Wednesday, June 9, 2010

New MiNTlib and Friends....

I'm lazy, so once again, I'll share a recent post to the MiNT List about building the latest release of the MiNTLib. Whats news? Support for GCC4 and also support for the Coldfire CPU for the upcoming Firebee! I am by no means as skillful or knowledgeable as Frank Naumann was, who is dearly missed for his support of the platform and shaping MiNTLib, and the kernel, to be what they are today.


Before this goes up into SpareMiNT, thought it should be tested to have a consensus that the build is good. I have not found any issues, but that does not mean that I did not overlook something obvious and my Falcon is pretty much a test/dev environment rather than 'production.'

Files are here:

[ ] mintbin-0.3-4.m68kmint.rpm 31-May-2010 11:54 579k
[ ] mintbin-0.3-4.src.rpm 31-May-2010 12:00 328k

Install this package first to fix a conflict between the older version and the latest MiNTLib.

[ ] mintlib-0.59.1-1.m68kmint.rpm 02-Jun-2010 16:11 961k
[ ] mintlib-0.59.1-1.src.rpm 02-Jun-2010 16:24 1.4M
[ ] mintlib-debug-0.59.1-1.m68kmint.rpm 02-Jun-2010 16:13 1.7M
[ ] mintlib-devel-0.59.1-1.m68kmint.rpm 02-Jun-2010 16:15 1.6M
[ ] mintlib-profile-0.59.1-1.m68kmint.rpm 02-Jun-2010 16:17 1.8M

Here is the MiNTLib packages themselves, install what is needed. They include the Coldfire libs, as well as the 68000 and 68020-60 libs.

[ ] fdlibm-5.3-2.m68kmint.rpm 04-Jun-2010 13:49 109k
[ ] fdlibm-5.3-2.src.rpm 04-Jun-2010 13:49 85k
[ ] pml-2.03-4.m68kmint.rpm 01-Jun-2010 18:10 25k
[ ] pml-2.03-4.src.rpm 01-Jun-2010 18:10 68k

The MiNTLib update will remove math.h, so whichever math lib is being used needs to be upgraded as well. Both PML and Fdlibm include libraries for 68000 / 68020-60 / Coldfire. PML's Coldfire support is from Vincent's patching. I added the Coldfire support to Fdlibm, and have not tested it yet.

[ ] gcc-4.5.0-1.m68kmint.rpm 01-Jun-2010 15:40 6.5M
[ ] gcc-4.5.0-1.src.rpm 03-Jun-2010 14:52 32.8M
[ ] gcc-c++-4.5.0-1.m68kmint.rpm 01-Jun-2010 15:51 9.5M

Finally, if you want to be brave, GCC 4.5.0 for FreeMiNT. I used this to built MiNTLib, so it seems pretty stable. Based on Vincent's latest patch. If this is installed, it will replace GCC 2.95.3, and you will want this MiNTLib and also Binutils 2.20.1.

I used 'black magic' to build this package. It was build with the Cross Compiler in PS3 Linux using a spec file modified to detect whether it is running natively or cross, and then I had to modify the RPM to be totally 3.0 compliant as it was built with RPM 4.4. There were some other tricks employed in there to fix problems with RPM noting the OS used for the build.


Hopefully today I will post a little tutorial (that might possibly be a tad incomplete, but we'll see) on how to try to setup an environment to try to cross compile RPMs. Right now, I am testing this process with the newly released OpenSSL 1.0.0a and will probably also try this with a rebuild of OpenSSH 5.5p1, to keep us up to date!

Tuesday, June 1, 2010

Multi-Target SpareMiNT (RPM) Thoughts - How to do it maybe...

This is the other list post about multiple targets for RPM....


This is extremely long, but hopefully I explain almost everything....

Basically, this is my thought towards a multi architectured SpareMiNT. Note that this can be tweaked, as I am just basically taking stabs to see what works where. And it is definitely not perfect.

Now, my opinion on libraries versus binaries is this, I think we should try to build the libraries all bundled into a single RPM, because a huge problem will arise with includes needing to overwrite each other. I have, so far, done this with both OpenSSL and Zlib. Binaries can each get their
own package, but the problem arise when lib/binary combination, and that is going to take some thought or fancy spec maneuvering. Anyway....

