3 * Copyright (C) 2009-2010 Red Hat Inc.
5 * This program is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 2 of the License, or
8 * (at your option) any later version.
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
15 * You should have received a copy of the GNU General Public License
16 * along with this program; if not, write to the Free Software
17 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20 (* This script generates language bindings and some documentation for
23 * After editing this file, run it (./generator/generator.ml) to
24 * regenerate all the output files. 'make' will rerun this
25 * automatically when necessary. Note that if you are using a separate
26 * build directory you must run generator.ml from the _source_
29 * IMPORTANT: This script should NOT print any warnings. If it prints
30 * warnings, you should treat them as errors.
32 * OCaml tips: (1) In emacs, install tuareg-mode to display and format
33 * OCaml code correctly. 'vim' comes with a good OCaml editing mode by
34 * default. (2) Read the resources at http://ocaml-tutorial.org/
39 #directory "+xml-light";;
40 #load "xml-light.cma";;
45 type style = ret * args
47 | RErr (* 0 = ok, -1 = error *)
48 | RHive (* Returns a hive_h or NULL. *)
49 | RNode (* Returns hive_node_h or 0. *)
50 | RNodeList (* Returns hive_node_h* or NULL. *)
51 | RValue (* Returns hive_value_h or 0. *)
52 | RValueList (* Returns hive_value_h* or NULL. *)
53 | RString (* Returns char* or NULL. *)
54 | RStringList (* Returns char** or NULL. *)
55 | RLenType (* See hivex_value_type. *)
56 | RLenTypeVal (* See hivex_value_value. *)
57 | RInt32 (* Returns int32. *)
58 | RInt64 (* Returns int64. *)
60 and args = argt list (* List of parameters. *)
62 and argt = (* Note, cannot be NULL/0 unless it
63 says so explicitly below. *)
65 | ANode of string (* hive_node_h *)
66 | AValue of string (* hive_value_h *)
67 | AString of string (* char* *)
68 | AStringNullable of string (* char* (can be NULL) *)
69 | AOpenFlags (* HIVEX_OPEN_* flags list. *)
70 | AUnusedFlags (* Flags arg that is always 0 *)
71 | ASetValues (* See hivex_node_set_values. *)
74 * https://secure.wikimedia.org/wikipedia/en/wiki/Windows_Registry#Keys_and_values
76 * It's unfortunate that in our original C binding we strayed away from
77 * the names that Windows uses (eg. REG_SZ for strings). We include
78 * both our names and the Windows names.
82 "Just a key without a value";
84 "A Windows string (encoding is unknown, but often UTF16-LE)";
85 2, "expand_string", "EXPAND_SZ",
86 "A Windows string that contains %env% (environment variable expansion)";
87 3, "binary", "BINARY",
90 "DWORD (32 bit integer), little endian";
91 5, "dword_be", "DWORD_BIG_ENDIAN",
92 "DWORD (32 bit integer), big endian";
94 "Symbolic link to another part of the registry tree";
95 7, "multiple_strings", "MULTI_SZ",
96 "Multiple Windows strings. See http://blogs.msdn.com/oldnewthing/archive/2009/10/08/9904646.aspx";
97 8, "resource_list", "RESOURCE_LIST",
99 9, "full_resource_description", "FULL_RESOURCE_DESCRIPTOR",
100 "Resource descriptor";
101 10, "resource_requirements_list", "RESOURCE_REQUIREMENTS_LIST",
102 "Resouce requirements list";
103 11, "qword", "QWORD",
104 "QWORD (64 bit integer), unspecified endianness but usually little endian"
107 (* Open flags (bitmask passed to AOpenFlags) *)
109 1, "VERBOSE", "Verbose messages";
110 2, "DEBUG", "Debug messages";
111 4, "WRITE", "Enable writes to the hive";
116 "open", (RHive, [AString "filename"; AOpenFlags]),
119 Opens the hive named C<filename> for reading.
121 Flags is an ORed list of the open flags (or C<0> if you don't
122 want to pass any flags). These flags are defined:
126 =item HIVEX_OPEN_VERBOSE
130 =item HIVEX_OPEN_DEBUG
132 Very verbose messages, suitable for debugging problems in the library
135 This is also selected if the C<HIVEX_DEBUG> environment variable
138 =item HIVEX_OPEN_WRITE
140 Open the hive for writing. If omitted, the hive is read-only.
142 See L<hivex(3)/WRITING TO HIVE FILES>.
146 "close", (RErr, [AHive]),
147 "close a hive handle",
149 Close a hive handle and free all associated resources.
151 Note that any uncommitted writes are I<not> committed by this call,
152 but instead are lost. See L<hivex(3)/WRITING TO HIVE FILES>.";
154 "root", (RNode, [AHive]),
155 "return the root node of the hive",
157 Return root node of the hive. All valid registries must contain
160 "node_name", (RString, [AHive; ANode "node"]),
161 "return the name of the node",
163 Return the name of the node.
165 Note that the name of the root node is a dummy, such as
166 C<$$$PROTO.HIV> (other names are possible: it seems to depend on the
167 tool or program that created the hive in the first place). You can
168 only know the \"real\" name of the root node by knowing which registry
169 file this hive originally comes from, which is knowledge that is
170 outside the scope of this library.";
172 "node_children", (RNodeList, [AHive; ANode "node"]),
173 "return children of node",
175 Return an array of nodes which are the subkeys
176 (children) of C<node>.";
