Per title. The create_config.py script does not work well for PSX binaries:
The initial_pc value from the header is the entrypoint for the binary, but the entrypoint is not the first function within the binary
The value of gp given in the header is nonsense, if used the $gp register's value is set in the entrypoint function
the data* and bss* values in the header also appear to be nonsense.
I have a PR incoming that addresses these in a best-efforts manner. We will scan through the binary until we find a threshold number of valid MIPS I instructions (and we will call this the start of the text segment), and continue until we reach the first invalid instruction (we will call this the end of the text segment and the beginning of the data segment).
This can perhaps be improved upon in due course, but is much better than the current version of create_config.py.
Per title. The create_config.py script does not work well for PSX binaries:
initial_pcvalue from the header is the entrypoint for the binary, but the entrypoint is not the first function within the binarygpgiven in the header is nonsense, if used the $gp register's value is set in the entrypoint functiondata*andbss*values in the header also appear to be nonsense.I have a PR incoming that addresses these in a best-efforts manner. We will scan through the binary until we find a threshold number of valid MIPS I instructions (and we will call this the start of the
textsegment), and continue until we reach the first invalid instruction (we will call this the end of thetextsegment and the beginning of thedatasegment).This can perhaps be improved upon in due course, but is much better than the current version of create_config.py.