1 /* A sometimes minimal FORTH compiler and tutorial for Linux / i386 systems. -*- asm -*-
2 By Richard W.M. Jones <rich@annexia.org> http://annexia.org/forth
3 This is PUBLIC DOMAIN (see public domain release statement below).
4 $Id: jonesforth.S,v 1.36 2007-09-27 23:09:39 rich Exp $
6 gcc -m32 -nostdlib -static -Wl,-Ttext,0 -o jonesforth jonesforth.S
10 INTRODUCTION ----------------------------------------------------------------------
12 FORTH is one of those alien languages which most working programmers regard in the same
13 way as Haskell, LISP, and so on. Something so strange that they'd rather any thoughts
14 of it just go away so they can get on with writing this paying code. But that's wrong
15 and if you care at all about programming then you should at least understand all these
16 languages, even if you will never use them.
18 LISP is the ultimate high-level language, and features from LISP are being added every
19 decade to the more common languages. But FORTH is in some ways the ultimate in low level
20 programming. Out of the box it lacks features like dynamic memory management and even
21 strings. In fact, at its primitive level it lacks even basic concepts like IF-statements
24 Why then would you want to learn FORTH? There are several very good reasons. First
25 and foremost, FORTH is minimal. You really can write a complete FORTH in, say, 2000
26 lines of code. I don't just mean a FORTH program, I mean a complete FORTH operating
27 system, environment and language. You could boot such a FORTH on a bare PC and it would
28 come up with a prompt where you could start doing useful work. The FORTH you have here
29 isn't minimal and uses a Linux process as its 'base PC' (both for the purposes of making
30 it a good tutorial). It's possible to completely understand the system. Who can say they
31 completely understand how Linux works, or gcc?
33 Secondly FORTH has a peculiar bootstrapping property. By that I mean that after writing
34 a little bit of assembly to talk to the hardware and implement a few primitives, all the
35 rest of the language and compiler is written in FORTH itself. Remember I said before
36 that FORTH lacked IF-statements and loops? Well of course it doesn't really because
37 such a lanuage would be useless, but my point was rather that IF-statements and loops are
38 written in FORTH itself.
40 Now of course this is common in other languages as well, and in those languages we call
41 them 'libraries'. For example in C, 'printf' is a library function written in C. But
42 in FORTH this goes way beyond mere libraries. Can you imagine writing C's 'if' in C?
43 And that brings me to my third reason: If you can write 'if' in FORTH, then why restrict
44 yourself to the usual if/while/for/switch constructs? You want a construct that iterates
45 over every other element in a list of numbers? You can add it to the language. What
46 about an operator which pulls in variables directly from a configuration file and makes
47 them available as FORTH variables? Or how about adding Makefile-like dependencies to
48 the language? No problem in FORTH. How about modifying the FORTH compiler to allow
49 complex inlining strategies -- simple. This concept isn't common in programming languages,
50 but it has a name (in fact two names): "macros" (by which I mean LISP-style macros, not
51 the lame C preprocessor) and "domain specific languages" (DSLs).
53 This tutorial isn't about learning FORTH as the language. I'll point you to some references
54 you should read if you're not familiar with using FORTH. This tutorial is about how to
55 write FORTH. In fact, until you understand how FORTH is written, you'll have only a very
56 superficial understanding of how to use it.
58 So if you're not familiar with FORTH or want to refresh your memory here are some online
61 http://en.wikipedia.org/wiki/Forth_%28programming_language%29
63 http://galileo.phys.virginia.edu/classes/551.jvn.fall01/primer.htm
65 http://wiki.laptop.org/go/Forth_Lessons
67 http://www.albany.net/~hello/simple.htm
69 Here is another "Why FORTH?" essay: http://www.jwdt.com/~paysan/why-forth.html
71 Discussion and criticism of this FORTH here: http://lambda-the-ultimate.org/node/2452
73 ACKNOWLEDGEMENTS ----------------------------------------------------------------------
75 This code draws heavily on the design of LINA FORTH (http://home.hccnet.nl/a.w.m.van.der.horst/lina.html)
76 by Albert van der Horst. Any similarities in the code are probably not accidental.
78 Some parts of this FORTH are also based on this IOCCC entry from 1992:
79 http://ftp.funet.fi/pub/doc/IOCCC/1992/buzzard.2.design.
80 I was very proud when Sean Barrett, the original author of the IOCCC entry, commented in the LtU thread
81 http://lambda-the-ultimate.org/node/2452#comment-36818 about this FORTH.
83 And finally I'd like to acknowledge the (possibly forgotten?) authors of ARTIC FORTH because their
84 original program which I still have on original cassette tape kept nagging away at me all these years.
85 http://en.wikipedia.org/wiki/Artic_Software
87 PUBLIC DOMAIN ----------------------------------------------------------------------
89 I, the copyright holder of this work, hereby release it into the public domain. This applies worldwide.
91 In case this is not legally possible, I grant any entity the right to use this work for any purpose,
92 without any conditions, unless such conditions are required by law.
94 SETTING UP ----------------------------------------------------------------------
96 Let's get a few housekeeping things out of the way. Firstly because I need to draw lots of
97 ASCII-art diagrams to explain concepts, the best way to look at this is using a window which
98 uses a fixed width font and is at least this wide:
100 <------------------------------------------------------------------------------------------------------------------------>
102 Secondly make sure TABS are set to 8 characters. The following should be a vertical
103 line. If not, sort out your tabs.
109 Thirdly I assume that your screen is at least 50 characters high.
111 ASSEMBLING ----------------------------------------------------------------------
113 If you want to actually run this FORTH, rather than just read it, you will need Linux on an
114 i386. Linux because instead of programming directly to the hardware on a bare PC which I
115 could have done, I went for a simpler tutorial by assuming that the 'hardware' is a Linux
116 process with a few basic system calls (read, write and exit and that's about all). i386
117 is needed because I had to write the assembly for a processor, and i386 is by far the most
118 common. (Of course when I say 'i386', any 32- or 64-bit x86 processor will do. I'm compiling
119 this on a 64 bit AMD Opteron).
121 Again, to assemble this you will need gcc and gas (the GNU assembler). The commands to
122 assemble and run the code (save this file as 'jonesforth.S') are:
124 gcc -m32 -nostdlib -static -Wl,-Ttext,0 -o jonesforth jonesforth.S
125 cat jonesforth.f - | ./jonesforth
127 If you want to run your own FORTH programs you can do:
129 cat jonesforth.f myprog.f | ./jonesforth
131 If you want to load your own FORTH code and then continue reading user commands, you can do:
133 cat jonesforth.f myfunctions.f - | ./jonesforth
135 ASSEMBLER ----------------------------------------------------------------------
137 (You can just skip to the next section -- you don't need to be able to read assembler to
138 follow this tutorial).
140 However if you do want to read the assembly code here are a few notes about gas (the GNU assembler):
142 (1) Register names are prefixed with '%', so %eax is the 32 bit i386 accumulator. The registers
143 available on i386 are: %eax, %ebx, %ecx, %edx, %esi, %edi, %ebp and %esp, and most of them
144 have special purposes.
146 (2) Add, mov, etc. take arguments in the form SRC,DEST. So mov %eax,%ecx moves %eax -> %ecx
148 (3) Constants are prefixed with '$', and you mustn't forget it! If you forget it then it
149 causes a read from memory instead, so:
150 mov $2,%eax moves number 2 into %eax
151 mov 2,%eax reads the 32 bit word from address 2 into %eax (ie. most likely a mistake)
153 (4) gas has a funky syntax for local labels, where '1f' (etc.) means label '1:' "forwards"
154 and '1b' (etc.) means label '1:' "backwards".
156 (5) 'ja' is "jump if above", 'jb' for "jump if below", 'je' "jump if equal" etc.
158 (6) gas has a reasonably nice .macro syntax, and I use them a lot to make the code shorter and
161 For more help reading the assembler, do "info gas" at the Linux prompt.
163 Now the tutorial starts in earnest.
165 THE DICTIONARY ----------------------------------------------------------------------
167 In FORTH as you will know, functions are called "words", and just as in other languages they
168 have a name and a definition. Here are two FORTH words:
170 : DOUBLE DUP + ; \ name is "DOUBLE", definition is "DUP +"
171 : QUADRUPLE DOUBLE DOUBLE ; \ name is "QUADRUPLE", definition is "DOUBLE DOUBLE"
173 Words, both built-in ones and ones which the programmer defines later, are stored in a dictionary
174 which is just a linked list of dictionary entries.
