1 /* BFD back-end for WebAssembly modules.
2 Copyright (C) 2017-2018 Free Software Foundation, Inc.
4 Based on srec.c, mmo.c, and binary.c
6 This file is part of BFD, the Binary File Descriptor library.
8 This program is free software; you can redistribute it and/or modify
9 it under the terms of the GNU General Public License as published by
10 the Free Software Foundation; either version 3 of the License, or
11 (at your option) any later version.
13 This program is distributed in the hope that it will be useful,
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 GNU General Public License for more details.
18 You should have received a copy of the GNU General Public License
19 along with this program; if not, write to the Free Software
20 Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston,
21 MA 02110-1301, USA. */
23 /* The WebAssembly module format is a simple object file format
24 including up to 11 numbered sections, plus any number of named
25 "custom" sections. It is described at:
26 https://github.com/WebAssembly/design/blob/master/BinaryEncoding.md. */
29 #include "alloca-conf.h"
32 #include "libiberty.h"
34 #include "wasm-module.h"
39 bfd_size_type symcount;
42 static const char * const wasm_numbered_sections[] =
44 NULL, /* Custom section, different layout. */
45 WASM_SECTION ( 1, "type"),
46 WASM_SECTION ( 2, "import"),
47 WASM_SECTION ( 3, "function"),
48 WASM_SECTION ( 4, "table"),
49 WASM_SECTION ( 5, "memory"),
50 WASM_SECTION ( 6, "global"),
51 WASM_SECTION ( 7, "export"),
52 WASM_SECTION ( 8, "start"),
53 WASM_SECTION ( 9, "element"),
54 WASM_SECTION (10, "code"),
55 WASM_SECTION (11, "data"),
58 #define WASM_NUMBERED_SECTIONS ARRAY_SIZE (wasm_numbered_sections)
60 /* Resolve SECTION_CODE to a section name if there is one, NULL
64 wasm_section_code_to_name (bfd_byte section_code)
66 if (section_code < WASM_NUMBERED_SECTIONS)
67 return wasm_numbered_sections[section_code];
72 /* Translate section name NAME to a section code, or 0 if it's a
76 wasm_section_name_to_code (const char *name)
80 for (i = 1; i < WASM_NUMBERED_SECTIONS; i++)
81 if (strcmp (name, wasm_numbered_sections[i]) == 0)
87 /* WebAssembly LEB128 integers are sufficiently like DWARF LEB128
88 integers that we use _bfd_safe_read_leb128, but there are two
91 - WebAssembly requires a 32-bit value to be encoded in at most 5
93 - _bfd_safe_read_leb128 accepts incomplete LEB128 encodings at the
94 end of the buffer, while these are invalid in WebAssembly.
96 Those differences mean that we will accept some files that are
97 invalid WebAssembly. */
99 /* Read an LEB128-encoded integer from ABFD's I/O stream, reading one
100 byte at a time. Set ERROR_RETURN if no complete integer could be
101 read, LENGTH_RETURN to the number of bytes read (including bytes in
102 incomplete numbers). SIGN means interpret the number as SLEB128. */
105 wasm_read_leb128 (bfd * abfd,
106 bfd_boolean * error_return,
107 unsigned int * length_return,
111 unsigned int num_read = 0;
112 unsigned int shift = 0;
113 unsigned char byte = 0;
114 bfd_boolean success = FALSE;
116 while (bfd_bread (&byte, 1, abfd) == 1)
120 result |= ((bfd_vma) (byte & 0x7f)) << shift;
123 if ((byte & 0x80) == 0)
130 if (length_return != NULL)
131 *length_return = num_read;
132 if (error_return != NULL)
133 *error_return = ! success;
135 if (sign && (shift < 8 * sizeof (result)) && (byte & 0x40))
136 result |= -((bfd_vma) 1 << shift);
141 /* Encode an integer V as LEB128 and write it to ABFD, return TRUE on
145 wasm_write_uleb128 (bfd *abfd, bfd_vma v)
149 bfd_byte c = v & 0x7f;
