output: aout -- Use nasm_x_space helpers
[nasm.git] / output / outmac64.c
blobd4ddeaa66bde118bb6c304b33fc78c4722887c82
1 /* ----------------------------------------------------------------------- *
2 *
3 * Copyright 1996-2013 The NASM Authors - All Rights Reserved
4 * See the file AUTHORS included with the NASM distribution for
5 * the specific copyright holders.
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following
9 * conditions are met:
11 * * Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer.
13 * * Redistributions in binary form must reproduce the above
14 * copyright notice, this list of conditions and the following
15 * disclaimer in the documentation and/or other materials provided
16 * with the distribution.
18 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
19 * CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES,
20 * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
21 * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
22 * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
23 * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
24 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
25 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
26 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
28 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
29 * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
30 * EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
32 * ----------------------------------------------------------------------- */
35 * outmac64.c output routines for the Netwide Assembler to produce
36 * NeXTstep/OpenStep/Rhapsody/Darwin/MacOS X (x86_64) object files
39 /* Most of this file is, like Mach-O itself, based on a.out. For more
40 * guidelines see outaout.c. */
42 #include "compiler.h"
44 #include <stdio.h>
45 #include <stdlib.h>
46 #include <string.h>
47 #include <ctype.h>
48 #include <inttypes.h>
50 #include "nasm.h"
51 #include "nasmlib.h"
52 #include "saa.h"
53 #include "raa.h"
54 #include "output/outform.h"
55 #include "output/outlib.h"
57 #if defined(OF_MACHO64)
59 /* Mach-O in-file header structure sizes */
60 #define MACHO_HEADER64_SIZE (32)
61 #define MACHO_SEGCMD64_SIZE (72)
62 #define MACHO_SECTCMD64_SIZE (80)
63 #define MACHO_SYMCMD_SIZE (24)
64 #define MACHO_NLIST64_SIZE (16)
65 #define MACHO_RELINFO64_SIZE (8)
67 /* Mach-O file header values */
68 #define MH_MAGIC_64 (0xfeedfacf)
69 #define CPU_TYPE_X86_64 (0x01000007) /* x86-64 platform */
70 #define CPU_SUBTYPE_I386_ALL (3) /* all-x86 compatible */
71 #define MH_OBJECT (0x1) /* object file */
73 #define LC_SEGMENT_64 (0x19) /* segment load command */
74 #define LC_SYMTAB (0x2) /* symbol table load command */
76 #define VM_PROT_NONE (0x00)
77 #define VM_PROT_READ (0x01)
78 #define VM_PROT_WRITE (0x02)
79 #define VM_PROT_EXECUTE (0x04)
81 #define VM_PROT_DEFAULT (VM_PROT_READ | VM_PROT_WRITE | VM_PROT_EXECUTE)
82 #define VM_PROT_ALL (VM_PROT_READ | VM_PROT_WRITE | VM_PROT_EXECUTE)
84 struct section {
85 /* nasm internal data */
86 struct section *next;
87 struct SAA *data;
88 int32_t index;
89 struct reloc *relocs;
90 int align;
92 /* data that goes into the file */
93 char sectname[16]; /* what this section is called */
94 char segname[16]; /* segment this section will be in */
95 uint64_t addr; /* in-memory address (subject to alignment) */
96 uint64_t size; /* in-memory and -file size */
97 uint32_t nreloc; /* relocation entry count */
98 uint32_t flags; /* type and attributes (masked) */
99 uint32_t extreloc; /* external relocations */
102 #define SECTION_TYPE 0x000000ff /* section type mask */
104 #define S_REGULAR (0x0) /* standard section */
105 #define S_ZEROFILL (0x1) /* zerofill, in-memory only */
107 #define SECTION_ATTRIBUTES_SYS 0x00ffff00 /* system setable attributes */
108 #define S_ATTR_SOME_INSTRUCTIONS 0x00000400 /* section contains some
109 machine instructions */
110 #define S_ATTR_EXT_RELOC 0x00000200 /* section has external
111 relocation entries */
112 #define S_ATTR_LOC_RELOC 0x00000100 /* section has local
113 relocation entries */
114 #define S_ATTR_PURE_INSTRUCTIONS 0x80000000 /* section uses pure
115 machine instructions */
117 static struct sectmap {
118 const char *nasmsect;
119 const char *segname;
120 const char *sectname;
121 const int32_t flags;
122 } sectmap[] = {
123 {".text", "__TEXT", "__text", S_REGULAR|S_ATTR_SOME_INSTRUCTIONS|S_ATTR_PURE_INSTRUCTIONS},
124 {".data", "__DATA", "__data", S_REGULAR},
125 {".rodata", "__DATA", "__const", S_REGULAR},
126 {".bss", "__DATA", "__bss", S_ZEROFILL},
127 {NULL, NULL, NULL, 0}
130 struct reloc {
131 /* nasm internal data */
132 struct reloc *next;
134 /* data that goes into the file */
135 int32_t addr; /* op's offset in section */
136 uint32_t snum:24, /* contains symbol index if
137 ** ext otherwise in-file
138 ** section number */
139 pcrel:1, /* relative relocation */
140 length:2, /* 0=byte, 1=word, 2=int32_t, 3=int64_t */
141 ext:1, /* external symbol referenced */
142 type:4; /* reloc type */
145 #define R_ABS 0 /* absolute relocation */
146 #define R_SCATTERED 0x80000000 /* reloc entry is scattered if
147 ** highest bit == 1 */
149 struct symbol {
150 /* nasm internal data */
151 struct symbol *next; /* next symbol in the list */
152 char *name; /* name of this symbol */
153 int32_t initial_snum; /* symbol number used above in
154 reloc */
155 int32_t snum; /* true snum for reloc */
157 /* data that goes into the file */
158 uint32_t strx; /* string table index */
159 uint8_t type; /* symbol type */
160 uint8_t sect; /* NO_SECT or section number */
161 uint16_t desc; /* for stab debugging, 0 for us */
162 uint64_t value; /* offset of symbol in section */
165 /* symbol type bits */
166 #define N_EXT 0x01 /* global or external symbol */
168 #define N_UNDF 0x0 /* undefined symbol | n_sect == */
169 #define N_ABS 0x2 /* absolute symbol | NO_SECT */
170 #define N_SECT 0xe /* defined symbol, n_sect holds
171 ** section number */
173 #define N_TYPE 0x0e /* type bit mask */
175 #define DEFAULT_SECTION_ALIGNMENT 0 /* byte (i.e. no) alignment */
177 /* special section number values */
178 #define NO_SECT 0 /* no section, invalid */
179 #define MAX_SECT 255 /* maximum number of sections */
181 static struct section *sects, **sectstail;
182 static struct symbol *syms, **symstail;
183 static uint32_t nsyms;
185 /* These variables are set by macho_layout_symbols() to organize
186 the symbol table and string table in order the dynamic linker
187 expects. They are then used in macho_write() to put out the
188 symbols and strings in that order.
