jvm開發筆記2---java反彙編器

做者：王智通（阿里雲安全工程師）

  這兩天在class文件解析器的基礎上， 加上了java反彙編的功能， 反彙編器是指令解釋器的基礎，經過編寫反彙編器能夠熟悉jvm的指令系統， 不過jvm的指令一共有201個，反彙編過程基本就是個體力活。在《java虛擬機規範》中對每一條指令都有了詳細的描述，下面說說我是如何解析 bytecode的：

一個java文件通過javac編譯後會生成class格式文件， 在class格式中method字段裏會有Code屬性，Code屬性包含了java的指令碼和長度。 首先用class解析器將指令碼提取出來， 舉個例子：

test.java

01	class aa {

02	int a = 6;

03

};

04

05	public class test {

06	public static void main(String args[]) {

07	int i = 0;

08

09	for (i = 0; i < 5; i++)

10	System.out.println("hehe");

11

}

12

}

咱們用class文件解析器把test對應的bytecode打印出來：
len: 5
0x2a0xb70x00x10xb1

這一串bytecode爲：0x2a0xb70x00x10xb1， 長度是5個字節。

對照《java虛擬機規範》咱們來一步步手工解析：

0x2a表明aload_0指令， 它將本地局部變量中的第一個變量壓入到堆棧裏。這個指令自己長度就是一個字節，沒有參數， 所以0x2a的解析就很是簡單， 直接在屏幕打印出aload_0便可：

printf("%s\n", symbol);

0xb7表明invokespecial 它用來調用超類構造方法，實例初始化方法， 私有方法。它的用法以下：
invokespecial indexbyte1 indexbyte2，indexbyte1和indexbyte2各佔一個字節，用(indexbyte1 << 8) | indexbyte2來構建一個常量池中的索引。每一個jvm指令自己都佔用一個字節，加上它的兩個參數， invokespecial語句它將佔用3個字節空間。 因此它的解析算法以下：

1

u2 index;

2

3	index = ((*(u1 *)(base + 1)) << 8) | (*(u1 *)(base + 2));

4	printf("%s #%x\n", symbol, index);

注意0xb7解析完後，咱們要跳過3個字節的地址，那麼就是0xb1了， 它是return指令，沒有參數，所以它的解析方法跟aload_0同樣：
printf("%s\n", symbol);

以上是咱們手工解析的過程， 可是jvm有201條指令， 咱們須要創建一個合適的數據結構:

1	typedef int (*interp_func)(u2 opcode_len, char *symbol, void *base);

2

3	typedef struct bytecode_st {

4	u2 opcode;                  // jvm的指令碼

5	u2 opcode_len;            // 指令總的長度，包括參數

6	char symbol[OPCODE_SYMBOL_LEN];    // 指令對應的助記符

7	interp_func func;         // 解析指令的回調函數

8	}BYTECODE;

