1 /* BFD back-end for WebAssembly modules.
2 Copyright (C) 2017-2021 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. */
30 #include "libiberty.h"
32 #include "wasm-module.h"
42 bfd_size_type symcount
;
45 static const char * const wasm_numbered_sections
[] =
47 NULL
, /* Custom section, different layout. */
48 WASM_SECTION ( 1, "type"),
49 WASM_SECTION ( 2, "import"),
50 WASM_SECTION ( 3, "function"),
51 WASM_SECTION ( 4, "table"),
52 WASM_SECTION ( 5, "memory"),
53 WASM_SECTION ( 6, "global"),
54 WASM_SECTION ( 7, "export"),
55 WASM_SECTION ( 8, "start"),
56 WASM_SECTION ( 9, "element"),
57 WASM_SECTION (10, "code"),
58 WASM_SECTION (11, "data"),
61 #define WASM_NUMBERED_SECTIONS ARRAY_SIZE (wasm_numbered_sections)
63 /* Resolve SECTION_CODE to a section name if there is one, NULL
67 wasm_section_code_to_name (bfd_byte section_code
)
69 if (section_code
< WASM_NUMBERED_SECTIONS
)
70 return wasm_numbered_sections
[section_code
];
75 /* Translate section name NAME to a section code, or 0 if it's a
79 wasm_section_name_to_code (const char *name
)
83 for (i
= 1; i
< WASM_NUMBERED_SECTIONS
; i
++)
84 if (strcmp (name
, wasm_numbered_sections
[i
]) == 0)
90 /* WebAssembly LEB128 integers are sufficiently like DWARF LEB128
91 integers that we use _bfd_safe_read_leb128, but there are two
94 - WebAssembly requires a 32-bit value to be encoded in at most 5
96 - _bfd_safe_read_leb128 accepts incomplete LEB128 encodings at the
97 end of the buffer, while these are invalid in WebAssembly.
99 Those differences mean that we will accept some files that are
100 invalid WebAssembly. */
102 /* Read an LEB128-encoded integer from ABFD's I/O stream, reading one
103 byte at a time. Set ERROR_RETURN if no complete integer could be
104 read, LENGTH_RETURN to the number of bytes read (including bytes in
105 incomplete numbers). SIGN means interpret the number as SLEB128. */
108 wasm_read_leb128 (bfd
*abfd
,
110 unsigned int *length_return
,
114 unsigned int num_read
= 0;
115 unsigned int shift
= 0;
116 unsigned char byte
= 0;
117 unsigned char lost
, mask
;
120 while (bfd_bread (&byte
, 1, abfd
) == 1)
124 if (shift
< CHAR_BIT
* sizeof (result
))
126 result
|= ((bfd_vma
) (byte
& 0x7f)) << shift
;
127 /* These bits overflowed. */
128 lost
= byte
^ (result
>> shift
);
129 /* And this is the mask of possible overflow bits. */
130 mask
= 0x7f ^ ((bfd_vma
) 0x7f << shift
>> shift
);
138 if ((lost
& mask
) != (sign
&& (bfd_signed_vma
) result
< 0 ? mask
: 0))
141 if ((byte
& 0x80) == 0)
144 if (sign
&& shift
< CHAR_BIT
* sizeof (result
) && (byte
& 0x40))
145 result
|= -((bfd_vma
) 1 << shift
);
150 if (length_return
!= NULL
)
151 *length_return
= num_read
;
152 if (error_return
!= NULL
)
153 *error_return
= status
!= 0;
158 /* Encode an integer V as LEB128 and write it to ABFD, return TRUE on
162 wasm_write_uleb128 (bfd
*abfd
, bfd_vma v
)
166 bfd_byte c
= v
& 0x7f;
172 if (bfd_bwrite (&c
, 1, abfd
) != 1)
180 /* Read the LEB128 integer at P, saving it to X; at end of buffer,
181 jump to error_return. */
182 #define READ_LEB128(x, p, end) \
187 (x) = _bfd_safe_read_leb128 (abfd, &(p), false, (end)); \
191 /* Verify the magic number at the beginning of a WebAssembly module
