snapshots/erofs: keep lowers stacked above a merged fsmeta

mounts() moved `first` to the merged fsmeta mount's index whenever
mountFsMeta matched, collapsing the overlay's lowerdir range to the
fsmeta alone and dropping any plain lowers already appended above it.

When fsmerge was added, fsmeta was only ever written for the top
parent of the chain being prepared, so the loop matched at i == 0,
where first already equalled the fsmeta's index and the reassignment
was a no-op. It only mattered once the fsmeta could sit below the top
parent, i.e. a chain extending an already-merged sub-chain. The
snapshotter no longer generates fsmeta itself, so today this only
affects externally supplied fsmeta files.

first only marks the start of the lowerdir range, so it must stay at
the first lower's index.

Signed-off-by: Derek McGowan <derek@mcg.dev>

plugins/snapshots/erofs/erofs.go

@@ -521,13 +521,14 @@ func (s *snapshotter) mounts(snap storage.Snapshot, info snapshots.Info) ([]moun
 		return []mount.Mount{m}, nil
 	}
 
+	// first marks the start of the lowerdir range. A merged fsmeta ends the
+	// range but never moves its start: lowers stacked above it stay in range.
 	first := len(mounts)
 	for i := range snap.ParentIDs {
 		// If a merged fsmeta is valid for this layer, skip the remaining bottom layers.
 		// Why? Because bottom layers have been flattened with the thin fsmeta.
 		if m, ok := s.mountFsMeta(snap, i); ok {
 			mounts = append(mounts, m)
-			first = len(mounts) - 1
 			break
 		}
 

plugins/snapshots/erofs/erofs_linux_test.go

@@ -820,6 +820,88 @@ func TestMountFsMeta(t *testing.T) {
 	})
 }
 
+// TestMountsWithMergedFsMeta covers s.mounts()'s assembly of the overlay
+// lowerdir when a merged fsmeta is present on a parent below the top of the
+// chain, i.e. one or more plain (non-merged) layers are stacked on top of a
+// merged fsmeta lower.
+func TestMountsWithMergedFsMeta(t *testing.T) {
+	root := t.TempDir()
+	s := &snapshotter{root: root}
+
+	// Chain (top to bottom): p0, p1, p2, p3. A merged fsmeta is only
+	// present for p2, flattening the sub-chain [p2, p3]. p0 and p1 are
+	// plain layers stacked on top of that merged base.
+	parents := []string{"p0", "p1", "p2", "p3"}
+	for _, id := range parents {
+		require.NoError(t, os.MkdirAll(filepath.Join(root, "snapshots", id), 0755))
+		require.NoError(t, os.WriteFile(s.layerBlobPath(id), []byte("layer"), 0644))
+	}
+	require.NoError(t, os.WriteFile(s.fsMetaPath("p2"), []byte("merged"), 0644))
+
+	snap := storage.Snapshot{Kind: snapshots.KindView, ParentIDs: parents}
+	info := snapshots.Info{}
+
+	mounts, err := s.mounts(snap, info)
+	require.NoError(t, err)
+
+	// Expect: [erofs(p0), erofs(p1), erofs(fsmeta p2, device=p3,p2), overlay]
+	require.Len(t, mounts, 4)
+
+	assert.Equal(t, "erofs", mounts[0].Type)
+	assert.Equal(t, s.layerBlobPath("p0"), mounts[0].Source)
+
+	assert.Equal(t, "erofs", mounts[1].Type)
+	assert.Equal(t, s.layerBlobPath("p1"), mounts[1].Source)
+
+	assert.Equal(t, "erofs", mounts[2].Type)
+	assert.Equal(t, s.fsMetaPath("p2"), mounts[2].Source)
+	assert.Equal(t, []string{
+		"ro", "loop",
+		"device=" + s.layerBlobPath("p3"),
+		"device=" + s.layerBlobPath("p2"),
+	}, mounts[2].Options)
+
+	// The overlay must span all three lowers (indices 0-2): the two plain
+	// top layers plus the merged fsmeta.
+	overlay := mounts[3]
+	assert.Equal(t, "format/mkdir/overlay", overlay.Type)
+	assert.Contains(t, overlay.Options, "lowerdir={{ overlay 0 2 }}")
+}
+
+// TestMountsWithMergedFsMetaOnTopParent covers the case where the merged
+// fsmeta is valid for the topmost parent, so it is the only lower. Since
+// overlayfs rejects a lowerdir with no upperdir, this collapses to a plain
+// bind mount instead.
+func TestMountsWithMergedFsMetaOnTopParent(t *testing.T) {
+	root := t.TempDir()
+	s := &snapshotter{root: root}
+
+	parents := []string{"p0", "p1"}
+	for _, id := range parents {
+		require.NoError(t, os.MkdirAll(filepath.Join(root, "snapshots", id), 0755))
+		require.NoError(t, os.WriteFile(s.layerBlobPath(id), []byte("layer"), 0644))
+	}
+	require.NoError(t, os.WriteFile(s.fsMetaPath("p0"), []byte("merged"), 0644))
+
+	snap := storage.Snapshot{Kind: snapshots.KindView, ParentIDs: parents}
+	info := snapshots.Info{}
+
+	mounts, err := s.mounts(snap, info)
+	require.NoError(t, err)
+
+	require.Len(t, mounts, 2)
+	assert.Equal(t, "erofs", mounts[0].Type)
+	assert.Equal(t, s.fsMetaPath("p0"), mounts[0].Source)
+	assert.Equal(t, []string{
+		"ro", "loop",
+		"device=" + s.layerBlobPath("p1"),
+		"device=" + s.layerBlobPath("p0"),
+	}, mounts[0].Options)
+
+	assert.Equal(t, "format/bind", mounts[1].Type)
+	assert.Equal(t, "{{ mount 0 }}", mounts[1].Source)
+}
+
 // --- layer content cache tests ---
 
 const (