>  TT> I also read that it ran fine on James Vahn's machine.. 
> Yes. I have compiled another version of my Linux-kernel (2.0.30),
> and now it doesn't generated a seg-fault. But I have (see a
> previous message of me) some other problems now.....
And like a kid with a new toy, I've been playing some more.. 
From what I've been finding is that while most libc calls are
straightforward in their useage of the stack, ctime isn't.
Further, it seems that ECX and EDX are usually modified on 
return from calls and meaningless, while EAX contains meaningful 
data (a pointer to the last result?).. EBX seems unchanged. 
EAX can be used to return an exitcode but I'm not sure of the 
format- experiments with gdb have been as confusing as usual.
; nasm ddate.asm 
; gcc ddate.o -o ddate
        global main
        extern time
        extern ctime
        extern printf
section .text                   
main:
        push ebp                ;Save these.
        mov ebp,esp             ;
        
        sub esp,4               ;Adjust stack.              
        push esp                ;Begin to build a string
        call time               ;  on the stack for printf.
        add esp,4
       
        mov eax,esp             ;Find our data on the stack.
        push eax                ;     
        call ctime              ;Convert time to ASCII.
        add esp,4               ;Adjust stack.
        push eax                ;Finish the printf string.
        push dword .printdec    ;   
        call printf
       
        xor eax,eax             ;Exitcode 0.            
        mov esp,ebp             ;Clean up and
        pop ebp                 ; exit.
        ret
section .data                   
.printdec       db "===========>>> %s",0
--- ifmail-tx (i386 Linux)
---------------