BREEBAAR%HLERUL5.BITNET@wiscvm.wisc.EDU (08/12/87)
I am currently working on an IBM AT with a Professional Graphics Display,
and am trying to rewrite my Turbo Pascal routines for communicating with
the PG-Adaptor in C, a language in which I have some programming
experience, but am no expert in.
I have to use absolute addressing, because of various buffers and variables
that both the PGA and I need to access. To illustrate, some of the original
Turbo Pascal declarations:
VAR outBuffer : ARRAY [0..255] OF BYTE ABSOLUTE $C000:$6000;
outHead : BYTE ABSOLUTE $C000:$6300;
outTail : BYTE ABSOLUTE $C000:$6301;
My problem ofcourse is: how do I do this in C? (I am using Turbo-C,
but have no manual (yet).)
Firstly I would like to know if Turbo-C has something equivalent to the
'ABSOLUTE' keyword. Probably not.
So secondly, I'd like the proper C solution to this.
I have some vague ideas myself, involving the assigning of values to pointers
but am at a loss at how to implement them, especially with IBM addresses
being 20 bits long etc. etc.
Can anyone out there help me?
Thanx.
Leo Breebaart (BREEBAART@HLERUL5.BITNET)
University of Leiden
The Netherlandsralf@nl.cs.cmu.edu (Ralf Brown) (08/14/87)
In article <8766@brl-adm.ARPA> BREEBAAR%HLERUL5.BITNET@wiscvm.wisc.EDU writes: >I have to use absolute addressing, because of various buffers and variables >that both the PGA and I need to access. To illustrate, some of the original >Turbo Pascal declarations: > > VAR outBuffer : ARRAY [0..255] OF BYTE ABSOLUTE $C000:$6000; > outHead : BYTE ABSOLUTE $C000:$6300; > outTail : BYTE ABSOLUTE $C000:$6301; > >My problem ofcourse is: how do I do this in C? (I am using Turbo-C, >but have no manual (yet).) >Firstly I would like to know if Turbo-C has something equivalent to the >'ABSOLUTE' keyword. Probably not. >So secondly, I'd like the proper C solution to this. > >Leo Breebaart (BREEBAART@HLERUL5.BITNET) >University of Leiden >The Netherlands The proper way to do this in C is to declare the type of the buffer, and then declare a pointer variable pointing to that type. The first thing to do before using the pointer variable is to assign it the address you need your buffers etc at. In Turbo C, the proper function is pointer = MK_FP( seg, ofs ) -- -=-=-=-=-=-=-=-= {harvard,seismo,ucbvax}!b.gp.cs.cmu.edu!ralf =-=-=-=-=-=-=-=- ARPAnet: RALF@B.GP.CS.CMU.EDU BITnet: RALF%B.GP.CS.CMU.EDU@CMUCCVMA AT&Tnet: (412) 268-3053 (school) FIDOnet: Ralf Brown at 129/31 DISCLAIMER? Who ever said I claimed anything? "I do not fear computers. I fear the lack of them..." -- Isaac Asimov
Wax.OsbuSouth@Xerox.COM (08/19/87)
If you assume that "long int" and "far *" are the same size, you could probably do:
union {
far unsigned char (*ptr)[]; /* or just "far unsigned char *ptr" */
long int outBufferAddress;
} outBuffer;
...
outBuffer.outBufferAddress = 0XC0006000L;
...
*outBuffer.ptr = ...
