Re: [MPTCP] [PATCH selftests 0/5] selftests: prepare for mptcp ipv6 support

Re: [MPTCP] [PATCH selftests 0/5] selftests: prepare for mptcp ipv6 support

[Florian Westphal]

[-- Attachment #1: Type: text/plain, Size: 319 bytes --]

Paolo Abeni <pabeni(a)redhat.com> wrote:
> I was also planning to try to estimate the effort there, but I sure do
> not want to arm-wrestle you! I'll fallback to other tasks - recvmsg
> cleanup? mp_join debug? - but please let me know if I can help somehow!

Someone could add sha256 support for mptcpv1 support.

Re: [MPTCP] [PATCH selftests 0/5] selftests: prepare for mptcp ipv6 support

[Peter Krystad]

[-- Attachment #1: Type: text/plain, Size: 1065 bytes --]

On Tue, 2019-09-17 at 11:28 +0200, Paolo Abeni wrote:
> Hi,
> 
> On Mon, 2019-09-16 at 13:50 -0700, Peter Krystad wrote:
> > On Fri, 2019-09-13 at 11:45 +0200, Florian Westphal wrote:
> > > This patch series inceases self test coverage:
> > > 
> > > 1. ipv6 support.  So all someone needs to do is actually implement ipv6
> > >    support in mptcp.  Then, set ipv6=true in the selftest script and it
> > >    should 'just work'.
> > 
> > I am planning to begin work on IPv6 support, unless someone else wants to
> > arm-wrestle me for it....
> 
> Thank you for doing that! 
> 
> I was also planning to try to estimate the effort there, but I sure do
> not want to arm-wrestle you! I'll fallback to other tasks - recvmsg
> cleanup? mp_join debug? - but please let me know if I can help somehow!
> 
Well, if we are going to go with v1 for our initial submission, maybe it makes
sense for me to work on the v1 MP_CAPABLE exchange since I did the original
MP_CAPABLE work and leave IPv6 for you or....

Peter.


> Thanks,
> 
> Paolo
> 

Re: [MPTCP] [PATCH selftests 0/5] selftests: prepare for mptcp ipv6 support

[Paolo Abeni]

[-- Attachment #1: Type: text/plain, Size: 744 bytes --]

Hi,

On Mon, 2019-09-16 at 13:50 -0700, Peter Krystad wrote:
> On Fri, 2019-09-13 at 11:45 +0200, Florian Westphal wrote:
> > This patch series inceases self test coverage:
> > 
> > 1. ipv6 support.  So all someone needs to do is actually implement ipv6
> >    support in mptcp.  Then, set ipv6=true in the selftest script and it
> >    should 'just work'.
> 
> I am planning to begin work on IPv6 support, unless someone else wants to
> arm-wrestle me for it....

Thank you for doing that! 

I was also planning to try to estimate the effort there, but I sure do
not want to arm-wrestle you! I'll fallback to other tasks - recvmsg
cleanup? mp_join debug? - but please let me know if I can help somehow!

Thanks,

Paolo

Re: [MPTCP] [PATCH selftests 0/5] selftests: prepare for mptcp ipv6 support

[Peter Krystad]

[-- Attachment #1: Type: text/plain, Size: 1057 bytes --]

On Fri, 2019-09-13 at 11:45 +0200, Florian Westphal wrote:
> This patch series inceases self test coverage:
> 
> 1. ipv6 support.  So all someone needs to do is actually implement ipv6
>    support in mptcp.  Then, set ipv6=true in the selftest script and it
>    should 'just work'.

I am planning to begin work on IPv6 support, unless someone else wants to
arm-wrestle me for it....

Peter.

> 2. Check accept/getpeername.  In case of connect side, we also check
>    that getpeername returned addresses are the ones we connected to.
> 3. less predictable topology: tso/gso/gro and tc parameters are more
>    random.
> 
> I suggest to apply them with no squashing, once we go for another RFC
> tests/selftest can be a single patch and all commits can be squashed.
> 
> If you disagree, I don't mind if this gets squashed right away.
> 
> No changes to mptcp stack itself here.
> 
> 
> _______________________________________________
> mptcp mailing list
> mptcp(a)lists.01.org
> https://lists.01.org/mailman/listinfo/mptcp

[MPTCP] [PATCH selftests 0/5] selftests: prepare for mptcp ipv6 support

[Florian Westphal]

[-- Attachment #1: Type: text/plain, Size: 699 bytes --]

This patch series inceases self test coverage:

1. ipv6 support.  So all someone needs to do is actually implement ipv6
   support in mptcp.  Then, set ipv6=true in the selftest script and it
   should 'just work'.
2. Check accept/getpeername.  In case of connect side, we also check
   that getpeername returned addresses are the ones we connected to.
3. less predictable topology: tso/gso/gro and tc parameters are more
   random.

I suggest to apply them with no squashing, once we go for another RFC
tests/selftest can be a single patch and all commits can be squashed.

If you disagree, I don't mind if this gets squashed right away.

No changes to mptcp stack itself here.