outrdf2.c

开源的nasm编译器源码,研究编译器原理很有帮且

/* outrdf2.c	output routines for the Netwide Assembler to produce
 *		RDOFF version 2 format object files (which are intended
 *		mainly for use in proprietary projects, as the code to 
 *		load and execute them is very simple). They will also be 
 *		used for device drivers and possibly some executable files
 *		in the MOSCOW operating system. See Rdoff.txt for
 *		details.
 *
 * The Netwide Assembler is copyright (C) 1996-1998 Simon Tatham and
 * Julian Hall. All rights reserved. The software is
 * redistributable under the licence given in the file "Licence"
 * distributed in the NASM archive.
 */

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <ctype.h>
#include <assert.h>

#include "nasm.h"
#include "nasmlib.h"
#include "outform.h"

/* VERBOSE_WARNINGS: define this to add some extra warnings... */
#define VERBOSE_WARNINGS     

#ifdef OF_RDF2

#define RDF_MAXSEGS 64	/* maximum number of segments - user configurable */

typedef unsigned short int16;
typedef unsigned char byte;

static const char *RDOFF2Id = "RDOFF2";	/* written to start of RDOFF files */


/* the records that can be found in the RDOFF header */

/* Note that whenever a segment is referred to in the RDOFF file, its number
 * is always half of the segment number that NASM uses to refer to it; this
 * is because NASM only allocates even numbered segments, so as to not
 * waste any of the 16 bits of segment number written to the file - this
 * allows up to 65533 external labels to be defined; otherwise it would be
 * 32764. */
 
#define RDFREC_RELOC		1
#define RDFREC_IMPORT		2
#define RDFREC_GLOBAL		3
#define RDFREC_DLL		4
#define RDFREC_BSS		5
#define RDFREC_SEGRELOC		6
#define RDFREC_FARIMPORT	7
#define RDFREC_MODNAME		8
#define RDFREC_COMMON		10
#define RDFREC_GENERIC		0


struct RelocRec {
  byte	type;		/* must be 1, or 6 for segment base ref */
  byte	reclen;		/* set to 8 */
  byte	segment;	/* only 0 for code, or 1 for data supported,
			 * but add 64 for relative refs (ie do not require
			 * reloc @ loadtime, only linkage) */
  long	offset;		/* from start of segment in which reference is loc'd */
  byte	length;		/* 1 2 or 4 bytes */
  int16	refseg;		/* segment to which reference refers to */
};

struct ImportRec {
  byte 	type;		/* must be 2, or 7 for FAR import */
  byte	reclen;		/* equals 3+label length */
  int16	segment;	/* segment number allocated to the label for reloc
			 * records - label is assumed to be at offset zero
			 * in this segment, so linker must fix up with offset
			 * of segment and of offset within segment */
  char	label[33];	/* zero terminated... should be written to file until
			 * the zero, but not after it - max len = 32 chars */
};

struct ExportRec {
  byte	type;		/* must be 3 */
  byte  reclen;		/* equals 7+label length */
  byte  flags;          /* SYM_* flags (see below) */
  byte	segment;	/* segment referred to (0/1) */
  long	offset;		/* offset within segment */
  char	label[33];	/* zero terminated as above. max len = 32 chars */
};

struct BSSRec {
  byte	type;		/* must be 5 */
  byte  reclen;		/* equals 4 */
  long	amount;		/* number of bytes BSS to reserve */
};

struct DLLModRec {
  byte type;		/* 4 for DLLRec, 8 for ModRec */
  byte reclen;		/* 1+lib name length for DLLRec, 1+mod name length */
  char name[128];	/* library to link at load time or module name */
};

struct CommonRec {
  byte	type;		/* must be 10 */
  byte  reclen;		/* equals 9+label length */
  int16 segment;	/* segment number */
  long	size;		/* size of common variable */
  int16 align;		/* alignment (power of two) */
  char	label[33];	/* zero terminated as above. max len = 32 chars */
};

/* Flags for ExportRec */
#define SYM_DATA	0x01
#define SYM_FUNCTION	0x02
#define SYM_GLOBAL	0x04