178 "node_get_child", (RNode, [AHive; ANode "node"; AString "name"]),
179 "return named child of node",
181 Return the child of node with the name C<name>, if it exists.
183 The name is matched case insensitively.
185 If the child node does not exist, this returns 0 without
188 "node_parent", (RNode, [AHive; ANode "node"]),
189 "return the parent of node",
191 Return the parent of C<node>.
193 The parent pointer of the root node in registry files that we
194 have examined seems to be invalid, and so this function will
195 return an error if called on the root node.";
197 "node_values", (RValueList, [AHive; ANode "node"]),
198 "return (key, value) pairs attached to a node",
200 Return the array of (key, value) pairs attached to this node.";
202 "node_get_value", (RValue, [AHive; ANode "node"; AString "key"]),
203 "return named key at node",
205 Return the value attached to this node which has the name C<key>,
208 The key name is matched case insensitively.
210 Note that to get the default key, you should pass the empty
211 string C<\"\"> here. The default key is often written C<\"@\">, but
212 inside hives that has no meaning and won't give you the
215 "value_key", (RString, [AHive; AValue "val"]),
216 "return the key of a (key, value) pair",
218 Return the key (name) of a (key, value) pair. The name
219 is reencoded as UTF-8 and returned as a string.
221 The string should be freed by the caller when it is no longer needed.
223 Note that this function can return a zero-length string. In the
224 context of Windows Registries, this means that this value is the
225 default key for this node in the tree. This is usually written
228 "value_type", (RLenType, [AHive; AValue "val"]),
229 "return data length and data type of a value",
231 Return the data length and data type of the value in this (key, value)
232 pair. See also C<hivex_value_value> which returns all this
233 information, and the value itself. Also, C<hivex_value_*> functions
234 below which can be used to return the value in a more useful form when
235 you know the type in advance.";
237 "value_value", (RLenTypeVal, [AHive; AValue "val"]),
238 "return data length, data type and data of a value",
240 Return the value of this (key, value) pair. The value should
241 be interpreted according to its type (see C<hive_type>).";
243 "value_string", (RString, [AHive; AValue "val"]),
244 "return value as a string",
246 If this value is a string, return the string reencoded as UTF-8
247 (as a C string). This only works for values which have type
248 C<hive_t_string>, C<hive_t_expand_string> or C<hive_t_link>.";
250 "value_multiple_strings", (RStringList, [AHive; AValue "val"]),
251 "return value as multiple strings",
253 If this value is a multiple-string, return the strings reencoded
254 as UTF-8 (as a NULL-terminated array of C strings). This only
255 works for values which have type C<hive_t_multiple_strings>.";
257 "value_dword", (RInt32, [AHive; AValue "val"]),
258 "return value as a DWORD",
260 If this value is a DWORD (Windows int32), return it. This only works
261 for values which have type C<hive_t_dword> or C<hive_t_dword_be>.";
263 "value_qword", (RInt64, [AHive; AValue "val"]),
264 "return value as a QWORD",
266 If this value is a QWORD (Windows int64), return it. This only
267 works for values which have type C<hive_t_qword>.";
269 "commit", (RErr, [AHive; AStringNullable "filename"; AUnusedFlags]),
270 "commit (write) changes to file",
272 Commit (write) any changes which have been made.
274 C<filename> is the new file to write. If C<filename> is NULL then we
275 overwrite the original file (ie. the file name that was passed to
276 C<hivex_open>). C<flags> is not used, always pass 0.
278 Note this does not close the hive handle. You can perform further
279 operations on the hive after committing, including making more
280 modifications. If you no longer wish to use the hive, call
281 C<hivex_close> after this.";
283 "node_add_child", (RNode, [AHive; ANode "parent"; AString "name"]),
286 Add a new child node named C<name> to the existing node C<parent>.
287 The new child initially has no subnodes and contains no keys or
288 values. The sk-record (security descriptor) is inherited from
291 The parent must not have an existing child called C<name>, so if you
292 want to overwrite an existing child, call C<hivex_node_delete_child>
295 "node_delete_child", (RErr, [AHive; ANode "node"]),
298 Delete the node C<node>. All values at the node and all subnodes are
299 deleted (recursively). The C<node> handle and the handles of all
300 subnodes become invalid. You cannot delete the root node.";
302 "node_set_values", (RErr, [AHive; ANode "node"; ASetValues; AUnusedFlags]),
303 "set (key, value) pairs at a node",
305 This call can be used to set all the (key, value) pairs stored in C<node>.
307 C<node> is the node to modify. C<values> is an array of (key, value)
308 pairs. There should be C<nr_values> elements in this array. C<flags>
309 is not used, always pass 0.
311 Any existing values stored at the node are discarded, and their
312 C<hive_value_h> handles become invalid. Thus you can remove all
313 values stored at C<node> by passing C<nr_values = 0>.
315 Note that this library does not offer a way to modify just a single
316 key at a node. We don't implement a way to do this efficiently.";
319 (* Used to memoize the result of pod2text. *)
320 let pod2text_memo_filename = "generator/.pod2text.data"
321 let pod2text_memo : ((int * string * string), string list) Hashtbl.t =
323 let chan = open_in pod2text_memo_filename in
324 let v = input_value chan in
328 _ -> Hashtbl.create 13
329 let pod2text_memo_updated () =
330 let chan = open_out pod2text_memo_filename in
331 output_value chan pod2text_memo;
335 * Note we don't want to use any external OCaml libraries which
336 * makes this a bit harder than it should be.
