1 /*
2  * Copyright (c) 2018, 2019, 2020 Stefan Sperling <stsp@openbsd.org>
3  *
4  * Permission to use, copy, modify, and distribute this software for any
5  * purpose with or without fee is hereby granted, provided that the above
6  * copyright notice and this permission notice appear in all copies.
7  *
8  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
9  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
10  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
11  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
12  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
13  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
14  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
15  */
16 
17 struct got_pack_privsep_child {
18 	int imsg_fd;
19 	pid_t pid;
20 	struct imsgbuf *ibuf;
21 };
22 
23 /* An open pack file. */
24 struct got_pack {
25 	char *path_packfile;
26 	int fd;
27 	uint8_t *map;
28 	off_t filesize;
29 	struct got_pack_privsep_child *privsep_child;
30 	int basefd;
31 	int accumfd;
32 	int child_has_tempfiles;
33 	int child_has_delta_outfd;
34 	struct got_delta_cache *delta_cache;
35 };
36 
37 struct got_packidx;
38 
39 const struct got_error *got_pack_start_privsep_child(struct got_pack *,
40     struct got_packidx *);
41 const struct got_error *got_pack_close(struct got_pack *);
42 
43 const struct got_error *got_pack_parse_offset_delta(off_t *, size_t *,
44     struct got_pack *, off_t, size_t);
45 const struct got_error *got_pack_parse_ref_delta(struct got_object_id *,
46     struct got_pack *, off_t, int);
47 const struct got_error *got_pack_resolve_delta_chain(struct got_delta_chain *,
48     struct got_packidx *, struct got_pack *, off_t, size_t, int, size_t,
49     unsigned int);
50 const struct got_error *got_pack_parse_object_type_and_size(uint8_t *,
51     uint64_t *, size_t *, struct got_pack *, off_t);
52 
53 #define GOT_PACK_PREFIX		"pack-"
54 #define GOT_PACKFILE_SUFFIX	".pack"
55 #define GOT_PACKIDX_SUFFIX		".idx"
56 #define GOT_PACKFILE_NAMELEN	(strlen(GOT_PACK_PREFIX) + \
57 				SHA1_DIGEST_STRING_LENGTH - 1 + \
58 				strlen(GOT_PACKFILE_SUFFIX))
59 #define GOT_PACKIDX_NAMELEN	(strlen(GOT_PACK_PREFIX) + \
60 				SHA1_DIGEST_STRING_LENGTH - 1 + \
61 				strlen(GOT_PACKIDX_SUFFIX))
62 
63 /* See Documentation/technical/pack-format.txt in Git. */
64 
65 struct got_packidx_trailer {
66 	u_int8_t	packfile_sha1[SHA1_DIGEST_LENGTH];
67 	u_int8_t	packidx_sha1[SHA1_DIGEST_LENGTH];
68 } __attribute__((__packed__));
69 
70 struct got_packidx_object_id {
71 	u_int8_t sha1[SHA1_DIGEST_LENGTH];
72 } __attribute__((__packed__));
73 
74 /* Ignore pack index version 1 which is no longer written by Git. */
75 #define GOT_PACKIDX_VERSION 2
76 
77 struct got_packidx_v2_hdr {
78 	uint32_t	*magic;		/* big endian */
79 #define GOT_PACKIDX_V2_MAGIC 0xff744f63	/* "\377t0c" */
80 	uint32_t	*version;
81 
82 	/*
83 	 * Each entry N in the fanout table contains the number of objects in
84 	 * the packfile whose SHA1 begins with a byte less than or equal to N.
85 	 * The last entry (index 255) contains the number of objects in the
86 	 * pack file whose first SHA1 byte is <= 0xff, and thus records the
87 	 * total number of objects in the pack file. All pointer variables
88 	 * below point to tables with a corresponding number of entries.
89 	 */
90 	uint32_t	*fanout_table;	/* values are big endian */
91 #define GOT_PACKIDX_V2_FANOUT_TABLE_ITEMS (0xff + 1)
92 
93 	/* Sorted SHA1 checksums for each object in the pack file. */
94 	struct got_packidx_object_id *sorted_ids;
95 
96 	/* CRC32 of the packed representation of each object. */
97 	uint32_t	*crc32;
98 
99 	/* Offset into the pack file for each object. */
100 	uint32_t	*offsets;		/* values are big endian */
101 #define GOT_PACKIDX_OFFSET_VAL_MASK		0x7fffffff
102 #define GOT_PACKIDX_OFFSET_VAL_IS_LARGE_IDX	0x80000000
103 
104 	/* Large offsets table is empty for pack files < 2 GB. */
105 	uint64_t	*large_offsets;		/* values are big endian */
106 
107 	struct got_packidx_trailer *trailer;
108 };
109 
110 struct got_pack_offset_index {
111 	uint32_t offset;
112 	uint32_t idx;
113 };
114 
115 struct got_pack_large_offset_index {
116 	uint64_t offset;
117 	uint32_t idx;
118 };
119 
120 /* An open pack index file. */
121 struct got_packidx {
122 	char *path_packidx; /* actual on-disk path */
123 	int fd;
124 	uint8_t *map;
125 	size_t len;
126 	size_t nlargeobj;
127 	struct got_packidx_v2_hdr hdr; /* convenient pointers into map */
128 	struct got_pack_offset_index *sorted_offsets;
129 	struct got_pack_large_offset_index *sorted_large_offsets;
130 };
131 
132 struct got_packfile_hdr {
133 	uint32_t	signature;
134 #define GOT_PACKFILE_SIGNATURE	0x5041434b	/* 'P' 'A' 'C' 'K' */
135 	uint32_t	version;	/* big endian */
136 #define GOT_PACKFILE_VERSION 2
137 	uint32_t	nobjects;	/* big endian */
138 };
139 
140 struct got_packfile_obj_hdr {
141 	/*
142 	 * The object size field uses a variable length encoding:
143 	 * size0...sizeN form a 4+7+7+...+7 bit integer, where size0 is the
144 	 * least significant part and sizeN is the most significant part.
