, we need to easily set and fit to other
SoCs (RK3036) which maybe have different GRF offset, and need adjust mac
TX/RX clock.
Signed-off-by: NXing Zheng <zhengxing@rock-chips.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

After enter arc_emac_probe, emac will get_phy_id, phy_poll_reset and
other connecting PHY via mdiobus_read, so we need to set correct
ref clock rate for emac before probe emac.
Signed-off-by: NXing Zheng <zhengxing@rock-chips.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Michael Chan says:

====================
bnxt_en: Zero pad fw messages and add fw reset.

2 patches related to firmware for net-next.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Use HWRM_FW_RESET command to request a self-reset of the embedded
processor(s) after successfully applying a firmware update. For boot
processor, the self-reset is currently deferred until the next PCIe reset.
Signed-off-by: NRob Swindell <swindell@broadcom.com>
Signed-off-by: NMichael Chan <mchan@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

For future compatibility, zero pad all messages that the driver sends
to the firmware to 128 bytes.  If these messages are extended in the
future with new byte enables, zero padding these messages now will
guarantee future compatibility.
Signed-off-by: NMichael Chan <mchan@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This work adds a generalization of the ingress qdisc as a qdisc holding
only classifiers. The clsact qdisc works on ingress, but also on egress.
In both cases, it's execution happens without taking the qdisc lock, and
the main difference for the egress part compared to prior version of [1]
is that this can be applied with _any_ underlying real egress qdisc (also
classless ones).

Besides solving the use-case of [1], that is, allowing for more programmability
on assigning skb->priority for the mqprio case that is supported by most
popular 10G+ NICs, it also opens up a lot more flexibility for other tc
applications. The main work on classification can already be done at clsact
egress time if the use-case allows and state stored for later retrieval
f.e. again in skb->priority with major/minors (which is checked by most
classful qdiscs before consulting tc_classify()) and/or in other skb fields
like skb->tc_index for some light-weight post-processing to get to the
eventual classid in case of a classful qdisc. Another use case is that
the clsact egress part allows to have a central egress counterpart to
the ingress classifiers, so that classifiers can easily share state (e.g.
in cls_bpf via eBPF maps) for ingress and egress.

Currently, default setups like mq + pfifo_fast would require for this to
use, for example, prio qdisc instead (to get a tc_classify() run) and to
duplicate the egress classifier for each queue. With clsact, it allows
for leaving the setup as is, it can additionally assign skb->priority to
put the skb in one of pfifo_fast's bands and it can share state with maps.
Moreover, we can access the skb's dst entry (f.e. to retrieve tclassid)
w/o the need to perform a skb_dst_force() to hold on to it any longer. In
lwt case, we can also use this facility to setup dst metadata via cls_bpf
(bpf_skb_set_tunnel_key()) without needing a real egress qdisc just for
that (case of IFF_NO_QUEUE devices, for example).

The realization can be done without any changes to the scheduler core
framework. All it takes is that we have two a-priori defined minors/child
classes, where we can mux between ingress and egress classifier list
(dev->ingress_cl_list and dev->egress_cl_list, latter stored close to
dev->_tx to avoid extra cacheline miss for moderate loads). The egress
part is a bit similar modelled to handle_ing() and patched to a noop in
case the functionality is not used. Both handlers are now called
sch_handle_ingress() and sch_handle_egress(), code sharing among the two
doesn't seem practical as there are various minor differences in both
paths, so that making them conditional in a single handler would rather
slow things down.

Full compatibility to ingress qdisc is provided as well. Since both
piggyback on TC_H_CLSACT, only one of them (ingress/clsact) can exist
per netdevice, and thus ingress qdisc specific behaviour can be retained
for user space. This means, either a user does 'tc qdisc add dev foo ingress'
and configures ingress qdisc as usual, or the 'tc qdisc add dev foo clsact'
alternative, where both, ingress and egress classifier can be configured
as in the below example. ingress qdisc supports attaching classifier to any
minor number whereas clsact has two fixed minors for muxing between the
lists, therefore to not break user space setups, they are better done as
two separate qdiscs.