338 module StringMap = Map.Make (String)
340 let failwithf fs = ksprintf failwith fs
342 let unique = let i = ref 0 in fun () -> incr i; !i
344 let replace_char s c1 c2 =
345 let s2 = String.copy s in
347 for i = 0 to String.length s2 - 1 do
348 if String.unsafe_get s2 i = c1 then (
349 String.unsafe_set s2 i c2;
353 if not !r then s else s2
357 (* || c = '\f' *) || c = '\n' || c = '\r' || c = '\t' (* || c = '\v' *)
359 let triml ?(test = isspace) str =
361 let n = ref (String.length str) in
362 while !n > 0 && test str.[!i]; do
367 else String.sub str !i !n
369 let trimr ?(test = isspace) str =
370 let n = ref (String.length str) in
371 while !n > 0 && test str.[!n-1]; do
374 if !n = String.length str then str
375 else String.sub str 0 !n
377 let trim ?(test = isspace) str =
378 trimr ~test (triml ~test str)
381 let len = String.length s in
382 let sublen = String.length sub in
384 if i <= len-sublen then (
387 if s.[i+j] = sub.[j] then loop2 (j+1)
393 if r = -1 then loop (i+1) else r
399 let rec replace_str s s1 s2 =
400 let len = String.length s in
401 let sublen = String.length s1 in
405 let s' = String.sub s 0 i in
406 let s'' = String.sub s (i+sublen) (len-i-sublen) in
407 s' ^ s2 ^ replace_str s'' s1 s2
410 let rec string_split sep str =
411 let len = String.length str in
412 let seplen = String.length sep in
413 let i = find str sep in
416 let s' = String.sub str 0 i in
417 let s'' = String.sub str (i+seplen) (len-i-seplen) in
418 s' :: string_split sep s''
421 let files_equal n1 n2 =
422 let cmd = sprintf "cmp -s %s %s" (Filename.quote n1) (Filename.quote n2) in
423 match Sys.command cmd with
426 | i -> failwithf "%s: failed with error code %d" cmd i
428 let rec filter_map f = function
432 | Some y -> y :: filter_map f xs
433 | None -> filter_map f xs
435 let rec find_map f = function
436 | [] -> raise Not_found
440 | None -> find_map f xs
443 let rec loop i = function
445 | x :: xs -> f i x; loop (i+1) xs
450 let rec loop i = function
452 | x :: xs -> let r = f i x in r :: loop (i+1) xs
456 let count_chars c str =
458 for i = 0 to String.length str - 1 do
459 if c = String.unsafe_get str i then incr count
463 let name_of_argt = function
465 | ANode n | AValue n | AString n | AStringNullable n -> n
466 | AOpenFlags | AUnusedFlags -> "flags"
467 | ASetValues -> "values"
469 (* Check function names etc. for consistency. *)
470 let check_functions () =
471 let contains_uppercase str =
472 let len = String.length str in
474 if i >= len then false
477 if c >= 'A' && c <= 'Z' then true
484 (* Check function names. *)
486 fun (name, _, _, _) ->
487 if String.length name >= 7 && String.sub name 0 7 = "hivex" then
488 failwithf "function name %s does not need 'hivex' prefix" name;
490 failwithf "function name is empty";
491 if name.[0] < 'a' || name.[0] > 'z' then
492 failwithf "function name %s must start with lowercase a-z" name;
493 if String.contains name '-' then
494 failwithf "function name %s should not contain '-', use '_' instead."
498 (* Check function parameter/return names. *)
500 fun (name, style, _, _) ->
501 let check_arg_ret_name n =
502 if contains_uppercase n then
503 failwithf "%s param/ret %s should not contain uppercase chars"
505 if String.contains n '-' || String.contains n '_' then
506 failwithf "%s param/ret %s should not contain '-' or '_'"
509 failwithf "%s has a param/ret called 'value', which causes conflicts in the OCaml bindings, use something like 'val' or a more descriptive name" name;
510 if n = "int" || n = "char" || n = "short" || n = "long" then
511 failwithf "%s has a param/ret which conflicts with a C type (eg. 'int', 'char' etc.)" name;
512 if n = "i" || n = "n" then
513 failwithf "%s has a param/ret called 'i' or 'n', which will cause some conflicts in the generated code" name;
514 if n = "argv" || n = "args" then
515 failwithf "%s has a param/ret called 'argv' or 'args', which will cause some conflicts in the generated code" name;
517 (* List Haskell, OCaml and C keywords here.
518 * http://www.haskell.org/haskellwiki/Keywords
519 * http://caml.inria.fr/pub/docs/manual-ocaml/lex.html#operator-char
520 * http://en.wikipedia.org/wiki/C_syntax#Reserved_keywords
521 * Formatted via: cat c haskell ocaml|sort -u|grep -vE '_|^val$' \
522 * |perl -pe 's/(.+)/"$1";/'|fmt -70
523 * Omitting _-containing words, since they're handled above.
524 * Omitting the OCaml reserved word, "val", is ok,
525 * and saves us from renaming several parameters.