176 <--- DICTIONARY ENTRY (HEADER) ----------------------->
177 +------------------------+--------+---------- - - - - +----------- - - - -
178 | LINK POINTER | LENGTH/| NAME | DEFINITION
180 +--- (4 bytes) ----------+- byte -+- n bytes - - - - +----------- - - - -
182 I'll come to the definition of the word later. For now just look at the header. The first
183 4 bytes are the link pointer. This points back to the previous word in the dictionary, or, for
184 the first word in the dictionary it is just a NULL pointer. Then comes a length/flags byte.
185 The length of the word can be up to 31 characters (5 bits used) and the top three bits are used
186 for various flags which I'll come to later. This is followed by the name itself, and in this
187 implementation the name is rounded up to a multiple of 4 bytes by padding it with zero bytes.
188 That's just to ensure that the definition starts on a 32 bit boundary.
190 A FORTH variable called LATEST contains a pointer to the most recently defined word, in
191 other words, the head of this linked list.
193 DOUBLE and QUADRUPLE might look like this:
195 pointer to previous word
198 +--|------+---+---+---+---+---+---+---+---+------------- - - - -
199 | LINK | 6 | D | O | U | B | L | E | 0 | (definition ...)
200 +---------+---+---+---+---+---+---+---+---+------------- - - - -
203 +--|------+---+---+---+---+---+---+---+---+---+---+---+---+------------- - - - -
204 | LINK | 9 | Q | U | A | D | R | U | P | L | E | 0 | 0 | (definition ...)
205 +---------+---+---+---+---+---+---+---+---+---+---+---+---+------------- - - - -
211 You should be able to see from this how you might implement functions to find a word in
212 the dictionary (just walk along the dictionary entries starting at LATEST and matching
213 the names until you either find a match or hit the NULL pointer at the end of the dictionary);
214 and add a word to the dictionary (create a new definition, set its LINK to LATEST, and set
215 LATEST to point to the new word). We'll see precisely these functions implemented in
216 assembly code later on.
218 One interesting consequence of using a linked list is that you can redefine words, and
219 a newer definition of a word overrides an older one. This is an important concept in
220 FORTH because it means that any word (even "built-in" or "standard" words) can be
221 overridden with a new definition, either to enhance it, to make it faster or even to
222 disable it. However because of the way that FORTH words get compiled, which you'll
223 understand below, words defined using the old definition of a word continue to use
224 the old definition. Only words defined after the new definition use the new definition.
226 DIRECT THREADED CODE ----------------------------------------------------------------------
228 Now we'll get to the really crucial bit in understanding FORTH, so go and get a cup of tea
229 or coffee and settle down. It's fair to say that if you don't understand this section, then you
230 won't "get" how FORTH works, and that would be a failure on my part for not explaining it well.
231 So if after reading this section a few times you don't understand it, please email me
234 Let's talk first about what "threaded code" means. Imagine a peculiar version of C where
235 you are only allowed to call functions without arguments. (Don't worry for now that such a
236 language would be completely useless!) So in our peculiar C, code would look like this:
245 and so on. How would a function, say 'f' above, be compiled by a standard C compiler?
246 Probably into assembly code like this. On the right hand side I've written the actual
250 CALL a E8 08 00 00 00
251 CALL b E8 1C 00 00 00
252 CALL c E8 2C 00 00 00
253 ; ignore the return from the function for now
255 "E8" is the x86 machine code to "CALL" a function. In the first 20 years of computing
256 memory was hideously expensive and we might have worried about the wasted space being used
257 by the repeated "E8" bytes. We can save 20% in code size (and therefore, in expensive memory)
258 by compressing this into just:
260 08 00 00 00 Just the function addresses, without
261 1C 00 00 00 the CALL prefix.
264 On a 16-bit machine like the ones which originally ran FORTH the savings are even greater - 33%.
266 [Historical note: If the execution model that FORTH uses looks strange from the following
267 paragraphs, then it was motivated entirely by the need to save memory on early computers.
268 This code compression isn't so important now when our machines have more memory in their L1
269 caches than those early computers had in total, but the execution model still has some
272 Of course this code won't run directly any more. Instead we need to write an interpreter
273 which takes each pair of bytes and calls it.
275 On an i386 machine it turns out that we can write this interpreter rather easily, in just
276 two assembly instructions which turn into just 3 bytes of machine code. Let's store the
277 pointer to the next word to execute in the %esi register:
279 08 00 00 00 <- We're executing this one now. %esi is the _next_ one to execute.
283 The all-important i386 instruction is called LODSL (or in Intel manuals, LODSW). It does
284 two things. Firstly it reads the memory at %esi into the accumulator (%eax). Secondly it
285 increments %esi by 4 bytes. So after LODSL, the situation now looks like this:
287 08 00 00 00 <- We're still executing this one
288 1C 00 00 00 <- %eax now contains this address (0x0000001C)
291 Now we just need to jump to the address in %eax. This is again just a single x86 instruction
292 written JMP *(%eax). And after doing the jump, the situation looks like:
295 1C 00 00 00 <- Now we're executing this subroutine.
298 To make this work, each subroutine is followed by the two instructions 'LODSL; JMP *(%eax)'
299 which literally make the jump to the next subroutine.
301 And that brings us to our first piece of actual code! Well, it's a macro.
310 /* The macro is called NEXT. That's a FORTH-ism. It expands to those two instructions.
312 Every FORTH primitive that we write has to be ended by NEXT. Think of it kind of like
315 The above describes what is known as direct threaded code.
317 To sum up: We compress our function calls down to a list of addresses and use a somewhat
318 magical macro to act as a "jump to next function in the list". We also use one register (%esi)
319 to act as a kind of instruction pointer, pointing to the next function in the list.
321 I'll just give you a hint of what is to come by saying that a FORTH definition such as:
323 : QUADRUPLE DOUBLE DOUBLE ;
325 actually compiles (almost, not precisely but we'll see why in a moment) to a list of
326 function addresses for DOUBLE, DOUBLE and a special function called EXIT to finish off.
328 At this point, REALLY EAGLE-EYED ASSEMBLY EXPERTS are saying "JONES, YOU'VE MADE A MISTAKE!".
330 I lied about JMP *(%eax).
332 INDIRECT THREADED CODE ----------------------------------------------------------------------
334 It turns out that direct threaded code is interesting but only if you want to just execute
335 a list of functions written in assembly language. So QUADRUPLE would work only if DOUBLE
336 was an assembly language function. In the direct threaded code, QUADRUPLE would look like:
339 | addr of DOUBLE --------------------> (assembly code to do the double)
340 +------------------+ NEXT
341 %esi -> | addr of DOUBLE |
344 We can add an extra indirection to allow us to run both words written in assembly language
345 (primitives written for speed) and words written in FORTH themselves as lists of addresses.
347 The extra indirection is the reason for the brackets in JMP *(%eax).
349 Let's have a look at how QUADRUPLE and DOUBLE really look in FORTH:
351 : QUADRUPLE DOUBLE DOUBLE ;
354 | codeword | : DOUBLE DUP + ;
356 | addr of DOUBLE ---------------> +------------------+
357 +------------------+ | codeword |
358 | addr of DOUBLE | +------------------+
359 +------------------+ | addr of DUP --------------> +------------------+
360 | addr of EXIT | +------------------+ | codeword -------+
361 +------------------+ %esi -> | addr of + --------+ +------------------+ |
362 +------------------+ | | assembly to <-----+
363 | addr of EXIT | | | implement DUP |
364 +------------------+ | | .. |
367 | +------------------+
369 +-----> +------------------+
371 +------------------+ |
372 | assembly to <------+
379 This is the part where you may need an extra cup of tea/coffee/favourite caffeinated
380 beverage. What has changed is that I've added an extra pointer to the beginning of
381 the definitions. In FORTH this is sometimes called the "codeword". The codeword is
382 a pointer to the interpreter to run the function. For primitives written in
383 assembly language, the "interpreter" just points to the actual assembly code itself.
