1 /*
2 * Copyright (C) 2008 The Android Open Source Project
3 *
4 * Licensed under the Apache License, Version 2.0 (the "License");
5 * you may not use this file except in compliance with the License.
6 * You may obtain a copy of the License at
7 *
8 * http://www.apache.org/licenses/LICENSE-2.0
9 *
10 * Unless required by applicable law or agreed to in writing, software
11 * distributed under the License is distributed on an "AS IS" BASIS,
12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 * See the License for the specific language governing permissions and
14 * limitations under the License.
15 */
16
17 /*
18 * Access .dex (Dalvik Executable Format) files. The code here assumes that
19 * the DEX file has been rewritten (byte-swapped, word-aligned) and that
20 * the contents can be directly accessed as a collection of C arrays. Please
21 * see docs/dalvik/dex-format.html for a detailed description.
22 *
23 * The structure and field names were chosen to match those in the DEX spec.
24 *
25 * It's generally assumed that the DEX file will be stored in shared memory,
26 * obviating the need to copy code and constant pool entries into newly
27 * allocated storage. Maintaining local pointers to items in the shared area
28 * is valid and encouraged.
29 *
30 * All memory-mapped structures are 32-bit aligned unless otherwise noted.
31 */
32
33 #ifndef _LIBDEX_DEXFILE
34 #define _LIBDEX_DEXFILE
35
36 #include "vm/Common.h" // basic type defs, e.g. u1/u2/u4/u8, and LOG
37 #include "libdex/SysUtil.h"
38
39 /*
40 * gcc-style inline management -- ensures we have a copy of all functions
41 * in the library, so code that links against us will work whether or not
42 * it was built with optimizations enabled.
43 */
44 #ifndef _DEX_GEN_INLINES /* only defined by DexInlines.c */
45 # define DEX_INLINE extern __inline__
46 #else
47 # define DEX_INLINE
48 #endif
49
50 /* DEX file magic number */
51 #define DEX_MAGIC "dex\n"
52 /* version, encoded in 4 bytes of ASCII */
53 #define DEX_MAGIC_VERS "035\0"
54
55 /* same, but for optimized DEX header */
56 #define DEX_OPT_MAGIC "dey\n"
57 #define DEX_OPT_MAGIC_VERS "036\0"
58
59 #define DEX_DEP_MAGIC "deps"
60
61 /*
62 * 160-bit SHA-1 digest.
63 */
64 enum { kSHA1DigestLen = 20,
65 kSHA1DigestOutputLen = kSHA1DigestLen*2 +1 };
66
67 /* general constants */
68 enum {
69 kDexEndianConstant = 0x12345678, /* the endianness indicator */
70 kDexNoIndex = 0xffffffff, /* not a valid index value */
71 };
72
73 /*
74 * access flags and masks; the "standard" ones are all <= 0x4000
75 *
76 * Note: There are related declarations in vm/oo/Object.h in the ClassFlags
77 * enum.
78 */
79 enum {
80 ACC_PUBLIC = 0x00000001, // class, field, method, ic
81 ACC_PRIVATE = 0x00000002, // field, method, ic
82 ACC_PROTECTED = 0x00000004, // field, method, ic
83 ACC_STATIC = 0x00000008, // field, method, ic
84 ACC_FINAL = 0x00000010, // class, field, method, ic
85 ACC_SYNCHRONIZED = 0x00000020, // method (only allowed on natives)
86 ACC_SUPER = 0x00000020, // class (not used in Dalvik)
87 ACC_VOLATILE = 0x00000040, // field
88 ACC_BRIDGE = 0x00000040, // method (1.5)
89 ACC_TRANSIENT = 0x00000080, // field
90 ACC_VARARGS = 0x00000080, // method (1.5)
91 ACC_NATIVE = 0x00000100, // method
92 ACC_INTERFACE = 0x00000200, // class, ic
93 ACC_ABSTRACT = 0x00000400, // class, method, ic
94 ACC_STRICT = 0x00000800, // method
95 ACC_SYNTHETIC = 0x00001000, // field, method, ic
96 ACC_ANNOTATION = 0x00002000, // class, ic (1.5)
97 ACC_ENUM = 0x00004000, // class, field, ic (1.5)
98 ACC_CONSTRUCTOR = 0x00010000, // method (Dalvik only)
99 ACC_DECLARED_SYNCHRONIZED =
100 0x00020000, // method (Dalvik only)
101 ACC_CLASS_MASK =
102 (ACC_PUBLIC | ACC_FINAL | ACC_INTERFACE | ACC_ABSTRACT
103 | ACC_SYNTHETIC | ACC_ANNOTATION | ACC_ENUM),
104 ACC_INNER_CLASS_MASK =
105 (ACC_CLASS_MASK | ACC_PRIVATE | ACC_PROTECTED | ACC_STATIC),
106 ACC_FIELD_MASK =
107 (ACC_PUBLIC | ACC_PRIVATE | ACC_PROTECTED | ACC_STATIC | ACC_FINAL
108 | ACC_VOLATILE | ACC_TRANSIENT | ACC_SYNTHETIC | ACC_ENUM),
109 ACC_METHOD_MASK =
110 (ACC_PUBLIC | ACC_PRIVATE | ACC_PROTECTED | ACC_STATIC | ACC_FINAL
111 | ACC_SYNCHRONIZED | ACC_BRIDGE | ACC_VARARGS | ACC_NATIVE
112 | ACC_ABSTRACT | ACC_STRICT | ACC_SYNTHETIC | ACC_CONSTRUCTOR
113 | ACC_DECLARED_SYNCHRONIZED),
114 };
115
116 /* annotation constants */
117 enum {
118 kDexVisibilityBuild = 0x00, /* annotation visibility */
119 kDexVisibilityRuntime = 0x01,
