Re: [storm] I-D Action: draft-ietf-storm-rdmap-ext-01.txt

Ok, agreed the document needs to make such a statement, and "natural alignment" is not sufficient especially when considered end-to-end.

Tom.

-----Original Message-----
From: david.black@emc.com [mailto:david.black@emc.com] 
Sent: Tuesday, September 13, 2011 10:50 AM
To: Tom Talpey; storm@ietf.org
Subject: RE: [storm] I-D Action: draft-ietf-storm-rdmap-ext-01.txt

Restating my concern:

If an alignment violation causes an error in this protocol, then this protocol must specify how to avoid causing an alignment violation.

Thanks,
--David

> -----Original Message-----
> From: Tom Talpey [mailto:ttalpey@microsoft.com]
> Sent: Tuesday, September 13, 2011 8:22 AM
> To: Black, David; storm@ietf.org
> Subject: RE: [storm] I-D Action: draft-ietf-storm-rdmap-ext-01.txt
> 
> Actually, this question goes deeper - alignment requirements such as 
> this are actually on the *application*. Are we requiring that they 
> align their atomics? Is 64-bits always enough to ensure correctness? 
> Also remember the peer performs no processing, it simply issues an operation to an application-provided value (the atomic's RDMA triplet).
> 
> Normally such a requirement would be specified in the Verbs. Since 
> local behavior is not in scope here, perhaps the document should make 
> no requirement at all, and simply give some advice. The mechanism to avoid it is necessarily at the upper layer.
> 
> Tom.
> 
> -----Original Message-----
> From: storm-bounces@ietf.org [mailto:storm-bounces@ietf.org] On Behalf 
> Of david.black@emc.com
> Sent: Monday, September 12, 2011 2:12 PM
> To: storm@ietf.org
> Subject: Re: [storm] I-D Action: draft-ietf-storm-rdmap-ext-01.txt
> 
> In looking at the changes in this draft from the -00 version, I 
> noticed something that needs attention, although it was already present in the -00 version:
> 
> The term "naturally aligned" needs to be defined for buffer addresses, 
> because failure to use a "naturally aligned" buffer address causes a 
> "Catastrophic Error" at the recipient.  For this reason, there needs to be a tight specification of what the sender needs to do to avoid that error.
> Specifying a boundary that is always "naturally aligned" (e.g., 
> 64-bit) should suffice (and is probably superior to providing a way to figure this out on-the-wire).
> 
> Thanks,
> --David
> 
> > -----Original Message-----
> > From: storm-bounces@ietf.org [mailto:storm-bounces@ietf.org] On 
> > Behalf Of internet-drafts@ietf.org
> > Sent: Monday, July 11, 2011 4:17 PM
> > To: i-d-announce@ietf.org
> > Cc: storm@ietf.org
> > Subject: [storm] I-D Action: draft-ietf-storm-rdmap-ext-01.txt
> >
> > A New Internet-Draft is available from the on-line Internet-Drafts 
> > directories. This draft is a work item of the STORage Maintenance Working Group of the IETF.
> >
> > 	Title           : RDMA Protocol Extensions
> > 	Author(s)       : Hemal Shah
> >                           Felix Marti
> >                           Wael Noureddine
> >                           Asgeir Eiriksson
> >                           Robert Sharp
> > 	Filename        : draft-ietf-storm-rdmap-ext-01.txt
> > 	Pages           : 30
> > 	Date            : 2011-07-11
> >
> >    This document specifies extensions to the IETF Remote Direct Memory
> >    Access Protocol (RDMAP [RFC5040]). RDMAP provides read and write
> >    services directly to applications and enables data to be transferred
> >    directly into Upper Layer Protocol (ULP) Buffers without
> >    intermediate data copies. The extensions specified in this document
> >    provide the following capabilities and/or improvements: Atomic
> >    Operations and Immediate Data.
> >
> >
> >
> >
> > A URL for this Internet-Draft is:
> > http://www.ietf.org/internet-drafts/draft-ietf-storm-rdmap-ext-01.tx
> > t
> >
> > Internet-Drafts are also available by anonymous FTP at:
> > ftp://ftp.ietf.org/internet-drafts/
> >
> > This Internet-Draft can be retrieved at:
> > ftp://ftp.ietf.org/internet-drafts/draft-ietf-storm-rdmap-ext-01.txt
> > _______________________________________________
> > storm mailing list
> > storm@ietf.org
> > https://www.ietf.org/mailman/listinfo/storm
> 
> _______________________________________________
> storm mailing list
> storm@ietf.org
> https://www.ietf.org/mailman/listinfo/storm
> 