From mboxrd@z Thu Jan  1 00:00:00 1970
From: keith.busch@intel.com (Keith Busch)
Date: Thu, 8 Mar 2018 15:45:58 -0700
Subject: [PATCH] Revert "nvme: create 'slaves' and 'holders' entries for
 hidden controllers"
In-Reply-To: <20180307150344.GE30975@localhost.localdomain>
References: <20180307131358.20275-1-hch@lst.de>
 <1c1332d4-1d92-af8f-974c-2eafcde741f2@suse.de>
 <20180307150344.GE30975@localhost.localdomain>
Message-ID: <20180308224558.GA14765@localhost.localdomain>

On Wed, Mar 07, 2018@08:03:45AM -0700, Keith Busch wrote:
> On Wed, Mar 07, 2018@02:40:39PM +0100, Hannes Reinecke wrote:
> > How do we detect the topology now?
> > Does nvme cli has the functionality?
> 
> Not in the way you'd probably like. It can provide all the information
> you'd need to put it together, but it doesn't provide a convenient
> method to get or visualize it. I'll add a note on the project to fix
> that.

Actually, nvmecli already has something kind of close, thanks
to Johannes! The 'list-subsys' command is most of the way there. It
currently looks like this, and just needs to go one level deeper to
append the namespaces to the output:

# nvme list-subsys -o json
{
  "Subsystems" : [
    {
      "Name" : "nvme-subsys0",
      "NQN" : "nqn.2014.08.org.nvmexpress:8086108ePHLE7200015N6P4B7335943:ICDPC5ED2ORA6.4T"
    },
    {
      "Paths" : [
        {
          "Name" : "nvme0",
          "Transport" : "pcie",
          "Address" : "0000:03:00.0"
        },
        {
          "Name" : "nvme1",
          "Transport" : "pcie",
          "Address" : "0000:04:00.0"
        }
      ]
    }
  ]
}

While this should readily work for fabrics, I needed to add a new kernel
driver patch to export the PCIe address for the above (patch staged
for 4.17).

From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <stable-owner@vger.kernel.org>
Received: from mga14.intel.com ([192.55.52.115]:29848 "EHLO mga14.intel.com"
        rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
        id S1750713AbeCHWoL (ORCPT <rfc822;stable@vger.kernel.org>);
        Thu, 8 Mar 2018 17:44:11 -0500
Date: Thu, 8 Mar 2018 15:45:58 -0700
From: Keith Busch <keith.busch@intel.com>
To: Hannes Reinecke <hare@suse.de>
Cc: axboe@kernel.dk, sagi@grimberg.me, bharat@chelsio.com,
        linux-nvme@lists.infradead.org, kzak@redhat.com,
        stable@vger.kernel.org, Christoph Hellwig <hch@lst.de>,
        sitsofe@gmail.com
Subject: Re: [PATCH] Revert "nvme: create 'slaves' and 'holders' entries for
 hidden controllers"
Message-ID: <20180308224558.GA14765@localhost.localdomain>
References: <20180307131358.20275-1-hch@lst.de>
 <1c1332d4-1d92-af8f-974c-2eafcde741f2@suse.de>
 <20180307150344.GE30975@localhost.localdomain>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20180307150344.GE30975@localhost.localdomain>
Sender: stable-owner@vger.kernel.org
List-ID: <stable.vger.kernel.org>

On Wed, Mar 07, 2018 at 08:03:45AM -0700, Keith Busch wrote:
> On Wed, Mar 07, 2018 at 02:40:39PM +0100, Hannes Reinecke wrote:
> > How do we detect the topology now?
> > Does nvme cli has the functionality?
> 
> Not in the way you'd probably like. It can provide all the information
> you'd need to put it together, but it doesn't provide a convenient
> method to get or visualize it. I'll add a note on the project to fix
> that.

Actually, nvmecli already has something kind of close, thanks
to Johannes! The 'list-subsys' command is most of the way there. It
currently looks like this, and just needs to go one level deeper to
append the namespaces to the output:

# nvme list-subsys -o json
{
  "Subsystems" : [
    {
      "Name" : "nvme-subsys0",
      "NQN" : "nqn.2014.08.org.nvmexpress:8086108ePHLE7200015N6P4B7335943:ICDPC5ED2ORA6.4T"
    },
    {
      "Paths" : [
        {
          "Name" : "nvme0",
          "Transport" : "pcie",
          "Address" : "0000:03:00.0"
        },
        {
          "Name" : "nvme1",
          "Transport" : "pcie",
          "Address" : "0000:04:00.0"
        }
      ]
    }
  ]
}

While this should readily work for fabrics, I needed to add a new kernel
driver patch to export the PCIe address for the above (patch staged
for 4.17).