1 | /* $Id: tstDisasm-2.cpp 8935 2008-05-19 20:39:21Z vboxsync $ */
|
---|
2 | /** @file
|
---|
3 | * Testcase - Generic Disassembler Tool.
|
---|
4 | */
|
---|
5 |
|
---|
6 | /*
|
---|
7 | * Copyright (C) 2008 Sun Microsystems, Inc.
|
---|
8 | *
|
---|
9 | * This file is part of VirtualBox Open Source Edition (OSE), as
|
---|
10 | * available from http://www.215389.xyz. This file is free software;
|
---|
11 | * you can redistribute it and/or modify it under the terms of the GNU
|
---|
12 | * General Public License (GPL) as published by the Free Software
|
---|
13 | * Foundation, in version 2 as it comes in the "COPYING" file of the
|
---|
14 | * VirtualBox OSE distribution. VirtualBox OSE is distributed in the
|
---|
15 | * hope that it will be useful, but WITHOUT ANY WARRANTY of any kind.
|
---|
16 | *
|
---|
17 | * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa
|
---|
18 | * Clara, CA 95054 USA or visit http://www.sun.com if you need
|
---|
19 | * additional information or have any questions.
|
---|
20 | */
|
---|
21 |
|
---|
22 | /*******************************************************************************
|
---|
23 | * Header Files *
|
---|
24 | *******************************************************************************/
|
---|
25 | #include <VBox/dis.h>
|
---|
26 | #include <iprt/stream.h>
|
---|
27 | #include <iprt/getopt.h>
|
---|
28 | #include <iprt/file.h>
|
---|
29 | #include <iprt/string.h>
|
---|
30 | #include <iprt/runtime.h>
|
---|
31 | #include <iprt/err.h>
|
---|
32 |
|
---|
33 |
|
---|
34 | /*******************************************************************************
|
---|
35 | * Structures and Typedefs *
|
---|
36 | *******************************************************************************/
|
---|
37 | typedef enum { kAsmStyle_Default, kAsmStyle_yasm, kAsmStyle_masm, kAsmStyle_invalid } ASMSTYLE;
|
---|
38 |
|
---|
39 | typedef struct MYDISSTATE
|
---|
40 | {
|
---|
41 | DISCPUSTATE Cpu;
|
---|
42 | uint64_t uAddress; /**< The current instruction address. */
|
---|
43 | uint8_t *pbInstr; /**< The current instruction (pointer). */
|
---|
44 | uint32_t cbInstr; /**< The size of the current instruction. */
|
---|
45 | bool fInvalid; /**< Whether the instruction is invalid/illegal or not. */
|
---|
46 | bool fRaw; /**< Whether invalid instructions are printed as byte defintions or not. */
|
---|
47 | int rc; /**< Set if we hit EOF. */
|
---|
48 | size_t cbLeft; /**< The number of bytes left. (read) */
|
---|
49 | uint8_t *pbNext; /**< The next byte. (read) */
|
---|
50 | uint64_t uNextAddr; /**< The address of the next byte. (read) */
|
---|
51 | char szLine[256]; /**< The disassembler text output. */
|
---|
52 | } MYDISSTATE;
|
---|
53 | typedef MYDISSTATE *PMYDISSTATE;
|
---|
54 |
|
---|
55 |
|
---|
56 | /**
|
---|
57 | * Default style.
|
---|
58 | *
|
---|
59 | * @param pState The disassembler state.
|
---|
60 | */
|
---|
61 | static void MyDisasDefaultFormatter(PMYDISSTATE pState)
|
---|
62 | {
|
---|
63 | RTPrintf("%s", pState->szLine);
|
---|
64 | }
|
---|
65 |
|
---|
66 |
|
---|
67 | /**
|
---|
68 | * Yasm style.
|
---|
69 | *
|
---|
70 | * @param pState The disassembler state.
