When looking at a function with an inline/absolute jump table in an .s file, it splits out and is treated like a separate entity, rather than being baked into the function.
However, if you don't mark it as a jump table, this happens:
The jump table gets baked into the function, but it's being interpreted as instruction bytes instead of raw memory addresses, resulting in those wrong loads.
jeff needs to be tweaked in such a way that it can recognize that this is a jump table (and the entries are interpreted as addresses), but not create a distinct object for it.
When looking at a function with an inline/absolute jump table in an .s file, it splits out and is treated like a separate entity, rather than being baked into the function.
However, if you don't mark it as a jump table, this happens:
The jump table gets baked into the function, but it's being interpreted as instruction bytes instead of raw memory addresses, resulting in those wrong loads.
jeff needs to be tweaked in such a way that it can recognize that this is a jump table (and the entries are interpreted as addresses), but not create a distinct object for it.