[PATCH bpf-next v2 2/2] bpf: Reject maps without iter_seq_info in bpf_iter_attach_map
From: Shihuang Liu
Date: Wed Nov 18 2026 - 11:17:03 EST
bpf_iter_attach_map() checks map_type only, so an offloaded ARRAY/HASH
map can be attached to a bpf_map_elem iterator. Offloaded maps have no
iter_seq_info, so __get_seq_info() falls back to the target's seq_info,
which is NULL for bpf_map_elem, and prepare_seq_file() dereferences it
when the iterator is opened.
Reject maps without ops->iter_seq_info. All map types accepted here
provide it.
Fixes: d6c4503cc296 ("bpf: Implement bpf iterator for hash maps")
Link: https://lore.kernel.org/all/20261112105620.57520-1-shlomojune6@xxxxxxxxx/
Signed-off-by: Shihuang Liu <shlomojune6@xxxxxxxxx>
---
kernel/bpf/map_iter.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/kernel/bpf/map_iter.c b/kernel/bpf/map_iter.c
index a58c148692e6c3..9e459a2d34c3aa 100644
--- a/kernel/bpf/map_iter.c
+++ b/kernel/bpf/map_iter.c
@@ -176,6 +176,10 @@ static int bpf_iter_attach_map(struct bpf_prog *prog,
map->map_type != BPF_MAP_TYPE_RHASH)
goto put_map;
+ /* An offloaded map has the map_type of a host map and its own ops */
+ if (!map->ops->iter_seq_info)
+ goto put_map;
+
aux->map = map;
err = bpf_iter_map_check_prog(prog, aux);
if (err)
--
2.43.0