OF series functions can be used only for DT case. Use unified
device property function instead to support both DT and ACPI.
Signed-off-by: Kejian Yan <yankejian@huawei.com>
Signed-off-by: Yisen Zhuang <Yisen.Zhuang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
hns-mdio needs to register itself to mii-bus. The info of the device can
be read by both DT and ACPI.
HNS tries to call Linux PHY driver to help access PHY-devices, the HNS
hardware topology is as below. The MDIO controller may control several
PHY-devices, and each PHY-device connects to a MAC device. The MDIO will
be registered to mdiobus, then PHY-devices will register when each mac
find PHY device.
cpu
|
|
-------------------------------------------
| | |
| | |
| dsaf |
MDIO | MDIO
| --------------------------- |
| | | | | |
| | | | | |
| MAC MAC MAC MAC |
| | | | | |
---- |-------- |-------- | | --------
|| || || ||
PHY PHY PHY PHY
And the driver can handle reset sequence by _RST method in DSDT in ACPI
case.
Signed-off-by: Kejian Yan <yankejian@huawei.com>
Signed-off-by: Yisen Zhuang <Yisen.Zhuang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Hns-mdio only supports DT case now. do some cleanup to prepare
for introducing other cases later, no functional change.
Signed-off-by: Kejian Yan <yankejian@huawei.com>
Signed-off-by: Yisen Zhuang <Yisen.Zhuang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
acpi_evaluate_dsm() will be used to handle the _DSM method in ACPI case.
It will be compiled in non-ACPI case, but the function is in acpi_bus.h
and acpi_bus.h can only be used in ACPI case, so this patch add the stub
function to linux/acpi.h to make compiled successfully in non-ACPI cases.
Cc: Rafael J. Wysocki <rjw@rjwysocki.net>
Signed-off-by: Kejian Yan <yankejian@huawei.com>
Signed-off-by: Yisen Zhuang <Yisen.Zhuang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
acpi_dev_found() will be used to detect if a given ACPI device is in the
system. It will be compiled in non-ACPI case, but the function is in
acpi_bus.h and acpi_bus.h can only be used in ACPI case, so this patch add
the stub function to linux/acpi.h to make compiled successfully in
non-ACPI cases.
Cc: Rafael J. Wysocki <rjw@rjwysocki.net>
Signed-off-by: Kejian Yan <yankejian@huawei.com>
Signed-off-by: Yisen Zhuang <Yisen.Zhuang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Lunn says:
====================
New DSA bind, switches as devices
The interesting patches here are the last three. They implement a new
binding for DSA, which removes a few limitations of the current DSA
binding. In particular, it allows switches to be true Linux devices.
These devices can be on any type of bus, unlike the old DSA binding
which assumes MDIO. See the commit log for more details. The second to
last patch modifies an existing boards device tree to use the new
binding, giving a good example of how switches can be true MDIO
devices. The last patch documents the new binding.
Thanks go to Florian and Vivien for reviewing, testing and bug fixing
these patches.
Tested-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Tested-by: Florian Fainelli <f.fainelli@gmail.com>
Since V1:
* Add lots of reviewed-by's
* Fix rtable comment
* dsa2: Clear cpu port mask in dsa_cpu_port_unapply()
* dsa2: Only set dsa_port_mask when port successfully configured
* dsa: clear {dsa|cpu}_port_mask on destroy
Since RFC:
* Split the mv88e6xxx MDIO refactor into a rename patch and a refactor
patch.
* Extend commit message with comment about wrong of_node_put()
* Fix destroy of cpu and dsa ports.
* Rename _DSA_TAG_LAST to DSA_TAG_LAST and add a comment.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Add the new binding to the documentation of the existing binding.
Mark the old binding as deprecated.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Hang the three switches of the three MDIO busses using the new DSA
binding. Also, make use of the mdio-bus and explicitly list the phys
on one device. This is not required, but good for testing.
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 existing DSA binding has a number of limitations and problems. The
main problem is that it cannot represent a switch as a linux device,
hanging off some bus. It is limited to one CPU port. The DSA platform
device is artificial, and does not really represent hardware.
