See my prior comment: IMO it's the caller's issue that (guaranteed) TCE is needed. The callee should not have to do anything.

One of the things I forget to give thanks to the Rust designers for is the default of immutability. I remember my days in a Java environment where policy was "everything needs to be declared final" (so "final" got spattered all over the place) with zero nostalgia. Requiring TCE callees to explicitly represent themselves as TCE-friendly would be the same sort of abomination.

I feel like I’ve seen elsewhere that the argument there is that you often must have this optimisation working in algos that rely on it or you will get stack overflows. Having a keyword to force it then becomes a very useful thing, vs relying on hopes that future compiler versions and different arch targets will all discover the optimisation opportunity.