[PATCH 000/227] use NUMA node terminology across the tree
chengwen.feng at linux.dev
chengwen.feng at linux.dev
Wed Nov 4 13:24:22 CET 2026
November 4, 2026 at 7:40 PM, "Thomas Monjalon" <thomas at monjalon.net mailto:thomas at monjalon.net?to=%22Thomas%20Monjalon%22%20%3Cthomas%40monjalon.net%3E > wrote:
>
> 04/11/2026 10:29, David Marchand:
>
> >
> > On Wed, 4 Nov 2026 at 09:56, fengchengwen <fengchengwen at huawei.com> wrote:
> > On 11/4/2026 3:34 PM, Stephen Hemminger wrote:
> > >> From: Chengwen Feng <fengchengwen at huawei.com>
> > >> 905 files changed, 6113 insertions(+), 5439 deletions(-)
> > >
> > > Maybe a single coccinelle script would help here? Rather than trying to look
> > > at so many patches.
> >
> > A coccinelle script is a helpful idea. I will try to prepare one.
> >
> > Well, maybe.
> > But never a 200+ patches series again.
> >
> > Putting aside the way it was submitted, it seems this change is only
> > cosmetic, and it was not announced in a *previous* deprecation notice.
> >
> > Why do we need this change?
> > Why now, in this release?
> >
> I assume it is not for 26.11.
> Chengwen wanted to fix the namespace of SOCKET_ID_ANY, adding RTE_.
> And I told him that we should avoid changing this API 2 times
> when we will want to fix the terminology
> around CPU socket and NUMA memory domain.
> I think we should do both at once.
> That's why we see this big change.
>
> BUT
>
> It would be more efficient for such a change to discuss it first.
> Then once we agree on the names and usage,
> we should make it in a way it can be reproduced, maintained
> and ease future backport of patches containing the new terminology.
> A coccinelle script looks appropriate.
> And of course, the patch series should not have so many patches,
> it does not make sense.
> A mechanical change should come in one patch.
> If the series deserves several different mechanical changes,
> few patches should be sent for them, not hundreds.
>
Thanks Thomas for explaining the background of this change.
This series was AI-generated following the one commit per moduleconvention
to support per-subsystem review, with the expectationthat these commits
would be squashed before merge.
I understand this split imposed heavy CI load now, and I will adjustthe patch
organisation in the next version.
Moving forward, I will prepare a coccinelle script for the mechanicalconversion
and follow the deprecation process before formal submission.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mails.dpdk.org/archives/dev/attachments/20261104/6b2faa2e/attachment-0001.htm>
More information about the dev
mailing list