hash.h

   1 #ifndef HASH_H
   2 #define HASH_H
   3
   4 #if defined(SHA1_APPLE)
   5 #include <CommonCrypto/CommonDigest.h>
   6 #elif defined(SHA1_OPENSSL)
   7 #  include <openssl/sha.h>
   8 #  if defined(OPENSSL_API_LEVEL) && OPENSSL_API_LEVEL >= 3
   9 #    define SHA1_NEEDS_CLONE_HELPER
  10 #    include "sha1/openssl.h"
  11 #  endif
  12 #elif defined(SHA1_DC)
  13 #include "sha1dc_git.h"
  14 #else /* SHA1_BLK */
  15 #include "block-sha1/sha1.h"
  16 #endif
  17
  18 #if defined(SHA1_APPLE_UNSAFE)
  19 #  include <CommonCrypto/CommonDigest.h>
  20 #  define platform_SHA_CTX_unsafe CC_SHA1_CTX
  21 #  define platform_SHA1_Init_unsafe CC_SHA1_Init
  22 #  define platform_SHA1_Update_unsafe CC_SHA1_Update
  23 #  define platform_SHA1_Final_unsafe CC_SHA1_Final
  24 #elif defined(SHA1_OPENSSL_UNSAFE)
  25 #  include <openssl/sha.h>
  26 #  if defined(OPENSSL_API_LEVEL) && OPENSSL_API_LEVEL >= 3
  27 #    define SHA1_NEEDS_CLONE_HELPER_UNSAFE
  28 #    include "sha1/openssl.h"
  29 #    define platform_SHA_CTX_unsafe openssl_SHA1_CTX
  30 #    define platform_SHA1_Init_unsafe openssl_SHA1_Init
  31 #    define platform_SHA1_Clone_unsafe openssl_SHA1_Clone
  32 #    define platform_SHA1_Update_unsafe openssl_SHA1_Update
  33 #    define platform_SHA1_Final_unsafe openssl_SHA1_Final
  34 #  else
  35 #    define platform_SHA_CTX_unsafe SHA_CTX
  36 #    define platform_SHA1_Init_unsafe SHA1_Init
  37 #    define platform_SHA1_Update_unsafe SHA1_Update
  38 #    define platform_SHA1_Final_unsafe SHA1_Final
  39 #  endif
  40 #elif defined(SHA1_BLK_UNSAFE)
  41 #  include "block-sha1/sha1.h"
  42 #  define platform_SHA_CTX_unsafe blk_SHA_CTX
  43 #  define platform_SHA1_Init_unsafe blk_SHA1_Init
  44 #  define platform_SHA1_Update_unsafe blk_SHA1_Update
  45 #  define platform_SHA1_Final_unsafe blk_SHA1_Final
  46 #endif
  47
  48 #if defined(SHA256_NETTLE)
  49 #include "sha256/nettle.h"
  50 #elif defined(SHA256_GCRYPT)
  51 #define SHA256_NEEDS_CLONE_HELPER
  52 #include "sha256/gcrypt.h"
  53 #elif defined(SHA256_OPENSSL)
  54 #  include <openssl/sha.h>
  55 #  if defined(OPENSSL_API_LEVEL) && OPENSSL_API_LEVEL >= 3
  56 #    define SHA256_NEEDS_CLONE_HELPER
  57 #    include "sha256/openssl.h"
  58 #  endif
  59 #else
  60 #include "sha256/block/sha256.h"
  61 #endif
  62
  63 #ifndef platform_SHA_CTX
  64 /*
  65  * platform's underlying implementation of SHA-1; could be OpenSSL,
  66  * blk_SHA, Apple CommonCrypto, etc...  Note that the relevant
  67  * SHA-1 header may have already defined platform_SHA_CTX for our
  68  * own implementations like block-sha1, so we list
  69  * the default for OpenSSL compatible SHA-1 implementations here.
