Re: [PATCH bpf-next v4 2/2] selftests/bpf: Add mptcp subflow subtest

[Martin KaFai Lau <martin.lau@linux.dev>]

On 8/17/24 7:13 PM, Geliang Tang wrote:
>> Take a look at a recent example [0]. The mptcp test is under a cgroup
>> already
>> and has the cgroup setup. An extra "cgroup/getsockopt" prog should be
>> enough.
>> That prog can walk the msk->conn_list and use bpf_rdonly_cast (or the
 >> bpf_core_cast macro in libbpf) to cast a pointer to tcp_sock for

[ ... ]

> SEC("cgroup/getsockopt")
> int _getsockopt(struct bpf_sockopt *ctx)
> {
>          struct mptcp_sock *msk = bpf_core_cast(ctx->sk, struct
> mptcp_sock);
>          struct mptcp_subflow_context *subflow;
>          __u32 token = 0;
> 
>          if (!msk || ctx->level != SOL_TCP ||
> 			ctx->optname != TCP_CONGESTION)
>                  return 1;
> 
>          subflow = list_first_entry(&msk->conn_list,
> 			struct mptcp_subflow_context, node);
>          token = subflow->token;
>          bpf_trace_printk(fmt, sizeof(fmt), msk, token);
> 
>          return 1;
> }
> 
> And got some access errors:
> 
> ; token = subflow->token; @ mptcp_subflow.c:92
> 13: (61) r4 = *(u32 *)(r1 +524)
> access beyond struct list_head at off 524 size 4

Similar to your bpf_core_cast() usage earlier that casts a sock ptr to 
mptcp_sock ptr. r1 is in list_head ptr type. It needs to cast to 
mptcp_subflow_context ptr.

The same cast to tcp_sock is needed when you try to get the tcp_sock ptr from 
subflow->tcp_sock which is in "struct sock *" type in the kernel instead of 
"struct tcp_sock *".