pack-check.c

   1 #include "cache.h"
   2 #include "repository.h"
   3 #include "pack.h"
   4 #include "pack-revindex.h"
   5 #include "progress.h"
   6 #include "packfile.h"
   7 #include "object-store.h"
   8
   9 struct idx_entry {
  10         off_t                offset;
  11         union idx_entry_object {
  12                 const unsigned char *hash;
  13                 struct object_id *oid;
  14         } oid;
  15         unsigned int nr;
  16 };
  17
  18 static int compare_entries(const void *e1, const void *e2)
  19 {
  20         const struct idx_entry *entry1 = e1;
  21         const struct idx_entry *entry2 = e2;
  22         if (entry1->offset < entry2->offset)
  23                 return -1;
  24         if (entry1->offset > entry2->offset)
  25                 return 1;
  26         return 0;
  27 }
  28
  29 int check_pack_crc(struct packed_git *p, struct pack_window **w_curs,
  30                    off_t offset, off_t len, unsigned int nr)
  31 {
  32         const uint32_t *index_crc;
  33         uint32_t data_crc = crc32(0, NULL, 0);
  34
  35         do {
  36                 unsigned long avail;
  37                 void *data = use_pack(p, w_curs, offset, &avail);
  38                 if (avail > len)
  39                         avail = len;
  40                 data_crc = crc32(data_crc, data, avail);
  41                 offset += avail;
  42                 len -= avail;
  43         } while (len);
  44
  45         index_crc = p->index_data;
  46         index_crc += 2 + 256 + p->num_objects * (the_hash_algo->rawsz/4) + nr;
  47
  48         return data_crc != ntohl(*index_crc);
  49 }
  50
  51 static int verify_packfile(struct repository *r,
  52                            struct packed_git *p,
  53                            struct pack_window **w_curs,
  54                            verify_fn fn,
  55                            struct progress *progress, uint32_t base_count)
  56
  57 {
  58         off_t index_size = p->index_size;
  59         const unsigned char *index_base = p->index_data;
  60         git_hash_ctx ctx;
  61         unsigned char hash[GIT_MAX_RAWSZ], *pack_sig;
  62         off_t offset = 0, pack_sig_ofs = 0;
  63         uint32_t nr_objects, i;
  64         int err = 0;
  65         struct idx_entry *entries;
  66
  67         if (!is_pack_valid(p))
  68                 return error("packfile %s cannot be accessed", p->pack_name);
  69
  70         r->hash_algo->init_fn(&ctx);
  71         do {
  72                 unsigned long remaining;
  73                 unsigned char *in = use_pack(p, w_curs, offset, &remaining);
  74                 offset += remaining;
  75                 if (!pack_sig_ofs)
  76                         pack_sig_ofs = p->pack_size - r->hash_algo->rawsz;
  77                 if (offset > pack_sig_ofs)
  78                         remaining -= (unsigned int)(offset - pack_sig_ofs);
  79                 r->hash_algo->update_fn(&ctx, in, remaining);
  80         } while (offset < pack_sig_ofs);
  81         r->hash_algo->final_fn(hash, &ctx);
  82         pack_sig = use_pack(p, w_curs, pack_sig_ofs, NULL);
  83         if (!hasheq(hash, pack_sig))
  84                 err = error("%s pack checksum mismatch",
  85                             p->pack_name);
  86         if (!hasheq(index_base + index_size - r->hash_algo->hexsz, pack_sig))
  87                 err = error("%s pack checksum does not match its index",
  88                             p->pack_name);
  89         unuse_pack(w_curs);
  90
  91         /* Make sure everything reachable from idx is valid.  Since we
  92          * have verified that nr_objects matches between idx and pack,
  93          * we do not do scan-streaming check on the pack file.
