1 | //===-- DataExtractor.cpp -------------------------------------------------===// |
2 | // |
3 | // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. |
4 | // See https://llvm.org/LICENSE.txt for license information. |
5 | // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception |
6 | // |
7 | //===----------------------------------------------------------------------===// |
8 | |
9 | #include "lldb/Utility/DataExtractor.h" |
10 | |
11 | #include "lldb/lldb-defines.h" |
12 | #include "lldb/lldb-enumerations.h" |
13 | #include "lldb/lldb-forward.h" |
14 | #include "lldb/lldb-types.h" |
15 | |
16 | #include "lldb/Utility/DataBuffer.h" |
17 | #include "lldb/Utility/DataBufferHeap.h" |
18 | #include "lldb/Utility/LLDBAssert.h" |
19 | #include "lldb/Utility/Log.h" |
20 | #include "lldb/Utility/Stream.h" |
21 | #include "lldb/Utility/StreamString.h" |
22 | #include "lldb/Utility/UUID.h" |
23 | |
24 | #include "llvm/ADT/ArrayRef.h" |
25 | #include "llvm/ADT/SmallVector.h" |
26 | #include "llvm/ADT/StringExtras.h" |
27 | #include "llvm/Support/LEB128.h" |
28 | #include "llvm/Support/MD5.h" |
29 | #include "llvm/Support/MathExtras.h" |
30 | |
31 | #include <algorithm> |
32 | #include <array> |
33 | #include <cassert> |
34 | #include <cstdint> |
35 | #include <string> |
36 | |
37 | #include <cctype> |
38 | #include <cinttypes> |
39 | #include <cstring> |
40 | |
41 | using namespace lldb; |
42 | using namespace lldb_private; |
43 | |
44 | static inline uint16_t ReadInt16(const unsigned char *ptr, offset_t offset) { |
45 | uint16_t value; |
46 | memcpy(dest: &value, src: ptr + offset, n: 2); |
47 | return value; |
48 | } |
49 | |
50 | static inline uint32_t ReadInt32(const unsigned char *ptr, |
51 | offset_t offset = 0) { |
52 | uint32_t value; |
53 | memcpy(dest: &value, src: ptr + offset, n: 4); |
54 | return value; |
55 | } |
56 | |
57 | static inline uint64_t ReadInt64(const unsigned char *ptr, |
58 | offset_t offset = 0) { |
59 | uint64_t value; |
60 | memcpy(dest: &value, src: ptr + offset, n: 8); |
61 | return value; |
62 | } |
63 | |
64 | static inline uint16_t ReadInt16(const void *ptr) { |
65 | uint16_t value; |
66 | memcpy(dest: &value, src: ptr, n: 2); |
67 | return value; |
68 | } |
69 | |
70 | static inline uint16_t ReadSwapInt16(const unsigned char *ptr, |
71 | offset_t offset) { |
72 | uint16_t value; |
73 | memcpy(dest: &value, src: ptr + offset, n: 2); |
74 | return llvm::byteswap<uint16_t>(V: value); |
75 | } |
76 | |
77 | static inline uint32_t ReadSwapInt32(const unsigned char *ptr, |
78 | offset_t offset) { |
79 | uint32_t value; |
80 | memcpy(dest: &value, src: ptr + offset, n: 4); |
81 | return llvm::byteswap<uint32_t>(V: value); |
82 | } |
83 | |
84 | static inline uint64_t ReadSwapInt64(const unsigned char *ptr, |
85 | offset_t offset) { |
86 | uint64_t value; |
87 | memcpy(dest: &value, src: ptr + offset, n: 8); |
88 | return llvm::byteswap<uint64_t>(V: value); |
89 | } |
90 | |
91 | static inline uint16_t ReadSwapInt16(const void *ptr) { |
92 | uint16_t value; |
93 | memcpy(dest: &value, src: ptr, n: 2); |
94 | return llvm::byteswap<uint16_t>(V: value); |
95 | } |
96 | |
97 | static inline uint32_t ReadSwapInt32(const void *ptr) { |
98 | uint32_t value; |
99 | memcpy(dest: &value, src: ptr, n: 4); |
100 | return llvm::byteswap<uint32_t>(V: value); |
101 | } |
102 | |
103 | static inline uint64_t ReadSwapInt64(const void *ptr) { |
104 | uint64_t value; |
105 | memcpy(dest: &value, src: ptr, n: 8); |
106 | return llvm::byteswap<uint64_t>(V: value); |
107 | } |
108 | |
109 | static inline uint64_t ReadMaxInt64(const uint8_t *data, size_t byte_size, |
110 | ByteOrder byte_order) { |
111 | uint64_t res = 0; |
112 | if (byte_order == eByteOrderBig) |
113 | for (size_t i = 0; i < byte_size; ++i) |
114 | res = (res << 8) | data[i]; |
115 | else { |
116 | assert(byte_order == eByteOrderLittle); |
117 | for (size_t i = 0; i < byte_size; ++i) |
118 | res = (res << 8) | data[byte_size - 1 - i]; |
119 | } |
120 | return res; |
121 | } |
122 | |
123 | DataExtractor::() |
124 | : m_byte_order(endian::InlHostByteOrder()), m_addr_size(sizeof(void *)), |
125 | m_data_sp() {} |
126 | |
127 | // This constructor allows us to use data that is owned by someone else. The |
128 | // data must stay around as long as this object is valid. |
129 | DataExtractor::(const void *data, offset_t length, |
130 | ByteOrder endian, uint32_t addr_size, |
131 | uint32_t target_byte_size /*=1*/) |
132 | : m_start(const_cast<uint8_t *>(static_cast<const uint8_t *>(data))), |
133 | m_end(const_cast<uint8_t *>(static_cast<const uint8_t *>(data)) + length), |
134 | m_byte_order(endian), m_addr_size(addr_size), m_data_sp(), |
135 | m_target_byte_size(target_byte_size) { |
136 | assert(addr_size >= 1 && addr_size <= 8); |
137 | } |
138 | |
139 | // Make a shared pointer reference to the shared data in "data_sp" and set the |
140 | // endian swapping setting to "swap", and the address size to "addr_size". The |
141 | // shared data reference will ensure the data lives as long as any |
142 | // DataExtractor objects exist that have a reference to this data. |
143 | DataExtractor::(const DataBufferSP &data_sp, ByteOrder endian, |
