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"
33 #include "bfd_stdint.h"
34 #include "libiberty.h"
36 #include "wasm-module.h"
41 bfd_size_type symcount;
44 static const char * const wasm_numbered_sections[] =
46 NULL, /* Custom section, different layout. */
47 WASM_SECTION ( 1, "type"),
48 WASM_SECTION ( 2, "import"),
49 WASM_SECTION ( 3, "function"),
50 WASM_SECTION ( 4, "table"),
51 WASM_SECTION ( 5, "memory"),
52 WASM_SECTION ( 6, "global"),
53 WASM_SECTION ( 7, "export"),
54 WASM_SECTION ( 8, "start"),
55 WASM_SECTION ( 9, "element"),
56 WASM_SECTION (10, "code"),
57 WASM_SECTION (11, "data"),
60 #define WASM_NUMBERED_SECTIONS ARRAY_SIZE (wasm_numbered_sections)
62 /* Resolve SECTION_CODE to a section name if there is one, NULL
66 wasm_section_code_to_name (bfd_byte section_code)
68 if (section_code < WASM_NUMBERED_SECTIONS)
69 return wasm_numbered_sections[section_code];
74 /* Translate section name NAME to a section code, or 0 if it's a
78 wasm_section_name_to_code (const char *name)
82 for (i = 1; i < WASM_NUMBERED_SECTIONS; i++)
83 if (strcmp (name, wasm_numbered_sections[i]) == 0)
89 /* WebAssembly LEB128 integers are sufficiently like DWARF LEB128
90 integers that we use _bfd_safe_read_leb128, but there are two
93 - WebAssembly requires a 32-bit value to be encoded in at most 5
95 - _bfd_safe_read_leb128 accepts incomplete LEB128 encodings at the
96 end of the buffer, while these are invalid in WebAssembly.
98 Those differences mean that we will accept some files that are
99 invalid WebAssembly. */
101 /* Read an LEB128-encoded integer from ABFD's I/O stream, reading one
102 byte at a time. Set ERROR_RETURN if no complete integer could be
103 read, LENGTH_RETURN to the number of bytes read (including bytes in
104 incomplete numbers). SIGN means interpret the number as SLEB128. */
107 wasm_read_leb128 (bfd * abfd,
108 bfd_boolean * error_return,
109 unsigned int * length_return,
113 unsigned int num_read = 0;
114 unsigned int shift = 0;
115 unsigned char byte = 0;
116 bfd_boolean success = FALSE;
118 while (bfd_bread (&byte, 1, abfd) == 1)
122 result |= ((bfd_vma) (byte & 0x7f)) << shift;
125 if ((byte & 0x80) == 0)
132 if (length_return != NULL)
133 *length_return = num_read;
134 if (error_return != NULL)
135 *error_return = ! success;
137 if (sign && (shift < 8 * sizeof (result)) && (byte & 0x40))
138 result |= -((bfd_vma) 1 << shift);
143 /* Encode an integer V as LEB128 and write it to ABFD, return TRUE on
147 wasm_write_uleb128 (bfd *abfd, bfd_vma v)
151 bfd_byte c = v & 0x7f;
157 if (bfd_bwrite (&c, 1, abfd) != 1)
165 /* Read the LEB128 integer at P, saving it to X; at end of buffer,
166 jump to error_return. */
167 #define READ_LEB128(x, p, end) \
170 unsigned int length_read; \
171 (x) = _bfd_safe_read_leb128 (abfd, (p), &length_read, \
173 (p) += length_read; \
174 if (length_read == 0) \
179 /* Verify the magic number at the beginning of a WebAssembly module
180 ABFD, setting ERRORPTR if there's a mismatch. */
183 wasm_read_magic (bfd *abfd, bfd_boolean *errorptr)
185 bfd_byte magic_const[SIZEOF_WASM_MAGIC] = WASM_MAGIC;
186 bfd_byte magic[SIZEOF_WASM_MAGIC];
188 if (bfd_bread (magic, sizeof (magic), abfd) == sizeof (magic)