155 if (bfd_bwrite (&c, 1, abfd) != 1)
163 /* Read the LEB128 integer at P, saving it to X; at end of buffer,
164 jump to error_return. */
165 #define READ_LEB128(x, p, end) \
168 unsigned int length_read; \
169 (x) = _bfd_safe_read_leb128 (abfd, (p), &length_read, \
171 (p) += length_read; \
172 if (length_read == 0) \
177 /* Verify the magic number at the beginning of a WebAssembly module
178 ABFD, setting ERRORPTR if there's a mismatch. */
181 wasm_read_magic (bfd *abfd, bfd_boolean *errorptr)
183 bfd_byte magic_const[SIZEOF_WASM_MAGIC] = WASM_MAGIC;
184 bfd_byte magic[SIZEOF_WASM_MAGIC];
186 if (bfd_bread (magic, sizeof (magic), abfd) == sizeof (magic)
187 && memcmp (magic, magic_const, sizeof (magic)) == 0)
194 /* Read the version number from ABFD, returning TRUE if it's a supported
195 version. Set ERRORPTR otherwise. */
198 wasm_read_version (bfd *abfd, bfd_boolean *errorptr)
200 bfd_byte vers_const[SIZEOF_WASM_VERSION] = WASM_VERSION;
201 bfd_byte vers[SIZEOF_WASM_VERSION];
203 if (bfd_bread (vers, sizeof (vers), abfd) == sizeof (vers)
204 /* Don't attempt to parse newer versions, which are likely to
205 require code changes. */
206 && memcmp (vers, vers_const, sizeof (vers)) == 0)
213 /* Read the WebAssembly header (magic number plus version number) from
214 ABFD, setting ERRORPTR to TRUE if there is a mismatch. */
217 wasm_read_header (bfd *abfd, bfd_boolean *errorptr)
219 if (! wasm_read_magic (abfd, errorptr))
222 if (! wasm_read_version (abfd, errorptr))
228 /* Scan the "function" subsection of the "name" section ASECT in the
229 wasm module ABFD. Create symbols. Return TRUE on success. */
232 wasm_scan_name_function_section (bfd *abfd, sec_ptr asect)
236 bfd_vma payload_size;
237 bfd_vma symcount = 0;
238 tdata_type *tdata = abfd->tdata.any;
239 asymbol *symbols = NULL;
240 sec_ptr space_function_index;
245 if (strcmp (asect->name, WASM_NAME_SECTION) != 0)
249 end = asect->contents + asect->size;
256 bfd_byte subsection_code = *p++;
257 if (subsection_code == WASM_FUNCTION_SUBSECTION)
260 /* subsection_code is documented to be a varuint7, meaning that
261 it has to be a single byte in the 0 - 127 range. If it isn't,
262 the spec must have changed underneath us, so give up. */
263 if (subsection_code & 0x80)
266 READ_LEB128 (payload_size, p, end);
268 if (p > p + payload_size)
277 READ_LEB128 (payload_size, p, end);
279 if (p > p + payload_size)
282 if (p + payload_size > end)
285 end = p + payload_size;
287 READ_LEB128 (symcount, p, end);
289 /* Sanity check: each symbol has at least two bytes. */
290 if (symcount > payload_size/2)
293 tdata->symcount = symcount;
295 space_function_index = bfd_make_section_with_flags
296 (abfd, WASM_SECTION_FUNCTION_INDEX, SEC_READONLY | SEC_CODE);
298 if (! space_function_index)
299 space_function_index = bfd_get_section_by_name (abfd, WASM_SECTION_FUNCTION_INDEX);
301 if (! space_function_index)
304 symbols = bfd_zalloc (abfd, tdata->symcount * sizeof (asymbol));
308 for (symcount = 0; p < end && symcount < tdata->symcount; symcount++)
315 READ_LEB128 (idx, p, end);
316 READ_LEB128 (len, p, end);
318 if (p + len < p || p + len > end)
321 name = bfd_zalloc (abfd, len + 1);
325 memcpy (name, p, len);
328 sym = &symbols[symcount];
332 sym->flags = BSF_GLOBAL | BSF_FUNCTION;
333 sym->section = space_function_index;
337 if (symcount < tdata->symcount)
340 tdata->symbols = symbols;
341 abfd->symcount = symcount;
347 bfd_release (abfd, (void *)symbols[--symcount].name);
348 bfd_release (abfd, symbols);
352 /* Read a byte from ABFD and return it, or EOF for EOF or error.
