Re: [PATCH net-next v11 0/5] Introducing P4TC (series 1)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Fri, Feb 23, 2024 at 10:38 AM Jakub Kicinski <kuba@xxxxxxxxxx> wrote:
>
> On Fri, 23 Feb 2024 08:17:23 -0500 Jamal Hadi Salim wrote:
> > This is the first patchset of two. In this patch we are only submitting 5
> > patches which touch the general TC code given these are trivial. We will be
> > posting a second patchset which handles the P4 objects and associated infra
> > (which includes 10 patches that we have already been posting to hit the 15
> > limit).
>
> Don't use the limit as a justification for questionable tactics :|

The discussion with Daniel was on removing the XDP referencing in the
filter which in my last email exchange i offered to remove. I believe
that removing the XDP reference should resolve the issue. Instead of
waiting for Daniel to respond (the last response took a while), at the
last minute i decided to only do the first five which are trivial and
have been posted for over a year now and have been reviewd by multiple
people. I had no intention to make this a conspiracy of any sort.

> If there's still BPF in it, it's not getting merged without BPF
> maintainers's acks. So we're going to merge these 5 and then what?
>

Look, this is getting to be too much navigation (which is what scares
most newbies from participating and i have been here for 100 years
now). I dont have a problem removing the XDP reference - but now we
are treading into subsystem territories; this is about P4 abstraction
and the infra around it, it is really not about eBPF. We dont need
anything "new" from ebpf i.e none of these patches make any eBPF
changes. This is a tc classifier that happens to use ebpf, whose
domain is that?  The exhausting part is some feedback is "you do it
our way or else" thing: I apologize i dont want to lump everyone in
that pool, and it is nothing to do specifically with ebpf people
rather a failure in reaching compromise within this community.

> BTW the diffstat at the end of the commit message is from the full set

Yeah, sorry - this was last minute "should i send all or just these
five to make space for the remainder 5 that we havent posted since
V8".
We have 5 more patches on top of what we posted on V10 - merging these
would open the door to post the rest with 15 limit.
I can repost all 15 if that works better.

cheers,
jamal





[Index of Archives]     [Linux Samsung SoC]     [Linux Rockchip SoC]     [Linux Actions SoC]     [Linux for Synopsys ARC Processors]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]


  Powered by Linux