This commit does three things that are bounded to each other
1) Moves configs from ConfigValues.inl to Config.json
2) Uses the json to generate a man file with the options inside of it
3) Generates the code required for FEXLoader to automatically parse defined options
Moving the configuration options to a parseable format was required to generate the man pages.
Can't really include an inl file in the man page
Man page gets generated and installed through the regular cmake install process
`man FEX` to get the man page
With this change, FEXLoader's argument parsing now will automatically be generated from the json.
This way whatever is in the json file matches what is in the man page, in the json, AND what is returned in FEXLoader --help.
It will stay in sync now.
FEXConfig is the only application that stays out of sync for now as it requires some more thought to plan out.
A minor improvement that this brings as well is that every boolean option that has a long argument also gains the inversion of that property.
aka, `--gdb` also gains `--no-gdb` The use case for this is minor but boolean arguments should always allow negated variants.
This was forcing every op to determine it had side effects, so DCE could
never actually remove any operations.
This fixes the typo which means each op declares side effects correctly
and DCE works again
Sadly this is a fairly substantial change that can't be broken in to
something smaller.
The gist of the problem is that most of the vector ops declared their
own uint8_t RegisterSize and ElementSize variables in their structs.
This was initially done because there were assumptions elsewhere in the
source that sometimes mixed up the IROp member `Size` and `Elements` as
sometimes meaning `Size` elements and `Elements` number of elements.
Other times there was an assumption that `Size` was the full register
size and `Elements` was the number of elements inside that size.
ex: `Size`: 16, `Elements` 8 would sometimes be interpreted as i8v16 OR i128v8
These assumptions are now removed and we also save on average 16bits per
vector operation. This is a necessary evil first step for upcoming IR
considerations.
This should have zero change on runtime behaviour and unit tests confirm
that.
Changes the IR register class definition to exist in the JSON file.
Removes a couple of IR ops that were unused and weren't necessary.
Changes the CPUID IR op to the new paired op which improves perf ever so
slightly.
Changes RA class handling in the RA pass to pull directly from the new
generated helper routine until it hits a complex op
Still need RA constraints to remove extraneous moves that would crop up
Ryan Houdek