My first thought is we basically have 3 targets:

m68kmint - 68000, basically the base
m68k020mint - 68020-60
cfmint - 5475

So, for the release of make, you have:

make-3.81-3.src.rpm The src RPM to rebuilt the packages
make-3.81-3.m68kmint.rpm The 68000 binary
make-3.81-3.m68k020mint.rpm The 68020-60 binary
make-3.81-3.cfmint.rpm The Colfire binary

Can these be built? Yes, but we have to make our systems allow them to be installed before we get to the build process.

All systems can install m68kmint.rpm packages, but there need to be changes to allow m68k020mint.rpm and/or cfmint.rpm packages to be installed and/or built.

Two files need to be edited. The first one is /etc/rpmrc.

I believe the original looks like this:

# /etc/rpmrc
optflags: -O2 -fomit-frame-pointer
Provides: /sbin/rmt


Basically this file is just specifying the CFLAGS (Compiler Optimizations)
to use when building a RPM pacakge. So, my modifications:

# /etc/rpmrc
optflags: m68kmint -O2 -fomit-frame-pointer
optflags: m68k020mint -m68020-60 -O2 -fomit-frame-pointer
optflags: cfmint -mcpu=5475 -O2 -fomit-frame-pointer
#optflags: -O2 -fomit-frame-pointer
Provides: /sbin/rmt


The above is setting the CFLAGS for the additional 2 targets so they are
CPU specific.

The other file that needs changes is the /usr/lib/rpm/rpmrc file. This
file is quite large. To be brief, here is the original snippet of lines
that contain mint and might be changed (but not all will):

arch_canon: m68kmint: m68kmint 13
arch_canon: atarist: m68kmint 13
arch_canon: atariste: m68kmint 13
arch_canon: ataritt: m68kmint 13
arch_canon: falcon: m68kmint 13
arch_canon: atariclone: m68kmint 13
arch_canon: milan: m68kmint 13
arch_canon: hades: m68kmint 13
buildarchtranslate: atarist: m68kmint
buildarchtranslate: atariste: m68kmint
buildarchtranslate: ataritt: m68kmint
buildarchtranslate: falcon: m68kmint
buildarchtranslate: atariclone: m68kmint
buildarchtranslate: milan: m68kmint
buildarchtranslate: hades: m68kmint
arch_compat: atarist: m68kmint noarch
arch_compat: atariste: m68kmint noarch
arch_compat: ataritt: m68kmint noarch
arch_compat: falcon: m68kmint noarch
arch_compat: atariclone: m68kmint noarch
arch_compat: milan: m68kmint noarch
arch_compat: hades: m68kmint noarch
buildarch_compat: atarist: m68kmint noarch
buildarch_compat: atariste: m68kmint noarch
buildarch_compat: ataritt: m68kmint noarch
buildarch_compat: falcon: m68kmint noarch
buildarch_compat: atariclone: m68kmint noarch
buildarch_compat: milan: m68kmint noarch
buildarch_compat: hades: m68kmint noarch


These are the definitions that allow rpm / sparemint software to be
installed on the various Atari systems. Now, here is my 'modified'
version, which allows the newer RPMs to install:

arch_canon: m68kmint: m68kmint 13
arch_canon: atarist: m68kmint 13
arch_canon: atariste: m68kmint 13
arch_canon: ataritt: m68kmint 13
arch_canon: falcon: m68kmint 13
arch_canon: atariclone: m68kmint 13
arch_canon: milan: m68kmint 13
arch_canon: hades: m68kmint 13
arch_canon: coldtari: m68kmint 13
buildarchtranslate: atarist: m68kmint
buildarchtranslate: atariste: m68kmint
buildarchtranslate: ataritt: m68kmint
buildarchtranslate: falcon: m68kmint
buildarchtranslate: atariclone: m68kmint
buildarchtranslate: milan: m68kmint
buildarchtranslate: hades: m68kmint
arch_compat: atarist: m68kmint noarch
arch_compat: atariste: m68kmint noarch
arch_compat: ataritt: m68kmint m68k020mint noarch
arch_compat: falcon: m68kmint m68k020mint noarch
arch_compat: atariclone: m68kmint m68k020mint noarch
arch_compat: milan: m68kmint m68k020mint noarch
arch_compat: hades: m68kmint m68k020mint noarch
arch_compat: coldtari: m68kmint cfmint noarch
buildarch_compat: atarist: m68kmint m68k020mint cfmint noarch
buildarch_compat: atariste: m68kmint m68k020mint cfmint noarch
buildarch_compat: ataritt: m68kmint m68k020mint cfmint noarch
buildarch_compat: falcon: m68kmint m68k020mint cfmint noarch
buildarch_compat: atariclone: m68kmint m68k020mint cfmint noarch
buildarch_compat: milan: m68kmint m68k020mint cfmint noarch
buildarch_compat: hades: m68kmint m68k020mint cfmint noarch
buildarch_compat: coldtari: m68kmint m68k020mint cfmint noarch


