File:  [gforth] / gforth / libcc.fs
Revision 1.20: download - view: text, annotated - select for diffs
Sat Jun 9 15:55:44 2007 UTC (12 years, 1 month ago) by anton
Branches: MAIN
CVS tags: HEAD
Documented and tested calling C function pointers

    1: \ libcc.fs	foreign function interface implemented using a C compiler
    2: 
    3: \ Copyright (C) 2006 Free Software Foundation, Inc.
    4: 
    5: \ This file is part of Gforth.
    6: 
    7: \ Gforth is free software; you can redistribute it and/or
    8: \ modify it under the terms of the GNU General Public License
    9: \ as published by the Free Software Foundation; either version 2
   10: \ of the License, or (at your option) any later version.
   11: 
   12: \ This program is distributed in the hope that it will be useful,
   13: \ but WITHOUT ANY WARRANTY; without even the implied warranty of
   14: \ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   15: \ GNU General Public License for more details.
   16: 
   17: \ You should have received a copy of the GNU General Public License
   18: \ along with this program; if not, write to the Free Software
   19: \ Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
   20: 
   21: 
   22: \ What this implementation does is this: if it sees a declaration like
   23: 
   24: \ \ something that tells it that the current library is libc
   25: \ \c #include <unistd.h>
   26: \ c-function dlseek lseek n d n -- d
   27: 
   28: \ it genererates C code similar to the following:
   29: 
   30: \ #include <gforth.h>
   31: \ #include <unistd.h>
   32: \ 
   33: \ void gforth_c_lseek_ndn_d(void)
   34: \ {
   35: \   Cell *sp = gforth_SP;
   36: \   Float *fp = gforth_FP;
   37: \   long long result;  /* longest type in C */
   38: \   gforth_ll2d(lseek(sp[3],gforth_d2ll(sp[2],sp[1]),sp[0]),sp[3],sp[2]);
   39: \   gforth_SP = sp+2;
   40: \ }
   41: 
   42: \ Then it compiles this code and dynamically links it into the Gforth
   43: \ system (batching and caching are future work).  It also dynamically
   44: \ links lseek.  Performing DLSEEK then puts the function pointer of
   45: \ the function pointer of gforth_c_lseek_ndn_d on the stack and
   46: \ calls CALL-C.
   47: 
   48: \ ToDo:
   49: 
   50: \ Batching, caching and lazy evaluation:
   51: 
   52: \ Batching:
   53: 
   54: \ New words are deferred, and the corresponding C functions are
   55: \ collected in one file, until the first word is EXECUTEd; then the
   56: \ file is compiled and linked into the system, and the word is
   57: \ resolved.
   58: 
   59: \ Caching:
   60: 
   61: \ Instead of compiling all this stuff anew for every execution, we
   62: \ keep the files around and have an index file containing the function
   63: \ names and their corresponding .so files.  If the needed wrapper name
   64: \ is already present, it is just linked instead of generating the
   65: \ wrapper again.  This is all done by loading the index file(s?),
   66: \ which define words for the wrappers in a separate wordlist.
   67: 
   68: \ The files are built in .../lib/gforth/$VERSION/libcc/ or
   69: \ ~/.gforth/libcc/$HOST/.
   70: 
   71: \ other things to do:
   72: 
   73: \ c-variable forth-name c-name
   74: \ c-constant forth-name c-name
   75: 
   76: \ Todo: conversion between function pointers and xts (both directions)
   77: 
   78: \ taking an xt and turning it into a function pointer:
   79: 
   80: \ e.g., assume we have the xt of + and want to create a C function int
   81: \ gforth_callback_plus(int, int), and then pass the pointer to that
   82: \ function:
   83: 
   84: \ There should be Forth code like this:
   85: \   ] + 0 (bye)
   86: \ Assume that the start of this code is START
   87:         
   88: \ Now, there should be a C function:
   89: 
   90: \ int gforth_callback_plus(int p1, int p2)
   91: \ {
   92: \   Cell   *sp = gforth_SP;
   93: \   Float  *fp = gforth_FP;
   94: \   Float  *fp = gforth_FP;
   95: \   Address lp = gforth_LP;
   96: \   sp -= 2;
   97: \   sp[0] = p1;
   98: \   sp[1] = p2;
   99: \   gforth_engine(START, sp, rp, fp, lp);
  100: \   sp += 1;
  101: \   gforth_RP = rp;
  102: \   gforth_SP = sp;
  103: \   gforth_FP = fp;
  104: \   gforth_LP = lp;
  105: \   return sp[0];
  106: \ }
  107: 
  108: \ and the pointer to that function is the C function pointer for the XT of +.