384 They don't need interpreting, they just run.
386 In words written in FORTH (like QUADRUPLE and DOUBLE), the codeword points to an interpreter
389 I'll show you the interpreter function shortly, but let's recall our indirect
390 JMP *(%eax) with the "extra" brackets. Take the case where we're executing DOUBLE
391 as shown, and DUP has been called. Note that %esi is pointing to the address of +
393 The assembly code for DUP eventually does a NEXT. That:
395 (1) reads the address of + into %eax %eax points to the codeword of +
396 (2) increments %esi by 4
397 (3) jumps to the indirect %eax jumps to the address in the codeword of +,
398 ie. the assembly code to implement +
403 | addr of DOUBLE ---------------> +------------------+
404 +------------------+ | codeword |
405 | addr of DOUBLE | +------------------+
406 +------------------+ | addr of DUP --------------> +------------------+
407 | addr of EXIT | +------------------+ | codeword -------+
408 +------------------+ | addr of + --------+ +------------------+ |
409 +------------------+ | | assembly to <-----+
410 %esi -> | addr of EXIT | | | implement DUP |
411 +------------------+ | | .. |
414 | +------------------+
416 +-----> +------------------+
418 +------------------+ |
419 now we're | assembly to <-----+
420 executing | implement + |
426 So I hope that I've convinced you that NEXT does roughly what you'd expect. This is
427 indirect threaded code.
429 I've glossed over four things. I wonder if you can guess without reading on what they are?
435 My list of four things are: (1) What does "EXIT" do? (2) which is related to (1) is how do
436 you call into a function, ie. how does %esi start off pointing at part of QUADRUPLE, but
437 then point at part of DOUBLE. (3) What goes in the codeword for the words which are written
438 in FORTH? (4) How do you compile a function which does anything except call other functions
439 ie. a function which contains a number like : DOUBLE 2 * ; ?
441 THE INTERPRETER AND RETURN STACK ------------------------------------------------------------
443 Going at these in no particular order, let's talk about issues (3) and (2), the interpreter
444 and the return stack.
446 Words which are defined in FORTH need a codeword which points to a little bit of code to
447 give them a "helping hand" in life. They don't need much, but they do need what is known
448 as an "interpreter", although it doesn't really "interpret" in the same way that, say,
449 Java bytecode used to be interpreted (ie. slowly). This interpreter just sets up a few
450 machine registers so that the word can then execute at full speed using the indirect
451 threaded model above.
453 One of the things that needs to happen when QUADRUPLE calls DOUBLE is that we save the old
454 %esi ("instruction pointer") and create a new one pointing to the first word in DOUBLE.
455 Because we will need to restore the old %esi at the end of DOUBLE (this is, after all, like
456 a function call), we will need a stack to store these "return addresses" (old values of %esi).
458 As you will have read, when reading the background documentation, FORTH has two stacks,
459 an ordinary stack for parameters, and a return stack which is a bit more mysterious. But
460 our return stack is just the stack I talked about in the previous paragraph, used to save
461 %esi when calling from a FORTH word into another FORTH word.
463 In this FORTH, we are using the normal stack pointer (%esp) for the parameter stack.
464 We will use the i386's "other" stack pointer (%ebp, usually called the "frame pointer")
465 for our return stack.
467 I've got two macros which just wrap up the details of using %ebp for the return stack.
468 You use them as for example "PUSHRSP %eax" (push %eax on the return stack) or "POPRSP %ebx"
469 (pop top of return stack into %ebx).
472 /* Macros to deal with the return stack. */
474 lea -4(%ebp),%ebp // push reg on to return stack
479 mov (%ebp),\reg // pop top of return stack to reg
484 And with that we can now talk about the interpreter.
486 In FORTH the interpreter function is often called DOCOL (I think it means "DO COLON" because
487 all FORTH definitions start with a colon, as in : DOUBLE DUP + ;
489 The "interpreter" (it's not really "interpreting") just needs to push the old %esi on the
490 stack and set %esi to the first word in the definition. Remember that we jumped to the
491 function using JMP *(%eax)? Well a consequence of that is that conveniently %eax contains
492 the address of this codeword, so just by adding 4 to it we get the address of the first
493 data word. Finally after setting up %esi, it just does NEXT which causes that first word
497 /* DOCOL - the interpreter! */
501 PUSHRSP %esi // push %esi on to the return stack
502 addl $4,%eax // %eax points to codeword, so make
503 movl %eax,%esi // %esi point to first data word
507 Just to make this absolutely clear, let's see how DOCOL works when jumping from QUADRUPLE
513 +------------------+ DOUBLE:
514 | addr of DOUBLE ---------------> +------------------+
515 +------------------+ %eax -> | addr of DOCOL |
516 %esi -> | addr of DOUBLE | +------------------+
517 +------------------+ | addr of DUP |
518 | addr of EXIT | +------------------+
519 +------------------+ | etc. |
521 First, the call to DOUBLE calls DOCOL (the codeword of DOUBLE). DOCOL does this: It
522 pushes the old %esi on the return stack. %eax points to the codeword of DOUBLE, so we
523 just add 4 on to it to get our new %esi:
528 +------------------+ DOUBLE:
529 | addr of DOUBLE ---------------> +------------------+
530 top of return +------------------+ %eax -> | addr of DOCOL |
531 stack points -> | addr of DOUBLE | + 4 = +------------------+
532 +------------------+ %esi -> | addr of DUP |
533 | addr of EXIT | +------------------+
534 +------------------+ | etc. |
536 Then we do NEXT, and because of the magic of threaded code that increments %esi again
539 Well, it seems to work.
541 One minor point here. Because DOCOL is the first bit of assembly actually to be defined
542 in this file (the others were just macros), and because I usually compile this code with the
543 text segment starting at address 0, DOCOL has address 0. So if you are disassembling the
544 code and see a word with a codeword of 0, you will immediately know that the word is
545 written in FORTH (it's not an assembler primitive) and so uses DOCOL as the interpreter.
547 STARTING UP ----------------------------------------------------------------------
549 Now let's get down to nuts and bolts. When we start the program we need to set up
550 a few things like the return stack. But as soon as we can, we want to jump into FORTH
551 code (albeit much of the "early" FORTH code will still need to be written as
552 assembly language primitives).
554 This is what the set up code does. Does a tiny bit of house-keeping, sets up the
555 separate return stack (NB: Linux gives us the ordinary parameter stack already), then
556 immediately jumps to a FORTH word called COLD. COLD stands for cold-start. In ISO
557 FORTH (but not in this FORTH), COLD can be called at any time to completely reset
558 the state of FORTH, and there is another word called WARM which does a partial reset.
561 /* ELF entry point. */
566 mov %esp,var_S0 // Store the initial data stack pointer.
567 mov $return_stack,%ebp // Initialise the return stack.
569 mov $cold_start,%esi // Initialise interpreter.
570 NEXT // Run interpreter!
573 cold_start: // High-level code without a codeword.
577 We also allocate some space for the return stack and some space to store user
578 definitions. These are static memory allocations using fixed-size buffers, but it
579 wouldn't be a great deal of work to make them dynamic.
583 /* FORTH return stack. */
584 .set RETURN_STACK_SIZE,8192
586 .space RETURN_STACK_SIZE
587 return_stack: // Initial top of return stack.
589 /* The user definitions area: space for user-defined words and general memory allocations. */
590 .set USER_DEFS_SIZE,65536
593 .space USER_DEFS_SIZE
595 /* This is used as a temporary input buffer when reading from files or the terminal. */
596 .set BUFFER_SIZE,4096
608 BUILT-IN WORDS ----------------------------------------------------------------------
610 Remember our dictionary entries (headers)? Let's bring those together with the codeword
611 and data words to see how : DOUBLE DUP + ; really looks in memory.
613 pointer to previous word
616 +--|------+---+---+---+---+---+---+---+---+------------+------------+------------+------------+
617 | LINK | 6 | D | O | U | B | L | E | 0 | DOCOL | DUP | + | EXIT |
618 +---------+---+---+---+---+---+---+---+---+------------+--|---------+------------+------------+
621 LINK in next word points to codeword of DUP
623 Initially we can't just write ": DOUBLE DUP + ;" (ie. that literal string) here because we
624 don't yet have anything to read the string, break it up at spaces, parse each word, etc. etc.