#define COUNT_SEGTYPES 9

static char * segmenttypes[COUNT_SEGTYPES] = {
  "null", "text", "code", "data", "comment", "lcomment", "pcomment",
  "symdebug", "linedebug" 
};

static int segmenttypenumbers[COUNT_SEGTYPES] = {
  0, 1, 1, 2, 3, 4, 5, 6, 7
};

/* code for managing buffers needed to seperate code and data into individual
 * sections until they are ready to be written to the file.
 * We'd better hope that it all fits in memory else we're buggered... */

#define BUF_BLOCK_LEN 4088		/* selected to match page size (4096)
                                         * on 80x86 machines for efficiency */

/***********************************************************************
 * Actual code to deal with RDOFF2 ouput format begins here...
 */

/* global variables set during the initialisation phase */

static struct SAA *seg[RDF_MAXSEGS];	/* seg 0 = code, seg 1 = data */
static struct SAA *header;	/* relocation/import/export records */

static FILE *ofile;

static efunc error;

static struct seginfo {
  char *segname;
  int   segnumber;
  int16 segtype;
  int16 segreserved;
  long  seglength;
} segments[RDF_MAXSEGS];

static int nsegments;

static long bsslength;
static long headerlength;

static void rdf2_init(FILE *fp, efunc errfunc, ldfunc ldef, evalfunc eval)
{
  int segtext, segdata, segbss;

  /* set up the initial segments */
  segments[0].segname = ".text"; 
  segments[0].segnumber = 0;
  segments[0].segtype = 1;
  segments[0].segreserved = 0;
  segments[0].seglength = 0;

  segments[1].segname = ".data";
  segments[1].segnumber = 1;
  segments[1].segtype = 2;
  segments[1].segreserved = 0;
  segments[1].seglength = 0;

  segments[2].segname = ".bss";
  segments[2].segnumber = 2;
  segments[2].segtype = 0xFFFF;	/* reserved - should never be produced */
  segments[2].segreserved = 0;
  segments[2].seglength = 0;

  nsegments = 3;

  ofile = fp;
  error = errfunc;

  seg[0] = saa_init(1L);
  seg[1] = saa_init(1L);
  seg[2] = NULL;		/* special case! */

  header = saa_init(1L);

  segtext = seg_alloc();
  segdata = seg_alloc();
  segbss = seg_alloc();
  if (segtext != 0 || segdata != 2 || segbss != 4)
    error(ERR_PANIC,"rdf segment numbers not allocated as expected (%d,%d,%d)",
	  segtext,segdata,segbss);
  bsslength=0;
  headerlength = 0;
}

static long rdf2_section_names(char *name, int pass, int *bits)
{
  int 	i;
  char 	* p, * q;
  int	code = -1;
  int 	reserved = 0;

  /*
   * Default is 32 bits, in the text segment.
   */
  if (!name) {
    *bits = 32;
    return 0;
  }

  /* look for segment type code following segment name */
  p = name;
  while (*p && !isspace(*p)) p++;
  if (*p) {	/* we're now in whitespace */
    *p++ = '\0';
    while (*p && isspace(80)) *p++ = '\0';
  }
  if (*p) {	/* we're now in an attribute value */
    /*
     * see if we have an optional ',number' following the type code
     */
    if ((q = strchr(p, ','))) {
      *q++ = '\0';

      reserved = readnum(q, &i);
      if (i) {
	  error(ERR_NONFATAL, "value following comma must be numeric");
	  reserved = 0;
      }
    }
    /*
     * check it against the text strings in segmenttypes 
     */

    for (i = 0; i < COUNT_SEGTYPES; i++)
      if (!nasm_stricmp(p, segmenttypes[i])) {
	code = segmenttypenumbers[i];
	break;
      }
    if (code == -1) {	/* didn't find anything */
      code = readnum(p, &i);
      if (i) {
	error(ERR_NONFATAL, "unrecognised RDF segment type (%s)",p);
	code = 3;
      }
    }
  }    
  for (i = 0; i < nsegments; i++) {
    if (!strcmp(name, segments[i].segname)) {
      if (code != -1 || reserved != 0) 
	error(ERR_NONFATAL, "segment attributes specified on"
	      " redeclaration of segment");
      return segments[i].segnumber * 2;
    }
  }