189 && memcmp (magic, magic_const, sizeof (magic)) == 0)
196 /* Read the version number from ABFD, returning TRUE if it's a supported
197 version. Set ERRORPTR otherwise. */
200 wasm_read_version (bfd *abfd, bfd_boolean *errorptr)
202 bfd_byte vers_const[SIZEOF_WASM_VERSION] = WASM_VERSION;
203 bfd_byte vers[SIZEOF_WASM_VERSION];
205 if (bfd_bread (vers, sizeof (vers), abfd) == sizeof (vers)
206 /* Don't attempt to parse newer versions, which are likely to
207 require code changes. */
208 && memcmp (vers, vers_const, sizeof (vers)) == 0)
215 /* Read the WebAssembly header (magic number plus version number) from
216 ABFD, setting ERRORPTR to TRUE if there is a mismatch. */
219 wasm_read_header (bfd *abfd, bfd_boolean *errorptr)
221 if (! wasm_read_magic (abfd, errorptr))
224 if (! wasm_read_version (abfd, errorptr))
230 /* Scan the "function" subsection of the "name" section ASECT in the
231 wasm module ABFD. Create symbols. Return TRUE on success. */
234 wasm_scan_name_function_section (bfd *abfd, sec_ptr asect)
238 bfd_vma payload_size;
239 bfd_vma symcount = 0;
240 tdata_type *tdata = abfd->tdata.any;
241 asymbol *symbols = NULL;
242 sec_ptr space_function_index;
247 if (strcmp (asect->name, WASM_NAME_SECTION) != 0)
251 end = asect->contents + asect->size;
258 bfd_byte subsection_code = *p++;
259 if (subsection_code == WASM_FUNCTION_SUBSECTION)
262 /* subsection_code is documented to be a varuint7, meaning that
263 it has to be a single byte in the 0 - 127 range. If it isn't,
264 the spec must have changed underneath us, so give up. */
265 if (subsection_code & 0x80)
268 READ_LEB128 (payload_size, p, end);
270 if (p > p + payload_size)
279 READ_LEB128 (payload_size, p, end);
281 if (p > p + payload_size)
284 if (p + payload_size > end)
287 end = p + payload_size;
289 READ_LEB128 (symcount, p, end);
291 /* Sanity check: each symbol has at least two bytes. */
292 if (symcount > payload_size/2)
295 tdata->symcount = symcount;
297 space_function_index = bfd_make_section_with_flags
298 (abfd, WASM_SECTION_FUNCTION_INDEX, SEC_READONLY | SEC_CODE);
300 if (! space_function_index)
301 space_function_index = bfd_get_section_by_name (abfd, WASM_SECTION_FUNCTION_INDEX);
303 if (! space_function_index)
306 symbols = bfd_zalloc (abfd, tdata->symcount * sizeof (asymbol));
310 for (symcount = 0; p < end && symcount < tdata->symcount; symcount++)
317 READ_LEB128 (idx, p, end);
318 READ_LEB128 (len, p, end);
320 if (p + len < p || p + len > end)
323 name = bfd_zalloc (abfd, len + 1);
327 memcpy (name, p, len);
330 sym = &symbols[symcount];
334 sym->flags = BSF_GLOBAL | BSF_FUNCTION;
335 sym->section = space_function_index;
339 if (symcount < tdata->symcount)
342 tdata->symbols = symbols;
343 abfd->symcount = symcount;
349 bfd_release (abfd, (void *)symbols[--symcount].name);
350 bfd_release (abfd, symbols);
354 /* Read a byte from ABFD and return it, or EOF for EOF or error.
355 Set ERRORPTR on non-EOF error. */
358 wasm_read_byte (bfd *abfd, bfd_boolean *errorptr)
362 if (bfd_bread (&byte, (bfd_size_type) 1, abfd) != 1)
364 if (bfd_get_error () != bfd_error_file_truncated)
372 /* Scan the wasm module ABFD, creating sections and symbols.
