[PATCH] avoid smscb test failure for ppc/ppc64 archi

[Sylvain Munaut]

I'm wondering if we shouldn't just refuse to build alltogether on big-endian.

It's known to be broken.

This might be the only instance picked up by the test case, but pretty
much the whole 'gsm' part is dependent on little endian snce we used
those 'struct' construct to represent gsm messages all over the place.

Cheers,

   Sylvain