On Mon, Jun 10, 2024 at 07:55:46AM +0200, Patrick Steinhardt wrote:

On Sun, Jun 09, 2024 at 11:27:35AM -0400, Taylor Blau wrote:

quoted

In 795006fff4 (pack-bitmap: gracefully handle missing BTMP chunks,
2024-04-15), we refactored the reuse_partial_packfile_from_bitmap()
function and stopped assigning the pack_int_id field when reusing only
the MIDX's preferred pack. This results in an uninitialized read down in
try_partial_reuse() like so:

I feel like I'm blind, but I cannot see how the patch changed what we do
with `pack_int_id`. It's not mentioned a single time in the diff, so how
did it have the effect of not setting it anymore?

It's because prior to 795006fff4, we handled reusing a single pack from
a MIDX differently than in the post-image of that commit. Prior to
795006fff4, the loop looked like:

    if (bitmap_is_midx(bitmap_git)) {
            for (i = 0; i < bitmap_git->midx->num_packs; i++) {
                    struct bitmapped_pack pack;
                    if (nth_bitmapped_pack(r, bitmap_git->midx, &pack, i) < 0) {
                          /* ... */
                          return;
                    }
                    if (!pack.bitmap_nr)
                          continue;
                    if (!multi_pack_reuse && pack.bitmap_pos)
                          continue;

                    ALLOC_GROW(packs, packs_nr + 1, packs_alloc);
                    memcpy(&packs[packs_nr++], &pack, sizeof(pack));
            }
    }

Since nth_bitmapped_pack() fills out the pack_int_id field, we got it
automatically since we just memcpy()'d the result of
nth_bitmapped_pack() into our array.

In the single pack bitmap case, we don't need to initialize the
pack_int_id field because we never read it, hence the lack of MSan
failures in that mode.

But since 795006fff4 combined these two single pack cases (that is,
single-pack bitmaps, and reusing only a single pack out of a MIDX
bitmap) into one, 795006fff4 neglected to initialize the pack_int_id
field, causing this issue.

Thanks,
Taylor