Update the reading of the Tx timestamp to account for a hardware issue
on how the fields and interrupt are cleared.  The "seconds" portion of
the timestamp should be read first, followed by the "nanoseconds" portion.
Reading the "nanoseconds" portion should clear the timestamp data and the
interrupt.  Because of an issue with the hardware this order is reversed
and reading the "seconds" portion actually clears the timestamp.  The code
currently follows this workaround, but to guard against future versions
where this is fixed add a field to the version data to indicate if the
workaround is required or not.
Signed-off-by: NTom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Due to a hardware issue, it is possible for interrupt events to be
incorrectly generated when performing a soft reset.  To guard against
this, perform the soft reset twice.
Signed-off-by: NTom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Jiri Benc says:

====================
openvswitch: support for layer 3 encapsulated packets

At the core of this patch set is removing the assumption in Open vSwitch
datapath that all packets have Ethernet header.

The implementation relies on the presence of pop_eth and push_eth actions
in datapath flows to facilitate adding and removing Ethernet headers as
appropriate. The construction of such flows is left up to user-space.

This series is based on work by Simon Horman, Lorand Jakab, Thomas Morin and
others. I kept Lorand's and Simon's s-o-b in the patches that are derived
from v11 to record their authorship of parts of the code.

Changes from v12 to v13:

* Addressed Pravin's feedback.
* Removed the GRE vport conversion patch; L3 GRE ports should be created by
  rtnetlink instead.

Main changes from v11 to v12:

* The patches were restructured and split differently for easier review.
* They were rebased and adjusted to the current net-next. Especially MPLS
  handling is different (and easier) thanks to the recent MPLS GSO rework.
* Several bugs were discovered and fixed. The most notable is fragment
  handling: header adjustment for ARPHRD_NONE devices on tx needs to be done
  after refragmentation, not before it. This required significant changes in
  the patchset. Another one is stricter checking of attributes (match on L2
  vs. L3 packet) at the kernel level.
* Instead of is_layer3 bool, a mac_proto field is used.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Allow ARPHRD_NONE interfaces to be added to ovs bridge.

Based on previous versions by Lorand Jakab and Simon Horman.
Signed-off-by: NLorand Jakab <lojakab@cisco.com>
Signed-off-by: NSimon Horman <simon.horman@netronome.com>
Signed-off-by: NJiri Benc <jbenc@redhat.com>
Acked-by: NPravin B Shelar <pshelar@ovn.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

It's not allowed to push Ethernet header in front of another Ethernet
header.

It's not allowed to pop Ethernet header if there's a vlan tag. This
preserves the invariant that L3 packet never has a vlan tag.

Based on previous versions by Lorand Jakab and Simon Horman.
Signed-off-by: NLorand Jakab <lojakab@cisco.com>
Signed-off-by: NSimon Horman <simon.horman@netronome.com>
Signed-off-by: NJiri Benc <jbenc@redhat.com>
Acked-by: NPravin B Shelar <pshelar@ovn.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Extend the ovs flow netlink protocol to support L3 packets. Packets without
OVS_KEY_ATTR_ETHERNET attribute specify L3 packets; for those, the
OVS_KEY_ATTR_ETHERTYPE attribute is mandatory.

Push/pop vlan actions are only supported for Ethernet packets.

Based on previous versions by Lorand Jakab and Simon Horman.
Signed-off-by: NLorand Jakab <lojakab@cisco.com>
Signed-off-by: NSimon Horman <simon.horman@netronome.com>
Signed-off-by: NJiri Benc <jbenc@redhat.com>
Acked-by: NPravin B Shelar <pshelar@ovn.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Support receiving, extracting flow key and sending of L3 packets (packets
without an Ethernet header).

Note that even after this patch, non-Ethernet interfaces are still not
allowed to be added to bridges. Similarly, netlink interface for sending and
receiving L3 packets to/from user space is not in place yet.

Based on previous versions by Lorand Jakab and Simon Horman.
Signed-off-by: NLorand Jakab <lojakab@cisco.com>
Signed-off-by: NSimon Horman <simon.horman@netronome.com>
Signed-off-by: NJiri Benc <jbenc@redhat.com>
Acked-by: NPravin B Shelar <pshelar@ovn.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Update Ethernet header only if there is one.
Signed-off-by: NJiri Benc <jbenc@redhat.com>
Acked-by: NPravin B Shelar <pshelar@ovn.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We'll need it to alter packets sent to ARPHRD_NONE interfaces.