Yes, the coldtari name needs to be fixed. :)

What this does now is this lets machines with 68030+ CPUs install the m68k020mint packages besides the standard ones. It also 'allows' the RPMS for all three targets to be built on the various systems.

So, now we have support for installing the RPMs, and support for building them, how do we build them?

So, take a standard RPM spec file. This is the file that has the instructions for building a package. We will use make.spec

For 68000 (The Default)
rpm -ba make.spec

For 68020-60
rpm -ba make.spec --target m68k020mint

For Coldfire
rpm -ba make.spec --target cfmint (Oops, this might fail, we'll get to that)

What this is doing is passing the compiler optimization for those targets to RPM, which, for make, this is how they are used:

make_cv_sys_gnu_glob=yes \
CFLAGS="${RPM_OPT_FLAGS}" ./configure \
--host=m68k-atari-mint \
--prefix=%{_prefix} --disable-nls \

Right here we are either setting no CPU target or -m68020-60 or -mcpu=5475 at the CFLAGS="${RPM_OPT_FLAGS}"

This will produce makefile to build that way and output a properly named RPM.

Did I just mention a possible problem with Coldfire? Well, the problem is that since Coldfire binaries will not run on anything other than Coldfire CPUs, this causes configure to error out when testing for a Cross Compiler. My 'current' solution for that is the above "--host=m68k-atari-mint" line, which causes configure to think it is cross compiling the Coldfire target on a 68060 system, which is sort of true. This option, I believe, is ignored for the 68000 and 68020-60 targets.

Anyway, if you are still with me, thats my thought. I'm sure a bunch of this is confusing, and people can tweak it, so, go for it!

In closing, let me share a large snippet of the Zlib spec to demo what I talked about in the intro:

# Coldfire
CFLAGS="-mcpu=5475 ${RPM_OPT_FLAGS}" \
./configure \
mv libz.a libzCF.a
# 68020-60
CFLAGS="-m68020-60 ${RPM_OPT_FLAGS}" \
./configure \
mv libz.a libz020.a
# 68000
./configure \

[ "${RPM_BUILD_ROOT}" != "/" ] && rm -rf ${RPM_BUILD_ROOT}

mkdir -p ${RPM_BUILD_ROOT}%{_prefix}
mkdir -p ${RPM_BUILD_ROOT}%{_prefix}/include
mkdir -p ${RPM_BUILD_ROOT}%{_prefix}/lib
mkdir -p ${RPM_BUILD_ROOT}%{_prefix}/lib/m68020-60
mkdir -p ${RPM_BUILD_ROOT}%{_prefix}/lib/m5475
mkdir -p ${RPM_BUILD_ROOT}%{_prefix}/share/man/man3

#make install \
# prefix=${RPM_BUILD_ROOT}%{_prefix}
install -m644 libz.a ${RPM_BUILD_ROOT}%{_prefix}/lib/libz.a
install -m644 libzCF.a ${RPM_BUILD_ROOT}%{_prefix}/lib/m5475/libz.a
install -m644 libz020.a ${RPM_BUILD_ROOT}%{_prefix}/lib/m68020-60/libz.a
install -m644 zlib.h ${RPM_BUILD_ROOT}%{_prefix}/include/zlib.h
install -m644 zconf.h ${RPM_BUILD_ROOT}%{_prefix}/include/zconf.h
install -m644 zutil.h ${RPM_BUILD_ROOT}%{_prefix}/include/zutil.h
install -m644 zlib.3 ${RPM_BUILD_ROOT}%{_prefix}/share/man/man3

# compress manpages
gzip -9nf ${RPM_BUILD_ROOT}%{_prefix}/share/man/man*/* ||:


If you can follow this, what is happening is Zlib is being compiled 3 times, and the libs for the first 2 compiles are moved out of the way. Finally, everything is installed manually, and the libs for all 3 targets are placed in their correct locations.