190 The order of the symbol table is:
191 local symbols
192 defined external symbols (sorted by name)
193 undefined external symbols (sorted by name)
195 The order of the string table is:
196 strings for external symbols
197 strings for local symbols
199 static uint32_t ilocalsym = 0;
200 static uint32_t iextdefsym = 0;
201 static uint32_t iundefsym = 0;
202 static uint32_t nlocalsym;
203 static uint32_t nextdefsym;
204 static uint32_t nundefsym;
205 static struct symbol **extdefsyms = NULL;
206 static struct symbol **undefsyms = NULL;
208 static struct RAA *extsyms;
209 static struct SAA *strs;
210 static uint32_t strslen;
212 extern struct ofmt of_macho64;
214 /* Global file information. This should be cleaned up into either
215 a structure or as function arguments. */
216 uint32_t head_ncmds64 = 0;
217 uint32_t head_sizeofcmds64 = 0;
218 uint64_t seg_filesize64 = 0;
219 uint64_t seg_vmsize64 = 0;
220 uint32_t seg_nsects64 = 0;
221 uint64_t rel_padcnt64 = 0;
224 #define xstrncpy(xdst, xsrc) \
225 memset(xdst, '\0', sizeof(xdst)); /* zero out whole buffer */ \
226 strncpy(xdst, xsrc, sizeof(xdst)); /* copy over string */ \
227 xdst[sizeof(xdst) - 1] = '\0'; /* proper null-termination */
229 #define alignint32_t(x) \
230 ALIGN(x, sizeof(int32_t)) /* align x to int32_t boundary */
232 #define alignint64_t(x) \
233 ALIGN(x, sizeof(int64_t)) /* align x to int64_t boundary */
235 static void debug_reloc (struct reloc *);
236 static void debug_section_relocs (struct section *) _unused;
238 static struct section *get_section_by_name(const char *segname,
239 const char *sectname)
241 struct section *s;
243 for (s = sects; s != NULL; s = s->next)
244 if (!strcmp(s->segname, segname) && !strcmp(s->sectname, sectname))
245 break;
247 return s;
250 static struct section *get_section_by_index(const int32_t index)
252 struct section *s;
254 for (s = sects; s != NULL; s = s->next)
255 if (index == s->index)
256 break;
258 return s;
261 static int32_t get_section_index_by_name(const char *segname,
262 const char *sectname)
264 struct section *s;
266 for (s = sects; s != NULL; s = s->next)
267 if (!strcmp(s->segname, segname) && !strcmp(s->sectname, sectname))
268 return s->index;
270 return -1;
273 static char *get_section_name_by_index(const int32_t index)
275 struct section *s;
277 for (s = sects; s != NULL; s = s->next)
278 if (index == s->index)
279 return s->sectname;
281 return NULL;
284 static uint8_t get_section_fileindex_by_index(const int32_t index)
286 struct section *s;
287 uint8_t i = 1;
289 for (s = sects; s != NULL && i < MAX_SECT; s = s->next, ++i)
290 if (index == s->index)
291 return i;
293 if (i == MAX_SECT)
294 nasm_error(ERR_WARNING,
295 "too many sections (>255) - clipped by fileindex");
297 return NO_SECT;
300 static struct symbol *get_closest_section_symbol_by_offset(uint8_t fileindex, int64_t offset)
302 struct symbol *nearest = NULL;
303 struct symbol *sym;
305 for (sym = syms; sym; sym = sym->next) {
306 if ((sym->sect != NO_SECT) && (sym->sect == fileindex)) {
307 if ((int64_t)sym->value > offset)
308 break;
309 nearest = sym;
313 if (!nearest)
314 nasm_error(ERR_FATAL, "No section for index %x offset %llx found\n",
315 fileindex, (long long)offset);
317 return nearest;
321 * Special section numbers which are used to define Mach-O special
322 * symbols, which can be used with WRT to provide PIC relocation
323 * types.