144 | uint32_t addr_size, |
145 | uint32_t target_byte_size /*=1*/) |
146 | : m_byte_order(endian), m_addr_size(addr_size), m_data_sp(), |
147 | m_target_byte_size(target_byte_size) { |
148 | assert(addr_size >= 1 && addr_size <= 8); |
149 | SetData(data_sp); |
150 | } |
151 | |
152 | // Initialize this object with a subset of the data bytes in "data". If "data" |
153 | // contains shared data, then a reference to this shared data will added and |
154 | // the shared data will stay around as long as any object contains a reference |
155 | // to that data. The endian swap and address size settings are copied from |
156 | // "data". |
157 | DataExtractor::(const DataExtractor &data, offset_t offset, |
158 | offset_t length, uint32_t target_byte_size /*=1*/) |
159 | : m_byte_order(data.m_byte_order), m_addr_size(data.m_addr_size), |
160 | m_data_sp(), m_target_byte_size(target_byte_size) { |
161 | assert(m_addr_size >= 1 && m_addr_size <= 8); |
162 | if (data.ValidOffset(offset)) { |
163 | offset_t bytes_available = data.GetByteSize() - offset; |
164 | if (length > bytes_available) |
165 | length = bytes_available; |
166 | SetData(data, offset, length); |
167 | } |
168 | } |
169 | |
170 | DataExtractor::(const DataExtractor &rhs) |
171 | : m_start(rhs.m_start), m_end(rhs.m_end), m_byte_order(rhs.m_byte_order), |
172 | m_addr_size(rhs.m_addr_size), m_data_sp(rhs.m_data_sp), |
173 | m_target_byte_size(rhs.m_target_byte_size) { |
174 | assert(m_addr_size >= 1 && m_addr_size <= 8); |
175 | } |
176 | |
177 | // Assignment operator |
178 | const DataExtractor &DataExtractor::(const DataExtractor &rhs) { |
179 | if (this != &rhs) { |
180 | m_start = rhs.m_start; |
181 | m_end = rhs.m_end; |
182 | m_byte_order = rhs.m_byte_order; |
183 | m_addr_size = rhs.m_addr_size; |
184 | m_data_sp = rhs.m_data_sp; |
185 | } |
186 | return *this; |
187 | } |
188 | |
189 | DataExtractor::() = default; |
190 | |
191 | // Clears the object contents back to a default invalid state, and release any |
192 | // references to shared data that this object may contain. |
193 | void DataExtractor::() { |
194 | m_start = nullptr; |
195 | m_end = nullptr; |
196 | m_byte_order = endian::InlHostByteOrder(); |
197 | m_addr_size = sizeof(void *); |
198 | m_data_sp.reset(); |
199 | } |
200 | |
201 | // If this object contains shared data, this function returns the offset into |
202 | // that shared data. Else zero is returned. |
203 | size_t DataExtractor::() const { |
204 | if (m_start != nullptr) { |
205 | const DataBuffer *data = m_data_sp.get(); |
206 | if (data != nullptr) { |
207 | const uint8_t *data_bytes = data->GetBytes(); |
208 | if (data_bytes != nullptr) { |
209 | assert(m_start >= data_bytes); |
210 | return m_start - data_bytes; |
211 | } |
212 | } |
213 | } |
214 | return 0; |
215 | } |
216 | |
217 | // Set the data with which this object will extract from to data starting at |
218 | // BYTES and set the length of the data to LENGTH bytes long. The data is |
219 | // externally owned must be around at least as long as this object points to |
220 | // the data. No copy of the data is made, this object just refers to this data |
221 | // and can extract from it. If this object refers to any shared data upon |
222 | // entry, the reference to that data will be released. Is SWAP is set to true, |
223 | // any data extracted will be endian swapped. |
224 | lldb::offset_t DataExtractor::(const void *bytes, offset_t length, |
225 | ByteOrder endian) { |
226 | m_byte_order = endian; |
227 | m_data_sp.reset(); |
228 | if (bytes == nullptr || length == 0) { |
229 | m_start = nullptr; |
230 | m_end = nullptr; |
231 | } else { |
232 | m_start = const_cast<uint8_t *>(static_cast<const uint8_t *>(bytes)); |
233 | m_end = m_start + length; |
234 | } |
235 | return GetByteSize(); |
236 | } |
237 | |
238 | // Assign the data for this object to be a subrange in "data" starting |
239 | // "data_offset" bytes into "data" and ending "data_length" bytes later. If |
240 | // "data_offset" is not a valid offset into "data", then this object will |
241 | // contain no bytes. If "data_offset" is within "data" yet "data_length" is too |
242 | // large, the length will be capped at the number of bytes remaining in "data". |
243 | // If "data" contains a shared pointer to other data, then a ref counted |
244 | // pointer to that data will be made in this object. If "data" doesn't contain |
245 | // a shared pointer to data, then the bytes referred to in "data" will need to |
246 | // exist at least as long as this object refers to those bytes. The address |
247 | // size and endian swap settings are copied from the current values in "data". |
248 | lldb::offset_t DataExtractor::(const DataExtractor &data, |
249 | offset_t data_offset, |
250 | offset_t data_length) { |
251 | m_addr_size = data.m_addr_size; |
252 | assert(m_addr_size >= 1 && m_addr_size <= 8); |
253 | // If "data" contains shared pointer to data, then we can use that |
254 | if (data.m_data_sp) { |
255 | m_byte_order = data.m_byte_order; |
256 | return SetData(data_sp: data.m_data_sp, offset: data.GetSharedDataOffset() + data_offset, |
