StackIR: Run StackIR during binary writing

[kripken]

Previously we had passes --generate-stack-ir, --optimize-stack-ir, --print-stack-ir
that could be run like any other passes. After generating StackIR it was stashed on
the function and invalidated if we modified BinaryenIR. If it wasn't invalidated then
it was used during binary writing. This PR switches things so that we optionally
generate, optimize, and print StackIR only during binary writing.

This is almost NFC, but there are some minor noticeable differences:

1. We no longer print has StackIR in the text format when we see it is there. It
   will not be there during normal printing, as it is only present during binary writing.
   (but --print-stack-ir still works as before; as mentioned above it runs during writing).
2. --generate/optimize/print-stack-ir change from being passes to being flags that
   control that behavior instead. As passes, their order on the commandline mattered,
   while now it does not, and they only "globally" affect things during writing.
3. The C API changes slightly, as there is no need to pass it an option "optimize" to
   the StackIR APIs. Whether we optimize is handled by --optimize-stack-ir which is
   set like other optimization flags on the PassOptions object, so we don't need the
   old option to those C APIs.

The benefits this means to bring are

1. A step towards Run (basic) StackIR optimizations in all binary writes? #6509 which suggests we run StackIR opts more often. By doing
   them during binary writing we can make that change if we want it: it would be a
   small change to the binary writing process to always generate StackIR and
   optimize it.
2. In Fix binary emitting of br_if with a refined value by emitting a cast #6510 we wanted to potentially depend on StackIR for lowering of br_if that
   needs extra work due to type system differences with the wasm spec. After this
   PR we can make such dependencies if we want them, both because of the previous
   point and also that we no longer "split" binary writing across the codebase: before
   the --generate-stack-ir pass began the binary writing process, in effect, and so it
   stashed StackIR for the writing later. If we want more things, like br_if lowering,
   then we'd have more to stash. Doing it all during binary writing avoids that, since
   nothing is done ahead of time.
3. This simplifies a bunch of things, e.g., this PR removes the logic to check and
   invalidate StackIR during passes (no longer needed since it happens after passes
   run), and the new wasm EH lowering logic no longer needs to do anything special
   for StackIR (things just work properly: if we are optimizing, then during binary
   writing we'll emit and optimize StackIR).

In more detail, this moves passes/StackIR.cpp to wasm/wasm-stack-opts.cpp.
Those are all the StackIR opts and they are unchanged. Otherwise the binary
writing logic calls a StackIROptimizer that does those opts (if we are doing
them). The binary writing logic still generates and optimizes StackIR in parallel,
so this should have no downside in compile times.

Several tests change due to losing (; has StackIR ;) comments, as explained
above. A more significant change happens in e.g. O2_print-stack-ir as the
--print-stack-ir flag is a global setting, affecting writing, which happens at the
end, so the order of the outputs is reversed (before it was a pass that executed
according to the order on the commandline); the contents are unchanged though.

[tlively]

Could we simplify the printing situation by generating StackIR both in the binary writer and also separately on demand when printing, rather than requiring the printer to depend on the binary writer?

[kripken]

Could we simplify the printing situation by generating StackIR both in the binary writer and also separately on demand when printing, rather than requiring the printer to depend on the binary writer?

Good idea, yeah, I can try that direction. Will take some work though...

[kripken]

Ok, I still need to clean up some comments and such, but I think this now has the right shape following your suggestion. I went all the way now and removed all StackIR from wasm.h, after which there is a utility to generate it in wasm-stack.h that is used by both binary writing and printing.

[tlively]

Nice, removing StackIR entirely from wasm.h seems like a huge win, conceptually, at least. I'd be happy landing this.

[kripken]

Yeah, I think the wasm.h simplification makes the other downsides moot This is simpler now, even if it may be doing slightly more work. I took some more measurements now and still can't see a speed downside in practice, so I was overly worried, I guess. Landing.

[kripken]

(Also fuzzed and verified no code size/metadce changes on Emscripten.)