|
---|
71 | */
|
---|
72 | static void MyDisasYasmFormatter(PMYDISSTATE pState)
|
---|
73 | {
|
---|
74 | RTPrintf("yasm not implemented: %s", pState->szLine);
|
---|
75 | }
|
---|
76 |
|
---|
77 |
|
---|
78 | /**
|
---|
79 | * Masm style.
|
---|
80 | *
|
---|
81 | * @param pState The disassembler state.
|
---|
82 | */
|
---|
83 | static void MyDisasMasmFormatter(PMYDISSTATE pState)
|
---|
84 | {
|
---|
85 | RTPrintf("masm not implemented: %s", pState->szLine);
|
---|
86 | }
|
---|
87 |
|
---|
88 |
|
---|
89 |
|
---|
90 | /**
|
---|
91 | * Callback for reading bytes.
|
---|
92 | */
|
---|
93 | static DECLCALLBACK(int) MyDisasInstrRead(RTUINTPTR uSrcAddr, uint8_t *pbDst, uint32_t cbRead, void *pvDisCpu)
|
---|
94 | {
|
---|
95 | PMYDISSTATE pState = (PMYDISSTATE)pvDisCpu;
|
---|
96 | if (RT_LIKELY( pState->uNextAddr == uSrcAddr
|
---|
97 | && pState->cbLeft >= cbRead))
|
---|
98 | {
|
---|
99 | /*
|
---|
100 | * Straight forward reading.
|
---|
101 | */
|
---|
102 | if (cbRead == 1)
|
---|
103 | {
|
---|
104 | pState->cbLeft--;
|
---|
105 | *pbDst = *pState->pbNext++;
|
---|
106 | pState->uNextAddr++;
|
---|
107 | }
|
---|
108 | else
|
---|
109 | {
|
---|
110 | memcpy(pbDst, pState->pbNext, cbRead);
|
---|
111 | pState->pbNext += cbRead;
|
---|
112 | pState->cbLeft -= cbRead;
|
---|
113 | pState->uNextAddr += cbRead;
|
---|
114 | }
|
---|
115 | }
|
---|
116 | else
|
---|
117 | {
|
---|
118 | /*
|
---|
119 | * Jumping up the stream.
|
---|
120 | */
|
---|
121 | uint64_t offReq64 = uSrcAddr - pState->uAddress;
|
---|
122 | if (offReq64 < 32)
|
---|
123 | {
|
---|
124 | uint32_t offReq = offReq64;
|
---|
125 | uintptr_t off = pState->pbNext - pState->pbInstr;
|
---|
126 | if (off + pState->cbLeft <= offReq)
|
---|
127 | {
|
---|
128 | pState->pbNext += pState->cbLeft;
|
---|
129 | pState->uNextAddr += pState->cbLeft;
|
---|
130 | pState->cbLeft = 0;
|
---|
131 |
|
---|
132 | memset(pbDst, 0xcc, cbRead);
|
---|
133 | pState->rc = VERR_EOF;
|
---|
134 | return VERR_EOF;
|
---|
135 | }
|
---|
136 |
|
---|
137 | /* reset the stream. */
|
---|
138 | pState->cbLeft += off;
|
---|
139 | pState->pbNext = pState->pbInstr;
|
---|
140 | pState->uNextAddr = pState->uAddress;
|
---|
141 |
|
---|
142 | /* skip ahead. */
|
---|
143 | pState->cbLeft -= offReq;
|
---|
144 | pState->pbNext += offReq;
|
---|
145 | pState->uNextAddr += offReq;
|
---|
146 |
|
---|
147 | /* do the reading. */
|
---|
148 | if (pState->cbLeft >= cbRead)
|
---|
149 | {
|
---|
150 | memcpy(pbDst, pState->pbNext, cbRead);
|
---|
151 | pState->cbLeft -= cbRead;
|
---|
152 | pState->pbNext += cbRead;
|
---|
153 | pState->uNextAddr += cbRead;
|
---|
154 | }
|
---|
155 | else
|
---|
156 | {
|
---|
157 | if (pState->cbLeft > 0)
|
---|
158 | {
|
---|
159 | memcpy(pbDst, pState->pbNext, pState->cbLeft);
|
---|
160 | pbDst += pState->cbLeft;