  94          */
  95         nr_objects = p->num_objects;
  96         ALLOC_ARRAY(entries, nr_objects + 1);
  97         entries[nr_objects].offset = pack_sig_ofs;
  98         /* first sort entries by pack offset, since unpacking them is more efficient that way */
  99         for (i = 0; i < nr_objects; i++) {
 100                 entries[i].oid.hash = nth_packed_object_sha1(p, i);
 101                 if (!entries[i].oid.hash)
 102                         die("internal error pack-check nth-packed-object");
 103                 entries[i].offset = nth_packed_object_offset(p, i);
 104                 entries[i].nr = i;
 105         }
 106         QSORT(entries, nr_objects, compare_entries);
 107
 108         for (i = 0; i < nr_objects; i++) {
 109                 void *data;
 110                 enum object_type type;
 111                 unsigned long size;
 112                 off_t curpos;
 113                 int data_valid;
 114
 115                 if (p->index_version > 1) {
 116                         off_t offset = entries[i].offset;
 117                         off_t len = entries[i+1].offset - offset;
 118                         unsigned int nr = entries[i].nr;
 119                         if (check_pack_crc(p, w_curs, offset, len, nr))
 120                                 err = error("index CRC mismatch for object %s "
 121                                             "from %s at offset %"PRIuMAX"",
 122                                             oid_to_hex(entries[i].oid.oid),
 123                                             p->pack_name, (uintmax_t)offset);
 124                 }
 125
 126                 curpos = entries[i].offset;
 127                 type = unpack_object_header(p, w_curs, &curpos, &size);
 128                 unuse_pack(w_curs);
 129
 130                 if (type == OBJ_BLOB && big_file_threshold <= size) {
 131                         /*
 132                          * Let check_object_signature() check it with
 133                          * the streaming interface; no point slurping
 134                          * the data in-core only to discard.
 135                          */
 136                         data = NULL;
 137                         data_valid = 0;
 138                 } else {
 139                         data = unpack_entry(r, p, entries[i].offset, &type, &size);
 140                         data_valid = 1;
 141                 }
 142
 143                 if (data_valid && !data)
 144                         err = error("cannot unpack %s from %s at offset %"PRIuMAX"",
 145                                     oid_to_hex(entries[i].oid.oid), p->pack_name,
 146                                     (uintmax_t)entries[i].offset);
 147                 else if (check_object_signature(r, entries[i].oid.oid, data, size, type_name(type)))
 148                         err = error("packed %s from %s is corrupt",
 149                                     oid_to_hex(entries[i].oid.oid), p->pack_name);
 150                 else if (fn) {
 151                         int eaten = 0;
 152                         err |= fn(entries[i].oid.oid, type, size, data, &eaten);
 153                         if (eaten)
 154                                 data = NULL;
 155                 }
 156                 if (((base_count + i) & 1023) == 0)
 157                         display_progress(progress, base_count + i);
 158                 free(data);
 159
 160         }
 161         display_progress(progress, base_count + i);
 162         free(entries);
 163
 164         return err;
 165 }
 166
 167 int verify_pack_index(struct packed_git *p)
 168 {
 169         off_t index_size;
 170         const unsigned char *index_base;
 171         git_hash_ctx ctx;
 172         unsigned char hash[GIT_MAX_RAWSZ];
 173         int err = 0;
 174
 175         if (open_pack_index(p))
 176                 return error("packfile %s index not opened", p->pack_name);
 177         index_size = p->index_size;
 178         index_base = p->index_data;
 179
 180         /* Verify SHA1 sum of the index file */
 181         the_hash_algo->init_fn(&ctx);
 182         the_hash_algo->update_fn(&ctx, index_base, (unsigned int)(index_size - the_hash_algo->rawsz));
 183         the_hash_algo->final_fn(hash, &ctx);
 184         if (!hasheq(hash, index_base + index_size - the_hash_algo->rawsz))
 185                 err = error("Packfile index for %s hash mismatch",
 186                             p->pack_name);
 187         return err;
 188 }
 189
 190 int verify_pack(struct repository *r, struct packed_git *p, verify_fn fn,
 191                 struct progress *progress, uint32_t base_count)
 192 {
 193         int err = 0;
 194         struct pack_window *w_curs = NULL;
 195
 196         err |= verify_pack_index(p);
 197         if (!p->index_data)
 198                 return -1;
 199
 200         err |= verify_packfile(r, p, &w_curs, fn, progress, base_count);
 201         unuse_pack(&w_curs);
 202
 203         return err;
 204 }