mirror/nasm
2
0
Fork 0
mirror of https://github.com/netwide-assembler/nasm.git synced 2024-12-21 09:19:31 +08:00
Code Issues Packages Projects Releases Wiki Activity
1b67bd25b2
nasm/output/outobj.c
Slavik Gnatenko 1b67bd25b2 BR 2010180: outobj: Garbage may be written in a last PUBDEF 
The testcase illustrates the problem. After "nasm -f obj
alonesym.nasm"
let's look to dump:

======
PUBDEF386(91) recnum:5, offset:0000005bh, len:03f9h, chksum:bbh(bb)
Group: 0, Seg: 1
00020000h - 'sym0000' Type:0
00020004h - 'sym0001' Type:0
....
00020134h - 'sym0077' Type:0

PUBDEF(90) recnum:6, offset:00000457h, len:000ah, chksum:b6h(b6)
Group: 0, Seg: 1
00000138h - 's' Type:2
0000b600h - '' Type:0
======

The problem is while 's' offset is 20138h it is marked as type 90h not
91h.  The root cause is located in obj_x():

static ObjRecord *obj_x(ObjRecord * orp, uint32_t val)
{
    if (orp->type & 1)
    	orp->x_size = 32;
    if (val > 0xFFFF)
        orp = obj_force(orp, 32);
    if (orp->x_size == 32)
        return (obj_dword(orp, val));
    orp->x_size = 16;
    return (obj_word(orp, val));
}

It sets up x_size and than writes data. In the testcase data are the
offset and this offset overflows a record. In this case the record is
emitted and its x_size is cleared. Because this is last PUBDEF the new
record with only 's' symbol is emitted also but its x_size is not 32
(it's still zero) so obj_fwrite doesn't switch to 91h type.

The problem seems to be very generic and expected to be occurred on
many other record types as well.

        ----

And the fix is simple:

if (orp->x_size == 32)
{
  ObjRecord * nxt = obj_dword(orp, val);
  nxt->x_size = 32; /* x_size is cleared when a record overflows */
  return nxt;
}
2008-07-19 19:27:41 -07:00

2542 lines
75 KiB
C
Raw Blame History

/* outobj.c output routines for the Netwide Assembler to produce
* .OBJ object files
*
* The Netwide Assembler is copyright (C) 1996 Simon Tatham and
* Julian Hall. All rights reserved. The software is
* redistributable under the license given in the file "LICENSE"
* distributed in the NASM archive.
*/
#include "compiler.h"
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <ctype.h>
#include <inttypes.h>
#include "nasm.h"
#include "nasmlib.h"
#include "stdscan.h"
#include "outform.h"
#ifdef OF_OBJ
/*
* outobj.c is divided into two sections. The first section is low level
* routines for creating obj records; It has nearly zero NASM specific
* code. The second section is high level routines for processing calls and
* data structures from the rest of NASM into obj format.
*
* It should be easy (though not zero work) to lift the first section out for
* use as an obj file writer for some other assembler or compiler.
*/
/*
* These routines are built around the ObjRecord data struture. An ObjRecord
* holds an object file record that may be under construction or complete.
*
* A major function of these routines is to support continuation of an obj
* record into the next record when the maximum record size is exceeded. The
* high level code does not need to worry about where the record breaks occur.
* It does need to do some minor extra steps to make the automatic continuation
* work. Those steps may be skipped for records where the high level knows no
* continuation could be required.
*
* 1) An ObjRecord is allocated and cleared by obj_new, or an existing ObjRecord
* is cleared by obj_clear.
*
* 2) The caller should fill in .type.
*
* 3) If the record is continuable and there is processing that must be done at
* the start of each record then the caller should fill in .ori with the
* address of the record initializer routine.
*
* 4) If the record is continuable and it should be saved (rather than emitted
* immediately) as each record is done, the caller should set .up to be a
* pointer to a location in which the caller keeps the master pointer to the
* ObjRecord. When the record is continued, the obj_bump routine will then
* allocate a new ObjRecord structure and update the master pointer.
*
* 5) If the .ori field was used then the caller should fill in the .parm with
* any data required by the initializer.
*
* 6) The caller uses the routines: obj_byte, obj_word, obj_rword, obj_dword,
* obj_x, obj_index, obj_value and obj_name to fill in the various kinds of
* data required for this record.
*
* 7) If the record is continuable, the caller should call obj_commit at each
* point where breaking the record is permitted.
*
* 8) To write out the record, the caller should call obj_emit2. If the
* caller has called obj_commit for all data written then he can get slightly
* faster code by calling obj_emit instead of obj_emit2.
*
* Most of these routines return an ObjRecord pointer. This will be the input
* pointer most of the time and will be the new location if the ObjRecord
* moved as a result of the call. The caller may ignore the return value in
* three cases: It is a "Never Reallocates" routine; or The caller knows
* continuation is not possible; or The caller uses the master pointer for the
* next operation.
*/
#define RECORD_MAX (1024-3) /* maximal size of any record except type+reclen */
#define OBJ_PARMS 3 /* maximum .parm used by any .ori routine */
#define FIX_08_LOW 0x8000 /* location type for various fixup subrecords */
#define FIX_16_OFFSET 0x8400
#define FIX_16_SELECTOR 0x8800
#define FIX_32_POINTER 0x8C00
#define FIX_08_HIGH 0x9000
#define FIX_32_OFFSET 0xA400
#define FIX_48_POINTER 0xAC00
enum RecordID { /* record ID codes */
THEADR = 0x80, /* module header */
COMENT = 0x88, /* comment record */
LINNUM = 0x94, /* line number record */
LNAMES = 0x96, /* list of names */
SEGDEF = 0x98, /* segment definition */
GRPDEF = 0x9A, /* group definition */
EXTDEF = 0x8C, /* external definition */
PUBDEF = 0x90, /* public definition */
COMDEF = 0xB0, /* common definition */
LEDATA = 0xA0, /* logical enumerated data */
FIXUPP = 0x9C, /* fixups (relocations) */
FIXU32 = 0x9D, /* 32-bit fixups (relocations) */
MODEND = 0x8A, /* module end */
MODE32 = 0x8B /* module end for 32-bit objects */
};
enum ComentID { /* ID codes for comment records */
dEXTENDED = 0xA1, /* tells that we are using translator-specific extensions */
dLINKPASS = 0xA2, /* link pass 2 marker */
dTYPEDEF = 0xE3, /* define a type */
dSYM = 0xE6, /* symbol debug record */
dFILNAME = 0xE8, /* file name record */
dCOMPDEF = 0xEA /* compiler type info */
};
typedef struct ObjRecord ObjRecord;
typedef void ORI(ObjRecord * orp);
struct ObjRecord {
ORI *ori; /* Initialization routine */
int used; /* Current data size */
int committed; /* Data size at last boundary */
int x_size; /* (see obj_x) */
unsigned int type; /* Record type */
ObjRecord *child; /* Associated record below this one */
ObjRecord **up; /* Master pointer to this ObjRecord */
ObjRecord *back; /* Previous part of this record */
uint32_t parm[OBJ_PARMS]; /* Parameters for ori routine */
uint8_t buf[RECORD_MAX + 3];
};
static void obj_fwrite(ObjRecord * orp);
static void ori_ledata(ObjRecord * orp);
static void ori_pubdef(ObjRecord * orp);
static void ori_null(ObjRecord * orp);
static ObjRecord *obj_commit(ObjRecord * orp);
static bool obj_uppercase; /* Flag: all names in uppercase */
static bool obj_use32; /* Flag: at least one segment is 32-bit */
/*
* Clear an ObjRecord structure. (Never reallocates).
* To simplify reuse of ObjRecord's, .type, .ori and .parm are not cleared.
*/
static ObjRecord *obj_clear(ObjRecord * orp)
{
orp->used = 0;
orp->committed = 0;
orp->x_size = 0;
orp->child = NULL;
orp->up = NULL;
orp->back = NULL;
return (orp);
}
/*
* Emit an ObjRecord structure. (Never reallocates).
* The record is written out preceeded (recursively) by its previous part (if
* any) and followed (recursively) by its child (if any).
* The previous part and the child are freed. The main ObjRecord is cleared,
* not freed.
*/
static ObjRecord *obj_emit(ObjRecord * orp)
{
if (orp->back) {
obj_emit(orp->back);
nasm_free(orp->back);
}
if (orp->committed)
obj_fwrite(orp);
if (orp->child) {
obj_emit(orp->child);
nasm_free(orp->child);
}
return (obj_clear(orp));
}
/*
* Commit and Emit a record. (Never reallocates).
*/
static ObjRecord *obj_emit2(ObjRecord * orp)
{
obj_commit(orp);
return (obj_emit(orp));
}
/*
* Allocate and clear a new ObjRecord; Also sets .ori to ori_null
*/
static ObjRecord *obj_new(void)
{
ObjRecord *orp;
orp = obj_clear(nasm_malloc(sizeof(ObjRecord)));
orp->ori = ori_null;
return (orp);
}
/*
* Advance to the next record because the existing one is full or its x_size
* is incompatible.
* Any uncommited data is moved into the next record.