  109: 
  110: \ Future problems:
  111: \   how to combine the Forth code generation with inlining
  112: \   START is not a constant across executions (when caching the C files)
  113: \      Solution: make START a variable, and store into it on startup with dlsym
  114: 
  115: \ Syntax:
  116: \  callback <rettype> <params> <paramtypes> -- <rettype>
  117: 
  118: 
  119: \ data structures
  120: 
  121: \ For every c-function, we have three words: two anonymous words
  122: \ created by c-function-ft (first time) and c-function-rt (run-time),
  123: \ and a named deferred word.  The deferred word first points to the
  124: \ first-time word, then to the run-time word; the run-time word calls
  125: \ the c function.
  126: 
  127: 
  128: require struct.fs
  129: 
  130:     \ counted-string
  131:     
  132: \ c-function-ft word body:
  133: struct
  134:     cell% field cff-cfr \ xt of c-function-rt word
  135:     cell% field cff-deferred \ xt of c-function deferred word
  136:     cell% field cff-lha \ address of the lib-handle for the lib that
  137:                         \ contains the wrapper function of the word
  138:     char% field cff-rtype  \ return type
  139:     char% field cff-np     \ number of parameters
  140:     1 0   field cff-ptypes \ #npar parameter types
  141:     \  counted string: c-name
  142: end-struct cff%
  143: 
  144: variable c-source-file-id \ contains the source file id of the current batch
  145: 0 c-source-file-id !
  146: variable lib-handle-addr \ points to the library handle of the current batch.
  147:                          \ the library handle is 0 if the current
  148:                          \ batch is not yet compiled.
  149: 2variable lib-filename \ filename without extension
  150: 
  151: : .nb ( n -- )
  152:     0 .r ;
  153: 
  154: : const+ ( n1 "name" -- n2 )
  155:     dup constant 1+ ;
  156: 
  157: : front-string { c-addr1 u1 c-addr2 u2 -- c-addr3 u3 }
  158:     \ insert string c-addr2 u2 in buffer c-addr1 u1; c-addr3 u3 is the
  159:     \ remainder of the buffer.
  160:     assert( u1 u2 u>= )
  161:     c-addr2 c-addr1 u2 move
  162:     c-addr1 u1 u2 /string ;
  163: 
  164: : front-char { c-addr1 u1 c -- c-addr3 u2 }
  165:     \ insert c in buffer c-addr1 u1; c-addr3 u3 is the remainder of
  166:     \ the buffer.
  167:     assert( u1 0 u> )
  168:     c c-addr1 c!
  169:     c-addr1 u1 1 /string ;
  170: 
  171: : s+ { addr1 u1 addr2 u2 -- addr u }
  172:     u1 u2 + allocate throw { addr }
  173:     addr1 addr u1 move
  174:     addr2 addr u1 + u2 move
  175:     addr u1 u2 +
  176: ;
  177: 
  178: : append { addr1 u1 addr2 u2 -- addr u }
  179:     addr1 u1 u2 + dup { u } resize throw { addr }
  180:     addr2 addr u1 + u2 move
  181:     addr u ;
  182: 
  183: \ linked list stuff (should go elsewhere)
  184: 
  185: struct
  186:     cell% field list-next
  187:     1 0   field list-payload
  188: end-struct list%
  189: 
  190: : list-insert { node list -- }
  191:     list list-next @ node list-next !
  192:     node list list-next ! ;
  193: 
  194: : list-append { node endlistp -- }
  195:     \ insert node at place pointed to by endlistp
  196:     node endlistp @ list-insert
  197:     node list-next endlistp ! ;
  198: 
  199: : list-map ( ... list xt -- ... )
  200:     \ xt ( ... node -- ... )
  201:     { xt } begin { node }
  202: 	node while
  203: 	    node xt execute
  204: 	    node list-next @
  205:     repeat ;
  206: 
  207: \ C prefix lines
  208: 
  209: \ linked list of longcstrings: [ link | count-cell | characters ]
  210: 
  211: list%
  212:     cell% field c-prefix-count
  213:     1 0   field c-prefix-chars
  214: end-struct c-prefix%
  215: 
  216: variable c-prefix-lines 0 c-prefix-lines !
