Segmentation in real mode
in [ASM]
|
Prev: stricmp
Next: Killing Explorer
From: Tom on 12 Mar 2005 12:27 Herman Dullink wrote: > > 1. Why do segment registers hold segments' paragraph numbers rather > > than their absolute addresses? > This is Intel's first attempt to break the 64KB limit of a 16-bit CPU. Using > paragraph addresses expands the total addressable space to 1MB. This allows > definition of several variable sized segments. Because of lack of > protection, it's possible to address outside the segment... > This is only the case on 8086 and compatible CPUs, and in Real Mode on 80286 > and compatible CPUs. In protected mode, segment registers hold an index in a > segment descriptor table, which not only hold a base (absolute) address, but > also its size (which is checked for every access) and privileges > (level,read/write/execute). > > > > What does the second > > "end Main" expression stand for? Why is this procedure not called? Or > > is it? > The "end" is to tell the assembler that this is the end of the source code. > The "Main" is to set the entry point. The linker will put this address in > the header of the executable file. DOS will then execute from this address > when this file is executed. > If e.g. you would put a 2nd procedure in this source file (MyMain) and use > the name of that 2nd procedure after end, then that 2nd procedure is run, > and not "Main". I see. Thank you. Tom
From: Tom on 12 Mar 2005 12:30 Frank Kotler wrote: > Tom wrote: > > > > Hello, > > For some time now I have been learning assembly on my own using > > Webster. So far I have managed to build a small paper-and-pencil uP and > > read/comprehend most of the material from the 16-bit DOS version of AoA > > (I understand DOS is obsolete, but I have found it simpler to start > > with, hope to do better in the future). > > I agree. You'll find some people feel that you shouldn't > "waste" any time at all learning dos. I'd agree that it > isn't worthwhile learning "dos", per se, in any depth, but I > agree with you that it's "simpler" to interface with dos > than with Windows or Linux. Linux has an "int 80h" interface > - similar to the dos "int 21h" interface, but Windows > requires calling a "Dynamically Linked Library". Easy enough > to *do*, but harder to understand what you're *doing* (IMO). Yes, thank you. I know, I agree. > 32-bit programming, in itself, is *easier* than 16-bit > programming - no worries about segments, and the addressing > modes are much more flexible. Once you feel ready to make > the switch (I'd learn how "call" and "ret" work - how to use > the stack to pass parameters, and keep the stack "sane"), I > think you'll like 32-bit programming... so don't waste *too* > much time with dos! I know this is a good suggestion. Thank you. > > Would you be kind enough as to help me clarify the following points > > (regarding real mode programming): > > > > 1. Why do segment registers hold segments' paragraph numbers rather > > than their absolute addresses? > > Funny you should ask... (see the thread about ".model" in > Tasm). It's either: > > 1) Because that's the chip the market *demanded* (at *that* > time). Easy "upgrade path" for 8080 code, and ability to use > more than 64k without "wasting" much memory. (my theory) Or: > > 2) Intel was really stupid. (Beth's theory) > > In any case, I think we agree that it wouldn't be a sensible > way to design a chip *now*. For purposes of *using* it, > rather than trying to improve on Intel's design, just assume > "that's the way it is". > > > 2. In Chapter 6 of AoA (The instruction set, 6.11.1 Simple arithmetic > > I) Mr. Hyde gives an example of code in which he declares a procedure > > MAIN, which is duly terminated with "Main endp". What does the second > > "end Main" expression stand for? > > It isn't really "second" - "end" is different from "ends". > "end" tells Masm that your code ends - anything after that > is ignored - *and* the label after "end" defines where the > entrypoint to the code is - this example probably doesn't do > it, but you could put subroutines *before* your "Main", and > execution would still start at "Main:"... > > For comparison - a ".com" file always starts at the > beginning of your code. Other assemblers indicate the > entrypoint differently - Nasm, for example, uses "..start:" > and doesn't use "end" or "ends" at all. > > > Why is this procedure not called? Or > > is it? > > No, it's not. Personally, I think "Main" is a potentially > confusing name to use, if it isn't a "C-style" main - that > is, called (so a "ret" or "retf" will exit it), and with > argc, argv, and envp on the stack. That isn't the situation > with an asm program unless you write the code to do it. > Or... you can write an asm file using "main" (probably have > to be "_main", actually), assemble it to an .obj file, and > use a C compiler to link it... *with* C startup code. In > this case, "main" needs to be named "main". Otherwise, *I'd* > call it something else. Right. Thank you. > > Thank you very much for all kind replies. > > You're welcome. Thank *you* for the on-topic question! Thank YOU. > Ignore any "less-than-kind" replies you get, and ignore any > newbie twirps who advise you to take your questions to clax! :-) > We *do* discuss assembly language around here... > sometimes... Yes, you do, and I am always most happy to read them. Thank you. Tom
