netdev - Re: [PATCHv19 bpf-next 2/6] bpf: add a new bpf argument type ARG_CONST_MAP_PTR_OR

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <737a6e4f-1551-e686-f180-e5d08347dbea@iogearbox.net>
Date:   Mon, 15 Feb 2021 23:03:21 +0100
From:   Daniel Borkmann <daniel@...earbox.net>
To:     Hangbin Liu <liuhangbin@...il.com>, bpf@...r.kernel.org
Cc:     netdev@...r.kernel.org,
        Toke Høiland-Jørgensen <toke@...hat.com>,
        Jiri Benc <jbenc@...hat.com>,
        Jesper Dangaard Brouer <brouer@...hat.com>,
        Eelco Chaudron <echaudro@...hat.com>, ast@...nel.org,
        Lorenzo Bianconi <lorenzo.bianconi@...hat.com>,
        David Ahern <dsahern@...il.com>,
        Andrii Nakryiko <andrii.nakryiko@...il.com>,
        Alexei Starovoitov <alexei.starovoitov@...il.com>,
        John Fastabend <john.fastabend@...il.com>,
        Maciej Fijalkowski <maciej.fijalkowski@...el.com>
Subject: Re: [PATCHv19 bpf-next 2/6] bpf: add a new bpf argument type
 ARG_CONST_MAP_PTR_OR_NULL

On 2/8/21 11:47 AM, Hangbin Liu wrote:
> Add a new bpf argument type ARG_CONST_MAP_PTR_OR_NULL which could be
> used when we want to allow NULL pointer for map parameter. The bpf helper
> need to take care and check if the map is NULL when use this type.
> 
> Acked-by: Toke Høiland-Jørgensen <toke@...hat.com>
> Acked-by: John Fastabend <john.fastabend@...il.com>
> Signed-off-by: Hangbin Liu <liuhangbin@...il.com>
[...]
> @@ -4259,9 +4261,9 @@ static int check_func_arg(struct bpf_verifier_env *env, u32 arg,
>   		meta->ref_obj_id = reg->ref_obj_id;
>   	}
>   
> -	if (arg_type == ARG_CONST_MAP_PTR) {
> -		/* bpf_map_xxx(map_ptr) call: remember that map_ptr */
> -		meta->map_ptr = reg->map_ptr;
> +	if (arg_type == ARG_CONST_MAP_PTR ||
> +	    arg_type == ARG_CONST_MAP_PTR_OR_NULL) {
> +		meta->map_ptr = register_is_null(reg) ? NULL : reg->map_ptr;

Sorry, but after re-reading this is unfortunately still broken :( Looking at your
helper func proto:

+static const struct bpf_func_proto bpf_xdp_redirect_map_multi_proto = {
+	.func           = bpf_xdp_redirect_map_multi,
+	.gpl_only       = false,
+	.ret_type       = RET_INTEGER,
+	.arg1_type      = ARG_CONST_MAP_PTR,
+	.arg2_type      = ARG_CONST_MAP_PTR_OR_NULL,
+	.arg3_type      = ARG_ANYTHING,
+};

So in check_helper_call() first meta->map_ptr is set to arg1 map, then when
checking arg2 map it can either be const NULL or a valid map ptr, but then
later on in check_map_func_compatibility() we only end up checking compatibility
of the /2nd/ map (e.g. on !meta->map_ptr we just return 0). This means, we
can pass whatever map type as arg1 map and it will pass the verifier just fine.