*/
static ObjRecord *obj_bump(ObjRecord * orp)
{
ObjRecord *nxt;
int used = orp->used;
int committed = orp->committed;
if (orp->up) {
*orp->up = nxt = obj_new();
nxt->ori = orp->ori;
nxt->type = orp->type;
nxt->up = orp->up;
nxt->back = orp;
memcpy(nxt->parm, orp->parm, sizeof(orp->parm));
} else
nxt = obj_emit(orp);
used -= committed;
if (used) {
nxt->committed = 1;
nxt->ori(nxt);
nxt->committed = nxt->used;
memcpy(nxt->buf + nxt->committed, orp->buf + committed, used);
nxt->used = nxt->committed + used;
}
return (nxt);
}
/*
* Advance to the next record if necessary to allow the next field to fit.
*/
static ObjRecord *obj_check(ObjRecord * orp, int size)
{
if (orp->used + size > RECORD_MAX)
orp = obj_bump(orp);
if (!orp->committed) {
orp->committed = 1;
orp->ori(orp);
orp->committed = orp->used;
}
return (orp);
}
/*
* All data written so far is commited to the current record (won't be moved to
* the next record in case of continuation).
*/
static ObjRecord *obj_commit(ObjRecord * orp)
{
orp->committed = orp->used;
return (orp);
}
/*
* Write a byte
*/
static ObjRecord *obj_byte(ObjRecord * orp, uint8_t val)
{
orp = obj_check(orp, 1);
orp->buf[orp->used] = val;
orp->used++;
return (orp);
}
/*
* Write a word
*/
static ObjRecord *obj_word(ObjRecord * orp, unsigned int val)
{
orp = obj_check(orp, 2);
orp->buf[orp->used] = val;
orp->buf[orp->used + 1] = val >> 8;
orp->used += 2;
return (orp);
}
/*
* Write a reversed word
*/
static ObjRecord *obj_rword(ObjRecord * orp, unsigned int val)
{
orp = obj_check(orp, 2);
orp->buf[orp->used] = val >> 8;
orp->buf[orp->used + 1] = val;
orp->used += 2;
return (orp);
}
/*
* Write a dword
*/
static ObjRecord *obj_dword(ObjRecord * orp, uint32_t val)
{
orp = obj_check(orp, 4);
orp->buf[orp->used] = val;
orp->buf[orp->used + 1] = val >> 8;
orp->buf[orp->used + 2] = val >> 16;
orp->buf[orp->used + 3] = val >> 24;
orp->used += 4;
return (orp);
}
/*
* All fields of "size x" in one obj record must be the same size (either 16
* bits or 32 bits). There is a one bit flag in each record which specifies
* which.
* This routine is used to force the current record to have the desired
* x_size. x_size is normally automatic (using obj_x), so that this
* routine should be used outside obj_x, only to provide compatibility with
* linkers that have bugs in their processing of the size bit.
*/
static ObjRecord *obj_force(ObjRecord * orp, int x)
{
if (orp->x_size == (x ^ 48))
orp = obj_bump(orp);
orp->x_size = x;
return (orp);
}
/*
* This routine writes a field of size x. The caller does not need to worry at
* all about whether 16-bits or 32-bits are required.
*/
static ObjRecord *obj_x(ObjRecord * orp, uint32_t val)
{
if (orp->type & 1)
orp->x_size = 32;
if (val > 0xFFFF)
orp = obj_force(orp, 32);
if (orp->x_size == 32) {
ObjRecord *nxt = obj_dword(orp, val);
nxt->x_size = 32; /* x_size is cleared when a record overflows */
return nxt;
}
orp->x_size = 16;
return (obj_word(orp, val));
}
/*
* Writes an index
*/
static ObjRecord *obj_index(ObjRecord * orp, unsigned int val)
{
if (val < 128)
return (obj_byte(orp, val));
return (obj_word(orp, (val >> 8) | (val << 8) | 0x80));
}
/*
* Writes a variable length value
*/
static ObjRecord *obj_value(ObjRecord * orp, uint32_t val)
{
if (val <= 128)
return (obj_byte(orp, val));
if (val <= 0xFFFF) {
orp = obj_byte(orp, 129);
return (obj_word(orp, val));
}
if (val <= 0xFFFFFF)
return (obj_dword(orp, (val << 8) + 132));
orp = obj_byte(orp, 136);
return (obj_dword(orp, val));
}
/*
* Writes a counted string
*/
static ObjRecord *obj_name(ObjRecord * orp, char *name)
{
int len = strlen(name);
uint8_t *ptr;
orp = obj_check(orp, len + 1);
ptr = orp->buf + orp->used;
*ptr++ = len;
orp->used += len + 1;
if (obj_uppercase)
while (--len >= 0) {
*ptr++ = toupper(*name);
name++;
} else
memcpy(ptr, name, len);
return (orp);
}
/*
* Initializer for an LEDATA record.
* parm[0] = offset
* parm[1] = segment index
* During the use of a LEDATA ObjRecord, parm[0] is constantly updated to
* represent the offset that would be required if the record were split at the
* last commit point.
* parm[2] is a copy of parm[0] as it was when the current record was initted.
*/
static void ori_ledata(ObjRecord * orp)
{
obj_index(orp, orp->parm[1]);
orp->parm[2] = orp->parm[0];
obj_x(orp, orp->parm[0]);
}
/*
* Initializer for a PUBDEF record.
* parm[0] = group index
* parm[1] = segment index
* parm[2] = frame (only used when both indexes are zero)
*/
static void ori_pubdef(ObjRecord * orp)
{
obj_index(orp, orp->parm[0]);
obj_index(orp, orp->parm[1]);
if (!(orp->parm[0] | orp->parm[1]))
obj_word(orp, orp->parm[2]);
}
/*
* Initializer for a LINNUM record.
* parm[0] = group index
* parm[1] = segment index
*/
static void ori_linnum(ObjRecord * orp)
{
obj_index(orp, orp->parm[0]);
obj_index(orp, orp->parm[1]);
}
/*
* Initializer for a local vars record.
*/
static void ori_local(ObjRecord * orp)
{
obj_byte(orp, 0x40);
obj_byte(orp, dSYM);
}
/*
* Null initializer for records that continue without any header info
*/
static void ori_null(ObjRecord * orp)
{
(void)orp; /* Do nothing */
}
/*
* This concludes the low level section of outobj.c
*/
static char obj_infile[FILENAME_MAX];
static efunc error;
static evalfunc evaluate;
static ldfunc deflabel;
static FILE *ofp;
static int32_t first_seg;
static bool any_segs;
static int passtwo;
static int arrindex;
#define GROUP_MAX 256 /* we won't _realistically_ have more
* than this many segs in a group */
#define EXT_BLKSIZ 256 /* block size for externals list */
struct Segment; /* need to know these structs exist */
struct Group;
struct LineNumber {
struct LineNumber *next;
struct Segment *segment;
int32_t offset;
int32_t lineno;
};
static struct FileName {
struct FileName *next;
char *name;
struct LineNumber *lnhead, **lntail;
int index;
} *fnhead, **fntail;
static struct Array {
struct Array *next;
unsigned size;
int basetype;
} *arrhead, **arrtail;
#define ARRAYBOT 31 /* magic number for first array index */
static struct Public {
struct Public *next;
char *name;
int32_t offset;
int32_t segment; /* only if it's far-absolute */
int type; /* only for local debug syms */
} *fpubhead, **fpubtail, *last_defined;
static struct External {
struct External *next;
char *name;
int32_t commonsize;
int32_t commonelem; /* element size if FAR, else zero */
int index; /* OBJ-file external index */
enum {
DEFWRT_NONE, /* no unusual default-WRT */
DEFWRT_STRING, /* a string we don't yet understand */
DEFWRT_SEGMENT, /* a segment */
DEFWRT_GROUP /* a group */
} defwrt_type;
union {
char *string;
struct Segment *seg;
struct Group *grp;
} defwrt_ptr;
struct External *next_dws; /* next with DEFWRT_STRING */
} *exthead, **exttail, *dws;
static int externals;
static struct ExtBack {
struct ExtBack *next;
struct External *exts[EXT_BLKSIZ];
} *ebhead, **ebtail;
static struct Segment {
struct Segment *next;
int32_t index; /* the NASM segment id */
int32_t obj_index; /* the OBJ-file segment index */
struct Group *grp; /* the group it beint32_ts to */
uint32_t currentpos;
int32_t align; /* can be SEG_ABS + absolute addr */
enum {
CMB_PRIVATE = 0,
CMB_PUBLIC = 2,
CMB_STACK = 5,
CMB_COMMON = 6
} combine;
bool use32; /* is this segment 32-bit? */
struct Public *pubhead, **pubtail, *lochead, **loctail;
char *name;
char *segclass, *overlay; /* `class' is a C++ keyword :-) */
ObjRecord *orp;
} *seghead, **segtail, *obj_seg_needs_update;
static struct Group {
struct Group *next;
char *name;
int32_t index; /* NASM segment id */
int32_t obj_index; /* OBJ-file group index */
int32_t nentries; /* number of elements... */
int32_t nindices; /* ...and number of index elts... */
union {
int32_t index;
