Porting C code to m68k Assembly

[Gampy]

Hi all,

Anyone know how to convert this C code into assembly ??
I have used intermediate files to see how it assembles ...

C Instruction to convert to assembly

Code: Select all

#define HARDWARE_IO     (*(unsigned short *)(0xFFFFE2FA))
HARDWARE_IO |= 0x0001;

Intermediate files show it done like,

Code: Select all

move.l #-7430,%d0
move.l #-7430,%d1
move.l %d1,%a0
move.w (%a0),%d1
or.w #1,%d1
move.l %d0,%a0
move.w %d1,(%a0)

And the oppasite

Code: Select all

#define HARDWARE_IO     (*(unsigned short *)(0xFFFFE2FA))
HARDWARE_IO &= 0xFFFE;

Intermediate files show it done like,

Code: Select all

move.l #-7430,%d0
move.l #-7430,%d1
move.l %d1,%a0
move.w (%a0),%d1
and.w #-2,%d1
move.l %d0,%a0
move.w %d1,(%a0)

Cannot figure out what,

-7430

equates too. Normally I can figure this out, for some reason I cannot with this one!

This is where http://www.godbolt.org would be extremely handy if it did m68k, unfortunately it takes Linux to do the testing, I don't run Linux or I would add it, it is supposed to be easy to add by using another GNU implementation as a template.
An awesome task for a Linux user looking to help the community ...

To read the Howto on installing a new compiler, Select menu Other, then Wiki, and on the right sidebar of the Wiki, Select How to install a compiler.
Want to do it and need help, ask, I'll do whatever I can to help!

-Enjoy

[yoda69]

Not that I understand the coding thought I’d throw it into ChatGPT and ask to convert the first one from C to M68K and it spat out the following:

MOVE.W #0xFFFF, A0 ; Load the base address of HARDWARE_IO into A0
MOVE.W #0xE2FA, D0 ; Load the offset of HARDWARE_IO into D0
ADD.L A0, D0 ; Add the offset to the base address to get the full address of HARDWARE_IO
MOVE.W (D0), D1 ; Load the current value of HARDWARE_IO into D1
ORI.W #0x0001, D1 ; Set the least significant bit of D1 to 1
MOVE.W D1, (D0) ; Write the new value of HARDWARE_IO back to memory

Note that this assumes a 16-bit m68k architecture, where the ori instruction is used to set the least significant bit of HARDWARE_IO to 1. If the architecture is different, the instructions may need to be modified accordingly.


Doesn’t help with your 7430 problem and not even sure it’s correct but was interesting to see it spat out commented code.

[kur4o]

Code: Select all

move.w  (word_FFFFE2FA).w,d3
bset    #0,d3
move.w  d3,(word_FFFFE2FA).w

Test in a timed loop when bit5 fa19=1, if loop expires without=1 exit with a error
btst    #5,(word_FFFFFA19).w

Code: Select all

move.w  (word_FFFFE2FA).w,d3
bclr    #0,d3
move.w  d3,(word_FFFFE2FA).w

test bit5 fa19=0
btst    #5,(word_FFFFFA19).w

I see what you want to do, enable vpp voltage with that register.

Code compiler don`t need that IO defined, as it use it as a define for the cpu, cpu don`t care about it.

Here is example what need to be done with that register, basically switching a bit on/off.

It may be optimized or switch to C code that will do something like that. Read register switch bit on/off write to a register

On some pcms enable vpp all the time bogs VPW bus and is used only for a write event and than switched off.

[In-Tech]

Hiya,
Shoot, I wish I had a bit of time. I love assembly. Equates for your assembler engine is your friend I will dig into it soon so I can hopefully help with some stuff. I had 5 boats to do up in the Delta near Stockton, CA, then went to Reno to do a couple vehicles. Feast or famine

I was a bit grouchy till this happened. I count cards digitally in assembly

Jackpot.jpg (138.77 KiB) Viewed 4127 times

[ironduke]

Nice pics...
As far as Grampy's problem.. I have the answer, sorta.. I can explain the 7430 at least..
It's -7430 decimal.. that is also 0xffffe2fa which is in your first statement..

Or did I misunderstand the question?? Happens often, lol..

[Gampy]

kur4o,
Thank you sir! It worked ...
See more below ...

yoda69,
Thank you sir!

Shoot, I wish I had a bit of time. I love assembly. Equates for your assembler engine is your friend I will dig into it soon so I can hopefully help with some stuff.

You're late to the party again ...

Thanks to kur4o for kicking my final hurdle down we now have Intel Flash Sector erase in Assembly!
AMD is already done, untested, still needs testing, I just haven't done that yet, I may rewrite it after learning some things doing Intel.
No P12 AMD Unlock/Lock yet, that has still to be worked out.

We are getting closer and closer to Assembly Kernel Writes ... A new era for PcmHammer!

Some damn nice rides ... Makes me jealous, I miss being out and about doing that shit!

As far as Grampy's problem.. I have the answer, sorta.. I can explain the 7430 at least..
It's -7430 decimal.. that is also 0xffffe2fa which is in your first statement..

Or did I misunderstand the question?? Happens often, lol..

Already figured it out, but you are 100% correct!
Don't know why I had so much problem, I must have been tired, because now my formula works ...

Thanks y'all!

-Enjoy

[antus]

Good stuff! I just saw the thread, and picked the endianness and decimal hex, took a screen shot of the calculator and then saw the next 10 posts and all the answers. Good work! Good to hear the intel is working now. I have not been able to figure out why my implementation appears to be crashing the DLC when it doesn't use it but I'll stand by now, with your progress.