It's the most recent build of 5.5 on and there is exactly one failing test.
What's better is that from what our main GIS developer is telling me, the output is actually correct. The failure is just because of a difference in the way DOUBLE numbers are handled compared to x86_64 which leads to a tiny difference in the test output (the testing program is expecting an "8" in the output and the PowerLinux build of MariaDB 5.5 is saying "7.999999999999999", as far as GIS is concerned, both answers are correct).
In other words, it's not really a failure, just a test that needs to be updated to account for the slightly different output.
So with less than a day of work, MariaDB 5.5 is building perfectly on Power7. :)
There's still a lot of work to do, naturally. There are quite a few failures with the MariaDB 10.0 builds, which we'll be working on over the coming days. We need to start doing debug builds to enable more extensive testing. And we need to configure our build system to start creating proper packages. But hey, basic builds from source of MariaDB 5.5 are looking good!