1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
|
/*
* Copyright (C) 2018 The Android Open Source Project
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "xz_utils.h"
#include <vector>
#include <mutex>
#include "base/array_ref.h"
#include "base/bit_utils.h"
#include "base/leb128.h"
#include "dwarf/writer.h"
// liblzma.
#include "7zCrc.h"
#include "Xz.h"
#include "XzCrc64.h"
#include "XzEnc.h"
namespace art {
namespace debug {
constexpr size_t kChunkSize = kPageSize;
static void XzInitCrc() {
static std::once_flag crc_initialized;
std::call_once(crc_initialized, []() {
CrcGenerateTable();
Crc64GenerateTable();
});
}
static void XzCompressChunk(ArrayRef<const uint8_t> src, std::vector<uint8_t>* dst) {
// Configure the compression library.
XzInitCrc();
CLzma2EncProps lzma2Props;
Lzma2EncProps_Init(&lzma2Props);
lzma2Props.lzmaProps.level = 1; // Fast compression.
Lzma2EncProps_Normalize(&lzma2Props);
CXzProps props;
XzProps_Init(&props);
props.lzma2Props = lzma2Props;
// Implement the required interface for communication (written in C so no virtual methods).
struct XzCallbacks : public ISeqInStream, public ISeqOutStream, public ICompressProgress {
static SRes ReadImpl(const ISeqInStream* p, void* buf, size_t* size) {
auto* ctx = static_cast<XzCallbacks*>(const_cast<ISeqInStream*>(p));
*size = std::min(*size, ctx->src_.size() - ctx->src_pos_);
memcpy(buf, ctx->src_.data() + ctx->src_pos_, *size);
ctx->src_pos_ += *size;
return SZ_OK;
}
static size_t WriteImpl(const ISeqOutStream* p, const void* buf, size_t size) {
auto* ctx = static_cast<const XzCallbacks*>(p);
const uint8_t* buffer = reinterpret_cast<const uint8_t*>(buf);
ctx->dst_->insert(ctx->dst_->end(), buffer, buffer + size);
return size;
}
static SRes ProgressImpl(const ICompressProgress* , UInt64, UInt64) {
return SZ_OK;
}
size_t src_pos_;
ArrayRef<const uint8_t> src_;
std::vector<uint8_t>* dst_;
};
XzCallbacks callbacks;
callbacks.Read = XzCallbacks::ReadImpl;
callbacks.Write = XzCallbacks::WriteImpl;
callbacks.Progress = XzCallbacks::ProgressImpl;
callbacks.src_pos_ = 0;
callbacks.src_ = src;
callbacks.dst_ = dst;
// Compress.
SRes res = Xz_Encode(&callbacks, &callbacks, &props, &callbacks);
CHECK_EQ(res, SZ_OK);
}
// Compress data while splitting it to smaller chunks to enable random-access reads.
// The XZ file format supports this well, but the compression library does not.
// Therefore compress the chunks separately and then glue them together manually.
//
// The XZ file format is described here: https://tukaani.org/xz/xz-file-format.txt
// In short, the file format is: [header] [compressed_block]* [index] [footer]
// Where [index] is: [num_records] ([compressed_size] [uncompressed_size])* [crc32]
//
void XzCompress(ArrayRef<const uint8_t> src, std::vector<uint8_t>* dst) {
uint8_t header[] = { 0xFD, '7', 'z', 'X', 'Z', 0, 0, 1, 0x69, 0x22, 0xDE, 0x36 };
uint8_t footer[] = { 0, 1, 'Y', 'Z' };
dst->insert(dst->end(), header, header + sizeof(header));
std::vector<uint8_t> tmp;
std::vector<uint32_t> index;
for (size_t offset = 0; offset < src.size(); offset += kChunkSize) {
size_t size = std::min(src.size() - offset, kChunkSize);
tmp.clear();
XzCompressChunk(src.SubArray(offset, size), &tmp);
DCHECK_EQ(memcmp(tmp.data(), header, sizeof(header)), 0);
DCHECK_EQ(memcmp(tmp.data() + tmp.size() - sizeof(footer), footer, sizeof(footer)), 0);
uint32_t* index_size = reinterpret_cast<uint32_t*>(tmp.data() + tmp.size() - 8);
DCHECK_ALIGNED(index_size, sizeof(uint32_t));
size_t index_offset = tmp.size() - 16 - *index_size * 4;
const uint8_t* index_ptr = tmp.data() + index_offset;
uint8_t index_indicator = *(index_ptr++);
CHECK_EQ(index_indicator, 0); // Mark the start of index (as opposed to compressed block).