char *name;
} segs[GROUP_MAX]; /* ...in this */
} *grphead, **grptail, *obj_grp_needs_update;
static struct ImpDef {
struct ImpDef *next;
char *extname;
char *libname;
unsigned int impindex;
char *impname;
} *imphead, **imptail;
static struct ExpDef {
struct ExpDef *next;
char *intname;
char *extname;
unsigned int ordinal;
int flags;
} *exphead, **exptail;
#define EXPDEF_FLAG_ORDINAL 0x80
#define EXPDEF_FLAG_RESIDENT 0x40
#define EXPDEF_FLAG_NODATA 0x20
#define EXPDEF_MASK_PARMCNT 0x1F
static int32_t obj_entry_seg, obj_entry_ofs;
struct ofmt of_obj;
/* The current segment */
static struct Segment *current_seg;
static int32_t obj_segment(char *, int, int *);
static void obj_write_file(int debuginfo);
static int obj_directive(char *, char *, int);
static void obj_init(FILE * fp, efunc errfunc, ldfunc ldef, evalfunc eval)
{
ofp = fp;
error = errfunc;
evaluate = eval;
deflabel = ldef;
first_seg = seg_alloc();
any_segs = false;
fpubhead = NULL;
fpubtail = &fpubhead;
exthead = NULL;
exttail = &exthead;
imphead = NULL;
imptail = &imphead;
exphead = NULL;
exptail = &exphead;
dws = NULL;
externals = 0;
ebhead = NULL;
ebtail = &ebhead;
seghead = obj_seg_needs_update = NULL;
segtail = &seghead;
grphead = obj_grp_needs_update = NULL;
grptail = &grphead;
obj_entry_seg = NO_SEG;
obj_uppercase = false;
obj_use32 = false;
passtwo = 0;
current_seg = NULL;
of_obj.current_dfmt->init(&of_obj, NULL, fp, errfunc);
}
static int obj_set_info(enum geninfo type, char **val)
{
(void)type;
(void)val;
return 0;
}
static void obj_cleanup(int debuginfo)
{
obj_write_file(debuginfo);
of_obj.current_dfmt->cleanup();
fclose(ofp);
while (seghead) {
struct Segment *segtmp = seghead;
seghead = seghead->next;
while (segtmp->pubhead) {
struct Public *pubtmp = segtmp->pubhead;
segtmp->pubhead = pubtmp->next;
nasm_free(pubtmp->name);
nasm_free(pubtmp);
}
nasm_free(segtmp->segclass);
nasm_free(segtmp->overlay);
nasm_free(segtmp);
}
while (fpubhead) {
struct Public *pubtmp = fpubhead;
fpubhead = fpubhead->next;
nasm_free(pubtmp->name);
nasm_free(pubtmp);
}
while (exthead) {
struct External *exttmp = exthead;
exthead = exthead->next;
nasm_free(exttmp);
}
while (imphead) {
struct ImpDef *imptmp = imphead;
imphead = imphead->next;
nasm_free(imptmp->extname);
nasm_free(imptmp->libname);
nasm_free(imptmp->impname); /* nasm_free won't mind if it's NULL */
nasm_free(imptmp);
}
while (exphead) {
struct ExpDef *exptmp = exphead;
exphead = exphead->next;
nasm_free(exptmp->extname);
nasm_free(exptmp->intname);
nasm_free(exptmp);
}
while (ebhead) {
struct ExtBack *ebtmp = ebhead;
ebhead = ebhead->next;
nasm_free(ebtmp);
}
while (grphead) {
struct Group *grptmp = grphead;
grphead = grphead->next;
nasm_free(grptmp);
}
}
static void obj_ext_set_defwrt(struct External *ext, char *id)
{
struct Segment *seg;
struct Group *grp;
for (seg = seghead; seg; seg = seg->next)
if (!strcmp(seg->name, id)) {
ext->defwrt_type = DEFWRT_SEGMENT;
ext->defwrt_ptr.seg = seg;
nasm_free(id);
return;
}
for (grp = grphead; grp; grp = grp->next)
if (!strcmp(grp->name, id)) {
ext->defwrt_type = DEFWRT_GROUP;
ext->defwrt_ptr.grp = grp;
nasm_free(id);
return;
}
ext->defwrt_type = DEFWRT_STRING;
ext->defwrt_ptr.string = id;
ext->next_dws = dws;
dws = ext;
}
static void obj_deflabel(char *name, int32_t segment,
int64_t offset, int is_global, char *special)
{
/*
* We have three cases:
*
* (i) `segment' is a segment-base. If so, set the name field
* for the segment or group structure it refers to, and then
* return.
*
* (ii) `segment' is one of our segments, or a SEG_ABS segment.
* Save the label position for later output of a PUBDEF record.
* (Or a MODPUB, if we work out how.)
*
* (iii) `segment' is not one of our segments. Save the label
* position for later output of an EXTDEF, and also store a
* back-reference so that we can map later references to this
* segment number to the external index.
*/
struct External *ext;
struct ExtBack *eb;
struct Segment *seg;
int i;
bool used_special = false; /* have we used the special text? */
#if defined(DEBUG) && DEBUG>2
fprintf(stderr,
" obj_deflabel: %s, seg=%ld, off=%ld, is_global=%d, %s\n",
name, segment, offset, is_global, special);
#endif
/*
* If it's a special-retry from pass two, discard it.
*/
if (is_global == 3)
return;
/*
* First check for the double-period, signifying something
* unusual.
*/
if (name[0] == '.' && name[1] == '.' && name[2] != '@') {
if (!strcmp(name, "..start")) {
obj_entry_seg = segment;
obj_entry_ofs = offset;
return;
}
error(ERR_NONFATAL, "unrecognised special symbol `%s'", name);
}
/*
* Case (i):
*/
if (obj_seg_needs_update) {
obj_seg_needs_update->name = name;
return;
} else if (obj_grp_needs_update) {
obj_grp_needs_update->name = name;
return;
}
if (segment < SEG_ABS && segment != NO_SEG && segment % 2)
return;
if (segment >= SEG_ABS || segment == NO_SEG) {
/*
* SEG_ABS subcase of (ii).
*/
if (is_global) {
struct Public *pub;
pub = *fpubtail = nasm_malloc(sizeof(*pub));
fpubtail = &pub->next;
pub->next = NULL;
pub->name = nasm_strdup(name);
pub->offset = offset;
pub->segment = (segment == NO_SEG ? 0 : segment & ~SEG_ABS);
}
if (special)
error(ERR_NONFATAL, "OBJ supports no special symbol features"
" for this symbol type");
return;
}
/*
* If `any_segs' is still false, we might need to define a
* default segment, if they're trying to declare a label in
* `first_seg'.
*/
if (!any_segs && segment == first_seg) {
int tempint; /* ignored */
if (segment != obj_segment("__NASMDEFSEG", 2, &tempint))
error(ERR_PANIC, "strange segment conditions in OBJ driver");
}
for (seg = seghead; seg && is_global; seg = seg->next)
if (seg->index == segment) {
struct Public *loc = nasm_malloc(sizeof(*loc));
/*
* Case (ii). Maybe MODPUB someday?
*/
*seg->pubtail = loc;
seg->pubtail = &loc->next;
loc->next = NULL;
loc->name = nasm_strdup(name);
loc->offset = offset;
if (special)
error(ERR_NONFATAL,
"OBJ supports no special symbol features"
" for this symbol type");
return;
}
/*
* Case (iii).
*/
if (is_global) {
ext = *exttail = nasm_malloc(sizeof(*ext));
ext->next = NULL;
exttail = &ext->next;
ext->name = name;
/* Place by default all externs into the current segment */
ext->defwrt_type = DEFWRT_NONE;
/* 28-Apr-2002 - John Coffman
The following code was introduced on 12-Aug-2000, and breaks fixups
on code passed thru the MSC 5.1 linker (3.66) and MSC 6.00A linker
(5.10). It was introduced after FIXUP32 was added, and may be needed
for 32-bit segments. The following will get 16-bit segments working
again, and maybe someone can correct the 'if' condition which is
actually needed.
*/
#if 0
if (current_seg) {
#else
if (current_seg && current_seg->use32) {
if (current_seg->grp) {
ext->defwrt_type = DEFWRT_GROUP;
ext->defwrt_ptr.grp = current_seg->grp;
} else {
ext->defwrt_type = DEFWRT_SEGMENT;
ext->defwrt_ptr.seg = current_seg;
}
}
#endif
if (is_global == 2) {
ext->commonsize = offset;
ext->commonelem = 1; /* default FAR */
} else
ext->commonsize = 0;
} else
return;
/*
* Now process the special text, if any, to find default-WRT
* specifications and common-variable element-size and near/far
* specifications.
*/
while (special && *special) {
used_special = true;
/*
* We might have a default-WRT specification.
*/
if (!nasm_strnicmp(special, "wrt", 3)) {
char *p;
int len;
special += 3;
special += strspn(special, " \t");
p = nasm_strndup(special, len = strcspn(special, ":"));
obj_ext_set_defwrt(ext, p);
special += len;
if (*special && *special != ':')
error(ERR_NONFATAL, "`:' expected in special symbol"
" text for `%s'", ext->name);
else if (*special == ':')
special++;
}
/*
* The NEAR or FAR keywords specify nearness or
* farness. FAR gives default element size 1.