Implement a new binding which can be embedded into any type of node on
a bus to represent one switch device, and its links to other switches.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Have the switch driver register its own MDIO bus. This allows for an
mdio property in the device tree, with child nodes for phys, which
can be referenced via phandles, etc.
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 switch implements a generic MDIO bus, which could host more than
PHYs. It is conventional to use _mdio_ or _mii_ in the function name,
so rename them. Also postfix make the historically first read/write
function with _direct, to help distinguish it from _indirect and _ppu.
While touching these functions, remove some of the _ prefixes, which
we are deprecating.
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 switch may want to instantiate its own MDIO bus. Only do it
centrally if the switch has not already created one, and the read op
is implemented.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Replace the two switch statements with an array lookup, and store the
result in the dsa tree structure. The drivers no longer need to know
the selected tag protocol, so remove it from the dsa switch structure.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The merged driver no longer offers the option to use DSA tagging. So
remove the code to setup the switch to do DSA tagging and hard code
the use of EDSA.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>y
Signed-off-by: David S. Miller <davem@davemloft.net>
Refactor the code to setup a single DSA/CPU port into a function of
its own, and export it, so it can be used by the new binding.
Similarly, refactor the destroy code into a function. When destroying
the ports, don't put the of node. They should be released at the end
along with the normal ports.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The new binding will not have a chip data structure, it will place the
routing directly into the switch structure. To enable backwards
compatibility, copy the routing from the chip data into the switch
structure.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
With a maximum of four switches, the size of the routing table is the
same as the pointer to it. Removing it makes the code simpler.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Move the port device node structure into the port structure, from the
chip data. This information is needed in the next step of implementing
the new binding.
The chip data structure is used while parsing the whole old binding,
before the individual switch structures exist. With the new bindings,
this is reversed, the switches exist first, and the interconnections
between the switches is derived from the individual switch
bindings. Thus this chip data structure becomes unneeded.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
eviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There are going to be more per-port members added to the switch
structure. So add a port structure and move the netdev into it.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The platform data nr_chips is used when validating a received packet,
to ensure it comes from a know switch chip. The number of possible
switches is limited to DSA_MAX_SWITCHES, so use this as the first
validation step. The new binding allows holes in the dst->ds[] array,
so also ensure ensure there is a valid dsa_switch for this packet.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The DSA layer should no longer assume the switch is connected to an
MDIO bus. As a result, we cannot use the address on the MDIO bus when
forming the name of the switches internal MDIO bus for its builtin and
possibly external PHYs. The switch index is sufficient to make the
name unique, so drop the MDIO address.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The new binding does not make use of dsa_chip_data, a.k.a cd. When
retrieving the size of the EEPROM attached to a switch, don't assume
there is a cd attached to the switch structure.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
size_t objects should be printed with %Z printf format.
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit a91eb52abb ("qed: Revisit chain implementation") contains an
incorrect implementation for BE platforms, as device's regpairs containing
addresses are LE and they're not converted correctly when read back.
In addition, it raises a compilation warning for 32-bit platforms where
dma_addr_t is a 32-bit variable.
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Yuval Mintz says:
====================
qed: RocE & iSCSI infrastructure
We plan on sending 2 new protocol drivers in the imminent future -
both our RoCE [qedr] and iSCSI [qedi] drivers. As both submissions
would be rather massive and in order to avoid collisions between them,
the common infrastructure on the qed side was prepared as an independent
patch-series to be sent ahead of those 2 submissions.
This patch series introduces in QED 2 new 'ids' - one for iscsi and
one for roce. It then goes and adds logic required for configuring
said protocols in HW. Notice it *doesn't* actually add any client using
said ids, but rather only the infrastructure to allow their later usage.
What this patch doesn't contain is the slowpath protocol-configuration
toward the firmware. I.e., it contains register-setting logic, memory
allocations, etc., but not actual flow-related configuration specific
to the protocl. Those would be sent as part of the protocol driver
submissions.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
RoCE and iSCSI would require some added/changed hw configuration in order
to properly run; The biggest single change being the requirement of
allocating and mapping host memory for several HW blocks that aren't being
used by qede [SRC, QM, TM, etc.].
In addition, whereas qede is only using context memory for HW blocks, the
new protocol would also require task memories to be added.
