1 /*
2  * Copyright (C) 2015 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 #ifndef LIBZIPARCHIVE_ZIPARCHIVECOMMON_H_
18 #define LIBZIPARCHIVE_ZIPARCHIVECOMMON_H_
19 
20 #include "android-base/macros.h"
21 
22 #include <inttypes.h>
23 
24 #include <optional>
25 
26 // The "end of central directory" (EOCD) record. Each archive
27 // contains exactly once such record which appears at the end of
28 // the archive. It contains archive wide information like the
29 // number of entries in the archive and the offset to the central
30 // directory of the offset.
31 struct EocdRecord {
32   static const uint32_t kSignature = 0x06054b50;
33 
34   // End of central directory signature, should always be
35   // |kSignature|.
36   uint32_t eocd_signature;
37   // The number of the current "disk", i.e, the "disk" that this
38   // central directory is on.
39   //
40   // This implementation assumes that each archive spans a single
41   // disk only. i.e, that disk_num == 1.
42   uint16_t disk_num;
43   // The disk where the central directory starts.
44   //
45   // This implementation assumes that each archive spans a single
46   // disk only. i.e, that cd_start_disk == 1.
47   uint16_t cd_start_disk;
48   // The number of central directory records on this disk.
49   //
50   // This implementation assumes that each archive spans a single
51   // disk only. i.e, that num_records_on_disk == num_records.
52   uint16_t num_records_on_disk;
53   // The total number of central directory records.
54   uint16_t num_records;
55   // The size of the central directory (in bytes).
56   uint32_t cd_size;
57   // The offset of the start of the central directory, relative
58   // to the start of the file.
59   uint32_t cd_start_offset;
60   // Length of the central directory comment.
61   uint16_t comment_length;
62 
63  private:
64   EocdRecord() = default;
65   DISALLOW_COPY_AND_ASSIGN(EocdRecord);
66 } __attribute__((packed));
67 
68 // A structure representing the fixed length fields for a single
69 // record in the central directory of the archive. In addition to
70 // the fixed length fields listed here, each central directory
71 // record contains a variable length "file_name" and "extra_field"
72 // whose lengths are given by |file_name_length| and |extra_field_length|
73 // respectively.
74 struct CentralDirectoryRecord {
75   static const uint32_t kSignature = 0x02014b50;
76 
77   // The start of record signature. Must be |kSignature|.
78   uint32_t record_signature;
79   // Source tool version. Top byte gives source OS.
80   uint16_t version_made_by;
81   // Tool version. Ignored by this implementation.
82   uint16_t version_needed;
83   // The "general purpose bit flags" for this entry. The only
84   // flag value that we currently check for is the "data descriptor"
85   // flag.
86   uint16_t gpb_flags;
87   // The compression method for this entry, one of |kCompressStored|
88   // and |kCompressDeflated|.
89   uint16_t compression_method;
90   // The file modification time and date for this entry.
91   uint16_t last_mod_time;
92   uint16_t last_mod_date;
93   // The CRC-32 checksum for this entry.
94   uint32_t crc32;
95   // The compressed size (in bytes) of this entry.
96   uint32_t compressed_size;
97   // The uncompressed size (in bytes) of this entry.
98   uint32_t uncompressed_size;
99   // The length of the entry file name in bytes. The file name
100   // will appear immediately after this record.
101   uint16_t file_name_length;
102   // The length of the extra field info (in bytes). This data
103   // will appear immediately after the entry file name.
104   uint16_t extra_field_length;
105   // The length of the entry comment (in bytes). This data will
106   // appear immediately after the extra field.
107   uint16_t comment_length;
108   // The start disk for this entry. Ignored by this implementation).
109   uint16_t file_start_disk;
110   // File attributes. Ignored by this implementation.
111   uint16_t internal_file_attributes;
112   // File attributes. For archives created on Unix, the top bits are the mode.
113   uint32_t external_file_attributes;
114   // The offset to the local file header for this entry, from the
115   // beginning of this archive.
116   uint32_t local_file_header_offset;
117 
118  private:
119   CentralDirectoryRecord() = default;
120   DISALLOW_COPY_AND_ASSIGN(CentralDirectoryRecord);
121 } __attribute__((packed));
122 
123 // The local file header for a given entry. This duplicates information
124 // present in the central directory of the archive. It is an error for
125 // the information here to be different from the central directory
126 // information for a given entry.
127 struct LocalFileHeader {
128   static const uint32_t kSignature = 0x04034b50;
129 
130   // The local file header signature, must be |kSignature|.
131   uint32_t lfh_signature;
132   // Tool version. Ignored by this implementation.
133   uint16_t version_needed;
134   // The "general purpose bit flags" for this entry. The only
135   // flag value that we currently check for is the "data descriptor"
136   // flag.
137   uint16_t gpb_flags;
138   // The compression method for this entry, one of |kCompressStored|
139   // and |kCompressDeflated|.
140   uint16_t compression_method;
141   // The file modification time and date for this entry.
142   uint16_t last_mod_time;
143   uint16_t last_mod_date;
144   // The CRC-32 checksum for this entry.
145   uint32_t crc32;
146   // The compressed size (in bytes) of this entry.
147   uint32_t compressed_size;
148   // The uncompressed size (in bytes) of this entry.
149   uint32_t uncompressed_size;
150   // The length of the entry file name in bytes. The file name
151   // will appear immediately after this record.
