midx: bounds-check large offset chunk

When we see a large offset bit in the regular midx offset table, we use the entry as an index into a separate large offset table (just like a pack idx does). But we don't bounds-check the access to that large offset table (nor even record its size when we parse the chunk!). The equivalent code for a regular pack idx is in check_pack_index_ptr(). But things are a bit simpler here because of the chunked format: we can just check our array index directly. As a bonus, we can get rid of the st_mult() here. If our array bounds-check is successful, then we know that the result will fit in a size_t (and the bounds check uses a division to avoid overflow entirely). Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Jeff King committed Oct 9, 2023 at 17:05 UTC 2abd56e9b2195c8111ff5d16efafabc5bccba92b
3 files changed +26 -3
midx.c
+5 -3
@@ -180,7 +180,8 @@ struct multi_pack_index *load_multi_pack_index(const char *object_dir, int local
180 if (read_chunk(cf, MIDX_CHUNKID_OBJECTOFFSETS, midx_read_object_offsets, m))
181 die(_("multi-pack-index required object offsets chunk missing or corrupted"));
182
183 - pair_chunk_unsafe(cf, MIDX_CHUNKID_LARGEOFFSETS, &m->chunk_large_offsets);
183 + pair_chunk(cf, MIDX_CHUNKID_LARGEOFFSETS, &m->chunk_large_offsets,
184 + &m->chunk_large_offsets_len);
185
186 if (git_env_bool("GIT_TEST_MIDX_READ_RIDX", 1))
187 pair_chunk_unsafe(cf, MIDX_CHUNKID_REVINDEX, &m->chunk_revindex);
@@ -303,8 +304,9 @@ off_t nth_midxed_offset(struct multi_pack_index *m, uint32_t pos)
304 die(_("multi-pack-index stores a 64-bit offset, but off_t is too small"));
305
306 offset32 ^= MIDX_LARGE_OFFSET_NEEDED;
306 - return get_be64(m->chunk_large_offsets +
307 - st_mult(sizeof(uint64_t), offset32));
307 + if (offset32 >= m->chunk_large_offsets_len / sizeof(uint64_t))
308 + die(_("multi-pack-index large offset out of bounds"));
309 + return get_be64(m->chunk_large_offsets + sizeof(uint64_t) * offset32);
310 }
311
312 return offset32;
midx.h
+1
@@ -37,6 +37,7 @@ struct multi_pack_index {
37 const unsigned char *chunk_oid_lookup;
38 const unsigned char *chunk_object_offsets;
39 const unsigned char *chunk_large_offsets;
40 + size_t chunk_large_offsets_len;
41 const unsigned char *chunk_revindex;
42
43 const char **pack_names;
t/t5319-multi-pack-index.sh
+20
@@ -1118,4 +1118,24 @@ test_expect_success 'reader notices too-small object offset chunk' '
1118 test_cmp expect err
1119 '
1120
1121 +test_expect_success 'reader bounds-checks large offset table' '
1122 + # re-use the objects64 dir here to cheaply get access to a midx
1123 + # with large offsets.
1124 + git init repo &&
1125 + test_when_finished "rm -rf repo" &&
1126 + (
1127 + cd repo &&
1128 + (cd ../objects64 && pwd) >.git/objects/info/alternates &&
1129 + git multi-pack-index --object-dir=../objects64 write &&
1130 + midx=../objects64/pack/multi-pack-index &&
1131 + corrupt_chunk_file $midx LOFF clear &&
1132 + test_must_fail git cat-file \
1133 + --batch-check --batch-all-objects 2>err &&
1134 + cat >expect <<-\EOF &&
1135 + fatal: multi-pack-index large offset out of bounds
1136 + EOF
1137 + test_cmp expect err
1138 + )
1139 +'
1140 +
1141 test_done