257 | length: data_length); |
258 | } |
259 | |
260 | // We have a DataExtractor object that just has a pointer to bytes |
261 | if (data.ValidOffset(offset: data_offset)) { |
262 | if (data_length > data.GetByteSize() - data_offset) |
263 | data_length = data.GetByteSize() - data_offset; |
264 | return SetData(bytes: data.GetDataStart() + data_offset, length: data_length, |
265 | endian: data.GetByteOrder()); |
266 | } |
267 | return 0; |
268 | } |
269 | |
270 | // Assign the data for this object to be a subrange of the shared data in |
271 | // "data_sp" starting "data_offset" bytes into "data_sp" and ending |
272 | // "data_length" bytes later. If "data_offset" is not a valid offset into |
273 | // "data_sp", then this object will contain no bytes. If "data_offset" is |
274 | // within "data_sp" yet "data_length" is too large, the length will be capped |
275 | // at the number of bytes remaining in "data_sp". A ref counted pointer to the |
276 | // data in "data_sp" will be made in this object IF the number of bytes this |
277 | // object refers to in greater than zero (if at least one byte was available |
278 | // starting at "data_offset") to ensure the data stays around as long as it is |
279 | // needed. The address size and endian swap settings will remain unchanged from |
280 | // their current settings. |
281 | lldb::offset_t DataExtractor::(const DataBufferSP &data_sp, |
282 | offset_t data_offset, |
283 | offset_t data_length) { |
284 | m_start = m_end = nullptr; |
285 | |
286 | if (data_length > 0) { |
287 | m_data_sp = data_sp; |
288 | if (data_sp) { |
289 | const size_t data_size = data_sp->GetByteSize(); |
290 | if (data_offset < data_size) { |
291 | m_start = data_sp->GetBytes() + data_offset; |
292 | const size_t bytes_left = data_size - data_offset; |
293 | // Cap the length of we asked for too many |
294 | if (data_length <= bytes_left) |
295 | m_end = m_start + data_length; // We got all the bytes we wanted |
296 | else |
297 | m_end = m_start + bytes_left; // Not all the bytes requested were |
298 | // available in the shared data |
299 | } |
300 | } |
301 | } |
302 | |
303 | size_t new_size = GetByteSize(); |
304 | |
305 | // Don't hold a shared pointer to the data buffer if we don't share any valid |
306 | // bytes in the shared buffer. |
307 | if (new_size == 0) |
308 | m_data_sp.reset(); |
309 | |
310 | return new_size; |
311 | } |
312 | |
313 | // Extract a single unsigned char from the binary data and update the offset |
314 | // pointed to by "offset_ptr". |
315 | // |
316 | // RETURNS the byte that was extracted, or zero on failure. |
317 | uint8_t DataExtractor::(offset_t *offset_ptr) const { |
318 | const uint8_t *data = static_cast<const uint8_t *>(GetData(offset_ptr, length: 1)); |
319 | if (data) |
320 | return *data; |
321 | return 0; |
322 | } |
323 | |
324 | // Extract "count" unsigned chars from the binary data and update the offset |
325 | // pointed to by "offset_ptr". The extracted data is copied into "dst". |
326 | // |
327 | // RETURNS the non-nullptr buffer pointer upon successful extraction of |
328 | // all the requested bytes, or nullptr when the data is not available in the |
329 | // buffer due to being out of bounds, or insufficient data. |
330 | void *DataExtractor::(offset_t *offset_ptr, void *dst, |
331 | uint32_t count) const { |
332 | const uint8_t *data = |
333 | static_cast<const uint8_t *>(GetData(offset_ptr, length: count)); |
334 | if (data) { |
335 | // Copy the data into the buffer |
336 | memcpy(dest: dst, src: data, n: count); |
337 | // Return a non-nullptr pointer to the converted data as an indicator of |
338 | // success |
339 | return dst; |
340 | } |
341 | return nullptr; |
342 | } |
343 | |
344 | // Extract a single uint16_t from the data and update the offset pointed to by |
345 | // "offset_ptr". |
346 | // |
347 | // RETURNS the uint16_t that was extracted, or zero on failure. |
348 | uint16_t DataExtractor::(offset_t *offset_ptr) const { |
349 | uint16_t val = 0; |
350 | const uint8_t *data = |
351 | static_cast<const uint8_t *>(GetData(offset_ptr, length: sizeof(val))); |
352 | if (data) { |
353 | if (m_byte_order != endian::InlHostByteOrder()) |
354 | val = ReadSwapInt16(ptr: data); |
355 | else |
356 | val = ReadInt16(ptr: data); |
357 | } |
358 | return val; |
359 | } |
360 | |
361 | uint16_t DataExtractor::(offset_t *offset_ptr) const { |
362 | uint16_t val; |
363 | if (m_byte_order == endian::InlHostByteOrder()) |
364 | val = ReadInt16(ptr: m_start, offset: *offset_ptr); |
365 | else |
366 | val = ReadSwapInt16(ptr: m_start, offset: *offset_ptr); |
367 | *offset_ptr += sizeof(val); |
368 | return val; |
369 | } |
370 | |
371 | uint32_t DataExtractor::(offset_t *offset_ptr) const { |
372 | uint32_t val; |
373 | if (m_byte_order == endian::InlHostByteOrder()) |
374 | val = ReadInt32(ptr: m_start, offset: *offset_ptr); |
375 | else |
376 | val = ReadSwapInt32(ptr: m_start, offset: *offset_ptr); |
377 | *offset_ptr += sizeof(val); |
378 | return val; |
379 | } |
380 | |
381 | uint64_t DataExtractor::(offset_t *offset_ptr) const { |
382 | uint64_t val; |
383 | if (m_byte_order == endian::InlHostByteOrder()) |