|
---|
161 | cbRead -= pState->cbLeft;
|
---|
162 | pState->pbNext += pState->cbLeft;
|
---|
163 | pState->uNextAddr += pState->cbLeft;
|
---|
164 | pState->cbLeft = 0;
|
---|
165 | }
|
---|
166 | memset(pbDst, 0xcc, cbRead);
|
---|
167 | pState->rc = VERR_EOF;
|
---|
168 | return VERR_EOF;
|
---|
169 | }
|
---|
170 | }
|
---|
171 | else
|
---|
172 | {
|
---|
173 | RTStrmPrintf(g_pStdErr, "Reading before current instruction!\n");
|
---|
174 | memset(pbDst, 0x90, cbRead);
|
---|
175 | pState->rc = VERR_INTERNAL_ERROR;
|
---|
176 | return VERR_INTERNAL_ERROR;
|
---|
177 | }
|
---|
178 | }
|
---|
179 |
|
---|
180 | return VINF_SUCCESS;
|
---|
181 | }
|
---|
182 |
|
---|
183 |
|
---|
184 | /**
|
---|
185 | * Disassembles a block of memory.
|
---|
186 | *
|
---|
187 | * @returns VBox status code.
|
---|
188 | * @param argv0 Program name (for errors and warnings).
|
---|
189 | * @param enmCpuMode The cpu mode to disassemble in.
|
---|
190 | * @param uAddress The address we're starting to disassemble at.
|
---|
191 | * @param pbFile Where to start disassemble.
|
---|
192 | * @param cbFile How much to disassemble.
|
---|
193 | * @param enmStyle The assembly output style.
|
---|
194 | * @param fListing Whether to print in a listing like mode.
|
---|
195 | * @param fRaw Whether to output byte definitions for invalid sequences.
|
---|
196 | * @param fAllInvalid Whether all instructions are expected to be invalid.
|
---|
197 | */
|
---|
198 | static int MyDisasmBlock(const char *argv0, DISCPUMODE enmCpuMode, uint64_t uAddress, uint8_t *pbFile, size_t cbFile,
|
---|
199 | ASMSTYLE enmStyle, bool fListing, bool fRaw, bool fAllInvalid)
|
---|
200 | {
|
---|
201 | /*
|
---|
202 | * Initialize the CPU context.
|
---|
203 | */
|
---|
204 | MYDISSTATE State;
|
---|
205 | State.Cpu.mode = enmCpuMode;
|
---|
206 | State.Cpu.pfnReadBytes = MyDisasInstrRead;
|
---|
207 | State.uAddress = uAddress;
|
---|
208 | State.pbInstr = pbFile;
|
---|
209 | State.cbInstr = 0;
|
---|
210 | State.fInvalid = false;
|
---|
211 | State.fRaw = fRaw;
|
---|
212 | State.rc = VINF_SUCCESS;
|
---|
213 | State.cbLeft = cbFile;
|
---|
214 | State.pbNext = pbFile;
|
---|
215 | State.uNextAddr = uAddress;
|
---|
216 |
|
---|
217 | void (*pfnFormatter)(PMYDISSTATE pState);
|
---|
218 | switch (enmStyle)
|
---|
219 | {
|
---|
220 | case kAsmStyle_Default:
|
---|
221 | pfnFormatter = MyDisasDefaultFormatter;
|
---|
222 | break;
|
---|
223 |
|
---|
224 | case kAsmStyle_yasm:
|
---|
225 | pfnFormatter = MyDisasYasmFormatter;
|
---|
226 | break;
|
---|
227 |
|
---|
228 | case kAsmStyle_masm:
|
---|
229 | pfnFormatter = MyDisasMasmFormatter;
|
---|
230 | break;
|
---|
231 |
|
---|
232 | default:
|
---|
233 | AssertFailedReturn(VERR_INTERNAL_ERROR);
|
---|
234 | }
|
---|
235 |
|
---|
236 | /*
|
---|
237 | * The loop.
