netdev - Re: [PATCH bpf-next 2/7] bpf: add bpf_skc_to_mptcp_sock

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: Windows password security audit tool. GUI, reports in PDF.

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

Date:   Mon, 25 Apr 2022 11:35:42 -0700 (PDT)
From:   Mat Martineau <mathew.j.martineau@...ux.intel.com>
To:     Geliang Tang <geliang.tang@...e.com>
cc:     Daniel Borkmann <daniel@...earbox.net>, netdev@...r.kernel.org,
        bpf@...r.kernel.org, ast@...nel.org, andrii@...nel.org,
        mptcp@...ts.linux.dev,
        Nicolas Rybowski <nicolas.rybowski@...sares.net>,
        Matthieu Baerts <matthieu.baerts@...sares.net>
Subject: Re: [PATCH bpf-next 2/7] bpf: add bpf_skc_to_mptcp_sock_proto

On Mon, 25 Apr 2022, Daniel Borkmann wrote:

> On 4/21/22 12:24 AM, Mat Martineau wrote:
> [...]
>>   static const struct bpf_func_proto *
>>   bpf_sk_base_func_proto(enum bpf_func_id func_id);
>> @@ -11279,6 +11280,19 @@ const struct bpf_func_proto 
>> bpf_skc_to_unix_sock_proto = {
>>   	.ret_btf_id		= &btf_sock_ids[BTF_SOCK_TYPE_UNIX],
>>   };
>>   +BPF_CALL_1(bpf_skc_to_mptcp_sock, struct sock *, sk)
>> +{
>> +	return (unsigned long)bpf_mptcp_sock_from_subflow(sk);
>> +}
>> +
>> +static const struct bpf_func_proto bpf_skc_to_mptcp_sock_proto = {
>> +	.func		= bpf_skc_to_mptcp_sock,
>> +	.gpl_only	= false,
>> +	.ret_type	= RET_PTR_TO_BTF_ID_OR_NULL,
>> +	.arg1_type	= ARG_PTR_TO_SOCK_COMMON,
>> +	.ret_btf_id	= &btf_sock_ids[BTF_SOCK_TYPE_MPTCP],
>> +};
>
> BPF CI 
> (https://github.com/kernel-patches/bpf/runs/6136052684?check_suite_focus=true) 
> fails with:
>
>  #7   base:FAIL
>  libbpf: prog '_sockops': BPF program load failed: Invalid argument
>  libbpf: prog '_sockops': -- BEGIN PROG LOAD LOG --
>  0: R1=ctx(off=0,imm=0) R10=fp0
>  ; int op = (int)ctx->op;
>  0: (61) r2 = *(u32 *)(r1 +0)          ; R1=ctx(off=0,imm=0) 
> R2_w=scalar(umax=4294967295,var_off=(0x0; 0xffffffff))
>  ; if (op != BPF_SOCK_OPS_TCP_CONNECT_CB)
>  1: (56) if w2 != 0x3 goto pc+50       ; R2_w=3
>  ; sk = ctx->sk;
>  2: (79) r6 = *(u64 *)(r1 +184)        ; R1=ctx(off=0,imm=0) 
> R6_w=sock_or_null(id=1,off=0,imm=0)
>  ; if (!sk)
>  3: (15) if r6 == 0x0 goto pc+48       ; R6_w=sock(off=0,imm=0)
>  ; tcp_sk = bpf_tcp_sock(sk);
>  4: (bf) r1 = r6                       ; R1_w=sock(off=0,imm=0) 
> R6_w=sock(off=0,imm=0)
>  5: (85) call bpf_tcp_sock#96          ; 
> R0_w=tcp_sock_or_null(id=2,off=0,imm=0)
>  6: (bf) r7 = r0                       ; 
> R0=tcp_sock_or_null(id=2,off=0,imm=0) R7=tcp_sock_or_null(id=2,off=0,imm=0)
>  ; if (!tcp_sk)
>  7: (15) if r7 == 0x0 goto pc+44       ; R7=tcp_sock(off=0,imm=0)
>  ; if (!tcp_sk->is_mptcp) {
>  8: (61) r1 = *(u32 *)(r7 +112)        ; 
> R1_w=scalar(umax=4294967295,var_off=(0x0; 0xffffffff)) 
> R7=tcp_sock(off=0,imm=0)
>  ; if (!tcp_sk->is_mptcp) {
>  9: (56) if w1 != 0x0 goto pc+14 24: R0=tcp_sock(off=0,imm=0) 
> R1_w=scalar(umax=4294967295,var_off=(0x0; 0xffffffff)) R6=sock(off=0,imm=0) 
> R7=tcp_sock(off=0,imm=0) R10=fp0
>  ; msk = bpf_skc_to_mptcp_sock(sk);
>  24: (bf) r1 = r6                      ; R1_w=sock(off=0,imm=0) 
> R6=sock(off=0,imm=0)
>  25: (85) call bpf_skc_to_mptcp_sock#194
>  invalid return type 8 of func bpf_skc_to_mptcp_sock#194
>  processed 34 insns (limit 1000000) max_states_per_insn 0 total_states 3 
> peak_states 3 mark_read 1
>  -- END PROG LOAD LOG --
>  libbpf: failed to load program '_sockops'
>  libbpf: failed to load object './mptcp_sock.o'
>  run_test:FAIL:165
>  test_base:FAIL:227
>  (network_helpers.c:88: errno: Protocol not supported) Failed to create 
> server socket
>  test_base:FAIL:241
>  RTNETLINK answers: No such file or directory
>  Error talking to the kernel
> [...]
>
> Looking at bpf_skc_to_tcp6_sock(), do we similarly need a BTF_TYPE_EMIT() 
> here?
>

Geliang, in addition to the BTF_TYPE_EMIT() can you also take a look at 
the places in kernel/trace/bpf_trace.c and kernel/bpf/verifier.c where 
bpf_skc_to_tcp6_sock and bpf_skc_to_tcp6_sock_proto are referenced?

--
Mat Martineau
Intel