325 static int32_t macho_gotpcrel_sect;
327 static void macho_init(void)
329 char zero = 0;
331 maxbits = 64;
333 sects = NULL;
334 sectstail = &sects;
336 syms = NULL;
337 symstail = &syms;
338 nsyms = 0;
339 nlocalsym = 0;
340 nextdefsym = 0;
341 nundefsym = 0;
343 extsyms = raa_init();
344 strs = saa_init(1L);
346 /* string table starts with a zero byte - don't ask why */
347 saa_wbytes(strs, &zero, sizeof(char));
348 strslen = 1;
350 /* add special symbol for ..gotpcrel */
351 macho_gotpcrel_sect = seg_alloc();
352 macho_gotpcrel_sect++;
353 define_label("..gotpcrel", macho_gotpcrel_sect, 0L, NULL, false, false);
356 static void sect_write(struct section *sect,
357 const uint8_t *data, uint32_t len)
359 saa_wbytes(sect->data, data, len);
360 sect->size += len;
363 static int32_t add_reloc(struct section *sect, int32_t section,
364 int pcrel, int bytes, int64_t reloff)
366 struct reloc *r;
367 struct symbol *sym;
368 int32_t fi;
369 int32_t adjustment = 0;
371 /* NeXT as puts relocs in reversed order (address-wise) into the
372 ** files, so we do the same, doesn't seem to make much of a
373 ** difference either way */
374 r = nasm_malloc(sizeof(struct reloc));
375 r->next = sect->relocs;
376 sect->relocs = r;
378 /* the current end of the section will be the symbol's address for
379 ** now, might have to be fixed by macho_fixup_relocs() later on. make
380 ** sure we don't make the symbol scattered by setting the highest
381 ** bit by accident */
382 r->addr = sect->size & ~R_SCATTERED;
383 r->ext = 1;
384 r->pcrel = (pcrel ? 1 : 0);
386 /* match byte count 1, 2, 4, 8 to length codes 0, 1, 2, 3 respectively */
387 switch(bytes){
388 case 1:
389 r->length = 0;
390 break;
391 case 2:
392 r->length = 1;
393 break;
394 case 4:
395 r->length = 2;
396 break;
397 case 8:
398 r->length = 3;
399 break;
400 default:
401 break;
404 /* set default relocation values */
405 r->type = 0; // X86_64_RELOC_UNSIGNED
406 r->snum = R_ABS; // Absolute Symbol (indicates no relocation)
408 /* absolute relocation */
409 if (pcrel == 0) {
411 /* intra-section */
412 if (section == NO_SEG) {
413 // r->snum = R_ABS; // Set above
415 /* inter-section */
416 } else {
417 fi = get_section_fileindex_by_index(section);
419 /* external */
420 if (fi == NO_SECT) {
421 r->snum = raa_read(extsyms, section);
423 /* local */
424 } else {
425 sym = get_closest_section_symbol_by_offset(fi, reloff);
426 r->snum = sym->initial_snum;
427 adjustment = sym->value;
431 /* relative relocation */
432 } else if (pcrel == 1) {
434 /* intra-section */
435 if (section == NO_SEG) {
436 r->type = 1; // X86_64_RELOC_SIGNED
438 /* inter-section */
439 } else {
440 r->type = 2; // X86_64_RELOC_BRANCH
441 fi = get_section_fileindex_by_index(section);
443 /* external */
444 if (fi == NO_SECT) {
445 sect->extreloc = 1;
446 r->snum = raa_read(extsyms, section);
448 /* local */
449 } else {
450 sym = get_closest_section_symbol_by_offset(fi, reloff);
451 r->snum = sym->initial_snum;
452 adjustment = sym->value;
456 /* subtractor */
457 } else if (pcrel == 2) {
458 r->pcrel = 0;
459 r->type = 5; // X86_64_RELOC_SUBTRACTOR
461 /* gotpcrel */
462 } else if (pcrel == 3) {
463 r->type = 4; // X86_64_RELOC_GOT
464 r->snum = macho_gotpcrel_sect;
466 /* gotpcrel MOVQ load */
467 } else if (pcrel == 4) {
468 r->type = 3; // X86_64_RELOC_GOT_LOAD
469 r->snum = macho_gotpcrel_sect;
472 ++sect->nreloc;
474 return adjustment;
477 static void macho_output(int32_t secto, const void *data,
478 enum out_type type, uint64_t size,
479 int32_t section, int32_t wrt)
481 struct section *s, *sbss;
482 int64_t addr;
483 uint8_t mydata[16], *p, gotload;
485 if (secto == NO_SEG) {
486 if (type != OUT_RESERVE)
487 nasm_error(ERR_NONFATAL, "attempt to assemble code in "
488 "[ABSOLUTE] space");
490 return;
493 s = get_section_by_index(secto);
495 if (s == NULL) {
496 nasm_error(ERR_WARNING, "attempt to assemble code in"
497 " section %d: defaulting to `.text'", secto);
498 s = get_section_by_name("__TEXT", "__text");
500 /* should never happen */
501 if (s == NULL)
502 nasm_error(ERR_PANIC, "text section not found");
505 sbss = get_section_by_name("__DATA", "__bss");
507 if (s == sbss && type != OUT_RESERVE) {
508 nasm_error(ERR_WARNING, "attempt to initialize memory in the"
509 " BSS section: ignored");
510 s->size += realsize(type, size);
511 return;
514 memset(mydata, 0, sizeof(mydata));
516 switch (type) {
517 case OUT_RESERVE:
518 if (s != sbss) {
519 nasm_error(ERR_WARNING, "uninitialized space declared in"
520 " %s section: zeroing",
521 get_section_name_by_index(secto));
523 sect_write(s, NULL, size);
524 } else
525 s->size += size;
527 break;
529 case OUT_RAWDATA:
530 if (section != NO_SEG)
531 nasm_error(ERR_PANIC, "OUT_RAWDATA with other than NO_SEG");
533 sect_write(s, data, size);
534 break;
536 case OUT_ADDRESS:
538 int asize = abs(size);
540 addr = *(int64_t *)data;
541 if (section != NO_SEG) {
542 if (section % 2) {
543 nasm_error(ERR_NONFATAL, "Mach-O format does not support"
544 " section base references");
545 } else {
546 if (wrt == NO_SEG) {
547 if (asize < 8) {
548 nasm_error(ERR_NONFATAL, "Mach-O 64-bit format does not support"
549 " 32-bit absolute addresses");
551 Seemingly, Mach-O's X86_64_RELOC_SUBTRACTOR would require