384 | val = ReadInt64(ptr: m_start, offset: *offset_ptr); |
385 | else |
386 | val = ReadSwapInt64(ptr: m_start, offset: *offset_ptr); |
387 | *offset_ptr += sizeof(val); |
388 | return val; |
389 | } |
390 | |
391 | // Extract "count" uint16_t values from the binary data and update the offset |
392 | // pointed to by "offset_ptr". The extracted data is copied into "dst". |
393 | // |
394 | // RETURNS the non-nullptr buffer pointer upon successful extraction of |
395 | // all the requested bytes, or nullptr when the data is not available in the |
396 | // buffer due to being out of bounds, or insufficient data. |
397 | void *DataExtractor::(offset_t *offset_ptr, void *void_dst, |
398 | uint32_t count) const { |
399 | const size_t src_size = sizeof(uint16_t) * count; |
400 | const uint16_t *src = |
401 | static_cast<const uint16_t *>(GetData(offset_ptr, length: src_size)); |
402 | if (src) { |
403 | if (m_byte_order != endian::InlHostByteOrder()) { |
404 | uint16_t *dst_pos = static_cast<uint16_t *>(void_dst); |
405 | uint16_t *dst_end = dst_pos + count; |
406 | const uint16_t *src_pos = src; |
407 | while (dst_pos < dst_end) { |
408 | *dst_pos = ReadSwapInt16(ptr: src_pos); |
409 | ++dst_pos; |
410 | ++src_pos; |
411 | } |
412 | } else { |
413 | memcpy(dest: void_dst, src: src, n: src_size); |
414 | } |
415 | // Return a non-nullptr pointer to the converted data as an indicator of |
416 | // success |
417 | return void_dst; |
418 | } |
419 | return nullptr; |
420 | } |
421 | |
422 | // Extract a single uint32_t from the data and update the offset pointed to by |
423 | // "offset_ptr". |
424 | // |
425 | // RETURNS the uint32_t that was extracted, or zero on failure. |
426 | uint32_t DataExtractor::(offset_t *offset_ptr) const { |
427 | uint32_t val = 0; |
428 | const uint8_t *data = |
429 | static_cast<const uint8_t *>(GetData(offset_ptr, length: sizeof(val))); |
430 | if (data) { |
431 | if (m_byte_order != endian::InlHostByteOrder()) { |
432 | val = ReadSwapInt32(ptr: data); |
433 | } else { |
434 | memcpy(dest: &val, src: data, n: 4); |
435 | } |
436 | } |
437 | return val; |
438 | } |
439 | |
440 | // Extract "count" uint32_t values from the binary data and update the offset |
441 | // pointed to by "offset_ptr". The extracted data is copied into "dst". |
442 | // |
443 | // RETURNS the non-nullptr buffer pointer upon successful extraction of |
444 | // all the requested bytes, or nullptr when the data is not available in the |
445 | // buffer due to being out of bounds, or insufficient data. |
446 | void *DataExtractor::(offset_t *offset_ptr, void *void_dst, |
447 | uint32_t count) const { |
448 | const size_t src_size = sizeof(uint32_t) * count; |
449 | const uint32_t *src = |
450 | static_cast<const uint32_t *>(GetData(offset_ptr, length: src_size)); |
451 | if (src) { |
452 | if (m_byte_order != endian::InlHostByteOrder()) { |
453 | uint32_t *dst_pos = static_cast<uint32_t *>(void_dst); |
454 | uint32_t *dst_end = dst_pos + count; |
455 | const uint32_t *src_pos = src; |
456 | while (dst_pos < dst_end) { |
457 | *dst_pos = ReadSwapInt32(ptr: src_pos); |
458 | ++dst_pos; |
459 | ++src_pos; |
460 | } |
461 | } else { |
462 | memcpy(dest: void_dst, src: src, n: src_size); |
463 | } |
464 | // Return a non-nullptr pointer to the converted data as an indicator of |
465 | // success |
466 | return void_dst; |
467 | } |
468 | return nullptr; |
469 | } |
470 | |
471 | // Extract a single uint64_t from the data and update the offset pointed to by |
472 | // "offset_ptr". |
473 | // |
474 | // RETURNS the uint64_t that was extracted, or zero on failure. |
475 | uint64_t DataExtractor::(offset_t *offset_ptr) const { |
476 | uint64_t val = 0; |
477 | const uint8_t *data = |
478 | static_cast<const uint8_t *>(GetData(offset_ptr, length: sizeof(val))); |
479 | if (data) { |
480 | if (m_byte_order != endian::InlHostByteOrder()) { |
481 | val = ReadSwapInt64(ptr: data); |
482 | } else { |
483 | memcpy(dest: &val, src: data, n: 8); |
484 | } |
485 | } |
486 | return val; |
487 | } |
488 | |
489 | // GetU64 |
490 | // |
491 | // Get multiple consecutive 64 bit values. Return true if the entire read |
492 | // succeeds and increment the offset pointed to by offset_ptr, else return |
493 | // false and leave the offset pointed to by offset_ptr unchanged. |
494 | void *DataExtractor::(offset_t *offset_ptr, void *void_dst, |
495 | uint32_t count) const { |
496 | const size_t src_size = sizeof(uint64_t) * count; |
497 | const uint64_t *src = |
498 | static_cast<const uint64_t *>(GetData(offset_ptr, length: src_size)); |
499 | if (src) { |
500 | if (m_byte_order != endian::InlHostByteOrder()) { |
501 | uint64_t *dst_pos = static_cast<uint64_t *>(void_dst); |
502 | uint64_t *dst_end = dst_pos + count; |
503 | const uint64_t *src_pos = src; |
504 | while (dst_pos < dst_end) { |
505 | *dst_pos = ReadSwapInt64(ptr: src_pos); |
506 | ++dst_pos; |
507 | ++src_pos; |
508 | } |
509 | } else { |
510 | memcpy(dest: void_dst, src: src, n: src_size); |
511 | } |
512 | // Return a non-nullptr pointer to the converted data as an indicator of |
513 | // success |
514 | return void_dst; |
515 | } |
