The entire argumentation here is ridiculous. There's a big jump from "IFUNC slightly undermines RELRO" to "IFUNC is the real culprit". You could have gotten all but the same effect spawning a thread from a plain init or C++ constructor. No one should think that any relro, r^x or aslr or anything like this is going to deter anyone who can literally control the contents of the libraries which are linked in. They could, literally, exec a copy of sshd with a patched config if necessary.
The title is just clickbait.
I think I triggered the posting of this article in a reply elsewhere, to which you gave the same comment. And dammit, you're totally right, there's lots of other ways to stick arbitrary code into the initialization process at load time. Maybe the real problem is linking dependencies by shoving them into the same address space as whatever is using them and calling it a day. Memory-safe languages can help protect that model against accident, but it still takes just one evil insider to steal the keys to the kingdom.