353 Set ERRORPTR on non-EOF error. */
356 wasm_read_byte (bfd *abfd, bfd_boolean *errorptr)
360 if (bfd_bread (&byte, (bfd_size_type) 1, abfd) != 1)
362 if (bfd_get_error () != bfd_error_file_truncated)
370 /* Scan the wasm module ABFD, creating sections and symbols.
371 Return TRUE on success. */
374 wasm_scan (bfd *abfd)
376 bfd_boolean error = FALSE;
377 /* Fake VMAs for now. Choose 0x80000000 as base to avoid clashes
378 with actual data addresses. */
379 bfd_vma vma = 0x80000000;
381 unsigned int bytes_read;
385 if (bfd_seek (abfd, (file_ptr) 0, SEEK_SET) != 0)
388 if (! wasm_read_header (abfd, &error))
391 while ((section_code = wasm_read_byte (abfd, &error)) != EOF)
393 if (section_code != 0)
395 const char *sname = wasm_section_code_to_name (section_code);
400 name = strdup (sname);
401 bfdsec = bfd_make_section_anyway_with_flags (abfd, name, SEC_HAS_CONTENTS);
408 bfdsec->size = wasm_read_leb128 (abfd, &error, &bytes_read, FALSE);
411 bfdsec->filepos = bfd_tell (abfd);
412 bfdsec->alignment_power = 0;
417 file_ptr section_start;
419 char *prefix = WASM_SECTION_PREFIX;
423 payload_len = wasm_read_leb128 (abfd, &error, &bytes_read, FALSE);
426 section_start = bfd_tell (abfd);
427 namelen = wasm_read_leb128 (abfd, &error, &bytes_read, FALSE);
428 if (error || namelen > payload_len)
430 name = bfd_zmalloc (namelen + strlen (prefix) + 1);
434 ret = sprintf (p, "%s", prefix);
435 if (ret < 0 || (bfd_vma) ret != strlen (prefix))
438 if (bfd_bread (p, namelen, abfd) != namelen)
441 bfdsec = bfd_make_section_anyway_with_flags (abfd, name, SEC_HAS_CONTENTS);
448 bfdsec->filepos = bfd_tell (abfd);
449 bfdsec->size = section_start + payload_len - bfdsec->filepos;
450 bfdsec->alignment_power = 0;
453 if (bfdsec->size != 0)
455 bfdsec->contents = bfd_zalloc (abfd, bfdsec->size);
456 if (! bfdsec->contents)
459 if (bfd_bread (bfdsec->contents, bfdsec->size, abfd) != bfdsec->size)
466 /* Make sure we're at actual EOF. There's no indication in the
467 WebAssembly format of how long the file is supposed to be. */
477 for (bfdsec = abfd->sections; bfdsec; bfdsec = bfdsec->next)
478 free ((void *) bfdsec->name);
483 /* Put a numbered section ASECT of ABFD into the table of numbered
484 sections pointed to by FSARG. */
487 wasm_register_section (bfd *abfd ATTRIBUTE_UNUSED,
491 sec_ptr *numbered_sections = fsarg;
492 int idx = wasm_section_name_to_code (asect->name);
497 numbered_sections[idx] = asect;
500 struct compute_section_arg
506 /* Compute the file position of ABFD's section ASECT. FSARG is a
507 pointer to the current file position.
509 We allow section names of the form .wasm.id to encode the numbered
510 section with ID id, if it exists; otherwise, a custom section with
511 ID "id" is produced. Arbitrary section names are for sections that
512 are assumed already to contain a section header; those are appended
513 to the WebAssembly module verbatim. */
516 wasm_compute_custom_section_file_position (bfd *abfd,
520 struct compute_section_arg *fs = fsarg;
526 idx = wasm_section_name_to_code (asect->name);
531 if (CONST_STRNEQ (asect->name, WASM_SECTION_PREFIX))
533 const char *name = asect->name + strlen (WASM_SECTION_PREFIX);
534 bfd_size_type payload_len = asect->size;
535 bfd_size_type name_len = strlen (name);
536 bfd_size_type nl = name_len;
538 payload_len += name_len;
547 bfd_seek (abfd, fs->pos, SEEK_SET);
548 if (! wasm_write_uleb128 (abfd, 0)
549 || ! wasm_write_uleb128 (abfd, payload_len)
550 || ! wasm_write_uleb128 (abfd, name_len)
551 || bfd_bwrite (name, name_len, abfd) != name_len)
553 fs->pos = asect->filepos = bfd_tell (abfd);
557 asect->filepos = fs->pos;
561 fs->pos += asect->size;
568 /* Compute the file positions for the sections of ABFD. Currently,
569 this writes all numbered sections first, in order, then all custom
570 sections, in section order.