516 | return nullptr; |
517 | } |
518 | |
519 | uint32_t DataExtractor::(offset_t *offset_ptr, |
520 | size_t byte_size) const { |
521 | lldbassert(byte_size > 0 && byte_size <= 4 && "GetMaxU32 invalid byte_size!" ); |
522 | return GetMaxU64(offset_ptr, byte_size); |
523 | } |
524 | |
525 | uint64_t DataExtractor::(offset_t *offset_ptr, |
526 | size_t byte_size) const { |
527 | lldbassert(byte_size > 0 && byte_size <= 8 && "GetMaxU64 invalid byte_size!" ); |
528 | switch (byte_size) { |
529 | case 1: |
530 | return GetU8(offset_ptr); |
531 | case 2: |
532 | return GetU16(offset_ptr); |
533 | case 4: |
534 | return GetU32(offset_ptr); |
535 | case 8: |
536 | return GetU64(offset_ptr); |
537 | default: { |
538 | // General case. |
539 | const uint8_t *data = |
540 | static_cast<const uint8_t *>(GetData(offset_ptr, length: byte_size)); |
541 | if (data == nullptr) |
542 | return 0; |
543 | return ReadMaxInt64(data, byte_size, byte_order: m_byte_order); |
544 | } |
545 | } |
546 | return 0; |
547 | } |
548 | |
549 | uint64_t DataExtractor::(offset_t *offset_ptr, |
550 | size_t byte_size) const { |
551 | switch (byte_size) { |
552 | case 1: |
553 | return GetU8_unchecked(offset_ptr); |
554 | case 2: |
555 | return GetU16_unchecked(offset_ptr); |
556 | case 4: |
557 | return GetU32_unchecked(offset_ptr); |
558 | case 8: |
559 | return GetU64_unchecked(offset_ptr); |
560 | default: { |
561 | uint64_t res = ReadMaxInt64(data: &m_start[*offset_ptr], byte_size, byte_order: m_byte_order); |
562 | *offset_ptr += byte_size; |
563 | return res; |
564 | } |
565 | } |
566 | return 0; |
567 | } |
568 | |
569 | int64_t DataExtractor::(offset_t *offset_ptr, size_t byte_size) const { |
570 | uint64_t u64 = GetMaxU64(offset_ptr, byte_size); |
571 | return llvm::SignExtend64(X: u64, B: 8 * byte_size); |
572 | } |
573 | |
574 | uint64_t DataExtractor::(offset_t *offset_ptr, size_t size, |
575 | uint32_t bitfield_bit_size, |
576 | uint32_t bitfield_bit_offset) const { |
577 | assert(bitfield_bit_size <= 64); |
578 | uint64_t uval64 = GetMaxU64(offset_ptr, byte_size: size); |
579 | |
580 | if (bitfield_bit_size == 0) |
581 | return uval64; |
582 | |
583 | int32_t lsbcount = bitfield_bit_offset; |
584 | if (m_byte_order == eByteOrderBig) |
585 | lsbcount = size * 8 - bitfield_bit_offset - bitfield_bit_size; |
586 | |
587 | if (lsbcount > 0) |
588 | uval64 >>= lsbcount; |
589 | |
590 | uint64_t bitfield_mask = |
591 | (bitfield_bit_size == 64 |
592 | ? std::numeric_limits<uint64_t>::max() |
593 | : ((static_cast<uint64_t>(1) << bitfield_bit_size) - 1)); |
594 | if (!bitfield_mask && bitfield_bit_offset == 0 && bitfield_bit_size == 64) |
595 | return uval64; |
596 | |
597 | uval64 &= bitfield_mask; |
598 | |
599 | return uval64; |
600 | } |
601 | |
602 | int64_t DataExtractor::(offset_t *offset_ptr, size_t size, |
603 | uint32_t bitfield_bit_size, |
604 | uint32_t bitfield_bit_offset) const { |
605 | assert(size >= 1 && "GetMaxS64Bitfield size must be >= 1" ); |
606 | assert(size <= 8 && "GetMaxS64Bitfield size must be <= 8" ); |
607 | int64_t sval64 = GetMaxS64(offset_ptr, byte_size: size); |
608 | if (bitfield_bit_size == 0) |
609 | return sval64; |
610 | int32_t lsbcount = bitfield_bit_offset; |
611 | if (m_byte_order == eByteOrderBig) |
612 | lsbcount = size * 8 - bitfield_bit_offset - bitfield_bit_size; |
613 | if (lsbcount > 0) |
614 | sval64 >>= lsbcount; |
615 | uint64_t bitfield_mask = llvm::maskTrailingOnes<uint64_t>(N: bitfield_bit_size); |
616 | sval64 &= bitfield_mask; |
617 | // sign extend if needed |
618 | if (sval64 & ((static_cast<uint64_t>(1)) << (bitfield_bit_size - 1))) |
619 | sval64 |= ~bitfield_mask; |
620 | return sval64; |
621 | } |
622 | |
623 | float DataExtractor::(offset_t *offset_ptr) const { |
624 | return Get<float>(offset_ptr, fail_value: 0.0f); |
625 | } |
626 | |
627 | double DataExtractor::(offset_t *offset_ptr) const { |
628 | return Get<double>(offset_ptr, fail_value: 0.0); |
629 | } |
630 | |
631 | long double DataExtractor::(offset_t *offset_ptr) const { |
632 | long double val = 0.0; |
633 | #if defined(__i386__) || defined(__amd64__) || defined(__x86_64__) || \ |
634 | defined(_M_IX86) || defined(_M_IA64) || defined(_M_X64) |
635 | *offset_ptr += CopyByteOrderedData(src_offset: *offset_ptr, src_len: 10, dst: &val, dst_len: sizeof(val), |
636 | dst_byte_order: endian::InlHostByteOrder()); |
637 | #else |
638 | *offset_ptr += CopyByteOrderedData(*offset_ptr, sizeof(val), &val, |
639 | sizeof(val), endian::InlHostByteOrder()); |
640 | #endif |
641 | return val; |
642 | } |
643 | |
644 | // Extract a single address from the data and update the offset pointed to by |
645 | // "offset_ptr". The size of the extracted address comes from the |
646 | // "this->m_addr_size" member variable and should be set correctly prior to |
647 | // extracting any address values. |
648 | // |
649 | // RETURNS the address that was extracted, or zero on failure. |
650 | uint64_t DataExtractor::(offset_t *offset_ptr) const { |
651 | assert(m_addr_size >= 1 && m_addr_size <= 8); |
652 | return GetMaxU64(offset_ptr, byte_size: m_addr_size); |
653 | } |
654 | |
655 | uint64_t DataExtractor::(offset_t *offset_ptr) const { |