|
---|
238 | */
|
---|
239 | int rcRet = VINF_SUCCESS;
|
---|
240 | while (State.cbLeft > 0)
|
---|
241 | {
|
---|
242 | /*
|
---|
243 | * Disassemble it.
|
---|
244 | */
|
---|
245 | State.cbInstr = 0;
|
---|
246 | State.cbLeft += State.pbNext - State.pbInstr;
|
---|
247 | State.uNextAddr = State.uAddress;
|
---|
248 | State.pbNext = State.pbInstr;
|
---|
249 |
|
---|
250 | int rc = DISInstr(&State.Cpu, State.uAddress, 0, &State.cbInstr, State.szLine);
|
---|
251 | if (RT_SUCCESS(rc))
|
---|
252 | {
|
---|
253 | State.fInvalid = State.Cpu.pCurInstr->opcode == OP_INVALID
|
---|
254 | || State.Cpu.pCurInstr->opcode == OP_ILLUD2;
|
---|
255 | if (!fAllInvalid || State.fInvalid)
|
---|
256 | pfnFormatter(&State);
|
---|
257 | else
|
---|
258 | {
|
---|
259 | RTPrintf("%s: error at %#RX64: unexpected valid instruction\n", argv0, State.uAddress);
|
---|
260 | pfnFormatter(&State);
|
---|
261 | rcRet = VERR_GENERAL_FAILURE;
|
---|
262 | }
|
---|
263 | }
|
---|
264 | else
|
---|
265 | {
|
---|
266 | State.cbInstr = State.pbNext - State.pbInstr;
|
---|
267 | if (!State.cbLeft)
|
---|
268 | RTPrintf("%s: error at %#RX64: read beyond the end (%Rrc)\n", argv0, State.uAddress, rc);
|
---|
269 | else if (State.cbInstr)
|
---|
270 | RTPrintf("%s: error at %#RX64: %Rrc cbInstr=%d\n", argv0, State.uAddress, rc, State.cbInstr);
|
---|
271 | else
|
---|
272 | {
|
---|
273 | RTPrintf("%s: error at %#RX64: %Rrc cbInstr=%d!\n", argv0, State.uAddress, rc, State.cbInstr);
|
---|
274 | if (rcRet == VINF_SUCCESS)
|
---|
275 | rcRet = rc;
|
---|
276 | break;
|
---|
277 | }
|
---|
278 | }
|
---|
279 |
|
---|
280 |
|
---|
281 | /* next */
|
---|
282 | State.uAddress += State.cbInstr;
|
---|
283 | State.pbInstr += State.cbInstr;
|
---|
284 | }
|
---|
285 |
|
---|
286 | return rcRet;
|
---|
287 | }
|
---|
288 |
|
---|
289 |
|
---|
290 | /**
|
---|
291 | * Prints usage info.
|
---|
292 | *
|
---|
293 | * @returns 1.
|
---|
294 | * @param argv0 The program name.
