04/08/2021 15:53, Kinsella, Ray: > On 04/08/2021 14:12, Thomas Monjalon wrote: > > 04/08/2021 15:00, Xueming(Steven) Li: > >> From: Kinsella, Ray <m...@ashroe.eu> > >>> On 04/08/2021 13:11, Xueming(Steven) Li wrote: > >>>> From: Kinsella, Ray <m...@ashroe.eu> > >>>>> Its not strictly a depreciation notice though, you are not breaking > >>>>> anything right. > >>>>> Since you are not breaking anything, don't think the notice is required > >>>>> in the 21.11 timeframe. > >>>>> > >>>>> Now if you where doing it in 21.08, it would be an ABI change and that > >>>>> would be a different story. > >>>> > >>>> Thanks for looking at this! > >>>> Yes, it targets to 21.11. The offloading flag is fine, but the > >>>> shared_group does break ABI, detail: > >>>> https://mails.dpdk.org/archives/dev/2021-July/215575.html > >>> > >>> Right ... its a new field, not a depreciation as such. > >>> What I mean by this is that no existing code is broken. > >>> > >>> 21.11 is a new ABI in any case and you are not depreciating anything, so > >>> no notice is required. > >> > >> Maybe it a new process, confirmed with Thomas, it's expected: > >> https://doc.dpdk.org/guides/contributing/abi_policy.html#abi-changes > > > > I think what Ray means is that it breaks ABI but not API, > > so he doesn't consider a notice is required. > > > My understanding of the policy is that *any* ABI change requires a notice. > > But if you want to make it lighter and allow any non-announced ABI change > > in an ABI-breaking release, I think I would vote for. > > Thanks for clarifying Thomas ... you are correct.
In the meantime, let's review and ack notices, even if ABI-only change: https://patches.dpdk.org/bundle/tmonjalo/deprecation-notices/ We'll discuss later if we can accept more ABI change, but we should try to be on the safe side for those already announced.