656 | assert(m_addr_size >= 1 && m_addr_size <= 8); |
657 | return GetMaxU64_unchecked(offset_ptr, byte_size: m_addr_size); |
658 | } |
659 | |
660 | size_t DataExtractor::(offset_t offset, offset_t length, |
661 | ByteOrder dst_byte_order, void *dst) const { |
662 | const uint8_t *src = PeekData(offset, length); |
663 | if (src) { |
664 | if (dst_byte_order != GetByteOrder()) { |
665 | // Validate that only a word- or register-sized dst is byte swapped |
666 | assert(length == 1 || length == 2 || length == 4 || length == 8 || |
667 | length == 10 || length == 16 || length == 32); |
668 | |
669 | for (uint32_t i = 0; i < length; ++i) |
670 | (static_cast<uint8_t *>(dst))[i] = src[length - i - 1]; |
671 | } else |
672 | ::memcpy(dest: dst, src: src, n: length); |
673 | return length; |
674 | } |
675 | return 0; |
676 | } |
677 | |
678 | // Extract data as it exists in target memory |
679 | lldb::offset_t DataExtractor::(offset_t offset, offset_t length, |
680 | void *dst) const { |
681 | const uint8_t *src = PeekData(offset, length); |
682 | if (src) { |
683 | ::memcpy(dest: dst, src: src, n: length); |
684 | return length; |
685 | } |
686 | return 0; |
687 | } |
688 | |
689 | // Extract data and swap if needed when doing the copy |
690 | lldb::offset_t |
691 | DataExtractor::(offset_t src_offset, offset_t src_len, |
692 | void *dst_void_ptr, offset_t dst_len, |
693 | ByteOrder dst_byte_order) const { |
694 | // Validate the source info |
695 | if (!ValidOffsetForDataOfSize(offset: src_offset, length: src_len)) |
696 | assert(ValidOffsetForDataOfSize(src_offset, src_len)); |
697 | assert(src_len > 0); |
698 | assert(m_byte_order == eByteOrderBig || m_byte_order == eByteOrderLittle); |
699 | |
700 | // Validate the destination info |
701 | assert(dst_void_ptr != nullptr); |
702 | assert(dst_len > 0); |
703 | assert(dst_byte_order == eByteOrderBig || dst_byte_order == eByteOrderLittle); |
704 | |
705 | // Validate that only a word- or register-sized dst is byte swapped |
706 | assert(dst_byte_order == m_byte_order || dst_len == 1 || dst_len == 2 || |
707 | dst_len == 4 || dst_len == 8 || dst_len == 10 || dst_len == 16 || |
708 | dst_len == 32); |
709 | |
710 | // Must have valid byte orders set in this object and for destination |
711 | if (!(dst_byte_order == eByteOrderBig || |
712 | dst_byte_order == eByteOrderLittle) || |
713 | !(m_byte_order == eByteOrderBig || m_byte_order == eByteOrderLittle)) |
714 | return 0; |
715 | |
716 | uint8_t *dst = static_cast<uint8_t *>(dst_void_ptr); |
717 | const uint8_t *src = PeekData(offset: src_offset, length: src_len); |
718 | if (src) { |
719 | if (dst_len >= src_len) { |
720 | // We are copying the entire value from src into dst. Calculate how many, |
721 | // if any, zeroes we need for the most significant bytes if "dst_len" is |
722 | // greater than "src_len"... |
723 | const size_t num_zeroes = dst_len - src_len; |
724 | if (dst_byte_order == eByteOrderBig) { |
725 | // Big endian, so we lead with zeroes... |
726 | if (num_zeroes > 0) |
727 | ::memset(s: dst, c: 0, n: num_zeroes); |
728 | // Then either copy or swap the rest |
729 | if (m_byte_order == eByteOrderBig) { |
730 | ::memcpy(dest: dst + num_zeroes, src: src, n: src_len); |
731 | } else { |
732 | for (uint32_t i = 0; i < src_len; ++i) |
733 | dst[i + num_zeroes] = src[src_len - 1 - i]; |
734 | } |
735 | } else { |
736 | // Little endian destination, so we lead the value bytes |
737 | if (m_byte_order == eByteOrderBig) { |
738 | for (uint32_t i = 0; i < src_len; ++i) |
739 | dst[i] = src[src_len - 1 - i]; |
740 | } else { |
741 | ::memcpy(dest: dst, src: src, n: src_len); |
742 | } |
743 | // And zero the rest... |
744 | if (num_zeroes > 0) |
745 | ::memset(s: dst + src_len, c: 0, n: num_zeroes); |
746 | } |
747 | return src_len; |
748 | } else { |
749 | // We are only copying some of the value from src into dst.. |
750 | |
751 | if (dst_byte_order == eByteOrderBig) { |
752 | // Big endian dst |
753 | if (m_byte_order == eByteOrderBig) { |
754 | // Big endian dst, with big endian src |
755 | ::memcpy(dest: dst, src: src + (src_len - dst_len), n: dst_len); |
756 | } else { |
757 | // Big endian dst, with little endian src |
758 | for (uint32_t i = 0; i < dst_len; ++i) |
759 | dst[i] = src[dst_len - 1 - i]; |
760 | } |
761 | } else { |
762 | // Little endian dst |
763 | if (m_byte_order == eByteOrderBig) { |
764 | // Little endian dst, with big endian src |
765 | for (uint32_t i = 0; i < dst_len; ++i) |
766 | dst[i] = src[src_len - 1 - i]; |
767 | } else { |
768 | // Little endian dst, with big endian src |
769 | ::memcpy(dest: dst, src: src, n: dst_len); |
770 | } |
771 | } |
772 | return dst_len; |
773 | } |
774 | } |
775 | return 0; |
776 | } |
777 | |
778 | // Extracts a variable length NULL terminated C string from the data at the |
779 | // offset pointed to by "offset_ptr". The "offset_ptr" will be updated with |
780 | // the offset of the byte that follows the NULL terminator byte. |
781 | // |
782 | // If the offset pointed to by "offset_ptr" is out of bounds, or if "length" is |
783 | // non-zero and there aren't enough available bytes, nullptr will be returned |