120 kDexVisibilitySystem = 0x02,
121
122 kDexAnnotationByte = 0x00,
123 kDexAnnotationShort = 0x02,
124 kDexAnnotationChar = 0x03,
125 kDexAnnotationInt = 0x04,
126 kDexAnnotationLong = 0x06,
127 kDexAnnotationFloat = 0x10,
128 kDexAnnotationDouble = 0x11,
129 kDexAnnotationString = 0x17,
130 kDexAnnotationType = 0x18,
131 kDexAnnotationField = 0x19,
132 kDexAnnotationMethod = 0x1a,
133 kDexAnnotationEnum = 0x1b,
134 kDexAnnotationArray = 0x1c,
135 kDexAnnotationAnnotation = 0x1d,
136 kDexAnnotationNull = 0x1e,
137 kDexAnnotationBoolean = 0x1f,
138
139 kDexAnnotationValueTypeMask = 0x1f, /* low 5 bits */
140 kDexAnnotationValueArgShift = 5,
141 };
142
143 /* map item type codes */
144 enum {
145 kDexTypeHeaderItem = 0x0000,
146 kDexTypeStringIdItem = 0x0001,
147 kDexTypeTypeIdItem = 0x0002,
148 kDexTypeProtoIdItem = 0x0003,
149 kDexTypeFieldIdItem = 0x0004,
150 kDexTypeMethodIdItem = 0x0005,
151 kDexTypeClassDefItem = 0x0006,
152 kDexTypeMapList = 0x1000,
153 kDexTypeTypeList = 0x1001,
154 kDexTypeAnnotationSetRefList = 0x1002,
155 kDexTypeAnnotationSetItem = 0x1003,
156 kDexTypeClassDataItem = 0x2000,
157 kDexTypeCodeItem = 0x2001,
158 kDexTypeStringDataItem = 0x2002,
159 kDexTypeDebugInfoItem = 0x2003,
160 kDexTypeAnnotationItem = 0x2004,
161 kDexTypeEncodedArrayItem = 0x2005,
162 kDexTypeAnnotationsDirectoryItem = 0x2006,
163 };
164
165 /* auxillary data section chunk codes */
166 enum {
167 kDexChunkClassLookup = 0x434c4b50, /* CLKP */
168 kDexChunkRegisterMaps = 0x524d4150, /* RMAP */
169
170 kDexChunkEnd = 0x41454e44, /* AEND */
171 };
172
173 /* debug info opcodes and constants */
174 enum {
175 DBG_END_SEQUENCE = 0x00,
176 DBG_ADVANCE_PC = 0x01,
177 DBG_ADVANCE_LINE = 0x02,
178 DBG_START_LOCAL = 0x03,
179 DBG_START_LOCAL_EXTENDED = 0x04,
180 DBG_END_LOCAL = 0x05,
181 DBG_RESTART_LOCAL = 0x06,
182 DBG_SET_PROLOGUE_END = 0x07,
183 DBG_SET_EPILOGUE_BEGIN = 0x08,
184 DBG_SET_FILE = 0x09,
185 DBG_FIRST_SPECIAL = 0x0a,
186 DBG_LINE_BASE = -4,
187 DBG_LINE_RANGE = 15,
188 };
189
190 /*
191 * Direct-mapped "header_item" struct.
192 */
193 typedef struct DexHeader {
194 u1 magic[8]; /* includes version number */
195 u4 checksum; /* adler32 checksum */
196 u1 signature[kSHA1DigestLen]; /* SHA-1 hash */
197 u4 fileSize; /* length of entire file */
198 u4 headerSize; /* offset to start of next section */
199 u4 endianTag;
200 u4 linkSize;
201 u4 linkOff;
202 u4 mapOff;
203 u4 stringIdsSize;
204 u4 stringIdsOff;
205 u4 typeIdsSize;
206 u4 typeIdsOff;
207 u4 protoIdsSize;
208 u4 protoIdsOff;
209 u4 fieldIdsSize;
210 u4 fieldIdsOff;
211 u4 methodIdsSize;
212 u4 methodIdsOff;
213 u4 classDefsSize;
214 u4 classDefsOff;
215 u4 dataSize;
216 u4 dataOff;
217 } DexHeader;
218
219 /*
220 * Direct-mapped "map_item".
221 */
222 typedef struct DexMapItem {
223 u2 type; /* type code (see kDexType* above) */
224 u2 unused;
225 u4 size; /* count of items of the indicated type */
226 u4 offset; /* file offset to the start of data */
227 } DexMapItem;
228
229 /*
230 * Direct-mapped "map_list".
231 */
232 typedef struct DexMapList {
233 u4 size; /* #of entries in list */
234 DexMapItem list[1]; /* entries */
235 } DexMapList;
236
237 /*
238 * Direct-mapped "string_id_item".
239 */
240 typedef struct DexStringId {
241 u4 stringDataOff; /* file offset to string_data_item */
242 } DexStringId;
243
244 /*
245 * Direct-mapped "type_id_item".
246 */
247 typedef struct DexTypeId {
248 u4 descriptorIdx; /* index into stringIds list for type descriptor */
249 } DexTypeId;
250
251 /*
252 * Direct-mapped "field_id_item".
253 */
254 typedef struct DexFieldId {
255 u2 classIdx; /* index into typeIds list for defining class */
256 u2 typeIdx; /* index into typeIds for field type */
257 u4 nameIdx; /* index into stringIds for field name */
258 } DexFieldId;
259
260 /*
261 * Direct-mapped "method_id_item".
262 */
263 typedef struct DexMethodId {
264 u2 classIdx; /* index into typeIds list for defining class */
265 u2 protoIdx; /* index into protoIds for method prototype */
266 u4 nameIdx; /* index into stringIds for method name */
267 } DexMethodId;
268
269 /*
270 * Direct-mapped "proto_id_item".
271 */
272 typedef struct DexProtoId {
273 u4 shortyIdx; /* index into stringIds for shorty descriptor */
274 u4 returnTypeIdx; /* index into typeIds list for return type */
275 u4 parametersOff; /* file offset to type_list for parameter types */
276 } DexProtoId;
277
278 /*
279 * Direct-mapped "class_def_item".