528 "and"; "as"; "asr"; "assert"; "auto"; "begin"; "break"; "case";
529 "char"; "class"; "const"; "constraint"; "continue"; "data";
530 "default"; "deriving"; "do"; "done"; "double"; "downto"; "else";
531 "end"; "enum"; "exception"; "extern"; "external"; "false"; "float";
532 "for"; "forall"; "foreign"; "fun"; "function"; "functor"; "goto";
533 "hiding"; "if"; "import"; "in"; "include"; "infix"; "infixl";
534 "infixr"; "inherit"; "initializer"; "inline"; "instance"; "int";
536 "land"; "lazy"; "let"; "long"; "lor"; "lsl"; "lsr"; "lxor";
537 "match"; "mdo"; "method"; "mod"; "module"; "mutable"; "new";
538 "newtype"; "object"; "of"; "open"; "or"; "private"; "qualified";
539 "rec"; "register"; "restrict"; "return"; "short"; "sig"; "signed";
540 "sizeof"; "static"; "struct"; "switch"; "then"; "to"; "true"; "try";
541 "type"; "typedef"; "union"; "unsigned"; "virtual"; "void";
542 "volatile"; "when"; "where"; "while";
544 if List.mem n reserved then
545 failwithf "%s has param/ret using reserved word %s" name n;
548 List.iter (fun arg -> check_arg_ret_name (name_of_argt arg)) (snd style)
551 (* Check short descriptions. *)
553 fun (name, _, shortdesc, _) ->
554 if shortdesc.[0] <> Char.lowercase shortdesc.[0] then
555 failwithf "short description of %s should begin with lowercase." name;
556 let c = shortdesc.[String.length shortdesc-1] in
557 if c = '\n' || c = '.' then
558 failwithf "short description of %s should not end with . or \\n." name
561 (* Check long dscriptions. *)
563 fun (name, _, _, longdesc) ->
564 if longdesc.[String.length longdesc-1] = '\n' then
565 failwithf "long description of %s should not end with \\n." name
568 (* 'pr' prints to the current output file. *)
569 let chan = ref Pervasives.stdout
574 let i = count_chars '\n' str in
576 output_string !chan str
579 let copyright_years =
580 let this_year = 1900 + (localtime (time ())).tm_year in
581 if this_year > 2009 then sprintf "2009-%04d" this_year else "2009"
583 (* Generate a header block in a number of standard styles. *)
585 CStyle | CPlusPlusStyle | HashStyle | OCamlStyle | HaskellStyle
586 type license = GPLv2plus | LGPLv2plus | GPLv2 | LGPLv2
588 let generate_header ?(extra_inputs = []) comment license =
589 let inputs = "generator/generator.ml" :: extra_inputs in
590 let c = match comment with
591 | CStyle -> pr "/* "; " *"
592 | CPlusPlusStyle -> pr "// "; "//"
593 | HashStyle -> pr "# "; "#"
594 | OCamlStyle -> pr "(* "; " *"
595 | HaskellStyle -> pr "{- "; " " in
596 pr "hivex generated file\n";
597 pr "%s WARNING: THIS FILE IS GENERATED FROM:\n" c;
598 List.iter (pr "%s %s\n" c) inputs;
599 pr "%s ANY CHANGES YOU MAKE TO THIS FILE WILL BE LOST.\n" c;
601 pr "%s Copyright (C) %s Red Hat Inc.\n" c copyright_years;
602 pr "%s Derived from code by Petter Nordahl-Hagen under a compatible license:\n" c;
603 pr "%s Copyright (c) 1997-2007 Petter Nordahl-Hagen.\n" c;
604 pr "%s Derived from code by Markus Stephany under a compatible license:\n" c;
605 pr "%s Copyright (c)2000-2004, Markus Stephany.\n" c;
609 pr "%s This program is free software; you can redistribute it and/or modify\n" c;
610 pr "%s it under the terms of the GNU General Public License as published by\n" c;
611 pr "%s the Free Software Foundation; either version 2 of the License, or\n" c;
612 pr "%s (at your option) any later version.\n" c;
614 pr "%s This program is distributed in the hope that it will be useful,\n" c;
615 pr "%s but WITHOUT ANY WARRANTY; without even the implied warranty of\n" c;
616 pr "%s MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the\n" c;
617 pr "%s GNU General Public License for more details.\n" c;
619 pr "%s You should have received a copy of the GNU General Public License along\n" c;
620 pr "%s with this program; if not, write to the Free Software Foundation, Inc.,\n" c;
621 pr "%s 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.\n" c;
624 pr "%s This library is free software; you can redistribute it and/or\n" c;
625 pr "%s modify it under the terms of the GNU Lesser General Public\n" c;
626 pr "%s License as published by the Free Software Foundation; either\n" c;
627 pr "%s version 2 of the License, or (at your option) any later version.\n" c;
629 pr "%s This library is distributed in the hope that it will be useful,\n" c;
630 pr "%s but WITHOUT ANY WARRANTY; without even the implied warranty of\n" c;
631 pr "%s MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU\n" c;
632 pr "%s Lesser General Public License for more details.\n" c;
634 pr "%s You should have received a copy of the GNU Lesser General Public\n" c;
635 pr "%s License along with this library; if not, write to the Free Software\n" c;
636 pr "%s Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA\n" c;
639 pr "%s This program is free software; you can redistribute it and/or modify\n" c;
640 pr "%s it under the terms of the GNU General Public License as published by\n" c;
641 pr "%s the Free Software Foundation; version 2 of the License only.\n" c;
643 pr "%s This program is distributed in the hope that it will be useful,\n" c;
644 pr "%s but WITHOUT ANY WARRANTY; without even the implied warranty of\n" c;
645 pr "%s MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the\n" c;
646 pr "%s GNU General Public License for more details.\n" c;
648 pr "%s You should have received a copy of the GNU General Public License along\n" c;
649 pr "%s with this program; if not, write to the Free Software Foundation, Inc.,\n" c;
650 pr "%s 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.\n" c;
653 pr "%s This library is free software; you can redistribute it and/or\n" c;
654 pr "%s modify it under the terms of the GNU Lesser General Public\n" c;