152   uint16_t file_name_length;
153   // The length of the extra field info (in bytes). This data
154   // will appear immediately after the entry file name.
155   uint16_t extra_field_length;
156 
157  private:
158   LocalFileHeader() = default;
159   DISALLOW_COPY_AND_ASSIGN(LocalFileHeader);
160 } __attribute__((packed));
161 
162 struct DataDescriptor {
163   // The *optional* data descriptor start signature.
164   static const uint32_t kOptSignature = 0x08074b50;
165 
166   // CRC-32 checksum of the entry.
167   uint32_t crc32;
168 
169   // For ZIP64 format archives, the compressed and uncompressed sizes are 8
170   // bytes each. Also, the ZIP64 format MAY be used regardless of the size
171   // of a file.  When extracting, if the zip64 extended information extra field
172   // is present for the file the compressed and uncompressed sizes will be 8
173   // byte values.
174 
175   // Compressed size of the entry, the field can be either 4 bytes or 8 bytes
176   // in the zip file.
177   uint64_t compressed_size;
178   // Uncompressed size of the entry, the field can be either 4 bytes or 8 bytes
179   // in the zip file.
180   uint64_t uncompressed_size;
181 
182  private:
183   DataDescriptor() = default;
184   DISALLOW_COPY_AND_ASSIGN(DataDescriptor);
185 };
186 
187 // The zip64 end of central directory locator helps to find the zip64 EOCD.
188 struct Zip64EocdLocator {
189   static constexpr uint32_t kSignature = 0x07064b50;
190 
191   // The signature of zip64 eocd locator, must be |kSignature|
192   uint32_t locator_signature;
193   // The start disk of the zip64 eocd. This implementation assumes that each
194   // archive spans a single disk only.
195   uint32_t eocd_start_disk;
196   // The offset offset of the zip64 end of central directory record.
197   uint64_t zip64_eocd_offset;
198   // The total number of disks. This implementation assumes that each archive
199   // spans a single disk only.
200   uint32_t num_of_disks;
201 
202  private:
203   Zip64EocdLocator() = default;
204   DISALLOW_COPY_AND_ASSIGN(Zip64EocdLocator);
205 } __attribute__((packed));
206 
207 // The optional zip64 EOCD. If one of the fields in the end of central directory
208 // record is too small to hold required data, the field SHOULD be  set to -1
209 // (0xFFFF or 0xFFFFFFFF) and the ZIP64 format record SHOULD be created.
210 struct Zip64EocdRecord {
211   static constexpr uint32_t kSignature = 0x06064b50;
212 
213   // The signature of zip64 eocd record, must be |kSignature|
214   uint32_t record_signature;
215   // Size of zip64 end of central directory record. It SHOULD be the size of the
216   // remaining record and SHOULD NOT include the leading 12 bytes.
217   uint64_t record_size;
218   // The version of the tool that make this archive.
219   uint16_t version_made_by;
220   // Tool version needed to extract this archive.
221   uint16_t version_needed;
222   // Number of this disk.
223   uint32_t disk_num;
224   // Number of the disk with the start of the central directory.
225   uint32_t cd_start_disk;
226   // Total number of entries in the central directory on this disk.
227   // This implementation assumes that each archive spans a single
228   // disk only. i.e, that num_records_on_disk == num_records.
229   uint64_t num_records_on_disk;
230   // The total number of central directory records.
231   uint64_t num_records;
232   // The size of the central directory in bytes.
233   uint64_t cd_size;
234   // The offset of the start of the central directory, relative to the start of
235   // the file.
236   uint64_t cd_start_offset;
237 
238  private:
239   Zip64EocdRecord() = default;
240   DISALLOW_COPY_AND_ASSIGN(Zip64EocdRecord);
241 } __attribute__((packed));
242 
243 // The possible contents of the Zip64 Extended Information Extra Field. It may appear in
244 // the 'extra' field of a central directory record or local file header. The order of
245 // the fields in the zip64 extended information record is fixed, but the fields MUST
246 // only appear if the corresponding local or central directory record field is set to
247 // 0xFFFF or 0xFFFFFFFF. And this entry in the Local header MUST include BOTH original
248 // and compressed file size fields.
249 struct Zip64ExtendedInfo {
250   static constexpr uint16_t kHeaderId = 0x0001;
251   // The header tag for this 'extra' block, should be |kHeaderId|.
252   uint16_t header_id;
253   // The size in bytes of the remaining data (excluding the top 4 bytes).
254   uint16_t data_size;
255   // Size in bytes of the uncompressed file.
256   std::optional<uint64_t> uncompressed_file_size;
257   // Size in bytes of the compressed file.
258   std::optional<uint64_t> compressed_file_size;
259   // Local file header offset relative to the start of the zip file.
260   std::optional<uint64_t> local_header_offset;
261 
262   // This implementation assumes that each archive spans a single disk only. So
263   // the disk_number is not used.
264   // uint32_t disk_num;
265  private:
266   Zip64ExtendedInfo() = default;
267   DISALLOW_COPY_AND_ASSIGN(Zip64ExtendedInfo);
268 };
269 
270 // mask value that signifies that the entry has a DD
271 static const uint32_t kGPBDDFlagMask = 0x0008;
272 
273 // The maximum size of a central directory or a file
274 // comment in bytes.
275 static const uint32_t kMaxCommentLen = 65535;
276 
277 #endif /* LIBZIPARCHIVE_ZIPARCHIVECOMMON_H_ */
278