Aurelio Colosimo authored 12 years ago and Alessandro Rubini committed 12 years ago

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

Aurelio Colosimo authored 12 years ago and Alessandro Rubini committed 12 years ago

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

After several back and forth, we agreed that softpll should live in
wrpc-sw, even if it is also used on the wr switch.  I don't pick up
the history, because the tool got renamed and moved several times.

This comes from commit 1e2c71d1f0e1094adc6c2eb46e30d4164262f85c, the
current master of ptp-noposix (the one that is currently a subproject)

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

The tool builds all releases travelling back from the current one up
to the named one. For example:

  ./tools/git-compileall master

  [... all the compiles ...]

   Number of failed builds: 0
   The log of all compiles is at /tmp/makeall.out.67Rf3n
   The md5sums of wrc.o are at /tmp/mdlog.sZXrAi

The tool was born for internal use, so it is misnamed and so on. But
it may be useful to others too.

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

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

Aurelio Colosimo authored 12 years ago and Alessandro Rubini committed 12 years ago

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

Aurelio Colosimo authored 12 years ago and Alessandro Rubini committed 12 years ago

This also changes ptp-noposix as needed

Aurelio Colosimo authored 12 years ago and Alessandro Rubini committed 12 years ago

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

Aurelio Colosimo authored 12 years ago and Alessandro Rubini committed 12 years ago

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

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

Calibration uses a lot of internal variables, and it used to overflow
even a 3kB stack (by a little amount). This patch shares a temporary
buffer between two functions, so it takes 128 bytes less of stack, and
3kB fit for the wr-nic release.

The problem must still be studied, to reduce stack use.

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

This fixes a bug with negative hex numbers (that must
be printed as unsigned hex, without the minux sign).

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 rule for "git submodule update" was wrong, so the command was
always performed.  This is a pain if you are using commits in the
submodule that are not yet committed in the supermodule. This always
happens during development.

I'm sorry for the inconvienence, and I thank Aurelio for noting this.

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 are always pretty near to the end of RAM, so now that we
have a reserved stack area, we can also check if it overlows.
When it happens, let's print a message forever, so a user who
comes and see what's up will see it eventually.

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

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

Pretty often it happens that during trials we pick wrong external
functions (e.g., "printf") and the final link claims horribly about
dozens of undefined symbols.  By having an intermediate wrc.o we can
easily check which are the symbols we really leave undefined for libc
and libgcc to fill, before they resolve them by asking for something
else.

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

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

This renames ram.ld to ram.ld.S, adding a preprocessing rule. The
commit has no techcnical effect: it is only preparation for the next one.

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

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

This adds copyright notes to all non-trivial source files,
unless they where already there (i.e. dev/endpoint.c alone).

I found authorship using the following script, run on commit "a2721762


documentation updated" (i.e, before automatic reindentation and other
trivial stuff by me):

  git grep -l . | grep -v sockitowm/ | \
  while read F; do
       echo "##### $F"
       git blame -w $F | \
         sed -e 's/^[^ ]* .//' -e 's/-[0-9][0-9]-[0-9][0-9] .*$//' | \
         sort | uniq -c | sort -rn
  done

Then I augmented each file with this boilerplate:

/*
 * This work is part of the White Rabbit project
 *
 * Copyright (C) 2011 CERN (www.cern.ch)
 * Copyright (C) 2011,2012 CERN (www.cern.ch)
 * Copyright (C) 2012 CERN (www.cern.ch)
 * Copyright (C) 2011 GSI (www.gsi.de)
 * Copyright (C) 2011,2012 GSI (www.gsi.de)
 * Copyright (C) 2012 GSI (www.gsi.de)
 * Author: Tomasz Wlostowski <tomasz.wlostowski@cern.ch>
 * Author: Grzegorz Daniluk <grzegorz.daniluk@cern.ch>
 * Author: Wesley W. Terpstra <w.terpstra@gsi.de>
 *
 * Released according to the GNU GPL, version 2 or any later version.
 */

Then I removed all the lines that didn't apply. Sometimes I reordered
the authors to reflect who is the main author.

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

Some individual files are LGPL, some are public domain,
some are GPL 2-only and most are GPL 2-or-later. You can
pick parts according to the respective license (marked
in the files themselves), but this one (GNU GPL version 2)
is what applies to the work as a whole.

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

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