192 ABFD, setting ERRORPTR if there's a mismatch. */
195 wasm_read_magic (bfd
*abfd
, bool *errorptr
)
197 bfd_byte magic_const
[SIZEOF_WASM_MAGIC
] = WASM_MAGIC
;
198 bfd_byte magic
[SIZEOF_WASM_MAGIC
];
200 if (bfd_bread (magic
, sizeof (magic
), abfd
) == sizeof (magic
)
201 && memcmp (magic
, magic_const
, sizeof (magic
)) == 0)
208 /* Read the version number from ABFD, returning TRUE if it's a supported
209 version. Set ERRORPTR otherwise. */
212 wasm_read_version (bfd
*abfd
, bool *errorptr
)
214 bfd_byte vers_const
[SIZEOF_WASM_VERSION
] = WASM_VERSION
;
215 bfd_byte vers
[SIZEOF_WASM_VERSION
];
217 if (bfd_bread (vers
, sizeof (vers
), abfd
) == sizeof (vers
)
218 /* Don't attempt to parse newer versions, which are likely to
219 require code changes. */
220 && memcmp (vers
, vers_const
, sizeof (vers
)) == 0)
227 /* Read the WebAssembly header (magic number plus version number) from
228 ABFD, setting ERRORPTR to TRUE if there is a mismatch. */
231 wasm_read_header (bfd
*abfd
, bool *errorptr
)
233 if (! wasm_read_magic (abfd
, errorptr
))
236 if (! wasm_read_version (abfd
, errorptr
))
242 /* Scan the "function" subsection of the "name" section ASECT in the
243 wasm module ABFD. Create symbols. Return TRUE on success. */
246 wasm_scan_name_function_section (bfd
*abfd
, sec_ptr asect
)
250 bfd_vma payload_size
;
251 bfd_vma symcount
= 0;
252 tdata_type
*tdata
= abfd
->tdata
.any
;
253 asymbol
*symbols
= NULL
;
254 sec_ptr space_function_index
;
258 end
= asect
->contents
+ asect
->size
;
265 bfd_byte subsection_code
= *p
++;
266 if (subsection_code
== WASM_FUNCTION_SUBSECTION
)
269 /* subsection_code is documented to be a varuint7, meaning that
270 it has to be a single byte in the 0 - 127 range. If it isn't,
271 the spec must have changed underneath us, so give up. */
272 if (subsection_code
& 0x80)
275 READ_LEB128 (payload_size
, p
, end
);
277 if (payload_size
> (size_t) (end
- p
))
286 READ_LEB128 (payload_size
, p
, end
);
288 if (payload_size
> (size_t) (end
- p
))
291 end
= p
+ payload_size
;
293 READ_LEB128 (symcount
, p
, end
);
295 /* Sanity check: each symbol has at least two bytes. */
296 if (symcount
> payload_size
/ 2)
299 tdata
->symcount
= symcount
;
302 = bfd_make_section_with_flags (abfd
, WASM_SECTION_FUNCTION_INDEX
,
303 SEC_READONLY
| SEC_CODE
);
305 if (!space_function_index
)
307 = bfd_get_section_by_name (abfd
, WASM_SECTION_FUNCTION_INDEX
);
309 if (!space_function_index
)
312 if (_bfd_mul_overflow (tdata
->symcount
, sizeof (asymbol
), &amt
))
314 bfd_set_error (bfd_error_file_too_big
);
317 symbols
= bfd_alloc (abfd
, amt
);
321 for (symcount
= 0; p
< end
&& symcount
< tdata
->symcount
; symcount
++)
328 READ_LEB128 (idx
, p
, end
);
329 READ_LEB128 (len
, p
, end
);
331 if (len
> (size_t) (end
- p
))
334 name
= bfd_alloc (abfd
, len
+ 1);
338 memcpy (name
, p
, len
);
342 sym
= &symbols
[symcount
];
346 sym
->flags
= BSF_GLOBAL
| BSF_FUNCTION
;
347 sym
->section
= space_function_index
;
351 if (symcount
< tdata
->symcount
)
354 tdata
->symbols
= symbols
;
355 abfd
->symcount
= symcount
;
360 bfd_release (abfd
, symbols
);
364 /* Read a byte from ABFD and return it, or EOF for EOF or error.