655 pr "%s License as published by the Free Software Foundation; either\n" c;
656 pr "%s version 2.1 of the License only.\n" c;
658 pr "%s This library is distributed in the hope that it will be useful,\n" c;
659 pr "%s but WITHOUT ANY WARRANTY; without even the implied warranty of\n" c;
660 pr "%s MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU\n" c;
661 pr "%s Lesser General Public License for more details.\n" c;
663 pr "%s You should have received a copy of the GNU Lesser General Public\n" c;
664 pr "%s License along with this library; if not, write to the Free Software\n" c;
665 pr "%s Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA\n" c;
668 | CStyle -> pr " */\n"
671 | OCamlStyle -> pr " *)\n"
672 | HaskellStyle -> pr "-}\n"
676 (* Start of main code generation functions below this line. *)
678 let rec generate_c_header () =
679 generate_header CStyle LGPLv2;
691 /* NOTE: This API is documented in the man page hivex(3). */
694 typedef struct hive_h hive_h;
696 /* Nodes and values. */
697 typedef size_t hive_node_h;
698 typedef size_t hive_value_h;
700 /* Pre-defined types. */
704 fun (t, old_style, new_style, description) ->
705 pr " /* %s */\n" description;
706 pr " hive_t_REG_%s,\n" new_style;
707 pr "#define hive_t_%s hive_t_REG_%s\n" old_style new_style;
713 typedef enum hive_type hive_type;
715 /* Bitmask of flags passed to hivex_open. */
718 fun (v, flag, description) ->
719 pr " /* %s */\n" description;
720 pr "#define HIVEX_OPEN_%-10s %d\n" flag v;
725 /* Array of (key, value) pairs passed to hivex_node_set_values. */
726 struct hive_set_value {
732 typedef struct hive_set_value hive_set_value;
735 (* Function declarations. *)
737 fun (shortname, style, _, _) ->
738 let name = "hivex_" ^ shortname in
739 generate_c_prototype ~extern:true name style
742 (* The visitor pattern. *)
744 /* Visit all nodes. This is specific to the C API and is not made
745 * available to other languages. This is because of the complexity
746 * of binding callbacks in other languages, but also because other
747 * languages make it much simpler to iterate over a tree.
749 struct hivex_visitor {
750 int (*node_start) (hive_h *, void *opaque, hive_node_h, const char *name);
751 int (*node_end) (hive_h *, void *opaque, hive_node_h, const char *name);
752 int (*value_string) (hive_h *, void *opaque, hive_node_h, hive_value_h, hive_type t, size_t len, const char *key, const char *str);
753 int (*value_multiple_strings) (hive_h *, void *opaque, hive_node_h, hive_value_h, hive_type t, size_t len, const char *key, char **argv);
754 int (*value_string_invalid_utf16) (hive_h *, void *opaque, hive_node_h, hive_value_h, hive_type t, size_t len, const char *key, const char *str);
755 int (*value_dword) (hive_h *, void *opaque, hive_node_h, hive_value_h, hive_type t, size_t len, const char *key, int32_t);
756 int (*value_qword) (hive_h *, void *opaque, hive_node_h, hive_value_h, hive_type t, size_t len, const char *key, int64_t);
757 int (*value_binary) (hive_h *, void *opaque, hive_node_h, hive_value_h, hive_type t, size_t len, const char *key, const char *value);
758 int (*value_none) (hive_h *, void *opaque, hive_node_h, hive_value_h, hive_type t, size_t len, const char *key, const char *value);
759 int (*value_other) (hive_h *, void *opaque, hive_node_h, hive_value_h, hive_type t, size_t len, const char *key, const char *value);
760 int (*value_any) (hive_h *, void *opaque, hive_node_h, hive_value_h, hive_type t, size_t len, const char *key, const char *value);
763 #define HIVEX_VISIT_SKIP_BAD 1
765 extern int hivex_visit (hive_h *h, const struct hivex_visitor *visitor, size_t len, void *opaque, int flags);
766 extern int hivex_visit_node (hive_h *h, hive_node_h node, const struct hivex_visitor *visitor, size_t len, void *opaque, int flags);
770 (* Finish the header file. *)
776 #endif /* HIVEX_H_ */
779 and generate_c_prototype ?(extern = false) name style =
780 if extern then pr "extern ";
781 (match fst style with
783 | RHive -> pr "hive_h *"
784 | RNode -> pr "hive_node_h "
785 | RNodeList -> pr "hive_node_h *"
786 | RValue -> pr "hive_value_h "
787 | RValueList -> pr "hive_value_h *"
788 | RString -> pr "char *"
789 | RStringList -> pr "char **"
790 | RLenType -> pr "int "
791 | RLenTypeVal -> pr "char *"
792 | RInt32 -> pr "int32_t "
793 | RInt64 -> pr "int64_t "
796 let comma = ref false in
799 if !comma then pr ", "; comma := true;
801 | AHive -> pr "hive_h *h"
802 | ANode n -> pr "hive_node_h %s" n
803 | AValue n -> pr "hive_value_h %s" n
804 | AString n | AStringNullable n -> pr "const char *%s" n
805 | AOpenFlags | AUnusedFlags -> pr "int flags"
806 | ASetValues -> pr "size_t nr_values, const hive_set_value *values"
808 (match fst style with
809 | RLenType | RLenTypeVal -> pr ", hive_type *t, size_t *len"
814 and generate_c_pod () =
820 hivex - Windows Registry \"hive\" extraction library
824 hive_h *hivex_open (const char *filename, int flags);
825 int hivex_close (hive_h *h);
829 libhivex is a library for extracting the contents of Windows Registry
830 \"hive\" files. It is designed to be secure against buggy or malicious
833 Unlike many other tools in this area, it doesn't use the textual .REG
834 format for output, because parsing that is as much trouble as parsing
835 the original binary format. Instead it makes the file available
836 through a C API, or there is a separate program to export the hive as
837 XML (see L<hivexml(1)>), or to get individual keys (see
846 This handle describes an open hive file.
850 This is a node handle, an integer but opaque outside the library.
851 Valid node handles cannot be 0. The library returns 0 in some
852 situations to indicate an error.
