[dm-devel] [PATCH v4 3/3] nvme: decouple basic ANA log page re-read support from native multipathing

[Mike Snitzer]

On Tue, Apr 20 2021 at  5:34am -0400,
Christoph Hellwig <hch at lst.de> wrote:

> On Fri, Apr 16, 2021 at 07:53:29PM -0400, Mike Snitzer wrote:
> > Whether or not ANA is present is a choice of the target implementation;
> > the host (and whether it supports multipathing) has _zero_ influence on
> > this. If the target declares a path as 'inaccessible' the path _is_
> > inaccessible to the host. As such, ANA support should be functional
> > even if native multipathing is not.

As you well know, ANA is decoupled from multipathing in the NVMe spec.
This fix illustrates that the existing Linux NVMe ANA handling is too
tightly coupled with native NVMe multipathing. Unfortunately, you've
forced this as a means to impose your political position:

> NAK.  nvme-multipathing is the only supported option for subsystems with
> multiple controllers.

And while this is largely irrelevant to the technical review of my fix:
native nvme-multipath may be the only supported option in your world. In
mine that isn't true and never has been.

Your political posturing doesn't replace technical justification. You
don't have a compelling technical case to take the stance you do, yet
you guard it like you do. Just a really stark inconsistency in your
technical leadership that is repeatedly left unchecked by others.

Alas.