365 Set ERRORPTR on non-EOF error. */
368 wasm_read_byte (bfd
*abfd
, bool *errorptr
)
372 if (bfd_bread (&byte
, (bfd_size_type
) 1, abfd
) != 1)
374 if (bfd_get_error () != bfd_error_file_truncated
)
382 /* Scan the wasm module ABFD, creating sections and symbols.
383 Return TRUE on success. */
386 wasm_scan (bfd
*abfd
)
389 /* Fake VMAs for now. Choose 0x80000000 as base to avoid clashes
390 with actual data addresses. */
391 bfd_vma vma
= 0x80000000;
393 unsigned int bytes_read
;
396 if (bfd_seek (abfd
, (file_ptr
) 0, SEEK_SET
) != 0)
399 if (!wasm_read_header (abfd
, &error
))
402 while ((section_code
= wasm_read_byte (abfd
, &error
)) != EOF
)
404 if (section_code
!= 0)
406 const char *sname
= wasm_section_code_to_name (section_code
);
411 bfdsec
= bfd_make_section_anyway_with_flags (abfd
, sname
,
416 bfdsec
->size
= wasm_read_leb128 (abfd
, &error
, &bytes_read
, false);
425 char *prefix
= WASM_SECTION_PREFIX
;
426 size_t prefixlen
= strlen (prefix
);
429 payload_len
= wasm_read_leb128 (abfd
, &error
, &bytes_read
, false);
432 namelen
= wasm_read_leb128 (abfd
, &error
, &bytes_read
, false);
433 if (error
|| bytes_read
> payload_len
434 || namelen
> payload_len
- bytes_read
)
436 payload_len
-= namelen
+ bytes_read
;
437 filesize
= bfd_get_file_size (abfd
);
438 if (filesize
!= 0 && namelen
> filesize
)
440 bfd_set_error (bfd_error_file_truncated
);
443 name
= bfd_alloc (abfd
, namelen
+ prefixlen
+ 1);
446 memcpy (name
, prefix
, prefixlen
);
447 if (bfd_bread (name
+ prefixlen
, namelen
, abfd
) != namelen
)
449 name
[prefixlen
+ namelen
] = 0;
451 bfdsec
= bfd_make_section_anyway_with_flags (abfd
, name
,
456 bfdsec
->size
= payload_len
;
461 bfdsec
->alignment_power
= 0;
462 bfdsec
->filepos
= bfd_tell (abfd
);
463 if (bfdsec
->size
!= 0)
465 bfdsec
->contents
= _bfd_alloc_and_read (abfd
, bfdsec
->size
,
467 if (!bfdsec
->contents
)
474 /* Make sure we're at actual EOF. There's no indication in the
475 WebAssembly format of how long the file is supposed to be. */
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 (startswith (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 size_t 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. */
745 wasm_object_p (bfd
*abfd
)
750 if (bfd_seek (abfd
, (file_ptr
) 0, SEEK_SET
) != 0)
753 if (!wasm_read_header (abfd
, &error
))
755 bfd_set_error (bfd_error_wrong_format
);
759 if (!wasm_mkobject (abfd
))
762 if (!wasm_scan (abfd
)
763 || !bfd_default_set_arch_mach (abfd
, bfd_arch_wasm32
, 0))
765 bfd_release (abfd
, abfd
->tdata
.any