572 The spec says that the numbered sections must appear in order of
573 their ids, but custom sections can appear in any position and any
574 order, and more than once. FIXME: support that. */
577 wasm_compute_section_file_positions (bfd *abfd)
579 bfd_byte magic[SIZEOF_WASM_MAGIC] = WASM_MAGIC;
580 bfd_byte vers[SIZEOF_WASM_VERSION] = WASM_VERSION;
581 sec_ptr numbered_sections[WASM_NUMBERED_SECTIONS];
582 struct compute_section_arg fs;
585 bfd_seek (abfd, (bfd_vma) 0, SEEK_SET);
587 if (bfd_bwrite (magic, sizeof (magic), abfd) != (sizeof magic)
588 || bfd_bwrite (vers, sizeof (vers), abfd) != sizeof (vers))
591 for (i = 0; i < WASM_NUMBERED_SECTIONS; i++)
592 numbered_sections[i] = NULL;
594 bfd_map_over_sections (abfd, wasm_register_section, numbered_sections);
596 fs.pos = bfd_tell (abfd);
597 for (i = 0; i < WASM_NUMBERED_SECTIONS; i++)
599 sec_ptr sec = numbered_sections[i];
605 if (bfd_seek (abfd, fs.pos, SEEK_SET) != 0)
607 if (! wasm_write_uleb128 (abfd, i)
608 || ! wasm_write_uleb128 (abfd, size))
610 fs.pos = sec->filepos = bfd_tell (abfd);
616 bfd_map_over_sections (abfd, wasm_compute_custom_section_file_position, &fs);
621 abfd->output_has_begun = TRUE;
627 wasm_set_section_contents (bfd *abfd,
629 const void *location,
636 if (! abfd->output_has_begun
637 && ! wasm_compute_section_file_positions (abfd))
640 if (bfd_seek (abfd, section->filepos + offset, SEEK_SET) != 0
641 || bfd_bwrite (location, count, abfd) != count)
648 wasm_write_object_contents (bfd* abfd)
650 bfd_byte magic[] = WASM_MAGIC;
651 bfd_byte vers[] = WASM_VERSION;
653 if (bfd_seek (abfd, 0, SEEK_SET) != 0)
656 if (bfd_bwrite (magic, sizeof (magic), abfd) != sizeof (magic)
657 || bfd_bwrite (vers, sizeof (vers), abfd) != sizeof (vers))
664 wasm_mkobject (bfd *abfd)
666 tdata_type *tdata = (tdata_type *) bfd_alloc (abfd, sizeof (tdata_type));
671 tdata->symbols = NULL;
674 abfd->tdata.any = tdata;
680 wasm_get_symtab_upper_bound (bfd *abfd)
682 tdata_type *tdata = abfd->tdata.any;
684 return (tdata->symcount + 1) * (sizeof (asymbol *));
688 wasm_canonicalize_symtab (bfd *abfd, asymbol **alocation)
690 tdata_type *tdata = abfd->tdata.any;
693 for (i = 0; i < tdata->symcount; i++)
694 alocation[i] = &tdata->symbols[i];
697 return tdata->symcount;
701 wasm_make_empty_symbol (bfd *abfd)
703 bfd_size_type amt = sizeof (asymbol);
704 asymbol *new_symbol = (asymbol *) bfd_zalloc (abfd, amt);
708 new_symbol->the_bfd = abfd;
713 wasm_print_symbol (bfd *abfd,
716 bfd_print_symbol_type how)
718 FILE *file = (FILE *) filep;
722 case bfd_print_symbol_name:
723 fprintf (file, "%s", symbol->name);
727 bfd_print_symbol_vandf (abfd, filep, symbol);
728 fprintf (file, " %-5s %s", symbol->section->name, symbol->name);
733 wasm_get_symbol_info (bfd *abfd ATTRIBUTE_UNUSED,
