v2 changes:
 - de-uglify YAML spef file, replacing '_' with '-' and avoiding
   unnecessary assignments in enums (Jakub Kicinski)
 - keep validating addr6 with NLA_POLICY_EXACT_LEN() (Jakub Kicinski)
 - Don't leave MPTCP_EVENT_UNSPEC undocumented (Simon Horman)

patch 1/5 will cause a small context mismatch in "net: mptcp: use policy
generated by YAML spec"

Davide Caratti (5):
  Squash-to: "net: mptcp: convert netlink from small_ops to ops"
  Squash-to: "Documentation: netlink: add a YAML spec for mptcp"
  Squash-to: "uapi: mptcp: use header file generated from YAML spec"
  Squash to: "net: mptcp: use policy generated by YAML spec"
  tools: ynl-gen: add support for exact-len validation

-- 
2.41.0

Hi Davide, Paolo,

On 16/10/2023 15:43, Davide Caratti wrote:
> v2 changes:
>  - de-uglify YAML spef file, replacing '_' with '-' and avoiding
>    unnecessary assignments in enums (Jakub Kicinski)
>  - keep validating addr6 with NLA_POLICY_EXACT_LEN() (Jakub Kicinski)
>  - Don't leave MPTCP_EVENT_UNSPEC undocumented (Simon Horman)
> 
> patch 1/5 will cause a small context mismatch in "net: mptcp: use policy
> generated by YAML spec"

Thank you for the new version.

This version looks good to me, but do you mind checking with Paolo if it
is also OK for him because he gave his Acked-by on the previous version.

When that will be done, I will apply the different patches and send them
to netdev. I can re-use your v2 changelog in the cover-letter. Or do you
want me to write something else?

One last thing: can I apply your last patch ("tools: ynl-gen: add
support for exact-len validation") between these two?

- "tools: ynl: add uns-admin-perm to genetlink legacy"
- "net: mptcp: convert netlink from small_ops to ops"

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

Hi Davide, Paolo,

On 19/10/2023 12:48, Matthieu Baerts wrote:
> Hi Davide, Paolo,
> 
> On 16/10/2023 15:43, Davide Caratti wrote:
>> v2 changes:
>>  - de-uglify YAML spef file, replacing '_' with '-' and avoiding
>>    unnecessary assignments in enums (Jakub Kicinski)
>>  - keep validating addr6 with NLA_POLICY_EXACT_LEN() (Jakub Kicinski)
>>  - Don't leave MPTCP_EVENT_UNSPEC undocumented (Simon Horman)
>>
>> patch 1/5 will cause a small context mismatch in "net: mptcp: use policy
>> generated by YAML spec"
> 
> Thank you for the new version.
> 
> This version looks good to me, but do you mind checking with Paolo if it
> is also OK for him because he gave his Acked-by on the previous version.

I hope that's OK: I just applied them in our tree. I will still wait for
Paolo's ACK before sending the v2 upstream

- 22e720cbc227: "squashed" patch 1/5 in "net: mptcp: convert netlink
from small_ops to ops"
- a1d16bbc8e6b: "squashed" patch 2/5 in "Documentation: netlink: add a
YAML spec for mptcp"
- 77fef1f8df04: "squashed" patch 3/5 in "uapi: mptcp: use header file
generated from YAML spec"

- dc863cd3742e: conflict in t/net-mptcp-use-policy-generated-by-YAML-spec

- e258da3fda0d: "squashed" patch 4/5 in "net: mptcp: use policy
generated by YAML spec"

- 3c523c3a147b: tools: ynl-gen: add support for exact-len validation

- Results: e5c88dc2d209..65682b5b41eb (export)

I had to resolved conflicts with patch 5/5 as you predicted. It was not
difficult but please double-check I did that correctly :)

I also checked that I could re-generate the files, and they would still
look the same.


Tests are now in progress:

https://cirrus-ci.com/github/multipath-tcp/mptcp_net-next/export/20231021T134222

Cheers,
Matt

hello Matthieu, thanks for looking at this!

On Thu, Oct 19, 2023 at 12:48 PM Matthieu Baerts <matttbe@kernel.org> wrote:
>
> Hi Davide, Paolo,
>
[...]

> When that will be done, I will apply the different patches and send them
> to netdev. I can re-use your v2 changelog in the cover-letter. Or do you
> want me to write something else?

Ok for me,

> One last thing: can I apply your last patch ("tools: ynl-gen: add
> support for exact-len validation") between these two?
>
> - "tools: ynl: add uns-admin-perm to genetlink legacy"
> - "net: mptcp: convert netlink from small_ops to ops"

that's ok also. Let's hope it won't conflict with [1] at the next sync
:) But if it does, let me know and I'll try another squash-to.

[1] https://lore.kernel.org/netdev/20231018163917.2514503-1-kuba@kernel.org/
thanks!
-- 
davide