552 pre-determined knowledge of where the image base would be,
553 making it impractical for use in intermediate object files
555 } else {
556 addr -= add_reloc(s, section, 0, asize, addr); // X86_64_RELOC_UNSIGNED
558 } else {
559 nasm_error(ERR_NONFATAL, "Mach-O format does not support"
560 " this use of WRT");
565 p = mydata;
566 WRITEADDR(p, addr, asize);
567 sect_write(s, mydata, asize);
568 break;
571 case OUT_REL2ADR:
572 p = mydata;
573 WRITESHORT(p, *(int64_t *)data);
575 if (section == secto)
576 nasm_error(ERR_PANIC, "intra-section OUT_REL2ADR");
578 if (section == NO_SEG) {
579 /* Do nothing */
580 } else if (section % 2) {
581 nasm_error(ERR_NONFATAL, "Mach-O format does not support"
582 " section base references");
583 } else {
584 nasm_error(ERR_NONFATAL, "Unsupported non-32-bit"
585 " Macho-O relocation [2]");
588 sect_write(s, mydata, 2L);
589 break;
591 case OUT_REL4ADR:
592 p = mydata;
593 WRITELONG(p, *(int64_t *)data + 4 - size);
595 if (section == secto)
596 nasm_error(ERR_PANIC, "intra-section OUT_REL4ADR");
598 if (section != NO_SEG && section % 2) {
599 nasm_error(ERR_NONFATAL, "Mach-O format does not support"
600 " section base references");
601 } else {
602 if (wrt == NO_SEG) {
603 *(int64_t *)mydata -= add_reloc(s, section, 1, 4, *(int64_t *)mydata); // X86_64_RELOC_SIGNED/BRANCH
604 } else if (wrt == macho_gotpcrel_sect) {
605 if (s->data->datalen > 1) {
606 saa_fread(s->data, s->data->datalen-2, &gotload, 1); // Retrieve Instruction Opcode
607 } else {
608 gotload = 0;
610 if (gotload == 0x8B) { // Check for MOVQ Opcode
611 *(int64_t *)mydata -= add_reloc(s, section, 4, 4, *(int64_t *)mydata); // X86_64_GOT_LOAD (MOVQ load)
612 } else {
613 *(int64_t *)mydata -= add_reloc(s, section, 3, 4, *(int64_t *)mydata); // X86_64_GOT
615 } else {
616 nasm_error(ERR_NONFATAL, "Mach-O format does not support"
617 " this use of WRT");
618 wrt = NO_SEG; /* we can at least _try_ to continue */
622 sect_write(s, mydata, 4L);
623 break;
625 default:
626 nasm_error(ERR_PANIC, "unknown output type?");
627 break;
631 static int32_t macho_section(char *name, int pass, int *bits)
633 int32_t index, originalIndex;
634 char *sectionAttributes;
635 struct sectmap *sm;
636 struct section *s;
638 (void)pass;
640 /* Default to 64 bits. */
641 if (!name) {
642 *bits = 64;
643 name = ".text";
644 sectionAttributes = NULL;
645 } else {
646 sectionAttributes = name;
647 name = nasm_strsep(&sectionAttributes, " \t");
650 for (sm = sectmap; sm->nasmsect != NULL; ++sm) {
651 /* make lookup into section name translation table */
652 if (!strcmp(name, sm->nasmsect)) {
653 char *currentAttribute;
655 /* try to find section with that name */
656 originalIndex = index = get_section_index_by_name(sm->segname,
657 sm->sectname);
659 /* create it if it doesn't exist yet */
660 if (index == -1) {
661 s = *sectstail = nasm_malloc(sizeof(struct section));
662 s->next = NULL;
663 sectstail = &s->next;
665 s->data = saa_init(1L);
666 s->index = seg_alloc();
667 s->relocs = NULL;
668 s->align = -1;
670 xstrncpy(s->segname, sm->segname);
671 xstrncpy(s->sectname, sm->sectname);
672 s->size = 0;
673 s->nreloc = 0;
674 s->flags = sm->flags;
676 index = s->index;
677 } else {
678 s = get_section_by_index(index);
681 while ((NULL != sectionAttributes)
682 && (currentAttribute = nasm_strsep(&sectionAttributes, " \t"))) {
683 if (0 != *currentAttribute) {
684 if (!nasm_strnicmp("align=", currentAttribute, 6)) {
685 char *end;
686 int newAlignment, value;
688 value = strtoul(currentAttribute + 6, (char**)&end, 0);
689 newAlignment = alignlog2_32(value);
691 if (0 != *end) {
692 nasm_error(ERR_PANIC,
693 "unknown or missing alignment value \"%s\" "
694 "specified for section \"%s\"",
695 currentAttribute + 6,
696 name);
697 return NO_SEG;
698 } else if (0 > newAlignment) {
699 nasm_error(ERR_PANIC,
700 "alignment of %d (for section \"%s\") is not "
701 "a power of two",
702 value,
703 name);
704 return NO_SEG;
707 if ((-1 != originalIndex)
708 && (s->align != newAlignment)
709 && (s->align != -1)) {
710 nasm_error(ERR_PANIC,
711 "section \"%s\" has already been specified "
712 "with alignment %d, conflicts with new "
713 "alignment of %d",
714 name,
715 (1 << s->align),
716 value);
717 return NO_SEG;
720 s->align = newAlignment;
721 } else if (!nasm_stricmp("data", currentAttribute)) {
722 /* Do nothing; 'data' is implicit */
723 } else {
724 nasm_error(ERR_PANIC,
725 "unknown section attribute %s for section %s",
726 currentAttribute,
727 name);
728 return NO_SEG;
733 return index;
737 nasm_error(ERR_PANIC, "invalid section name %s", name);
738 return NO_SEG;
741 static void macho_symdef(char *name, int32_t section, int64_t offset,
742 int is_global, char *special)
744 struct symbol *sym;
746 if (special) {
747 nasm_error(ERR_NONFATAL, "The Mach-O output format does "
748 "not support any special symbol types");
749 return;
752 if (is_global == 3) {
753 nasm_error(ERR_NONFATAL, "The Mach-O format does not "
754 "(yet) support forward reference fixups.");
755 return;
758 if (name[0] == '.' && name[1] == '.' && name[2] != '@') {
760 * This is a NASM special symbol. We never allow it into
761 * the Macho-O symbol table, even if it's a valid one. If it
762 * _isn't_ a valid one, we should barf immediately.