625 So instead we will have to define built-in words using the GNU assembler data constructors
626 (like .int, .byte, .string, .ascii and so on -- look them up in the gas info page if you are
629 The long way would be:
630 .int <link to previous word>
632 .ascii "DOUBLE" // string
634 DOUBLE: .int DOCOL // codeword
635 .int DUP // pointer to codeword of DUP
636 .int PLUS // pointer to codeword of +
637 .int EXIT // pointer to codeword of EXIT
639 That's going to get quite tedious rather quickly, so here I define an assembler macro
640 so that I can just write:
642 defword "DOUBLE",6,,DOUBLE
645 and I'll get exactly the same effect.
647 Don't worry too much about the exact implementation details of this macro - it's complicated!
650 /* Flags - these are discussed later. */
653 .set F_LENMASK,0x1f // length mask
655 // Store the chain of links.
658 .macro defword name, namelen, flags=0, label
664 .set link,name_\label
665 .byte \flags+\namelen // flags + length byte
666 .ascii "\name" // the name
670 .int DOCOL // codeword - the interpreter
671 // list of word pointers follow
675 Similarly I want a way to write words written in assembly language. There will quite a few
676 of these to start with because, well, everything has to start in assembly before there's
677 enough "infrastructure" to be able to start writing FORTH words, but also I want to define
678 some common FORTH words in assembly language for speed, even though I could write them in FORTH.
680 This is what DUP looks like in memory:
682 pointer to previous word
685 +--|------+---+---+---+---+------------+
686 | LINK | 3 | D | U | P | code_DUP ---------------------> points to the assembly
687 +---------+---+---+---+---+------------+ code used to write DUP,
688 ^ len codeword which ends with NEXT.
692 Again, for brevity in writing the header I'm going to write an assembler macro called defcode.
695 .macro defcode name, namelen, flags=0, label
701 .set link,name_\label
702 .byte \flags+\namelen // flags + length byte
703 .ascii "\name" // the name
707 .int code_\label // codeword
711 code_\label : // assembler code follows
715 Now some easy FORTH primitives. These are written in assembly for speed. If you understand
716 i386 assembly language then it is worth reading these. However if you don't understand assembly
717 you can skip the details.
721 pop %eax // duplicate top of stack
726 defcode "DROP",4,,DROP
727 pop %eax // drop top of stack
730 defcode "SWAP",4,,SWAP
731 pop %eax // swap top of stack
737 defcode "OVER",4,,OVER
738 mov 4(%esp),%eax // get the second element of stack
739 push %eax // and push it on top
751 defcode "-ROT",4,,NROT
761 incl (%esp) // increment top of stack
765 decl (%esp) // decrement top of stack
768 defcode "4+",2,,INCR4
769 addl $4,(%esp) // add 4 to top of stack
772 defcode "4-",2,,DECR4
773 subl $4,(%esp) // subtract 4 from top of stack
777 pop %eax // get top of stack
778 addl %eax,(%esp) // and add it to next word on stack
782 pop %eax // get top of stack
783 subl %eax,(%esp) // and subtract it from next word on stack
790 push %eax // ignore overflow
794 In this FORTH, only /MOD is primitive. Later we will define the / and MOD words in
795 terms of the primitive /MOD. The design of the i386 assembly instruction idiv which
796 leaves both quotient and remainder makes this obvious choice.
799 defcode "/MOD",4,,DIVMOD
804 push %edx // push remainder
805 push %eax // push quotient
808 defcode "=",1,,EQU // top two words are equal?
818 defcode "<>",2,,NEQU // top two words are not equal?
868 defcode "0=",2,,ZEQU // top of stack equals 0?
877 defcode "0<>",3,,ZNEQU // top of stack not 0?
886 defcode "0<",2,,ZLT // comparisons with 0
922 defcode "AND",3,,AND // bitwise AND
927 defcode "OR",2,,OR // bitwise OR
932 defcode "XOR",3,,XOR // bitwise XOR
937 defcode "INVERT",6,,INVERT // this is the FORTH bitwise "NOT" function (cf. NEGATE)
942 RETURNING FROM FORTH WORDS ----------------------------------------------------------------------
944 Time to talk about what happens when we EXIT a function. In this diagram QUADRUPLE has called
945 DOUBLE, and DOUBLE is about to exit (look at where %esi is pointing):
950 +------------------+ DOUBLE
951 | addr of DOUBLE ---------------> +------------------+
952 +------------------+ | codeword |
953 | addr of DOUBLE | +------------------+
954 +------------------+ | addr of DUP |
955 | addr of EXIT | +------------------+
956 +------------------+ | addr of + |
958 %esi -> | addr of EXIT |
961 What happens when the + function does NEXT? Well, the following code is executed.
964 defcode "EXIT",4,,EXIT
965 POPRSP %esi // pop return stack into %esi
969 EXIT gets the old %esi which we saved from before on the return stack, and puts it in %esi.
970 So after this (but just before NEXT) we get:
975 +------------------+ DOUBLE
976 | addr of DOUBLE ---------------> +------------------+
977 +------------------+ | codeword |
978 %esi -> | addr of DOUBLE | +------------------+
979 +------------------+ | addr of DUP |
980 | addr of EXIT | +------------------+
981 +------------------+ | addr of + |
986 And NEXT just completes the job by, well, in this case just by calling DOUBLE again :-)
988 LITERALS ----------------------------------------------------------------------
990 The final point I "glossed over" before was how to deal with functions that do anything
991 apart from calling other functions. For example, suppose that DOUBLE was defined like this:
995 It does the same thing, but how do we compile it since it contains the literal 2? One way
996 would be to have a function called "2" (which you'd have to write in assembler), but you'd need
997 a function for every single literal that you wanted to use.
999 FORTH solves this by compiling the function using a special word called LIT:
1001 +---------------------------+-------+-------+-------+-------+-------+
1002 | (usual header of DOUBLE) | DOCOL | LIT | 2 | * | EXIT |
1003 +---------------------------+-------+-------+-------+-------+-------+
1005 LIT is executed in the normal way, but what it does next is definitely not normal. It
1006 looks at %esi (which now points to the literal 2), grabs it, pushes it on the stack, then
1007 manipulates %esi in order to skip the literal as if it had never been there.
1009 What's neat is that the whole grab/manipulate can be done using a single byte single
1010 i386 instruction, our old friend LODSL. Rather than me drawing more ASCII-art diagrams,
1011 see if you can find out how LIT works:
1014 defcode "LIT",3,,LIT
1015 // %esi points to the next command, but in this case it points to the next
1016 // literal 32 bit integer. Get that literal into %eax and increment %esi.
1017 // On x86, it's a convenient single byte instruction! (cf. NEXT macro)
1019 push %eax // push the literal number on to stack
1023 MEMORY ----------------------------------------------------------------------
1025 As important point about FORTH is that it gives you direct access to the lowest levels
1026 of the machine. Manipulating memory directly is done frequently in FORTH, and these are
1027 the primitive words for doing it.
1030 defcode "!",1,,STORE
1031 pop %ebx // address to store at
1032 pop %eax // data to store there
1033 mov %eax,(%ebx) // store it
1036 defcode "@",1,,FETCH
1037 pop %ebx // address to fetch
1038 mov (%ebx),%eax // fetch it
1039 push %eax // push value onto stack
1042 defcode "+!",2,,ADDSTORE
1044 pop %eax // the amount to add
1045 addl %eax,(%ebx) // add it
1048 defcode "-!",2,,SUBSTORE
1050 pop %eax // the amount to subtract
1051 subl %eax,(%ebx) // add it
1055 ! and @ (STORE and FETCH) store 32-bit words. It's also useful to be able to read and write bytes
1056 so we also define standard words C@ and C!.
