Add test to check Q-in-VNI traffic with IPv6 underlay and overlay.
The test is similar to the existing IPv4 test.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

In a similar fashion to the asymmetric test, add a test for symmetric
routing. In symmetric routing both the ingress and egress VTEPs perform
routing in the overlay network into / from the VxLAN tunnel. Packets in
different directions use the same VNI - the L3 VNI.
Different tenants (VRFs) use different L3 VNIs.

Add a test which is similar to the existing IPv4 test to check IPv6.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

In asymmetric routing the ingress VTEP routes the packet into the
correct VxLAN tunnel, whereas the egress VTEP only bridges the packet to
the correct host. Therefore, packets in different directions use
different VNIs - the target VNI.

Add a test which is similar to the existing IPv4 test to check IPv6.

The test uses a simple topology with two VTEPs and two VNIs and verifies
that ping passes between hosts (local / remote) in the same VLAN (VNI)
and in different VLANs belonging to the same tenant (VRF).

While the test does not check VM mobility, it does configure an anycast
gateway using a macvlan device on both VTEPs.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Remove `vxlan_ping_test()` which is not used and probably was copied
mistakenly from vxlan_bridge_1d.sh.

This was found while adding an equivalent test for IPv6.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

The tests are very similar to their VLAN-unaware counterpart
(vxlan_bridge_1d_ipv6.sh and vxlan_bridge_1d_port_8472_ipv6.sh),
but instead of using multiple VLAN-unaware bridges, a single VLAN-aware
bridge is used with multiple VLANs.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Add tests similar to vxlan_bridge_1d.sh and vxlan_bridge_1d_port_8472.sh.

The tests set up a topology with three VxLAN endpoints: one
"local", possibly offloaded, and two "remote", formed using veth pairs
and likely purely software bridges. The "local" endpoint is connected to
host systems by a VLAN-unaware bridge.

Since VxLAN tunnels must be unique per namespace, each of the "remote"
endpoints is in its own namespace. H3 forms the bridge between the three
domains.

Send IPv4 packets and IPv6 packets with IPv6 underlay.
Use `TC_FLAG`, which is defined in `forwarding.config` file, for TC
checks. `TC_FLAG` allows testing that on HW datapath, the traffic
actually goes through HW.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Currently `ping_do()` and `ping6_do()` send 10 packets.

There are cases that it is not possible to catch only the interesting
packets using tc rule, so then, it is possible to send many packets and
verify that at least this amount of packets hit the rule.

Add `PING_COUNT` variable, which is set to 10 by default, to allow tests
sending more than 10 packets using the existing ping API.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Spectrum ASICs do not support matching of VLAN ID at egress.
Currently, mlxsw driver forbids matching of all VLAN related fields at
egress, which is too strict check.

For example, the following filter is not supported by the driver:
$ tc filter add dev swpX egress protocol 802.1q pref 1 handle 101 flower
vlan_ethtype ipv4 src_ip .. dst_ip .. skip_sw action pass
Error: mlxsw_spectrum: vlan_id key is not supported on egress.
We have an error talking to the kernel

The filter above does not match on VLAN ID, but is bounced anyway.

Make the check more specific, forbid only matching of 'vlan_id' at egress.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Saeed Mahameed says:

====================
mlx5-updates-2021-12-21

1) From Shay Drory: Devlink user knobs to control device's EQ size

This series provides knobs which will enable users to
minimize memory consumption of mlx5 Functions (PF/VF/SF).
mlx5 exposes two new generic devlink params for EQ size
configuration and uses devlink generic param max_macs.

LINK: https://lore.kernel.org/netdev/20211208141722.13646-1-shayd@nvidia.com/

2) From Tariq and Lama, allocate software channel objects and statistics
  of a mlx5 netdevice private data dynamically upon first demand to save on
  memory.

* tag 'mlx5-updates-2021-12-21' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux:
  net/mlx5e: Take packet_merge params directly from the RX res struct
  net/mlx5e: Allocate per-channel stats dynamically at first usage
  net/mlx5e: Use dynamic per-channel allocations in stats
  net/mlx5e: Allow profile-specific limitation on max num of channels
  net/mlx5e: Save memory by using dynamic allocation in netdev priv
  net/mlx5e: Add profile indications for PTP and QOS HTB features
  net/mlx5e: Use bitmap field for profile features
  net/mlx5: Remove the repeated declaration
  net/mlx5: Let user configure max_macs generic param
  devlink: Clarifies max_macs generic devlink param
  net/mlx5: Let user configure event_eq_size param
  devlink: Add new "event_eq_size" generic device param
  net/mlx5: Let user configure io_eq_size param
  devlink: Add new "io_eq_size" generic device param
====================

Link: https://lore.kernel.org/r/20211222031604.14540-1-saeed@kernel.orgSigned-off-by: Jakub Kicinski <kuba@kernel.org>

Commit d068ca2a ("codel: split into multiple files") moved all
Qdisc-related code to codel_qdisc.h, move the include of pkt_sched.h
as well.