764 if (strcmp(name, "..gotpcrel"))
765 nasm_error(ERR_NONFATAL, "unrecognized special symbol `%s'", name);
766 return;
769 sym = *symstail = nasm_malloc(sizeof(struct symbol));
770 sym->next = NULL;
771 symstail = &sym->next;
773 sym->name = name;
774 sym->strx = strslen;
775 sym->type = 0;
776 sym->desc = 0;
777 sym->value = offset;
778 sym->initial_snum = -1;
780 /* external and common symbols get N_EXT */
781 if (is_global != 0) {
782 sym->type |= N_EXT;
785 if (section == NO_SEG) {
786 /* symbols in no section get absolute */
787 sym->type |= N_ABS;
788 sym->sect = NO_SECT;
789 } else {
790 sym->type |= N_SECT;
792 /* get the in-file index of the section the symbol was defined in */
793 sym->sect = get_section_fileindex_by_index(section);
795 /* track the initially allocated symbol number for use in future fix-ups */
796 sym->initial_snum = nsyms;
798 if (sym->sect == NO_SECT) {
800 /* remember symbol number of references to external
801 ** symbols, this works because every external symbol gets
802 ** its own section number allocated internally by nasm and
803 ** can so be used as a key */
804 extsyms = raa_write(extsyms, section, nsyms);
806 switch (is_global) {
807 case 1:
808 case 2:
809 /* there isn't actually a difference between global
810 ** and common symbols, both even have their size in
811 ** sym->value */
812 sym->type = N_EXT;
813 break;
815 default:
816 /* give an error on unfound section if it's not an
817 ** external or common symbol (assemble_file() does a
818 ** seg_alloc() on every call for them) */
819 nasm_error(ERR_PANIC, "in-file index for section %d not found",
820 section);
824 ++nsyms;
827 static void macho_sectalign(int32_t seg, unsigned int value)
829 struct section *s;
831 list_for_each(s, sects) {
832 if (s->index == seg)
833 break;
836 if (!s || !is_power2(value))
837 return;
839 value = alignlog2_32(value);
840 if (s->align < (int)value)
841 s->align = value;
844 static int32_t macho_segbase(int32_t section)
846 return section;
849 static void macho_filename(char *inname, char *outname)
851 standard_extension(inname, outname, ".o");
854 extern macros_t macho_stdmac[];
856 /* Comparison function for qsort symbol layout. */
857 static int layout_compare (const struct symbol **s1,
858 const struct symbol **s2)
860 return (strcmp ((*s1)->name, (*s2)->name));
863 /* The native assembler does a few things in a similar function
865 * Remove temporary labels
866 * Sort symbols according to local, external, undefined (by name)
867 * Order the string table
869 We do not remove temporary labels right now.
871 numsyms is the total number of symbols we have. strtabsize is the
872 number entries in the string table. */
874 static void macho_layout_symbols (uint32_t *numsyms,
875 uint32_t *strtabsize)
877 struct symbol *sym, **symp;
878 uint32_t i,j;
880 *numsyms = 0;
881 *strtabsize = sizeof (char);
883 symp = &syms;
885 while ((sym = *symp)) {
886 /* Undefined symbols are now external. */
887 if (sym->type == N_UNDF)
888 sym->type |= N_EXT;
890 if ((sym->type & N_EXT) == 0) {
891 sym->snum = *numsyms;
892 *numsyms = *numsyms + 1;
893 nlocalsym++;
895 else {
896 if ((sym->type & N_TYPE) != N_UNDF) {
897 nextdefsym++;
898 } else {
899 nundefsym++;
902 /* If we handle debug info we'll want
903 to check for it here instead of just
904 adding the symbol to the string table. */
905 sym->strx = *strtabsize;
906 saa_wbytes (strs, sym->name, (int32_t)(strlen(sym->name) + 1));
907 *strtabsize += strlen(sym->name) + 1;
909 symp = &(sym->next);
912 /* Next, sort the symbols. Most of this code is a direct translation from
913 the Apple cctools symbol layout. We need to keep compatibility with that. */
914 /* Set the indexes for symbol groups into the symbol table */
915 ilocalsym = 0;
916 iextdefsym = nlocalsym;
917 iundefsym = nlocalsym + nextdefsym;
919 /* allocate arrays for sorting externals by name */
920 extdefsyms = nasm_malloc(nextdefsym * sizeof(struct symbol *));
921 undefsyms = nasm_malloc(nundefsym * sizeof(struct symbol *));
923 i = 0;
924 j = 0;
926 symp = &syms;
928 while ((sym = *symp)) {
930 if((sym->type & N_EXT) == 0) {
931 sym->strx = *strtabsize;
932 saa_wbytes (strs, sym->name, (int32_t)(strlen (sym->name) + 1));
933 *strtabsize += strlen(sym->name) + 1;
935 else {
936 if((sym->type & N_TYPE) != N_UNDF) {
937 extdefsyms[i++] = sym;
938 } else {
939 undefsyms[j++] = sym;
942 symp = &(sym->next);
945 qsort(extdefsyms, nextdefsym, sizeof(struct symbol *),
946 (int (*)(const void *, const void *))layout_compare);
