RDROYA01@ULKYVX.BITNET (Robert Royar) (05/25/86)
Hello again,
I now have the problem with background CP/M licked, I believe. This
morning I worked out a way to let the ccp have control until a ^C is typed
and I patched the CPM.SYS file to leave traps 0 & 1 alone. This means that,
if they are not set by some program, they point to $0. Could this be why
the BDOS re-inits them each time, just so they will point to the default
exception handler? Now, if I call a trap #0 before installing the handler,
I get a "User Exception 3 at xxxxx Aborted." So simply changing the
compare in _initexc is not the perfect solution. But what I have gotten
for this is a way of running the system from within a loaded program
without having to rewrite the ccp, or set-up LPBs, FCBs, or DMAs.
The submit function works, and since warmboot is side-stepped (what
potential havock is this leading to?), loads and exits are much faster. I
have been able to compile, link, reloc, and load a test file all in one
background submit function, hit ^C and pop back into the source code. The
debugger will not load. It aborts with an address error, but a ^C takes
back me to the source code from that error also. Since STAT must reset the
login vector to get the Disk parameter block, file operations may not go as
planned there needs to be a return login vector befor entering the CCP and a
replace original before calling another program. The difficulty here comes
when you actually want to change the logged drive. It's easier just to load
all files into the calling program with the disk specification explicitly
stated and to call all background operations in the same way. Since STAT
seems to be the only problem here, resetting the drives manually after
calling STAT to check on a drive that is not the default seems a resonible
alternative to a lot of extra testing in the code.
For any interested programmers here is the code I now have:
.globl _initv
_initv: move.l $8c,$80 * set up trap #0 from trap #3
move.l $88,$84 * set up trap #1 from trap #2
move.l #_ccp,$90 * set up trap #4
move.l #hndlr,$88 * set new trap #2 handler
rts
hndlr: movem.l d1-d7/a0-a6,regs1 * handles calls to reset
cmp.w #$0,d0 * request for warm boot?
beq ccp * must have been from another program
trap #1 * just goto ccp
movem.l regs1,d1-d7/a0-a6
rte
.globl _ccp
.globl _fix
_ccp: movem.l d0-d7/a1-a7,regs3 * access to ccp through trap #4
move.l usp,a0 * save USP for later return
move.l a0,_usp
move.l #bios_tr,$8c * set temporary trap #3 handler
ccp: jsr $234a * address of _ccpmain on HSC board
bra ccp: * loop forever (until ^C)
_fix: move.l $80,$8c * reset the original trap #3 handler
movem.l regs3,d0-d7/a1-a7 * gets out of disasterous mistakes
move.l _usp,a0 * replace user stack pointer
move.l a0,usp
rte
bios_tr:
cmp #3,d0 * was it a conin request?
beq conin
trap #0 * no, process the request
rte * return
conin: trap #0 * get the character
cmp #3,d0 * was it a ^C?
beq _fix * yes, go back to calling program
rte * no return from exception d0=requested char
regs1: ds.l 14 * register storage for trap handler
regs3: ds.l 15 * register storage for _ccp and _fix
_usp: ds.l 1 * value of USP for _ccp and _fix