737 bfd_symbol_info (symbol, ret);
740 /* Check whether ABFD is a WebAssembly module; if so, scan it. */
742 static const bfd_target *
743 wasm_object_p (bfd *abfd)
747 if (bfd_seek (abfd, (file_ptr) 0, SEEK_SET) != 0)
750 if (! wasm_read_header (abfd, &error))
752 bfd_set_error (bfd_error_wrong_format);
756 if (! wasm_mkobject (abfd) || ! wasm_scan (abfd))
759 if (! bfd_default_set_arch_mach (abfd, bfd_arch_wasm32, 0))
762 if (wasm_scan_name_function_section (abfd, bfd_get_section_by_name (abfd, WASM_NAME_SECTION)))
763 abfd->flags |= HAS_SYMS;
768 /* BFD_JUMP_TABLE_WRITE */
769 #define wasm_set_arch_mach _bfd_generic_set_arch_mach
771 /* BFD_JUMP_TABLE_SYMBOLS */
772 #define wasm_get_symbol_version_string _bfd_nosymbols_get_symbol_version_string
773 #define wasm_bfd_is_local_label_name bfd_generic_is_local_label_name
774 #define wasm_bfd_is_target_special_symbol _bfd_bool_bfd_asymbol_false
775 #define wasm_get_lineno _bfd_nosymbols_get_lineno
776 #define wasm_find_nearest_line _bfd_nosymbols_find_nearest_line
777 #define wasm_find_line _bfd_nosymbols_find_line
778 #define wasm_find_inliner_info _bfd_nosymbols_find_inliner_info
779 #define wasm_bfd_make_debug_symbol _bfd_nosymbols_bfd_make_debug_symbol
780 #define wasm_read_minisymbols _bfd_generic_read_minisymbols
781 #define wasm_minisymbol_to_symbol _bfd_generic_minisymbol_to_symbol
783 const bfd_target wasm_vec =
786 bfd_target_unknown_flavour,
789 (HAS_SYMS | WP_TEXT), /* Object flags. */
790 (SEC_CODE | SEC_DATA | SEC_HAS_CONTENTS), /* Section flags. */
791 0, /* Leading underscore. */
792 ' ', /* AR_pad_char. */
793 255, /* AR_max_namelen. */
794 0, /* Match priority. */
795 /* Routines to byte-swap various sized integers from the data sections. */
796 bfd_getl64, bfd_getl_signed_64, bfd_putl64,
797 bfd_getl32, bfd_getl_signed_32, bfd_putl32,
798 bfd_getl16, bfd_getl_signed_16, bfd_putl16,
800 /* Routines to byte-swap various sized integers from the file headers. */
801 bfd_getl64, bfd_getl_signed_64, bfd_putl64,
802 bfd_getl32, bfd_getl_signed_32, bfd_putl32,
803 bfd_getl16, bfd_getl_signed_16, bfd_putl16,
807 wasm_object_p, /* bfd_check_format. */
812 _bfd_bool_bfd_false_error,
814 _bfd_generic_mkarchive,
815 _bfd_bool_bfd_false_error,
817 { /* bfd_write_contents. */
818 _bfd_bool_bfd_false_error,
819 wasm_write_object_contents,
820 _bfd_write_archive_contents,
821 _bfd_bool_bfd_false_error,
824 BFD_JUMP_TABLE_GENERIC (_bfd_generic),
825 BFD_JUMP_TABLE_COPY (_bfd_generic),
826 BFD_JUMP_TABLE_CORE (_bfd_nocore),
827 BFD_JUMP_TABLE_ARCHIVE (_bfd_noarchive),
828 BFD_JUMP_TABLE_SYMBOLS (wasm),
829 BFD_JUMP_TABLE_RELOCS (_bfd_norelocs),
830 BFD_JUMP_TABLE_WRITE (wasm),
831 BFD_JUMP_TABLE_LINK (_bfd_nolink),
832 BFD_JUMP_TABLE_DYNAMIC (_bfd_nodynamic),