咱們能夠直接創建一個大的BYTECODE數組：

001	BYTECODE jvm_byte_code[OPCODE_LEN] = {

002	{0x00,  1,      "nop",          jvm_interp_nop},

003	{0x01,  1,      "aconst_null",          jvm_interp_aconst_null},

004	{0x02,  1,      "iconst_m1",            jvm_interp_iconst_m1},

005	{0x03,  1,      "iconst_0",             jvm_interp_iconst_0},

006	{0x04,  1,      "iconst_1",             jvm_interp_iconst_1},

007	{0x05,  1,      "iconst_2",             jvm_interp_iconst_2},

008	{0x06,  1,      "iconst_3",             jvm_interp_iconst_3},

009	{0x07,  1,      "iconst_4",             jvm_interp_iconst_4},

010	{0x08,  1,      "iconst_5",             jvm_interp_iconst_5},

011	{0x09,  1,      "lconst_0",             jvm_interp_lconst_0},

012	{0x0a,  1,      "lconst_1",             jvm_interp_lconst_1},

013	{0x0b,  1,      "fconst_0",             jvm_interp_fconst_0},

014	{0x0c,  1,      "fconst_1",             jvm_interp_fconst_1},

015	{0x0d,  1,      "fconst_2",             jvm_interp_fconst_2},

016	{0x0e,  1,      "dconst_0",             jvm_interp_dconst_0},

017	{0x0f,  1,      "dconst_1",             jvm_interp_dconst_1},

018	{0x10,  1,      "bipush",               jvm_interp_bipush},

019	{0x11,  1,      "sipush",               jvm_interp_sipush},

020	{0x12,  2,      "ldc",          jvm_interp_ldc},

021	{0x13,  1,      "ldc_w",                jvm_interp_ldc_w},

022	{0x14,  1,      "ldc2_w",               jvm_interp_ldc2_w},

023	{0x15,  1,      "iload",                jvm_interp_iload},

024	{0x16,  1,      "lload",                jvm_interp_lload},

025	{0x17,  1,      "fload",                jvm_interp_fload},

026	{0x18,  1,      "dload",                jvm_interp_dload},

027	{0x19,  1,      "aload",                jvm_interp_aload},

028	{0x1a,  1,      "iload_0",              jvm_interp_iload_0},

029	{0x1b,  1,      "iload_1",              jvm_interp_iload_1},

030	{0x1c,  1,      "iload_2",              jvm_interp_iload_2},

031	{0x1d,  1,      "iload_3",              jvm_interp_iload_3},

032	{0x1e,  1,      "lload_0",              jvm_interp_lload_0},

033	{0x1f,  1,      "lload_1",              jvm_interp_lload_1},

034	{0x20,  1,      "lload_2",              jvm_interp_lload_2},

035	{0x21,  1,      "lload_3",              jvm_interp_lload_3},

036	{0x22,  1,      "fload_0",              jvm_interp_fload_0},

037	{0x23,  1,      "fload_1",              jvm_interp_fload_1},

038	{0x24,  1,      "fload_2",              jvm_interp_fload_2},

039	{0x25,  1,      "fload_3",              jvm_interp_fload_3},

040	{0x26,  1,      "dload_0",              jvm_interp_dload_0},

041	{0x27,  1,      "dload_1",              jvm_interp_dload_1},

042	{0x28,  1,      "dload_2",              jvm_interp_dload_2},

043	{0x29,  1,      "dload_3",              jvm_interp_dload_3},

044	{0x2a,  1,      "aload_0",              jvm_interp_aload_0},

045	{0x2b,  1,      "aload_1",              jvm_interp_aload_1},

046	{0x2c,  1,      "aload_2",              jvm_interp_aload_2},

047	{0x2d,  1,      "aload_3",              jvm_interp_aload_3},

048	{0x2e,  1,      "iaload",               jvm_interp_iaload},

049	{0x2f,  1,      "laload",               jvm_interp_laload},

050	{0x30,  1,      "faload",               jvm_interp_faload},

051	{0x31,  1,      "daload",               jvm_interp_daload},

052	{0x32,  1,      "aaload",               jvm_interp_aaload},

053	{0x33,  1,      "baload",               jvm_interp_baload},

054	{0x34,  1,      "caload",               jvm_interp_caload},

055	{0x35,  1,      "saload",               jvm_interp_saload},

056	{0x36,  1,      "istore",               jvm_interp_istore},

057	{0x37,  1,      "lstore",               jvm_interp_lstore},

058	{0x38,  1,      "fstore",               jvm_interp_fstore},

059	{0x39,  1,      "dstore",               jvm_interp_dstore},

060	{0x3a,  1,      "astore",               jvm_interp_astore},

061	{0x3b,  1,      "istore_0",             jvm_interp_istore_0},

062	{0x3c,  1,      "istore_1",             jvm_interp_istore_1},

063	{0x3d,  1,      "istore_2",             jvm_interp_istore_2},

064	{0x3e,  1,      "istore_3",             jvm_interp_istore_3},

065	{0x3f,  1,      "lstore_0",             jvm_interp_lstore_0},

066	{0x40,  1,      "lstore_1",             jvm_interp_lstore_1},

067	{0x41,  1,      "lstore_2",             jvm_interp_lstore_2},

068	{0x42,  1,      "lstore_3",             jvm_interp_lstore_3},

069	{0x43,  1,      "fstore_0",             jvm_interp_fstore_0},

070	{0x44,  1,      "fstore_1",             jvm_interp_fstore_1},

071	{0x45,  1,      "fstore_2",             jvm_interp_fstore_2},

072	{0x46,  1,      "fstore_3",             jvm_interp_fstore_3},

073	{0x47,  1,      "dstore_0",             jvm_interp_dstore_0},

074	{0x48,  1,      "dstore_1",             jvm_interp_dstore_1},

075	{0x49,  1,      "dstore_2",             jvm_interp_dstore_2},

076	{0x4a,  1,      "dstore_3",             jvm_interp_dstore_3},

