The interesting bit is that Pi avoids these attacks mostly by not doing speculation in the first place, rather than by adding some special Spectre mitigation.
Incremental seems to keep winning the easy wins, so the interesting part is whether the remaining compile-time still lives in the same places as last year.
Feels like the useful benchmark would measure code size and i-cache effects together, since a bit of block cloning could beat a dispatcher until larger hot loops flip that tradeoff.
Wasm’s structured blocks are a nice constraint, but irreducible CFGs are where the compiler has to choose between cloning code and introducing a dispatcher.