This is similar to the previous commit, although we don't care as
much about incremental builds after pkt_sched.h was touched itself
it is included by net/sch_generic.h which is modified ~20 times
a year.

This decreases the incremental build size after touching pkt_sched.h
from 1592 to 617 objects.

Fix unmasked missing includes in WiFi drivers.
Acked-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20211221193941.3805147-2-kuba@kernel.orgSigned-off-by: Jakub Kicinski <kuba@kernel.org>

Since sock.h is modified relatively often (60 times in the last
12 months) it seems worthwhile to decrease the incremental build
work.

CoDel's header includes net/inet_ecn.h which in turn includes net/sock.h.
codel.h is itself included by mac80211 which is included by much of
the WiFi stack and drivers. Removing the net/inet_ecn.h include from
CoDel breaks the dependecy between WiFi and sock.h.

Commit d068ca2a ("codel: split into multiple files") moved all
the code which actually needs ECN helpers out to net/codel_impl.h,
the include can be moved there as well.

This decreases the incremental build size after touching sock.h
from 4999 objects to 4051 objects.

Fix unmasked missing includes in WiFi drivers.
Acked-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20211221193941.3805147-1-kuba@kernel.orgSigned-off-by: Jakub Kicinski <kuba@kernel.org>

The variable bytes is being used to summate slot lengths,
however the value is never used afterwards. The summation
is redundant so remove variable bytes.
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Link: https://lore.kernel.org/r/20211222003937.727325-1-colin.i.king@gmail.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

Fix an odd indent where some code was left indented, and causes smatch
to warn:
ice_log_pkg_init() warn: inconsistent indenting

While here, for consistency, add a break after the default case.

This commit has a Fixes: but we caught this while it was only in net-next.

Fixes: 247dd97d ("ice: Refactor status flow for DDP load")
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Link: https://lore.kernel.org/r/20211221230538.2546315-1-jesse.brandeburg@intel.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

Tony Nguyen says:

====================
100GbE Intel Wired LAN Driver Updates 2021-12-21

This series contains updates to ice driver only.

Karol modifies the reset flow to correct issues with PTP reset.

Jake extends PTP support for E822 based devices. This includes a few
cleanup patches, that fix some minor issues. In addition, there are some
slight refactors to ease the addition of E822 support, followed by adding
the new hardware implementation ice_ptp_hw.c.

There are a few major differences with E822 support compared to E810
support:

*) The E822 device has a Clock Generation Unit which must be initialized in
order to generate proper clock frequencies on the output that drives the PTP
hardware clock registers

*) The E822 PHY is a bit different and requires a more complex
initialization procedure which must be rerun any time the link configuration
changes.

*) The E822 devices support enhanced timestamp calibration by making use of
a process called Vernier offset measurement. This allows the hardware to
measure phase offset related to the PHY clocks for Serdes and FEC, reducing
the inaccuracy of the timestamp relative to the actual packet transmission
and receipt. Making use of this requires data gathered from the first
transmitted and received packets, and waiting for the PHY to complete the
calibration measurements. This is done as part of a new kthread, ov_work.
Note that to avoid delay in enabling timestamps, we start the PHY in
'bypass' mode which allows timestamps to be captured without the Vernier
calibration measurement. Once the first packets have been sent and received,
we then complete the calibration setup and exit bypass mode and begin using
the more precise timestamps. According to the datasheet, timestamps without
calibration data can be incorrect relative to actual receipt or transmission
by up to 1 clock cycle (~1.25 nanoseconds), while calibrated timestamps
should be correct to within 1/8th of a clock cycle (~0.15 nanoseconds).

*) E822 devices support crosstimestamping via PCIe PTM, which we enable when
available on the platform.

There is a fair amount of logic required to perform PHY and CGU
initialization, which is the vast majority of the new code, but it is fairly
self contained within ice_ptp_hw.c, with the exception of monitoring for
offset validity being handled by a kthread.

* '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/next-queue:
  ice: support crosstimestamping on E822 devices if supported
  ice: exit bypass mode once hardware finishes timestamp calibration
  ice: ensure the hardware Clock Generation Unit is configured
  ice: implement basic E822 PTP support
  ice: convert clk_freq capability into time_ref
  ice: introduce ice_ptp_init_phc function
  ice: use 'int err' instead of 'int status' in ice_ptp_hw.c
  ice: PTP: move setting of tstamp_config
  ice: introduce ice_base_incval function
  ice: Fix E810 PTP reset flow
====================

Link: https://lore.kernel.org/r/20211221174845.3063640-1-anthony.l.nguyen@intel.comAcked-by: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

As packet_merge params structure is saved on the RX resources structure, there
is no need to pass it separately.
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Make stats allocation per-channel dynamic on demand, at channel open
operation.