*/
if (!nasm_strnicmp(special, "far", 3)) {
if (ext->commonsize)
ext->commonelem = 1;
else
error(ERR_NONFATAL,
"`%s': `far' keyword may only be applied"
" to common variables\n", ext->name);
special += 3;
special += strspn(special, " \t");
} else if (!nasm_strnicmp(special, "near", 4)) {
if (ext->commonsize)
ext->commonelem = 0;
else
error(ERR_NONFATAL,
"`%s': `far' keyword may only be applied"
" to common variables\n", ext->name);
special += 4;
special += strspn(special, " \t");
}
/*
* If it's a common, and anything else remains on the line
* before a further colon, evaluate it as an expression and
* use that as the element size. Forward references aren't
* allowed.
*/
if (*special == ':')
special++;
else if (*special) {
if (ext->commonsize) {
expr *e;
struct tokenval tokval;
stdscan_reset();
stdscan_bufptr = special;
tokval.t_type = TOKEN_INVALID;
e = evaluate(stdscan, NULL, &tokval, NULL, 1, error, NULL);
if (e) {
if (!is_simple(e))
error(ERR_NONFATAL, "cannot use relocatable"
" expression as common-variable element size");
else
ext->commonelem = reloc_value(e);
}
special = stdscan_bufptr;
} else {
error(ERR_NONFATAL,
"`%s': element-size specifications only"
" apply to common variables", ext->name);
while (*special && *special != ':')
special++;
if (*special == ':')
special++;
}
}
}
i = segment / 2;
eb = ebhead;
if (!eb) {
eb = *ebtail = nasm_malloc(sizeof(*eb));
eb->next = NULL;
ebtail = &eb->next;
}
while (i >= EXT_BLKSIZ) {
if (eb && eb->next)
eb = eb->next;
else {
eb = *ebtail = nasm_malloc(sizeof(*eb));
eb->next = NULL;
ebtail = &eb->next;
}
i -= EXT_BLKSIZ;
}
eb->exts[i] = ext;
ext->index = ++externals;
if (special && !used_special)
error(ERR_NONFATAL, "OBJ supports no special symbol features"
" for this symbol type");
}
/* forward declaration */
static void obj_write_fixup(ObjRecord * orp, int bytes,
int segrel, int32_t seg, int32_t wrt,
struct Segment *segto);
static void obj_out(int32_t segto, const void *data,
enum out_type type, uint64_t size,
int32_t segment, int32_t wrt)
{
const uint8_t *ucdata;
int32_t ldata;
struct Segment *seg;
ObjRecord *orp;
/*
* handle absolute-assembly (structure definitions)
*/
if (segto == NO_SEG) {
if (type != OUT_RESERVE)
error(ERR_NONFATAL, "attempt to assemble code in [ABSOLUTE]"
" space");
return;
}
/*
* If `any_segs' is still false, we must define a default
* segment.
*/
if (!any_segs) {
int tempint; /* ignored */
if (segto != obj_segment("__NASMDEFSEG", 2, &tempint))
error(ERR_PANIC, "strange segment conditions in OBJ driver");
}
/*
* Find the segment we are targetting.
*/
for (seg = seghead; seg; seg = seg->next)
if (seg->index == segto)
break;
if (!seg)
error(ERR_PANIC, "code directed to nonexistent segment?");
orp = seg->orp;
orp->parm[0] = seg->currentpos;
if (type == OUT_RAWDATA) {
ucdata = data;
while (size > 0) {
unsigned int len;
orp = obj_check(seg->orp, 1);
len = RECORD_MAX - orp->used;
if (len > size)
len = size;
memcpy(orp->buf + orp->used, ucdata, len);
orp->committed = orp->used += len;
orp->parm[0] = seg->currentpos += len;
ucdata += len;
size -= len;
}
} else if (type == OUT_ADDRESS || type == OUT_REL2ADR ||
type == OUT_REL4ADR) {
int rsize;
if (segment == NO_SEG && type != OUT_ADDRESS)
error(ERR_NONFATAL, "relative call to absolute address not"
" supported by OBJ format");
if (segment >= SEG_ABS)
error(ERR_NONFATAL, "far-absolute relocations not supported"
" by OBJ format");
ldata = *(int64_t *)data;
if (type == OUT_REL2ADR) {
ldata += (size - 2);
size = 2;
}
if (type == OUT_REL4ADR) {
ldata += (size - 4);
size = 4;
}
if (size == 2)
orp = obj_word(orp, ldata);
else
orp = obj_dword(orp, ldata);
rsize = size;
if (segment < SEG_ABS && (segment != NO_SEG && segment % 2) &&
size == 4) {
/*
* This is a 4-byte segment-base relocation such as
* `MOV EAX,SEG foo'. OBJ format can't actually handle
* these, but if the constant term has the 16 low bits
* zero, we can just apply a 2-byte segment-base
* relocation to the low word instead.
*/
rsize = 2;
if (ldata & 0xFFFF)
error(ERR_NONFATAL, "OBJ format cannot handle complex"
" dword-size segment base references");
}
if (segment != NO_SEG)
obj_write_fixup(orp, rsize,
(type == OUT_ADDRESS ? 0x4000 : 0),
segment, wrt, seg);
seg->currentpos += size;
} else if (type == OUT_RESERVE) {
if (orp->committed)
orp = obj_bump(orp);
seg->currentpos += size;
}
obj_commit(orp);
}
static void obj_write_fixup(ObjRecord * orp, int bytes,
int segrel, int32_t seg, int32_t wrt,
struct Segment *segto)
{
unsigned locat;
int method;
int base;
int32_t tidx, fidx;
struct Segment *s = NULL;
struct Group *g = NULL;
struct External *e = NULL;
ObjRecord *forp;
if (bytes == 1) {
error(ERR_NONFATAL, "`obj' output driver does not support"
" one-byte relocations");
return;
}
forp = orp->child;
if (forp == NULL) {
orp->child = forp = obj_new();
forp->up = &(orp->child);
/* We should choose between FIXUPP and FIXU32 record type */
/* If we're targeting a 32-bit segment, use a FIXU32 record */
if (segto->use32)
forp->type = FIXU32;
else
forp->type = FIXUPP;
}
if (seg % 2) {
base = true;
locat = FIX_16_SELECTOR;
seg--;
if (bytes != 2)
error(ERR_PANIC, "OBJ: 4-byte segment base fixup got"
" through sanity check");
} else {
base = false;
locat = (bytes == 2) ? FIX_16_OFFSET : FIX_32_OFFSET;
if (!segrel)
/*
* There is a bug in tlink that makes it process self relative
* fixups incorrectly if the x_size doesn't match the location
* size.
*/
forp = obj_force(forp, bytes << 3);
}
forp = obj_rword(forp, locat | segrel | (orp->parm[0] - orp->parm[2]));
tidx = fidx = -1, method = 0; /* placate optimisers */
/*
* See if we can find the segment ID in our segment list. If
* so, we have a T4 (LSEG) target.
*/
for (s = seghead; s; s = s->next)
if (s->index == seg)
break;
if (s)
method = 4, tidx = s->obj_index;
else {
for (g = grphead; g; g = g->next)
if (g->index == seg)
break;
if (g)
method = 5, tidx = g->obj_index;
else {
int32_t i = seg / 2;
struct ExtBack *eb = ebhead;
while (i >= EXT_BLKSIZ) {
if (eb)
eb = eb->next;
else
break;
i -= EXT_BLKSIZ;
}
if (eb)
method = 6, e = eb->exts[i], tidx = e->index;
else
error(ERR_PANIC,
"unrecognised segment value in obj_write_fixup");
}
}
/*
* If no WRT given, assume the natural default, which is method
* F5 unless:
*
* - we are doing an OFFSET fixup for a grouped segment, in
* which case we require F1 (group).
*
* - we are doing an OFFSET fixup for an external with a
* default WRT, in which case we must honour the default WRT.
*/
if (wrt == NO_SEG) {
if (!base && s && s->grp)
method |= 0x10, fidx = s->grp->obj_index;
else if (!base && e && e->defwrt_type != DEFWRT_NONE) {
if (e->defwrt_type == DEFWRT_SEGMENT)
method |= 0x00, fidx = e->defwrt_ptr.seg->obj_index;
else if (e->defwrt_type == DEFWRT_GROUP)
method |= 0x10, fidx = e->defwrt_ptr.grp->obj_index;
else {
error(ERR_NONFATAL, "default WRT specification for"
" external `%s' unresolved", e->name);
method |= 0x50, fidx = -1; /* got to do _something_ */
}
} else
method |= 0x50, fidx = -1;
} else {
/*
* See if we can find the WRT-segment ID in our segment
* list. If so, we have a F0 (LSEG) frame.