1058 Byte-oriented operations only work on architectures which permit them (i386 is one of those).
1061 defcode "C!",2,,STOREBYTE
1062 pop %ebx // address to store at
1063 pop %eax // data to store there
1064 movb %al,(%ebx) // store it
1067 defcode "C@",2,,FETCHBYTE
1068 pop %ebx // address to fetch
1070 movb (%ebx),%al // fetch it
1071 push %eax // push value onto stack
1075 BUILT-IN VARIABLES ----------------------------------------------------------------------
1077 These are some built-in variables and related standard FORTH words. Of these, the only one that we
1078 have discussed so far was LATEST, which points to the last (most recently defined) word in the
1079 FORTH dictionary. LATEST is also a FORTH word which pushes the address of LATEST (the variable)
1080 on to the stack, so you can read or write it using @ and ! operators. For example, to print
1081 the current value of LATEST (and this can apply to any FORTH variable) you would do:
1085 To make defining variables shorter, I'm using a macro called defvar, similar to defword and
1086 defcode above. (In fact the defvar macro uses defcode to do the dictionary header).
1089 .macro defvar name, namelen, flags=0, label, initial=0
1090 defcode \name,\namelen,\flags,\label
1100 The built-in variables are:
1102 STATE Is the interpreter executing code (0) or compiling a word (non-zero)?
1103 LATEST Points to the latest (most recently defined) word in the dictionary.
1104 HERE Points to the next free byte of memory. When compiling, compiled words go here.
1105 _X These are three scratch variables, used by some standard dictionary words.
1108 S0 Stores the address of the top of the parameter stack.
1109 BASE The current base for printing and reading numbers.
1112 defvar "STATE",5,,STATE
1113 defvar "HERE",4,,HERE,user_defs_start
1114 defvar "LATEST",6,,LATEST,name_SYSEXIT // SYSEXIT must be last in built-in dictionary
1119 defvar "BASE",4,,BASE,10
1122 BUILT-IN CONSTANTS ----------------------------------------------------------------------
1124 It's also useful to expose a few constants to FORTH. When the word is executed it pushes a
1125 constant value on the stack.
1127 The built-in constants are:
1129 VERSION Is the current version of this FORTH.
1130 R0 The address of the top of the return stack.
1131 DOCOL Pointer to DOCOL.
1132 F_IMMED The IMMEDIATE flag's actual value.
1133 F_HIDDEN The HIDDEN flag's actual value.
1134 F_LENMASK The length mask in the flags/len byte.
1137 .macro defconst name, namelen, flags=0, label, value
1138 defcode \name,\namelen,\flags,\label
1143 defconst "VERSION",7,,VERSION,JONES_VERSION
1144 defconst "R0",2,,RZ,return_stack
1145 defconst "DOCOL",5,,__DOCOL,DOCOL
1146 defconst "F_IMMED",7,,__F_IMMED,F_IMMED
1147 defconst "F_HIDDEN",8,,__F_HIDDEN,F_HIDDEN
1148 defconst "F_LENMASK",9,,__F_LENMASK,F_LENMASK
1151 RETURN STACK ----------------------------------------------------------------------
1153 These words allow you to access the return stack. Recall that the register %ebp always points to
1154 the top of the return stack.
1158 pop %eax // pop parameter stack into %eax
1159 PUSHRSP %eax // push it on to the return stack
1162 defcode "R>",2,,FROMR
1163 POPRSP %eax // pop return stack on to %eax
1164 push %eax // and push on to parameter stack
1167 defcode "RSP@",4,,RSPFETCH
1171 defcode "RSP!",4,,RSPSTORE
1175 defcode "RDROP",5,,RDROP
1176 lea 4(%ebp),%ebp // pop return stack and throw away
1180 PARAMETER (DATA) STACK ----------------------------------------------------------------------
1182 These functions allow you to manipulate the parameter stack. Recall that Linux sets up the parameter
1183 stack for us, and it is accessed through %esp.
1186 defcode "DSP@",4,,DSPFETCH
1191 defcode "DSP!",4,,DSPSTORE
1196 INPUT AND OUTPUT ----------------------------------------------------------------------
1198 These are our first really meaty/complicated FORTH primitives. I have chosen to write them in
1199 assembler, but surprisingly in "real" FORTH implementations these are often written in terms
1200 of more fundamental FORTH primitives. I chose to avoid that because I think that just obscures
1201 the implementation. After all, you may not understand assembler but you can just think of it
1202 as an opaque block of code that does what it says.
1204 Let's discuss input first.
1206 The FORTH word KEY reads the next byte from stdin (and pushes it on the parameter stack).
1207 So if KEY is called and someone hits the space key, then the number 32 (ASCII code of space)
1208 is pushed on the stack.
1210 In FORTH there is no distinction between reading code and reading input. We might be reading
1211 and compiling code, we might be reading words to execute, we might be asking for the user
1212 to type their name -- ultimately it all comes in through KEY.
1214 The implementation of KEY uses an input buffer of a certain size (defined at the end of the
1215 program). It calls the Linux read(2) system call to fill this buffer and tracks its position
1216 in the buffer using a couple of variables, and if it runs out of input buffer then it refills
1217 it automatically. The other thing that KEY does is if it detects that stdin has closed, it
1218 exits the program, which is why when you hit ^D the FORTH system cleanly exits.
1221 #include <asm-i386/unistd.h>
1223 defcode "KEY",3,,KEY
1225 push %eax // push return value on stack
1237 1: // out of input; use read(2) to fetch more input from stdin
1238 xor %ebx,%ebx // 1st param: stdin
1239 mov $buffer,%ecx // 2nd param: buffer
1241 mov $buffend-buffer,%edx // 3rd param: max length
1242 mov $__NR_read,%eax // syscall: read
1244 test %eax,%eax // If %eax <= 0, then exit.
1246 addl %eax,%ecx // buffer+%eax = bufftop
1250 2: // error or out of input: exit
1252 mov $__NR_exit,%eax // syscall: exit
1256 By contrast, output is much simpler. The FORTH word EMIT writes out a single byte to stdout.
1257 This implementation just uses the write system call. No attempt is made to buffer output, but
1258 it would be a good exercise to add it.
1261 defcode "EMIT",4,,EMIT
1266 mov $1,%ebx // 1st param: stdout
1268 // write needs the address of the byte to write
1270 mov $2f,%ecx // 2nd param: address
1272 mov $1,%edx // 3rd param: nbytes = 1
1274 mov $__NR_write,%eax // write syscall
1279 2: .space 1 // scratch used by EMIT
1282 Back to input, WORD is a FORTH word which reads the next full word of input.
1284 What it does in detail is that it first skips any blanks (spaces, tabs, newlines and so on).
1285 Then it calls KEY to read characters into an internal buffer until it hits a blank. Then it
1286 calculates the length of the word it read and returns the address and the length as
1287 two words on the stack (with the length at the top of stack).
1289 Notice that WORD has a single internal buffer which it overwrites each time (rather like
1290 a static C string). Also notice that WORD's internal buffer is just 32 bytes long and
1291 there is NO checking for overflow. 31 bytes happens to be the maximum length of a
1292 FORTH word that we support, and that is what WORD is used for: to read FORTH words when
1293 we are compiling and executing code. The returned strings are not NUL-terminated, so
1294 in some crazy-world you could define FORTH words containing ASCII NULs, although why
1295 you'd want to is a bit beyond me.
1297 WORD is not suitable for just reading strings (eg. user input) because of all the above
1298 peculiarities and limitations.
1300 Note that when executing, you'll see:
1302 which puts "FOO" and length 3 on the stack, but when compiling:
1304 is an error (or at least it doesn't do what you might expect). Later we'll talk about compiling
1305 and immediate mode, and you'll understand why.
1308 defcode "WORD",4,,WORD
1310 push %edi // push base address
1311 push %ecx // push length
1315 /* Search for first non-blank character. Also skip \ comments. */
1317 call _KEY // get next key, returned in %eax
1318 cmpb $'\\',%al // start of a comment?
1319 je 3f // if so, skip the comment
1321 jbe 1b // if so, keep looking
1323 /* Search for the end of the word, storing chars as we go. */
1324 mov $5f,%edi // pointer to return buffer
1326 stosb // add character to return buffer
1327 call _KEY // get next key, returned in %al
1328 cmpb $' ',%al // is blank?
1329 ja 2b // if not, keep looping
1331 /* Return the word (well, the static buffer) and length. */
1333 mov %edi,%ecx // return length of the word
1334 mov $5f,%edi // return address of the word
1337 /* Code to skip \ comments to end of the current line. */
1340 cmpb $'\n',%al // end of line yet?
1345 // A static buffer where WORD returns. Subsequent calls
1346 // overwrite this buffer. Maximum word length is 32 chars.
