The unix time and network operations are going to be used as a fall-back
solution by other timing engines.

For example, hardware-specific timestamping capabilities can be
selected (such as the Linux ioctl API) but if the Ethernet interface
(and thus the pp_instance) doesn't support the feature, related code
will be able to set back the unix operations in place.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

This fixes the correct behaviour of ppsi in arch-wrpc, where adjust_freq
is not available.
Still some more fixes may be needed in servo.c. I'm not sure about the
parameter -adj to be passed both in adjust_freq and adjust_offset case.
Anyway, this has been tested on two SPECs and reaches synchronization.
Signed-off-by: Aurelio Colosimo <aurelio@aureliocolosimo.it>

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

rt_opts parameter was always passed as NULL. With this patch, users who do not
want default rt_opts provided by proto-standard need to set their rt_opts
explicitly.
Signed-off-by: Aurelio Colosimo <aurelio@aureliocolosimo.it>

Command line now acts globally. If multilink context (ppsi.conf defined and
cointaining more than one link), the behaviour of -g and -e flags is common
to each link. On the other way, -b flag can not be used in this case.
Signed-off-by: Aurelio Colosimo <aurelio@aureliocolosimo.it>

GDSDEF macro returns the defaultDS field of struct pp_globals. Actually,
it is the same pointer returned by DSDEF, since the default dataset is unique.
The DSDEF macro is not removed both to preserve commit history and to increase
readability (we do not take ppg pointer around and DSDEF(ppi) is shorter
than GDSDEF(GLOBS(ppi))).
Signed-off-by: Aurelio Colosimo <aurelio@aureliocolosimo.it>

GOPTS macro returns the rt_opts field of struct pp_globals. Actually,
in the current implementation it is the same pointer returned by OPTS, since
each pp_instance has no "local" rt_opt, but points to the global one.
Signed-off-by: Aurelio Colosimo <aurelio@aureliocolosimo.it>

New function receives a struct pp_globals, coherently with pp_open_globals.
Signed-off-by: Aurelio Colosimo <aurelio@aureliocolosimo.it>

pp_open_instance received a struct pp_globals, so rename it coherently.
We have no pp_open_instance alone because an instance can live only inside its
pp_globals struct.
Signed-off-by: Aurelio Colosimo <aurelio@aureliocolosimo.it>

pp_parse_conf will receive the txt data read from conf file and will compile
the ppg->links structs. File open is supposed to be done outside the proper
parser here defined, so that any architecture, even those with no filesystems,
can make use of the parser, supposing the configuration txt data is saved
elsewhere (e.g. in a E2PROM)
Signed-off-by: Aurelio Colosimo <aurelio@aureliocolosimo.it>

Default Data Set is common to all ports, so it must be defined globally.
pp_open_instance now gets a pp_globals object; the same should be done for
pp_close_intance, but is not urgent since it is currently used by wrpc arch
only, which is a single port arch.
Signed-off-by: Aurelio Colosimo <aurelio@aureliocolosimo.it>

Signed-off-by: Aurelio Colosimo <aurelio@aureliocolosimo.it>

Signed-off-by: Aurelio Colosimo <aurelio@aureliocolosimo.it>

This patch is temporary and is thought only to guarantee the correct
convergence of the algorithm. Some of the latest changes in servo.c, about
how time is adjusted and how the max value for observed drift is set,
made it unstable. With this patch, the behaviour is restored to the same as
ptpd, though it still needs to be reviewed to increase performance, readability
and easier maintainability.
Signed-off-by: Aurelio Colosimo <aurelio@aureliocolosimo.it>

Signed-off-by: Aurelio Colosimo <aurelio@aureliocolosimo.it>

Signed-off-by: Aurelio Colosimo <aurelio@aureliocolosimo.it>

I initially looked at int64_to_TimeInternal to swap arguments, since
to_TimeInternal as timeinternal (the output) as first argument, and
they appeared in pairs.  However, int64_to_TimeInternal is only used
for correction fields, and they are scaled nanoseconds (page 126), not
nanoseconds.

Thus, this commit fixes the conversion and renames the function to
"cField_to_TimeInternal" to better tell the reader what it is.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Now that it's clear pp_update_offset() receives t1 and t2, which are
fields of ppi itself, avoid passing them.