*/
for (s = seghead; s; s = s->next)
if (s->index == wrt - 1)
break;
if (s)
method |= 0x00, fidx = s->obj_index;
else {
for (g = grphead; g; g = g->next)
if (g->index == wrt - 1)
break;
if (g)
method |= 0x10, fidx = g->obj_index;
else {
int32_t i = wrt / 2;
struct ExtBack *eb = ebhead;
while (i >= EXT_BLKSIZ) {
if (eb)
eb = eb->next;
else
break;
i -= EXT_BLKSIZ;
}
if (eb)
method |= 0x20, fidx = eb->exts[i]->index;
else
error(ERR_PANIC,
"unrecognised WRT value in obj_write_fixup");
}
}
}
forp = obj_byte(forp, method);
if (fidx != -1)
forp = obj_index(forp, fidx);
forp = obj_index(forp, tidx);
obj_commit(forp);
}
static int32_t obj_segment(char *name, int pass, int *bits)
{
/*
* We call the label manager here to define a name for the new
* segment, and when our _own_ label-definition stub gets
* called in return, it should register the new segment name
* using the pointer it gets passed. That way we save memory,
* by sponging off the label manager.
*/
#if defined(DEBUG) && DEBUG>=3
fprintf(stderr, " obj_segment: < %s >, pass=%d, *bits=%d\n",
name, pass, *bits);
#endif
if (!name) {
*bits = 16;
current_seg = NULL;
return first_seg;
} else {
struct Segment *seg;
struct Group *grp;
struct External **extp;
int obj_idx, i, attrs;
bool rn_error;
char *p;
/*
* Look for segment attributes.
*/
attrs = 0;
while (*name == '.')
name++; /* hack, but a documented one */
p = name;
while (*p && !nasm_isspace(*p))
p++;
if (*p) {
*p++ = '\0';
while (*p && nasm_isspace(*p))
*p++ = '\0';
}
while (*p) {
while (*p && !nasm_isspace(*p))
p++;
if (*p) {
*p++ = '\0';
while (*p && nasm_isspace(*p))
*p++ = '\0';
}
attrs++;
}
obj_idx = 1;
for (seg = seghead; seg; seg = seg->next) {
obj_idx++;
if (!strcmp(seg->name, name)) {
if (attrs > 0 && pass == 1)
error(ERR_WARNING, "segment attributes specified on"
" redeclaration of segment: ignoring");
if (seg->use32)
*bits = 32;
else
*bits = 16;
current_seg = seg;
return seg->index;
}
}
*segtail = seg = nasm_malloc(sizeof(*seg));
seg->next = NULL;
segtail = &seg->next;
seg->index = (any_segs ? seg_alloc() : first_seg);
seg->obj_index = obj_idx;
seg->grp = NULL;
any_segs = true;
seg->name = NULL;
seg->currentpos = 0;
seg->align = 1; /* default */
seg->use32 = false; /* default */
seg->combine = CMB_PUBLIC; /* default */
seg->segclass = seg->overlay = NULL;
seg->pubhead = NULL;
seg->pubtail = &seg->pubhead;
seg->lochead = NULL;
seg->loctail = &seg->lochead;
seg->orp = obj_new();
seg->orp->up = &(seg->orp);
seg->orp->ori = ori_ledata;
seg->orp->type = LEDATA;
seg->orp->parm[1] = obj_idx;
/*
* Process the segment attributes.
*/
p = name;
while (attrs--) {
p += strlen(p);
while (!*p)
p++;
/*
* `p' contains a segment attribute.
*/
if (!nasm_stricmp(p, "private"))
seg->combine = CMB_PRIVATE;
else if (!nasm_stricmp(p, "public"))
seg->combine = CMB_PUBLIC;
else if (!nasm_stricmp(p, "common"))
seg->combine = CMB_COMMON;
else if (!nasm_stricmp(p, "stack"))
seg->combine = CMB_STACK;
else if (!nasm_stricmp(p, "use16"))
seg->use32 = false;
else if (!nasm_stricmp(p, "use32"))
seg->use32 = true;
else if (!nasm_stricmp(p, "flat")) {
/*
* This segment is an OS/2 FLAT segment. That means
* that its default group is group FLAT, even if
* the group FLAT does not explicitly _contain_ the
* segment.
*
* When we see this, we must create the group
* `FLAT', containing no segments, if it does not
* already exist; then we must set the default
* group of this segment to be the FLAT group.
*/
struct Group *grp;
for (grp = grphead; grp; grp = grp->next)
if (!strcmp(grp->name, "FLAT"))
break;
if (!grp) {
obj_directive("group", "FLAT", 1);
for (grp = grphead; grp; grp = grp->next)
if (!strcmp(grp->name, "FLAT"))
break;
if (!grp)
error(ERR_PANIC, "failure to define FLAT?!");
}
seg->grp = grp;
} else if (!nasm_strnicmp(p, "class=", 6))
seg->segclass = nasm_strdup(p + 6);
else if (!nasm_strnicmp(p, "overlay=", 8))
seg->overlay = nasm_strdup(p + 8);
else if (!nasm_strnicmp(p, "align=", 6)) {
seg->align = readnum(p + 6, &rn_error);
if (rn_error) {
seg->align = 1;
error(ERR_NONFATAL, "segment alignment should be"
" numeric");
}
switch ((int)seg->align) {
case 1: /* BYTE */
case 2: /* WORD */
case 4: /* DWORD */
case 16: /* PARA */
case 256: /* PAGE */
case 4096: /* PharLap extension */
break;
case 8:
error(ERR_WARNING,
"OBJ format does not support alignment"
" of 8: rounding up to 16");
seg->align = 16;
break;
case 32:
case 64:
case 128:
error(ERR_WARNING,
"OBJ format does not support alignment"
" of %d: rounding up to 256", seg->align);
seg->align = 256;
break;
case 512:
case 1024:
case 2048:
error(ERR_WARNING,
"OBJ format does not support alignment"
" of %d: rounding up to 4096", seg->align);
seg->align = 4096;
break;
default:
error(ERR_NONFATAL, "invalid alignment value %d",
seg->align);
seg->align = 1;
break;
}
} else if (!nasm_strnicmp(p, "absolute=", 9)) {
seg->align = SEG_ABS + readnum(p + 9, &rn_error);
if (rn_error)
error(ERR_NONFATAL, "argument to `absolute' segment"
" attribute should be numeric");
}
}
/* We need to know whenever we have at least one 32-bit segment */
obj_use32 |= seg->use32;
obj_seg_needs_update = seg;
if (seg->align >= SEG_ABS)
deflabel(name, NO_SEG, seg->align - SEG_ABS,
NULL, false, false, &of_obj, error);
else
deflabel(name, seg->index + 1, 0L,
NULL, false, false, &of_obj, error);
obj_seg_needs_update = NULL;
/*
* See if this segment is defined in any groups.
*/
for (grp = grphead; grp; grp = grp->next) {
for (i = grp->nindices; i < grp->nentries; i++) {
if (!strcmp(grp->segs[i].name, seg->name)) {
nasm_free(grp->segs[i].name);
grp->segs[i] = grp->segs[grp->nindices];
grp->segs[grp->nindices++].index = seg->obj_index;
if (seg->grp)
error(ERR_WARNING,
"segment `%s' is already part of"
" a group: first one takes precedence",
seg->name);
else
seg->grp = grp;
}
}
}
/*
* Walk through the list of externals with unresolved
* default-WRT clauses, and resolve any that point at this
* segment.
*/
extp = &dws;
while (*extp) {
if ((*extp)->defwrt_type == DEFWRT_STRING &&
!strcmp((*extp)->defwrt_ptr.string, seg->name)) {
nasm_free((*extp)->defwrt_ptr.string);
(*extp)->defwrt_type = DEFWRT_SEGMENT;
(*extp)->defwrt_ptr.seg = seg;
*extp = (*extp)->next_dws;
} else
extp = &(*extp)->next_dws;
}
if (seg->use32)
*bits = 32;
else
*bits = 16;
current_seg = seg;
return seg->index;
}
}
static int obj_directive(char *directive, char *value, int pass)
{
if (!strcmp(directive, "group")) {
char *p, *q, *v;
if (pass == 1) {
struct Group *grp;
struct Segment *seg;
struct External **extp;
int obj_idx;
q = value;
while (*q == '.')
q++; /* hack, but a documented one */
v = q;
while (*q && !nasm_isspace(*q))
q++;
if (nasm_isspace(*q)) {
*q++ = '\0';
while (*q && nasm_isspace(*q))
q++;
}
/*
* Here we used to sanity-check the group directive to
* ensure nobody tried to declare a group containing no
* segments. However, OS/2 does this as standard
* practice, so the sanity check has been removed.
*
* if (!*q) {
* error(ERR_NONFATAL,"GROUP directive contains no segments");
* return 1;
* }
*/
obj_idx = 1;
for (grp = grphead; grp; grp = grp->next) {
obj_idx++;
if (!strcmp(grp->name, v)) {
error(ERR_NONFATAL, "group `%s' defined twice", v);
return 1;
}
}
*grptail = grp = nasm_malloc(sizeof(*grp));
grp->next = NULL;
grptail = &grp->next;
grp->index = seg_alloc();
grp->obj_index = obj_idx;
grp->nindices = grp->nentries = 0;
grp->name = NULL;
obj_grp_needs_update = grp;
deflabel(v, grp->index + 1, 0L,
NULL, false, false, &of_obj, error);
obj_grp_needs_update = NULL;
while (*q) {
p = q;
while (*q && !nasm_isspace(*q))
q++;
if (nasm_isspace(*q)) {
*q++ = '\0';
while (*q && nasm_isspace(*q))
q++;
}
/*
* Now p contains a segment name. Find it.