784 | // and "offset_ptr" will not be updated. |
785 | const char *DataExtractor::(offset_t *offset_ptr) const { |
786 | const char *start = reinterpret_cast<const char *>(PeekData(offset: *offset_ptr, length: 1)); |
787 | // Already at the end of the data. |
788 | if (!start) |
789 | return nullptr; |
790 | |
791 | const char *end = reinterpret_cast<const char *>(m_end); |
792 | |
793 | // Check all bytes for a null terminator that terminates a C string. |
794 | const char *terminator_or_end = std::find(first: start, last: end, val: '\0'); |
795 | |
796 | // We didn't find a null terminator, so return nullptr to indicate that there |
797 | // is no valid C string at that offset. |
798 | if (terminator_or_end == end) |
799 | return nullptr; |
800 | |
801 | // Update offset_ptr for the caller to point to the data behind the |
802 | // terminator (which is 1 byte long). |
803 | *offset_ptr += (terminator_or_end - start + 1UL); |
804 | return start; |
805 | } |
806 | |
807 | // Extracts a NULL terminated C string from the fixed length field of length |
808 | // "len" at the offset pointed to by "offset_ptr". The "offset_ptr" will be |
809 | // updated with the offset of the byte that follows the fixed length field. |
810 | // |
811 | // If the offset pointed to by "offset_ptr" is out of bounds, or if the offset |
812 | // plus the length of the field is out of bounds, or if the field does not |
813 | // contain a NULL terminator byte, nullptr will be returned and "offset_ptr" |
814 | // will not be updated. |
815 | const char *DataExtractor::(offset_t *offset_ptr, offset_t len) const { |
816 | const char *cstr = reinterpret_cast<const char *>(PeekData(offset: *offset_ptr, length: len)); |
817 | if (cstr != nullptr) { |
818 | if (memchr(s: cstr, c: '\0', n: len) == nullptr) { |
819 | return nullptr; |
820 | } |
821 | *offset_ptr += len; |
822 | return cstr; |
823 | } |
824 | return nullptr; |
825 | } |
826 | |
827 | // Peeks at a string in the contained data. No verification is done to make |
828 | // sure the entire string lies within the bounds of this object's data, only |
829 | // "offset" is verified to be a valid offset. |
830 | // |
831 | // Returns a valid C string pointer if "offset" is a valid offset in this |
832 | // object's data, else nullptr is returned. |
833 | const char *DataExtractor::(offset_t offset) const { |
834 | return reinterpret_cast<const char *>(PeekData(offset, length: 1)); |
835 | } |
836 | |
837 | // Extracts an unsigned LEB128 number from this object's data starting at the |
838 | // offset pointed to by "offset_ptr". The offset pointed to by "offset_ptr" |
839 | // will be updated with the offset of the byte following the last extracted |
840 | // byte. |
841 | // |
842 | // Returned the extracted integer value. |
843 | uint64_t DataExtractor::(offset_t *offset_ptr) const { |
844 | const uint8_t *src = PeekData(offset: *offset_ptr, length: 1); |
845 | if (src == nullptr) |
846 | return 0; |
847 | |
848 | unsigned byte_count = 0; |
849 | uint64_t result = llvm::decodeULEB128(p: src, n: &byte_count, end: m_end); |
850 | *offset_ptr += byte_count; |
851 | return result; |
852 | } |
853 | |
854 | // Extracts an signed LEB128 number from this object's data starting at the |
855 | // offset pointed to by "offset_ptr". The offset pointed to by "offset_ptr" |
856 | // will be updated with the offset of the byte following the last extracted |
857 | // byte. |
858 | // |
859 | // Returned the extracted integer value. |
860 | int64_t DataExtractor::(offset_t *offset_ptr) const { |
861 | const uint8_t *src = PeekData(offset: *offset_ptr, length: 1); |
862 | if (src == nullptr) |
863 | return 0; |
864 | |
865 | unsigned byte_count = 0; |
866 | int64_t result = llvm::decodeSLEB128(p: src, n: &byte_count, end: m_end); |
867 | *offset_ptr += byte_count; |
868 | return result; |
869 | } |
870 | |
871 | // Skips a ULEB128 number (signed or unsigned) from this object's data starting |
872 | // at the offset pointed to by "offset_ptr". The offset pointed to by |
873 | // "offset_ptr" will be updated with the offset of the byte following the last |
874 | // extracted byte. |
875 | // |
876 | // Returns the number of bytes consumed during the extraction. |
877 | uint32_t DataExtractor::(offset_t *offset_ptr) const { |
878 | uint32_t bytes_consumed = 0; |
879 | const uint8_t *src = PeekData(offset: *offset_ptr, length: 1); |
880 | if (src == nullptr) |
881 | return 0; |
882 | |
883 | const uint8_t *end = m_end; |
884 | |
885 | if (src < end) { |
886 | const uint8_t *src_pos = src; |
887 | while ((src_pos < end) && (*src_pos++ & 0x80)) |
888 | ++bytes_consumed; |
889 | *offset_ptr += src_pos - src; |
890 | } |
891 | return bytes_consumed; |
892 | } |
893 | |
894 | // Dumps bytes from this object's data to the stream "s" starting |
895 | // "start_offset" bytes into this data, and ending with the byte before |
896 | // "end_offset". "base_addr" will be added to the offset into the dumped data |
897 | // when showing the offset into the data in the output information. |
898 | // "num_per_line" objects of type "type" will be dumped with the option to |