I decided to extend the sch_ingress module with clsact functionality so
that commonly used code can be reused, the module is being aliased with
sch_clsact so that it can be auto-loaded properly. Alternative would have been
to add a flag when initializing ingress to alter its behaviour plus aliasing
to a different name (as it's more than just ingress). However, the first would
end up, based on the flag, choosing the new/old behaviour by calling different
function implementations to handle each anyway, the latter would require to
register ingress qdisc once again under different alias. So, this really begs
to provide a minimal, cleaner approach to have Qdisc_ops and Qdisc_class_ops
by its own that share callbacks used by both.

Example, adding qdisc:

   # tc qdisc add dev foo clsact
   # tc qdisc show dev foo
   qdisc mq 0: root
   qdisc pfifo_fast 0: parent :1 bands 3 priomap  1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
   qdisc pfifo_fast 0: parent :2 bands 3 priomap  1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
   qdisc pfifo_fast 0: parent :3 bands 3 priomap  1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
   qdisc pfifo_fast 0: parent :4 bands 3 priomap  1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
   qdisc clsact ffff: parent ffff:fff1

Adding filters (deleting, etc works analogous by specifying ingress/egress):

   # tc filter add dev foo ingress bpf da obj bar.o sec ingress
   # tc filter add dev foo egress  bpf da obj bar.o sec egress
   # tc filter show dev foo ingress
   filter protocol all pref 49152 bpf
   filter protocol all pref 49152 bpf handle 0x1 bar.o:[ingress] direct-action
   # tc filter show dev foo egress
   filter protocol all pref 49152 bpf
   filter protocol all pref 49152 bpf handle 0x1 bar.o:[egress] direct-action

A 'tc filter show dev foo' or 'tc filter show dev foo parent ffff:' will
show an empty list for clsact. Either using the parent names (ingress/egress)
or specifying the full major/minor will then show the related filter lists.

Prior work on a mqprio prequeue() facility [1] was done mainly by John Fastabend.

  [1] http://patchwork.ozlabs.org/patch/512949/Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>
Acked-by: NJohn Fastabend <john.r.fastabend@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The warning about being able to read any MDIO device, not just the
attached ethernet devices PHY applies to all MDIO drivers. So remove
it. This also removes a reference to a member in phy_device which has
moved.
Signed-off-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Fix a number of build errors due to moving the phy_map and centralizing
interrupt allocation.
Signed-off-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Use phy_find_first() to find the first phy device instead of
open coding it.

Cc: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Acked-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Commit 7f854420 ("phy: Add API for {un}registering an mdio device to
a bus") introduces an API to access mii_bus structures, but missed to
update the sb1250 driver. This results in the following build error.

drivers/net/ethernet/broadcom/sb1250-mac.c: In function 'sbmac_mii_probe':
drivers/net/ethernet/broadcom/sb1250-mac.c:2360:24: error:
	'struct mii_bus' has no member named 'phy_map'

Use phy_find_first() instead of open coding it.

Commit 2220943a ("phy: Centralise print about attached phy") introduces
the following build error.

drivers/net/ethernet/broadcom/sb1250-mac.c: In function 'sbmac_mii_probe':
drivers/net/ethernet/broadcom/sb1250-mac.c:2383:20: error: 'phydev' undeclared

Fixes: 7f854420 ("phy: Add API for {un}registering an mdio device to a bus")
Fixes: 2220943a ("phy: Centralise print about attached phy")
Cc: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Acked-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Andrew Lunn says:

====================
Fix breakage from mdio device

These two patches fix MIPS platforms which got broken by
the recent mdio device patchset.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The mdio device patches moved the bus member in phy_device into a
substructure. This driver got missed. Fix it.
Signed-off-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Make use of the helper to find the first phy device.
This also fixes the compile breakage.
Signed-off-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Originally, most of the platforms using this driver did not define an mdio subnode
in the devicetree. Commit e34d65 ("stmmac: create of compatible mdio bus for stmmac driver")
introduced a backward compatibily issue by using of_mdiobus_register explicitly
with an mdio subnode. This patch fixes the issue by calling the function
mdiobus_register, when mdio subnode is not found. The driver is now compatible
with both modes.

