With the patch
commit f8aa2696b4af ("esp: check the NETIF_F_HW_ESP_TX_CSUM bit before segmenting")
we no longer need to protect ourself from checksum
offload requests on IPsec packets, so we can remove
the check in our .ndo_features_check callback.
Signed-off-by: Shannon Nelson <shannon.nelson@oracle.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Replaced an assignment operation with an OR operation.
The variable assignment was overwriting the value read from the PHY
register. The OR operation sets only the intended register bits.
The bits that were being overwritten are reserved, so the assignment had no
functional impact.
Reported by: Shannon Nelson <shannon.nelson@oracle.com>
Signed-off-by: Paul Greenwalt <paul.greenwalt@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Add status register reads and delay between reads to ixgbe_check_remove.
Registers can read 0xFFFFFFFF during PCI reset, which causes the driver
to remove the adapter. The additional status register reads can reduce the
chance of this race condition.
If the status register is not 0xFFFFFFFF, then ixgbe_check_remove returns
the value of the register being read.
Signed-off-by: Paul Greenwalt <paul.greenwalt@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The phys embedded into the v1.1 of the VR9 SoC are using different phy
ids. Add the phy ids to use the driver for this VR9 version as well.
Signed-off-by: Mathias Kresin <dev@kresin.me>
Signed-off-by: David S. Miller <davem@davemloft.net>
The VR9 phy ids are matching only for the SoC version 1.2. Rename the
macros and change the names to take this into account.
Signed-off-by: Mathias Kresin <dev@kresin.me>
Signed-off-by: David S. Miller <davem@davemloft.net>
Assign true or false to boolean variables instead of an integer value.
This issue was detected with the help of Coccinelle.
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
_rule_ is being freed and then dereferenced by accessing rule->ctx
Fix this by copying the value returned by PTR_ERR(rule->ctx) into a local
variable for its safe use after freeing _rule_
Addresses-Coverity-ID: 1466041 ("Read from pointer after free")
Fixes: 05564d0ae0 ("net/mlx5: Add flow-steering commands for FPGA IPSec implementation")
Reviewed-by: Yuval Shaia <yuval.shaia@oracle.com>
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Acked-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since the first element of struct mlxsw_sp_span_parms is a pointer,
to zero-initialize this structure the correct notation is not = {0}, but
rather = {NULL}, as reported by sparse.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Update MTU of overlay loopback in accordance with the setting on the
tunnel netdevice.
Fixes: 0063587d35 ("mlxsw: spectrum: Support decap-only IP-in-IP tunnels")
Signed-off-by: Petr Machata <petrm@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Move the function so that it can be called without forward declaration
from a function that will be added in a follow-up patch.
Fixes: 0063587d35 ("mlxsw: spectrum: Support decap-only IP-in-IP tunnels")
Signed-off-by: Petr Machata <petrm@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add the SPDX identifiers to all the Intel wired LAN driver files, as
outlined in Documentation/process/license-rules.rst.
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The operstate update logic will leave an interface in the
default UNKNOWN operstate if the interface carrier state never changes
from the default carrier up state set at creation. This includes the
case of an explicit call to netif_carrier_on, as the carrier on to on
transition has no effect on operstate.
This affects virtio-net for the case that the virtio peer does
not support VIRTIO_NET_F_STATUS (the feature that provides carrier state
updates). Without this feature, the virtio specification states that
"the link should be assumed active," so, logically, the operstate should
be UP instead of UNKNOWN. This has impact on user space applications
that use the operstate to make availability decisions for the interface.
Resolve this by changing the virtio probe logic slightly to call
netif_carrier_off for both the "with" and "without" VIRTIO_NET_F_STATUS
cases, and then the existing call to netif_carrier_on for the "without"
case will cause an operstate transition.
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Jay Vosburgh <jay.vosburgh@canonical.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fun set of conflict resolutions here...
For the mac80211 stuff, these were fortunately just parallel
adds. Trivially resolved.
In drivers/net/phy/phy.c we had a bug fix in 'net' that moved the
function phy_disable_interrupts() earlier in the file, whilst in
'net-next' the phy_error() call from this function was removed.
In net/ipv4/xfrm4_policy.c, David Ahern's changes to remove the
'rt_table_id' member of rtable collided with a bug fix in 'net' that
added a new struct member "rt_mtu_locked" which needs to be copied
over here.
The mlxsw driver conflict consisted of net-next separating
the span code and definitions into separate files, whilst
a 'net' bug fix made some changes to that moved code.
The mlx5 infiniband conflict resolution was quite non-trivial,
the RDMA tree's merge commit was used as a guide here, and
here are their notes:
====================
Due to bug fixes found by the syzkaller bot and taken into the for-rc
branch after development for the 4.17 merge window had already started
being taken into the for-next branch, there were fairly non-trivial
merge issues that would need to be resolved between the for-rc branch
and the for-next branch. This merge resolves those conflicts and
provides a unified base upon which ongoing development for 4.17 can
be based.
