Please don't take this comment the wrong way. Can DoD meaningfully control for this risk? No. What they're doing is no doubt stupid and unhelpful. But the original notion, that maybe "foreigners" shouldn't be touching the IPv6 stack of something DoD depends on, is valid.
Having said that: just because you can inspect the source code doesn't mean it's safe to run. I don't believe for 1 millisecond that an entire industry of "security-conscious American programmers" would be effective at assuring the Linux IPv6 stack --- not without an actual overt program to arrange for that to happen.
The top vulnerability researchers routinely find game-over flaws in open source code long after the code hits the release branch, and often long after the code has been vetted several times over by other vulnerability researchers.
If my goal was to mess with the US by backdooring critical infrastructure software, I'd almost certainly not do that by inserting code like "look for secret passphrase at offset X from IPv6 extension header Y". Instead, I'd introduce extremely subtle bugs, perhaps depending object lifecycles and concurrency, and I'd make sure they were planted in changesets that improved code quality overall and/or added valuable capabilities to the code.
You sinister bastard. ;) Nuff said, point taken.