[PATCH] nvme-core: update NS Attr Changed AEN handling for ANA group

[Knight, Frederick]

Here are the authors of the original TP:

Technical Proposal Author(s)
Name	Company
Fred Knight	NetApp
David Black	Dell EMC
Curtis Ballard	HPE
Christoph Hellwig	WDC

Myself, David Black, Curtis Ballard and the rest of the weekly FMDS meeting all agreed that an AEN is prohibited during this circumstance ("due to the creation of a namespace").  All three agree that when ANA groups are created, they already have a state and there is NO transition.  They all agree if you want to change that, you need to bring a TPAR to change it.

The fourth author, Christoph, also stated during the TP development that he did NOT want 2 AENs when a namespace was created.  He specifically required there be only 1.

	Fred

-----Original Message-----
From: Keith Busch <kbusch at kernel.org> 
Sent: Tuesday, December 8, 2020 4:47 PM
To: Knight, Frederick <Frederick.Knight at netapp.com>
Cc: hch at lst.de; Meneghini, John <John.Meneghini at netapp.com>; George, Martin <Martin.George at netapp.com>; linux-nvme at lists.infradead.org; sagi at grimberg.me; Hannes Reinecke <hare at suse.com>
Subject: Re: [PATCH] nvme-core: update NS Attr Changed AEN handling for ANA group

NetApp Security WARNING: This is an external email. Do not click links or open attachments unless you recognize the sender and know the content is safe.




On Tue, Dec 08, 2020 at 07:13:00PM +0000, Knight, Frederick wrote:
> Setting up a new ANA Group does NOT cause a transition from <UNKNOWN state> to <some known state>.  The establishment of initial state is NOT a change of state as defined in section 8.20.3.6.
>
> An ANA AEN is prohibited if the ANA change is due to the creation of a namespace.

Yes, the AEN is prohibited if the ANA change is due to the creation of a namespace. In this scenario, though, the ANA change is due to a group entering a state, and you're required to notify that as long as ANA events are enabled. There is no text in section 8.20.3.6 or anywhere else that I can find that says creating a group with a valid state disqualifies it as "entering" that state. However, I do find that section 8.20.3.5 says transitions may occur from states not visible to the host, so <UNKNOWN state> to <known state> is a transition according to the spec.