Add a new DAE pass with a fixed point analysis

[tlively]

DAE can be slow because it performs several rounds of interleaved
analysis and optimization. On top of this, the analysis it performs is
not as precise as it could be because it never removes parameters from
referenced functions and it cannot optimize unused parameters or results
that are forwarded through recursive cycles.

Start improving both the performance and the power of DAE by creating a
new pass, called DAE2 for now. DAE2 performs a single parallel walk of
the module to collect information with which it performs a fixed point
analysis to find unused parameters, then does a single parallel walk of
the module to optimize based on this analysis.

[tlively]

This replaces #8085 because so much has changed in the meantime.

[kripken]

There is a lot of test code here - is it copied from the existing tests? If so, can we use another lit run line for it? If that isn't practical yet because of differences, I still wonder if we can reuse the existing tests somehow, even if the two passes have differing output in some cases (can use a different check prefix for those tests, maybe move them out to another file, etc.).

On another topic, what is our overall plan here? Do you want to land this even before we know if we will be replacing the old pass?

[tlively]

The tests here are all new. We could additionally run the existing DAE tests, but we wouldn't be able to remove many of the new tests that way.

[tlively]

As far as the overall plan goes, I'd like to have this landed so we can consider making it more powerful iteratively. My hope is that if it does more work DAE currently does, then we'll start seeing better speedups. We could also use it immediately to replace the unprofitable bailout from DAE, but that's probably not worth it.