IT should be changed to non-HSR ports only forward non-HSR frames.
HSR tag in PTP frame must be removed in final version (this is just for
debugging)

Forcing SWO from SW.
Next step is using comma symbols.

Correction field for P2P TC/HY now follows the PTP standard (13.3.2.7).

In common-fun.c, all pdelay_request/respond do not check parent, this is
for some stuff related to HSR ring topologies. It will be fixed in the
future.

Fixing pmaster state machine after LOCK

It now forwards from non-HSR external nodes to internal HSR nodes and
also from HSR nodes to HSR nodes.

When a slave is connected to the ring, it would receive two PTP frames,
it only receives now one.
It is done at PTP level since it is not nowadays at GW level.
Wr-servo.c for TC exit has been fixed (it was exiting at a wrong line)

When getting out of the ring part of the ETH HEADER was missing. Not
fixed.

Adding TC behaviour. WRS can be configured as a transparent clock in
ppsi.conf ("transpclock [enabled|disabled]"). In case it is enabled, the
WRS sintonizes but it never applies PTP to the servo.
Missing: it should be done with clock-class and not in new line. And also variables from the standard probably.

Minor changes to use read values for forwarding and HSR from ppsi.conf.

HSR protocol and forwarding capabilities are now set in ppsi.conf file.
Variables added to globals, pp_argname, etc.
Conf.c now supports this configuration:
"wr-hsr [enabled|disabled]"
"forwarding [enabled|disabled]"

wrs-socket.c send function must take care whether a frame comes from a
HSR node or not. In addition, peerdelay messages also include HSR tags
now.

HSR sequence number must be called after forwarding in all ports the
same message, not when sending in wrs-socket.c. (now increased in
common-fun.c

pp-instance.h: adding is_HSR field.
forwarding: now it also copies ppi->rx_ptp, if not, the conversion doesn't work (???)
unix-socket: it accepts all type of packets (we need both PTP and HSR frames)
wrs-socket: duplication of wrs_net_send, if for HSR and else for non-HSR (GM)

Porting jlgutierrez-road2hsr-150427 here.
- Missing? When outer master node sends non-HSR PTP, first node in ring
shall consume, add HSR tag and forward to the HSR0&HSR1 ports.
Also check the sender to drop it from the network

We need to know whether a WRS is conected to a master outside the ring
to start the backup<->backup synchronization

I need to know which backup port is really active.

TRACK PHASE in both WRS are working correctly, but PPSs are doing
whatever they want.
It seems it is something related with "current_slave". I think I have to
develop something like "first slave" to know who is the really first
slave...

+ backup-backup works in tree topology
- when new linkUp, PPS is gone, but clock is synchronized.

Two slave/backup ports shall exchange WR messages. Trying to make this
work

Modifying switchover's ppsi to use ppsi and p2p. This is a temporal
branch...

[PSU] added functions to directly control PSU -- this needs to be done nicer, for the time being it works

master port on GM should copy DefaultDS.clockQuality (e.g. clockClass)
to ParentDS.grandmaster.clockQuality so that it is send out in Announce
messages. BC should not do that, as it copies clockQuality received
on the slave port into the ParentDS.grandmaster.ClockQuality and sends
out on its master ports this info.

presumably-bug commented out: some-kind of holdover degradation of ClockClass done in very bad place

The degradation of class should be managed more globally, is specific
to holdover (which is now being implemented) and should not inform about
unlocking if no holdover implemented. Will be remmoved once holdover
works.

m1() was used always when entering MASTER on Announce timeout - this
caused  "comming up" port (which usually enters MASTER) to use m1()
to overwrite the parentDS->grandmasterClockClass with its local
ClockClass. This should not happen on BC which is already synchronize.

instead of GM clockQuality info, the information from Default DS
was sent in announce. Corrected according to Table 25, p129 of PTP

bugfix: the classClass of a current GM was overwriten by the default clockClass each time new port went up

the bmc data copying functions (e.g. m1(), s1()) should be called only as a result of BMCA. This is rather irrelevant for WR, but important
for PTP networks

bugfix of multi-backup servo: now the active port is updated only in precise moments so if we update the phase of the backup, it will finish the process before switchover is registered