Comment 4 for bug 1075111

Fix committed. We're now using the memcpy from master, seems to be about the same speed in my benchmarks.
Still using cortex-strings memcmp and memset, they're faster than the new stuff that landed on master.