1350 As well as reading in words we'll need to read in numbers and for that we are using a function
1351 called SNUMBER. This parses a numeric string such as one returned by WORD and pushes the
1352 number on the parameter stack.
1354 This function does absolutely no error checking, and in particular the length of the string
1355 must be >= 1 bytes, and should contain only digits 0-9. If it doesn't you'll get random results.
1357 This function is only used when reading literal numbers in code, and shouldn't really be used
1358 in user code at all.
1360 defcode "SNUMBER",7,,SNUMBER
1370 imull $10,%eax // %eax *= 10
1373 subb $'0',%bl // ASCII -> digit
1380 DICTIONARY LOOK UPS ----------------------------------------------------------------------
1382 We're building up to our prelude on how FORTH code is compiled, but first we need yet more infrastructure.
1384 The FORTH word FIND takes a string (a word as parsed by WORD -- see above) and looks it up in the
1385 dictionary. What it actually returns is the address of the dictionary header, if it finds it,
1388 So if DOUBLE is defined in the dictionary, then WORD DOUBLE FIND returns the following pointer:
1394 +---------+---+---+---+---+---+---+---+---+------------+------------+------------+------------+
1395 | LINK | 6 | D | O | U | B | L | E | 0 | DOCOL | DUP | + | EXIT |
1396 +---------+---+---+---+---+---+---+---+---+------------+------------+------------+------------+
1398 See also >CFA and >DFA.
1400 FIND doesn't find dictionary entries which are flagged as HIDDEN. See below for why.
1403 defcode "FIND",4,,FIND
1404 pop %ecx // %ecx = length
1405 pop %edi // %edi = address
1407 push %eax // %eax = address of dictionary entry (or NULL)
1411 push %esi // Save %esi so we can use it in string comparison.
1413 // Now we start searching backwards through the dictionary for this word.
1414 mov var_LATEST,%edx // LATEST points to name header of the latest word in the dictionary
1416 test %edx,%edx // NULL pointer? (end of the linked list)
1419 // Compare the length expected and the length of the word.
1420 // Note that if the F_HIDDEN flag is set on the word, then by a bit of trickery
1421 // this won't pick the word (the length will appear to be wrong).
1423 movb 4(%edx),%al // %al = flags+length field
1424 andb $(F_HIDDEN|F_LENMASK),%al // %al = name length
1425 cmpb %cl,%al // Length is the same?
1428 // Compare the strings in detail.
1429 push %ecx // Save the length
1430 push %edi // Save the address (repe cmpsb will move this pointer)
1431 lea 5(%edx),%esi // Dictionary string we are checking against.
1432 repe cmpsb // Compare the strings.
1435 jne 2f // Not the same.
1437 // The strings are the same - return the header pointer in %eax
1443 mov (%edx),%edx // Move back through the link field to the previous word
1444 jmp 1b // .. and loop.
1448 xor %eax,%eax // Return zero to indicate not found.
1452 FIND returns the dictionary pointer, but when compiling we need the codeword pointer (recall
1453 that FORTH definitions are compiled into lists of codeword pointers). The standard FORTH
1454 word >CFA turns a dictionary pointer into a codeword pointer.
1456 The example below shows the result of:
1458 WORD DOUBLE FIND >CFA
1460 FIND returns a pointer to this
1461 | >CFA converts it to a pointer to this
1464 +---------+---+---+---+---+---+---+---+---+------------+------------+------------+------------+
1465 | LINK | 6 | D | O | U | B | L | E | 0 | DOCOL | DUP | + | EXIT |
1466 +---------+---+---+---+---+---+---+---+---+------------+------------+------------+------------+
1470 Because names vary in length, this isn't just a simple increment.
1472 In this FORTH you cannot easily turn a codeword pointer back into a dictionary entry pointer, but
1473 that is not true in most FORTH implementations where they store a back pointer in the definition
1474 (with an obvious memory/complexity cost). The reason they do this is that it is useful to be
1475 able to go backwards (codeword -> dictionary entry) in order to decompile FORTH definitions.
1477 What does CFA stand for? My best guess is "Code Field Address".
1480 defcode ">CFA",4,,TCFA
1487 add $4,%edi // Skip link pointer.
1488 movb (%edi),%al // Load flags+len into %al.
1489 inc %edi // Skip flags+len byte.
1490 andb $F_LENMASK,%al // Just the length, not the flags.
1491 add %eax,%edi // Skip the name.
1492 addl $3,%edi // The codeword is 4-byte aligned.
1497 Related to >CFA is >DFA which takes a dictionary entry address as returned by FIND and
1498 returns a pointer to the first data field.
1500 FIND returns a pointer to this
1501 | >CFA converts it to a pointer to this
1503 | | >DFA converts it to a pointer to this
1506 +---------+---+---+---+---+---+---+---+---+------------+------------+------------+------------+
1507 | LINK | 6 | D | O | U | B | L | E | 0 | DOCOL | DUP | + | EXIT |
1508 +---------+---+---+---+---+---+---+---+---+------------+------------+------------+------------+
1510 (Note to those following the source of FIG-FORTH / ciforth: My >DFA definition is
1511 different from theirs, because they have an extra indirection).
1513 You can see that >DFA is easily defined in FORTH just by adding 4 to the result of >CFA.
1516 defword ">DFA",4,,TDFA
1517 .int TCFA // >CFA (get code field address)
1518 .int INCR4 // 4+ (add 4 to it to get to next word)
1519 .int EXIT // EXIT (return from FORTH word)
1522 COMPILING ----------------------------------------------------------------------
1524 Now we'll talk about how FORTH compiles words. Recall that a word definition looks like this:
1528 and we have to turn this into:
1530 pointer to previous word
1533 +--|------+---+---+---+---+---+---+---+---+------------+------------+------------+------------+
1534 | LINK | 6 | D | O | U | B | L | E | 0 | DOCOL | DUP | + | EXIT |
1535 +---------+---+---+---+---+---+---+---+---+------------+--|---------+------------+------------+
1536 ^ len pad codeword |
1538 LATEST points here points to codeword of DUP
1540 There are several problems to solve. Where to put the new word? How do we read words? How
1541 do we define the words : (COLON) and ; (SEMICOLON)?
1543 FORTH solves this rather elegantly and as you might expect in a very low-level way which
1544 allows you to change how the compiler works on your own code.
1546 FORTH has an INTERPRETER function (a true interpreter this time, not DOCOL) which runs in a
1547 loop, reading words (using WORD), looking them up (using FIND), turning them into codeword
1548 pointers (using >CFA) and deciding what to do with them.
1550 What it does depends on the mode of the interpreter (in variable STATE).
1552 When STATE is zero, the interpreter just runs each word as it looks them up. This is known as
1555 The interesting stuff happens when STATE is non-zero -- compiling mode. In this mode the
1556 interpreter appends the codeword pointer to user memory (the HERE variable points to the next
1557 free byte of user memory).
1559 So you may be able to see how we could define : (COLON). The general plan is:
1561 (1) Use WORD to read the name of the function being defined.
1563 (2) Construct the dictionary entry -- just the header part -- in user memory:
1565 pointer to previous word (from LATEST) +-- Afterwards, HERE points here, where
1566 ^ | the interpreter will start appending
1568 +--|------+---+---+---+---+---+---+---+---+------------+
1569 | LINK | 6 | D | O | U | B | L | E | 0 | DOCOL |
1570 +---------+---+---+---+---+---+---+---+---+------------+
1573 (3) Set LATEST to point to the newly defined word, ...
1575 (4) .. and most importantly leave HERE pointing just after the new codeword. This is where
1576 the interpreter will append codewords.
1578 (5) Set STATE to 1. This goes into compile mode so the interpreter starts appending codewords to
1579 our partially-formed header.
1581 After : has run, our input is here:
1586 Next byte returned by KEY will be the 'D' character of DUP
1588 so the interpreter (now it's in compile mode, so I guess it's really the compiler) reads "DUP",
1589 looks it up in the dictionary, gets its codeword pointer, and appends it:
1591 +-- HERE updated to point here.
1594 +---------+---+---+---+---+---+---+---+---+------------+------------+
1595 | LINK | 6 | D | O | U | B | L | E | 0 | DOCOL | DUP |
1596 +---------+---+---+---+---+---+---+---+---+------------+------------+
1599 Next we read +, get the codeword pointer, and append it:
1601 +-- HERE updated to point here.
