[PATCH] nvme: NVME_TCP_OFFLOAD should not default to m

[PATCH] nvme: NVME_TCP_OFFLOAD should not default to m

[Geert Uytterhoeven]

The help text for the symbol controlling support for the NVM Express
over Fabrics TCP offload common layer suggests to not enable this
support when unsure.

Hence drop the "default m", which actually means "default y" if
CONFIG_MODULES is not enabled.

Fixes: f0e8cb6106da2703 ("nvme-tcp-offload: Add nvme-tcp-offload - NVMeTCP HW offload ULP")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
---
 drivers/nvme/host/Kconfig | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/nvme/host/Kconfig b/drivers/nvme/host/Kconfig
index 9c6f4d776daf14cf..f76cc4690bfc37bc 100644
--- a/drivers/nvme/host/Kconfig
+++ b/drivers/nvme/host/Kconfig
@@ -88,7 +88,6 @@ config NVME_TCP
 
 config NVME_TCP_OFFLOAD
 	tristate "NVM Express over Fabrics TCP offload common layer"
-	default m
 	depends on BLOCK
 	depends on INET
 	select NVME_CORE
-- 
2.25.1

Re: [PATCH] nvme: NVME_TCP_OFFLOAD should not default to m

[Christoph Hellwig]

On Tue, Jun 08, 2021 at 12:56:09PM +0200, Geert Uytterhoeven wrote:
> The help text for the symbol controlling support for the NVM Express
> over Fabrics TCP offload common layer suggests to not enable this
> support when unsure.
> 
> Hence drop the "default m", which actually means "default y" if
> CONFIG_MODULES is not enabled.
> 
> Fixes: f0e8cb6106da2703 ("nvme-tcp-offload: Add nvme-tcp-offload - NVMeTCP HW offload ULP")
> Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>

Err, where did this appear from?  This code has not been accepted into
the NVMe tree and is indeed not acceptable in this form.

Re: [PATCH] nvme: NVME_TCP_OFFLOAD should not default to m

[Geert Uytterhoeven]

Hi Christoph,

On Tue, Jun 8, 2021 at 2:19 PM Christoph Hellwig <hch@lst.de> wrote:
> On Tue, Jun 08, 2021 at 12:56:09PM +0200, Geert Uytterhoeven wrote:
> > The help text for the symbol controlling support for the NVM Express
> > over Fabrics TCP offload common layer suggests to not enable this
> > support when unsure.
> >
> > Hence drop the "default m", which actually means "default y" if
> > CONFIG_MODULES is not enabled.
> >
> > Fixes: f0e8cb6106da2703 ("nvme-tcp-offload: Add nvme-tcp-offload - NVMeTCP HW offload ULP")
> > Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
>
> Err, where did this appear from?  This code has not been accepted into
> the NVMe tree and is indeed not acceptable in this form.

It was applied to net-next.

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

please drop the nvme code from net-next

[Christoph Hellwig]

Hi Dave,

please drop the nvme-offload code from net-next.  Code for drivers/nvme/
needs ACKs from us nvme maintainers and for something this significant
also needs to go through the NVMe tree.  And this code is not ready yet.

Re: please drop the nvme code from net-next

[David Miller]

From: Christoph Hellwig <hch@lst.de>
Date: Tue, 8 Jun 2021 15:43:03 +0200

> please drop the nvme-offload code from net-next.  Code for drivers/nvme/
> needs ACKs from us nvme maintainers and for something this significant
> also needs to go through the NVMe tree.  And this code is not ready yet.

Please send me a revert, and I will apply it, thank you.

It's tricky because at least one driver uses the new interfaces.

Thank you.

Re: [PATCH] nvme: NVME_TCP_OFFLOAD should not default to m

[patchwork-bot+netdevbpf]

Hello:

This patch was applied to netdev/net-next.git (refs/heads/master):

On Tue,  8 Jun 2021 12:56:09 +0200 you wrote:
> The help text for the symbol controlling support for the NVM Express
> over Fabrics TCP offload common layer suggests to not enable this
> support when unsure.
> 
> Hence drop the "default m", which actually means "default y" if
> CONFIG_MODULES is not enabled.
> 
> [...]