uint32_t num_records = DecodeUnsignedLeb128(&index_ptr);
for (uint32_t i = 0; i < num_records; i++) {
index.push_back(DecodeUnsignedLeb128(&index_ptr)); // Compressed size.
index.push_back(DecodeUnsignedLeb128(&index_ptr)); // Uncompressed size.
}
// Copy the raw compressed block(s) located between the header and index.
dst->insert(dst->end(), tmp.data() + sizeof(header), tmp.data() + index_offset);
}
// Write the index.
uint32_t index_size_in_words;
{
tmp.clear();
dwarf::Writer<> writer(&tmp);
writer.PushUint8(0); // Index indicator.
writer.PushUleb128(static_cast<uint32_t>(index.size()) / 2); // Record count.
for (uint32_t i : index) {
writer.PushUleb128(i);
}
writer.Pad(4);
index_size_in_words = writer.size() / sizeof(uint32_t);
writer.PushUint32(CrcCalc(tmp.data(), tmp.size()));
dst->insert(dst->end(), tmp.begin(), tmp.end());
}
// Write the footer.
{
tmp.clear();
dwarf::Writer<> writer(&tmp);
writer.PushUint32(0); // CRC placeholder.
writer.PushUint32(index_size_in_words);
writer.PushData(footer, sizeof(footer));
writer.UpdateUint32(0, CrcCalc(tmp.data() + 4, 6));
dst->insert(dst->end(), tmp.begin(), tmp.end());
}
// Decompress the data back and check that we get the original.
if (kIsDebugBuild) {
std::vector<uint8_t> decompressed;
XzDecompress(ArrayRef<const uint8_t>(*dst), &decompressed);
DCHECK_EQ(decompressed.size(), src.size());
DCHECK_EQ(memcmp(decompressed.data(), src.data(), src.size()), 0);
}
}
void XzDecompress(ArrayRef<const uint8_t> src, std::vector<uint8_t>* dst) {
XzInitCrc();
std::unique_ptr<CXzUnpacker> state(new CXzUnpacker());
ISzAlloc alloc;
alloc.Alloc = [](ISzAllocPtr, size_t size) { return malloc(size); };
alloc.Free = [](ISzAllocPtr, void* ptr) { return free(ptr); };
XzUnpacker_Construct(state.get(), &alloc);
size_t src_offset = 0;
size_t dst_offset = 0;
ECoderStatus status;
do {
dst->resize(RoundUp(dst_offset + kPageSize / 4, kPageSize));
size_t src_remaining = src.size() - src_offset;
size_t dst_remaining = dst->size() - dst_offset;
int return_val = XzUnpacker_Code(state.get(),
dst->data() + dst_offset,
&dst_remaining,
src.data() + src_offset,
&src_remaining,
true,
CODER_FINISH_ANY,
&status);
CHECK_EQ(return_val, SZ_OK);
src_offset += src_remaining;
dst_offset += dst_remaining;
} while (status == CODER_STATUS_NOT_FINISHED);
CHECK_EQ(src_offset, src.size());
CHECK(XzUnpacker_IsStreamWasFinished(state.get()));
XzUnpacker_Free(state.get());
dst->resize(dst_offset);
}
} // namespace debug
} // namespace art
|