  70  */
  71 #define platform_SHA_CTX        SHA_CTX
  72 #define platform_SHA1_Init      SHA1_Init
  73 #define platform_SHA1_Update    SHA1_Update
  74 #define platform_SHA1_Final     SHA1_Final
  75 #endif
  76
  77 #ifndef platform_SHA_CTX_unsafe
  78 #  define platform_SHA_CTX_unsafe      platform_SHA_CTX
  79 #  define platform_SHA1_Init_unsafe    platform_SHA1_Init
  80 #  define platform_SHA1_Update_unsafe  platform_SHA1_Update
  81 #  define platform_SHA1_Final_unsafe   platform_SHA1_Final
  82 #  ifdef platform_SHA1_Clone
  83 #    define platform_SHA1_Clone_unsafe platform_SHA1_Clone
  84 #  endif
  85 #  ifdef SHA1_NEEDS_CLONE_HELPER
  86 #    define SHA1_NEEDS_CLONE_HELPER_UNSAFE
  87 #  endif
  88 #endif
  89
  90 #define git_SHA_CTX             platform_SHA_CTX
  91 #define git_SHA1_Init           platform_SHA1_Init
  92 #define git_SHA1_Update         platform_SHA1_Update
  93 #define git_SHA1_Final          platform_SHA1_Final
  94
  95 #define git_SHA_CTX_unsafe      platform_SHA_CTX_unsafe
  96 #define git_SHA1_Init_unsafe    platform_SHA1_Init_unsafe
  97 #define git_SHA1_Update_unsafe  platform_SHA1_Update_unsafe
  98 #define git_SHA1_Final_unsafe   platform_SHA1_Final_unsafe
  99
 100 #ifdef platform_SHA1_Clone
 101 #define git_SHA1_Clone  platform_SHA1_Clone
 102 #endif
 103 #ifdef platform_SHA1_Clone_unsafe
 104 #  define git_SHA1_Clone_unsafe platform_SHA1_Clone_unsafe
 105 #endif
 106
 107 #ifndef platform_SHA256_CTX
 108 #define platform_SHA256_CTX     SHA256_CTX
 109 #define platform_SHA256_Init    SHA256_Init
 110 #define platform_SHA256_Update  SHA256_Update
 111 #define platform_SHA256_Final   SHA256_Final
 112 #endif
 113
 114 #define git_SHA256_CTX          platform_SHA256_CTX
 115 #define git_SHA256_Init         platform_SHA256_Init
 116 #define git_SHA256_Update       platform_SHA256_Update
 117 #define git_SHA256_Final        platform_SHA256_Final
 118
 119 #ifdef platform_SHA256_Clone
 120 #define git_SHA256_Clone        platform_SHA256_Clone
 121 #endif
 122
 123 #ifdef SHA1_MAX_BLOCK_SIZE
 124 #include "compat/sha1-chunked.h"
 125 #undef git_SHA1_Update
 126 #define git_SHA1_Update         git_SHA1_Update_Chunked
 127 #endif
 128
 129 #ifndef SHA1_NEEDS_CLONE_HELPER
 130 static inline void git_SHA1_Clone(git_SHA_CTX *dst, const git_SHA_CTX *src)
 131 {
 132         memcpy(dst, src, sizeof(*dst));
 133 }
 134 #endif
 135 #ifndef SHA1_NEEDS_CLONE_HELPER_UNSAFE
 136 static inline void git_SHA1_Clone_unsafe(git_SHA_CTX_unsafe *dst,
 137                                        const git_SHA_CTX_unsafe *src)
 138 {
 139         memcpy(dst, src, sizeof(*dst));
 140 }
 141 #endif
 142
 143 #ifndef SHA256_NEEDS_CLONE_HELPER
 144 static inline void git_SHA256_Clone(git_SHA256_CTX *dst, const git_SHA256_CTX *src)
 145 {
 146         memcpy(dst, src, sizeof(*dst));
 147 }
 148 #endif
 149
 150 /*
 151  * Note that these constants are suitable for indexing the hash_algos array and
 152  * comparing against each other, but are otherwise arbitrary, so they should not
 153  * be exposed to the user or serialized to disk.  To know whether a
 154  * git_hash_algo struct points to some usable hash function, test the format_id
 155  * field for being non-zero.  Use the name field for user-visible situations and
 156  * the format_id field for fixed-length fields on disk.