1604 +---------+---+---+---+---+---+---+---+---+------------+------------+------------+
1605 | LINK | 6 | D | O | U | B | L | E | 0 | DOCOL | DUP | + |
1606 +---------+---+---+---+---+---+---+---+---+------------+------------+------------+
1609 The issue is what happens next. Obviously what we _don't_ want to happen is that we
1610 read ";" and compile it and go on compiling everything afterwards.
1612 At this point, FORTH uses a trick. Remember the length byte in the dictionary definition
1613 isn't just a plain length byte, but can also contain flags. One flag is called the
1614 IMMEDIATE flag (F_IMMED in this code). If a word in the dictionary is flagged as
1615 IMMEDIATE then the interpreter runs it immediately _even if it's in compile mode_.
1617 This is how the word ; (SEMICOLON) works -- as a word flagged in the dictionary as IMMEDIATE.
1618 And all it does is append the codeword for EXIT on to the current definition and switch
1619 back to immediate mode (set STATE back to 0). Shortly we'll see the actual definition
1620 of ; and we'll see that it's really a very simple definition, declared IMMEDIATE.
1622 After the interpreter reads ; and executes it 'immediately', we get this:
1624 +---------+---+---+---+---+---+---+---+---+------------+------------+------------+------------+
1625 | LINK | 6 | D | O | U | B | L | E | 0 | DOCOL | DUP | + | EXIT |
1626 +---------+---+---+---+---+---+---+---+---+------------+------------+------------+------------+
1633 And that's it, job done, our new definition is compiled, and we're back in immediate mode
1634 just reading and executing words, perhaps including a call to test our new word DOUBLE.
1636 The only last wrinkle in this is that while our word was being compiled, it was in a
1637 half-finished state. We certainly wouldn't want DOUBLE to be called somehow during
1638 this time. There are several ways to stop this from happening, but in FORTH what we
1639 do is flag the word with the HIDDEN flag (F_HIDDEN in this code) just while it is
1640 being compiled. This prevents FIND from finding it, and thus in theory stops any
1641 chance of it being called.
1643 The above explains how compiling, : (COLON) and ; (SEMICOLON) works and in a moment I'm
1644 going to define them. The : (COLON) function can be made a little bit more general by writing
1645 it in two parts. The first part, called CREATE, makes just the header:
1647 +-- Afterwards, HERE points here.
1650 +---------+---+---+---+---+---+---+---+---+
1651 | LINK | 6 | D | O | U | B | L | E | 0 |
1652 +---------+---+---+---+---+---+---+---+---+
1655 and the second part, the actual definition of : (COLON), calls CREATE and appends the
1656 DOCOL codeword, so leaving:
1658 +-- Afterwards, HERE points here.
1661 +---------+---+---+---+---+---+---+---+---+------------+
1662 | LINK | 6 | D | O | U | B | L | E | 0 | DOCOL |
1663 +---------+---+---+---+---+---+---+---+---+------------+
1666 CREATE is a standard FORTH word and the advantage of this split is that we can reuse it to
1667 create other types of words (not just ones which contain code, but words which contain variables,
1668 constants and other data).
1671 defcode "CREATE",6,,CREATE
1674 call _WORD // Returns %ecx = length, %edi = pointer to word.
1675 mov %edi,%ebx // %ebx = address of the word
1678 movl var_HERE,%edi // %edi is the address of the header
1679 movl var_LATEST,%eax // Get link pointer
1680 stosl // and store it in the header.
1682 // Length byte and the word itself.
1683 mov %cl,%al // Get the length.
1684 stosb // Store the length/flags byte.
1686 mov %ebx,%esi // %esi = word
1687 rep movsb // Copy the word
1689 addl $3,%edi // Align to next 4 byte boundary.
1692 // Update LATEST and HERE.
1694 movl %eax,var_LATEST
1699 Because I want to define : (COLON) in FORTH, not assembler, we need a few more FORTH words
1702 The first is , (COMMA) which is a standard FORTH word which appends a 32 bit integer to the user
1703 data area pointed to by HERE, and adds 4 to HERE. So the action of , (COMMA) is:
1705 previous value of HERE
1708 +---------+---+---+---+---+---+---+---+---+-- - - - - --+------------+
1709 | LINK | 6 | D | O | U | B | L | E | 0 | | <data> |
1710 +---------+---+---+---+---+---+---+---+---+-- - - - - --+------------+
1715 and <data> is whatever 32 bit integer was at the top of the stack.
1717 , (COMMA) is quite a fundamental operation when compiling. It is used to append codewords
1718 to the current word that is being compiled.
1721 defcode ",",1,,COMMA
1722 pop %eax // Code pointer to store.
1726 movl var_HERE,%edi // HERE
1728 movl %edi,var_HERE // Update HERE (incremented)
1732 Our definitions of : (COLON) and ; (SEMICOLON) will need to switch to and from compile mode.
1734 Immediate mode vs. compile mode is stored in the global variable STATE, and by updating this
1735 variable we can switch between the two modes.
1737 For various reasons which may become apparent later, FORTH defines two standard words called
1738 [ and ] (LBRAC and RBRAC) which switch between modes:
1740 Word Assembler Action Effect
1741 [ LBRAC STATE := 0 Switch to immediate mode.
1742 ] RBRAC STATE := 1 Switch to compile mode.
1744 [ (LBRAC) is an IMMEDIATE word. The reason is as follows: If we are in compile mode and the
1745 interpreter saw [ then it would compile it rather than running it. We would never be able to
1746 switch back to immediate mode! So we flag the word as IMMEDIATE so that even in compile mode
1747 the word runs immediately, switching us back to immediate mode.
1750 defcode "[",1,F_IMMED,LBRAC
1752 movl %eax,var_STATE // Set STATE to 0.
1755 defcode "]",1,,RBRAC
1756 movl $1,var_STATE // Set STATE to 1.
1760 Now we can define : (COLON) using CREATE. It just calls CREATE, appends DOCOL (the codeword), sets
1761 the word HIDDEN and goes into compile mode.
1764 defword ":",1,,COLON
1765 .int CREATE // CREATE the dictionary entry / header
1766 .int LIT, DOCOL, COMMA // Append DOCOL (the codeword).
1767 .int LATEST, FETCH, HIDDEN // Make the word hidden (see below for definition).
1768 .int RBRAC // Go into compile mode.
1769 .int EXIT // Return from the function.
1772 ; (SEMICOLON) is also elegantly simple. Notice the F_IMMED flag.
1775 defword ";",1,F_IMMED,SEMICOLON
1776 .int LIT, EXIT, COMMA // Append EXIT (so the word will return).
1777 .int LATEST, FETCH, HIDDEN // Toggle hidden flag -- unhide the word (see below for definition).
1778 .int LBRAC // Go back to IMMEDIATE mode.
1779 .int EXIT // Return from the function.
1782 EXTENDING THE COMPILER ----------------------------------------------------------------------
1784 Words flagged with IMMEDIATE (F_IMMED) aren't just for the FORTH compiler to use. You can define
1785 your own IMMEDIATE words too, and this is a crucial aspect when extending basic FORTH, because
1786 it allows you in effect to extend the compiler itself. Does gcc let you do that?
1788 Standard FORTH words like IF, WHILE, ." and so on are all written as extensions to the basic
1789 compiler, and are all IMMEDIATE words.
1791 The IMMEDIATE word toggles the F_IMMED (IMMEDIATE flag) on the most recently defined word,
1792 or on the current word if you call it in the middle of a definition.
1796 : MYIMMEDWORD IMMEDIATE
1800 but some FORTH programmers write this instead:
1806 The two usages are equivalent, to a first approximation.
1809 defcode "IMMEDIATE",9,F_IMMED,IMMEDIATE
1810 movl var_LATEST,%edi // LATEST word.
1811 addl $4,%edi // Point to name/flags byte.
1812 xorb $F_IMMED,(%edi) // Toggle the IMMED bit.
