Accessing data structures in the host to get target fields is broken:
we may have different data size or alignment, or whatever. The hack
of using "-m32" to be able to access lm32 fields from x86-64 is not
clean (we imply the i386 and target have the same size/alignment).
Even worse, not everybody has an x86-64 cross-compiler for i386.

Thus, we use OFFSET_OF in the target, with the lm32 cross compiler,
to build a binary table that is then accessed by the host.

In order to link the data structure in the host we use the assembler
instead of objcopy. With objcopy we'd need to spell out the bfd name
of the host, and we'd get horrible names like
   struct dump_info _binary_dump_info_host_bin_start[];

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

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

There was a hole at the beginning of the generated binary, so this
commit includes some data to identify the endianness of the dump file
and host pointers to the important structures that we dump.
For wr-switch, the ppi pointer is just set to zero.

The change is backward compatible, because we just fill 32 bytes of
data where we had a bunch of zeroes.

The next commit will use such information when available in the dump file.

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

SQUASH

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

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

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

Use an environment variable to check for spec byte order, and
fix the buffer before using it.

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 commit has no effect on the final binary file.

The state structures for softpll_ng must be accessed by tools/wrpc-dump,
so we need them in an header (not in sofpll_mg.c itself).

Unfortunately, struct softpll_state includes a lot of other structures,
so all dependent headers must be included by softpll_ng.h.  Thus,
all C files should now include softpll_ng.h alone, not the other ones.

I also had to remove inclusion of unneeded <stdint.h> and <stdlib.h>
from spll_common.h, because stdlib pulled the prototype for abs()
which is #defined in ppsi (my bad!), and we would have a build error.

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

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

It's still suboptimal, we need to autodetect with a in-binary helper,
but by now I'm analyzing old logs so I'd better not change wrpc
itself.

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

This fixes the problem with data sizes when building on a 64-bit host.
Now the dump of ppi makes sense, besides endianness.

Please note how this requires a cross-compile environment for i386
when you build on x86-64. Later on this requirement is lifted.

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

The tool is meant to print internal data structures of the wrpc,
by running either on a dump file or a live system.

The code is an edited copy of wr-switch-sw/userspace/tools/wrs_dump_shmem.c,
so the data-dumping mechanism is from there.

bugs at this point:
   - all endianness stuff is bad (in wrs it was local, here we cross-dump)
   - data sizes are wrong (see next commit)
   - we miss checks for consistency
   - we can't open /dev/mem (for spec for example) or a elf file
   - docs are missing too

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

CONFIG_WRNIC was not used, so it is now removed.
CONFIG_NIC_PFILTER does it all

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

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

so that is works both on WRS and WRPC

Needed mainly for SPEC, where Spartan6 PLL does not report unlock when clk_in
(10MHz) is unplugged. In addition to that, datasheet says that PLL should be
reset after losing clk_in.

The last commit in branch pfilter-cleanup, where the built-time tables
are built, was bugged in patching the local mac address in the table.
Thus, ptp was working, but IP was not.

Thanks Greg for making the proper checks.

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

This commit fixes all copyright notes, using "git blame" as a
reference.  For simple files I just added a 2-lines boilerplate about
being WR and GPL.

Some parts I left untouched:

  trivial-init and related files (unclear to me what this is,
  and pf-microcode.h is magic binary pfilter values).

  tools/lm32-etheruart: I don't know what this is and who the author is

  libsdbfs: this comes from another repo and I'd better not change it

  sfp_lib.h and hal_shmem.h: I lost track about the origin of this,
  among the several packages.

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

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

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

We have not been using mprintf for ages now, and it's better for the
code to explicitly call pp_printf, so people is aware of it.

We might "#define printf pp_printf", but currently we'd better not.
Maybe when we turn this into a real operating system, next millennium...

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

In this commit:

   40b87d1c

 trivial: remove debug prints from pfilter

I made an error. By defining pfilter_dbg to nothing the arguments remained,
a a parens-enclosed comma expression with no effects. This fixes the
mishap and the related compiler warnings.

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 gui is timely: it refreshes every that many seconds, but I'd
better see a change immediated when it happens.  This commit
introduces monitoring of the wr-servo update count, so if anything
changed an information scree is dumped.

This is different from stats (two commits ago), because we want to
refresh timely in any case, so the user sees the thing is active.

You can verify the effect of the commit after "refresh 4": timely
refresh is now slower than servo updated and you'll see how gui
updates are synced with the ptp protocol.

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

With the new "only print if sth happens" way, sometimes stats is
just silent forever. This happens if we are master or if no slave
is active (well, and during initial setting up the wr link).

So now typing "stat" or "stat on" will turn on stats and print a line
of statistics, every time -- or the line "statistics now off".
This happens by changing the internal status of monitor_ppsi.c,
now exported to the cmd_stat.c.

This is the result on a master system:

   wrc# stat
   lnk:1 rx:0 tx:3 lock:1 sv:0 ss:'' aux:1 sec:5 [...]
   wrc#
   wrc#
   wrc# stat
   statistics now off
   wrc#
   wrc#
   wrc# stat
   lnk:1 rx:0 tx:11 lock:1 sv:0 ss:'' aux:1 sec:8 [...]
   wrc#
   wrc#

And this is what happens when we become slave:

   wrc# mode slave
   Locking PLL
   wrc# ptp start
   Slave Only, clock class set to 255
   wrc# lnk:1 rx:57 tx:60 lock:1 sv:1 ss:'SYNC_SEC' aux:1 sec:49 [...]
   lnk:1 rx:70 tx:64 lock:1 sv:1 ss:'SYNC_NSEC' aux:1 sec:1444901411 [...]
   lnk:1 rx:77 tx:66 lock:1 sv:1 ss:'SYNC_PHASE' aux:1 sec:1444901414 [...]
   lnk:1 rx:83 tx:68 lock:1 sv:1 ss:'WAIT_OFFSET_STABLE' aux:1 sec:1444901416 [...]
   lnk:1 rx:87 tx:69 lock:1 sv:1 ss:'WAIT_OFFSET_STABLE' aux:1 sec:1444901417 [...]
   lnk:1 rx:90 tx:70 lock:1 sv:1 ss:'TRACK_PHASE' aux:1 sec:1444901418 [...]
   lnk:1 rx:94 tx:71 lock:1 sv:1 ss:'TRACK_PHASE' aux:1 sec:1444901419 [...]

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

This change monitors the update_count of wr_servo, and only prints
a statistics line when something happens.

This means we won't have long "TRACK_PHASE" listings when the master
is dead. Moreover, we see changes when they happen, not at a random
time during the next second.

The bad side of the thing is that if nothing happens in the wr servo,
nothing is ever printed. But see next commit.

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