... = *outBuffer.ptr
Allan Wax
Wax.OSBUSouth@Xerox.COMbpm@psueclb.BITNET (08/22/87)
>[I] am trying to rewrite my Turbo Pascal routines for communicating with >the PG-Adaptor in C, a language in which I have some programming >experience, but am no expert in. >I have to use absolute addressing, because of various buffers and variables >that both the PGA and I need to access. To illustrate, some of the original >Turbo Pascal declarations: > > VAR outBuffer : ARRAY [0..255] OF BYTE ABSOLUTE $C000:$6000; > outHead : BYTE ABSOLUTE $C000:$6300; > outTail : BYTE ABSOLUTE $C000:$6301; > To declare an absolute pointer in Turbo C, use the "far" adjective. For example, the equivalent to the above is: char far outBuffer[256] = 0xC0006000; char far *outHead = 0xC0006300; char far *outTail = 0xC0006301; --whiffle
BREEBAAR%HLERUL5.BITNET@wiscvm.wisc.EDU (09/03/87)
My sincere thanks to all the kind people who took time to answer my question about absolute adressing in Turbo-C. For those interested, the -for me- handiest way to say Turbo Pascal's VAR outHead : BYTE ABOLUTE $C000:$C300; in C is: #define outHead ( *(unsigned char far *)0xC000C300L ) Of course I had not realized at first that 'BYTE' should be translated as 'unsigned char', instead of just 'char', to get the range 0..255; and it took me half an hour to discover this. Well, I *said* I was no expert. But I feel somewhat comforted when I realize that of the thirteen people who responded, only *one* pointed it out. I also have another question. It's probably a very old one, but please bear in mind that I have no local C-gurus to consult, and that reference books never seem to say anything definite about the subject (at least the ones I've read, don't, and I just *can't* afford K&R; why is that book so horribly expensive?). It's this: Just WHY is it in C 'better' to write: *(a+n) instead of: a[n], where 'a' is an array of something? And how much 'better' is it? I remember reading on this list some time back that this does not extend to arrays of more than one dimension. Why? Just because the expressions will get too complex? Hope you will help me with this too. Leo Breebaart (BREEBAAR@HLERUL5.BITNET) Univ. of Leiden The Netherlands
guy%gorodish@Sun.COM (Guy Harris) (09/03/87)
> It's this: Just WHY is it in C 'better' to write: *(a+n) instead of: a[n], > where 'a' is an array of something? And how much 'better' is it? It isn't any "better"; in fact, it's exactly the same. "a[n]" is defined as "*(a+n)". The only ways in which one would be "better" would either be that one is somehow stylistically, or that one generates better code. Different people have different preferences in style, and different compilers may well not always generate better code for the same version (our compiler generates the same code for both of them), so it's not clear there is a single style that could be considered "better" in any objective sense. > I remember reading on this list some time back that this does not extend to > arrays of more than one dimension. Why? Just because the expressions will > get too complex? No, this extends quite well to arrays of more than one dimension: "a[i][j]" is (a[i])[j], and thus is *((*(a+i)) + j). There may be some confusion because "*(a+i)" is an array-valued expression, and K&R doesn't discuss this notion; however, ANSI C does - array-valued expressions, in most contexts, are converted to pointer-valued expressions whose value is a pointer to the first element of the array. This is the obvious generalization of the K&R rule that the name of an array, in most contexts, is converted to a pointer to the first member of that array; this generalization is implemented by our C compiler, and probably by most if not all other PCC-based compilers and most likely by many other compilers as well. Guy Harris {ihnp4, decvax, seismo, decwrl, ...}!sun!guy guy@sun.com
gwyn@brl-smoke.ARPA (Doug Gwyn ) (09/04/87)
In article <9140@brl-adm.ARPA> BREEBAAR%HLERUL5.BITNET@wiscvm.wisc.EDU writes: >It's this: Just WHY is it in C 'better' to write: *(a+n) instead of: a[n], >where 'a' is an array of something? And how much 'better' is it? It's NOT "better". It's not particularly worse, either, although if a is declared as an array (rather than a pointer) the a[n] form is generally better style.
bright@dataio.Data-IO.COM (Walter Bright) (09/04/87)
In article <9140@brl-adm.ARPA> BREEBAAR%HLERUL5.BITNET@wiscvm.wisc.EDU writes:
#It's this: Just WHY is it in C 'better' to write: *(a+n) instead of: a[n],
#where 'a' is an array of something? And how much 'better' is it?
#I remember reading on this list some time back that this does not extend to
#arrays of more than one dimension. Why? Just because the expressions will
#get too complex?
*(a+n) and a[n] are exactly equivalent expressions. Neither is better as
far as the compiler is concerned. As a programming style issue, however:
*(a+n) should be used when a is a pointer
a[n] should be used when a is an array
This makes the code a little more readable.
In fact, I would have enforced this in the language if I had designed it.
This would have eliminated the 'pointer-array equivalence' kludge and
all its associated kludges and confusion in the language.