280 */
281 typedef struct DexClassDef {
282 u4 classIdx; /* index into typeIds for this class */
283 u4 accessFlags;
284 u4 superclassIdx; /* index into typeIds for superclass */
285 u4 interfacesOff; /* file offset to DexTypeList */
286 u4 sourceFileIdx; /* index into stringIds for source file name */
287 u4 annotationsOff; /* file offset to annotations_directory_item */
288 u4 classDataOff; /* file offset to class_data_item */
289 u4 staticValuesOff; /* file offset to DexEncodedArray */
290 } DexClassDef;
291
292 /*
293 * Direct-mapped "type_item".
294 */
295 typedef struct DexTypeItem {
296 u2 typeIdx; /* index into typeIds */
297 } DexTypeItem;
298
299 /*
300 * Direct-mapped "type_list".
301 */
302 typedef struct DexTypeList {
303 u4 size; /* #of entries in list */
304 DexTypeItem list[1]; /* entries */
305 } DexTypeList;
306
307 /*
308 * Direct-mapped "code_item".
309 *
310 * The "catches" table is used when throwing an exception,
311 * "debugInfo" is used when displaying an exception stack trace or
312 * debugging. An offset of zero indicates that there are no entries.
313 */
314 typedef struct DexCode {
315 u2 registersSize;
316 u2 insSize;
317 u2 outsSize;
318 u2 triesSize;
319 u4 debugInfoOff; /* file offset to debug info stream */
320 u4 insnsSize; /* size of the insns array, in u2 units */
321 u2 insns[1];
322 /* followed by optional u2 padding */
323 /* followed by try_item[triesSize] */
324 /* followed by uleb128 handlersSize */
325 /* followed by catch_handler_item[handlersSize] */
326 } DexCode;
327
328 /*
329 * Direct-mapped "try_item".
330 */
331 typedef struct DexTry {
332 u4 startAddr; /* start address, in 16-bit code units */
333 u2 insnCount; /* instruction count, in 16-bit code units */
334 u2 handlerOff; /* offset in encoded handler data to handlers */
335 } DexTry;
336
337 /*
338 * Link table. Currently undefined.
339 */
340 typedef struct DexLink {
341 u1 bleargh;
342 } DexLink;
343
344
345 /*
346 * Direct-mapped "annotations_directory_item".
347 */
348 typedef struct DexAnnotationsDirectoryItem {
349 u4 classAnnotationsOff; /* offset to DexAnnotationSetItem */
350 u4 fieldsSize; /* count of DexFieldAnnotationsItem */
351 u4 methodsSize; /* count of DexMethodAnnotationsItem */
352 u4 parametersSize; /* count of DexParameterAnnotationsItem */
353 /* followed by DexFieldAnnotationsItem[fieldsSize] */
354 /* followed by DexMethodAnnotationsItem[methodsSize] */
355 /* followed by DexParameterAnnotationsItem[parametersSize] */
356 } DexAnnotationsDirectoryItem;
357
358 /*
359 * Direct-mapped "field_annotations_item".
360 */
361 typedef struct DexFieldAnnotationsItem {
362 u4 fieldIdx;
363 u4 annotationsOff; /* offset to DexAnnotationSetItem */
364 } DexFieldAnnotationsItem;
365
366 /*
367 * Direct-mapped "method_annotations_item".
368 */
369 typedef struct DexMethodAnnotationsItem {
370 u4 methodIdx;
371 u4 annotationsOff; /* offset to DexAnnotationSetItem */
372 } DexMethodAnnotationsItem;
373
374 /*
375 * Direct-mapped "parameter_annotations_item".
376 */
377 typedef struct DexParameterAnnotationsItem {
378 u4 methodIdx;
379 u4 annotationsOff; /* offset to DexAnotationSetRefList */
380 } DexParameterAnnotationsItem;
381
382 /*
383 * Direct-mapped "annotation_set_ref_item".
384 */
385 typedef struct DexAnnotationSetRefItem {
386 u4 annotationsOff; /* offset to DexAnnotationSetItem */
387 } DexAnnotationSetRefItem;
388
389 /*
390 * Direct-mapped "annotation_set_ref_list".
391 */
392 typedef struct DexAnnotationSetRefList {
393 u4 size;
394 DexAnnotationSetRefItem list[1];
395 } DexAnnotationSetRefList;
396
397 /*
398 * Direct-mapped "anotation_set_item".
399 */
400 typedef struct DexAnnotationSetItem {
401 u4 size;
402 u4 entries[1]; /* offset to DexAnnotationItem */
403 } DexAnnotationSetItem;
404
405 /*
406 * Direct-mapped "annotation_item".
407 *
408 * NOTE: this structure is byte-aligned.
409 */
410 typedef struct DexAnnotationItem {
411 u1 visibility;
412 u1 annotation[1]; /* data in encoded_annotation format */
413 } DexAnnotationItem;
414
415 /*
416 * Direct-mapped "encoded_array".
417 *
418 * NOTE: this structure is byte-aligned.
419 */
420 typedef struct DexEncodedArray {
421 u1 array[1]; /* data in encoded_array format */
422 } DexEncodedArray;
423
424 /*
425 * Lookup table for classes. It provides a mapping from class name to
426 * class definition. Used by dexFindClass().
427 *
428 * We calculate this at DEX optimization time and embed it in the file so we
429 * don't need the same hash table in every VM. This is slightly slower than
430 * a hash table with direct pointers to the items, but because it's shared
431 * there's less of a penalty for using a fairly sparse table.
