[llvm] [WebAssembly] Fix lowering of (extending) loads from addrspace(1) globals (PR #155937)

[Demetrius Kanios via llvm-commits]

QuantumSegfault wrote:

@sparker-arm @dschuff 

So what's the verdict here? I'm guessing since this PR solves a less common use case than SIMD, the SIMD optimizations are prioritized. And since it has so far proven difficult to get this to not interfere with the other optimizers, should I just close this PR?

I'm still holding out hope that transitioning to GlobalISel will help solve this long-term...

https://github.com/llvm/llvm-project/pull/155937