Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Thu, 18 Aug 2022 20:21:15 +0530
From:   Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
To:     Jon Mason <jdmason@kudzu.us>
Cc:     ntb@lists.linux.dev, Bjorn Helgaas <helgaas@kernel.org>,
        Frank.Li@nxp.com, linux-pci@vger.kernel.org,
        linux-kernel@vger.kernel.org, kishon@ti.com
Subject: Re: [PATCH] MAINTAINERS: add PCI Endpoint NTB drivers to NTB files
Message-ID: <20220818145115.GA111116@thinkpad>
References: <20220812194205.388967-1-jdmason@kudzu.us>
 <20220818060230.GA12008@thinkpad>
 <Yv5EA8uuhwn049jx@kudzu.us>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <Yv5EA8uuhwn049jx@kudzu.us>
Precedence: bulk

On Thu, Aug 18, 2022 at 09:52:03AM -0400, Jon Mason wrote:
> On Thu, Aug 18, 2022 at 11:32:30AM +0530, Manivannan Sadhasivam wrote:
> > + Kishon (PCI EP Maintainer)
> > 
> > On Fri, Aug 12, 2022 at 03:42:05PM -0400, Jon Mason wrote:
> > > The PCI Endpoint NTB drivers are under the NTB umbrella.  Add an entry
> > > there to allow for notification of changes for it.
> > > 
> > > Signed-off-by: Jon Mason <jdmason@kudzu.us>
> > 
> > Hi Jason,
> 
> I assume you mean me.  Odd that you got my name wrong 2 lines below it
> being properly written out.
> 

Terribly sorry about that! I was reading another thread just before this
and misspelled your name.

> > I know that this patch is already in Linus's tree but I think this PCI Endpoint
> > VNTB driver is not going in a correct path. First, Kishon is not convinced with
> > the way the PCI Endpoint VNTB function driver is written currently. He prefers
> > the VirtIO approach over the current one [1].
> 
> To your point, this is already in Linus' tree.  If it is not the way
> people want it, patches accepted.
> 
> Kishon (in the thread) recommended doing it one way, and Frank
> responded he liked doing it another.  Kishon didn't respond to that
> last email.  To me, this is an acceptable technical disagreement that
> can be addressed in the future and no need to prevent working patches
> from being accepted.
> 

Kishon being the maintainer proposed an entirely different way of representing
the driver. I agree that the patch is working but maintainer's view matters and
if you don't hear from the maintainer for some time, you'll ping them (Frank
did ping but there is something called RESEND).

I'm not sure that merging the patches without an ACK from the relevant subsystem
maintainer is the right thing to do.

> > But while the conversation was still going on, the series got merged via NTB
> > tree without any ACKs from the PCI/PCI_EP maintainers. Also, note that there
> > was a patch touching the PCI Controller driver as well and that was also not
> > ACKed [2].
> 
> I put the series in my ntb-next branch, which was pulled into linux-next
> for roughly 3 months, and he did not object then (though likely he did
> not notice).  Multiple patches were submitted to the relevant mailing
> lists to address minor issues in the series (from being in linux-next)
> and most/all of those hit the PCI mailing list.  Bjorn responded to
> all of them saying they needed to go through the ntb tree (because of
> the dependency on Frank Li's original series).  So while not an
> explicit ack, it was implicit to me in that he was aware of the
> series.
> 
> Given the length of time and the public work on the series, how much
> longer should I have waited for a nack?
> 

I'd argue that you should've waited for the ACK first. I've seen and
experienced patch series hanging there for multiple releases. I'm not in favour
of not responding to the patches, maintainers do have their own work to do but
merging the patches touching the different subsystem without an ACK doesn't
sound good to me.

I don't know why he didn't object when the series got merged in this manner :/

> > If this trend is going to continue in the coming days, then I'm afraid that NTB
> > might end up being a backdoor for PCI/PCI_EP patches :(
> 
> Completely unfounded, per Bjorn's comment on
> https://lore.kernel.org/all/20220815183920.GA1960006@bhelgaas/
> 

It's now fine that NTB related PCI patches can be merged through NTB tree but
please wait for an ACK for patches touching the non-NTB drivers. If you ask me
how long you should wait, then I don't have an answer, but atleast give a
notice before doing so that it can catch the proper eyes.

Thanks,
Mani

> Thanks,
> Jon
> 
> > 
> > Thanks,
> > Mani
> > 
> > [1] https://lore.kernel.org/all/20220222162355.32369-4-Frank.Li@nxp.com
> > [2] https://lore.kernel.org/all/20220222162355.32369-2-Frank.Li@nxp.com
> > 
> > > ---
> > >  MAINTAINERS | 1 +
> > >  1 file changed, 1 insertion(+)
> > > 
> > > diff --git a/MAINTAINERS b/MAINTAINERS
> > > index 64379c699903..47e9f86bd712 100644
> > > --- a/MAINTAINERS
> > > +++ b/MAINTAINERS
> > > @@ -14254,6 +14254,7 @@ W:	https://github.com/jonmason/ntb/wiki
> > >  T:	git git://github.com/jonmason/ntb.git
> > >  F:	drivers/net/ntb_netdev.c
> > >  F:	drivers/ntb/
> > > +F:	drivers/pci/endpoint/functions/pci-epf-*ntb.c
> > >  F:	include/linux/ntb.h
> > >  F:	include/linux/ntb_transport.h
> > >  F:	tools/testing/selftests/ntb/
> > > -- 
> > > 2.30.2
> > > 
> > 
> > -- 
> > மணிவண்ணன் சதாசிவம்

-- 
மணிவண்ணன் சதாசிவம்