432 */
433 typedef struct DexClassLookup {
434 int size; // total size, including "size"
435 int numEntries; // size of table[]; always power of 2
436 struct {
437 u4 classDescriptorHash; // class descriptor hash code
438 int classDescriptorOffset; // in bytes, from start of DEX
439 int classDefOffset; // in bytes, from start of DEX
440 } table[1];
441 } DexClassLookup;
442
443 /*
444 * Header added by DEX optimization pass. Values are always written in
445 * local byte and structure padding. The first field (magic + version)
446 * is guaranteed to be present and directly readable for all expected
447 * compiler configurations; the rest is version-dependent.
448 *
449 * Try to keep this simple and fixed-size.
450 */
451 typedef struct DexOptHeader {
452 u1 magic[8]; /* includes version number */
453
454 u4 dexOffset; /* file offset of DEX header */
455 u4 dexLength;
456 u4 depsOffset; /* offset of optimized DEX dependency table */
457 u4 depsLength;
458 u4 optOffset; /* file offset of optimized data tables */
459 u4 optLength;
460
461 u4 flags; /* some info flags */
462 u4 checksum; /* adler32 checksum covering deps/opt */
463
464 /* pad for 64-bit alignment if necessary */
465 } DexOptHeader;
466
467 #define DEX_FLAG_VERIFIED (1) /* tried to verify all classes */
468 #define DEX_OPT_FLAG_BIG (1<<1) /* swapped to big-endian */
469 #define DEX_OPT_FLAG_FIELDS (1<<2) /* field access optimized */
470 #define DEX_OPT_FLAG_INVOCATIONS (1<<3) /* method calls optimized */
471
472 #define DEX_INTERFACE_CACHE_SIZE 128 /* must be power of 2 */
473
474 /*
475 * Structure representing a DEX file.
476 *
477 * Code should regard DexFile as opaque, using the API calls provided here
478 * to access specific structures.
479 */
480 typedef struct DexFile {
481 /* directly-mapped "opt" header */
482 const DexOptHeader* pOptHeader;
483
484 /* pointers to directly-mapped structs and arrays in base DEX */
485 const DexHeader* pHeader;
486 const DexStringId* pStringIds;
487 const DexTypeId* pTypeIds;
488 const DexFieldId* pFieldIds;
489 const DexMethodId* pMethodIds;
490 const DexProtoId* pProtoIds;
491 const DexClassDef* pClassDefs;
492 const DexLink* pLinkData;
493
494 /*
495 * These are mapped out of the "auxillary" section, and may not be
496 * included in the file.
497 */
498 const DexClassLookup* pClassLookup;
499 const void* pRegisterMapPool; // RegisterMapClassPool
500
501 /* points to start of DEX file data */
502 const u1* baseAddr;
503
504 /* track memory overhead for auxillary structures */
505 int overhead;
506
507 /* additional app-specific data structures associated with the DEX */
508 //void* auxData;
509 } DexFile;
510
511 /*
512 * Utility function -- rounds up to the nearest power of 2.
513 */
514 u4 dexRoundUpPower2(u4 val);
515
516 /*
517 * Parse an optimized or unoptimized .dex file sitting in memory.
518 *
519 * On success, return a newly-allocated DexFile.
520 */
521 DexFile* dexFileParse(const u1* data, size_t length, int flags);
522
523 /* bit values for "flags" argument to dexFileParse */
524 enum {
525 kDexParseDefault = 0,
526 kDexParseVerifyChecksum = 1,
527 kDexParseContinueOnError = (1 << 1),
528 };
529
530 /*
531 * Fix the byte ordering of all fields in the DEX file, and do
532 * structural verification. This is only required for code that opens
533 * "raw" DEX files, such as the DEX optimizer.
534 *
535 * Return 0 on success.
536 */
537 int dexSwapAndVerify(u1* addr, int len);
538
539 /*
540 * Detect the file type of the given memory buffer via magic number.
541 * Call dexSwapAndVerify() on an unoptimized DEX file, do nothing
542 * but return successfully on an optimized DEX file, and report an
543 * error for all other cases.
544 *
545 * Return 0 on success.
546 */
547 int dexSwapAndVerifyIfNecessary(u1* addr, int len);
548
549 /*
550 * Compute DEX checksum.
551 */
552 u4 dexComputeChecksum(const DexHeader* pHeader);
553
554 /*
555 * Free a DexFile structure, along with any associated structures.
556 */
557 void dexFileFree(DexFile* pDexFile);
558
559 /*
560 * Create class lookup table.
561 */
562 DexClassLookup* dexCreateClassLookup(DexFile* pDexFile);
563
564 /*
565 * Find a class definition by descriptor.
566 */
567 const DexClassDef* dexFindClass(const DexFile* pFile, const char* descriptor);
568
569 /*
570 * Set up the basic raw data pointers of a DexFile. This function isn't
571 * meant for general use.
572 */
573 void dexFileSetupBasicPointers(DexFile* pDexFile, const u1* data);
574
575 /* return the DexMapList of the file, if any */
dexGetMap(const DexFile * pDexFile)576 DEX_INLINE const DexMapList* dexGetMap(const DexFile* pDexFile) {
577 u4 mapOff = pDexFile->pHeader->mapOff;
578
579 if (mapOff == 0) {
580 return NULL;
581 } else {
582 return (const DexMapList*) (pDexFile->baseAddr + mapOff);
583 }
584 }
585
586 /* return the const char* string data referred to by the given string_id */
dexGetStringData(const DexFile * pDexFile,const DexStringId * pStringId)587 DEX_INLINE const char* dexGetStringData(const DexFile* pDexFile,
588 const DexStringId* pStringId) {
589 const u1* ptr = pDexFile->baseAddr + pStringId->stringDataOff;
590
591 // Skip the uleb128 length.
