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>

If we are master (or slave with no master running) the next commit
makes stats completely silent (nothing happens: nothing printed).
A user typing "stat" won't know if it turned them on or off. So
turning on is now printed:

   wrc# stat
   wrc#
   wrc#
   wrc# stat
   statistics now off
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

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

"stat" enables statistics, without interrupting shell interaction.
Thus, there is no "esc to terminate" any more.

Another "stat" command with turn stats off. As an alternative, "stat on"
and "stat off" is supported.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

The function was only called with "0" as argument. Drop it.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

This reverts commit 03062ea3.

The patch is wrong, because it duplicates

    9491e61b ppsi: pick new master and fix building accordingly

which does the "fixing" in a more elegant way. With both in history,
we now have

   ./ppsi/include/generated/autoconf.h:13:0: warning:
           "CONFIG_VLAN_ARRAY_SIZE" redefined
   generated/autoconf.h:14:0: note: this is the previous definition

Besides, the reverted commit was not even using new ppsi master (we
did it before this commit).
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Commit 70b42d0d is currently proposed_master, so by our agreements with
GSI we can't remove it from history.  The change itself is already
in master (but with typos fixed in the commit message). So let's
pin the commit by merging, even if it will appear twice in the history.

This just adds "void" in prototypes. I did the pfilter on a different
branch, and lost these changes of commit be2bd234 wyhile moving pfilter
generation to tools/
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Use the rules from the array we are passed in the object file.
We'll use the feature when enabling vlans: at that point we'll have
two sets to choose from, according to run-time configuration.

Using the build-time configuration avoids the hairy ifdef and saves
almost 1kB of binary size.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

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

This changes pfilter-builder.c so it really build in user space, but removing
register setting and creating output files instead.

Running this program creates 3 files, that correspond to the previous
three #ifdef conditions.  The files are:

      rules-plain.bin    this is used for the no-etherbone configuration
      rules-ebone.bin    etherbone setup
      rules-e+nic.bin    etherbon plus 7solution's wr-nic packet filter

The files begin with a magix 0x11223344 word, that allows the soft-core
to fix any endianness difference (so no hairy mishaps are expected when
switching to a different soft-core). Then the 40-bit command words are
saved as 64-bit vaules, LSB-first. The output file is thus an odd number
of words and no 64-bit alignment is required.

The first three instructions of the packet filter are used to compare
the destination mac address of the frame.  We now use a fake mac address,
and the LM32 code will change it while programming the binary.

Please look at this commit while ignoring white-space, as it changes
indentation while turning #ifdef/#else into if()/else .
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

We are going to have a user-space tool to build the packet filter
images. The first step is copying dev/ep_pfilter.c, so we can
"git diff" the changes.  At this point the file is not
even built.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

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

Simplify access to $CONFIG_RAMSIZE. Actually, the previous way
was not working for me, as my $SHELL is not bash.

The bug is uncovered by the next commit, because the "-s" argument
to genramvhd has always been wrong for me, but I didn't notice
(so maybe the "-s" argument has no effect? I can't tell at this point)
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

If we compare for the local MAC first, it's easier to patch
the mac address at run time.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

We now explain how the class bits are used after the pfilter is over.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>
Acked-by: Grzegorz Daniluk <grzegorz.daniluk@cern.ch>

There is no need to assign register names to numbers explicitly,
because there's always the risk to make errors.

This also adds a build-time check for any overflow of the 22 registers:
if we allocate more than 22, it won't build.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

This removes the register numbers from comments, and rewords
comments in general to match the fact that registers are now named.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

The current code was reusing R10 and R11 for a different role.
However, we have several free registers, so this allocates two new
more, by giving them a meaningful name.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Uses of R20 and R12 are temporaries, and very local. Thus,
we can reuse R14, already blessed R_TMP.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

The existing code used R6 both for ARP and for ETHERBONE.  Allocate a
new register, since we have plenty. This clarifies the code (by using
proper symbolic name) but changes the generated binaries.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

This commit is not changing the generated code at all (the
bits generated are exactly the same as in the previous commit),
in all three cases of #ifdef.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

By naming registers the code is much more readable in my opinion.
And we avoid, in future modifications, to mistake a frame offset for
a register.

The enumeration of names starts from 1024 to deny use of simple
numbers. Also, the enumeration is verified to be exactly 32 names,
by raising a build error if not.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

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