Maelgrum wrote on 2023-10-04, 13:06:

Logs are dumped on each line
Scanning process not stops on first found

log attached from that version - attached - all failed. currently testing Starge5, will report back soon...

One more time please !
I feel we are very very close ...

Maelgrum wrote on 2023-10-04, 13:17:

And stage 5, with another search.
This log and previous log (one post up) i hope gives exact address location.

Stage 5 log, same results as with Stage 4. Now testing the recent versions of both.

Gmlb256 wrote on 2023-10-04, 13:33:

Stage 5 log, same results as with Stage 4. Now testing the recent versions of both.

same here, but attaching it anyway... OK, going to run "sbcrack_st5.zip " and "sbcrack_st4.zip"...

Maelgrum wrote on 2023-10-04, 13:28:

One more time please !
I feel we are very very close ...

New logs, no success from both.

I guess good news from my side: "sbcrack_st5.zip" - gives several passes - the log as usual - attached. P.S. "sbcrack_st4.zip" - are all "failed" for me.

[EDIT] but we have maybe now too many "passes", I counted 37:

1$ cat LOG.TXT | grep pass | wc -l
237

Gmlb256 wrote on 2023-10-04, 13:53:

Maelgrum wrote on 2023-10-04, 13:28:

One more time please !
I feel we are very very close ...

New logs, no success from both.

On stage 4 - It must see old matches - 0x1183, 0x116c
Looks like it in invalid internal state.
Will previous version give positive match - 0x1183, as it was ?

Maelgrum wrote on 2023-10-04, 14:10:

Will previous version give positive match - 0x1183, as it was ?

0x1183 is failed for me with ST4 version.

mattw wrote on 2023-10-04, 14:12:

Maelgrum wrote on 2023-10-04, 14:10:

Will previous version give positive match - 0x1183, as it was ?

0x1183 is failed for me with ST4 version.

May be power off/on needed ?

Why not try address search on 4.13 and see if sbcrack finds correct addresses that you already know?

Maelgrum wrote on 2023-10-04, 14:14:

May be power off/on needed ?

always, do that.

georgel wrote on 2023-10-04, 14:15:

Why not try address search on 4.13 and see if sbcrack finds correct addresses that you already know?

I was doing to suggest the same, i.e. test if "E3-search-algo" works with V4.13, because V4.13 we know what it is supposed to return.

I will try again to see if there is any difference with the recent versions.

georgel wrote on 2023-10-04, 14:15:

Why not try address search on 4.13 and see if sbcrack finds correct addresses that you already know?

If needed i can put in a 4.13 one so you dont have to keep switching stuff around.
Just need to know which version (stage) to use for probing.

Ok, we can try to make it with 4.13. Its reasonable suggestion. But tomorrow ))

What do you assume SP to be at MIDI byte write routine in 4.16?

Maelgrum wrote on 2023-10-04, 14:21:

But tomorrow ))

yep, lets rest and think about it for awhile. BTW, about 1183 - I mentioned here that it stopped giving "passed" result:

Re: The Soundblaster DSP project

but the 1st version still gives "passed" for me on 1183.

mattw wrote on 2023-10-04, 14:29:

yep, lets rest and think about it for awhile. BTW, about 1183 - I mentioned here that it stopped giving "passed" result: […]
Show full quote

Maelgrum wrote on 2023-10-04, 14:21:

But tomorrow ))

yep, lets rest and think about it for awhile. BTW, about 1183 - I mentioned here that it stopped giving "passed" result:

Re: The Soundblaster DSP project

but the 1st version still gives "passed" for me on 1183.

Best would be to go by stage i recon, using 4.13 as 'known result' and put 4.16 next to it.
The offset should be known (as that was the first test i believe?) and based on the known code decompile for 4.13 you can approximate in which section of the code you are. (assuming it didnt change drastically)
(Hopefully that makes sense)

georgel wrote on 2023-10-04, 14:29:

What do you assume SP to be at MIDI byte write routine in 4.16?

0xC0 at command loop
0xC2 at entering midi loop
0xC4 in midi byte write routine

Thanks to all this massive testing !

mattw wrote on 2023-10-04, 14:29:

yep, lets rest and think about it for awhile. BTW, about 1183 - I mentioned here that it stopped giving "passed" result: […]
Show full quote

Maelgrum wrote on 2023-10-04, 14:21:

But tomorrow ))

yep, lets rest and think about it for awhile. BTW, about 1183 - I mentioned here that it stopped giving "passed" result:

Re: The Soundblaster DSP project

but the 1st version still gives "passed" for me on 1183.

Thats strange, what last version gives nonsense, but old gives something more sane. Will look into this.