Fixes: e34d6569 ("stmmac: create of compatible mdio bus for stmmac driver")
Signed-off-by: NRomain Perier <romain.perier@gmail.com>
Tested-by: NPhil Reid <preid@electromag.com.au>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Daniel Borkmann says:

====================
BPF update

Fixes a csum issue on ingress. As mentioned previously, net-next
seems just fine imho. Later on, will follow up with couple of
replacements like ovs_skb_postpush_rcsum() etc.

Thanks!

v1 -> v2:
  - Added patch 1 with helper
  - Implemented Hannes' idea to just use csum_partial, thanks!
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add a small helper skb_postpush_rcsum() and fix up redirect locations
that need CHECKSUM_COMPLETE fixups on ingress. dev_forward_skb() expects
a proper csum that covers also Ethernet header, f.e. since 2c26d34b
("net/core: Handle csum for CHECKSUM_COMPLETE VXLAN forwarding"), we
also do skb_postpull_rcsum() after pulling Ethernet header off via
eth_type_trans().

When using eBPF in a netns setup f.e. with vxlan in collect metadata mode,
I can trigger the following csum issue with an IPv6 setup:

  [  505.144065] dummy1: hw csum failure
  [...]
  [  505.144108] Call Trace:
  [  505.144112]  <IRQ>  [<ffffffff81372f08>] dump_stack+0x44/0x5c
  [  505.144134]  [<ffffffff81607cea>] netdev_rx_csum_fault+0x3a/0x40
  [  505.144142]  [<ffffffff815fee3f>] __skb_checksum_complete+0xcf/0xe0
  [  505.144149]  [<ffffffff816f0902>] nf_ip6_checksum+0xb2/0x120
  [  505.144161]  [<ffffffffa08c0e0e>] icmpv6_error+0x17e/0x328 [nf_conntrack_ipv6]
  [  505.144170]  [<ffffffffa0898eca>] ? ip6t_do_table+0x2fa/0x645 [ip6_tables]
  [  505.144177]  [<ffffffffa08c0725>] ? ipv6_get_l4proto+0x65/0xd0 [nf_conntrack_ipv6]
  [  505.144189]  [<ffffffffa06c9a12>] nf_conntrack_in+0xc2/0x5a0 [nf_conntrack]
  [  505.144196]  [<ffffffffa08c039c>] ipv6_conntrack_in+0x1c/0x20 [nf_conntrack_ipv6]
  [  505.144204]  [<ffffffff8164385d>] nf_iterate+0x5d/0x70
  [  505.144210]  [<ffffffff816438d6>] nf_hook_slow+0x66/0xc0
  [  505.144218]  [<ffffffff816bd302>] ipv6_rcv+0x3f2/0x4f0
  [  505.144225]  [<ffffffff816bca40>] ? ip6_make_skb+0x1b0/0x1b0
  [  505.144232]  [<ffffffff8160b77b>] __netif_receive_skb_core+0x36b/0x9a0
  [  505.144239]  [<ffffffff8160bdc8>] ? __netif_receive_skb+0x18/0x60
  [  505.144245]  [<ffffffff8160bdc8>] __netif_receive_skb+0x18/0x60
  [  505.144252]  [<ffffffff8160ccff>] process_backlog+0x9f/0x140
  [  505.144259]  [<ffffffff8160c4a5>] net_rx_action+0x145/0x320
  [...]

What happens is that on ingress, we push Ethernet header back in, either
from cls_bpf or right before skb_do_redirect(), but without updating csum.
The "hw csum failure" can be fixed by using the new skb_postpush_rcsum()
helper for the dev_forward_skb() case to correct the csum diff again.

Thanks to Hannes Frederic Sowa for the csum_partial() idea!

Fixes: 3896d655 ("bpf: introduce bpf_clone_redirect() helper")
Fixes: 27b29f63 ("bpf: add bpf_redirect() helper")
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>
Acked-by: NAlexei Starovoitov <ast@kernel.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>