bdshemu: Fix edge case with unsupported instrs

[ScimitarEnjoyer]

Fixes #106

Adjusted this PR to be more consistent with the code style.

Note: InstructionsCount is incremented before stop on exploit is checked. If shellcode is detected after n instructions, InstructionsCount will be n+1, even though instruction n+1 was never emulated. Context->InstructionsCount++ could be done after the stop on exploit check in order to be more accurate, but this is nitpicking.

[vlutas]

This was intended from the beginning (i.e., cycle the emulator for these instructions), but this may also be problematic, because when we reach the maximum number of instruction, we would exit with SHEMU_SUCCESS, confusing the caller.
Your change means that in those cases, we would get fewer emulated instructions, which will also cause the tests to fail (they will have to be adjusted accordingly).
I'll give this some thinking time, to see if the change can be merged or if there are better ways to handle this.
Thanks for the PR!

[ScimitarEnjoyer]

Yeah, I was calling ShemuEmulate in a loop with MaxInstructionsCount = 1 in order to be able to "hook" into the emulator and inspect the context after every instruction when I realized that it was succeeding when it shouldn't have (like on syscalls).

[vlutas]

So in conclusion, the change looks fine, and we can simply return STATUS_ABORT_CANT_EMULATE before cycling the emulator, as in the PR.
However, please make sure to also run the tests, and adjust those that fail (since now the number of emulated instructions will be different).
Once you also fix the tests, I will merge this.
Thanks again for the PR!
LE: To run the tests, check out the tests/main.py script. To regenerate the result files, check out the --regenerate option, which will re-generate all the output files to match the latest emulator output.

[ScimitarEnjoyer]

I have checked and regenerated the tests, should be good to merge now :)
Thanks!

[vlutas]

The change will be part of a future v2.2.1 release, as the v2.2.0 was already drafted. Thanks again for the PR!