947 qsort(undefsyms, nundefsym, sizeof(struct symbol *),
948 (int (*)(const void *, const void *))layout_compare);
950 for(i = 0; i < nextdefsym; i++) {
951 extdefsyms[i]->snum = *numsyms;
952 *numsyms += 1;
954 for(j = 0; j < nundefsym; j++) {
955 undefsyms[j]->snum = *numsyms;
956 *numsyms += 1;
960 /* Calculate some values we'll need for writing later. */
962 static void macho_calculate_sizes (void)
964 struct section *s;
966 /* count sections and calculate in-memory and in-file offsets */
967 for (s = sects; s != NULL; s = s->next) {
968 uint64_t pad = 0;
970 /* zerofill sections aren't actually written to the file */
971 if ((s->flags & SECTION_TYPE) != S_ZEROFILL)
972 seg_filesize64 += s->size;
974 /* recalculate segment address based on alignment and vm size */
975 s->addr = seg_vmsize64;
976 /* we need section alignment to calculate final section address */
977 if (s->align == -1)
978 s->align = DEFAULT_SECTION_ALIGNMENT;
979 if(s->align) {
980 uint64_t newaddr = ALIGN(s->addr, 1 << s->align);
981 pad = newaddr - s->addr;
982 s->addr = newaddr;
985 seg_vmsize64 += s->size + pad;
986 ++seg_nsects64;
989 /* calculate size of all headers, load commands and sections to
990 ** get a pointer to the start of all the raw data */
991 if (seg_nsects64 > 0) {
992 ++head_ncmds64;
993 head_sizeofcmds64 +=
994 MACHO_SEGCMD64_SIZE + seg_nsects64 * MACHO_SECTCMD64_SIZE;
997 if (nsyms > 0) {
998 ++head_ncmds64;
999 head_sizeofcmds64 += MACHO_SYMCMD_SIZE;
1003 /* Write out the header information for the file. */
1005 static void macho_write_header (void)
1007 fwriteint32_t(MH_MAGIC_64, ofile); /* magic */
1008 fwriteint32_t(CPU_TYPE_X86_64, ofile); /* CPU type */
1009 fwriteint32_t(CPU_SUBTYPE_I386_ALL, ofile); /* CPU subtype */
1010 fwriteint32_t(MH_OBJECT, ofile); /* Mach-O file type */
1011 fwriteint32_t(head_ncmds64, ofile); /* number of load commands */
1012 fwriteint32_t(head_sizeofcmds64, ofile); /* size of load commands */
1013 fwriteint32_t(0, ofile); /* no flags */
1014 fwriteint32_t(0, ofile); /* reserved for future use */
1017 /* Write out the segment load command at offset. */
1019 static uint32_t macho_write_segment (uint64_t offset)
1021 uint64_t rel_base = alignint64_t (offset + seg_filesize64);
1022 uint32_t s_reloff = 0;
1023 struct section *s;
1025 fwriteint32_t(LC_SEGMENT_64, ofile); /* cmd == LC_SEGMENT_64 */
1027 /* size of load command including section load commands */
1028 fwriteint32_t(MACHO_SEGCMD64_SIZE + seg_nsects64 *
1029 MACHO_SECTCMD64_SIZE, ofile);
1031 /* in an MH_OBJECT file all sections are in one unnamed (name
1032 ** all zeros) segment */
1033 fwritezero(16, ofile);
1034 fwriteint64_t(0, ofile); /* in-memory offset */
1035 fwriteint64_t(seg_vmsize64, ofile); /* in-memory size */
1036 fwriteint64_t(offset, ofile); /* in-file offset to data */
1037 fwriteint64_t(seg_filesize64, ofile); /* in-file size */
1038 fwriteint32_t(VM_PROT_DEFAULT, ofile); /* maximum vm protection */
1039 fwriteint32_t(VM_PROT_DEFAULT, ofile); /* initial vm protection */
1040 fwriteint32_t(seg_nsects64, ofile); /* number of sections */
1041 fwriteint32_t(0, ofile); /* no flags */
1043 /* emit section headers */
1044 for (s = sects; s != NULL; s = s->next) {
1045 nasm_write(s->sectname, sizeof(s->sectname), ofile);
1046 nasm_write(s->segname, sizeof(s->segname), ofile);
1047 fwriteint64_t(s->addr, ofile);
1048 fwriteint64_t(s->size, ofile);
1050 /* dummy data for zerofill sections or proper values */
1051 if ((s->flags & SECTION_TYPE) != S_ZEROFILL) {
1052 fwriteint32_t(offset, ofile);
1053 /* Write out section alignment, as a power of two.
1054 e.g. 32-bit word alignment would be 2 (2^2 = 4). */
1055 if (s->align == -1)
1056 s->align = DEFAULT_SECTION_ALIGNMENT;
1057 fwriteint32_t(s->align, ofile);
1058 /* To be compatible with cctools as we emit
1059 a zero reloff if we have no relocations. */
1060 fwriteint32_t(s->nreloc ? rel_base + s_reloff : 0, ofile);
1061 fwriteint32_t(s->nreloc, ofile);
1063 offset += s->size;
1064 s_reloff += s->nreloc * MACHO_RELINFO64_SIZE;
1065 } else {
1066 fwriteint32_t(0, ofile);
1067 fwriteint32_t(0, ofile);
1068 fwriteint32_t(0, ofile);
1069 fwriteint32_t(0, ofile);
1072 if (s->nreloc) {
1073 s->flags |= S_ATTR_LOC_RELOC;
1074 if (s->extreloc)
1075 s->flags |= S_ATTR_EXT_RELOC;
1078 fwriteint32_t(s->flags, ofile); /* flags */
1079 fwriteint32_t(0, ofile); /* reserved */
1080 fwriteint32_t(0, ofile); /* reserved */
1082 fwriteint32_t(0, ofile); /* align */
1085 rel_padcnt64 = rel_base - offset;
1086 offset = rel_base + s_reloff;