Change do_output() to use the actual L2 header size of the packet when
deciding on the minimum cutlen. The assumption here is that what matters is
not the output interface hard_header_len but rather the L2 header of the
particular packet. For example, ARPHRD_NONE tunnels that encapsulate
Ethernet should get at least the Ethernet header.
Signed-off-by: NJiri Benc <jbenc@redhat.com>
Acked-by: NPravin B Shelar <pshelar@ovn.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Use a hole in the structure. We support only Ethernet so far and will add
a support for L2-less packets shortly. We could use a bool to indicate
whether the Ethernet header is present or not but the approach with the
mac_proto field is more generic and occupies the same number of bytes in the
struct, while allowing later extensibility. It also makes the code in the
next patches more self explaining.

It would be nice to use ARPHRD_ constants but those are u16 which would be
waste. Thus define our own constants.

Another upside of this is that we can overload this new field to also denote
whether the flow key is valid. This has the advantage that on
refragmentation, we don't have to reparse the packet but can rely on the
stored eth.type. This is especially important for the next patches in this
series - instead of adding another branch for L2-less packets before calling
ovs_fragment, we can just remove all those branches completely.
Signed-off-by: NJiri Benc <jbenc@redhat.com>
Acked-by: NPravin B Shelar <pshelar@ovn.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

On tx, use hard_header_len while deciding whether to refragment or drop the
packet. That way, all combinations are calculated correctly:

* L2 packet going to L2 interface (the L2 header len is subtracted),
* L2 packet going to L3 interface (the L2 header is included in the packet
  lenght),
* L3 packet going to L3 interface.
Signed-off-by: NJiri Benc <jbenc@redhat.com>
Acked-by: NPravin B Shelar <pshelar@ovn.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Commit 67f8b1dc ("net/mlx4_en: Refactor the XDP forwarding rings
scheme") added a bug in that the prog's reference count is not dropped
in the error path when mlx4_en_try_alloc_resources() is failing from
mlx4_xdp_set().

We previously took bpf_prog_add(prog, priv->rx_ring_num - 1), that we
need to release again. Earlier in the call path, dev_change_xdp_fd()
itself holds a reference to the prog as well (hence the '- 1' in the
bpf_prog_add()), so a simple atomic_sub() is safe to use here. When
an error is propagated, then bpf_prog_put() is called eventually from
dev_change_xdp_fd()

Fixes: 67f8b1dc ("net/mlx4_en: Refactor the XDP forwarding rings scheme")
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>
Acked-by: NAlexei Starovoitov <ast@kernel.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

While create/destroy channel operation memory is not freed. It was
supposed that memory is freed while driver remove. But a channel
can be created and destroyed many times while changing number of
channels with ethtool.

Based on net-next/master
Signed-off-by: NIvan Khoronzhuk <ivan.khoronzhuk@linaro.org>
Reviewed-by: NGrygorii Strashko <grygorii.strashko@ti.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Salil Mehta says:

====================
Bug fixes & Code improvements in HNS driver

This patch-set introduces some bug fixes and code improvements.
These have been identified during internal review or testing of
the driver by internal Hisilicon teams.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch adds the support to add or remove the unicast entries
to the table and remove from the table.
Reported-by: NDaode Huang <huangdaode@hisilicon.com>
Signed-off-by: NKejian Yan <yankejian@huawei.com>
Reviewed-by: NYisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: NSalil Mehta <salil.mehta@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

There is no clear operation before add a new multicast tcam table,
so the tcam table will be overflow when add more entries.
Reported-by: NDaode Huang <huangdaode@hisilicon.com>
Signed-off-by: NKejian Yan <yankejian@huawei.com>
Reviewed-by: NYisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: NSalil Mehta <salil.mehta@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The packets of wrong mac address(only the last bit is different) can be
received in Big-endian by current definition of mask_key. Thus it needs
to be modified to support Big-endian and ensure Big-endian normal.
Signed-off-by: NQianqian Xie <xieqianqian@huawei.com>
Reviewed-by: NYisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: NSalil Mehta <salil.mehta@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The current definition of mac_mc_entry is only suitable for
Little-endian. Thus it needs to modify tcam table of mac mc-entry
to support both Little-endian and Big-endian.
Signed-off-by: NQianqian Xie <xieqianqian@huawei.com>
Reviewed-by: NYisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: NSalil Mehta <salil.mehta@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Little-endian is only supported by current tcam table to add
or delete mac mc-port. This patch makes it support both
Little-endian and Big-endian.
Signed-off-by: NQianqian Xie <xieqianqian@huawei.com>
Reviewed-by: NYisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: NSalil Mehta <salil.mehta@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Big-endian is not supported by the current definition of table index to get
mac entry. It needs to be modified to support both Little-endian
and Big-endian.
Signed-off-by: NQianqian Xie <xieqianqian@huawei.com>
Reviewed-by: NYisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: NSalil Mehta <salil.mehta@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The current definition of mac_uc_entry is only suitable for
Little-endian. Thus it needs to modify tcam table of mac uc-entry
to support both Little-endian and Big-endian.
Signed-off-by: NQianqian Xie <xieqianqian@huawei.com>
Reviewed-by: NYisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: NSalil Mehta <salil.mehta@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The current definition of dsaf_drv_tbl_tcam_key is only suitable for
Little-endian. If data is stored in Big-endian, this may lead to
error in data use. Shift operation can make it work normally in both
Big-endian and Little-endian.
Signed-off-by: NQianqian Xie <xieqianqian@huawei.com>
Reviewed-by: NYisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: NSalil Mehta <salil.mehta@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Hardware ring buffer data is stored in Little-endian. Thus cpu data
should be modified to Little-endian.
Signed-off-by: NQianqian Xie <xieqianqian@huawei.com>
Reviewed-by: NYisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: NSalil Mehta <salil.mehta@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In current scenario, when the interface is disabled we reset the XGMAC
RX/TX functionality. This operation does not affects the PHY layer/SFP
and which appears UP to the remote end(this behaviour is unlike GMAC).
The result is remote end keeps on sending the packets which gets partly
processed by XMAC and dropped. Since these are partly processed these
appears as errored packets in the packet counter statistics.

