From mboxrd@z Thu Jan  1 00:00:00 1970
From: willy@linux.intel.com (Matthew Wilcox)
Date: Wed, 18 Jun 2014 12:25:47 -0400
Subject: [PATCH] NVMe: Fail SCSI->NVMe translation for UNMAP when anchor
 is set
In-Reply-To: <B8DB29EA17EB2742AC3405C31A9B9EE9614C0D68@ORSMSX111.amr.corp.intel.com>
References: <1403082188-4852-1-git-send-email-daniel.mcleran@intel.com>
 <alpine.LRH.2.03.1406180908390.4699@AMR>
 <20140618152708.GH12025@linux.intel.com>
 <alpine.WNT.2.00.1406180929430.12984@DMCLERAN-DESK.amr.corp.intel.com>
 <20140618154532.GI12025@linux.intel.com>
 <B8DB29EA17EB2742AC3405C31A9B9EE9614C0D68@ORSMSX111.amr.corp.intel.com>
Message-ID: <20140618162547.GJ12025@linux.intel.com>

On Wed, Jun 18, 2014@03:51:52PM +0000, Mayes, Barrett N wrote:
> > On the one hand, it says that ANC_SUP shall be set to 0.  
> 
> LSI proposed the following and asked for feedback by June 9.  I don't recall seeing any comments to this on the public reflector.
> 
> ANC_SUC
> Shall be set to 0, to indicate that setting the ANCHOR bit in UNMAP is not supported if the namespace is not resource or thin provisioned.
> Shall be set to 1, to indicate that setting the ANCHOR bit in UNMAP is supported if the namespace is resource provisioned. May be set to 1 if the namespace is thin provisioned.

Thanks!  Just found that unread email ;-)

I'm not sure about Guy's reasoning that NVMe supports three different
provisioning models like SCSI does.  Should probably take this to the
NVMe mailing list now ...