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds in the ecore 2 new personalities in addition to
QED_PCI_ETH - QED_PCI_ISCSI and QED_PCI_ETH_ROCE.
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This adds the qed portion of the RoCE & iSCSI firmware HSI,
as well as adding several new common HSI files which would be required
by both qed and qed* protocols.
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
RoCE driver is going to need a 32-bit chain [current chain implementation
for qed* currently supports only 16-bit producer/consumer chains].
This patch adds said support, as well as doing other slight tweaks and
modifications to qed's chain API.
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There is no reason in this lock. At least for now.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronzhuk@linaro.org>
Reviewed-by: Grygorii Strashko <grygorii.strashko@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use the more common kernel logging style and reduce object size.
The logging message prefix changes from a mixture of
"RxRPC:" and "RXRPC:" to "af_rxrpc: ".
$ size net/rxrpc/built-in.o*
text data bss dec hex filename
64172 1972 8304 74448 122d0 net/rxrpc/built-in.o.new
67512 1972 8304 77788 12fdc net/rxrpc/built-in.o.old
Miscellanea:
o Consolidate the ASSERT macros to use a single pr_err call with
decimal and hexadecimal output and a stringified #OP argument
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Added a condition to avoid vlan devices with same MAC registering
as VF.
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Reviewed-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Marcelo Ricardo Leitner says:
====================
sctp: Add GSO support
This patchset adds sctp GSO support.
Performance tests indicates that increases throughput by 10% if using
bigger chunk sizes, specially if bigger than MTU. For small chunks, it
doesn't help much if not using heavy firewall rules.
For small chunks it will probably be of more use once we get something
like MSG_MORE as David Laight had suggested.
overall changes:
v1->v2:
Added support for receiving GSO frames on SCTP stack, as requested by
Dave Miller.
v2->v3:
Consider sctphdr size in skb_gso_transport_seglen()
rebased due to 5c7cdf339a ("gso: Remove arbitrary checks for
unsupported GSO")
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
This is useful for debugging packet sizes.
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Tested-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
SCTP has this pecualiarity that its packets cannot be just segmented to
(P)MTU. Its chunks must be contained in IP segments, padding respected.
So we can't just generate a big skb, set gso_size to the fragmentation
point and deliver it to IP layer.
This patch takes a different approach. SCTP will now build a skb as it
would be if it was received using GRO. That is, there will be a cover
skb with protocol headers and children ones containing the actual
segments, already segmented to a way that respects SCTP RFCs.
With that, we can tell skb_segment() to just split based on frag_list,
trusting its sizes are already in accordance.
This way SCTP can benefit from GSO and instead of passing several
packets through the stack, it can pass a single large packet.
v2:
- Added support for receiving GSO frames, as requested by Dave Miller.
- Clear skb->cb if packet is GSO (otherwise it's not used by SCTP)
- Added heuristics similar to what we have in TCP for not generating
single GSO packets that fills cwnd.
v3:
- consider sctphdr size in skb_gso_transport_seglen()
- rebased due to 5c7cdf339a ("gso: Remove arbitrary checks for
unsupported GSO")
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Tested-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch is a preparation for the GSO one. In order to successfully
handle GSO packets on rx path we must not call skb_linearize, otherwise
it defeats any gain GSO may have had.
This patch thus delays as much as possible the call to skb_linearize,
leaving it to sctp_inq_pop() moment. For that the sanity checks
performed now know how to deal with fragments.
One positive side-effect of this is that if the socket is backlogged it
will have the chance of doing it on backlog processing instead of
during softirq.
With this move, it's evident that a check for non-linearity in
sctp_inq_pop was ineffective and is now removed. Note that a similar
check is performed a bit below this one.
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Tested-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
skb_gso_network_seglen is not enough for checking fragment sizes if
skb is using GSO_BY_FRAGS as we have to check frag per frag.
This patch introduces skb_gso_validate_mtu, based on the former, which
will wrap the use case inside it as all calls to skb_gso_network_seglen
were to validate if it fits on a given TMU, and improve the check.
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Tested-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch allows segmenting a skb based on its frags sizes instead of
based on a fixed value.
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Tested-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
sctp GSO requires it and sctp can be compiled as a module, so we need to
export this function.