899 | // override the format for each object with "type_format". "type_format" is a |
900 | // printf style formatting string. If "type_format" is nullptr, then an |
901 | // appropriate format string will be used for the supplied "type". If the |
902 | // stream "s" is nullptr, then the output will be send to Log(). |
903 | lldb::offset_t DataExtractor::(Log *log, offset_t start_offset, |
904 | offset_t length, uint64_t base_addr, |
905 | uint32_t num_per_line, |
906 | DataExtractor::Type type) const { |
907 | if (log == nullptr) |
908 | return start_offset; |
909 | |
910 | offset_t offset; |
911 | offset_t end_offset; |
912 | uint32_t count; |
913 | StreamString sstr; |
914 | for (offset = start_offset, end_offset = offset + length, count = 0; |
915 | ValidOffset(offset) && offset < end_offset; ++count) { |
916 | if ((count % num_per_line) == 0) { |
917 | // Print out any previous string |
918 | if (sstr.GetSize() > 0) { |
919 | log->PutString(str: sstr.GetString()); |
920 | sstr.Clear(); |
921 | } |
922 | // Reset string offset and fill the current line string with address: |
923 | if (base_addr != LLDB_INVALID_ADDRESS) |
924 | sstr.Printf(format: "0x%8.8" PRIx64 ":" , |
925 | static_cast<uint64_t>(base_addr + (offset - start_offset))); |
926 | } |
927 | |
928 | switch (type) { |
929 | case TypeUInt8: |
930 | sstr.Printf(format: " %2.2x" , GetU8(offset_ptr: &offset)); |
931 | break; |
932 | case TypeChar: { |
933 | char ch = GetU8(offset_ptr: &offset); |
934 | sstr.Printf(format: " %c" , llvm::isPrint(C: ch) ? ch : ' '); |
935 | } break; |
936 | case TypeUInt16: |
937 | sstr.Printf(format: " %4.4x" , GetU16(offset_ptr: &offset)); |
938 | break; |
939 | case TypeUInt32: |
940 | sstr.Printf(format: " %8.8x" , GetU32(offset_ptr: &offset)); |
941 | break; |
942 | case TypeUInt64: |
943 | sstr.Printf(format: " %16.16" PRIx64, GetU64(offset_ptr: &offset)); |
944 | break; |
945 | case TypePointer: |
946 | sstr.Printf(format: " 0x%" PRIx64, GetAddress(offset_ptr: &offset)); |
947 | break; |
948 | case TypeULEB128: |
949 | sstr.Printf(format: " 0x%" PRIx64, GetULEB128(offset_ptr: &offset)); |
950 | break; |
951 | case TypeSLEB128: |
952 | sstr.Printf(format: " %" PRId64, GetSLEB128(offset_ptr: &offset)); |
953 | break; |
954 | } |
955 | } |
956 | |
957 | if (!sstr.Empty()) |
958 | log->PutString(str: sstr.GetString()); |
959 | |
960 | return offset; // Return the offset at which we ended up |
961 | } |
962 | |
963 | size_t DataExtractor::(DataExtractor &dest_data) const { |
964 | if (m_data_sp) { |
965 | // we can pass along the SP to the data |
966 | dest_data.SetData(data_sp: m_data_sp); |
967 | } else { |
968 | const uint8_t *base_ptr = m_start; |
969 | size_t data_size = GetByteSize(); |
970 | dest_data.SetData(data_sp: DataBufferSP(new DataBufferHeap(base_ptr, data_size))); |
971 | } |
972 | return GetByteSize(); |
973 | } |
974 | |
975 | bool DataExtractor::(DataExtractor &rhs) { |
976 | if (rhs.GetByteOrder() != GetByteOrder()) |
977 | return false; |
978 | |
979 | if (rhs.GetByteSize() == 0) |
980 | return true; |
981 | |
982 | if (GetByteSize() == 0) |
983 | return (rhs.Copy(dest_data&: *this) > 0); |
984 | |
985 | size_t bytes = GetByteSize() + rhs.GetByteSize(); |
986 | |
987 | DataBufferHeap *buffer_heap_ptr = nullptr; |
988 | DataBufferSP buffer_sp(buffer_heap_ptr = new DataBufferHeap(bytes, 0)); |
989 | |
990 | if (!buffer_sp || buffer_heap_ptr == nullptr) |
991 | return false; |
992 | |
993 | uint8_t *bytes_ptr = buffer_heap_ptr->GetBytes(); |
994 | |
995 | memcpy(dest: bytes_ptr, src: GetDataStart(), n: GetByteSize()); |
996 | memcpy(dest: bytes_ptr + GetByteSize(), src: rhs.GetDataStart(), n: rhs.GetByteSize()); |
997 | |
998 | SetData(data_sp: buffer_sp); |
999 | |
1000 | return true; |
1001 | } |
1002 | |
1003 | bool DataExtractor::(void *buf, offset_t length) { |
1004 | if (buf == nullptr) |
1005 | return false; |
1006 | |
1007 | if (length == 0) |
1008 | return true; |
1009 | |
1010 | size_t bytes = GetByteSize() + length; |
1011 | |
1012 | DataBufferHeap *buffer_heap_ptr = nullptr; |
1013 | DataBufferSP buffer_sp(buffer_heap_ptr = new DataBufferHeap(bytes, 0)); |
1014 | |
1015 | if (!buffer_sp || buffer_heap_ptr == nullptr) |
1016 | return false; |
1017 | |
1018 | uint8_t *bytes_ptr = buffer_heap_ptr->GetBytes(); |
1019 | |
1020 | if (GetByteSize() > 0) |
1021 | memcpy(dest: bytes_ptr, src: GetDataStart(), n: GetByteSize()); |
1022 | |
1023 | memcpy(dest: bytes_ptr + GetByteSize(), src: buf, n: length); |
1024 | |
1025 | SetData(data_sp: buffer_sp); |
1026 | |
1027 | return true; |
1028 | } |
1029 | |
1030 | void DataExtractor::(llvm::SmallVectorImpl<uint8_t> &dest, |
1031 | uint64_t max_data) { |
1032 | if (max_data == 0) |
1033 | max_data = GetByteSize(); |
1034 | else |
1035 | max_data = std::min(a: max_data, b: GetByteSize()); |
1036 | |
1037 | llvm::MD5 md5; |
1038 | |
1039 | const llvm::ArrayRef<uint8_t> data(GetDataStart(), max_data); |
1040 | md5.update(Data: data); |
1041 | |
1042 | llvm::MD5::MD5Result result; |
1043 | md5.final(Result&: result); |
1044 | |
1045 | dest.clear(); |
1046 | dest.append(in_start: result.begin(), in_end: result.end()); |
1047 | } |
1048 | |