Conflicts:
drivers/infiniband/hw/mlx5/main.c - Commit 42cea83f95
(IB/mlx5: Fix cleanup order on unload) added to for-rc and
commit b5ca15ad7e (IB/mlx5: Add proper representors support)
add as part of the devel cycle both needed to modify the
init/de-init functions used by mlx5. To support the new
representors, the new functions added by the cleanup patch
needed to be made non-static, and the init/de-init list
added by the representors patch needed to be modified to
match the init/de-init list changes made by the cleanup
patch.
Updates:
drivers/infiniband/hw/mlx5/mlx5_ib.h - Update function
prototypes added by representors patch to reflect new function
names as changed by cleanup patch
drivers/infiniband/hw/mlx5/ib_rep.c - Update init/de-init
stage list to match new order from cleanup patch
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull networking fixes from David Miller:
1) Always validate XFRM esn replay attribute, from Florian Westphal.
2) Fix RCU read lock imbalance in xfrm_get_tos(), from Xin Long.
3) Don't try to get firmware dump if not loaded in iwlwifi, from Shaul
Triebitz.
4) Fix BPF helpers to deal with SCTP GSO SKBs properly, from Daniel
Axtens.
5) Fix some interrupt handling issues in e1000e driver, from Benjamin
Poitier.
6) Use strlcpy() in several ethtool get_strings methods, from Florian
Fainelli.
7) Fix rhlist dup insertion, from Paul Blakey.
8) Fix SKB leak in netem packet scheduler, from Alexey Kodanev.
9) Fix driver unload crash when link is up in smsc911x, from Jeremy
Linton.
10) Purge out invalid socket types in l2tp_tunnel_create(), from Eric
Dumazet.
11) Need to purge the write queue when TCP connections are aborted,
otherwise userspace using MSG_ZEROCOPY can't close the fd. From
Soheil Hassas Yeganeh.
12) Fix double free in error path of team driver, from Arkadi
Sharshevsky.
13) Filter fixes for hv_netvsc driver, from Stephen Hemminger.
14) Fix non-linear packet access in ipv6 ndisc code, from Lorenzo
Bianconi.
15) Properly filter out unsupported feature flags in macvlan driver,
from Shannon Nelson.
16) Don't request loading the diag module for a protocol if the protocol
itself is not even registered. From Xin Long.
17) If datagram connect fails in ipv6, make sure the socket state is
consistent afterwards. From Paolo Abeni.
18) Use after free in qed driver, from Dan Carpenter.
19) If received ipv4 PMTU is less than the min pmtu, lock the mtu in the
entry. From Sabrina Dubroca.
20) Fix sleep in atomic in tg3 driver, from Jonathan Toppins.
21) Fix vlan in vlan untagging in some situations, from Toshiaki Makita.
22) Fix double SKB free in genlmsg_mcast(). From Nicolas Dichtel.
23) Fix NULL derefs in error paths of tcf_*_init(), from Davide Caratti.
24) Unbalanced PM runtime calls in FEC driver, from Florian Fainelli.
25) Memory leak in gemini driver, from Igor Pylypiv.
26) IDR leaks in error paths of tcf_*_init() functions, from Davide
Caratti.
27) Need to use GFP_ATOMIC in seg6_build_state(), from David Lebrun.
28) Missing dev_put() in error path of macsec_newlink(), from Dan
Carpenter.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (201 commits)
macsec: missing dev_put() on error in macsec_newlink()
net: dsa: Fix functional dsa-loop dependency on FIXED_PHY
hv_netvsc: common detach logic
hv_netvsc: change GPAD teardown order on older versions
hv_netvsc: use RCU to fix concurrent rx and queue changes
hv_netvsc: disable NAPI before channel close
net/ipv6: Handle onlink flag with multipath routes
ppp: avoid loop in xmit recursion detection code
ipv6: sr: fix NULL pointer dereference when setting encap source address
ipv6: sr: fix scheduling in RCU when creating seg6 lwtunnel state
net: aquantia: driver version bump
net: aquantia: Implement pci shutdown callback
net: aquantia: Allow live mac address changes
net: aquantia: Add tx clean budget and valid budget handling logic
net: aquantia: Change inefficient wait loop on fw data reads
net: aquantia: Fix a regression with reset on old firmware
net: aquantia: Fix hardware reset when SPI may rarely hangup
s390/qeth: on channel error, reject further cmd requests
s390/qeth: lock read device while queueing next buffer
s390/qeth: when thread completes, wake up all waiters
...
PF needs to assert the VF reset when it receives the request to
reset from VF. After receiving request PF ackknowledges the
request by replying back MBX_ASSERTING_RESET message to VF.