Here is the summary with links:
  - nvme: NVME_TCP_OFFLOAD should not default to m
    https://git.kernel.org/netdev/net-next/c/762411542050

You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html

Re: please drop the nvme code from net-next

[Keith Busch]

On Tue, Jun 08, 2021 at 12:08:39PM -0700, David Miller wrote:
> From: Christoph Hellwig <hch@lst.de>
> Date: Tue, 8 Jun 2021 15:43:03 +0200
> 
> > please drop the nvme-offload code from net-next.  Code for drivers/nvme/
> > needs ACKs from us nvme maintainers and for something this significant
> > also needs to go through the NVMe tree.  And this code is not ready yet.
> 
> Please send me a revert, and I will apply it, thank you.
> 
> It's tricky because at least one driver uses the new interfaces.

Shouldn't whoever merged un-ACK'ed patches from a different subsystem
get to own the tricky revert?

RE: please drop the nvme code from net-next

[Shai Malin]

On Tue, Jun 08, 2021 at 10:51:00PM +0200, Shai Malin wrote:
> On Tue, Jun 08, 2021 at 10:41:00PM -0700, Keith Busch wrote:
> > On Tue, Jun 08, 2021 at 12:08:39PM -0700, David Miller wrote:
> > > From: Christoph Hellwig <hch@lst.de>
> > > Date: Tue, 8 Jun 2021 15:43:03 +0200
> > >
> > > > please drop the nvme-offload code from net-next.  Code for drivers/nvme/
> > > > needs ACKs from us nvme maintainers and for something this significant
> > > > also needs to go through the NVMe tree.  And this code is not ready yet.
> > >
> > > Please send me a revert, and I will apply it, thank you.
> > >
> > > It's tricky because at least one driver uses the new interfaces.
> >
> > Shouldn't whoever merged un-ACK'ed patches from a different subsystem
> > get to own the tricky revert?
> 
> Dave, we will provide a revert patch.

The revert patch was sent to net-next.
https://lore.kernel.org/netdev/20210609104918.10329-1-smalin@marvell.com/

Re: please drop the nvme code from net-next

[Shai Malin]

On Tue, Jun 08, 2021 at 10:41:00PM -0700, Keith Busch wrote:
> On Tue, Jun 08, 2021 at 12:08:39PM -0700, David Miller wrote:
> > From: Christoph Hellwig <hch@lst.de>
> > Date: Tue, 8 Jun 2021 15:43:03 +0200
> >
> > > please drop the nvme-offload code from net-next.  Code for drivers/nvme/
> > > needs ACKs from us nvme maintainers and for something this significant
> > > also needs to go through the NVMe tree.  And this code is not ready yet.
> >
> > Please send me a revert, and I will apply it, thank you.
> >
> > It's tricky because at least one driver uses the new interfaces.
> 
> Shouldn't whoever merged un-ACK'ed patches from a different subsystem
> get to own the tricky revert?

Dave, we will provide a revert patch.

RE: please drop the nvme code from net-next

[Shai Malin]

On Tue, Jun 8, 2021 at 4:38 PM Christoph Hellwig <hch@lst.de> wrote:
> Hi Dave,
> 
> please drop the nvme-offload code from net-next.  Code for drivers/nvme/
> needs ACKs from us nvme maintainers and for something this significant
> also needs to go through the NVMe tree.  And this code is not ready yet.

Hi all,

Sorry for any confusion we may have caused. Our plan was for the qed series 
to be considered for net-next, and for the nvme-tcp-offload series to be 
considered for nvme mailing list.
We attempted to communicate this in the cover letter and the addresses 
in to: vs. those in cc:, but perhaps this was not clear enough.

The plan was detailed in the cover latter under the "upstream plan" section 
https://lore.kernel.org/netdev/20210531225222.16992-1-smalin@marvell.com/
The series is structured in a modular way so that part 1 (nvme-tcp-offload) 
and part 2 (qed) are independent and part 3 (qedn) depends on both parts 1+2.

We have sent the first 2 parts which are independent:

- QED NVMeTCP Offload - https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/commit/?id=eda1bc65b0dc1b03006e427430ba23746ec44714
  This part includes the qed infrastructure which was discussed over the RFC.
  Our intent for this part was for it to be accepted to net-next, which it was.
  Dave, from our perspective this piece can stay in net-next.

- NVMeTCP Offload ULP - https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/commit/?id=5ff5622ea1f16d535f1be4e478e712ef48fe183b
  This is the nvme-tcp-offload ULP which we intended the NVMe tree,
  and it shouldn't be merged to net-next.
  Dave, please revert this from net-next until nvme maintainers are completely 
  satisfied with it.

Christoph, we would be more than happy to incorporate any feedback you may 
provide for any part of the series.