However now it's clear that this function might be seriously wrong,
and needs review (ofst_first_updated should die, I think, preferring
simply code to unneeded contortions).
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

I'm adding generic dump code to the receive and send code paths,
so I need the length of the incoming frame when first "unpacking" it,
to dump the whole frame using code from tools/dump-funcs.c
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

This is the result, on a master, with "-d 02":

diag-time-2-eth0: posix_time_get:  1362675879.940187923
diag-time-1-eth0: timeout expired: PP_TO_ANN_INTERVAL
diag-time-2-eth0: posix_time_get:  1362675880.081818439
diag-time-2-eth0: posix_time_get:  1362675880.233227916
diag-time-2-eth0: recv stamp: 1362675880.233227916 (user)
diag-time-2-eth0: posix_time_get:  1362675880.233499803
diag-time-1-eth0: timeout expired: PP_TO_SYNC

(time level 1 reports timeouts and set, level 2 reports get-time too).
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Unfortunately, we need to pass ppi to the time operations, so we
can have per-instance logging options. This is the second time I found
the need for such extra argument: the first time I worked it around, now
I surrender.

Actually, I don't find it unlikely for a multi-instance process to
spit a lot of diagnostics about timestamps. A developer may need to
filter messages according to who is tiemstamping/timing out etc.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

I need to refer to ppi->flags from inside diag-macros.h, so
this splits out a new header, which is expected to only be included
by ppsi.h (no change for our developers).
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Actually, until this commit we could not use pp_printf inside
ppsi.h, in inline functions or whatever.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

The new function reports how many ms in the future is a timeout event,
to be able to correctly time when we should re-entering the state machine
and correctly spread our frames according to the specification.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

All archs allocated one foreign master, but the default was 5 and
the command line could even change it.  Hardwire the number to 5
and skip the allocation by folding the array in pp_instance itself.

Anyways, the add_foreign() function needs serious review, later on.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

the pp instance is bound to a network interface, so avoid a separate
allocation.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

This changes some things to make code more readable (I really didn't
know what the tmp_header was: now it is the "received_ptp_header".
The commit simplifies allocation, because the tx and rx buffers are
now part of ppi itself.

We now have tx_ptp and tx_frame, both pointing within tx_buffer (and
the same for rx). The pointers are the protocol side (aligned) and the
argument passed to send/recv (which may be not be 4-aligned).

At this point nothing changes, i.e. both gnu-linux flavours work and
bare-i386 is not able to received.

This commit has no technical effect: what works still works, and what
doesn't work still doesn't work.  Still, the inlines will do
type-checking at a later time when we have a structure for the ptp
header.

Moreover the renames allow me to trace all uses in order to change
them; it also generates build errors in wrpc-sw, forcing me to fix the new
use in there as well, when I update ppsi as submodule.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

This commit moves the time operations inside the instance.  It
allows to have different time operations for each instance, which will be
useful in the future. For example, one ethernet card may White Rabbit,
and another may not be.

This commit, like moving network operations inside ppi, requires a
change in wrpc-sw.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

This commit moves the network operations inside the instance.  It
allows to have different operations for each instance, which will be
useful in the future. For example, one ethernet card may have hardware
timestamping and another may not.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

I'm going to split diagnostics into groups based on flags. And
with different diagnostics for each pp instance, to be ready to
debug multi-link situations.  This commit adds the flags in the proper
places and the first flag value, used in the next commit.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

By moving calc_timeout into the time functions, we simplify
the split-out of ./time-<arch> as a separate subdirectory.

As a side effect, calc_timeout disappears from arch-wrpc, and
wrpc-sw needs to be adapted to this commit, to provide it
inside its own time_ops.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

The structure was separately allocated with a size of 16.  This avoids
the alloc and makes the array the right size from the enum of message
types. We thus a remove a number of magic "16" in the code.
wrpc-sw must be updated too, in order to use this commit.

There's still a problem: the code claims the number to be "last sent",
but actually the value is incremented after sending. I'll need to pass
again on this thing.

ps: I changed name from "sent_seq_id" to "sent_seq" to the compiler
complains for previous users, otherwise undetectable.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

The function st_com_restart_annrec_timer is now just a one-liner, and
it really is just a timeout operation.

So call it pp_timeout_restart_annrec instead, and have it inline in
ppsi.h, removing it from common-fun.c
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>