From: Tom on 12 Mar 2005 12:37 Beth wrote: > Hi :) Hello Beth, > > For some time now I have been learning assembly on my own using > > Webster. So far I have managed to build a small paper-and-pencil uP and > > read/comprehend most of the material from the 16-bit DOS version of AoA > > (I understand DOS is obsolete, but I have found it simpler to start > > with, hope to do better in the future). > > Cool :) The coolest of all thinks one can do. I know. :-) > > Would you be kind enough as to help me clarify the following points > > (regarding real mode programming): > > > > 1. Why do segment registers hold segments' paragraph numbers rather > > than their absolute addresses? > > Well, 16-bits (the size of the original 8086's registers) can only address > 64KB maximum (2^16 = 65536 bytes or 64KB)...some means is, therefore, > required to "extend" the accessible range to beyond the 64KB that 16-bits > alone can address... > > As to why Intel used _this particular method_ of doing it rather than some > other schemes...well, goodness knows why... I see. :-) > it doesn't seem particularly sensible to me at all... Right. :-) [snipped the stuff I learnt from AoA] > Absolute address = (segment * 16) + offset Yes. Why "the miserable 16"? Rather than: > Absolute address = segment + offset [...] > Right...basically because "endp" is for ending the procedure..."end" is for > ending the source file...different purposes (though, admittedly, the way > they read, it does seem very odd indeed...the way the "end" directive > specifies the _start_ of all things is, indeed, a rather confusing facet of > this kind of syntax...the NASM assembler (and others similar to it) don't > bother with the "end" directive at all because it is kind of confusing and > seems "unnecessary" with modern machines to have to put "end" at the end of > the source file, rather than simply using the "end of file" as the end > :)... Yes. Thank you. > In generic terms: > > <label> endp - ends procedure called "label" > > end [<label>] - ends source file and optionally specifies the > "entry-point" - start address - for execution of the code That's it! Thank you! Loud and clear! :-) > Simply, the two serves completely different purposes...and it's just > "confusing phrasiology" that it does, indeed, _LOOK_ like they seem to be > saying the same thing...just a poor choice of name for the directives, > really...yes, as weird as it is, "end Main" is NOT actually saying "end of > main", it's saying "end of source file...oh, by the way, start the program > at Main"...but, yeah, it sure doesn't "read" that way, does it? Just a > "quirk" of this style of syntax... This is just that I am still a novice, imho. Thank you very much for taking your time with me. Tom
From: Tom on 12 Mar 2005 12:39 '\\o//'annabee wrote: > Theres a very good explaination of this on Alex's homepage, here : > http://alexfru.narod.ru/os/c16/c16.html#SegOfsPair He calls that a trick. Thank you very much indeed for the link. Tom
From: Betov on 12 Mar 2005 12:50
"Tom" <tmyslovsky(a)yahoo.com> ýcrivait news:1110648637.261048.308660 @l41g2000cwc.googlegroups.com: >> I agree. You'll find some people feel that you shouldn't >> "waste" any time at all learning dos. I'd agree that it >> isn't worthwhile learning "dos", per se, in any depth, but I >> agree with you that it's "simpler" to interface with dos >> than with Windows or Linux. Linux has an "int 80h" interface >> - similar to the dos "int 21h" interface, but Windows >> requires calling a "Dynamically Linked Library". Easy enough >> to *do*, but harder to understand what you're *doing* (IMO). > > Yes, thank you. I know, I agree. .... and this is all of the sad problem: You agrea with something that is utterly wrong and completely absurd. Simply, as DOS is evidently "simpler", at a technology point of view, you fall easily into the illusion that it should be easier to program and easier to understand. This is wrong and the reverse is true. There is absolutely nothing different between calling a DOS interruption, that is a Black Box, doing some operation, and calling a Win32 Api, that is also a Black Box, doing some operation. Whatever you _call_ for, be it an INT or an API, black is black. DOS INTs are, in no way lesser black than black. Plus, the developements under a modern OS are incomparably easier than with DOS. Because of the Tools environements. Because of the various Debug thingies (including the default OS securities). Because of the real power of these various "Black Boxes". And so on. Betov. < http://rosasm.org/ > |