bpf: fix out-of-bounds read in __bpf_skc_lookup
authorLorenz Bauer <lmb@cloudflare.com>
Tue, 21 May 2019 07:52:38 +0000 (08:52 +0100)
committerDaniel Borkmann <daniel@iogearbox.net>
Tue, 21 May 2019 14:25:09 +0000 (16:25 +0200)
commit9b28ae243ef3b13d8a88b5451d025475c75ebdef
tree9f4de06e34f7e88112812578ad1a9a922b358197
parent221fb7268d67c0867a93f23586bd53c3c3969eee
bpf: fix out-of-bounds read in __bpf_skc_lookup

__bpf_skc_lookup takes a socket tuple and the length of the
tuple as an argument. Based on the length, it decides which
address family to pass to the helper function sk_lookup.

In case of AF_INET6, it fails to verify that the length
of the tuple is long enough. sk_lookup may therefore access
data past the end of the tuple.

Fixes: 6acc9b432e67 ("bpf: Add helper to retrieve socket in BPF")
Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
net/core/filter.c