*/
for (seg = seghead; seg; seg = seg->next)
if (!strcmp(seg->name, p))
break;
if (seg) {
/*
* We have a segment index. Shift a name entry
* to the end of the array to make room.
*/
grp->segs[grp->nentries++] = grp->segs[grp->nindices];
grp->segs[grp->nindices++].index = seg->obj_index;
if (seg->grp)
error(ERR_WARNING,
"segment `%s' is already part of"
" a group: first one takes precedence",
seg->name);
else
seg->grp = grp;
} else {
/*
* We have an as-yet undefined segment.
* Remember its name, for later.
*/
grp->segs[grp->nentries++].name = nasm_strdup(p);
}
}
/*
* Walk through the list of externals with unresolved
* default-WRT clauses, and resolve any that point at
* this group.
*/
extp = &dws;
while (*extp) {
if ((*extp)->defwrt_type == DEFWRT_STRING &&
!strcmp((*extp)->defwrt_ptr.string, grp->name)) {
nasm_free((*extp)->defwrt_ptr.string);
(*extp)->defwrt_type = DEFWRT_GROUP;
(*extp)->defwrt_ptr.grp = grp;
*extp = (*extp)->next_dws;
} else
extp = &(*extp)->next_dws;
}
}
return 1;
}
if (!strcmp(directive, "uppercase")) {
obj_uppercase = true;
return 1;
}
if (!strcmp(directive, "import")) {
char *q, *extname, *libname, *impname;
if (pass == 2)
return 1; /* ignore in pass two */
extname = q = value;
while (*q && !nasm_isspace(*q))
q++;
if (nasm_isspace(*q)) {
*q++ = '\0';
while (*q && nasm_isspace(*q))
q++;
}
libname = q;
while (*q && !nasm_isspace(*q))
q++;
if (nasm_isspace(*q)) {
*q++ = '\0';
while (*q && nasm_isspace(*q))
q++;
}
impname = q;
if (!*extname || !*libname)
error(ERR_NONFATAL, "`import' directive requires symbol name"
" and library name");
else {
struct ImpDef *imp;
bool err = false;
imp = *imptail = nasm_malloc(sizeof(struct ImpDef));
imptail = &imp->next;
imp->next = NULL;
imp->extname = nasm_strdup(extname);
imp->libname = nasm_strdup(libname);
imp->impindex = readnum(impname, &err);
if (!*impname || err)
imp->impname = nasm_strdup(impname);
else
imp->impname = NULL;
}
return 1;
}
if (!strcmp(directive, "export")) {
char *q, *extname, *intname, *v;
struct ExpDef *export;
int flags = 0;
unsigned int ordinal = 0;
if (pass == 2)
return 1; /* ignore in pass two */
intname = q = value;
while (*q && !nasm_isspace(*q))
q++;
if (nasm_isspace(*q)) {
*q++ = '\0';
while (*q && nasm_isspace(*q))
q++;
}
extname = q;
while (*q && !nasm_isspace(*q))
q++;
if (nasm_isspace(*q)) {
*q++ = '\0';
while (*q && nasm_isspace(*q))
q++;
}
if (!*intname) {
error(ERR_NONFATAL, "`export' directive requires export name");
return 1;
}
if (!*extname) {
extname = intname;
intname = "";
}
while (*q) {
v = q;
while (*q && !nasm_isspace(*q))
q++;
if (nasm_isspace(*q)) {
*q++ = '\0';
while (*q && nasm_isspace(*q))
q++;
}
if (!nasm_stricmp(v, "resident"))
flags |= EXPDEF_FLAG_RESIDENT;
else if (!nasm_stricmp(v, "nodata"))
flags |= EXPDEF_FLAG_NODATA;
else if (!nasm_strnicmp(v, "parm=", 5)) {
bool err = false;
flags |= EXPDEF_MASK_PARMCNT & readnum(v + 5, &err);
if (err) {
error(ERR_NONFATAL,
"value `%s' for `parm' is non-numeric", v + 5);
return 1;
}
} else {
bool err = false;
ordinal = readnum(v, &err);
if (err) {
error(ERR_NONFATAL,
"unrecognised export qualifier `%s'", v);
return 1;
}
flags |= EXPDEF_FLAG_ORDINAL;
}
}
export = *exptail = nasm_malloc(sizeof(struct ExpDef));
exptail = &export->next;
export->next = NULL;
export->extname = nasm_strdup(extname);
export->intname = nasm_strdup(intname);
export->ordinal = ordinal;
export->flags = flags;
return 1;
}
return 0;
}
static int32_t obj_segbase(int32_t segment)
{
struct Segment *seg;
/*
* Find the segment in our list.
*/
for (seg = seghead; seg; seg = seg->next)
if (seg->index == segment - 1)
break;
if (!seg) {
/*
* Might be an external with a default WRT.
*/
int32_t i = segment / 2;
struct ExtBack *eb = ebhead;
struct External *e;
while (i >= EXT_BLKSIZ) {
if (eb)
eb = eb->next;
else
break;
i -= EXT_BLKSIZ;
}
if (eb) {
e = eb->exts[i];
if (e->defwrt_type == DEFWRT_NONE)
return segment; /* fine */
else if (e->defwrt_type == DEFWRT_SEGMENT)
return e->defwrt_ptr.seg->index + 1;
else if (e->defwrt_type == DEFWRT_GROUP)
return e->defwrt_ptr.grp->index + 1;
else
return NO_SEG; /* can't tell what it is */
}
return segment; /* not one of ours - leave it alone */
}
if (seg->align >= SEG_ABS)
return seg->align; /* absolute segment */
if (seg->grp)
return seg->grp->index + 1; /* grouped segment */
return segment; /* no special treatment */
}
static void obj_filename(char *inname, char *outname, efunc lerror)
{
strcpy(obj_infile, inname);
standard_extension(inname, outname, ".obj", lerror);
}
static void obj_write_file(int debuginfo)
{
struct Segment *seg, *entry_seg_ptr = 0;
struct FileName *fn;
struct LineNumber *ln;
struct Group *grp;
struct Public *pub, *loc;
struct External *ext;
struct ImpDef *imp;
struct ExpDef *export;
static char boast[] = "The Netwide Assembler " NASM_VER;
int lname_idx;
ObjRecord *orp;
/*
* Write the THEADR module header.
*/
orp = obj_new();
orp->type = THEADR;
obj_name(orp, obj_infile);
obj_emit2(orp);
/*
* Write the NASM boast comment.
*/
orp->type = COMENT;
obj_rword(orp, 0); /* comment type zero */
obj_name(orp, boast);
obj_emit2(orp);
orp->type = COMENT;
/*
* Write the IMPDEF records, if any.
*/
for (imp = imphead; imp; imp = imp->next) {
obj_rword(orp, 0xA0); /* comment class A0 */
obj_byte(orp, 1); /* subfunction 1: IMPDEF */
if (imp->impname)
obj_byte(orp, 0); /* import by name */
else
obj_byte(orp, 1); /* import by ordinal */
obj_name(orp, imp->extname);
obj_name(orp, imp->libname);
if (imp->impname)
obj_name(orp, imp->impname);
else
obj_word(orp, imp->impindex);
obj_emit2(orp);
}
/*
* Write the EXPDEF records, if any.
*/
for (export = exphead; export; export = export->next) {
obj_rword(orp, 0xA0); /* comment class A0 */
obj_byte(orp, 2); /* subfunction 2: EXPDEF */
obj_byte(orp, export->flags);
obj_name(orp, export->extname);
obj_name(orp, export->intname);
if (export->flags & EXPDEF_FLAG_ORDINAL)
obj_word(orp, export->ordinal);
obj_emit2(orp);
}
/* we're using extended OMF if we put in debug info */
if (debuginfo) {
orp->type = COMENT;
obj_byte(orp, 0x40);
obj_byte(orp, dEXTENDED);
obj_emit2(orp);
}
/*
* Write the first LNAMES record, containing LNAME one, which
* is null. Also initialize the LNAME counter.
*/
orp->type = LNAMES;
obj_byte(orp, 0);
lname_idx = 1;
/*
* Write some LNAMES for the segment names
*/
for (seg = seghead; seg; seg = seg->next) {
orp = obj_name(orp, seg->name);
if (seg->segclass)
orp = obj_name(orp, seg->segclass);
if (seg->overlay)
orp = obj_name(orp, seg->overlay);
obj_commit(orp);
}
/*
* Write some LNAMES for the group names
*/
for (grp = grphead; grp; grp = grp->next) {
orp = obj_name(orp, grp->name);
obj_commit(orp);
}
obj_emit(orp);
/*
* Write the SEGDEF records.
