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

[Toni Wilen <twilen@xxxxxxxxxx>]

>> I'll see whether i can update the test program
> 
> I've done that now. You'll have to change your test program to pass the
> CPU type just before the included *.sa files for this to work (see
> https://github.com/th-otto/fpsp/blob/f247747c396dc1811c30b09f3606b161d22850b6/fpsp060/fpsptst.s#L562
> <https://github.com/th-otto/fpsp/blob/f247747c396dc1811c30b09f3606b161d22850b6/fpsp060/fpsptst.s#L562>
> ) . Or just ignore the 3 failures for now ;)
> 
> Changed *.sa can now be found here
> <http://tho-otto.de/download/fpspsa.zip> (are you still not able to
> compile that yourself? Maybe i can include the amiga test program in the
> repo, too)
> 
> The only remaining failure now is the SNAN test.

I get expected SNAN exception if I run following code (important bits
copied from test code):

	lea data,a0
	fmove.l #$00004000,fpcr
	move.l	#$ffff0000,(a0)
	move.l	#$00000000,4(a0)
	move.l	#$00000001,8(a0)
	fmovem.x (a0),fp0
	fadd.b #2,fp0
	fmovem.x fp0,(a0)

FPIAR is also set correctly.

FPU  arithmetic exception pending: FPSR: 09004080, FPCR: 4000 (vector:
54) PC=0865cebc!
FPU ARITHMETIC EXCEPTION (54) PC=0840209c

NaN test does look correct in softfloat emulation, I mistook invalid
encoding test (that does not check for SNaN) as NaN test.

(I'll try to compile the test program later)