592 while (*(ptr++) > 0x7f) /* empty */ ;
593
594 return (const char*) ptr;
595 }
596 /* return the StringId with the specified index */
dexGetStringId(const DexFile * pDexFile,u4 idx)597 DEX_INLINE const DexStringId* dexGetStringId(const DexFile* pDexFile, u4 idx) {
598 assert(idx < pDexFile->pHeader->stringIdsSize);
599 return &pDexFile->pStringIds[idx];
600 }
601 /* return the UTF-8 encoded string with the specified string_id index */
dexStringById(const DexFile * pDexFile,u4 idx)602 DEX_INLINE const char* dexStringById(const DexFile* pDexFile, u4 idx) {
603 const DexStringId* pStringId = dexGetStringId(pDexFile, idx);
604 return dexGetStringData(pDexFile, pStringId);
605 }
606
607 /* Return the UTF-8 encoded string with the specified string_id index,
608 * also filling in the UTF-16 size (number of 16-bit code points).*/
609 const char* dexStringAndSizeById(const DexFile* pDexFile, u4 idx,
610 u4* utf16Size);
611
612 /* return the TypeId with the specified index */
dexGetTypeId(const DexFile * pDexFile,u4 idx)613 DEX_INLINE const DexTypeId* dexGetTypeId(const DexFile* pDexFile, u4 idx) {
614 assert(idx < pDexFile->pHeader->typeIdsSize);
615 return &pDexFile->pTypeIds[idx];
616 }
617
618 /*
619 * Get the descriptor string associated with a given type index.
620 * The caller should not free() the returned string.
621 */
dexStringByTypeIdx(const DexFile * pDexFile,u4 idx)622 DEX_INLINE const char* dexStringByTypeIdx(const DexFile* pDexFile, u4 idx) {
623 const DexTypeId* typeId = dexGetTypeId(pDexFile, idx);
624 return dexStringById(pDexFile, typeId->descriptorIdx);
625 }
626
627 /* return the MethodId with the specified index */
dexGetMethodId(const DexFile * pDexFile,u4 idx)628 DEX_INLINE const DexMethodId* dexGetMethodId(const DexFile* pDexFile, u4 idx) {
629 assert(idx < pDexFile->pHeader->methodIdsSize);
630 return &pDexFile->pMethodIds[idx];
631 }
632
633 /* return the FieldId with the specified index */
dexGetFieldId(const DexFile * pDexFile,u4 idx)634 DEX_INLINE const DexFieldId* dexGetFieldId(const DexFile* pDexFile, u4 idx) {
635 assert(idx < pDexFile->pHeader->fieldIdsSize);
636 return &pDexFile->pFieldIds[idx];
637 }
638
639 /* return the ProtoId with the specified index */
dexGetProtoId(const DexFile * pDexFile,u4 idx)640 DEX_INLINE const DexProtoId* dexGetProtoId(const DexFile* pDexFile, u4 idx) {
641 assert(idx < pDexFile->pHeader->protoIdsSize);
642 return &pDexFile->pProtoIds[idx];
643 }
644
645 /*
646 * Get the parameter list from a ProtoId. The returns NULL if the ProtoId
647 * does not have a parameter list.
648 */
dexGetProtoParameters(const DexFile * pDexFile,const DexProtoId * pProtoId)649 DEX_INLINE const DexTypeList* dexGetProtoParameters(
650 const DexFile *pDexFile, const DexProtoId* pProtoId) {
651 if (pProtoId->parametersOff == 0) {
652 return NULL;
653 }
654 return (const DexTypeList*)
655 (pDexFile->baseAddr + pProtoId->parametersOff);
656 }
657
658 /* return the ClassDef with the specified index */
dexGetClassDef(const DexFile * pDexFile,u4 idx)659 DEX_INLINE const DexClassDef* dexGetClassDef(const DexFile* pDexFile, u4 idx) {
660 assert(idx < pDexFile->pHeader->classDefsSize);
661 return &pDexFile->pClassDefs[idx];
662 }
663
664 /* given a ClassDef pointer, recover its index */
dexGetIndexForClassDef(const DexFile * pDexFile,const DexClassDef * pClassDef)665 DEX_INLINE u4 dexGetIndexForClassDef(const DexFile* pDexFile,
666 const DexClassDef* pClassDef)
667 {
668 assert(pClassDef >= pDexFile->pClassDefs &&
669 pClassDef < pDexFile->pClassDefs + pDexFile->pHeader->classDefsSize);
670 return pClassDef - pDexFile->pClassDefs;
671 }
672
673 /* get the interface list for a DexClass */
dexGetInterfacesList(const DexFile * pDexFile,const DexClassDef * pClassDef)674 DEX_INLINE const DexTypeList* dexGetInterfacesList(const DexFile* pDexFile,
675 const DexClassDef* pClassDef)
676 {
677 if (pClassDef->interfacesOff == 0)
678 return NULL;
679 return (const DexTypeList*)
680 (pDexFile->baseAddr + pClassDef->interfacesOff);
681 }
682 /* return the Nth entry in a DexTypeList. */
dexGetTypeItem(const DexTypeList * pList,u4 idx)683 DEX_INLINE const DexTypeItem* dexGetTypeItem(const DexTypeList* pList,
684 u4 idx)
685 {
686 assert(idx < pList->size);
687 return &pList->list[idx];
688 }
689 /* return the type_idx for the Nth entry in a TypeList */
dexTypeListGetIdx(const DexTypeList * pList,u4 idx)690 DEX_INLINE u4 dexTypeListGetIdx(const DexTypeList* pList, u4 idx) {
691 const DexTypeItem* pItem = dexGetTypeItem(pList, idx);
692 return pItem->typeIdx;
693 }
694
695 /* get the static values list for a DexClass */
dexGetStaticValuesList(const DexFile * pDexFile,const DexClassDef * pClassDef)696 DEX_INLINE const DexEncodedArray* dexGetStaticValuesList(
697 const DexFile* pDexFile, const DexClassDef* pClassDef)
698 {
699 if (pClassDef->staticValuesOff == 0)
700 return NULL;
701 return (const DexEncodedArray*)
702 (pDexFile->baseAddr + pClassDef->staticValuesOff);
703 }
704
705 /* get the annotations directory item for a DexClass */
dexGetAnnotationsDirectoryItem(const DexFile * pDexFile,const DexClassDef * pClassDef)706 DEX_INLINE const DexAnnotationsDirectoryItem* dexGetAnnotationsDirectoryItem(
707 const DexFile* pDexFile, const DexClassDef* pClassDef)
708 {
709 if (pClassDef->annotationsOff == 0)
710 return NULL;
711 return (const DexAnnotationsDirectoryItem*)
712 (pDexFile->baseAddr + pClassDef->annotationsOff);
713 }
714
715 /* get the source file string */
dexGetSourceFile(const DexFile * pDexFile,const DexClassDef * pClassDef)716 DEX_INLINE const char* dexGetSourceFile(
717 const DexFile* pDexFile, const DexClassDef* pClassDef)
718 {
719 if (pClassDef->sourceFileIdx == 0xffffffff)
720 return NULL;
721 return dexStringById(pDexFile, pClassDef->sourceFileIdx);
722 }
723
724 /* get the size, in bytes, of a DexCode */
725 size_t dexGetDexCodeSize(const DexCode* pCode);
726
727 /* Get the list of "tries" for the given DexCode. */
dexGetTries(const DexCode * pCode)728 DEX_INLINE const DexTry* dexGetTries(const DexCode* pCode) {
729 const u2* insnsEnd = &pCode->insns[pCode->insnsSize];
730
731 // Round to four bytes.
