[PATCH bpf 0/2] Check timer_off for map_in_map only when map value has timer

[Hengqi Chen <hengqi.chen@xxxxxxxxx>]

The timer_off value could be -EINVAL or -ENOENT when map value of
inner map is struct and contains no bpf_timer.

The EINVAL case happens when the map is created without BTF key/value
info, map->timer_off is set to -EINVAL in map_create(). For example:

    map_fd = bpf_map_create(BPF_MAP_TYPE_LRU_HASH, NULL, 4, 4, 1, NULL);

The ENOENT case happens when the map is created with BTF key/value
info (e.g. from BPF skeleton), map->timer_off is set to -ENOENT as
what btf_find_timer() returns. For example, map created from BPF skeleton:

    struct inner_key {
    	__u32 x;
    };

    struct inner_value {
    	__u32 y;
    };

    struct inner {
    	__uint(type, BPF_MAP_TYPE_LRU_HASH);
    	__uint(max_entries, 1);
    	__type(key, struct inner_key);
    	__type(value, struct inner_value);
    } inner SEC(".maps");

    struct {
    	__uint(type, BPF_MAP_TYPE_ARRAY_OF_MAPS);
    	__uint(max_entries, 1);
    	__type(key, __u32);
    	__array(values, struct inner);
    } outer SEC(".maps");

Since timer_off is different, the map_in_map outer in the above case
can NOT be updated using map_fd in the first case. This patch tries
to fix such restriction.

Hengqi Chen (2):
  bpf: Check timer_off for map_in_map only when map value have timer
  selftests/bpf: Update map_in_map using map without BTF key/value info

 kernel/bpf/map_in_map.c                       |  9 ++++++-
 .../selftests/bpf/prog_tests/btf_map_in_map.c | 27 +++++++++++++++++++
 .../selftests/bpf/progs/test_btf_map_in_map.c | 22 +++++++++++++++
 3 files changed, 57 insertions(+), 1 deletion(-)

--
2.34.1