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Tested-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
NETIF_F_GSO_SOFTWARE was defined to list all GSO software types, so lets
make use of it in loopback code. Note that veth/vxlan/others already
uses it.
Within this patch series, this patch causes lo to pick up SCTP GSO feature
automatically (as it's added to NETIF_F_GSO_SOFTWARE) and thus avoiding
segmentation if possible.
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Tested-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
alloc_workqueue replaces deprecated create_workqueue().
The workqueue adapter->txrx_wq has workitem
&adapter->raise_intr_rxdata_task per adapter. Extended Socket Network
Device is shared memory based, so someone's transmission denotes other's
reception. raise_intr_rxdata_task raises interruption of receivers from
the sender in order to notify receivers.
The workqueue adapter->control_wq has workitem
&adapter->interrupt_watch_task per adapter. interrupt_watch_task is used
to prevent delay of interrupts.
Dedicated workqueues have been used in both cases since the workitems
on the workqueues are involved in normal device operation and require
forward progress under memory pressure.
max_active has been set to 0 since there is no need for throttling
the number of active work items.
Since network devices may be used for memory reclaim,
WQ_MEM_RECLAIM has been set to guarantee forward progress.
Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The New QED firmware contains several fixes, including:
- Wrong classification of packets in 4-port devices.
- Anti-spoof interoperability with encapsulated packets.
- Tx-switching of encapsulated packets.
It also slightly improves Tx performance of the device.
In addition, this firmware contains the necessary logic for
supporting iscsi & rdma, for which we plan on pushing protocol
drivers in the imminent future.
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The VRF device exists to define L3 domains and guide FIB lookups. As
such its operstate is not relevant. Seeing 'state UNKNOWN' in the
output of 'ip link show' can be confusing, so set operstate at link
create.
Similarly, the MTU for a VRF device is not used; any fragmentation
of the payload is done on the output path based on the real egress
device. An MTU of 1500 on the VRF device while enslaved devices
have a higher MTU can lead to confusion. Since the VRF MTU is not
relevant set to 64k similar to what is done for loopback.
Signed-off-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Set name_assign_type of internal port to NET_NAME_USER.
Signed-off-by: Zhang Shengju <zhangshengju@cmss.chinamobile.com>
Acked-by: Pravin B Shelar <pshelar@ovn.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
alloc_workqueue replaces deprecated create_workqueue().
A dedicated workqueue has been used since the workitems are involved
in normal device operation. Workitems &priv->rx_work and &priv->tx_work,
map to w5100_rx_work and w5100_tx_work respectively and are involved in
receiving and transmitting packets. Forward progress under
memory pressure is a requirement here.
create_workqueue has been replaced with alloc_workqueue with max_active
as 0 since there is no need for throttling the number of active work
items.
Since the driver may be used in memory reclaim path,
WQ_MEM_RECLAIM has been set to guarantee forward progress.
flush_workqueue is unnecessary since destroy_workqueue() itself calls
drain_workqueue() which flushes repeatedly till the workqueue
becomes empty. Hence the call to flush_workqueue() has been dropped.
Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Writing a non-zero value to the manual PAUSE frame register (MPR) starts
the transmission of a PAUSE frame.
A PAUSE frame is sent in ravb_emac_init(), but it is not expected behavior.
Signed-off-by: Masaru Nagai <masaru.nagai.vx@renesas.com>
Signed-off-by: Kazuya Mizuguchi <kazuya.mizuguchi.ks@renesas.com>
Signed-off-by: Yoshihiro Kaneko <ykaneko0929@gmail.com>
Acked-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There's not much point, except compile test, enabling the stmmac
platform drivers unless their actual SoC is enabled. They're not
useful without it.
Signed-off-by: Peter Robinson <pbrobinson@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Herbert Xu says:
====================
macvlan: Avoid unnecessary multicast cloning
This patch tries to improve macvlan multicast performance by
maintaining a filter hash at the macvlan_port level so that we
can quickly determine whether a given packet is needed or not.
It is preceded by a patch that fixes a potential use-after-free
bug that I discovered while looking over this.
v2 fixed a bug where promiscuous/allmulti settings weren't handled
correctly.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