|
---|
295 | */
|
---|
296 | static int Usage(const char *argv0)
|
---|
297 | {
|
---|
298 | RTStrmPrintf(g_pStdErr,
|
---|
299 | "usage: %s [options] <file1> [file2..fileN]\n"
|
---|
300 | " or: %s <--help|-h>\n"
|
---|
301 | "\n"
|
---|
302 | "Options:\n"
|
---|
303 | " --address|-a <address>\n"
|
---|
304 | " The base address. Default: 0\n"
|
---|
305 | " --max-bytes|-b <bytes>\n"
|
---|
306 | " The maximum number of bytes to disassemble. Default: 1GB\n"
|
---|
307 | " --cpumode|-c <16|32|64>\n"
|
---|
308 | " The cpu mode. Default: 32\n"
|
---|
309 | " --all-invalid|-i\n"
|
---|
310 | " When specified all instructions are expected to be invalid.\n"
|
---|
311 | " --listing|-l, --no-listing|-L\n"
|
---|
312 | " Enables or disables listing mode. Default: --no-listing\n"
|
---|
313 | " --offset|-o <offset>\n"
|
---|
314 | " The file offset at which to start disassembling. Default: 0\n"
|
---|
315 | " --raw|-r, --no-raw|-R\n"
|
---|
316 | " Whether to employ byte defines for unknown bits. Default: --no-raw\n"
|
---|
317 | " --style|-s <default|yasm|masm>\n"
|
---|
318 | " The assembly output style. Default: default\n"
|
---|
319 | , argv0, argv0);
|
---|
320 | return 1;
|
---|
321 | }
|
---|
322 |
|
---|
323 |
|
---|
324 | int main(int argc, char **argv)
|
---|
325 | {
|
---|
326 | RTR3Init();
|
---|
327 |
|
---|
328 | /* options */
|
---|
329 | uint64_t uAddress = 0;
|
---|
330 | ASMSTYLE enmStyle = kAsmStyle_Default;
|
---|
331 | bool fListing = true;
|
---|
332 | bool fRaw = false;
|
---|
333 | bool fAllInvalid = false;
|
---|
334 | DISCPUMODE enmCpuMode = CPUMODE_32BIT;
|
---|
335 | RTFOFF off = 0;
|
---|
336 | RTFOFF cbMax = _1G;
|
---|
337 |
|
---|
338 | /*
|
---|
339 | * Parse arguments.
|
---|
340 | */
|
---|
341 | static const RTOPTIONDEF g_aOptions[] =
|
---|
342 | {
|
---|
343 | { "--address", 'a', RTGETOPT_REQ_UINT64 },
|
---|
344 | { "--cpumode", 'c', RTGETOPT_REQ_UINT32 },
|
---|
345 | { "--help", 'h', 0 },
|
---|
346 | { "--bytes", 'b', RTGETOPT_REQ_INT64 },
|
---|
347 | { "--all-invalid", 'i', 0, },
|
---|
348 | { "--listing", 'l', 0 },
|
---|
349 | { "--no-listing", 'L', 0 },
|
---|
350 | { "--offset", 'o', RTGETOPT_REQ_INT64 },
|
---|
351 | { "--raw", 'r', 0 },
|
---|
352 | { "--no-raw", 'R', 0 },
|
---|
353 | { "--style", 's', RTGETOPT_REQ_STRING },
|
---|
354 | };
|
---|
355 |
|
---|
356 | int ch;
|
---|
357 | int iArg = 1;
|
---|
358 | RTOPTIONUNION ValueUnion;
|
---|
359 | while ((ch = RTGetOpt(argc, argv, g_aOptions, RT_ELEMENTS(g_aOptions), &iArg, &ValueUnion)))
|
---|
360 | {
|
---|
361 | switch (ch)
|
---|
362 | {
|
---|
363 | case 'a':
|
---|
364 | uAddress = ValueUnion.u64;
|
---|
365 | break;
|
---|
366 |
|
---|
367 | case 'b':
|
---|
368 | cbMax = ValueUnion.i;
|
---|
369 | break;
|
---|
370 |
|
---|
371 | case 'c':
|
---|