732 if ((((u4) insnsEnd) & 3) != 0) {
733 insnsEnd++;
734 }
735
736 return (const DexTry*) insnsEnd;
737 }
738
739 /* Get the base of the encoded data for the given DexCode. */
dexGetCatchHandlerData(const DexCode * pCode)740 DEX_INLINE const u1* dexGetCatchHandlerData(const DexCode* pCode) {
741 const DexTry* pTries = dexGetTries(pCode);
742 return (const u1*) &pTries[pCode->triesSize];
743 }
744
745 /* get a pointer to the start of the debugging data */
dexGetDebugInfoStream(const DexFile * pDexFile,const DexCode * pCode)746 DEX_INLINE const u1* dexGetDebugInfoStream(const DexFile* pDexFile,
747 const DexCode* pCode)
748 {
749 if (pCode->debugInfoOff == 0) {
750 return NULL;
751 } else {
752 return pDexFile->baseAddr + pCode->debugInfoOff;
753 }
754 }
755
756 /*
757 * Callback for "new position table entry".
758 * Returning non-0 causes the decoder to stop early.
759 */
760 typedef int (*DexDebugNewPositionCb)(void *cnxt, u4 address, u4 lineNum);
761
762 /*
763 * Callback for "new locals table entry". "signature" is an empty string
764 * if no signature is available for an entry.
765 */
766 typedef void (*DexDebugNewLocalCb)(void *cnxt, u2 reg, u4 startAddress,
767 u4 endAddress, const char *name, const char *descriptor,
768 const char *signature);
769
770 /*
771 * Decode debug info for method.
772 *
773 * posCb is called in ascending address order.
774 * localCb is called in order of ascending end address.
775 */
776 void dexDecodeDebugInfo(
777 const DexFile* pDexFile,
778 const DexCode* pDexCode,
779 const char* classDescriptor,
780 u4 protoIdx,
781 u4 accessFlags,
782 DexDebugNewPositionCb posCb, DexDebugNewLocalCb localCb,
783 void* cnxt);
784
785 /* DexClassDef convenience - get class descriptor */
dexGetClassDescriptor(const DexFile * pDexFile,const DexClassDef * pClassDef)786 DEX_INLINE const char* dexGetClassDescriptor(const DexFile* pDexFile,
787 const DexClassDef* pClassDef)
788 {
789 return dexStringByTypeIdx(pDexFile, pClassDef->classIdx);
790 }
791
792 /* DexClassDef convenience - get superclass descriptor */
dexGetSuperClassDescriptor(const DexFile * pDexFile,const DexClassDef * pClassDef)793 DEX_INLINE const char* dexGetSuperClassDescriptor(const DexFile* pDexFile,
794 const DexClassDef* pClassDef)
795 {
796 if (pClassDef->superclassIdx == 0)
797 return NULL;
798 return dexStringByTypeIdx(pDexFile, pClassDef->superclassIdx);
799 }
800
801 /* DexClassDef convenience - get class_data_item pointer */
dexGetClassData(const DexFile * pDexFile,const DexClassDef * pClassDef)802 DEX_INLINE const u1* dexGetClassData(const DexFile* pDexFile,
803 const DexClassDef* pClassDef)
804 {
805 if (pClassDef->classDataOff == 0)
806 return NULL;
807 return (const u1*) (pDexFile->baseAddr + pClassDef->classDataOff);
808 }
809
810 /* Get an annotation set at a particular offset. */
dexGetAnnotationSetItem(const DexFile * pDexFile,u4 offset)811 DEX_INLINE const DexAnnotationSetItem* dexGetAnnotationSetItem(
812 const DexFile* pDexFile, u4 offset)
813 {
814 return (const DexAnnotationSetItem*) (pDexFile->baseAddr + offset);
815 }
816 /* get the class' annotation set */
dexGetClassAnnotationSet(const DexFile * pDexFile,const DexAnnotationsDirectoryItem * pAnnoDir)817 DEX_INLINE const DexAnnotationSetItem* dexGetClassAnnotationSet(
818 const DexFile* pDexFile, const DexAnnotationsDirectoryItem* pAnnoDir)
819 {
820 if (pAnnoDir->classAnnotationsOff == 0)
821 return NULL;
822 return dexGetAnnotationSetItem(pDexFile, pAnnoDir->classAnnotationsOff);
823 }
824
825 /* get the class' field annotation list */
dexGetFieldAnnotations(const DexFile * pDexFile,const DexAnnotationsDirectoryItem * pAnnoDir)826 DEX_INLINE const DexFieldAnnotationsItem* dexGetFieldAnnotations(
827 const DexFile* pDexFile, const DexAnnotationsDirectoryItem* pAnnoDir)
828 {
829 if (pAnnoDir->fieldsSize == 0)
830 return NULL;
831
832 // Skip past the header to the start of the field annotations.
