RE: [PATCH net-next v2] macsec: Add support for IFLA_MACSEC_OFFLOAD in the netlink layer

[Emeel Hakim]

> -----Original Message-----
> From: Sabrina Dubroca <sd@xxxxxxxxxxxxxxx>
> Sent: Tuesday, 6 December 2022 18:11
> To: Jiri Pirko <jiri@xxxxxxxxxxx>
> Cc: Emeel Hakim <ehakim@xxxxxxxxxx>; linux-kernel@xxxxxxxxxxxxxxx; Raed
> Salem <raeds@xxxxxxxxxx>; davem@xxxxxxxxxxxxx; edumazet@xxxxxxxxxx;
> kuba@xxxxxxxxxx; pabeni@xxxxxxxxxx; netdev@xxxxxxxxxxxxxxx
> Subject: Re: [PATCH net-next v2] macsec: Add support for IFLA_MACSEC_OFFLOAD
> in the netlink layer
> 
> External email: Use caution opening links or attachments
> 
> 
> 2022-12-06, 14:35:23 +0100, Jiri Pirko wrote:
> > Tue, Dec 06, 2022 at 01:31:54PM CET, ehakim@xxxxxxxxxx wrote:
> > >> Tue, Dec 06, 2022 at 09:57:57AM CET, ehakim@xxxxxxxxxx wrote:
> > >> >From: Emeel Hakim <ehakim@xxxxxxxxxx>
> > >> >
> > >> >This adds support for configuring Macsec offload through the
> > >>
> > >> Tell the codebase what to do. Be imperative in your patch
> > >> descriptions so it is clear what are the intensions of the patch.
> > >
> > >Ack
> > >
> > >>
> > >>
> > >> >netlink layer by:
> > >> >- Considering IFLA_MACSEC_OFFLOAD in macsec_fill_info.
> > >> >- Handling IFLA_MACSEC_OFFLOAD in macsec_changelink.
> > >> >- Adding IFLA_MACSEC_OFFLOAD to the netlink policy.
> > >> >- Adjusting macsec_get_size.
> > >>
> > >> 4 patches then?
> > >
> > >Ack, I will change the commit message to be imperative and will replace the list
> with a good description.
> > >I still believe it should be a one patch since splitting this could break a bisect
> process.
> >
> > Well, when you split, you have to make sure you don't break bisection,
> > always. Please try to figure that out.
> 
> I think this can be split pretty nicely into 3 patches:
>  - add IFLA_MACSEC_OFFLOAD to macsec_rtnl_policy (probably for net
>    with a Fixes tag on the commit that introduced IFLA_MACSEC_OFFLOAD)
>  - add offload to macsec_fill_info/macsec_get_size
>  - add IFLA_MACSEC_OFFLOAD support to changelink
> 
> The subject of the last patch should also make it clear that it's only adding
> IFLA_MACSEC_OFFLOAD to changelink. As it's written, someone could assume
> there's no support at all via rtnl ops and wonder why this patch isn't doing anything
> to newlink, and whether/why this IFLA_MACSEC_OFFLOAD already exists.

Ack , I will split the patch an send the new patches.

> --
> Sabrina