The only other way to cleanly support the multilibs I can see, besides having all three loaded into 1 package is to make the 68000 package the main package, and then the other 2 packages sub to it, so you would have:


Anyway, this is long enough, but this is my thought process. Any and all thoughts except on how to make this better.


MiNTLib / Kernel Future and also SpareMiNT......

Since I have been remiss with updates, been doing work with some progress, let me post some of my list postings. Will be 2, one about the future (this post) and the next about SpareMiNT for multiple targets......


Hi All,

Well, not sure where to go with this conversation, but we need to start it.

First being, we need a person, or a group of people, to steer MiNTLib and the Kernel. Mainly, I think, these people need to decide when it is time to do a release, maybe a little project planning, but I do not believe we need too much. I think this needs to more or less be a decision by the group.

It seems to me that it was more or less decided we should release MiNTLib at 0.59, should this be done now, or with a past build? I know I agree with this since some of the modern tools (gmp/mpfr/mpc/gcc) need support from functionality added and fixed in the CVS/developmental branch of the MiNTLib.

I would propose that we do a release now, and then plan the 0.60 version of MiNTLib to be released around the time of the Firebee release (possibly) in case things need to be fixed or added for Coldfire.

As for the kernel, I can not speak to it's status for release. I would recommend a road map or maybe a list of goals to prepare it for release, mainly bug squishing, and then expand that list to include what features, additions, etc, would go into the next revision.

Now, the other question is Sparemint. Where should it head? Are people still willing to build RPMs and support the software, and do people still want the software supported?

The other issue with Sparemint (well RPM), at least in my opinion, is architecture support. By default, unless currently noted, RPMs are built for a 68000 target. Should that remain the same, or should this expand so that packages are also target at 68020-60 and then later at the Coldfire? If so, how?

I have my own ideas / thoughts for this, but not much point in executing it if it would be of no use.

Remember, Alan is also working on a Gentoo MiNT distribution. I just want to basically see programming resources not go to waste, and get some real 'direction.'

Sorry for writing a book. Back to working on MPFR RPM.....


Monday, May 24, 2010

The world of RPMs and SpareMiNT with FreeMiNT....... Specs and multiple targets...

RPMs, at least for users, make it easy for them to add or remove software, update, etc. More modern versions of RPM even make net updates pretty easy.

For developers though, RPM is hit or miss. Personally, I dislike the spec file. The spec file is basically the recipe, or instructions, for RPM to take a source archive, and possible a patch, and to successfully compile it, install it, and then package it for use by users.

Where am I going with this? Well, my latest adventure in trying to solve a major issue with SpareMiNT, how do you support multiple targets, and differentiate between them. Multiple targets meaning 68000, 68020-60, and 5475 (Coldfire).

68000 will work anywhere. On the Coldfire CPUs, though, these packages degrade performance because they must be run through an emulation layer to emulate the difference between the two architectures, since Coldfire is missing 68k instructions. So, when possible, you want software targeted for Coldfire. 68000 is compatible with 68020-60, but there are various optimizations gained from targeting the series of CPUs, so while you might not need all native software targeted for these higher end CPUs, there are some nice performance boost given in different areas.

68020-60 will only work on these CPUs and not Coldfire (I could be wrong). On that Atari ST platform, 68020 are rare, 68030 are common, 68040 are uncommon, and 68060 is pretty common now with the CT60.

5475 will only work on Coldfire series chips, such as the Firebee that is currently being developed / tested / in production. This series of chips is basically the successor of the m68k series. There is no telling how common this CPU will become, but tools are already supporting this target. I have built OpenSSL to include libraries supporting the target, and the GCC port also supports it, along with MiNTLib.

So, we have a quick and dirty summary of our targets. If you are a developer, you probably want or should have libraries built for each target, while if you are a user, you most likely only want binaries targeted at the systems CPU. This is where things get messy.

