On Mon, 14 Sep 2020 11:46:30 -0700 Yonghong Song wrote:
> Currently, we use bucket_lock when traversing bpf_sk_storage_map
> elements. Since bpf_iter programs cannot use bpf_sk_storage_get()
> and bpf_sk_storage_delete() helpers which may also grab bucket lock,
> we do not have a deadlock issue which exists for hashmap when
> using bucket_lock ([1]).
>
> If a bucket contains a lot of sockets, during bpf_iter traversing
> a bucket, concurrent bpf_sk_storage_{get,delete}() may experience
> some undesirable delays. Using rcu_read_lock() is a reasonable
> compromise here. Although it may lose some precision, e.g.,
> access stale sockets, but it will not hurt performance of other
> bpf programs.
>
> [1] https://lore.kernel.org/bpf/20200902235341.2001534-1-...@fb.com
>
> Cc: Martin KaFai Lau <ka...@fb.com>
> Signed-off-by: Yonghong Song <y...@fb.com>
Sparse is not happy about it. Could you add some annotations, perhaps?
include/linux/rcupdate.h:686:9: warning: context imbalance in
'bpf_sk_storage_map_seq_find_next' - unexpected unlock
include/linux/rcupdate.h:686:9: warning: context imbalance in
'bpf_sk_storage_map_seq_stop' - unexpected unlock