1816 'addr HIDDEN' toggles the hidden flag (F_HIDDEN) of the word defined at addr. To hide the
1817 most recently defined word (used above in : and ; definitions) you would do:
1821 Setting this flag stops the word from being found by FIND, and so can be used to make 'private'
1822 words. For example, to break up a large word into smaller parts you might do:
1824 : SUB1 ... subword ... ;
1825 : SUB2 ... subword ... ;
1826 : SUB3 ... subword ... ;
1827 : MAIN ... defined in terms of SUB1, SUB2, SUB3 ... ;
1828 WORD SUB1 FIND HIDDEN \ Hide SUB1
1829 WORD SUB2 FIND HIDDEN \ Hide SUB2
1830 WORD SUB3 FIND HIDDEN \ Hide SUB3
1832 After this, only MAIN is 'exported' or seen by the rest of the program.
1835 defcode "HIDDEN",6,,HIDDEN
1836 pop %edi // Dictionary entry.
1837 addl $4,%edi // Point to name/flags byte.
1838 xorb $F_HIDDEN,(%edi) // Toggle the HIDDEN bit.
1842 ' (TICK) is a standard FORTH word which returns the codeword pointer of the next word.
1844 The common usage is:
1848 which appends the codeword of FOO to the current word we are defining (this only works in compiled code).
1850 You tend to use ' in IMMEDIATE words. For example an alternate (and rather useless) way to define
1851 a literal 2 might be:
1854 ' LIT , \ Appends LIT to the currently-being-defined word
1855 2 , \ Appends the number 2 to the currently-being-defined word
1862 (If you don't understand how LIT2 works, then you should review the material about compiling words
1863 and immediate mode).
1865 This definition of ' uses a cheat which I copied from buzzard92. As a result it only works in
1866 compiled code. It is possible to write a version of ' based on WORD, FIND, >CFA which works in
1870 lodsl // Get the address of the next word and skip it.
1871 pushl %eax // Push it on the stack.
1875 BRANCHING ----------------------------------------------------------------------
1877 It turns out that all you need in order to define looping constructs, IF-statements, etc.
1880 BRANCH is an unconditional branch. 0BRANCH is a conditional branch (it only branches if the
1881 top of stack is zero).
1883 The diagram below shows how BRANCH works in some imaginary compiled word. When BRANCH executes,
1884 %esi starts by pointing to the offset field (compare to LIT above):
1886 +---------------------+-------+---- - - ---+------------+------------+---- - - - ----+------------+
1887 | (Dictionary header) | DOCOL | | BRANCH | offset | (skipped) | word |
1888 +---------------------+-------+---- - - ---+------------+-----|------+---- - - - ----+------------+
1891 | +-----------------------+
1892 %esi added to offset
1894 The offset is added to %esi to make the new %esi, and the result is that when NEXT runs, execution
1895 continues at the branch target. Negative offsets work as expected.
1897 0BRANCH is the same except the branch happens conditionally.
1899 Now standard FORTH words such as IF, THEN, ELSE, WHILE, REPEAT, etc. can be implemented entirely
1900 in FORTH. They are IMMEDIATE words which append various combinations of BRANCH or 0BRANCH
1901 into the word currently being compiled.
1903 As an example, code written like this:
1905 condition-code IF true-part THEN rest-code
1909 condition-code 0BRANCH OFFSET true-part rest-code
1915 defcode "BRANCH",6,,BRANCH
1916 add (%esi),%esi // add the offset to the instruction pointer
1919 defcode "0BRANCH",7,,ZBRANCH
1921 test %eax,%eax // top of stack is zero?
1922 jz code_BRANCH // if so, jump back to the branch function above
1923 lodsl // otherwise we need to skip the offset
1927 PRINTING STRINGS ----------------------------------------------------------------------
1929 LITSTRING and EMITSTRING are primitives used to implement the ." and S" operators
1930 (which are written in FORTH). See the definition of those operators below.
1933 defcode "LITSTRING",9,,LITSTRING
1934 lodsl // get the length of the string
1935 push %esi // push the address of the start of the string
1936 push %eax // push it on the stack
1937 addl %eax,%esi // skip past the string
1938 addl $3,%esi // but round up to next 4 byte boundary
1942 defcode "EMITSTRING",10,,EMITSTRING
1943 mov $1,%ebx // 1st param: stdout
1944 pop %edx // 3rd param: length of string
1945 pop %ecx // 2nd param: address of string
1946 mov $__NR_write,%eax // write syscall
1951 COLD START AND INTERPRETER ----------------------------------------------------------------------
1953 COLD is the first FORTH function called, almost immediately after the FORTH system "boots".
1955 INTERPRETER is the FORTH interpreter ("toploop", "toplevel" or "REPL" might be a more accurate
1956 description -- see: http://en.wikipedia.org/wiki/REPL).
1959 // COLD must not return (ie. must not call EXIT).
1960 defword "COLD",4,,COLD
1961 .int INTERPRETER // call the interpreter loop (never returns)
1962 .int LIT,1,SYSEXIT // hmmm, but in case it does, exit(1).
1964 /* This interpreter is pretty simple, but remember that in FORTH you can always override
1965 * it later with a more powerful one!
1967 defword "INTERPRETER",11,,INTERPRETER
1968 .int INTERPRET,RDROP,INTERPRETER
1970 defcode "INTERPRET",9,,INTERPRET
1971 call _WORD // Returns %ecx = length, %edi = pointer to word.
1973 // Is it in the dictionary?
1975 movl %eax,interpret_is_lit // Not a literal number (not yet anyway ...)
1976 call _FIND // Returns %eax = pointer to header or 0 if not found.
1977 test %eax,%eax // Found?
1980 // In the dictionary. Is it an IMMEDIATE codeword?
1981 mov %eax,%edi // %edi = dictionary entry
1982 movb 4(%edi),%al // Get name+flags.
1983 push %ax // Just save it for now.
1984 call _TCFA // Convert dictionary entry (in %edi) to codeword pointer.
1986 andb $F_IMMED,%al // Is IMMED flag set?
1988 jnz 4f // If IMMED, jump straight to executing.
1992 1: // Not in the dictionary (not a word) so assume it's a literal number.
1993 incl interpret_is_lit
1994 call _SNUMBER // Returns the parsed number in %eax
1996 mov $LIT,%eax // The word is LIT
1998 2: // Are we compiling or executing?
2001 jz 4f // Jump if executing.
2003 // Compiling - just append the word to the current dictionary definition.
2005 mov interpret_is_lit,%ecx // Was it a literal?
2008 mov %ebx,%eax // Yes, so LIT is followed by a number.
2012 4: // Executing - run it!
2013 mov interpret_is_lit,%ecx // Literal?
2014 test %ecx,%ecx // Literal?
2017 // Not a literal, execute it now. This never returns, but the codeword will
2018 // eventually call NEXT which will reenter the loop in INTERPRETER.
2021 5: // Executing a literal, which means push it on the stack.
2028 .int 0 // Flag used to record if reading a literal
2031 ODDS AND ENDS ----------------------------------------------------------------------
2033 CHAR puts the ASCII code of the first character of the following word on the stack. For example
2034 CHAR A puts 65 on the stack.
2036 SYSEXIT exits the process using Linux exit syscall.
2038 In this FORTH, SYSEXIT must be the last word in the built-in (assembler) dictionary because we
2039 initialise the LATEST variable to point to it. This means that if you want to extend the assembler
2040 part, you must put new words before SYSEXIT, or else change how LATEST is initialised.
2043 defcode "CHAR",4,,CHAR
2044 call _WORD // Returns %ecx = length, %edi = pointer to word.
2046 movb (%edi),%al // Get the first character of the word.
2047 push %eax // Push it onto the stack.
2050 // NB: SYSEXIT must be the last entry in the built-in dictionary.
2051 defcode SYSEXIT,7,,SYSEXIT
2057 START OF FORTH CODE ----------------------------------------------------------------------
2059 We've now reached the stage where the FORTH system is running and self-hosting. All further
2060 words can be written as FORTH itself, including words like IF, THEN, .", etc which in most
2061 languages would be considered rather fundamental.
2063 I used to append this here in the assembly file, but I got sick of fighting against gas's
2064 stupid (lack of) multiline string syntax. So now that is in a separate file called jonesforth.f
2066 If you don't already have that file, download it from http://annexia.org/forth in order
2067 to continue the tutorial.
2070 /* END OF jonesforth.S */