  217: variable c-prefix-lines-end c-prefix-lines c-prefix-lines-end !
  218: 
  219: : print-c-prefix-line ( node -- )
  220:     dup c-prefix-chars swap c-prefix-count @ type cr ;
  221: 
  222: : print-c-prefix-lines ( -- )
  223:     c-prefix-lines @ ['] print-c-prefix-line list-map ;
  224: 
  225: : save-c-prefix-line ( c-addr u -- )
  226:     c-source-file-id @ ?dup-if
  227: 	>r 2dup r> write-line throw
  228:     then
  229:     align here 0 , c-prefix-lines-end list-append ( c-addr u )
  230:     longstring, ;
  231: 
  232: : \c ( "rest-of-line" -- ) \ gforth backslash-c
  233:     \G One line of C declarations for the C interface
  234:     -1 parse save-c-prefix-line ;
  235: 
  236: s" #include <gforth/" version-string s+ s" /libcc.h>" append ( c-addr u )
  237:   2dup save-c-prefix-line drop free throw
  238: 
  239: \ Types (for parsing)
  240: 
  241: wordlist constant libcc-types
  242: 
  243: get-current libcc-types set-current
  244: 
  245: \ index values
  246: -1
  247: const+ -- \ end of arguments
  248: const+ n \ integer cell
  249: const+ a \ address cell
  250: const+ d \ double
  251: const+ r \ float
  252: const+ func \ C function pointer
  253: const+ void
  254: drop
  255: 
  256: set-current
  257: 
  258: : parse-libcc-type ( "libcc-type" -- u )
  259:     parse-name libcc-types search-wordlist 0= -13 and throw execute ;
  260: 
  261: : parse-function-types ( "{libcc-type}" "--" "libcc-type" -- )
  262:     here 2 chars allot here begin
  263: 	parse-libcc-type dup 0>= while
  264: 	    c,
  265:     repeat
  266:     drop here swap - over char+ c!
  267:     parse-libcc-type dup 0< -32 and throw swap c! ;
  268: 
  269: : type-letter ( n -- c )
  270:     chars s" nadrfv" drop + c@ ;
  271: 
  272: \ count-stacks
  273: 
  274: : count-stacks-n ( fp-change1 sp-change1 -- fp-change2 sp-change2 )
  275:     1+ ;
  276: 
  277: : count-stacks-a ( fp-change1 sp-change1 -- fp-change2 sp-change2 )
  278:     1+ ;
  279: 
  280: : count-stacks-d ( fp-change1 sp-change1 -- fp-change2 sp-change2 )
  281:     2 + ;
  282: 
  283: : count-stacks-r ( fp-change1 sp-change1 -- fp-change2 sp-change2 )
  284:     swap 1+ swap ;
  285: 
  286: : count-stacks-func ( fp-change1 sp-change1 -- fp-change2 sp-change2 )
  287:     1+ ;
  288: 
  289: : count-stacks-void ( fp-change1 sp-change1 -- fp-change2 sp-change2 )
  290: ;
  291: 
  292: create count-stacks-types
  293: ' count-stacks-n ,
  294: ' count-stacks-a ,
  295: ' count-stacks-d ,
  296: ' count-stacks-r ,
  297: ' count-stacks-func ,
  298: ' count-stacks-void ,
  299: 
  300: : count-stacks ( pars -- fp-change sp-change )
  301:     \ pars is an addr u pair
  302:     0 0 2swap over + swap u+do
  303: 	i c@ cells count-stacks-types + @ execute
  304:     loop ;
  305: 
  306: \ gen-pars
  307: 
  308: : gen-par-n ( fp-depth1 sp-depth1 -- fp-depth2 sp-depth2 )
  309:     ." sp[" 1- dup .nb ." ]" ;
  310: 
  311: : gen-par-a ( fp-depth1 sp-depth1 -- fp-depth2 sp-depth2 )
  312:     ." (void *)(" gen-par-n ." )" ;
  313: 
  314: : gen-par-d ( fp-depth1 sp-depth1 -- fp-depth2 sp-depth2 )
  315:     ." gforth_d2ll(" gen-par-n ." ," gen-par-n ." )" ;
  316: 
  317: : gen-par-r ( fp-depth1 sp-depth1 -- fp-depth2 sp-depth2 )
  318:     swap 1- tuck ." fp[" .nb ." ]" ;
  319: 
  320: : gen-par-func ( fp-depth1 sp-depth1 -- fp-depth2 sp-depth2 )
  321:     gen-par-a ;
  322: 
  323: : gen-par-void ( fp-depth1 sp-depth1 -- fp-depth2 sp-depth2 )
  324:     -32 throw ;
  325: 
  326: create gen-par-types
  327: ' gen-par-n ,
  328: ' gen-par-a ,
  329: ' gen-par-d ,
  330: ' gen-par-r ,
