The segment prefix only overrides one side of the string instruction.
Passing in the flags was causing the instruction to have the segment override on both sides of the copy.
So something like `movsb es:[rdi], fs:[rsi]` was being interpreted as `movsb fs:[rdi], fs:[rsi]`
Same for all the string ops which is now fixed.
We can't unit test these currently since we can't set host segment register for confirmed host behaviour.
Introduces three new environment variables that need to live outside of the scope
of the regular argument loader path.
This will allow external applications to adjust FEX parameters in interesting ways.
FEX_APP_CONFIG: Allows you to override where Config.json lives
FEX_APP_CONFIG_LOCATION: Allows you to provide an entire folder of app profiles
FEX_APP_DATA_LOCATION: Allows you to override where the data gets stored and loaded from
Fixes#572
In the case of a shift value larger than the element size then the shift will zero the element.
Incoming shift amount is always a 64bit value which means that the value can become quite large on shifts.
Checks the incoming source and sets the maximum to be the element size, which works for AArch64's vector shifts
Which only operate on 8bit incoming shift amounts
In a world with NaNs we were hitting some undefined behaviour with fmin/fmax
Needed to change these over to something more complex to handle the NaN case.
ARMv8.7 alternative FP mode would solve this for the scalar case, but it still doesn't work for the vecto case sadly.
If the relative or absolute folder doesn't exist then FEX will search in the data folder for a rootFS named the same thing
If the path exists then it is used.
For example if my data folder contains `$HOME/.fex-emu/RootFS/Ubuntu_main/` and I set the RootFS option to `Ubuntu_main`
Then this rootfs will be used. Allows you to easily select a rootfs in that folder without having the full file path.
emplace() only performs an insertion or assignment if the key doesn't
already exist within the map, which is at odds with what the comment
above the line indicates should happen.
insert_or_assign() better models what the comment indicates.
Prevents a class of sneaky logic bugs from slipping through into the
codebase.
This also resolves a case of such a bug within the Decorder's ReadData()
where all 3 byte reads would be performed as if they were a 4 byte read.
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.
There are a couple of warnings remaining but need restructuring to solve.
I believe @phire is going to fix the one warning about a virtual destructor in a upcoming PR.
Fixes#679
Ryan Houdek