hoford@sequoia.circ.upenn.edu (John Hoford) (04/30/91)
I have been trying to get a handle on writing asmbly code to for the hp48 and I have a few questions. Q2 When I convert an object to asm using "->asc" I get the object plus a checksum appended to the end. How do I calculate the checksum To hand code my own objects (using "asc->")? Q2 I have been using "MLDL 1.01" by Jan Brittenson to look at machine language code in the hp rom. This is the most useful program for learning asm, great work! My problem is it does not have a simple way of getting the starting address of a program in a variable. For Ml programs the only way I can do this now use "asm->" and then hand dissemble the code. So the question: Does anyone have a program That returns the address of the start of a stored program ? ie. 1: label ->addr 1: #213FF Would the address change to quickly to be useful? This would also help me check my code after I write it. I would be great to have a library that had all of the MLDL code "asc->" "->asc" "peek" @ some sort peek function "->addr" @as described above Q3 The STAR (thanks again Jan Brittenson) assemblers seem to produce raw ml. Is their a progam which will put them into a "asc->" or a "hp/kermit file" format? Optionally put a #02DCC (code object) in front of it Thanks John D. Hoford
bson@rice-chex.ai.mit.edu (Jan Brittenson) (04/30/91)
In a posting of [29 Apr 91 22:05:09 GMT] hoford@sequoia.circ.upenn.edu (John Hoford) writes: > Q2 > I have been using "MLDL 1.01" ... My problem is it does not have a > simple way of getting the starting address of a program in a variable. Recall the variable and pass the CODE object as an argument on level 1. > So the question: Does anyone have a program That returns the address > of the start of a stored program ?' In MLDL 1.02 the XLIB 1092 7 takes a code object and returns its address plus 10 in both level 1 and 2. So, to get the address of a CODE object in level 1 the following program can be used: \<< 1092 # 18CEAh SYSEVAL # 4035h SYSEVAL # 7E50h SYSEVAL EVAL DROP # 59CCh SYSEVAL 10 - \>> Notice that MLDL 1.02 must be attached, and it will *not* work with 1.01. > ie. > label ->addr 1: #213FF > Would the address change to quickly to be useful? It's a ROM address. It won't change. If it were a RAM address, then generally, the answer is no. Programs and directories move around much less than other objects. > The STAR assemblers seem to produce raw ml. Is their a progam which > will put them into a "asc->" or a "hp/kermit file" format? You can use the HEADER macro defined in hp48.star, it will create a kermit header. If you want a code object, you can use the CODE...ENDCODE macro. Just make sure to assemble with hp48.star. You can use the following skeleton: ;;+ ;; HP-48 ML Skeleton Program ;; ;; Assemble with hp48.star: ;; ;; % star listing binary hp48 skeleton ;; ;;- ; This is the kermit header. header ; This is a CODE...ENDCODE macro invocation. ; Don't break the code object into several files. code ; the code goes here move.a @d0, a ; Continue RPL thread add 5, d0 ; -- "" -- jump @a ; -- "" -- endcode ;;--- -- Jan Brittenson bson@ai.mit.edu
akcs.falco@hpcvbbs.UUCP (Andrey Dolgachev) (05/01/91)
Jan, for those of us who don't have STAR, could you please also post you
your machine-code routines in a format which can be simply d/l'd and
ASC-> ed? Hmm, I was thinking along the lines of putting the ML code
alongside your documentation, but it would prob. be easier to just ASC->
the CODE and attach that at the end of the program. How about it, I
would appreciate and I'm sure others would also.
Thanks in advance,
--Falcobson@rice-chex.ai.mit.edu (Jan Brittenson) (05/01/91)
In a posting of [30 Apr 91 22:40:19 GMT] akcs.falco@hpcvbbs.UUCP (Andrey Dolgachev) writes: > Jan, for those of us who don't have STAR, could you please also post > you your machine-code routines in a format which can be simply d/l'd > and ASC-> ed? Hmm... it took me a while, but I finally think I know what you are refering to. Some guy at U Penn asked how to create kermit headers and code objects with STAR. What I provided was a skeleton to be filled in with the actual code. The skeleton doesn't do anything by itself, and was intended only as an example. I apologize for not making this clear in the message. -- Jan Brittenson bson@ai.mit.edu
herman@corpane.uucp (Harry Herman) (05/02/91)
In <15465@life.ai.mit.edu> bson@rice-chex.ai.mit.edu (Jan Brittenson) writes: >In a posting of [30 Apr 91 22:40:19 GMT] > akcs.falco@hpcvbbs.UUCP (Andrey Dolgachev) writes: > > Jan, for those of us who don't have STAR, could you please also post > > you your machine-code routines in a format which can be simply d/l'd > > and ASC-> ed? > Hmm... it took me a while, but I finally think I know what you are >refering to. Some guy at U Penn asked how to create kermit headers and >code objects with STAR. What I provided was a skeleton to be filled in >with the actual code. The skeleton doesn't do anything by itself, and >was intended only as an example. I apologize for not making this clear >in the message. > -- Jan Brittenson > bson@ai.mit.edu Jan, I suspect he may have been talking about your machine code examples for things like the keyboard buffer routines you posted a couple of weeks ago. They are apparantly written in STAR and for those of us who do not have STAR and don't want to hand assemble the routines, having ASC-> files available would be nice. Harry Herman herman@corpane
akcs.falco@hpcvbbs.UUCP (Andrey Dolgachev) (05/03/91)
Jan, sorry I didn't make it clear , but Harry Herman was right, I meant
that for those of us who don't have STAR, we can't use your examples of
machine code like the keyboard buffer, or the sample for i/o, or any of
the other ML code that you post. If you could post the machine code
instruction alongside for hand assembly, or better yet, just post your
machine code programs in ASC format, I would be very grateful.
--Falco