[v4] Splitting commits for IP_TOS support in setsockopt()

[PATCH mptcp-next v4 0/2] Splitting commits for IP_TOS support in setsockopt()

Poorva Sonparote posted 2 patches 2 years, 5 months ago

Diff against v3
Download series mbox

Patches applied successfully (tree, apply log)
git fetch https://github.com/multipath-tcp/mptcp_net-next tags/patchew/cover.1635367735.git.psonparo@redhat.com

Maintainers: "David S. Miller" <davem@davemloft.net>, Jakub Kicinski <kuba@kernel.org>, Mat Martineau <mathew.j.martineau@linux.intel.com>, Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>, David Ahern <dsahern@kernel.org>, Matthieu Baerts <matthieu.baerts@tessares.net>

include/net/ip.h       |  1 +
net/ipv4/ip_sockglue.c |  2 +-
net/mptcp/sockopt.c    | 40 ++++++++++++++++++++++++++++++++++++++++
3 files changed, 42 insertions(+), 1 deletion(-)

Expand all Fold all

[PATCH mptcp-next v4 0/2] Splitting commits for IP_TOS support in setsockopt()

Posted by Poorva Sonparote 2 years, 5 months ago

- patch 1/2 Add Support for IP_TOS for MPTCP setsockopt()
- patch 2/2 Exposing  __ip_sock_set_tos() in ip.h

Poorva Sonparote (2):
  Exposing __ip_sock_set_tos() in ip.h
  Support for IP_TOS for MPTCP setsockopt()

 include/net/ip.h       |  1 +
 net/ipv4/ip_sockglue.c |  2 +-
 net/mptcp/sockopt.c    | 40 ++++++++++++++++++++++++++++++++++++++++
 3 files changed, 42 insertions(+), 1 deletion(-)

-- 
2.18.2

Re: [PATCH mptcp-next v4 0/2] Splitting commits for IP_TOS support in setsockopt()

Posted by Mat Martineau 2 years, 5 months ago

On Wed, 27 Oct 2021, Poorva Sonparote wrote:

> - patch 1/2 Add Support for IP_TOS for MPTCP setsockopt()
> - patch 2/2 Exposing  __ip_sock_set_tos() in ip.h
>
> Poorva Sonparote (2):
>  Exposing __ip_sock_set_tos() in ip.h
>  Support for IP_TOS for MPTCP setsockopt()
>
> include/net/ip.h       |  1 +
> net/ipv4/ip_sockglue.c |  2 +-
> net/mptcp/sockopt.c    | 40 ++++++++++++++++++++++++++++++++++++++++
> 3 files changed, 42 insertions(+), 1 deletion(-)
>
> -- 
> 2.18.2

Hi Poorva -

Thanks! Looks like this is ready for the export branch:

Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>

It's not necessary to use getsockopt to implement a test for this, you 
could also use 'ss --tos' to see the tos value on the mptcp socket and the 
subflows. Would be good to have a selftest for this before upstreaming.

--
Mat Martineau
Intel

Re: [PATCH mptcp-next v4 0/2] Splitting commits for IP_TOS support in setsockopt()

Posted by Matthieu Baerts 2 years, 5 months ago

Hi Poorva, Mat,

On 28/10/2021 02:50, Mat Martineau wrote:
> On Wed, 27 Oct 2021, Poorva Sonparote wrote:
> 
>> - patch 1/2 Add Support for IP_TOS for MPTCP setsockopt()
>> - patch 2/2 Exposing  __ip_sock_set_tos() in ip.h
>>
>> Poorva Sonparote (2):
>>  Exposing __ip_sock_set_tos() in ip.h
>>  Support for IP_TOS for MPTCP setsockopt()

Thank you for the patches and the reviews!

> Thanks! Looks like this is ready for the export branch:
> 
> Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>

Now in our tree (features for net-next) + added prefixes in your patches
("ipv4:" and "mptcp:"):

- f80cd2ae7b35: Exposing __ip_sock_set_tos() in ip.h
- dcb0d4eb4318: Support for IP_TOS for MPTCP setsockopt()
- 758c711752bb: tg:msg: prefix with ipv4 and add Mat's ACK
- 76fecc38fcfb: tg:msg: prefix with mptcp and add Mat's RvB
- Results: 4af534378cb0..a1638f497c0c

Builds and tests are now in progress:

https://cirrus-ci.com/github/multipath-tcp/mptcp_net-next/export/20211028T092928
https://github.com/multipath-tcp/mptcp_net-next/actions/workflows/build-validation.yml?query=branch:export

> It's not necessary to use getsockopt to implement a test for this, you
> could also use 'ss --tos' to see the tos value on the mptcp socket and
> the subflows. Would be good to have a selftest for this before upstreaming.

Should we add something in the selftests for each "simple" sockopt? I
mean: it can be quite "heavy" to create E2E connections, wait for them
to be created and be long enough, check for an output with external
tool(s), etc.

For more "complex" sockopts like MPTCP_INFO, it makes sense because it
would imply a lot of modifications on packetdrill side.

Should we discuss about that at our weekly meeting?

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