1(* Capstone Disassembly Engine
2 * By Nguyen Anh Quynh <aquynh@gmail.com>, 2013-2014 *)
3
4open Arm
5open Arm64
6open Mips
7open Ppc
8open X86
9open Sparc
10open Systemz
11open Xcore
12open Printf	(* debug *)
13
14(* Hardware architectures *)
15type arch =
16  | CS_ARCH_ARM
17  | CS_ARCH_ARM64
18  | CS_ARCH_MIPS
19  | CS_ARCH_X86
20  | CS_ARCH_PPC
21  | CS_ARCH_SPARC
22  | CS_ARCH_SYSZ
23  | CS_ARCH_XCORE
24
25(* Hardware modes *)
26type mode =
27  |	CS_MODE_LITTLE_ENDIAN	(* little-endian mode (default mode) *)
28  |	CS_MODE_ARM			(* ARM mode *)
29  |	CS_MODE_16			(* 16-bit mode (for X86) *)
30  |	CS_MODE_32			(* 32-bit mode (for X86) *)
31  |	CS_MODE_64			(* 64-bit mode (for X86, PPC) *)
32  |	CS_MODE_THUMB		(* ARM's Thumb mode, including Thumb-2 *)
33  |	CS_MODE_MCLASS		(* ARM's MClass mode *)
34  |	CS_MODE_V8    		(* ARMv8 A32 encodings for ARM *)
35  |	CS_MODE_MICRO		(* MicroMips mode (MIPS architecture) *)
36  |	CS_MODE_MIPS3		(* Mips3 mode (MIPS architecture) *)
37  |	CS_MODE_MIPS32R6	(* Mips32-R6 mode (MIPS architecture) *)
38  |	CS_MODE_MIPSGP64	(* MipsGP64 mode (MIPS architecture) *)
39  |	CS_MODE_V9			(* SparcV9 mode (Sparc architecture) *)
40  |	CS_MODE_BIG_ENDIAN	(* big-endian mode *)
41  |	CS_MODE_MIPS32		(* Mips32 mode (for Mips) *)
42  |	CS_MODE_MIPS64		(* Mips64 mode (for Mips) *)
43
44
45(* Runtime option for the disassembled engine *)
46type opt_type =
47  |	CS_OPT_SYNTAX		(*  Asssembly output syntax *)
48  |	CS_OPT_DETAIL		(* Break down instruction structure into details *)
49  |	CS_OPT_MODE		(* Change engine's mode at run-time *)
50  |	CS_OPT_MEM		(* User-defined dynamic memory related functions *)
51  |	CS_OPT_SKIPDATA		(* Skip data when disassembling. Then engine is in SKIPDATA mode. *)
52  |	CS_OPT_SKIPDATA_SETUP 	(* Setup user-defined function for SKIPDATA option *)
53
54
55(* Runtime option value (associated with option type above) *)
56let _CS_OPT_OFF = 0L;; (* Turn OFF an option - default option of CS_OPT_DETAIL, CS_OPT_SKIPDATA. *)
57let _CS_OPT_ON = 3L;;  (* Turn ON an option (CS_OPT_DETAIL, CS_OPT_SKIPDATA). *)
58let _CS_OPT_SYNTAX_DEFAULT = 0L;; (* Default asm syntax (CS_OPT_SYNTAX). *)
59let _CS_OPT_SYNTAX_INTEL = 1L;; (* X86 Intel asm syntax - default on X86 (CS_OPT_SYNTAX). *)
60let _CS_OPT_SYNTAX_ATT = 2L;; (* X86 ATT asm syntax (CS_OPT_SYNTAX). *)
61let _CS_OPT_SYNTAX_NOREGNAME = 3L;; (* Prints register name with only number (CS_OPT_SYNTAX) *)
62
63(* Common instruction operand types - to be consistent across all architectures. *)
64let _CS_OP_INVALID = 0;;  (* uninitialized/invalid operand. *)
65let _CS_OP_REG     = 1;;  (* Register operand. *)
66let _CS_OP_IMM     = 2;;  (* Immediate operand. *)
67let _CS_OP_MEM     = 3;;  (* Memory operand. *)
68let _CS_OP_FP      = 4;;  (* Floating-Point operand. *)
69
70(* Common instruction groups - to be consistent across all architectures. *)
71let _CS_GRP_INVALID = 0;;  (* uninitialized/invalid group. *)
72let _CS_GRP_JUMP    = 1;;  (* all jump instructions (conditional+direct+indirect jumps) *)
73let _CS_GRP_CALL    = 2;;  (* all call instructions *)
74let _CS_GRP_RET     = 3;;  (* all return instructions *)
75let _CS_GRP_INT     = 4;;  (* all interrupt instructions (int+syscall) *)
76let _CS_GRP_IRET    = 5;;  (* all interrupt return instructions *)
77
78type cs_arch =
