It addresses:
* .cvsignore (only minor changes to yours, as they were pretty clean,
most changes result from my mkcvsignore script being overly pedantic
sorting entries alphabetically.)
* timing.t handling and time<BSP>_.texi handling in supplements
* network.t handling in bsp_howto
* RTEMS_DATE and RTEMS_UPDATE handling in configure.in. [It actually is
a hack, as we could apply automake's version.texi handling instead - I
know think to understand what automake does with it.]
* avoid using temporary files; Therefore a new tool called bmenu2,
derived from bemenu, is introduced, which reads a single file from stdin
and writes to stdio.
To apply:
cd rtemsdoc
patch -p1 < rtemsdoc-4.5.0-rc-3.diff
cvs rm -f bsp_howto/network.t
cvs add tools/bmenu/bmenu2.c
./bootstrap
BTW: word-replace now is unused. It could be removed if you like to.
that addresses the following:
The patch below should
* fix the *pdf cleanup handling
* fix dependencies between pdf-documents and texi-files
It also contains an attempt to let networking/schedule.t use states.pdf
instead of ascii graphics, but though each pdf image looks nice, it
comes out weirdly distorted in the main pdf document.
which contains the bulk of converting the documentation tree to automake
and GNU conventions. Comments follow:
This is the automake port of rtemsdoc.
To apply:
cvs co rtemsdoc
cd rtemsdoc
sh cvs-rm.sh
patch -p0 < rtemsdoc-4.5.0-rc-0.diff
sh cvs-add.sh
[Attention: cvs-rm.sh and cvs-add.sh directly modify cvs]
Known bugs:
1) src2html is not supported (yet? - Is this supposed to work?)
2) all *.pdf images now are generated on-the-fly, but not yet deleted
during "make distclean"
3) All supplements, including the templated ones, get build and
installed.
4) Building outside of the source tree is completely untested and very
likely does not work.
5) Make [ps|pdf] are not (yet) supported, make [dvi|info] are supported
by automake's default texinfo rules.
Fixing 2, 3 and 5 is almost trivial and needs to be done.
4) is a matter of testing and tool-properties, for now it is simply
untested.
General issues:
* gif vs jpg vs png. I would recommend to replace all images with pngs
to avoid potential copyright issues (gif) or lack in quality (jpg, jpg
is good for real world photographs, but extremely poor on artificial
images, graphs).
* pdf images do net get placed correctly in pdf-documents.
* texinfo: We now use a local copy of texinfo-4.0's texinfo.tex in
texinfo/texinfo.tex for generating infos. However pdftex's system-wide
texinfo.tex and pdftexinfo.tex are used for generating *.dvi, *.ps, *.pdf.
* .cvsignore files still missing.
* I have renamed the supplements filename not to use c_<supplement>,
because automake seems to have problems with it.
Notes:
* Again, I recommend not to put any generated files into CVS. Here, this
comprises some *texi, all *.pdf and many *.html pages. Ie. I recommend
to run make maintainer-clean before checking in any files.
* To get building started, this should be sufficient:
./bootstrap
./configure
cd tools; make; cd ..
make info
* To make a public tarball:
[cvs co ; ./bootstrap]
./configure
cd tools; make; cd ..
make info
[make clean]
make dist
=> This generates a rtems-<version>.tar.gz in the toplevel directory.
=> Building the tools only is required after a "cvs co", but not in a
distribution tarball.
Gauthier <charles.gauthier@nrc.ca>. The power supply in their
VMEbus cage died so it was submitted at this point. It passes ttcp
which is a really good sign but they would like to do more testing
and cleanup once their hardware is functional again. Please
contact them if you are interested in using or fixing this driver.
As their comments indicate, the performance is actually quite good
even at this point as indicated by the ttcp results.
Please note that this is by no means a final version. The code is still
fairly ugly, and will require some further fixing. On the bright side, what
is attached works. I finally ran the test programs successfully with
optimized code and data cache enabled:
netdemos worked, but failed on the UDP transfer (runs out of buffers).
ttcp worked, with something like 1036 KB/sec Rx, 952 KB/sec Tx.
tftp worked
linkcmds updated, simio references removed, and switch to libchip for
serial ports from simio.
Added a MEMORY_MAP file to capture information about the various
addresses on this board.
In addition, many of the beta patches are now included.
Bob Wisdon <bobwis@ascweb.co.uk> and Chris Johns <ccj@acm.org>
to resolve a random network lockup problem.
ckinit.c:
Occasional network lockups have been noted when the PIT has a higher
interrupt request level than the CPM. The SCC1 bit in the CISR is set
even though the SCC1 interrupt handler is not active. This blocks
interrupts from SCC1 (and all other CPM sources) and locks up the
system. It has not been determined whether the error is within the
68360 or in the RTEMS interrupt support assembler code. The solution,
for now, is to set both PIT and CPM interrupt request levels to the same
value (4).
network.c:
Set CPM transmitter buffer pointer (_tbptr) to beginning of frame
before restarting transmitter. Don't retire transmitter buffer
descriptors belonging to the restarted frame.
that contains:
* Removes remaining (now illegal) references to $(SRC) from a couple of
Makefile.ams
* Removes duplicate AC_CONFIG_SUBDIRS macro from c/configure.in
* Moves ENABLE_LIBCDIR into RTEMS_PROG_C[C|XX]_FOR_TARGET (hides LIBCDIR
from most configure scripts, i.e. LIBCDIR becomes less visible)
* Adds RTEMS_PROG_C[C|XX]_FOR_TARGET and RTEMS_CANONICALIZE_TOOLS to
libbsp/*/configure.ins (A minor bug in previous implementations, which
only has an impact when switching to GNU/Cygnus canonicalization)
* Cleans up several bogus comments.
* Removes MKLIB
* Switches the version number to 4.5.0 (for testing version number
handling)