*/
orp->type = SEGDEF;
for (seg = seghead; seg; seg = seg->next) {
int acbp;
uint32_t seglen = seg->currentpos;
acbp = (seg->combine << 2); /* C field */
if (seg->use32)
acbp |= 0x01; /* P bit is Use32 flag */
else if (seglen == 0x10000L) {
seglen = 0; /* This special case may be needed for old linkers */
acbp |= 0x02; /* B bit */
}
/* A field */
if (seg->align >= SEG_ABS)
/* acbp |= 0x00 */ ;
else if (seg->align >= 4096) {
if (seg->align > 4096)
error(ERR_NONFATAL, "segment `%s' requires more alignment"
" than OBJ format supports", seg->name);
acbp |= 0xC0; /* PharLap extension */
} else if (seg->align >= 256) {
acbp |= 0x80;
} else if (seg->align >= 16) {
acbp |= 0x60;
} else if (seg->align >= 4) {
acbp |= 0xA0;
} else if (seg->align >= 2) {
acbp |= 0x40;
} else
acbp |= 0x20;
obj_byte(orp, acbp);
if (seg->align & SEG_ABS) {
obj_x(orp, seg->align - SEG_ABS); /* Frame */
obj_byte(orp, 0); /* Offset */
}
obj_x(orp, seglen);
obj_index(orp, ++lname_idx);
obj_index(orp, seg->segclass ? ++lname_idx : 1);
obj_index(orp, seg->overlay ? ++lname_idx : 1);
obj_emit2(orp);
}
/*
* Write the GRPDEF records.
*/
orp->type = GRPDEF;
for (grp = grphead; grp; grp = grp->next) {
int i;
if (grp->nindices != grp->nentries) {
for (i = grp->nindices; i < grp->nentries; i++) {
error(ERR_NONFATAL, "group `%s' contains undefined segment"
" `%s'", grp->name, grp->segs[i].name);
nasm_free(grp->segs[i].name);
grp->segs[i].name = NULL;
}
}
obj_index(orp, ++lname_idx);
for (i = 0; i < grp->nindices; i++) {
obj_byte(orp, 0xFF);
obj_index(orp, grp->segs[i].index);
}
obj_emit2(orp);
}
/*
* Write the PUBDEF records: first the ones in the segments,
* then the far-absolutes.
*/
orp->type = PUBDEF;
orp->ori = ori_pubdef;
for (seg = seghead; seg; seg = seg->next) {
orp->parm[0] = seg->grp ? seg->grp->obj_index : 0;
orp->parm[1] = seg->obj_index;
for (pub = seg->pubhead; pub; pub = pub->next) {
orp = obj_name(orp, pub->name);
orp = obj_x(orp, pub->offset);
orp = obj_byte(orp, 0); /* type index */
obj_commit(orp);
}
obj_emit(orp);
}
orp->parm[0] = 0;
orp->parm[1] = 0;
for (pub = fpubhead; pub; pub = pub->next) { /* pub-crawl :-) */
if (orp->parm[2] != (uint32_t)pub->segment) {
obj_emit(orp);
orp->parm[2] = pub->segment;
}
orp = obj_name(orp, pub->name);
orp = obj_x(orp, pub->offset);
orp = obj_byte(orp, 0); /* type index */
obj_commit(orp);
}
obj_emit(orp);
/*
* Write the EXTDEF and COMDEF records, in order.
*/
orp->ori = ori_null;
for (ext = exthead; ext; ext = ext->next) {
if (ext->commonsize == 0) {
if (orp->type != EXTDEF) {
obj_emit(orp);
orp->type = EXTDEF;
}
orp = obj_name(orp, ext->name);
orp = obj_index(orp, 0);
} else {
if (orp->type != COMDEF) {
obj_emit(orp);
orp->type = COMDEF;
}
orp = obj_name(orp, ext->name);
orp = obj_index(orp, 0);
if (ext->commonelem) {
orp = obj_byte(orp, 0x61); /* far communal */
orp = obj_value(orp, (ext->commonsize / ext->commonelem));
orp = obj_value(orp, ext->commonelem);
} else {
orp = obj_byte(orp, 0x62); /* near communal */
orp = obj_value(orp, ext->commonsize);
}
}
obj_commit(orp);
}
obj_emit(orp);
/*
* Write a COMENT record stating that the linker's first pass
* may stop processing at this point. Exception is if our
* MODEND record specifies a start point, in which case,
* according to some variants of the documentation, this COMENT
* should be omitted. So we'll omit it just in case.
* But, TASM puts it in all the time so if we are using
* TASM debug stuff we are putting it in
*/
if (debuginfo || obj_entry_seg == NO_SEG) {
orp->type = COMENT;
obj_byte(orp, 0x40);
obj_byte(orp, dLINKPASS);
obj_byte(orp, 1);
obj_emit2(orp);
}
/*
* 1) put out the compiler type
* 2) Put out the type info. The only type we are using is near label #19
*/
if (debuginfo) {
int i;
struct Array *arrtmp = arrhead;
orp->type = COMENT;
obj_byte(orp, 0x40);
obj_byte(orp, dCOMPDEF);
obj_byte(orp, 4);
obj_byte(orp, 0);
obj_emit2(orp);
obj_byte(orp, 0x40);
obj_byte(orp, dTYPEDEF);
obj_word(orp, 0x18); /* type # for linking */
obj_word(orp, 6); /* size of type */
obj_byte(orp, 0x2a); /* absolute type for debugging */
obj_emit2(orp);
obj_byte(orp, 0x40);
obj_byte(orp, dTYPEDEF);
obj_word(orp, 0x19); /* type # for linking */
obj_word(orp, 0); /* size of type */
obj_byte(orp, 0x24); /* absolute type for debugging */
obj_byte(orp, 0); /* near/far specifier */
obj_emit2(orp);
obj_byte(orp, 0x40);
obj_byte(orp, dTYPEDEF);
obj_word(orp, 0x1A); /* type # for linking */
obj_word(orp, 0); /* size of type */
obj_byte(orp, 0x24); /* absolute type for debugging */
obj_byte(orp, 1); /* near/far specifier */
obj_emit2(orp);
obj_byte(orp, 0x40);
obj_byte(orp, dTYPEDEF);
obj_word(orp, 0x1b); /* type # for linking */
obj_word(orp, 0); /* size of type */
obj_byte(orp, 0x23); /* absolute type for debugging */
obj_byte(orp, 0);
obj_byte(orp, 0);
obj_byte(orp, 0);
obj_emit2(orp);
obj_byte(orp, 0x40);
obj_byte(orp, dTYPEDEF);
obj_word(orp, 0x1c); /* type # for linking */
obj_word(orp, 0); /* size of type */
obj_byte(orp, 0x23); /* absolute type for debugging */
obj_byte(orp, 0);
obj_byte(orp, 4);
obj_byte(orp, 0);
obj_emit2(orp);
obj_byte(orp, 0x40);
obj_byte(orp, dTYPEDEF);
obj_word(orp, 0x1d); /* type # for linking */
obj_word(orp, 0); /* size of type */
obj_byte(orp, 0x23); /* absolute type for debugging */
obj_byte(orp, 0);
obj_byte(orp, 1);
obj_byte(orp, 0);
obj_emit2(orp);
obj_byte(orp, 0x40);
obj_byte(orp, dTYPEDEF);
obj_word(orp, 0x1e); /* type # for linking */
obj_word(orp, 0); /* size of type */
obj_byte(orp, 0x23); /* absolute type for debugging */
obj_byte(orp, 0);
obj_byte(orp, 5);
obj_byte(orp, 0);
obj_emit2(orp);
/* put out the array types */
for (i = ARRAYBOT; i < arrindex; i++) {
obj_byte(orp, 0x40);
obj_byte(orp, dTYPEDEF);
obj_word(orp, i); /* type # for linking */
obj_word(orp, arrtmp->size); /* size of type */
obj_byte(orp, 0x1A); /* absolute type for debugging (array) */
obj_byte(orp, arrtmp->basetype); /* base type */
obj_emit2(orp);
arrtmp = arrtmp->next;
}
}
/*
* write out line number info with a LINNUM record
* switch records when we switch segments, and output the
* file in a pseudo-TASM fashion. The record switch is naive; that
* is that one file may have many records for the same segment
* if there are lots of segment switches
*/
if (fnhead && debuginfo) {
seg = fnhead->lnhead->segment;
for (fn = fnhead; fn; fn = fn->next) {
/* write out current file name */
orp->type = COMENT;
orp->ori = ori_null;
obj_byte(orp, 0x40);
obj_byte(orp, dFILNAME);
obj_byte(orp, 0);
obj_name(orp, fn->name);
obj_dword(orp, 0);
obj_emit2(orp);
/* write out line numbers this file */
orp->type = LINNUM;
orp->ori = ori_linnum;
for (ln = fn->lnhead; ln; ln = ln->next) {
if (seg != ln->segment) {
/* if we get here have to flush the buffer and start
* a new record for a new segment
*/
seg = ln->segment;
obj_emit(orp);
}
orp->parm[0] = seg->grp ? seg->grp->obj_index : 0;
orp->parm[1] = seg->obj_index;
orp = obj_word(orp, ln->lineno);
orp = obj_x(orp, ln->offset);
obj_commit(orp);
}
obj_emit(orp);
}
}
/*
* we are going to locate the entry point segment now
* rather than wait until the MODEND record, because,
* then we can output a special symbol to tell where the
* entry point is.
