[jsa]

[CF75] contains a value of 1sec

and SAD shows its use like this

Code: Select all

b1d3: d3,03             jnc   b1d8           if ((uns) 0 < cf75) goto b1d8;

I am thinking it should be

Code: Select all

b1d3: d3,03             jnc   b1d8           if ((uns) 0 < [cf75]) goto b1d8;

Like

Code: Select all

b1f0: d9,0b             jgtu  b1fd           if ((uns) Rbe > [cf7f] ) goto b1fd

[sailorbob]

The jnc is jumping according to the state of the carry flag, the state of which is determined by what's happened in the subroutine at 0x0x8099 not the contents of 0xCF75.

[jsa]

My query is not about the reason for the jump, but SAD not applying a defined symbol.

Here's an example without jump.

Directive contains

Code: Select all

SYM c20a "SomeSym"

SAD outputs

Code: Select all

be9b: 45,46,00,f2,38    ad3w  R38,Rf2,46     R38 = c20a;
bea0: ef,fb,b9          call  789e           STabLU();

This would be consistent with other opcodes

Code: Select all

be9b: 45,46,00,f2,38    ad3w  R38,Rf2,46     R38 = SomeSym;
bea0: ef,fb,b9          call  789e           STabLU();

[sailorbob]

I don't think you can realistically expect the disassembler to apply a symbol, or an address, in line 0xB1D3 as it does not know what state the carry flag will be in after the subroutine at 0x8099 has been executed (and would not know unless it simulated the code running with values for the various times and temperatures the routine uses).

[jsa]

SAD recognises the lookup subroutines and outputs the row and column registers to the msg file. So why not do that for the address input register, r38 in this case, then step back through the code to the previous R38= to nail the address and apply the symbol.

[sailorbob]

If you look at the subroutine at 0x8099 you'll see that the value in the entry address passed by 0x30 is not necessarily the one that gets used to determine the state of the carry flag. The following four addresses also get used in the subroutine.

[tvrfan]

-- BUMP --

Just pushing this title to the top to let everyone know that the latest version of SAD is now posted to Github, and has source code,
exe for Windows and 64 bit linux, and (should) have all reported bugs fixed...

I have not changed the command structure so all existing directive files, comment files etc should work just as before (unless I've introduced a bug somewhere). Download address in my sig below.....