Re: [PATCH mptcp-next] selftests: bpf: exercise bpf_mptcp_sock()

[Geliang Tang <geliang.tang@suse.com>]

Hi Matt,

On Tue, Mar 01, 2022 at 12:56:08PM +0100, Matthieu Baerts wrote:
> Hi Geliang,
> 
> On 24/02/2022 16:21, Geliang Tang wrote:
> > This patch extended the MPTCP test base, to exercise bpf_mptcp_sock() from
> > C test as Alexei suggested in [1].
> 
> Thank you for this patch!
> 
> > diff --git a/tools/testing/selftests/bpf/prog_tests/mptcp.c b/tools/testing/selftests/bpf/prog_tests/mptcp.c
> > index 04aef0f147dc..eba1b6d12a8c 100644
> > --- a/tools/testing/selftests/bpf/prog_tests/mptcp.c
> > +++ b/tools/testing/selftests/bpf/prog_tests/mptcp.c
> > @@ -6,6 +6,7 @@
> >  struct mptcp_storage {
> >  	__u32 invoked;
> >  	__u32 is_mptcp;
> > +	__u32 token;
> >  };
> >  
> >  static int verify_sk(int map_fd, int client_fd, const char *msg, __u32 is_mptcp)
> 
> Do you think we could verify that the token has been properly read and
> stored here? I know there were some limitations but probably we can work
> around them, no? We could store items on the msk and not the ssk for
> MPTCP connections, would that not help?

Yes, we should verify the token. Could you please give me more details
about how to store items on the msk? I don't know how the items are stored
on the ssk either, please show me which code is used to implement this.

Thanks,
-Geliang

> 
> Cheers,
> Matt
> -- 
> Tessares | Belgium | Hybrid Access Solutions
> www.tessares.net
>