*
*/
if (obj_entry_seg != NO_SEG) {
for (seg = seghead; seg; seg = seg->next) {
if (seg->index == obj_entry_seg) {
entry_seg_ptr = seg;
break;
}
}
if (!seg)
error(ERR_NONFATAL, "entry point is not in this module");
}
/*
* get ready to put out symbol records
*/
orp->type = COMENT;
orp->ori = ori_local;
/*
* put out a symbol for the entry point
* no dots in this symbol, because, borland does
* not (officially) support dots in label names
* and I don't know what various versions of TLINK will do
*/
if (debuginfo && obj_entry_seg != NO_SEG) {
orp = obj_name(orp, "start_of_program");
orp = obj_word(orp, 0x19); /* type: near label */
orp = obj_index(orp, seg->grp ? seg->grp->obj_index : 0);
orp = obj_index(orp, seg->obj_index);
orp = obj_x(orp, obj_entry_ofs);
obj_commit(orp);
}
/*
* put out the local labels
*/
for (seg = seghead; seg && debuginfo; seg = seg->next) {
/* labels this seg */
for (loc = seg->lochead; loc; loc = loc->next) {
orp = obj_name(orp, loc->name);
orp = obj_word(orp, loc->type);
orp = obj_index(orp, seg->grp ? seg->grp->obj_index : 0);
orp = obj_index(orp, seg->obj_index);
orp = obj_x(orp, loc->offset);
obj_commit(orp);
}
}
if (orp->used)
obj_emit(orp);
/*
* Write the LEDATA/FIXUPP pairs.
*/
for (seg = seghead; seg; seg = seg->next) {
obj_emit(seg->orp);
nasm_free(seg->orp);
}
/*
* Write the MODEND module end marker.
*/
orp->type = obj_use32 ? MODE32 : MODEND;
orp->ori = ori_null;
if (entry_seg_ptr) {
orp->type = entry_seg_ptr->use32 ? MODE32 : MODEND;
obj_byte(orp, 0xC1);
seg = entry_seg_ptr;
if (seg->grp) {
obj_byte(orp, 0x10);
obj_index(orp, seg->grp->obj_index);
} else {
/*
* the below changed to prevent TLINK crashing.
* Previous more efficient version read:
*
* obj_byte (orp, 0x50);
*/
obj_byte(orp, 0x00);
obj_index(orp, seg->obj_index);
}
obj_index(orp, seg->obj_index);
obj_x(orp, obj_entry_ofs);
} else
obj_byte(orp, 0);
obj_emit2(orp);
nasm_free(orp);
}
static void obj_fwrite(ObjRecord * orp)
{
unsigned int cksum, len;
uint8_t *ptr;
cksum = orp->type;
if (orp->x_size == 32)
cksum |= 1;
fputc(cksum, ofp);
len = orp->committed + 1;
cksum += (len & 0xFF) + ((len >> 8) & 0xFF);
fwriteint16_t(len, ofp);
fwrite(orp->buf, 1, len - 1, ofp);
for (ptr = orp->buf; --len; ptr++)
cksum += *ptr;
fputc((-cksum) & 0xFF, ofp);
}
extern macros_t obj_stdmac[];
void dbgbi_init(struct ofmt *of, void *id, FILE * fp, efunc error)
{
(void)of;
(void)id;
(void)fp;
(void)error;
fnhead = NULL;
fntail = &fnhead;
arrindex = ARRAYBOT;
arrhead = NULL;
arrtail = &arrhead;
}
static void dbgbi_cleanup(void)
{
struct Segment *segtmp;
while (fnhead) {
struct FileName *fntemp = fnhead;
while (fnhead->lnhead) {
struct LineNumber *lntemp = fnhead->lnhead;
fnhead->lnhead = lntemp->next;
nasm_free(lntemp);
}
fnhead = fnhead->next;
nasm_free(fntemp->name);
nasm_free(fntemp);
}
for (segtmp = seghead; segtmp; segtmp = segtmp->next) {
while (segtmp->lochead) {
struct Public *loctmp = segtmp->lochead;
segtmp->lochead = loctmp->next;
nasm_free(loctmp->name);
nasm_free(loctmp);
}
}
while (arrhead) {
struct Array *arrtmp = arrhead;
arrhead = arrhead->next;
nasm_free(arrtmp);
}
}
static void dbgbi_linnum(const char *lnfname, int32_t lineno, int32_t segto)
{
struct FileName *fn;
struct LineNumber *ln;
struct Segment *seg;
if (segto == NO_SEG)
return;
/*
* If `any_segs' is still false, we must define a default
* segment.
*/
if (!any_segs) {
int tempint; /* ignored */
if (segto != obj_segment("__NASMDEFSEG", 2, &tempint))
error(ERR_PANIC, "strange segment conditions in OBJ driver");
}
/*
* Find the segment we are targetting.
*/
for (seg = seghead; seg; seg = seg->next)
if (seg->index == segto)
break;
if (!seg)
error(ERR_PANIC, "lineno directed to nonexistent segment?");
/* for (fn = fnhead; fn; fn = fnhead->next) */
for (fn = fnhead; fn; fn = fn->next) /* fbk - Austin Lunnen - John Fine */
if (!nasm_stricmp(lnfname, fn->name))
break;
if (!fn) {
fn = nasm_malloc(sizeof(*fn));
fn->name = nasm_malloc(strlen(lnfname) + 1);
strcpy(fn->name, lnfname);
fn->lnhead = NULL;
fn->lntail = &fn->lnhead;
fn->next = NULL;
*fntail = fn;
fntail = &fn->next;
}
ln = nasm_malloc(sizeof(*ln));
ln->segment = seg;
ln->offset = seg->currentpos;
ln->lineno = lineno;
ln->next = NULL;
*fn->lntail = ln;
fn->lntail = &ln->next;
}
static void dbgbi_deflabel(char *name, int32_t segment,
int64_t offset, int is_global, char *special)
{
struct Segment *seg;
(void)special;
/*
* If it's a special-retry from pass two, discard it.
*/
if (is_global == 3)
return;
/*
* First check for the double-period, signifying something
* unusual.
*/
if (name[0] == '.' && name[1] == '.' && name[2] != '@') {
return;
}
/*
* Case (i):
*/
if (obj_seg_needs_update) {
return;
} else if (obj_grp_needs_update) {
return;
}
if (segment < SEG_ABS && segment != NO_SEG && segment % 2)
return;
if (segment >= SEG_ABS || segment == NO_SEG) {
return;
}
/*
* If `any_segs' is still false, we might need to define a
* default segment, if they're trying to declare a label in
* `first_seg'. But the label should exist due to a prior
* call to obj_deflabel so we can skip that.
*/
for (seg = seghead; seg; seg = seg->next)
if (seg->index == segment) {
struct Public *loc = nasm_malloc(sizeof(*loc));
/*
* Case (ii). Maybe MODPUB someday?
*/
last_defined = *seg->loctail = loc;
seg->loctail = &loc->next;
loc->next = NULL;
loc->name = nasm_strdup(name);
loc->offset = offset;
}
}
static void dbgbi_typevalue(int32_t type)
{
int vsize;
int elem = TYM_ELEMENTS(type);
type = TYM_TYPE(type);
if (!last_defined)
return;
switch (type) {
case TY_BYTE:
last_defined->type = 8; /* uint8_t */
vsize = 1;
break;
case TY_WORD:
last_defined->type = 10; /* unsigned word */
vsize = 2;
break;
case TY_DWORD:
last_defined->type = 12; /* unsigned dword */
vsize = 4;
break;
case TY_FLOAT:
last_defined->type = 14; /* float */
vsize = 4;
break;
case TY_QWORD:
last_defined->type = 15; /* qword */
vsize = 8;
break;
case TY_TBYTE:
last_defined->type = 16; /* TBYTE */
vsize = 10;
break;
default:
last_defined->type = 0x19; /*label */
vsize = 0;
break;
}
if (elem > 1) {
struct Array *arrtmp = nasm_malloc(sizeof(*arrtmp));
int vtype = last_defined->type;
arrtmp->size = vsize * elem;
arrtmp->basetype = vtype;
arrtmp->next = NULL;
last_defined->type = arrindex++;
*arrtail = arrtmp;
arrtail = &(arrtmp->next);
}
last_defined = NULL;
}
static void dbgbi_output(int output_type, void *param)
{
(void)output_type;
(void)param;
}
static struct dfmt borland_debug_form = {
"Borland Debug Records",
"borland",
dbgbi_init,
dbgbi_linnum,
dbgbi_deflabel,
null_debug_routine,
dbgbi_typevalue,
dbgbi_output,
dbgbi_cleanup,
};
static struct dfmt *borland_debug_arr[3] = {
&borland_debug_form,
&null_debug_form,
NULL
};
struct ofmt of_obj = {
"MS-DOS 16-bit/32-bit OMF object files",
"obj",
NULL,
borland_debug_arr,
&null_debug_form,
obj_stdmac,
obj_init,
obj_set_info,
obj_out,
obj_deflabel,
obj_segment,
obj_segbase,
obj_directive,
obj_filename,
obj_cleanup
};
#endif /* OF_OBJ */
Reference in New Issue View Git Blame Copy Permalink