The code, when receiving announce, was setting a flag in order to run
bmc at the next iteration, when that happened, the frame just received
was discarded in favour for the programmed state update.

This commits remove some fields and one helper function, so the new
announce is immediately processed and no frame is lost.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

the portState field is probably a relic of ptpd, then the state
machine was a single switch() statement.  Many places were still
setting it, but nobody was ever reading the field.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

There is no need for a "msg_tmp" in the pp-instance. All uses are
really temporary to a function, so we can use local variables instead.

(This is not the same as what was called tmp_header, which was
referenced all over message processing, so it still exists, though
renamed).
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

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

A large number of functions here got hdr and ann, but struct
pp_frgn_master includes such fields, and most callers passed hdr and
ann from the same structure.

So I chose to pass a foreign master instead of the two components, but
I'm still unable to remove the two full structures inside until I
clean up the msg.c mess.   So, at the end, this is not really a step
forward, but it helps such later step.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Since WR only needs one byte of information in the announce structure,
we can make is generic by calling it differently.

This changes only the name (and the size: it is now unsigned long),
but has no effect on the marshalling/unmarshalling of such data.

I'm touching this because I'm working on announce messages (as related
to foreign masters).
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

I renamed port_identity to port_id so the compiler helped me.
Then, since both fields are used at the same time, compare/copy
them in a single shot.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

The field was only to be set and reported, not really used.  We
definitely need a way to age potential masters, but counting announce
messages without expiration is not the way.
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>

Trivially, return what we have, becuase less than zero is less than
zero.  No, the wouldn't pass my exam.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

When the "if" branch returns, we don't need an "else" branch.  These
students wouldn't pass my exam.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

You may be bored by these small changes, but I want to show that
I didn't introduce errors. In any case, each such change is reducing
both entropy in the source and binary size in the output.
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>

This "type" was the only type represented as an array instead of
being a structure. This leads to inconsistent use: if you memcpy
the other ones you need "&", but not here.

So this commit uses an encompassing structure, to have ClockIdentity
behave like other composite fields.  The commit has no effect at all
on the generated binary files for wrpc but it unexpectedly reduces the
size of the output on the PC. Anyways, the commit is trivial, and all
changes can be easily verified to have no effect.

While this increases entropy, the reasoning is that I want to assign
structures instead of running zillions of memcpy functions, with the
hairy PP_CLOCK_IDENTITY_LENGTH.  So later commits, based on this one,
reduce entropy.
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>

One I forgot to remove when cleaning arch-spec, the oter I re-added by
error quite some commits ago, after it was moved to
lib-bare/bare-time.c
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

This kills the global build-time macro pp_verbose_servo and uses flags
instead.  Again, until I review the servo, stuff is not always
per-instance.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

This kills the global build-time macro pp_verbose_dump to use flags
instead.  In msg.c this commit fakes the old pp_verbose_dump, which
however is extracted from the global flags.

So, setting "frames" level to 2 prints information on frames being
received ("-d 002").

Until I get to integrate tools/ptpdump in the daemon itself, we have
no instance-specific dump, nor symmetric dumping of input and output.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

An error of mine. Removed, but make pp_error add "ERROR: " in front of
the format string it receives.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

This converts fsm.c to use the new diagnostics. We keep printing
the timestamp for fsm messages, which means diag.c::__pp_diag is
somehow duplicated here (in what was called pp_timed_printf).

This is the output with "-d 1" (or "-d "100") on a forced-slave ("-g"):

  diag level: 10000000
  diag-fsm-1-eth0: 1362666577.043: ENTER initializing, packet len 0
  diag-fsm-1-eth0: 1362666577.045: LEAVE initializing (next:   4 in 0 ms)
  diag-fsm-1-eth0: 1362666577.045: ENTER listening, packet len 0
  diag-fsm-1-eth0: 1362666577.045: listening: reenter in 11999 ms
  diag-fsm-1-eth0: 1362666587.204: listening: reenter in 1840 ms
  diag-fsm-1-eth0: 1362666588.388: LEAVE listening (next:   9 in 0 ms)
  diag-fsm-1-eth0: 1362666588.388: ENTER slave, packet len 44
  diag-fsm-1-eth0: 1362666588.389: slave: reenter in 889 ms
  diag-fsm-1-eth0: 1362666589.219: slave: reenter in 59 ms
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

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

By passing "-V -d 001" now I get level-1 diagnostics for frames (i.e.
sending and receiving):

Slave:

   diag-frames-1-eth0: RECV 44 bytes at 1362665418.025198568 (type 0)
   diag-frames-1-eth0: RECV 44 bytes at 1362665418.025970153 (type 8)
   diag-frames-1-eth0: SENT 44 bytes at 1362665416.806670510 (delay_req)
   diag-frames-1-eth0: RECV 54 bytes at 1362665416.808217467 (type 9)

Master:

   diag-frames-1-eth0: SENT 64 bytes at 1362665415.430554014 (announce)
   diag-frames-1-eth0: SENT 44 bytes at 1362665416.612818573 (sync)
   diag-frames-1-eth0: SENT 44 bytes at 1362665416.613056524 (follow_up)
   diag-frames-1-eth0: RECV 44 bytes at 1362665416.807983194 (type 1)
   diag-frames-1-eth0: SENT 54 bytes at 1362665416.808206209 (delay_resp)

(note: we don't tell names at receive time because the frame is still
to be validated).
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

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

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

This only introduces the header files, the actual code is still
missing, and is added by later commits. I still want this to be by
itself because it explains the thing (please read the comments in
diag-macros.h, that I won't copy in this commit message).
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>

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

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

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>

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

This adds the interface name where the error happened (we'll need it
for multi-link setups, that Aurelio is going to implement soon).

Moreover, this kills the last pp_diag_error call
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

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