833 return (const DexFieldAnnotationsItem*) &pAnnoDir[1];
834 }
835
836 /* get field annotation list size */
dexGetFieldAnnotationsSize(const DexFile * pDexFile,const DexAnnotationsDirectoryItem * pAnnoDir)837 DEX_INLINE int dexGetFieldAnnotationsSize(const DexFile* pDexFile,
838 const DexAnnotationsDirectoryItem* pAnnoDir)
839 {
840 return pAnnoDir->fieldsSize;
841 }
842
843 /* return a pointer to the field's annotation set */
dexGetFieldAnnotationSetItem(const DexFile * pDexFile,const DexFieldAnnotationsItem * pItem)844 DEX_INLINE const DexAnnotationSetItem* dexGetFieldAnnotationSetItem(
845 const DexFile* pDexFile, const DexFieldAnnotationsItem* pItem)
846 {
847 return dexGetAnnotationSetItem(pDexFile, pItem->annotationsOff);
848 }
849
850 /* get the class' method annotation list */
dexGetMethodAnnotations(const DexFile * pDexFile,const DexAnnotationsDirectoryItem * pAnnoDir)851 DEX_INLINE const DexMethodAnnotationsItem* dexGetMethodAnnotations(
852 const DexFile* pDexFile, const DexAnnotationsDirectoryItem* pAnnoDir)
853 {
854 if (pAnnoDir->methodsSize == 0)
855 return NULL;
856
857 /*
858 * Skip past the header and field annotations to the start of the
859 * method annotations.
860 */
861 const u1* addr = (const u1*) &pAnnoDir[1];
862 addr += pAnnoDir->fieldsSize * sizeof (DexFieldAnnotationsItem);
863 return (const DexMethodAnnotationsItem*) addr;
864 }
865
866 /* get method annotation list size */
dexGetMethodAnnotationsSize(const DexFile * pDexFile,const DexAnnotationsDirectoryItem * pAnnoDir)867 DEX_INLINE int dexGetMethodAnnotationsSize(const DexFile* pDexFile,
868 const DexAnnotationsDirectoryItem* pAnnoDir)
869 {
870 return pAnnoDir->methodsSize;
871 }
872
873 /* return a pointer to the method's annotation set */
dexGetMethodAnnotationSetItem(const DexFile * pDexFile,const DexMethodAnnotationsItem * pItem)874 DEX_INLINE const DexAnnotationSetItem* dexGetMethodAnnotationSetItem(
875 const DexFile* pDexFile, const DexMethodAnnotationsItem* pItem)
876 {
877 return dexGetAnnotationSetItem(pDexFile, pItem->annotationsOff);
878 }
879
880 /* get the class' parameter annotation list */
dexGetParameterAnnotations(const DexFile * pDexFile,const DexAnnotationsDirectoryItem * pAnnoDir)881 DEX_INLINE const DexParameterAnnotationsItem* dexGetParameterAnnotations(
882 const DexFile* pDexFile, const DexAnnotationsDirectoryItem* pAnnoDir)
883 {
884 if (pAnnoDir->parametersSize == 0)
885 return NULL;
886
887 /*
888 * Skip past the header, field annotations, and method annotations
889 * to the start of the parameter annotations.
890 */
891 const u1* addr = (const u1*) &pAnnoDir[1];
892 addr += pAnnoDir->fieldsSize * sizeof (DexFieldAnnotationsItem);
893 addr += pAnnoDir->methodsSize * sizeof (DexMethodAnnotationsItem);
894 return (const DexParameterAnnotationsItem*) addr;
895 }
896
897 /* get method annotation list size */
dexGetParameterAnnotationsSize(const DexFile * pDexFile,const DexAnnotationsDirectoryItem * pAnnoDir)898 DEX_INLINE int dexGetParameterAnnotationsSize(const DexFile* pDexFile,
899 const DexAnnotationsDirectoryItem* pAnnoDir)
900 {
901 return pAnnoDir->parametersSize;
902 }
903
904 /* return the parameter annotation ref list */
dexGetParameterAnnotationSetRefList(const DexFile * pDexFile,const DexParameterAnnotationsItem * pItem)905 DEX_INLINE const DexAnnotationSetRefList* dexGetParameterAnnotationSetRefList(
906 const DexFile* pDexFile, const DexParameterAnnotationsItem* pItem)
907 {
908 return (const DexAnnotationSetRefList*)
909 (pDexFile->baseAddr + pItem->annotationsOff);
910 }
911
912 /* get method annotation list size */
dexGetParameterAnnotationSetRefSize(const DexFile * pDexFile,const DexParameterAnnotationsItem * pItem)913 DEX_INLINE int dexGetParameterAnnotationSetRefSize(const DexFile* pDexFile,
914 const DexParameterAnnotationsItem* pItem)
915 {
916 if (pItem->annotationsOff == 0)
917 return 0;
918 return dexGetParameterAnnotationSetRefList(pDexFile, pItem)->size;
919 }
920
921 /* return the Nth entry from an annotation set ref list */
dexGetParameterAnnotationSetRef(const DexAnnotationSetRefList * pList,u4 idx)922 DEX_INLINE const DexAnnotationSetRefItem* dexGetParameterAnnotationSetRef(
923 const DexAnnotationSetRefList* pList, u4 idx)
924 {
925 assert(idx < pList->size);
926 return &pList->list[idx];
927 }
928
929 /* given a DexAnnotationSetRefItem, return the DexAnnotationSetItem */
dexGetSetRefItemItem(const DexFile * pDexFile,const DexAnnotationSetRefItem * pItem)930 DEX_INLINE const DexAnnotationSetItem* dexGetSetRefItemItem(
931 const DexFile* pDexFile, const DexAnnotationSetRefItem* pItem)
932 {
933 return dexGetAnnotationSetItem(pDexFile, pItem->annotationsOff);
934 }
935
936 /* return the Nth annotation offset from a DexAnnotationSetItem */
dexGetAnnotationOff(const DexAnnotationSetItem * pAnnoSet,u4 idx)937 DEX_INLINE u4 dexGetAnnotationOff(
938 const DexAnnotationSetItem* pAnnoSet, u4 idx)
939 {
940 assert(idx < pAnnoSet->size);
941 return pAnnoSet->entries[idx];
942 }
943
944 /* return the Nth annotation item from a DexAnnotationSetItem */
dexGetAnnotationItem(const DexFile * pDexFile,const DexAnnotationSetItem * pAnnoSet,u4 idx)945 DEX_INLINE const DexAnnotationItem* dexGetAnnotationItem(
946 const DexFile* pDexFile, const DexAnnotationSetItem* pAnnoSet, u4 idx)
947 {
948 return (const DexAnnotationItem*)
949 (pDexFile->baseAddr + dexGetAnnotationOff(pAnnoSet, idx));
950 }
951
952
953 /*
954 * ===========================================================================
955 * Utility Functions
956 * ===========================================================================
957 */
958
959 /*
960 * Retrieve the next UTF-16 character from a UTF-8 string.
