>Do you mean that in the next iteration you are going to remove the 'enum >addr_type_t type' filed from struct inet_fill_args? That would be good.
Sorry, I already sent out the v6 patch. I should wait for more time. PTAL at the v6 patch thread and we can continue the discussion there. >As per the header file, I tend to think that the better choice in this >case would probably be creating a new one. Having the `struct inet_fill_args` in igmp.h will make the IPv4 logic consistent with IPv6 logic. IPv4: `struct inet_fill_args` in igmp.h and use it in igmp.c/devinet.c IPv6: ``struct inet6_fill_args` in addrconf.h and use it in mld.c/addrconf.c Thanks, Yuyang On Fri, Jan 17, 2025 at 6:53 PM Paolo Abeni <pab...@redhat.com> wrote: > > On 1/17/25 7:56 AM, Yuyang Huang wrote: > > Hi Paolo > > > > Thanks for the review feedback, I will adjust them in the patch v6. > > > >> Why moving the struct definition here? IMHO addrconf.h is better suited > >> and will avoid additional headers dep. > > > > The `struct inet_fill_args` is moved from devinet.c to igmp.h to make > > it usable in both devinet.c and igmp.c. > > I feel it is incorrect to move `struct inet_fill_args` to addrconf.h > > because that file contains IPv6 specific definition and it also > > contains `struct inet6_fill_args`. After refactoring, I will remove > > the usage of enum addr_type_t type, so we don't need to import > > addrconf.h anymore. > > Do you mean that in the next iteration you are going to remove the 'enum > addr_type_t type' filed from struct inet_fill_args? That would be good. > > As per the header file, I tend to think that the better choice in this > case would probably be creating a new one. > > /P >