 157  */
 158 /* An unknown hash function. */
 159 #define GIT_HASH_UNKNOWN 0
 160 /* SHA-1 */
 161 #define GIT_HASH_SHA1 1
 162 /* SHA-256  */
 163 #define GIT_HASH_SHA256 2
 164 /* Number of algorithms supported (including unknown). */
 165 #define GIT_HASH_NALGOS (GIT_HASH_SHA256 + 1)
 166
 167 /* "sha1", big-endian */
 168 #define GIT_SHA1_FORMAT_ID 0x73686131
 169
 170 /* The length in bytes and in hex digits of an object name (SHA-1 value). */
 171 #define GIT_SHA1_RAWSZ 20
 172 #define GIT_SHA1_HEXSZ (2 * GIT_SHA1_RAWSZ)
 173 /* The block size of SHA-1. */
 174 #define GIT_SHA1_BLKSZ 64
 175
 176 /* "s256", big-endian */
 177 #define GIT_SHA256_FORMAT_ID 0x73323536
 178
 179 /* The length in bytes and in hex digits of an object name (SHA-256 value). */
 180 #define GIT_SHA256_RAWSZ 32
 181 #define GIT_SHA256_HEXSZ (2 * GIT_SHA256_RAWSZ)
 182 /* The block size of SHA-256. */
 183 #define GIT_SHA256_BLKSZ 64
 184
 185 /* The length in byte and in hex digits of the largest possible hash value. */
 186 #define GIT_MAX_RAWSZ GIT_SHA256_RAWSZ
 187 #define GIT_MAX_HEXSZ GIT_SHA256_HEXSZ
 188 /* The largest possible block size for any supported hash. */
 189 #define GIT_MAX_BLKSZ GIT_SHA256_BLKSZ
 190
 191 struct object_id {
 192         unsigned char hash[GIT_MAX_RAWSZ];
 193         int algo;       /* XXX requires 4-byte alignment */
 194 };
 195
 196 #define GET_OID_QUIETLY           01
 197 #define GET_OID_COMMIT            02
 198 #define GET_OID_COMMITTISH        04
 199 #define GET_OID_TREE             010
 200 #define GET_OID_TREEISH          020
 201 #define GET_OID_BLOB             040
 202 #define GET_OID_FOLLOW_SYMLINKS 0100
 203 #define GET_OID_RECORD_PATH     0200
 204 #define GET_OID_ONLY_TO_DIE    04000
 205 #define GET_OID_REQUIRE_PATH  010000
 206 #define GET_OID_HASH_ANY      020000
 207
 208 #define GET_OID_DISAMBIGUATORS \
 209         (GET_OID_COMMIT | GET_OID_COMMITTISH | \
 210         GET_OID_TREE | GET_OID_TREEISH | \
 211         GET_OID_BLOB)
 212
 213 enum get_oid_result {
 214         FOUND = 0,
 215         MISSING_OBJECT = -1, /* The requested object is missing */
 216         SHORT_NAME_AMBIGUOUS = -2,
 217         /* The following only apply when symlinks are followed */
 218         DANGLING_SYMLINK = -4, /*
 219                                 * The initial symlink is there, but
 220                                 * (transitively) points to a missing
 221                                 * in-tree file
 222                                 */
 223         SYMLINK_LOOP = -5,
 224         NOT_DIR = -6, /*
 225                        * Somewhere along the symlink chain, a path is
 226                        * requested which contains a file as a
 227                        * non-final element.
