Re: [hatari-devel] Possible bug in 060 exception stack frames

[ Thread Index | Date Index | More Archives ]

 > Yes, I tested it with real 68040, those 3 instructions won't modify


Thats strange and does not make much sense to me, especially because it apparently only affects 040. I'll see whether i can update the test program so it does not report failures for this when run on 040.

68881/68882 also does not update FPIAR in this situation.

FPIAR behavior has changed between 6888x and 68040/060.

68881/2: FPIAR is only modified if instruction can generate arithmetic exception and at least one arithmetic exception is enabled in FPCR. This is also mentioned in 68881/2 documentation.

But 68040 and 68060 will always modify FPIAR if instruction can generate any arithmetic exception, even if arithmetic exceptions are disabled. FPIAR is also modified if instruction generates non-arithmetic exception that 6888x does not generate, for example if instruction or addressing mode is unimplemented (that is implemented in 6888x).

I can't be 100% sure but so far this seems to match real hardware behavior.

Mail converted by MHonArc 2.6.19+