077	{0x4b,  1,      "astore_0",             jvm_interp_astore_0},

078	{0x4c,  1,      "astore_1",             jvm_interp_astore_1},

079	{0x4d,  1,      "astore_2",             jvm_interp_astore_2},

080	{0x4e,  1,      "astore_3",             jvm_interp_astore_3},

081	{0x4f,  1,      "iastore",              jvm_interp_iastore},

082	{0x50,  1,      "lastore",              jvm_interp_lastore},

083	{0x51,  1,      "fastore",              jvm_interp_fastore},

084	{0x52,  1,      "dastore",              jvm_interp_dastore},

085	{0x53,  1,      "aastore",              jvm_interp_aastore},

086	{0x54,  1,      "bastore",              jvm_interp_bastore},

087	{0x55,  1,      "castore",              jvm_interp_castore},

088	{0x56,  1,      "sastore",              jvm_interp_sastore},

089	{0x57,  1,      "pop",          jvm_interp_pop},

090	{0x58,  1,      "pop2",         jvm_interp_pop2},

091	{0x59,  1,      "dup",          jvm_interp_dup},

092	{0x5a,  1,      "dup_x1",               jvm_interp_dup_x1},

093	{0x5b,  1,      "dup_x2",               jvm_interp_dup_x2},

094	{0x5c,  1,      "dup2",         jvm_interp_dup2},

095	{0x5d,  1,      "dup2_x1",              jvm_interp_dup2_x1},

096	{0x5e,  1,      "dup2_x2",              jvm_interp_dup2_x2},

097	{0x5f,  1,      "swap",         jvm_interp_swap},

098	{0x60,  1,      "iadd",         jvm_interp_iadd},

099	{0x61,  1,      "ladd",         jvm_interp_ladd},

100	{0x62,  1,      "fadd",         jvm_interp_fadd},

101	{0x63,  1,      "dadd",         jvm_interp_dadd},

102	{0x64,  1,      "isub",         jvm_interp_isub},

103	{0x65,  1,      "lsub",         jvm_interp_lsub},

104	{0x66,  1,      "fsub",         jvm_interp_fsub},

105	{0x67,  1,      "dsub",         jvm_interp_dsub},

106	{0x68,  1,      "imul",         jvm_interp_imul},

107	{0x69,  1,      "lmul",         jvm_interp_lmul},

108	{0x6a,  1,      "fmul",         jvm_interp_fmul},

109	{0x6b,  1,      "dmul",         jvm_interp_dmul},

110	{0x6c,  1,      "idiv",         jvm_interp_idiv},

111	{0x6d,  1,      "ldiv",         jvm_interp_ldiv},

112	{0x6e,  1,      "fdiv",         jvm_interp_fdiv},

113	{0x6f,  1,      "ddiv",         jvm_interp_ddiv},

114	{0x70,  1,      "irem",         jvm_interp_irem},

115	{0x71,  1,      "lrem",         jvm_interp_lrem},

116	{0x72,  1,      "frem",         jvm_interp_frem},

117	{0x73,  1,      "drem",         jvm_interp_drem},

118	{0x74,  1,      "ineg",         jvm_interp_ineg},

119	{0x75,  1,      "lneg",         jvm_interp_lneg},

120	{0x76,  1,      "fneg",         jvm_interp_fneg},

121	{0x77,  1,      "dneg",         jvm_interp_dneg},

122	{0x78,  1,      "ishl",         jvm_interp_ishl},

123	{0x79,  1,      "lshl",         jvm_interp_lshl},

124	{0x7a,  1,      "ishr",         jvm_interp_ishr},

125	{0x7b,  1,      "lshr",         jvm_interp_lshr},

126	{0x7c,  1,      "iushr",                jvm_interp_iushr},

127	{0x7d,  1,      "lushr",                jvm_interp_lushr},

128	{0x7e,  1,      "iand",         jvm_interp_iand},

129	{0x7f,  1,      "land",         jvm_interp_land},

130	{0x80,  1,      "ior",          jvm_interp_ior},

131	{0x81,  1,      "lor",          jvm_interp_lor},

132	{0x82,  1,      "ixor",         jvm_interp_ixor},

133	{0x83,  1,      "lxor",         jvm_interp_lxor},

134	{0x84,  3,      "iinc",         jvm_interp_iinc},

135	{0x85,  1,      "i2l",          jvm_interp_i2l},

136	{0x86,  1,      "i2f",          jvm_interp_i2f},

137	{0x87,  1,      "i2d",          jvm_interp_i2d},

138	{0x88,  1,      "l2i",          jvm_interp_l2i},

139	{0x89,  1,      "l2f",          jvm_interp_l2f},

140	{0x8a,  1,      "l2d",          jvm_interp_l2d},

141	{0x8b,  1,      "f2i",          jvm_interp_f2i},

142	{0x8c,  1,      "f2l",          jvm_interp_f2l},

143	{0x8d,  1,      "f2d",          jvm_interp_f2d},

144	{0x8e,  1,      "d2i",          jvm_interp_d2i},

145	{0x8f,  1,      "d2l",          jvm_interp_d2l},

146	{0x90,  1,      "d2f",          jvm_interp_d2f},

147	{0x91,  1,      "i2b",          jvm_interp_i2b},

148	{0x92,  1,      "i2c",          jvm_interp_i2c},

149	{0x93,  1,      "i2s",          jvm_interp_i2s},

150	{0x94,  1,      "lcmp",         jvm_interp_lcmp},

151	{0x95,  1,      "fcmpl",                jvm_interp_fcmpl},

152	{0x96,  1,      "fcmpg",                jvm_interp_fcmpg},

153	{0x97,  1,      "dcmpl",                jvm_interp_dcmpl},

154	{0x98,  1,      "dcmpg",                jvm_interp_dcmpg},

155	{0x99,  1,      "ifeq",                 jvm_interp_ifeq},

156	{0x9a,  1,      "ifne",                 jvm_interp_ifne},

157	{0x9b,  1,      "iflt",                 jvm_interp_iflt},

158	{0x9c,  1,      "ifge",                 jvm_interp_ifge},

159	{0x9d,  1,      "ifgt",                 jvm_interp_ifgt},

160	{0x9e,  1,      "ifle",                 jvm_interp_ifle},

161	{0x9f,  1,      "if_icmpeq",            jvm_interp_if_icmpeq},

162	{0xa0,  1,      "if_icmpne",            jvm_interp_if_icmpne},