 228                        */
 229 };
 230
 231 #ifdef USE_THE_REPOSITORY_VARIABLE
 232 # include "repository.h"
 233 # define the_hash_algo the_repository->hash_algo
 234 #endif
 235
 236 /* A suitably aligned type for stack allocations of hash contexts. */
 237 union git_hash_ctx {
 238         git_SHA_CTX sha1;
 239         git_SHA_CTX_unsafe sha1_unsafe;
 240
 241         git_SHA256_CTX sha256;
 242 };
 243 typedef union git_hash_ctx git_hash_ctx;
 244
 245 typedef void (*git_hash_init_fn)(git_hash_ctx *ctx);
 246 typedef void (*git_hash_clone_fn)(git_hash_ctx *dst, const git_hash_ctx *src);
 247 typedef void (*git_hash_update_fn)(git_hash_ctx *ctx, const void *in, size_t len);
 248 typedef void (*git_hash_final_fn)(unsigned char *hash, git_hash_ctx *ctx);
 249 typedef void (*git_hash_final_oid_fn)(struct object_id *oid, git_hash_ctx *ctx);
 250
 251 struct git_hash_algo {
 252         /*
 253          * The name of the algorithm, as appears in the config file and in
 254          * messages.
 255          */
 256         const char *name;
 257
 258         /* A four-byte version identifier, used in pack indices. */
 259         uint32_t format_id;
 260
 261         /* The length of the hash in binary. */
 262         size_t rawsz;
 263
 264         /* The length of the hash in hex characters. */
 265         size_t hexsz;
 266
 267         /* The block size of the hash. */
 268         size_t blksz;
 269
 270         /* The hash initialization function. */
 271         git_hash_init_fn init_fn;
 272
 273         /* The hash context cloning function. */
 274         git_hash_clone_fn clone_fn;
 275
 276         /* The hash update function. */
 277         git_hash_update_fn update_fn;
 278
 279         /* The hash finalization function. */
 280         git_hash_final_fn final_fn;
 281
 282         /* The hash finalization function for object IDs. */
 283         git_hash_final_oid_fn final_oid_fn;
 284
 285         /* The non-cryptographic hash initialization function. */
 286         git_hash_init_fn unsafe_init_fn;
 287
 288         /* The non-cryptographic hash context cloning function. */
 289         git_hash_clone_fn unsafe_clone_fn;
 290
 291         /* The non-cryptographic hash update function. */
 292         git_hash_update_fn unsafe_update_fn;
 293
 294         /* The non-cryptographic hash finalization function. */
 295         git_hash_final_fn unsafe_final_fn;
 296
 297         /* The non-cryptographic hash finalization function. */
 298         git_hash_final_oid_fn unsafe_final_oid_fn;
 299
 300         /* The OID of the empty tree. */
 301         const struct object_id *empty_tree;
 302
 303         /* The OID of the empty blob. */
 304         const struct object_id *empty_blob;
 305
 306         /* The all-zeros OID. */
 307         const struct object_id *null_oid;
 308 };
 309 extern const struct git_hash_algo hash_algos[GIT_HASH_NALGOS];
 310
 311 /*
 312  * Return a GIT_HASH_* constant based on the name.  Returns GIT_HASH_UNKNOWN if
 313  * the name doesn't match a known algorithm.
 314  */
 315 int hash_algo_by_name(const char *name);
 316 /* Identical, except based on the format ID. */
 317 int hash_algo_by_id(uint32_t format_id);
 318 /* Identical, except based on the length. */
 319 int hash_algo_by_length(int len);
 320 /* Identical, except for a pointer to struct git_hash_algo. */
 321 static inline int hash_algo_by_ptr(const struct git_hash_algo *p)
 322 {
 323         return p - hash_algos;
 324 }
 325
 326 const struct object_id *null_oid(void);
 327
 328 static inline int hashcmp(const unsigned char *sha1, const unsigned char *sha2, const struct git_hash_algo *algop)
 329 {
 330         /*
 331          * Teach the compiler that there are only two possibilities of hash size
 332          * here, so that it can optimize for this case as much as possible.
 333          */
 334         if (algop->rawsz == GIT_MAX_RAWSZ)
 335                 return memcmp(sha1, sha2, GIT_MAX_RAWSZ);
 336         return memcmp(sha1, sha2, GIT_SHA1_RAWSZ);
 337 }
 338
 339 static inline int hasheq(const unsigned char *sha1, const unsigned char *sha2, const struct git_hash_algo *algop)
 340 {
 341         /*
 342          * We write this here instead of deferring to hashcmp so that the
 343          * compiler can properly inline it and avoid calling memcmp.