373 Return TRUE on success. */
376 wasm_scan (bfd *abfd)
378 bfd_boolean error = FALSE;
379 /* Fake VMAs for now. Choose 0x80000000 as base to avoid clashes
380 with actual data addresses. */
381 bfd_vma vma = 0x80000000;
383 unsigned int bytes_read;
387 if (bfd_seek (abfd, (file_ptr) 0, SEEK_SET) != 0)
390 if (! wasm_read_header (abfd, &error))
393 while ((section_code = wasm_read_byte (abfd, &error)) != EOF)
395 if (section_code != 0)
397 const char *sname = wasm_section_code_to_name (section_code);
402 name = strdup (sname);
403 bfdsec = bfd_make_section_anyway_with_flags (abfd, name, SEC_HAS_CONTENTS);
410 bfdsec->size = wasm_read_leb128 (abfd, &error, &bytes_read, FALSE);
413 bfdsec->filepos = bfd_tell (abfd);
414 bfdsec->alignment_power = 0;
419 file_ptr section_start;
421 char *prefix = WASM_SECTION_PREFIX;
425 payload_len = wasm_read_leb128 (abfd, &error, &bytes_read, FALSE);
428 section_start = bfd_tell (abfd);
429 namelen = wasm_read_leb128 (abfd, &error, &bytes_read, FALSE);
430 if (error || namelen > payload_len)
432 name = bfd_zmalloc (namelen + strlen (prefix) + 1);
436 ret = sprintf (p, "%s", prefix);
437 if (ret < 0 || (bfd_vma) ret != strlen (prefix))
440 if (bfd_bread (p, namelen, abfd) != namelen)
443 bfdsec = bfd_make_section_anyway_with_flags (abfd, name, SEC_HAS_CONTENTS);
450 bfdsec->filepos = bfd_tell (abfd);
451 bfdsec->size = section_start + payload_len - bfdsec->filepos;
452 bfdsec->alignment_power = 0;
455 if (bfdsec->size != 0)
457 bfdsec->contents = bfd_zalloc (abfd, bfdsec->size);
458 if (! bfdsec->contents)
461 if (bfd_bread (bfdsec->contents, bfdsec->size, abfd) != bfdsec->size)
468 /* Make sure we're at actual EOF. There's no indication in the
469 WebAssembly format of how long the file is supposed to be. */
479 for (bfdsec = abfd->sections; bfdsec; bfdsec = bfdsec->next)
480 free ((void *) bfdsec->name);
485 /* Put a numbered section ASECT of ABFD into the table of numbered
486 sections pointed to by FSARG. */
489 wasm_register_section (bfd *abfd ATTRIBUTE_UNUSED,
493 sec_ptr *numbered_sections = fsarg;
494 int idx = wasm_section_name_to_code (asect->name);
499 numbered_sections[idx] = asect;
502 struct compute_section_arg
508 /* Compute the file position of ABFD's section ASECT. FSARG is a
509 pointer to the current file position.
511 We allow section names of the form .wasm.id to encode the numbered
512 section with ID id, if it exists; otherwise, a custom section with
513 ID "id" is produced. Arbitrary section names are for sections that
514 are assumed already to contain a section header; those are appended
515 to the WebAssembly module verbatim. */
518 wasm_compute_custom_section_file_position (bfd *abfd,
522 struct compute_section_arg *fs = fsarg;
528 idx = wasm_section_name_to_code (asect->name);
533 if (CONST_STRNEQ (asect->name, WASM_SECTION_PREFIX))
535 const char *name = asect->name + strlen (WASM_SECTION_PREFIX);
536 bfd_size_type payload_len = asect->size;
537 bfd_size_type name_len = strlen (name);
538 bfd_size_type nl = name_len;
540 payload_len += name_len;
549 bfd_seek (abfd, fs->pos, SEEK_SET);
550 if (! wasm_write_uleb128 (abfd, 0)
551 || ! wasm_write_uleb128 (abfd, payload_len)
552 || ! wasm_write_uleb128 (abfd, name_len)
553 || bfd_bwrite (name, name_len, abfd) != name_len)
555 fs->pos = asect->filepos = bfd_tell (abfd);
559 asect->filepos = fs->pos;
563 fs->pos += asect->size;
570 /* Compute the file positions for the sections of ABFD. Currently,
571 this writes all numbered sections first, in order, then all custom
572 sections, in section order.