856 The enum below describes the possible types for the value(s)
857 stored at each node. Note that you should not trust the
858 type field in a Windows Registry, as it very often has no
859 relationship to reality. Some applications use their own
860 types. The encoding of strings is not specified. Some
861 programs store everything (including strings) in binary blobs.
866 fun (t, _, new_style, description) ->
867 pr " /* %s */\n" description;
868 pr " hive_t_REG_%s = %d,\n" new_style t
875 This is a value handle, an integer but opaque outside the library.
876 Valid value handles cannot be 0. The library returns 0 in some
877 situations to indicate an error.
881 The typedef C<hive_set_value> is used in conjunction with the
882 C<hivex_node_set_values> call described below.
884 struct hive_set_value {
885 char *key; /* key - a UTF-8 encoded ASCIIZ string */
886 hive_type t; /* type of value field */
887 size_t len; /* length of value field in bytes */
888 char *value; /* value field */
890 typedef struct hive_set_value hive_set_value;
892 To set the default value for a node, you have to pass C<key = \"\">.
894 Note that the C<value> field is just treated as a list of bytes, and
895 is stored directly in the hive. The caller has to ensure correct
896 encoding and endianness, for example converting dwords to little
899 The correct type and encoding for values depends on the node and key
900 in the registry, the version of Windows, and sometimes even changes
901 between versions of Windows for the same key. We don't document it
902 here. Often it's not documented at all.
912 fun (shortname, style, _, longdesc) ->
913 let name = "hivex_" ^ shortname in
914 pr "=item %s\n" name;
916 generate_c_prototype ~extern:false name style;
920 (match fst style with
923 Returns 0 on success.
924 On error this returns -1 and sets errno.\n\n"
927 Returns a new hive handle.
928 On error this returns NULL and sets errno.\n\n"
931 Returns a node handle.
932 On error this returns 0 and sets errno.\n\n"
935 Returns a 0-terminated array of nodes.
936 The array must be freed by the caller when it is no longer needed.
937 On error this returns NULL and sets errno.\n\n"
940 Returns a value handle.
941 On error this returns 0 and sets errno.\n\n"
944 Returns a 0-terminated array of values.
945 The array must be freed by the caller when it is no longer needed.
946 On error this returns NULL and sets errno.\n\n"
950 The string must be freed by the caller when it is no longer needed.
951 On error this returns NULL and sets errno.\n\n"
954 Returns a NULL-terminated array of C strings.
955 The strings and the array must all be freed by the caller when
956 they are no longer needed.
957 On error this returns NULL and sets errno.\n\n"
960 Returns 0 on success.
961 On error this returns NULL and sets errno.\n\n"
964 The value is returned as an array of bytes (of length C<len>).
965 The value must be freed by the caller when it is no longer needed.
966 On error this returns NULL and sets errno.\n\n"
967 | RInt32 | RInt64 -> ()
974 =head2 WRITING TO HIVE FILES
976 The hivex library supports making limited modifications to hive files.
977 We have tried to implement this very conservatively in order to reduce
978 the chance of corrupting your registry. However you should be careful
979 and take back-ups, since Microsoft has never documented the hive
980 format, and so it is possible there are nuances in the
981 reverse-engineered format that we do not understand.
983 To be able to modify a hive, you must pass the C<HIVEX_OPEN_WRITE>
984 flag to C<hivex_open>, otherwise any write operation will return with
987 The write operations shown below do not modify the on-disk file
988 immediately. You must call C<hivex_commit> in order to write the
989 changes to disk. If you call C<hivex_close> without committing then
990 any writes are discarded.
992 Hive files internally consist of a \"memory dump\" of binary blocks
993 (like the C heap), and some of these blocks can be unused. The hivex
994 library never reuses these unused blocks. Instead, to ensure
995 robustness in the face of the partially understood on-disk format,
996 hivex only allocates new blocks after the end of the file, and makes
997 minimal modifications to existing structures in the file to point to
998 these new blocks. This makes hivex slightly less disk-efficient than
999 it could be, but disk is cheap, and registry modifications tend to be
1002 When deleting nodes, it is possible that this library may leave
1003 unreachable live blocks in the hive. This is because certain parts of
1004 the hive disk format such as security (sk) records and big data (db)
1005 records and classname fields are not well understood (and not
1006 documented at all) and we play it safe by not attempting to modify
1007 them. Apart from wasting a little bit of disk space, it is not
1008 thought that unreachable blocks are a problem.
1010 =head3 WRITE OPERATIONS WHICH ARE NOT SUPPORTED
1016 Changing the root node.
1020 Creating a new hive file from scratch. This is impossible at present
1021 because not all fields in the header are understood.
1025 Modifying or deleting single values at a node.
1029 Modifying security key (sk) records or classnames.
1030 Previously we did not understand these records. However now they
1031 are well-understood and we could add support if it was required
1032 (but nothing much really uses them).