  331: ' gen-par-func ,
  332: ' gen-par-void ,
  333: 
  334: : gen-par ( fp-depth1 sp-depth1 partype -- fp-depth2 sp-depth2 )
  335:     cells gen-par-types + @ execute ;
  336: 
  337: \ the call itself
  338: 
  339: : gen-wrapped-call { d: pars d: c-name fp-change1 sp-change1 -- }
  340:     c-name type ." ("
  341:     fp-change1 sp-change1 pars over + swap u+do 
  342: 	i c@ gen-par
  343: 	i 1+ i' < if
  344: 	    ." ,"
  345: 	endif
  346:     loop
  347:     2drop ." )" ;
  348: 
  349: \ calls for various kinds of return values
  350: 
  351: : gen-wrapped-void ( pars c-name fp-change1 sp-change1 -- fp-change sp-change )
  352:     2dup 2>r gen-wrapped-call 2r> ;
  353: 
  354: : gen-wrapped-n ( pars c-name fp-change1 sp-change1 -- fp-change sp-change )
  355:     2dup gen-par-n 2>r ." =" gen-wrapped-call 2r> ;
  356: 
  357: : gen-wrapped-a ( pars c-name fp-change1 sp-change1 -- fp-change sp-change )
  358:     2dup gen-par-n 2>r ." =(Cell)" gen-wrapped-call 2r> ;
  359: 
  360: : gen-wrapped-d ( pars c-name fp-change1 sp-change1 -- fp-change sp-change )
  361:     ." gforth_ll2d(" gen-wrapped-void
  362:     ." ," gen-par-n ." ," gen-par-n ." )" ;
  363: 
  364: : gen-wrapped-r ( pars c-name fp-change1 sp-change1 -- fp-change sp-change )
  365:     2dup gen-par-r 2>r ." =" gen-wrapped-void 2r> ;
  366: 
  367: : gen-wrapped-func ( pars c-name fp-change1 sp-change1 -- fp-change sp-change )
  368:     gen-wrapped-a ;
  369: 
  370: create gen-wrapped-types
  371: ' gen-wrapped-n ,
  372: ' gen-wrapped-a ,
  373: ' gen-wrapped-d ,
  374: ' gen-wrapped-r ,
  375: ' gen-wrapped-func ,
  376: ' gen-wrapped-void ,
  377: 
  378: : gen-wrapped-stmt ( pars c-name fp-change1 sp-change1 ret -- fp-change sp-change )
  379:     cells gen-wrapped-types + @ execute ;
  380: 
  381: : wrapper-function-name ( addr -- c-addr u )
  382:     \ addr points to the return type index of a c-function descriptor
  383:     count { r-type } count { d: pars }
  384:     pars + count { d: c-name }
  385:     s" gforth_c_" { d: prefix }
  386:     prefix nip c-name nip + pars nip + 3 + { u }
  387:     u allocate throw { c-addr }
  388:     c-addr u
  389:     prefix front-string c-name front-string '_ front-char
  390:     pars bounds u+do
  391: 	i c@ type-letter front-char
  392:     loop
  393:     '_ front-char r-type type-letter front-char assert( dup 0= )
  394:     2drop c-addr u ;
  395: 
  396: : gen-wrapper-function ( addr -- )
  397:     \ addr points to the return type index of a c-function descriptor
  398:     dup { descriptor }
  399:     count { ret } count 2dup { d: pars } chars + count { d: c-name }
  400:     ." void " descriptor wrapper-function-name 2dup type drop free throw
  401:     .\" (void)\n"
  402:     .\" {\n  Cell MAYBE_UNUSED *sp = gforth_SP;\n  Float MAYBE_UNUSED *fp = gforth_FP;\n  "
  403:     pars c-name 2over count-stacks ret gen-wrapped-stmt .\" ;\n"
  404:     ?dup-if
  405: 	."   gforth_SP = sp+" .nb .\" ;\n"
  406:     endif
  407:     ?dup-if
  408: 	."   gforth_FP = fp+" .nb .\" ;\n"
  409:     endif
  410:     .\" }\n" ;
  411: 
  412: : tempdir ( -- c-addr u )
  413:     s" TMPDIR" getenv dup 0= if
  414:         2drop s" /tmp"
  415:     then ;
  416: 
  417: : gen-filename ( x -- c-addr u )
  418:     \ generates a filename without extension for lib-handle-addr X
  419:     0 <<# ['] #s $10 base-execute #> 
  420:     tempdir s" /gforth-c-" s+ 2swap append #>> ;
  421: 
  422: : init-c-source-file ( -- )
  423:     c-source-file-id @ 0= if
  424: 	here 0 , dup lib-handle-addr ! gen-filename 2dup lib-filename 2!
