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 "progress.h"
   7 #include "packfile.h"
   8 #include "object-file.h"
   9 #include "object-store-ll.h"
  10
  11 struct idx_entry {
  12         off_t                offset;
  13         unsigned int nr;
  14 };
  15
  16 static int compare_entries(const void *e1, const void *e2)
  17 {
  18         const struct idx_entry *entry1 = e1;
  19         const struct idx_entry *entry2 = e2;
  20         if (entry1->offset < entry2->offset)
  21                 return -1;
  22         if (entry1->offset > entry2->offset)
  23                 return 1;
  24         return 0;
  25 }
  26
  27 int check_pack_crc(struct packed_git *p, struct pack_window **w_curs,
  28                    off_t offset, off_t len, unsigned int nr)
  29 {
  30         const uint32_t *index_crc;
  31         uint32_t data_crc = crc32(0, NULL, 0);
  32
  33         do {
  34                 unsigned long avail;
  35                 void *data = use_pack(p, w_curs, offset, &avail);
  36                 if (avail > len)
  37                         avail = len;
  38                 data_crc = crc32(data_crc, data, avail);
  39                 offset += avail;
  40                 len -= avail;
  41         } while (len);
  42
  43         index_crc = p->index_data;
  44         index_crc += 2 + 256 + (size_t)p->num_objects * (the_hash_algo->rawsz/4) + nr;
  45
  46         return data_crc != ntohl(*index_crc);
  47 }
  48
  49 static int verify_packfile(struct repository *r,
  50                            struct packed_git *p,
  51                            struct pack_window **w_curs,
  52                            verify_fn fn,
  53                            struct progress *progress, uint32_t base_count)
  54
  55 {
  56         off_t index_size = p->index_size;
  57         const unsigned char *index_base = p->index_data;
  58         git_hash_ctx ctx;
  59         unsigned char hash[GIT_MAX_RAWSZ], *pack_sig;
  60         off_t offset = 0, pack_sig_ofs = 0;
  61         uint32_t nr_objects, i;
  62         int err = 0;
  63         struct idx_entry *entries;
  64
  65         if (!is_pack_valid(p))
  66                 return error("packfile %s cannot be accessed", p->pack_name);
  67
  68         r->hash_algo->init_fn(&ctx);
  69         do {
  70                 unsigned long remaining;
  71                 unsigned char *in = use_pack(p, w_curs, offset, &remaining);
  72                 offset += remaining;
  73                 if (!pack_sig_ofs)
  74                         pack_sig_ofs = p->pack_size - r->hash_algo->rawsz;
  75                 if (offset > pack_sig_ofs)
  76                         remaining -= (unsigned int)(offset - pack_sig_ofs);
  77                 r->hash_algo->update_fn(&ctx, in, remaining);
  78         } while (offset < pack_sig_ofs);
  79         r->hash_algo->final_fn(hash, &ctx);
  80         pack_sig = use_pack(p, w_curs, pack_sig_ofs, NULL);
  81         if (!hasheq(hash, pack_sig))
  82                 err = error("%s pack checksum mismatch",
  83                             p->pack_name);
  84         if (!hasheq(index_base + index_size - r->hash_algo->hexsz, pack_sig))
  85                 err = error("%s pack checksum does not match its index",
  86                             p->pack_name);
  87         unuse_pack(w_curs);
  88
  89         /* Make sure everything reachable from idx is valid.  Since we
  90          * have verified that nr_objects matches between idx and pack,
  91          * we do not do scan-streaming check on the pack file.
  92          */
  93         nr_objects = p->num_objects;
  94         ALLOC_ARRAY(entries, nr_objects + 1);
  95         entries[nr_objects].offset = pack_sig_ofs;
  96         /* first sort entries by pack offset, since unpacking them is more efficient that way */
  97         for (i = 0; i < nr_objects; i++) {
  98                 entries[i].offset = nth_packed_object_offset(p, i);
  99                 entries[i].nr = i;
 100         }
 101         QSORT(entries, nr_objects, compare_entries);
 102
 103         for (i = 0; i < nr_objects; i++) {
 104                 void *data;
 105                 struct object_id oid;
 106                 enum object_type type;
 107                 unsigned long size;
 108                 off_t curpos;
 109                 int data_valid;
 110
 111                 if (nth_packed_object_id(&oid, p, entries[i].nr) < 0)
 112                         BUG("unable to get oid of object %lu from %s",
 113                             (unsigned long)entries[i].nr, p->pack_name);
 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(&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 stream_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(&oid), p->pack_name,
 146                                     (uintmax_t)entries[i].offset);
 147                 else if (data && check_object_signature(r, &oid, data, size,
 148                                                         type) < 0)
 149                         err = error("packed %s from %s is corrupt",
 150                                     oid_to_hex(&oid), p->pack_name);
 151                 else if (!data && stream_object_signature(r, &oid) < 0)
 152                         err = error("packed %s from %s is corrupt",
 153                                     oid_to_hex(&oid), p->pack_name);
 154                 else if (fn) {
 155                         int eaten = 0;
 156                         err |= fn(&oid, type, size, data, &eaten);
 157                         if (eaten)
 158                                 data = NULL;
 159                 }
 160                 if (((base_count + i) & 1023) == 0)
 161                         display_progress(progress, base_count + i);
 162                 free(data);
 163
 164         }
 165         display_progress(progress, base_count + i);
 166         free(entries);
 167
 168         return err;
 169 }
 170
 171 int verify_pack_index(struct packed_git *p)
 172 {
 173         int err = 0;
 174
 175         if (open_pack_index(p))
 176                 return error("packfile %s index not opened", p->pack_name);
 177
 178         /* Verify SHA1 sum of the index file */
 179         if (!hashfile_checksum_valid(p->index_data, p->index_size))
 180                 err = error("Packfile index for %s hash mismatch",
 181                             p->pack_name);
 182         return err;
 183 }
 184
 185 int verify_pack(struct repository *r, struct packed_git *p, verify_fn fn,
 186                 struct progress *progress, uint32_t base_count)
 187 {
 188         int err = 0;
 189         struct pack_window *w_curs = NULL;
 190
 191         err |= verify_pack_index(p);
 192         if (!p->index_data)
 193                 return -1;
 194
 195         err |= verify_packfile(r, p, &w_curs, fn, progress, base_count);
 196         unuse_pack(&w_curs);
 197
 198         return err;
 199 }