pack-check.c

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