  /* declaring a new segment! */

  if (code == -1) {
    error(ERR_NONFATAL, "new segment declared without type code");
    code = 3;
  }
  if (nsegments == RDF_MAXSEGS) {
    error(ERR_FATAL, "reached compiled-in maximum segment limit (%d)",
	  RDF_MAXSEGS);
    return NO_SEG;
  }

  segments[nsegments].segname = nasm_strdup(name);
  i = seg_alloc();
  if (i % 2 != 0)
    error(ERR_PANIC, "seg_alloc() returned odd number");
  segments[nsegments].segnumber = i >> 1;
  segments[nsegments].segtype = code;
  segments[nsegments].segreserved = reserved;
  segments[nsegments].seglength = 0;

  seg[nsegments] = saa_init(1L);

  return i;
}

static void write_reloc_rec(struct RelocRec *r)
{
  char buf[4],*b;

  if (r->refseg != (int16)NO_SEG && (r->refseg & 1)) /* segment base ref */
      r->type = RDFREC_SEGRELOC;

  r->refseg >>= 1;    /* adjust segment nos to RDF rather than NASM */

  saa_wbytes(header,&r->type,1);
  saa_wbytes(header,&r->reclen,1);
  saa_wbytes(header,&r->segment,1);
  b = buf; WRITELONG(b,r->offset);
  saa_wbytes(header,buf,4);
  saa_wbytes(header,&r->length,1);
  b = buf; WRITESHORT(b,r->refseg);
  saa_wbytes(header,buf,2);
  headerlength += r->reclen + 2;
}

static void write_export_rec(struct ExportRec *r)
{
  char buf[4], *b;

  r->segment >>= 1;

  saa_wbytes(header,&r->type,1);
  saa_wbytes(header,&r->reclen,1);
  saa_wbytes(header,&r->flags,1);
  saa_wbytes(header,&r->segment,1);
  b = buf; WRITELONG(b,r->offset);
  saa_wbytes(header,buf,4);
  saa_wbytes(header,r->label,strlen(r->label) + 1);
  headerlength += r->reclen + 2;
}

static void write_import_rec(struct ImportRec *r)
{
  char buf[4], *b;

  r->segment >>= 1;

  saa_wbytes(header,&r->type,1);
  saa_wbytes(header,&r->reclen,1);
  b = buf; WRITESHORT(b,r->segment);
  saa_wbytes(header,buf,2);
  saa_wbytes(header,r->label,strlen(r->label) + 1);
  headerlength += r->reclen + 2;
}

static void write_bss_rec(struct BSSRec *r)
{
    char buf[4], *b;

    saa_wbytes(header,&r->type,1);
    saa_wbytes(header,&r->reclen,1);
    b = buf; WRITELONG(b,r->amount);
    saa_wbytes(header,buf,4);
    headerlength += r->reclen + 2;
}

/*
 * Write common variable record.
 */
static void write_common_rec(struct CommonRec *r)
{
  char buf[4], *b;

  r->segment >>= 1;

  saa_wbytes(header,&r->type,1);
  saa_wbytes(header,&r->reclen,1);
  b = buf; WRITESHORT(b,r->segment);
  saa_wbytes(header,buf,2);
  b = buf; WRITELONG(b,r->size);
  saa_wbytes(header,buf,4);
  b = buf; WRITESHORT(b,r->align);
  saa_wbytes(header,buf,2);
  saa_wbytes(header,r->label,strlen(r->label) + 1);
  headerlength += r->reclen + 2;
}

/*
 * Write library record. Also used for module name records.
 */
static void write_dllmod_rec(struct DLLModRec *r)
{
    saa_wbytes(header,&r->type,1);
    saa_wbytes(header,&r->reclen,1);
    saa_wbytes(header,r->name,strlen(r->name) + 1);
    headerlength += r->reclen + 2;
}

static void rdf2_deflabel(char *name, long segment, long offset,
			 int is_global, char *special)
{
  struct ExportRec r;
  struct ImportRec ri;
  struct CommonRec ci;
  static int farsym = 0;
  static int i;
  byte export_flags = 0;

  if (is_global == 2) {
    /* Common variable */
    ci.type = RDFREC_COMMON;