Chris Fallin

Ah, force-pushed after my review; could you do subsequent updates as additional commits? Let me know when stable and I can re-review.

Sure; to summarize here, the issue is that alignment is now correct with respect to the start of stack frame, but start of stack frame is only aligned as per...

> I had to pass -O0 to wasm-ld (which, while we're at it, is an incredibly unintuitive flag--it took me literal years and the compiler driver source to find out...

> If it's not, should we not just fix that? Feel free to add more data upstream! I think several folks have pushed on this (see above thread) but additional...

``` # good.wasm 29: 44 88 11 mov %r10b,(%rcx) 2c: c7 01 00 00 00 00 movl $0x0,(%rcx) 32: c7 01 00 00 00 00 movl $0x0,(%rcx) 38: c7 01...

On the flip-side, if you're proposing altering the generated code to assist debugging observability @jameysharp, there is a large design space that we haven't really explored. A relatively simple change...

@RyanTorok there are a lot of interesting ideas in your comment (I have to admit that I skimmed it in parts; I'd encourage a "tl;dr" of points for comments this...

Specifically the debug RFC from Nick, Trevor and Rainy [here](https://github.com/elliottt/rfcs/blob/trevor/wasmtime-debugging/accepted/wasmtime-debugging.md) covers our overall plans, with coredump fidelity (providing all Wasm virtual machine state) as a necessary prerequisite for our further-along...

> I'm assuming adding some kind of pseudo instructions in Cranelift for extra bookkeeping (when coredump-on-trap is set) has an acceptable perf impact. That would be the first question to...

> (the PBL changes have since landed upstream). Well, *most* of them -- I have to get to the last part of the weval-specific patches and submit them upstream if...