372 | if (ValueUnion.u32 == 16)
|
---|
373 | enmCpuMode = CPUMODE_16BIT;
|
---|
374 | else if (ValueUnion.u32 == 32)
|
---|
375 | enmCpuMode = CPUMODE_32BIT;
|
---|
376 | else if (ValueUnion.u32 == 64)
|
---|
377 | enmCpuMode = CPUMODE_64BIT;
|
---|
378 | else
|
---|
379 | {
|
---|
380 | RTStrmPrintf(g_pStdErr, "%s: Invalid CPU mode value %RU32\n", argv[0], ValueUnion.u32);
|
---|
381 | return 1;
|
---|
382 | }
|
---|
383 | break;
|
---|
384 |
|
---|
385 | case 'h':
|
---|
386 | return Usage(argv[0]);
|
---|
387 |
|
---|
388 | case 'i':
|
---|
389 | fAllInvalid = true;
|
---|
390 | break;
|
---|
391 |
|
---|
392 | case 'l':
|
---|
393 | fListing = true;
|
---|
394 | break;
|
---|
395 |
|
---|
396 | case 'L':
|
---|
397 | fListing = false;
|
---|
398 | break;
|
---|
399 |
|
---|
400 | case 'o':
|
---|
401 | off = ValueUnion.i;
|
---|
402 | break;
|
---|
403 |
|
---|
404 | case 'r':
|
---|
405 | fRaw = true;
|
---|
406 | break;
|
---|
407 |
|
---|
408 | case 'R':
|
---|
409 | fRaw = false;
|
---|
410 | break;
|
---|
411 |
|
---|
412 | case 's':
|
---|
413 | if (!strcmp(ValueUnion.psz, "default"))
|
---|
414 | enmStyle = kAsmStyle_Default;
|
---|
415 | else if (!strcmp(ValueUnion.psz, "yasm"))
|
---|
416 | enmStyle = kAsmStyle_yasm;
|
---|
417 | else if (!strcmp(ValueUnion.psz, "masm"))
|
---|
418 | {
|
---|
419 | enmStyle = kAsmStyle_masm;
|
---|
420 | RTStrmPrintf(g_pStdErr, "%s: masm style isn't implemented yet\n", argv[0]);
|
---|
421 | return 1;
|
---|
422 | }
|
---|
423 | else
|
---|
424 | {
|
---|
425 | RTStrmPrintf(g_pStdErr, "%s: unknown assembly style: %s\n", argv[0], ValueUnion.psz);
|
---|
426 | return 1;
|
---|
427 | }
|
---|
428 | break;
|
---|
429 |
|
---|
430 | default:
|
---|
431 | RTStrmPrintf(g_pStdErr, "%s: syntax error: %Rrc\n", argv[0], ch);
|
---|
432 | return 1;
|
---|
433 | }
|
---|
434 | }
|
---|
435 | if (iArg >= argc)
|
---|
436 | return Usage(argv[0]);
|
---|
437 |
|
---|
438 | /*
|
---|
439 | * Process the files.
|
---|
440 | */
|
---|
441 | int rc = VINF_SUCCESS;
|
---|
442 | for ( ; iArg < argc; iArg++)
|
---|
443 | {
|
---|
444 | /*
|
---|
445 | * Read the file into memory.
|
---|
446 | */
|
---|
447 | void *pvFile;
|
---|
448 | size_t cbFile;
|
---|
449 | rc = RTFileReadAllEx(argv[iArg], off, cbMax, 0, &pvFile, &cbFile);
|
---|
450 | if (RT_FAILURE(rc))
|
---|
451 | {
|
---|
452 | RTStrmPrintf(g_pStdErr, "%s: %s: %Rrc\n", argv[0], argv[iArg], rc);
|
---|
453 | break;
|
---|
454 | }
|
---|
455 |
|
---|
456 | /*
|
---|
457 | * Disassemble it.
|
---|
458 | */
|
---|
459 | rc = MyDisasmBlock(argv[0], enmCpuMode, uAddress, (uint8_t *)pvFile, cbFile, enmStyle, fListing, fRaw, fAllInvalid);
|
---|
460 | if (RT_FAILURE(rc))
|
---|
461 | break;
|
---|
462 | }
|
---|
463 |
|
---|
464 | return RT_SUCCESS(rc) ? 0 : 1;
|
---|
465 | }
|
---|
466 |
|
---|