Previously the stats array was pre-allocated for the maximum possible
number of channels. Here we defer the per-channel stats instance allocation
upon its first usage, so that it's allocated only if really needed.

Allocating stats on demand helps maintain a more memory-efficient code,
as we're saving memory when the used number of channels is smaller than
the maximum.

The stats memory instances are still freed in mlx5e_priv_arrays_free(),
so that they are persistent to channels' closure.

Memory size allocated for struct mlx5e_channel_stats is 3648 bytes.
If maximum number of channel stands for 64, the total memory space
allocated for stats is 3648x64 = 228K bytes. In scenarios where the
number of channels in use is significantly smaller than maximum number,
the memory saved can be remarkable.
Signed-off-by: Lama Kayal <lkayal@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Make stats array an array of pointer. This patch comes in to prepare for
the next patch where allocations of the stats are to be performed
dynamically on first usage.
Signed-off-by: Lama Kayal <lkayal@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Let SF/VF representor's netdev use profile-specific limitation on
max_nch to reduce its memory and HW resources consumption.

This is particularly important for environments with limited memory
and high number of SFs.
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Vu Pham <vuhuong@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Parav Pandit <parav@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Many arrays in priv are statically allocated with a pre-defined maximum
(for num channels, num TCs, etc...), that is in some cases significantly
larger than the actual maximum. Examples:
- The more VFs are supported, the less MSIX vectors each of them could
  have. This limits the max_nch for each.
- Systems with limited number of cores or MSIX (< 64).
- Netdev profiles that do not support: QoS (DCB / HTB), PTP TX port
  timestamping.

Here we save some amount of memory by moving several structures
and arrays to follow the actual maximum instead.
This patch also prepares the code for even more savings to follow.

For example, on a system where the maximum num of channel is 8,
the channels stats structs alone go down from 3648*64 = 228 KB to
3648*8 = 28.5 KB per interface.

This is important for environments with high number of VFs/SFs or
limited memory.
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Parav Pandit <parav@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Let the profile indicate support of the PTP and HTB (QOS) features.
This unifies the logic that calculates the number of netdev queues needed
for the features, and allows simplification of mlx5e_create_netdev(),
which no longer requires number of rx/tx queues as parameters.
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Aya Levin <ayal@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Use a features bitmap field in mlx5e_profile to declare profile support
state of the different features.  Let it replace the existing
rx_ptp_support boolean. It will be extended to cover more features in a
downstream patch.
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Function 'mlx5_esw_vport_match_metadata_supported' and
'mlx5_esw_offloads_vport_metadata_set' are declared twice, so remove
the repeated declaration and blank line.

Cc: Saeed Mahameed <saeedm@nvidia.com>
Cc: Leon Romanovsky <leon@kernel.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Shaokun Zhang <zhangshaokun@hisilicon.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Currently, max_macs is taking 70Kbytes of memory per function. This
size is not needed in all use cases, and is critical with large scale.
Hence, allow user to configure the number of max_macs.

For example, to reduce the number of max_macs to 1, execute::
$ devlink dev param set pci/0000:00:0b.0 name max_macs value 1 \
              cmode driverinit
$ devlink dev reload pci/0000:00:0b.0
Signed-off-by: Shay Drory <shayd@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Parav Pandit <parav@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

The generic param max_macs documentation isn't clear.
Replace it with a more descriptive documentation
Signed-off-by: Shay Drory <shayd@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Event EQ is an EQ which received the notification of almost all the
events generated by the NIC.
Currently, each event EQ is taking 512KB of memory. This size is not
needed in most use cases, and is critical with large scale. Hence,
allow user to configure the size of the event EQ.

For example to reduce event EQ size to 64, execute::
$ devlink dev param set pci/0000:00:0b.0 name event_eq_size value 64 \
              cmode driverinit
$ devlink dev reload pci/0000:00:0b.0
Signed-off-by: Shay Drory <shayd@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Add new device generic parameter to determine the size of the
asynchronous control events EQ.

For example, to reduce event EQ size to 64, execute:
$ devlink dev param set pci/0000:06:00.0 \
              name event_eq_size value 64 cmode driverinit
$ devlink dev reload pci/0000:06:00.0
Signed-off-by: Shay Drory <shayd@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Currently, each I/O EQ is taking 128KB of memory. This size
is not needed in all use cases, and is critical with large scale.
Hence, allow user to configure the size of I/O EQs.