  425: 	s" .c" s+ 2dup w/o create-file throw dup c-source-file-id !
  426:         ['] print-c-prefix-lines swap outfile-execute
  427:         drop free throw
  428:     endif ;
  429: 
  430: : c-source-file ( -- file-id )
  431:     c-source-file-id @ assert( dup ) ;
  432: 
  433: : compile-wrapper-function ( -- )
  434:     c-source-file close-file throw
  435:     0 c-source-file-id !
  436:     s" gcc -fPIC -shared -Wl,-soname," lib-filename 2@ s+
  437:     s" .so.1 -Wl,-export_dynamic -o " append lib-filename 2@ append
  438:     [ s" .so.1 -O -I " s" includedir" getenv append s"  " append ] sliteral
  439:     append lib-filename 2@ append s" .c" append ( c-addr u )
  440:     2dup system drop free throw
  441:     $? abort" compiler generated error" \ !! call dlerror
  442:     lib-filename 2@ s" .so.1" s+
  443:     2dup open-lib dup 0= abort" open-lib failed" \ !! call dlerror
  444:     ( lib-handle ) lib-handle-addr @ !
  445:     2dup delete-file throw drop free throw
  446:     lib-filename 2@ s" .c" s+ 2dup delete-file throw drop free throw
  447:     lib-filename 2@ drop free throw 0 0 lib-filename 2! ;
  448: \    s" ar rcs xxx.a xxx.o" system
  449: \    $? abort" ar generated error" ;
  450: 
  451: : link-wrapper-function { cff -- sym }
  452:     cff cff-rtype wrapper-function-name { d: wrapper-name }
  453:     wrapper-name cff cff-lha @ @ assert( dup ) lib-sym dup 0= -&32 and throw
  454:     wrapper-name drop free throw ;
  455: 
  456: : c-function-ft ( xt-defr xt-cfr "c-name" "{libcc-type}" "--" "libcc-type" -- )
  457:     \ build time/first time action for c-function
  458:     init-c-source-file
  459:     noname create 2, lib-handle-addr @ ,
  460:     parse-name { d: c-name }
  461:     here parse-function-types c-name string,
  462:     ['] gen-wrapper-function c-source-file outfile-execute
  463:   does> ( ... -- ... )
  464:     dup 2@ { xt-defer xt-cfr }
  465:     dup cff-lha @ @ 0= if
  466: 	compile-wrapper-function
  467:     endif
  468:     link-wrapper-function xt-cfr >body !
  469:     xt-cfr xt-defer defer!
  470:     xt-cfr execute ;
  471: 
  472: : c-function-rt ( -- )
  473:     \ run-time definition for c function; addr is the address where
  474:     \ the sym should be stored
  475:     noname create 0 ,
  476:   does> ( ... -- ... )
  477:     @ call-c ;
  478: 
  479: : c-function ( "forth-name" "c-name" "@{type@}" "--" "type" -- ) \ gforth
  480:     \G Define a Forth word @i{forth-name}.  @i{Forth-name} has the
  481:     \G specified stack effect and calls the C function @code{c-name}.
  482:     defer lastxt dup c-function-rt lastxt c-function-ft
  483:     lastxt swap defer! ;

FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>