);
766 abfd
->tdata
.any
= NULL
;
770 s
= bfd_get_section_by_name (abfd
, WASM_NAME_SECTION
);
771 if (s
!= NULL
&& wasm_scan_name_function_section (abfd
, s
))
772 abfd
->flags
|= HAS_SYMS
;
774 return _bfd_no_cleanup
;
777 /* BFD_JUMP_TABLE_WRITE */
778 #define wasm_set_arch_mach _bfd_generic_set_arch_mach
780 /* BFD_JUMP_TABLE_SYMBOLS */
781 #define wasm_get_symbol_version_string _bfd_nosymbols_get_symbol_version_string
782 #define wasm_bfd_is_local_label_name bfd_generic_is_local_label_name
783 #define wasm_bfd_is_target_special_symbol _bfd_bool_bfd_asymbol_false
784 #define wasm_get_lineno _bfd_nosymbols_get_lineno
785 #define wasm_find_nearest_line _bfd_nosymbols_find_nearest_line
786 #define wasm_find_line _bfd_nosymbols_find_line
787 #define wasm_find_inliner_info _bfd_nosymbols_find_inliner_info
788 #define wasm_bfd_make_debug_symbol _bfd_nosymbols_bfd_make_debug_symbol
789 #define wasm_read_minisymbols _bfd_generic_read_minisymbols
790 #define wasm_minisymbol_to_symbol _bfd_generic_minisymbol_to_symbol
792 const bfd_target wasm_vec
=
795 bfd_target_unknown_flavour
,
798 (HAS_SYMS
| WP_TEXT
), /* Object flags. */
799 (SEC_CODE
| SEC_DATA
| SEC_HAS_CONTENTS
), /* Section flags. */
800 0, /* Leading underscore. */
801 ' ', /* AR_pad_char. */
802 255, /* AR_max_namelen. */
803 0, /* Match priority. */
804 TARGET_KEEP_UNUSED_SECTION_SYMBOLS
, /* keep unused section symbols. */
805 /* Routines to byte-swap various sized integers from the data sections. */
806 bfd_getl64
, bfd_getl_signed_64
, bfd_putl64
,
807 bfd_getl32
, bfd_getl_signed_32
, bfd_putl32
,
808 bfd_getl16
, bfd_getl_signed_16
, bfd_putl16
,
810 /* Routines to byte-swap various sized integers from the file headers. */
811 bfd_getl64
, bfd_getl_signed_64
, bfd_putl64
,
812 bfd_getl32
, bfd_getl_signed_32
, bfd_putl32
,
813 bfd_getl16
, bfd_getl_signed_16
, bfd_putl16
,
817 wasm_object_p
, /* bfd_check_format. */
822 _bfd_bool_bfd_false_error
,
824 _bfd_generic_mkarchive
,
825 _bfd_bool_bfd_false_error
,
827 { /* bfd_write_contents. */
828 _bfd_bool_bfd_false_error
,
829 wasm_write_object_contents
,
830 _bfd_write_archive_contents
,
831 _bfd_bool_bfd_false_error
,
834 BFD_JUMP_TABLE_GENERIC (_bfd_generic
),
835 BFD_JUMP_TABLE_COPY (_bfd_generic
),
836 BFD_JUMP_TABLE_CORE (_bfd_nocore
),
837 BFD_JUMP_TABLE_ARCHIVE (_bfd_noarchive
),
838 BFD_JUMP_TABLE_SYMBOLS (wasm
),
839 BFD_JUMP_TABLE_RELOCS (_bfd_norelocs
),
840 BFD_JUMP_TABLE_WRITE (wasm
),
841 BFD_JUMP_TABLE_LINK (_bfd_nolink
),
842 BFD_JUMP_TABLE_DYNAMIC (_bfd_nodynamic
),