VF then goes to pending state and wait for hardware to complete
the reset.
This patch contains code to handle the received VF message, inform
the VF of assertion and reset the VF using cmdq interface.
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Reset Asserting message is forwarded by PF to inform VF about
the hardware reset which is about to happen. This might be due
to the earlier VF reset request received by the PF or because PF
for any reason decides to undergo reset. This message results in
VF to go in pending state in which it polls the hardware to
complete the reset and then further resets/tears its own stack.
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Current mailbox CRQ could consists of both synchronous and async
responses from the PF. Synchronous responses are time critical
and should be handed over to the waiting tasks/context as quickly
as possible otherwise timeout occurs.
Above problem gets accentuated if CRQ consists of even single
async message. Hence, it is important to have quick handling of
synchronous messages and maybe deferred handling of async messages
This patch introduces separate ARQ(async receive queues) for the
async messages. These messages are processed later with repsect
to mailbox task while synchronous messages still gets processed
in context to mailbox interrupt.
ARQ is important as VF reset introduces some new async messages
like MBX_ASSERTING_RESET which adds up to the presssure on the
responses for synchronousmessages and they timeout even more
quickly.
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After the hardware reset we should re-fetch the configuration from
PF like queue info and tc info. This might have impact on allocations
made like that of TQPs. Hence, we should release all such allocations
and re-allocate fresh according to new fetched configuration after
reset.
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After VF driver knows that hardware reset has been performed
successfully, it should proceed ahead and reset the enet layer.
This primarily consists of bringing down interface, clearing
TX/RX rings, disassociating vectors from ring etc.
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
VF driver depends upon PF to eventually reset the hardware. This
request is made using the mailbox command. This patch adds the
required function to acheive above.
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This introduces the hclge device reset states of "requested" and
"pending" and also its handling in context to Reset Service Task.
Device gets into requested state because of any VF reset request
asserted from upper layers, for example due to watchdog timeout
expiration. Requested state would result in eventually forwarding
the VF reset request to PF which would actually reset the VF.
Device will get into pending state if:
1. VF receives the acknowledgement from PF for the VF reset
request it originally sent to PF.
2. Reset Service Task detects that after asserting VF reset for
certain times the data-path is not working and device then
decides to assert full VF reset(this means also resetting the
PCIe interface).
3. PF intimates the VF that it has undergone reset.
Pending state would result in VF to poll for hardware reset
completion status and then resetting the stack/enet layer, which
in turn means reinitializing the ring management/enet layer.
Note: we would be adding support of 3. later as a separate patch.
This decision should not affect VF reset as its event handling
is generic in nature.
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
VF reset would involve handling of different reset related events
from the stack, physical function, mailbox etc. Reset service task
would be used in servicing such reset event requests and later
handling the hardware completions waits and initiating the stack
resets.
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
HNS3 drivers enet layer, used for the ring management and stack
interaction, is common to both VF and PF. PF already supports reset
functionality to handle the network stack watchdog timeout trigger
but the existing code is not generic enough to be used to support VF
reset as well.
This patch does following:
1. Makes the existing watchdog timeout handler in enet layer generic
i.e. suitable for both VF and PF and
2. Introduces the new reset event handler for the VF code.
3. Changes existing reset event handler of PF code to initialize the
reset level
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When a VF is trusted, all promiscuous traffic will only be sent to that VF.
In normal operation promiscuous traffic is sent to the PF. There can be
only one trusted VF per PF
Signed-off-by: Intiyaz Basha <intiyaz.basha@cavium.com>
Signed-off-by: Felix Manlunas <felix.manlunas@cavium.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This is needed to query the mux_id and flags of a rmnet device.
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Define new netlink attributes for rmnet mux_id and flags. These
flags / mux_id were earlier using vlan flags / id respectively.
The flag bits are also moved to uapi and are renamed with
prefix RMNET_FLAG_*.
Also add the rmnet policy to handle the new netlink attributes.
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Device of the de-aggregated skb is correctly assigned after inspecting
the mux_id, so remove the assignment in rmnet_map_deaggregate().
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix warnings which were reported when running with sparse
(make C=1 CF=-D__CHECK_ENDIAN__)
drivers/net/ethernet/qualcomm/rmnet/rmnet_handlers.c:81:15:
warning: cast to restricted __be16
drivers/net/ethernet/qualcomm/rmnet/rmnet_map_data.c:271:37:
warning: incorrect type in assignment (different base types)
expected unsigned short [unsigned] [usertype] pkt_len
got restricted __be16 [usertype] <noident>
drivers/net/ethernet/qualcomm/rmnet/rmnet_map_data.c:287:29:
warning: incorrect type in assignment (different base types)
expected unsigned short [unsigned] [usertype] pkt_len
got restricted __be16 [usertype] <noident>
drivers/net/ethernet/qualcomm/rmnet/rmnet_map_data.c:310:22:
warning: cast to restricted __be16
drivers/net/ethernet/qualcomm/rmnet/rmnet_map_data.c:319:13:
warning: cast to restricted __be16
drivers/net/ethernet/qualcomm/rmnet/rmnet_map_command.c:49:18:
warning: cast to restricted __be16
drivers/net/ethernet/qualcomm/rmnet/rmnet_map_command.c:50:18:
warning: cast to restricted __be32
drivers/net/ethernet/qualcomm/rmnet/rmnet_map_command.c:74:21:
warning: cast to restricted __be16
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Array mvpp2_pools is being indexed by long_log_pool, however this
looks like a cut-n-paste bug and in fact should be short_log_pool.
