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

[Eero Tamminen <oak@xxxxxxxxxxxxxx>]

Hi,

Non-booting 060 demos work with FPSP060.
I've updated the docs accordingly:
https://git.tuxfamily.org/hatari/hatari.git/commit/?id=465915e1d5c93ee4cbd752802fac9bdd1cda38b3

Fit's demos have rendering issues.  Some of
them look like they could be due to incorrect
calculation results.


	- Eero

On 1/18/21 8:53 PM, Eero Tamminen wrote:
> Building requires:
>      ln -s src/hdr.fpu .
>
> Otherwise I get [1].
>
> [1]
> $ git clone https://github.com/th-otto/fpsp.git
> Cloning into 'fpsp'...
> remote: Enumerating objects: 334, done.
> remote: Counting objects: 100% (334/334), done.
> remote: Compressing objects: 100% (172/172), done.
> remote: Total 334 (delta 166), reused 329 (delta 161), pack-reused 0
> Receiving objects: 100% (334/334), 858.99 KiB | 1.07 MiB/s, done.
> Resolving deltas: 100% (166/166), done.
> $ cd fpsp/fpsp060/
> fpsp/fpsp060$ ls
> CHANGES  doc          fpsp060.s    fpsptst.s  MISC    src
> dfpsp.c  fpsp060.prj  fpsptst.prj  Makefile   README
> fpsp/fpsp060$ make
> cat src/freal.S src/x_fovfl.S src/x_funfl.S src/x_funsupp.S 
> src/x_effadd.S src/x_foperr.S src/x_fsnan.S src/x_finex.S src/x_fdz.S 
> src/x_fline.S src/x_funimp.S src/fsin.S src/ftan.S src/fatan.S 
> src/fasin.S src/facos.S src/fetox.S src/fgetem.S src/fcosh.S src/fsinh.S 
> src/ftanh.S src/flogn.S src/fatanh.S src/flog2.S src/ftwotox.S 
> src/fmovecr.S src/fscale.S src/frem_mod.S src/fkern.S src/fgen_except.S 
> src/foptbl.S src/fmul.S src/fin.S src/fdiv.S src/fneg.S src/ftst.S 
> src/fint.S src/fintrz.S src/fabs.S src/fcmp.S src/fsglmul.S 
> src/fsgldiv.S src/fadd.S src/fsub.S src/fsqrt.S src/fmisc.S src/fdbcc.S 
> src/ftrapcc.S src/fscc.S src/fmovem.S src/fctrl.S src/fcalc_ea.S 
> src/fmem.S src/fout.S src/ireg.S src/fdenorm.S src/fround.S src/fnorm.S 
> src/foptag_set.S src/fresult.S src/fpack.S src/fdecbin.S src/fbindec.S 
> src/fbinstr.S src/faccess.S  > fpsp.S
> m68k-atari-mint-gcc -m68060 -Isrc -Wall -Werror -s -Wl,--oformat,binary 
> -o fpsp.bin -nostartfiles -nostdlib fpsp.S
> /tmp/cc9R0J34.s: Assembler messages:
> /tmp/cc9R0J34.s:300: Error: can't open hdr.fpu for reading: No such file 
> or directory
> /tmp/cc9R0J34.s:479: Error: can't open hdr.fpu for reading: No such file 
> or directory
> /tmp/cc9R0J34.s:673: Error: can't open hdr.fpu for reading: No such file 
> or directory
> ...
> (lots of errors)
>
> On 1/18/21 6:29 PM, Thorsten Otto wrote:
> > On Montag, 18. Januar 2021 17:20:20 CET Thorsten Otto wrote:
> > > I've also pushed the repo with the recent c
> >
> > Forgot to mention: the debug output uses 2 new callouts in the test 
> > program.
> > So if you include the *.sa files in your amiga program, you have to 
> > define
> > them. The actual functions start at https://github.com/th-otto/fpsp/blob/
> > c75e8b20fddc3ede91198a8e587bbe1d0b8db460/fpsp060/fpsptst.s#L127
> > and have to be added to the callout table 
> > https://github.com/th-otto/fpsp/
> > blob/c75e8b20fddc3ede91198a8e587bbe1d0b8db460/fpsp060/fpsptst.s#L554