 344          */
 345         if (algop->rawsz == GIT_MAX_RAWSZ)
 346                 return !memcmp(sha1, sha2, GIT_MAX_RAWSZ);
 347         return !memcmp(sha1, sha2, GIT_SHA1_RAWSZ);
 348 }
 349
 350 static inline void hashcpy(unsigned char *sha_dst, const unsigned char *sha_src,
 351                            const struct git_hash_algo *algop)
 352 {
 353         memcpy(sha_dst, sha_src, algop->rawsz);
 354 }
 355
 356 static inline void hashclr(unsigned char *hash, const struct git_hash_algo *algop)
 357 {
 358         memset(hash, 0, algop->rawsz);
 359 }
 360
 361 static inline int oidcmp(const struct object_id *oid1, const struct object_id *oid2)
 362 {
 363         return memcmp(oid1->hash, oid2->hash, GIT_MAX_RAWSZ);
 364 }
 365
 366 static inline int oideq(const struct object_id *oid1, const struct object_id *oid2)
 367 {
 368         return !memcmp(oid1->hash, oid2->hash, GIT_MAX_RAWSZ);
 369 }
 370
 371 static inline void oidcpy(struct object_id *dst, const struct object_id *src)
 372 {
 373         memcpy(dst->hash, src->hash, GIT_MAX_RAWSZ);
 374         dst->algo = src->algo;
 375 }
 376
 377 static inline void oidread(struct object_id *oid, const unsigned char *hash,
 378                            const struct git_hash_algo *algop)
 379 {
 380         memcpy(oid->hash, hash, algop->rawsz);
 381         if (algop->rawsz < GIT_MAX_RAWSZ)
 382                 memset(oid->hash + algop->rawsz, 0, GIT_MAX_RAWSZ - algop->rawsz);
 383         oid->algo = hash_algo_by_ptr(algop);
 384 }
 385
 386 static inline void oidclr(struct object_id *oid,
 387                           const struct git_hash_algo *algop)
 388 {
 389         memset(oid->hash, 0, GIT_MAX_RAWSZ);
 390         oid->algo = hash_algo_by_ptr(algop);
 391 }
 392
 393 static inline struct object_id *oiddup(const struct object_id *src)
 394 {
 395         struct object_id *dst = xmalloc(sizeof(struct object_id));
 396         oidcpy(dst, src);
 397         return dst;
 398 }
 399
 400 static inline void oid_set_algo(struct object_id *oid, const struct git_hash_algo *algop)
 401 {
 402         oid->algo = hash_algo_by_ptr(algop);
 403 }
 404
 405 /*
 406  * Converts a cryptographic hash (e.g. SHA-1) into an int-sized hash code
 407  * for use in hash tables. Cryptographic hashes are supposed to have
 408  * uniform distribution, so in contrast to `memhash()`, this just copies
 409  * the first `sizeof(int)` bytes without shuffling any bits. Note that
 410  * the results will be different on big-endian and little-endian
 411  * platforms, so they should not be stored or transferred over the net.
 412  */
 413 static inline unsigned int oidhash(const struct object_id *oid)
 414 {
 415         /*
 416          * Equivalent to 'return *(unsigned int *)oid->hash;', but safe on
 417          * platforms that don't support unaligned reads.
 418          */
 419         unsigned int hash;
 420         memcpy(&hash, oid->hash, sizeof(hash));
 421         return hash;
 422 }
 423
 424 static inline int is_null_oid(const struct object_id *oid)
 425 {
 426         static const unsigned char null_hash[GIT_MAX_RAWSZ];
 427         return !memcmp(oid->hash, null_hash, GIT_MAX_RAWSZ);
 428 }
 429
 430 const char *empty_tree_oid_hex(const struct git_hash_algo *algop);
 431
 432 static inline int is_empty_blob_oid(const struct object_id *oid,
 433                                     const struct git_hash_algo *algop)
 434 {
 435         return oideq(oid, algop->empty_blob);
 436 }
 437
 438 static inline int is_empty_tree_oid(const struct object_id *oid,
 439                                     const struct git_hash_algo *algop)
 440 {
 441         return oideq(oid, algop->empty_tree);
 442 }
 443
 444 #endif