Detected by CoverityScan, CID#1466113 ("Copy-paste error")
Fixes: 576193f2d5 ("net: mvpp2: jumbo frames support")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Acked-by: Antoine Tenart <antoine.tenart@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We moved the dev_hold(real_dev); call earlier in the function but forgot
to update the error paths.
Fixes: 0759e552bc ("macsec: fix negative refcnt on parent link")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
request_firmware_cache() will ensure the firmware is available on resume
from suspend if on reboot the device retains the firmware.
This optimization is in place given otherwise on reboot we have to
reload the firmware, the opmization saves us about max 1s, minimum 10ms.
Cantabile has reported back this fixes his woes with both suspend and
hibernation.
Reported-by: Cantabile <cantabile.desu@gmail.com>
Tested-by: Cantabile <cantabile.desu@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Johannes Berg says:
====================
Two more fixes (in three patches):
* ath9k_htc doesn't like QoS NDP frames, use regular ones
* hwsim: set up wmediumd for radios created later
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
In 'auto-neg off' the device have sent AN (auto-negotiation) frames
with the forced speed. Thus, fix it using an_disable_admin field in
Port type and speed (PTYS) register. This field indicates if speed
negotiation frames would be send by the port or not.
Add the field and enable/disable it for 'auto-neg on/off', make the
port to start/stop sending AN (auto-negotiation) frames. Note that for
SwitchX2 the behavior doesn't change (i.e support only AN enabled with
forced speed).
Signed-off-by: Tal Bar <talb@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When phy exists, phy_ethtool_ksettings_get function is enough to get the
link ksettings. If the phy exists, get_link_ksettings function can return
directly after phy_ethtool_ksettings_get is called.
Signed-off-by: Fuyun Liang <liangfuyun1@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fixed link mode is returned by hns3_get_link_ksettings. It is
unreasonable.
This patch fixes it by adding some related functions to get link
mode from hardware.
Signed-off-by: Fuyun Liang <liangfuyun1@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since we change the update rate of int_gl from every interrupt to every
one hundred interrupts, the old way to get time interval by int_gl value
is not accurate. This patch calculates the time interval using the jiffies
value.
Signed-off-by: Fuyun Liang <liangfuyun1@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The interrupt coalescing self-adaptive function updates the int_gl every
interrupt. The GL update rate is too faster to get a better new GL value.
This patch changes the GL update rate to every one hundred interrupts.
The GL update rate is defined by HNS3_INT_ADAPT_DOWN_START.
Signed-off-by: Fuyun Liang <liangfuyun1@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
It may need more time for IMP handle some command, such as reset.
This patch enlarges the max time for cmd timeout.
Driver will check the IMP result every us, it may break through the
loop when get the right result. So not all command need the max time.
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There is no module that is dependent on hclge or hclgevf's symbol,
but hns_enet need them to provide ops for it to run. When there is
a need to auto load the hns3 driver, the auto load will fail because
hclge or hclgevf is not loaded.
Hns_enet has already exported the pci table, so this patch exports
the pci table for hclge and hclgevf module too.
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The vlan table in hardware is clear after PF/Core/IMP/Global
reset, which will cause vlan tagged packets not being received
problem.
This patch fixes it by restoring the vlan table after reset.
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
VF queue reset flow is different from PF queue reset flow.
VF driver should stop VF queue first, then send message to PF
and PF do the reset. PF should send a response to VF after
PF complete the queue reset, VF can initialize the queue hw
after get the response.
This patch fixes the VF queue reset flow as the correct step.
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When changing the mtu, the max frame size also will be changed. The tx
buffer size and the rx buffer size to be allocated are determined by max
frame size. So when max frame size is changed, the tx buffer and rx buffer
need to be reallocated.
When the tc_num is changed, the tx buffer and rx buffer need to be
reallocated too. So calling set_mtu and buffer_alloc separately is better.
Signed-off-by: Fuyun Liang <liangfuyun1@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
top_hierarchy arg can be determined by comparing parent_resource_id to
DEVLINK_RESOURCE_ID_PARENT_TOP so it does not need to be a separate
argument.
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