574 The spec says that the numbered sections must appear in order of
575 their ids, but custom sections can appear in any position and any
576 order, and more than once. FIXME: support that. */
579 wasm_compute_section_file_positions (bfd *abfd)
581 bfd_byte magic[SIZEOF_WASM_MAGIC] = WASM_MAGIC;
582 bfd_byte vers[SIZEOF_WASM_VERSION] = WASM_VERSION;
583 sec_ptr numbered_sections[WASM_NUMBERED_SECTIONS];
584 struct compute_section_arg fs;
587 bfd_seek (abfd, (bfd_vma) 0, SEEK_SET);
589 if (bfd_bwrite (magic, sizeof (magic), abfd) != (sizeof magic)
590 || bfd_bwrite (vers, sizeof (vers), abfd) != sizeof (vers))
593 for (i = 0; i < WASM_NUMBERED_SECTIONS; i++)
594 numbered_sections[i] = NULL;
596 bfd_map_over_sections (abfd, wasm_register_section, numbered_sections);
598 fs.pos = bfd_tell (abfd);
599 for (i = 0; i < WASM_NUMBERED_SECTIONS; i++)
601 sec_ptr sec = numbered_sections[i];
607 if (bfd_seek (abfd, fs.pos, SEEK_SET) != 0)
609 if (! wasm_write_uleb128 (abfd, i)
610 || ! wasm_write_uleb128 (abfd, size))
612 fs.pos = sec->filepos = bfd_tell (abfd);
618 bfd_map_over_sections (abfd, wasm_compute_custom_section_file_position, &fs);
623 abfd->output_has_begun = TRUE;
629 wasm_set_section_contents (bfd *abfd,
631 const void *location,
638 if (! abfd->output_has_begun
639 && ! wasm_compute_section_file_positions (abfd))
642 if (bfd_seek (abfd, section->filepos + offset, SEEK_SET) != 0
643 || bfd_bwrite (location, count, abfd) != count)
650 wasm_write_object_contents (bfd* abfd)
652 bfd_byte magic[] = WASM_MAGIC;
653 bfd_byte vers[] = WASM_VERSION;
655 if (bfd_seek (abfd, 0, SEEK_SET) != 0)
658 if (bfd_bwrite (magic, sizeof (magic), abfd) != sizeof (magic)
659 || bfd_bwrite (vers, sizeof (vers), abfd) != sizeof (vers))
666 wasm_mkobject (bfd *abfd)
668 tdata_type *tdata = (tdata_type *) bfd_alloc (abfd, sizeof (tdata_type));
673 tdata->symbols = NULL;
676 abfd->tdata.any = tdata;
682 wasm_get_symtab_upper_bound (bfd *abfd)
684 tdata_type *tdata = abfd->tdata.any;
686 return (tdata->symcount + 1) * (sizeof (asymbol *));
690 wasm_canonicalize_symtab (bfd *abfd, asymbol **alocation)
692 tdata_type *tdata = abfd->tdata.any;
695 for (i = 0; i < tdata->symcount; i++)
696 alocation[i] = &tdata->symbols[i];
699 return tdata->symcount;
703 wasm_make_empty_symbol (bfd *abfd)
705 bfd_size_type amt = sizeof (asymbol);
706 asymbol *new_symbol = (asymbol *) bfd_zalloc (abfd, amt);
710 new_symbol->the_bfd = abfd;
715 wasm_print_symbol (bfd *abfd,
718 bfd_print_symbol_type how)
720 FILE *file = (FILE *) filep;
724 case bfd_print_symbol_name:
725 fprintf (file, "%s", symbol->name);
729 bfd_print_symbol_vandf (abfd, filep, symbol);
730 fprintf (file, " %-5s %s", symbol->section->name, symbol->name);
735 wasm_get_symbol_info (bfd *abfd ATTRIBUTE_UNUSED,