1088 return offset;
1091 /* For a given chain of relocs r, write out the entire relocation
1092 chain to the object file. */
1094 static void macho_write_relocs (struct reloc *r)
1096 while (r) {
1097 uint32_t word2;
1099 fwriteint32_t(r->addr, ofile); /* reloc offset */
1101 word2 = r->snum;
1102 word2 |= r->pcrel << 24;
1103 word2 |= r->length << 25;
1104 word2 |= r->ext << 27;
1105 word2 |= r->type << 28;
1106 fwriteint32_t(word2, ofile); /* reloc data */
1107 r = r->next;
1111 /* Write out the section data. */
1112 static void macho_write_section (void)
1114 struct section *s, *s2;
1115 struct reloc *r;
1116 uint8_t fi, *p, *q, blk[8];
1117 int32_t len;
1118 int64_t l;
1120 for (s = sects; s != NULL; s = s->next) {
1121 if ((s->flags & SECTION_TYPE) == S_ZEROFILL)
1122 continue;
1124 /* no padding needs to be done to the sections */
1126 /* Like a.out Mach-O references things in the data or bss
1127 * sections by addresses which are actually relative to the
1128 * start of the _text_ section, in the _file_. See outaout.c
1129 * for more information. */
1130 saa_rewind(s->data);
1131 for (r = s->relocs; r != NULL; r = r->next) {
1132 len = (int32_t)r->length << 1;
1133 if(len > 4) len = 8;
1134 saa_fread(s->data, r->addr, blk, len);
1135 p = q = blk;
1136 l = *p++;
1138 /* get offset based on relocation type */
1139 if (r->length > 0) {
1140 l += ((int64_t)*p++) << 8;
1142 if (r->length > 1) {
1143 l += ((int64_t)*p++) << 16;
1144 l += ((int64_t)*p++) << 24;
1147 if (r->length > 2) {
1148 l += ((int64_t)*p++) << 32;
1149 l += ((int64_t)*p++) << 40;
1150 l += ((int64_t)*p++) << 48;
1151 l += ((int64_t)*p++) << 56;
1157 /* If the relocation is internal add to the current section
1158 offset. Otherwise the only value we need is the symbol
1159 offset which we already have. The linker takes care
1160 of the rest of the address. */
1161 if (!r->ext) {
1162 /* generate final address by section address and offset */
1163 for (s2 = sects, fi = 1;
1164 s2 != NULL; s2 = s2->next, fi++) {
1165 if (fi == r->snum) {
1166 l += s2->addr;
1167 break;
1172 /* write new offset back */
1173 if (r->length == 3)
1174 WRITEDLONG(q, l);
1175 else if (r->length == 2)
1176 WRITELONG(q, l);
1177 else if (r->length == 1)
1178 WRITESHORT(q, l);
1179 else
1180 *q++ = l & 0xFF;
1182 saa_fwrite(s->data, r->addr, blk, len);
1185 /* dump the section data to file */
1186 saa_fpwrite(s->data, ofile);
1189 /* pad last section up to reloc entries on int64_t boundary */
1190 fwritezero(rel_padcnt64, ofile);
1192 /* emit relocation entries */
1193 for (s = sects; s != NULL; s = s->next)
1194 macho_write_relocs (s->relocs);
1197 /* Write out the symbol table. We should already have sorted this
1198 before now. */
1199 static void macho_write_symtab (void)
1201 struct symbol *sym;
1202 struct section *s;
1203 int64_t fi;
1204 uint64_t i;
1206 /* we don't need to pad here since MACHO_RELINFO_SIZE == 8 */
1208 for (sym = syms; sym != NULL; sym = sym->next) {
1209 if ((sym->type & N_EXT) == 0) {
1210 fwriteint32_t(sym->strx, ofile); /* string table entry number */
1211 nasm_write(&sym->type, 1, ofile); /* symbol type */
1212 nasm_write(&sym->sect, 1, ofile); /* section */
1213 fwriteint16_t(sym->desc, ofile); /* description */
1215 /* Fix up the symbol value now that we know the final section
1216 sizes. */
1217 if (((sym->type & N_TYPE) == N_SECT) && (sym->sect != NO_SECT)) {
1218 for (s = sects, fi = 1; s != NULL; s = s->next, fi++) {
1219 if (fi == sym->sect) {
1220 sym->value += s->addr;
1221 break;
1226 fwriteint64_t(sym->value, ofile); /* value (i.e. offset) */
1230 for (i = 0; i < nextdefsym; i++) {
1231 sym = extdefsyms[i];
1232 fwriteint32_t(sym->strx, ofile);
1233 nasm_write(&sym->type, 1, ofile); /* symbol type */
1234 nasm_write(&sym->sect, 1, ofile); /* section */
1235 fwriteint16_t(sym->desc, ofile); /* description */
1237 /* Fix up the symbol value now that we know the final section
1238 sizes. */
1239 if (((sym->type & N_TYPE) == N_SECT) && (sym->sect != NO_SECT)) {
1240 for (s = sects, fi = 1;
1241 s != NULL && fi < sym->sect; s = s->next, ++fi)
1242 sym->value += s->size;
1245 fwriteint64_t(sym->value, ofile); /* value (i.e. offset) */
1248 for (i = 0; i < nundefsym; i++) {
1249 sym = undefsyms[i];
1250 fwriteint32_t(sym->strx, ofile);
1251 nasm_write(&sym->type, 1, ofile); /* symbol type */
1252 nasm_write(&sym->sect, 1, ofile); /* section */
1253 fwriteint16_t(sym->desc, ofile); /* description */
1255 // Fix up the symbol value now that we know the final section sizes.