145 	 * If the MSB of a size byte is set, an additional size byte follows.
146 	 * Of the 7 remaining bits of size0, the first 3 bits indicate the
147 	 * object's type, and the remaining 4 bits contribute to the size.
148 	 */
149 	uint8_t *size;		/* variable length */
150 #define GOT_PACK_OBJ_SIZE_MORE		0x80
151 #define GOT_PACK_OBJ_SIZE0_TYPE_MASK	0x70 /* See struct got_object->type */
152 #define GOT_PACK_OBJ_SIZE0_TYPE_MASK_SHIFT	4
153 #define GOT_PACK_OBJ_SIZE0_VAL_MASK	0x0f
154 #define GOT_PACK_OBJ_SIZE_VAL_MASK	0x7f
155 };
156 
157 /* If object is not a DELTA type. */
158 struct got_packfile_object_data {
159 	uint8_t *data;	/* compressed */
160 };
161 
162 /* If object is of type	GOT_OBJ_TYPE_REF_DELTA. */
163 struct got_packfile_object_data_ref_delta {
164 	uint8_t sha1[SHA1_DIGEST_LENGTH];
165 	uint8_t *delta_data;		/* compressed */
166 };
167 
168 /* If object is of type GOT_OBJ_TYPE_OFFSET_DELTA. */
169 struct got_packfile_object_data_offset_delta {
170 	/*
171 	 * This offset is interpreted as a negative offset from
172 	 * the got_packfile_obj_hdr corresponding to this object.
173 	 * The size provided in the header specifies the amount
174 	 * of compressed delta data that follows.
175 	 *
176 	 * This field uses a variable length encoding of N bytes,
177 	 * where the MSB is always set except for the last byte.
178 	 * The value is encoded as a series of N 7 bit integers,
179 	 * which are concatenated, and if N > 1 the value 2^7 +
180 	 * 2^14 + ... + 2^(7 * (n-1)) is added to the result.
181 	 */
182 	uint8_t *offset;	/* variable length */
183 #define GOT_PACK_OBJ_DELTA_OFF_MORE		0x80
184 #define GOT_PACK_OBJ_DELTA_OFF_VAL_MASK		0x7f
185 	uint8_t *delta_data;		/* compressed */
186 };
187 
188 struct got_packfile_obj_data {
189 	union {
190 		struct got_packfile_object_data data;
191 		struct got_packfile_object_data_ref_delta ref_delta;
192 		struct got_packfile_object_data_offset_delta offset_delta;
193 	} __attribute__((__packed__));
194 } __attribute__((__packed__));
195 
196 const struct got_error *got_packidx_init_hdr(struct got_packidx *, int, off_t);
197 const struct got_error *got_packidx_open(struct got_packidx **,
198     int, const char *, int);
199 const struct got_error *got_packidx_close(struct got_packidx *);
200 const struct got_error *got_packidx_get_packfile_path(char **, const char *);
201 off_t got_packidx_get_object_offset(struct got_packidx *, int idx);
202 int got_packidx_get_object_idx(struct got_packidx *, struct got_object_id *);
203 const struct got_error *got_packidx_get_offset_idx(int *, struct got_packidx *,
204     off_t);
205 const struct got_error *got_packidx_get_object_id(struct got_object_id *,
206     struct got_packidx *, int);
207 const struct got_error *got_packidx_match_id_str_prefix(
208     struct got_object_id_queue *, struct got_packidx *, const char *);
209 
210 const struct got_error *got_packfile_open_object(struct got_object **,
211     struct got_pack *, struct got_packidx *, int, struct got_object_id *);
212 const struct got_error *got_pack_get_delta_chain_max_size(uint64_t *,
213     struct got_delta_chain *, struct got_pack *);
214 const struct got_error *got_pack_get_max_delta_object_size(uint64_t *,
215     struct got_object *, struct got_pack *);
216 const struct got_error *got_pack_dump_delta_chain_to_file(size_t *,
217     struct got_delta_chain *, struct got_pack *, FILE *, FILE *, FILE *);
218 const struct got_error *got_pack_dump_delta_chain_to_mem(uint8_t **, size_t *,
219     struct got_delta_chain *, struct got_pack *);
220 const struct got_error *got_packfile_extract_object(struct got_pack *,
221     struct got_object *, FILE *, FILE *, FILE *);
222 const struct got_error *got_packfile_extract_object_to_mem(uint8_t **, size_t *,
223     struct got_object *, struct got_pack *);
224 const struct got_error *got_packfile_extract_raw_delta(uint8_t **, size_t *,
225     size_t *, off_t *, off_t *, off_t *, struct got_object_id *, uint64_t *,
226     uint64_t *, struct got_pack *, struct got_packidx *, int);