739 bfd_symbol_info (symbol, ret);
742 /* Check whether ABFD is a WebAssembly module; if so, scan it. */
744 static const bfd_target *
745 wasm_object_p (bfd *abfd)
749 if (bfd_seek (abfd, (file_ptr) 0, SEEK_SET) != 0)
752 if (! wasm_read_header (abfd, &error))
754 bfd_set_error (bfd_error_wrong_format);
758 if (! wasm_mkobject (abfd) || ! wasm_scan (abfd))
761 if (! bfd_default_set_arch_mach (abfd, bfd_arch_wasm32, 0))
764 if (wasm_scan_name_function_section (abfd, bfd_get_section_by_name (abfd, WASM_NAME_SECTION)))
765 abfd->flags |= HAS_SYMS;
770 /* BFD_JUMP_TABLE_WRITE */
771 #define wasm_set_arch_mach _bfd_generic_set_arch_mach
773 /* BFD_JUMP_TABLE_SYMBOLS */
774 #define wasm_get_symbol_version_string _bfd_nosymbols_get_symbol_version_string
775 #define wasm_bfd_is_local_label_name bfd_generic_is_local_label_name
776 #define wasm_bfd_is_target_special_symbol _bfd_bool_bfd_asymbol_false
777 #define wasm_get_lineno _bfd_nosymbols_get_lineno
778 #define wasm_find_nearest_line _bfd_nosymbols_find_nearest_line
779 #define wasm_find_line _bfd_nosymbols_find_line
780 #define wasm_find_inliner_info _bfd_nosymbols_find_inliner_info
781 #define wasm_bfd_make_debug_symbol _bfd_nosymbols_bfd_make_debug_symbol
782 #define wasm_read_minisymbols _bfd_generic_read_minisymbols
783 #define wasm_minisymbol_to_symbol _bfd_generic_minisymbol_to_symbol
785 const bfd_target wasm_vec =
788 bfd_target_unknown_flavour,
791 (HAS_SYMS | WP_TEXT), /* Object flags. */
792 (SEC_CODE | SEC_DATA | SEC_HAS_CONTENTS), /* Section flags. */
793 0, /* Leading underscore. */
794 ' ', /* AR_pad_char. */
795 255, /* AR_max_namelen. */
796 0, /* Match priority. */
797 /* Routines to byte-swap various sized integers from the data sections. */
798 bfd_getl64, bfd_getl_signed_64, bfd_putl64,
799 bfd_getl32, bfd_getl_signed_32, bfd_putl32,
800 bfd_getl16, bfd_getl_signed_16, bfd_putl16,
802 /* Routines to byte-swap various sized integers from the file headers. */
803 bfd_getl64, bfd_getl_signed_64, bfd_putl64,
804 bfd_getl32, bfd_getl_signed_32, bfd_putl32,
805 bfd_getl16, bfd_getl_signed_16, bfd_putl16,
809 wasm_object_p, /* bfd_check_format. */
814 _bfd_bool_bfd_false_error,
816 _bfd_generic_mkarchive,
817 _bfd_bool_bfd_false_error,
819 { /* bfd_write_contents. */
820 _bfd_bool_bfd_false_error,
821 wasm_write_object_contents,
822 _bfd_write_archive_contents,
823 _bfd_bool_bfd_false_error,
826 BFD_JUMP_TABLE_GENERIC (_bfd_generic),
827 BFD_JUMP_TABLE_COPY (_bfd_generic),
828 BFD_JUMP_TABLE_CORE (_bfd_nocore),
829 BFD_JUMP_TABLE_ARCHIVE (_bfd_noarchive),
830 BFD_JUMP_TABLE_SYMBOLS (wasm),
831 BFD_JUMP_TABLE_RELOCS (_bfd_norelocs),
832 BFD_JUMP_TABLE_WRITE (wasm),
833 BFD_JUMP_TABLE_LINK (_bfd_nolink),
834 BFD_JUMP_TABLE_DYNAMIC (_bfd_nodynamic),