I originally sent this to the mailing list because I was not familiar with the bug tracker, but it belongs here, so I am reposting it here.

There is an odd interaction between the kernel and glibc on gcc49. I have confirmed failures in coreutils 8.12, coreutils 8.16 and m4 1.4.16.

I asked the Gentoo MIPS team lead about the failure and he had this to say:

23:13 <@mattst88> yeah, that's right
23:14 < ryao> The standard prefix bootstrap doesn't work. make[3]: vfork: Function not implemented
23:15 < ryao> That is when building coreutils 8.16 as part of the
procedure. I am going to try altering it to make it work.
23:26 <@mattst88> I remember seeing that error
23:27 < ryao> When you saw it, was there an issue in glibc?
23:28 <@mattst88> 22:10 < mattst88> | for my future self: the vfork problem appears on .36 (and perhaps earlier) when glibc is compiled with new linux-headers, I think
23:28 <@mattst88> 22:10 < mattst88> | booting .37 solves the issue
23:28 <@mattst88> that was what I said more than a year ago in May!

The kernel is 2.6.36.3+, which corresponds with his notes on this issue. Upgrading to 2.6.37.y might fix it.