This patch fixes this behaviour and adds local-fault and remote-fault
functionality which can be used to intimate the remote peer whenever
the state of the interface changes. This patch also removes the
existing hns_dsaf_xge_core_srst_by_port function which was being used
to reset the RX/TX functionality at XGE Core.
Reported-by: NJun He <hjat2005@huawei.com>
Signed-off-by: NDaode Huang <huangdaode@hisilicon.com>
Reviewed-by: NYisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: NSalil Mehta <salil.mehta@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch modify the gmac_rx_filt_pkt and gmac_rx_octets_total_filt
statistics value. The two statistics is inconsistent with register,
and just the opposite.
Signed-off-by: NQianqian Xie <xieqianqian@huawei.com>
Signed-off-by: NJun He <hjat2005@huawei.com>
Reviewed-by: NYisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: NSalil Mehta <salil.mehta@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch deletes redundant macro definitions in hns drivers.
And change the .h file containing relation to make the layers
more clearly
Signed-off-by: NQianqian Xie <xieqianqian@huawei.com>
Signed-off-by: NWeiwei Deng <dengweiwei@huawei.com>
Reviewed-by: NYisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: NSalil Mehta <salil.mehta@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When set auto-negotiation off and duplex half, if run "ethtool -r ethX"
on port with phy, then the port will be failed to work. It should
forbid to start auto-negotiation when auto-negotiate is off. This
patch add the limited condition.
Reported-by: NJinchuang Tian <tianjinchuang1@huawei.com>
Signed-off-by: NDaode Huang <huangdaode@hisilicon.com>
Reviewed-by: NYisen Zhuang <yisen.zhuang@huawei.com>
Reviewed-by: Nlipeng <lipeng321@huawei.com>
Signed-off-by: NSalil Mehta <salil.mehta@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The default mac pause time set to 0xff which is too short for pausing,
this patch change it to the max value 0xffff.
Signed-off-by: NDaode Huang <huangdaode@hisilicon.com>
Reviewed-by: NYisen Zhuang <yisen.zhuang@huawei.com>
Reviewed-by: Nlipeng <lipeng321@huawei.com>
Signed-off-by: NSalil Mehta <salil.mehta@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If set promisc mode when there is some traffic, The service nic will
cause system halted. We reserve the last 6 tcam entry for the 6 ports.
If promisc mode is enabled, we can config the relative tcam as fuzzy
matching and set to be valid, or set the tcam to be invalid
Signed-off-by: NKejian Yan <yankejian@huawei.com>
Reviewed-by: NYisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: NSalil Mehta <salil.mehta@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Since there is not enough tcam table entries for vlan and multicast
address, HNSv2 needs to add support of fuzzy matching of TCAM tables.
To add fuzzy match of TCAM, we Add the property to mask the bits to
be fuzzy matched
Signed-off-by: NKejian Yan <yankejian@huawei.com>
Reviewed-by: NYisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: NSalil Mehta <salil.mehta@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Since there is not enough tcam table entries for every vlan and multicast
address, HNS needs to add support of fuzzy matching of TCAM tables. Adding
the property to mask the bits to be fuzzy matched, so update the bindings
document
Signed-off-by: NKejian Yan <yankejian@huawei.com>
Reviewed-by: NYisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: NSalil Mehta <salil.mehta@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