961 *
962 * Advances "*pUtf8Ptr" to the start of the next character.
963 *
964 * WARNING: If a string is corrupted by dropping a '\0' in the middle
965 * of a 3-byte sequence, you can end up overrunning the buffer with
966 * reads (and possibly with the writes if the length was computed and
967 * cached before the damage). For performance reasons, this function
968 * assumes that the string being parsed is known to be valid (e.g., by
969 * already being verified). Most strings we process here are coming
970 * out of dex files or other internal translations, so the only real
971 * risk comes from the JNI NewStringUTF call.
972 */
dexGetUtf16FromUtf8(const char ** pUtf8Ptr)973 DEX_INLINE u2 dexGetUtf16FromUtf8(const char** pUtf8Ptr)
974 {
975 unsigned int one, two, three;
976
977 one = *(*pUtf8Ptr)++;
978 if ((one & 0x80) != 0) {
979 /* two- or three-byte encoding */
980 two = *(*pUtf8Ptr)++;
981 if ((one & 0x20) != 0) {
982 /* three-byte encoding */
983 three = *(*pUtf8Ptr)++;
984 return ((one & 0x0f) << 12) |
985 ((two & 0x3f) << 6) |
986 (three & 0x3f);
987 } else {
988 /* two-byte encoding */
989 return ((one & 0x1f) << 6) |
990 (two & 0x3f);
991 }
992 } else {
993 /* one-byte encoding */
994 return one;
995 }
996 }
997
998 /* Compare two '\0'-terminated modified UTF-8 strings, using Unicode
999 * code point values for comparison. This treats different encodings
1000 * for the same code point as equivalent, except that only a real '\0'
1001 * byte is considered the string terminator. The return value is as
1002 * for strcmp(). */
1003 int dexUtf8Cmp(const char* s1, const char* s2);
1004
1005
1006 /* for dexIsValidMemberNameUtf8(), a bit vector indicating valid low ascii */
1007 extern u4 DEX_MEMBER_VALID_LOW_ASCII[4];
1008
1009 /* Helper for dexIsValidMemberUtf8(); do not call directly. */
1010 bool dexIsValidMemberNameUtf8_0(const char** pUtf8Ptr);
1011
1012 /* Return whether the pointed-at modified-UTF-8 encoded character is
1013 * valid as part of a member name, updating the pointer to point past
1014 * the consumed character. This will consume two encoded UTF-16 code
1015 * points if the character is encoded as a surrogate pair. Also, if
1016 * this function returns false, then the given pointer may only have
1017 * been partially advanced. */
dexIsValidMemberNameUtf8(const char ** pUtf8Ptr)1018 DEX_INLINE bool dexIsValidMemberNameUtf8(const char** pUtf8Ptr) {
1019 u1 c = (u1) **pUtf8Ptr;
1020 if (c <= 0x7f) {
1021 // It's low-ascii, so check the table.
1022 u4 wordIdx = c >> 5;
1023 u4 bitIdx = c & 0x1f;
1024 (*pUtf8Ptr)++;
1025 return (DEX_MEMBER_VALID_LOW_ASCII[wordIdx] & (1 << bitIdx)) != 0;
1026 }
1027
1028 /*
1029 * It's a multibyte encoded character. Call a non-inline function
1030 * for the heavy lifting.
1031 */
1032 return dexIsValidMemberNameUtf8_0(pUtf8Ptr);
1033 }
1034
1035 /* Return whether the given string is a valid field or method name. */
1036 bool dexIsValidMemberName(const char* s);
1037
1038 /* Return whether the given string is a valid type descriptor. */
1039 bool dexIsValidTypeDescriptor(const char* s);
1040
1041 /* Return whether the given string is a valid reference descriptor. This
1042 * is true if dexIsValidTypeDescriptor() returns true and the descriptor
1043 * is for a class or array and not a primitive type. */
1044 bool dexIsReferenceDescriptor(const char* s);
1045
1046 /* Return whether the given string is a valid class descriptor. This
1047 * is true if dexIsValidTypeDescriptor() returns true and the descriptor
1048 * is for a class and not an array or primitive type. */
1049 bool dexIsClassDescriptor(const char* s);
1050
1051 /* Return whether the given string is a valid field type descriptor. This
1052 * is true if dexIsValidTypeDescriptor() returns true and the descriptor
1053 * is for anything but "void". */
1054 bool dexIsFieldDescriptor(const char* s);
1055
1056 #endif /*_LIBDEX_DEXFILE*/
1057