<img height="1" width="1" style="display:none" src="https://www.facebook.com/tr?id=267568703576098&ev=PageView&noscript=1" />
mkroot is almost usable now
May 1 at 6:44am
Aboriginal Linux hasn't had an update in over a year because that project has come to an end, and its replacement is https://github.com/landley/mkroot as described here:

http://lists.landley.net/pipermail/aboriginal-landley.net/2017-January/002594.html

tl;dr the kernel broke my old lastgplv2 toolchain packages four times in two releases, and somebody else is providing usable cross and native toolchains based on current package versions now. When I moved to an external toolchain 2/3 of the old infrastructure had no reason to exist anymore, so I stated over with a simpler design.

I've done a lot of commits to mkroot this week, and it's about halfway to usable now. It can build kernels and qemu boot scripts for armv5l, aarrch64, mips, mipsel, powerpc-32, sh4, and x86-64. I still need to do kernel and qemu configs for armv7l, s390, i486, i686, microblaze, mips64, powerpc64, and sh2eb, but I've got compilers for all that. (If we want m68k/coldfire, sparc, alpha, and so on, we need to add support to musl-libc first. The new one doesn't use uClibc at all.)

I haven't added a module/compiler to install native compilers yet, but the mcm-buildall.sh script is building them in musl-cross-make. I also haven't done the second pass on my kernel configs and made sure they all have ethernet and block devices and battery backed up clock and everything. (They boot to a shell prompt in an initramfs, that's a start.)

This is also relevant to toybox because one of my _big_ todo items is getting a proper test environment set up where I can test things like mount/ps/modprobe. A known system where I can run stuff as root under an emulator and get consistent results without damaging my development environment. Aboriginal Linux being up on cinderblocks got in the way of that, now I'm building a toybox test harness on top of mkroot. (Elliott's flood of test suite patches may have had something to do with this. :)

My real question is should I try to rename the Aboriginal Linux list, or should I start a new list for mkroot and let interested parties subscribe? There's a ton of old documentation attached to the aboriginal linux webpage that applies to mkroot too, but if I'm rewriting it I wonder if it should go in some other format?

I tried to do a mkroot tutorial at ELC, and although it starts slowly (so much jetlag) it's not so bad once it gets going, you can watch it at https://www.youtube.com/watch?v=Sk9TatW9ino if you're interested.

Tag icon
Heart icon
Log in to comment ...
Rewards
Pledge $1 or more per month
2 patrons
Do the thing. (General expression of support.)
Pledge $5 or more per month
0 patrons
Do the thing! (Louder expression of support.)
Pledge $5 or more per month
1 patron
Do toybox. (Encourage me to spend more time on toybox, which has its own roadmap.)
Pledge $5 or more per month
2 patrons
Do system bootstrapping. (Encourage me to spend more time on Aboriginal Linux, which includes getting the Android Open Source Project to build under the minimal development environment it provides.)
Pledge $5 or more per month
0 patrons
Do toolchains. (Come up with cross and native compiler/linker/assemblers that can build a runnable system and are freely binary distributable without obvious risk of license lawsuit. This includes writing a new "make", "lex", "yacc", and so on.)
Pledge $25 or more per month
3 patrons
Do all the things! (Enthusiastic expression of support.)
Pledge $25 or more per month
0 patrons
Do toybox! (Louder encouragement to spend time on toybox.)
Pledge $25 or more per month
0 patrons
Do system bootstrapping! (Louder encouragement to spend time on Aboriginal Linux and cleaning up AOSP.)
Pledge $25 or more per month
0 patrons
Do toolchains! (Lounder encouragement to spend time getting a new toolchain working.)
Pledge $2,000 or more per month
0 of 5 patrons
Corporate sponsorship. Send me your todo list, I'll review it (via email, phone, or google hangouts), then take your priorities into account when choosing what order to work on things in one or more of the above projects, and send you monthly reports on progress towards your goals.

This won't change the nature of the projects, but can encourage me to include features I wouldn't otherwise bother with.