1036 =head2 VISITING ALL NODES
1038 The visitor pattern is useful if you want to visit all nodes
1039 in the tree or all nodes below a certain point in the tree.
1041 First you set up your own C<struct hivex_visitor> with your
1044 Each of these callback functions should return 0 on success or -1
1045 on error. If any callback returns -1, then the entire visit
1046 terminates immediately. If you don't need a callback function at
1047 all, set the function pointer to NULL.
1049 struct hivex_visitor {
1050 int (*node_start) (hive_h *, void *opaque, hive_node_h, const char *name);
1051 int (*node_end) (hive_h *, void *opaque, hive_node_h, const char *name);
1052 int (*value_string) (hive_h *, void *opaque, hive_node_h, hive_value_h,
1053 hive_type t, size_t len, const char *key, const char *str);
1054 int (*value_multiple_strings) (hive_h *, void *opaque, hive_node_h,
1055 hive_value_h, hive_type t, size_t len, const char *key, char **argv);
1056 int (*value_string_invalid_utf16) (hive_h *, void *opaque, hive_node_h,
1057 hive_value_h, hive_type t, size_t len, const char *key,
1059 int (*value_dword) (hive_h *, void *opaque, hive_node_h, hive_value_h,
1060 hive_type t, size_t len, const char *key, int32_t);
1061 int (*value_qword) (hive_h *, void *opaque, hive_node_h, hive_value_h,
1062 hive_type t, size_t len, const char *key, int64_t);
1063 int (*value_binary) (hive_h *, void *opaque, hive_node_h, hive_value_h,
1064 hive_type t, size_t len, const char *key, const char *value);
1065 int (*value_none) (hive_h *, void *opaque, hive_node_h, hive_value_h,
1066 hive_type t, size_t len, const char *key, const char *value);
1067 int (*value_other) (hive_h *, void *opaque, hive_node_h, hive_value_h,
1068 hive_type t, size_t len, const char *key, const char *value);
1069 /* If value_any callback is not NULL, then the other value_*
1070 * callbacks are not used, and value_any is called on all values.
1072 int (*value_any) (hive_h *, void *opaque, hive_node_h, hive_value_h,
1073 hive_type t, size_t len, const char *key, const char *value);
1080 int hivex_visit (hive_h *h, const struct hivex_visitor *visitor, size_t len, void *opaque, int flags);
1082 Visit all the nodes recursively in the hive C<h>.
1084 C<visitor> should be a C<hivex_visitor> structure with callback
1085 fields filled in as required (unwanted callbacks can be set to
1086 NULL). C<len> must be the length of the 'visitor' struct (you
1087 should pass C<sizeof (struct hivex_visitor)> for this).
1089 This returns 0 if the whole recursive visit was completed
1090 successfully. On error this returns -1. If one of the callback
1091 functions returned an error than we don't touch errno. If the
1092 error was generated internally then we set errno.
1094 You can skip bad registry entries by setting C<flag> to
1095 C<HIVEX_VISIT_SKIP_BAD>. If this flag is not set, then a bad registry
1096 causes the function to return an error immediately.
1098 This function is robust if the registry contains cycles or
1099 pointers which are invalid or outside the registry. It detects
1100 these cases and returns an error.
1102 =item hivex_visit_node
1104 int hivex_visit_node (hive_h *h, hive_node_h node, const struct hivex_visitor *visitor, size_t len, void *opaque);
1106 Same as C<hivex_visit> but instead of starting out at the root, this
1111 =head1 THE STRUCTURE OF THE WINDOWS REGISTRY
1113 Note: To understand the relationship between hives and the common
1114 Windows Registry keys (like C<HKEY_LOCAL_MACHINE>) please see the
1115 Wikipedia page on the Windows Registry.
1117 The Windows Registry is split across various binary files, each
1118 file being known as a \"hive\". This library only handles a single
1119 hive file at a time.
1121 Hives are n-ary trees with a single root. Each node in the tree
1124 Each node in the tree (including non-leaf nodes) may have an
1125 arbitrary list of (key, value) pairs attached to it. It may
1126 be the case that one of these pairs has an empty key. This
1127 is referred to as the default key for the node.
1129 The (key, value) pairs are the place where the useful data is
1130 stored in the registry. The key is always a string (possibly the
1131 empty string for the default key). The value is a typed object
1132 (eg. string, int32, binary, etc.).
1134 =head2 RELATIONSHIP TO .REG FILES
1136 Although this library does not care about or deal with Windows reg
1137 files, it's useful to look at the relationship between the registry
1138 itself and reg files because they are so common.
1140 A reg file is a text representation of the registry, or part of the
1141 registry. The actual registry hives that Windows uses are binary
1142 files. There are a number of Windows and Linux tools that let you
1143 generate reg files, or merge reg files back into the registry hives.