For example, to reduce I/O EQ size to 64, execute:
$ devlink dev param set pci/0000:00:0b.0 name io_eq_size value 64 \
              cmode driverinit
$ devlink dev reload pci/0000:00:0b.0
Signed-off-by: Shay Drory <shayd@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Add new device generic parameter to determine the size of the
I/O completion EQs.

For example, to reduce I/O EQ size to 64, execute:
$ devlink dev param set pci/0000:06:00.0 \
              name io_eq_size value 64 cmode driverinit
$ devlink dev reload pci/0000:06:00.0
Signed-off-by: Shay Drory <shayd@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Tony Nguyen says:

====================
1GbE Intel Wired LAN Driver Updates 2021-12-21

This series contains updates to igc, igb, igbvf, and fm10k drivers.

Sasha removes unused defines and enum values from igc driver.

Jason Wang removes a variable whose value never changes and, instead,
returns the value directly for igb.

Karen adjusts a reset message from warning to info for igbvf.

Xiang wangx removes a repeated word for fm10k.

* '1GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/next-queue:
  fm10k: Fix syntax errors in comments
  igbvf: Refactor trace
  igb: remove never changed variable `ret_val'
  igc: Remove obsolete define
  igc: Remove obsolete mask
  igc: Remove obsolete nvm type
  igc: Remove unused phy type
  igc: Remove unused _I_PHY_ID define
====================

Link: https://lore.kernel.org/r/20211221180200.3176851-1-anthony.l.nguyen@intel.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

This patch add support for Link up or Link down
interrupt support in LAN8814 Quad phy
Signed-off-by: Divya Koppera <Divya.Koppera@microchip.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/20211221112217.9502-1-Divya.Koppera@microchip.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

Delete the redundant word 'by'.
Signed-off-by: Xiang wangx <wangxiang@cdjrlc.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

Refactoring "PF still resetting" message, because previous version looked
like a bug - it informed about changes that worked as designed but might
confuse users. Changes requested to make message more user-friendly.
Signed-off-by: Karen Sornek <karen.sornek@intel.com>
Tested-by: Tony Brelinski <tony.brelinski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

The variable used for return status in `igb_write_xmdio_reg' function
is never changed  and this function is just need return 0. Thus, the
`ret_val' can be removed and return 0 at the end of the
`igb_write_xmdio_reg' function.
Signed-off-by: Jason Wang <wangborong@cdjrlc.com>
Tested-by: Gurucharan G <gurucharanx.g@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

'MII_CR_FULL_DUPLEX' define not in use. This patch comes to tidy up
 obsolete define.
Signed-off-by: Sasha Neftin <sasha.neftin@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

'IGC_CTRL_EXT_LINK_MODE_MASK' not in use. This patch comes to tidy up
obsolete define.
Signed-off-by: Sasha Neftin <sasha.neftin@intel.com>
Tested-by: Nechama Kraus <nechamax.kraus@linux.intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

i225 devices use only spi nvm type. This patch comes to tidy up
obsolete nvm types.
Signed-off-by: Sasha Neftin <sasha.neftin@intel.com>
Tested-by: Nechama Kraus <nechamax.kraus@linux.intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

_phy_none type not in use. Clean up the code accordingly,
and get rid of the unused enum line
Signed-off-by: Sasha Neftin <sasha.neftin@intel.com>
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Tested-by: Nechama Kraus <nechamax.kraus@linux.intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

_I_PHY_ID not in use. Clean up the code accordingly,
and get rid of the unused define
Signed-off-by: Sasha Neftin <sasha.neftin@intel.com>
Tested-by: Nechama Kraus <nechamax.kraus@linux.intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

E822 devices on supported platforms can generate a cross timestamp
between the platform ART and the device time. This process allows for
very precise measurement of the difference between the PTP hardware
clock and the platform time.

This is only supported if we know the TSC frequency relative to ART, so
we do not enable this unless the boot CPU has a known TSC frequency (as
required by convert_art_ns_to_tsc).

Because PCIe PTM support is not available on all platforms, introduce
CONFIG_ICE_HWTS and make it depend on X86 where we know the support
exists.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Gurucharan G <gurucharanx.g@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>

Once the E822 device has sent and received one packet, the hardware
computes the internal delay of the PHY using a process known as Vernier
calibration. This calibration calculates a more accurate offset for the
Tx and Rx timestamps. To make use of this offset, we need to exit the
bypass mode. This cannot be done until the PHY has completed offset
calibration, as indicated by the offset valid bits.

To handle this, introduce a kthread work item which will poll the offset
valid bits every few milliseconds seeing if it is safe to exit bypass
mode.

Once we have finished calibrating the offsets, we can program the total
Tx and Rx offset registers and turn off the bypass bit. This allows the
hardware to include the more precise vernier calibration offset, and
improves the timestamp precision.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Gurucharan G <gurucharanx.g@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>