On 2/25/25 2:11 PM, Dumitru Ceara wrote:
> On 2/25/25 2:04 PM, martin.kal...@canonical.com wrote:
>> On Tue, 2025-02-25 at 00:01 +0100, Dumitru Ceara wrote:
>>> It makes more sense to make these static (internal) in the module
>>> where
>>> they're actually used. They use "public" APIs to add advertised
>>> dynamic
>>> routes.
>>>
>>> Fixes: cd4ad2f56179 ("northd: Redistribution of NAT/LB routes.")
>>> Signed-off-by: Dumitru Ceara <dce...@redhat.com>
>>
>> I have no objections to moving these functions out of northd.c, so
>> overall this can have Acked-by me. However, since we are at it,
>> wouldn't it make sense to also move out the `build_parsed_routes` (and
>> related functions), as they too, seem to be only used from `en-
>> northd.c`.
>>
>
> Good point, I'll move those too in v2.
>
Looking more closely at it, I think we could go even further and move
all the parsed_route types and functions to a separate parsed-route.[ch]
module. However, I'd like to do that (and move build_parsed_routes()
too) as a follow up, on main. That would allow us to minimize the set
of changes we need to backport to 25.03 now.
What do you think?
Thanks,
Dumitru
_______________________________________________
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
