Re: [PATCH blktests v5 5/5] nvme/055: add test for nvme-tcp zero-copy offload

[Shinichiro Kawasaki <shinichiro.kawasaki@wdc.com>]

On Jan 28, 2026 / 15:28, John Garry wrote:
> On 06/12/2024 13:51, Aurelien Aptel wrote:
> > +++ b/tests/nvme/055
> > @@ -0,0 +1,268 @@
> > +#!/bin/bash
> > +# SPDX-License-Identifier: GPL-3.0+
> > +# Copyright (C) 2024 Aurelien Aptel<aaptel@nvidia.com>
> > +#
> > +# zero-copy offload
> > +
> > +. tests/nvme/rc
> > +
> > +DESCRIPTION="enable zero copy offload and run rw traffic"
> > +TIMED=1
> > +
> > +iface_idx=""
> > +
> > +# these vars get updated after each call to connect_run_disconnect()
> > +nb_packets=0
> > +nb_bytes=0
> > +nb_offload_packets=0
> > +nb_offload_bytes=0
> > +offload_bytes_ratio=0
> > +offload_packets_ratio=0
> > +
> > +requires() {
> > +	_nvme_requires
> > +	_require_remote_nvme_target
> > +	_require_nvme_trtype tcp
> > +	_have_kernel_option ULP_DDP
> 
> I notice that this kernel config option is still not included in mainline,
> over 1 year later.
> 
> It is strange to me to merge tests for features which are yet to be
> accepted.

When this test case was added, I was aware of that the DDP offload feature was
not yet accepted [1]. At that time, I expected that the feature would be applied
soon.

This test case was the first use case of the blktests feature
"NVME_TARGET_CONTROL" which allows setting up nvme targets for testing in
flexible manner [2]. The test case was useful to demonstrate the value of the
feature. That's why I merged the test for "features which are yet to be
accepted".

Said that, assuming the DDP offload feature is not going to be mainlined, this
test case has lost its value. I think it is reasonable to remove the test case
at this moment. A patch post for removal will be helpful to confirm consensus.

[1] https://lore.kernel.org/all/2537c8iv1kq.fsf@nvidia.com/
[2] https://lore.kernel.org/linux-block/20241206135120.5141-2-aaptel@nvidia.com/