1144 Notable amongst them is Microsoft's REGEDIT program (formerly known as
1147 A typical reg file will contain many sections looking like this:
1149 [HKEY_LOCAL_MACHINE\\SOFTWARE\\Classes\\Stack]
1150 \"@\"=\"Generic Stack\"
1151 \"TileInfo\"=\"prop:System.FileCount\"
1152 \"TilePath\"=str(2):\"%%systemroot%%\\\\system32\"
1153 \"ThumbnailCutoff\"=dword:00000000
1154 \"FriendlyTypeName\"=hex(2):40,00,25,00,53,00,79,00,73,00,74,00,65,00,6d,00,52,00,6f,00,\\
1155 6f,00,74,00,25,00,5c,00,53,00,79,00,73,00,74,00,65,00,6d,00,\\
1156 33,00,32,00,5c,00,73,00,65,00,61,00,72,00,63,00,68,00,66,00,\\
1157 6f,00,6c,00,64,00,65,00,72,00,2e,00,64,00,6c,00,6c,00,2c,00,\\
1158 2d,00,39,00,30,00,32,00,38,00,00,00,d8
1160 Taking this one piece at a time:
1162 [HKEY_LOCAL_MACHINE\\SOFTWARE\\Classes\\Stack]
1164 This is the path to this node in the registry tree. The first part,
1165 C<HKEY_LOCAL_MACHINE\\SOFTWARE> means that this comes from a hive
1166 (file) called C<SOFTWARE>. C<\\Classes\\Stack> is the real path part,
1167 starting at the root node of the C<SOFTWARE> hive.
1169 Below the node name is a list of zero or more key-value pairs. Any
1170 interior or leaf node in the registry may have key-value pairs
1173 \"@\"=\"Generic Stack\"
1175 This is the \"default key\". In reality (ie. inside the binary hive)
1176 the key string is the empty string. In reg files this is written as
1177 C<@> but this has no meaning either in the hives themselves or in this
1178 library. The value is a string (type 1 - see C<enum hive_type>
1181 \"TileInfo\"=\"prop:System.FileCount\"
1183 This is a regular (key, value) pair, with the value being a type 1
1184 string. Note that inside the binary file the string is likely to be
1185 UTF-16 encoded. This library converts to and from UTF-8 strings
1188 \"TilePath\"=str(2):\"%%systemroot%%\\\\system32\"
1190 The value in this case has type 2 (expanded string) meaning that some
1191 %%...%% variables get expanded by Windows. (This library doesn't know
1192 or care about variable expansion).
1194 \"ThumbnailCutoff\"=dword:00000000
1196 The value in this case is a dword (type 4).
1198 \"FriendlyTypeName\"=hex(2):40,00,....
1200 This value is an expanded string (type 2) represented in the reg file
1201 as a series of hex bytes. In this case the string appears to be a
1204 =head1 NOTE ON THE USE OF ERRNO
1206 Many functions in this library set errno to indicate errors. These
1207 are the values of errno you may encounter (this list is not
1214 Corrupt or unsupported Registry file format.
1222 Passed an invalid argument to the function.
1226 Followed a Registry pointer which goes outside
1227 the registry or outside a registry block.
1231 Registry contains cycles.
1235 Field in the registry out of range.
1239 Registry key already exists.
1243 Tried to write to a registry which is not opened for writing.
1247 =head1 ENVIRONMENT VARIABLES
1253 Setting HIVEX_DEBUG=1 will enable very verbose messages. This is
1254 useful for debugging problems with the library itself.
1264 L<http://libguestfs.org/>,
1267 L<http://en.wikipedia.org/wiki/Windows_Registry>.
1271 Richard W.M. Jones (C<rjones at redhat dot com>)
1275 Copyright (C) 2009-2010 Red Hat Inc.
1277 Derived from code by Petter Nordahl-Hagen under a compatible license:
1278 Copyright (C) 1997-2007 Petter Nordahl-Hagen.
1280 Derived from code by Markus Stephany under a compatible license:
1281 Copyright (C) 2000-2004 Markus Stephany.
1283 This library is free software; you can redistribute it and/or
1284 modify it under the terms of the GNU Lesser General Public
1285 License as published by the Free Software Foundation;
1286 version 2.1 of the License.
1288 This library is distributed in the hope that it will be useful,
1289 but WITHOUT ANY WARRANTY; without even the implied warranty of
1290 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
1291 Lesser General Public License for more details.
1293 See file LICENSE for the full license.
1296 let output_to filename k =
1297 let filename_new = filename ^ ".new" in
1298 chan := open_out filename_new;
1301 chan := Pervasives.stdout;
1303 (* Is the new file different from the current file? *)
1304 if Sys.file_exists filename && files_equal filename filename_new then
1305 unlink filename_new (* same, so skip it *)
1307 (* different, overwrite old one *)
1308 (try chmod filename 0o644 with Unix_error _ -> ());
1309 rename filename_new filename;
1310 chmod filename 0o444;
1311 printf "written %s\n%!" filename;
1314 let perror msg = function
1315 | Unix_error (err, _, _) ->
1316 eprintf "%s: %s\n" msg (error_message err)
1318 eprintf "%s: %s\n" msg (Printexc.to_string exn)
1323 try openfile "configure.ac" [O_RDWR] 0
1325 | Unix_error (ENOENT, _, _) ->
1327 You are probably running this from the wrong directory.
1328 Run it from the top source directory using the command
1329 generator/generator.ml
1333 perror "open: configure.ac" exn;
1336 (* Acquire a lock so parallel builds won't try to run the generator
1337 * twice at the same time. Subsequent builds will wait for the first
1338 * one to finish. Note the lock is released implicitly when the
1341 (try lockf lock_fd F_LOCK 1
1343 perror "lock: configure.ac" exn;
1348 output_to "hivex/hivex.h" generate_c_header;
1349 output_to "hivex/hivex.pod" generate_c_pod;
1351 (* Always generate this file last, and unconditionally. It's used
1352 * by the Makefile to know when we must re-run the generator.
1354 let chan = open_out "generator/stamp-generator" in
1358 printf "generated %d lines of code\n" !lines