1256 if (((sym->type & N_TYPE) == N_SECT) && (sym->sect != NO_SECT)) {
1257 for (s = sects, fi = 1;
1258 s != NULL && fi < sym->sect; s = s->next, ++fi)
1259 sym->value += s->size;
1262 fwriteint64_t(sym->value, ofile); // value (i.e. offset)
1267 /* Fixup the snum in the relocation entries, we should be
1268 doing this only for externally referenced symbols. */
1269 static void macho_fixup_relocs (struct reloc *r)
1271 struct symbol *sym;
1273 while (r != NULL) {
1274 if (r->ext) {
1275 for (sym = syms; sym != NULL; sym = sym->next) {
1276 if (sym->initial_snum == r->snum) {
1277 r->snum = sym->snum;
1278 break;
1282 r = r->next;
1286 /* Write out the object file. */
1288 static void macho_write (void)
1290 uint64_t offset = 0;
1292 /* mach-o object file structure:
1294 ** mach header
1295 ** uint32_t magic
1296 ** int cpu type
1297 ** int cpu subtype
1298 ** uint32_t mach file type
1299 ** uint32_t number of load commands
1300 ** uint32_t size of all load commands
1301 ** (includes section struct size of segment command)
1302 ** uint32_t flags
1304 ** segment command
1305 ** uint32_t command type == LC_SEGMENT_64
1306 ** uint32_t size of load command
1307 ** (including section load commands)
1308 ** char[16] segment name
1309 ** uint64_t in-memory offset
1310 ** uint64_t in-memory size
1311 ** uint64_t in-file offset to data area
1312 ** uint64_t in-file size
1313 ** (in-memory size excluding zerofill sections)
1314 ** int maximum vm protection
1315 ** int initial vm protection
1316 ** uint32_t number of sections
1317 ** uint32_t flags
1319 ** section commands
1320 ** char[16] section name
1321 ** char[16] segment name
1322 ** uint64_t in-memory offset
1323 ** uint64_t in-memory size
1324 ** uint32_t in-file offset
1325 ** uint32_t alignment
1326 ** (irrelevant in MH_OBJECT)
1327 ** uint32_t in-file offset of relocation entires
1328 ** uint32_t number of relocations
1329 ** uint32_t flags
1330 ** uint32_t reserved
1331 ** uint32_t reserved
1333 ** symbol table command
1334 ** uint32_t command type == LC_SYMTAB
1335 ** uint32_t size of load command
1336 ** uint32_t symbol table offset
1337 ** uint32_t number of symbol table entries
1338 ** uint32_t string table offset
1339 ** uint32_t string table size
1341 ** raw section data
1343 ** padding to int64_t boundary
1345 ** relocation data (struct reloc)
1346 ** int32_t offset
1347 ** uint data (symbolnum, pcrel, length, extern, type)
1349 ** symbol table data (struct nlist)
1350 ** int32_t string table entry number
1351 ** uint8_t type
1352 ** (extern, absolute, defined in section)
1353 ** uint8_t section
1354 ** (0 for global symbols, section number of definition (>= 1, <=
1355 ** 254) for local symbols, size of variable for common symbols
1356 ** [type == extern])
1357 ** int16_t description
1358 ** (for stab debugging format)
1359 ** uint64_t value (i.e. file offset) of symbol or stab offset
1361 ** string table data
1362 ** list of null-terminated strings
1365 /* Emit the Mach-O header. */
1366 macho_write_header();
1368 offset = MACHO_HEADER64_SIZE + head_sizeofcmds64;
1370 /* emit the segment load command */
1371 if (seg_nsects64 > 0)
1372 offset = macho_write_segment (offset);
1373 else
1374 nasm_error(ERR_WARNING, "no sections?");
1376 if (nsyms > 0) {
1377 /* write out symbol command */
1378 fwriteint32_t(LC_SYMTAB, ofile); /* cmd == LC_SYMTAB */
1379 fwriteint32_t(MACHO_SYMCMD_SIZE, ofile); /* size of load command */
1380 fwriteint32_t(offset, ofile); /* symbol table offset */
1381 fwriteint32_t(nsyms, ofile); /* number of symbol
1382 ** table entries */
1384 offset += nsyms * MACHO_NLIST64_SIZE;
1385 fwriteint32_t(offset, ofile); /* string table offset */
1386 fwriteint32_t(strslen, ofile); /* string table size */
1389 /* emit section data */
1390 if (seg_nsects64 > 0)
1391 macho_write_section ();
1393 /* emit symbol table if we have symbols */
1394 if (nsyms > 0)
1395 macho_write_symtab ();
1397 /* we don't need to pad here since MACHO_NLIST64_SIZE == 16 */
1399 /* emit string table */
1400 saa_fpwrite(strs, ofile);
1402 /* We do quite a bit here, starting with finalizing all of the data
1403 for the object file, writing, and then freeing all of the data from
1404 the file. */
1406 static void macho_cleanup(int debuginfo)
1408 struct section *s;
1409 struct reloc *r;
1410 struct symbol *sym;
1412 (void)debuginfo;
1414 /* Sort all symbols. */
1415 macho_layout_symbols (&nsyms, &strslen);
1417 /* Fixup relocation entries */
1418 for (s = sects; s != NULL; s = s->next) {
1419 macho_fixup_relocs (s->relocs);
1422 /* First calculate and finalize needed values. */
1423 macho_calculate_sizes();
1424 macho_write();
1426 /* free up everything */
1427 while (sects->next) {
1428 s = sects;
1429 sects = sects->next;
1431 saa_free(s->data);
1432 while (s->relocs != NULL) {
1433 r = s->relocs;
1434 s->relocs = s->relocs->next;
1435 nasm_free(r);
1438 nasm_free(s);
1441 saa_free(strs);
1442 raa_free(extsyms);
1444 if (syms) {
1445 while (syms->next) {
1446 sym = syms;
1447 syms = syms->next;
1449 nasm_free (sym);
1454 /* Debugging routines. */
1455 static void debug_reloc (struct reloc *r)
1457 fprintf (stdout, "reloc:\n");
1458 fprintf (stdout, "\taddr: %"PRId32"\n", r->addr);
1459 fprintf (stdout, "\tsnum: %d\n", r->snum);
1460 fprintf (stdout, "\tpcrel: %d\n", r->pcrel);
1461 fprintf (stdout, "\tlength: %d\n", r->length);
1462 fprintf (stdout, "\text: %d\n", r->ext);
1463 fprintf (stdout, "\ttype: %d\n", r->type);
1466 static void debug_section_relocs (struct section *s)
1468 struct reloc *r = s->relocs;
1470 fprintf (stdout, "relocs for section %s:\n\n", s->sectname);
1472 while (r != NULL) {
1473 debug_reloc (r);
1474 r = r->next;
1478 struct ofmt of_macho64 = {
1479 "NeXTstep/OpenStep/Rhapsody/Darwin/MacOS X (x86_64) object files",
1480 "macho64",
1482 null_debug_arr,
1483 &null_debug_form,
1484 macho_stdmac,
1485 macho_init,
1486 null_setinfo,
1487 macho_output,
1488 macho_symdef,
1489 macho_section,
1490 macho_sectalign,
1491 macho_segbase,
1492 null_directive,
1493 macho_filename,
1494 macho_cleanup
1497 #endif
1500 * Local Variables:
1501 * mode:c
1502 * c-basic-offset:4
1503 * End:
1505 * end of file */