79	| CS_INFO_ARM of cs_arm
80	| CS_INFO_ARM64 of cs_arm64
81	| CS_INFO_MIPS of cs_mips
82	| CS_INFO_X86 of cs_x86
83	| CS_INFO_PPC of cs_ppc
84	| CS_INFO_SPARC of cs_sparc
85	| CS_INFO_SYSZ of cs_sysz
86	| CS_INFO_XCORE of cs_xcore
87
88
89type csh = {
90	h: Int64.t;
91	a: arch;
92}
93
94type cs_insn0 = {
95	id: int;
96	address: int;
97	size: int;
98	bytes: int array;
99	mnemonic: string;
100	op_str: string;
101	regs_read: int array;
102	regs_write: int array;
103	groups: int array;
104	arch: cs_arch;
105}
106
107external _cs_open: arch -> mode list -> Int64.t option = "ocaml_open"
108external cs_disasm_quick: arch -> mode list -> string -> Int64.t -> Int64.t -> cs_insn0 list = "ocaml_cs_disasm"
109external _cs_disasm_internal: arch -> Int64.t -> string -> Int64.t -> Int64.t -> cs_insn0 list = "ocaml_cs_disasm_internal"
110external _cs_reg_name: Int64.t -> int -> string = "ocaml_register_name"
111external _cs_insn_name: Int64.t -> int -> string = "ocaml_instruction_name"
112external _cs_group_name: Int64.t -> int -> string = "ocaml_group_name"
113external cs_version: unit -> int = "ocaml_version"
114external _cs_option: Int64.t -> opt_type -> Int64.t -> int = "ocaml_option"
115external _cs_close: Int64.t -> int = "ocaml_close"
116
117
118let cs_open _arch _mode: csh = (
119	let _handle = _cs_open _arch _mode in (
120	match _handle with
121	| None -> { h = 0L; a = _arch }
122	| Some v -> { h = v; a = _arch }
123	);
124);;
125
126let cs_close handle = (
127	_cs_close handle.h;
128)
129
130let cs_option handle opt value = (
131	_cs_option handle.h opt value;
132);;
133
134let cs_disasm handle code address count = (
135	_cs_disasm_internal handle.a handle.h code address count;
136);;
137
138let cs_reg_name handle id = (
139	_cs_reg_name handle.h id;
140);;
141
142let cs_insn_name handle id = (
143	_cs_insn_name handle.h id;
144);;
145
146let cs_group_name handle id = (
147	_cs_group_name handle.h id;
148);;
149
150class cs_insn c a =
151	let csh = c in
152	let (id, address, size, bytes, mnemonic, op_str, regs_read,
153        regs_write, groups, arch) =
154        (a.id, a.address, a.size, a.bytes, a.mnemonic, a.op_str,
155        a.regs_read, a.regs_write, a.groups, a.arch) in
156	object
157		method id = id;
158		method address = address;
159		method size = size;
160	        method bytes = bytes;
161		method mnemonic = mnemonic;
162		method op_str = op_str;
163		method regs_read = regs_read;
164		method regs_write = regs_write;
165		method groups = groups;
166		method arch = arch;
167		method reg_name id = _cs_reg_name csh.h id;
168		method insn_name id = _cs_insn_name csh.h id;
169		method group_name id = _cs_group_name csh.h id;
170	end;;
171
172let cs_insn_group handle insn group_id =
173	List.exists (fun g -> g == group_id) (Array.to_list insn.groups);;
174
175let cs_reg_read handle insn reg_id =
176	List.exists (fun g -> g == reg_id) (Array.to_list insn.regs_read);;
177
178let cs_reg_write handle insn reg_id =
179	List.exists (fun g -> g == reg_id) (Array.to_list insn.regs_write);;
180
181
182class cs a m =
183	let mode = m and arch = a in
184	let handle = cs_open arch mode in
185	object
186		method disasm code offset count =
187			let insns = (_cs_disasm_internal arch handle.h code offset count) in
188			List.map (fun x -> new cs_insn handle x) insns;
189
190	end;;
191