For developers, having the libraries separated is easy, currently, 68000 libs are the default and reside in /usr/lib and /usr/local/lib. The other targets reside in m68020-60 and m5475 directories in each of those directories. When software is built and the compiler is told to build for a CPU target or defaults to 68000 if nothing is specified, and it searched those directories first for supporting libraries needed, and then falls back to the 68000 libs if needed (and this behavior may change). So, this makes it sort of easy for developers. Only sort of, because when the libraries are installed doing the standard 'make install' they never actually go to the right home, so that needs to be done manually, and the '.la' file with the lib, if it has one, needs to be updated as well for the relocation. If I am incorrect with this statement, boy have I been making additional work for myself!

Anyway, the point is, we have an easy way to separate the libraries, and keep track. And, all of the libraries can be easily packed into a single -dev RPM for specific software. My recent Zlib build includes libs for all three targets, but only binaries for 68000.

This is where we run into problems. We have no easy way to separate the binaries. All of the binaries are just dumped into /usr/bin or /bin. Now, we could add CPU targeted subdirectories, but I do not see that as being worthwhile. My thought right now has been that the suffix is modified on an RPM for a user package so they can choose which version they want. For instance, I am using m68kmint (what has been used, 68000), m68k020mint (68020-60 target) and cfmint (5475). Not the most pretty solution, but we need something.

I guess that is it for rambles.....

Currently I am testing to see if I can build a Coldfire RPM of OpenSSH by just adding "--host=m68k-atari-mint" to configure. Configure is tricked into thinking it is cross compiling on my Falcon, which I guess is not a bad thing. I actually got a decent distance once with the compilation today before the dreaded malloc bug hit. Maybe that is my current issue with RPM, no luck building OpenSSL so far in RPM form because of the same bug..... RPM basically forces you to complete the process from the beginning to the end, no restarting without a lot of monkey business and tricking RPM.

Sunday, May 23, 2010

Zlib 1.2.5 and OpenSSH 5.5p1 (Compiled against OpenSSL 1.0.0) RPMs.....

First we have Zlib. The binaries are built for 68000, and libs are included for 68000 / 68020-60 / 5475.

And here is OpenSSH 5.5p1, built for either 68000 or 68020-60. To install the 68020-60 versions, modifications need to be made to Sparemint's rpmrc file adding the m68k020mint target.
is a modified rpmrc file for /etc and
is the modified version of the file in /usr/lib/rpm/

This also adds a cfmint target as well.

No 5475 version yet as it needs to be either cross compiled or built natively on a Coldfire board.

If the above rpmrc files are installed, it is possible to built architecture specific RPMs, with properly formatted .spec files with:

rpm --target m68k020mint -ba openssh.spec

Basically, the CPU target is specified as the CFLAGs, and the spec file needs to read this variable and provide it to the configuration mechanism used for building the source.

Friday, May 21, 2010

m68k-atari-mint Cross Compiler (GCC 4.5.0 / Binutils 2.20.1) for PS3 YellowDog Linux 6

Just like Vincent's cywin cross compiler, this installs in /opt/cross-mint. Just extract the 3 files in /. You will also need at least MiNTLib and PML, which can be grabbed from Vincent's page.

OpenTTD, Maelstrom and other jabbers.....

As far as I can recall, Maelstrom was stable, except for maybe a problem with writing scores to disk. I'm thinking it is time for me to give it a go again, build it with the latest tools and SDL, and see how it performs. Hopefully I can even get it going in TOS, which would be nice. Here is the old version:

And here is the latest 'working' version of OpenTTD.

I should probably qualify that, working for me. I had this run on my Falcon for about 36 hours without a problem. Since this revision, they have switched the code to C++, and every time I build, the resulting binaries fail. It seems to be some weird string issue, but I never get far investigating cause I shift my time back towards ScummVM or other MiNT items like the modern GCC running natively and OpenSSL/SSH. I do have the latest (1.01) source tree sitting compiled on the Falcon just waiting for debugging to see if I can figure out what the woes are.

This info is mostly from this ancient web page:

And, for the curious.... This is a game I want to get ported and working, and so far have not been able to for lack of hacking time mostly:

Two others would be Rise of the Triad and Jagged Alliance 2, both of which lack real websites. I did a TON of work on JA2 originally to make the released source compile with GCC on Windows removing many of the, um, Microsoftisms of the C code.