After commit 6ed46d12 ("sock_diag: align nlattr properly when
needed"), tcp_get_info() gets 64bit aligned memory, so we can avoid
the unaligned helpers.
Suggested-by: NDavid Miller <davem@davemloft.net>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Acked-by: NSoheil Hassas Yeganeh <soheil@google.com>
Acked-by: NYuchung Cheng <ycheng@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Simon Wunderlich says:

====================
pull request for net-next: batman-adv 2016-11-08 v2

This feature and cleanup patchset includes the following changes:

 - netlink and code cleanups by Sven Eckelmann (3 patches)

 - Cleanup and minor fixes by Linus Luessing (3 patches)

 - Speed up multicast update intervals, by Linus Luessing

 - Avoid (re)broadcast in meshes for some easy cases,
   by Linus Luessing

 - Clean up tx return state handling, by Sven Eckelmann (6 patches)

 - Fix some special mac address handling cases, by Sven Eckelmann
   (3 patches)
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Richard Cochran says:

====================
PHC frequency fine tuning

This series expands the PTP Hardware Clock subsystem by adding a
method that passes the frequency tuning word to the the drivers
without dropping the low order bits.  Keeping those bits is useful for
drivers whose frequency resolution is higher than 1 ppb.

The appended script (below) runs a simple demonstration of the
improvement.  This test needs two Intel i210 PCIe cards installed in
the same PC, with their SDP0 pins connected by copper wire.  Measuring
the estimated offset (from the ptp4l servo) and the true offset (from
the PPS) over one hour yields the following statistics.

|        |   Est. Before |    Est. After |   True Before |    True After |
|--------+---------------+---------------+---------------+---------------|
| min    | -5.200000e+01 | -1.600000e+01 | -3.100000e+01 | -1.000000e+00 |
| max    | +5.700000e+01 | +2.500000e+01 | +8.500000e+01 | +4.000000e+01 |
| pk-pk: | +1.090000e+02 | +4.100000e+01 | +1.160000e+02 | +4.100000e+01 |
| mean   | +6.472222e-02 | +1.277778e-02 | +2.422083e+01 | +1.826083e+01 |
| stddev | +1.158006e+01 | +4.581982e+00 | +1.207708e+01 | +4.981435e+00 |

Here the numbers in units of nanoseconds, and the ~20 nanosecond PPS
offset is due to input/output delays on the i210's external interface
logic.

With the series applied, both the peak to peak error and the standard
deviation improve by a factor of more than two.  These two graphs show
the improvement nicely.

  http://linuxptp.sourceforge.net/fine-tuning/fine-est.png

  http://linuxptp.sourceforge.net/fine-tuning/fine-tru.png
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The dp83640 has a frequency resolution of about 0.029 ppb.
This patch lets users of the device benefit from the
increased frequency resolution when tuning the clock.
Signed-off-by: NRichard Cochran <richardcochran@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The 82580 and related devices offer a frequency resolution of about
0.029 ppb.  This patch lets users of the device benefit from the
increased frequency resolution when tuning the clock.
Signed-off-by: NRichard Cochran <richardcochran@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The internal PTP Hardware Clock (PHC) interface limits the resolution for
frequency adjustments to one part per billion.  However, some hardware
devices allow finer adjustment, and making use of the increased resolution
improves synchronization measurably on such devices.

This patch adds an alternative method that allows finer frequency tuning
by passing the scaled ppm value to PHC drivers.  This value comes from
user space, and it has a resolution of about 0.015 ppb.  We also deprecate
the older method, anticipating its removal once existing drivers have been
converted over.
Signed-off-by: NRichard Cochran <richardcochran@gmail.com>
Suggested-by: NUlrik De Bie <ulrik.debie-os@e2big.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

There are no more users except from net/core/dev.c
napi_hash_add() can now be static.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Cc: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This is automatically done from netif_napi_add(), and we want to not
export napi_hash_add() anymore in the following patch.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Cc: Michael Chan <michael.chan@broadcom.com>
Acked-by: NMichael Chan <michael.chan@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Remove the unused but set variables min_set and max_set in
adjust_reg_min_max_vals to fix the following warning when building with
'W=1':

  kernel/bpf/verifier.c:1483:7: warning: variable ‘min_set’ set but not used [-Wunused-but-set-variable]

There is no warning about max_set being unused, but since it is only
used in the assignment of min_set it can be removed as well.

They were introduced in commit 48461135 ("bpf: allow access into map
value arrays") but seem to have never been used.

Cc: Josef Bacik <jbacik@fb.com>
Signed-off-by: NTobias Klauser <tklauser@distanz.ch>
Acked-by: NAlexei Starovoitov <ast@kernel.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>