Western Digital issues with 'operations' definitions in SBC-3 r35j

Ralph Weber Ralph.Weber at wdc.com
Mon Oct 14 12:28:08 PDT 2013


Formatted message: <a href="http://www.t10.org/cgi-bin/ac.pl?t=r&f=r1310148_f.htm">HTML-formatted message</a>

Since I want to deprive Kevin Marks of the need to look for r1 excuses, a new
revision will be posted ASAP.
I am pleased that you approve of the basic logic of the proposal. It is a
reassuring sign that I have not screwed up yet another time.
All the best,
.Ralph
________________________________
From: Penokie, George [George.Penokie at lsi.com]
Sent: Monday, October 14, 2013 2:07 PM
To: Ralph Weber; t10 at t10.org
Subject: RE: Western Digital issues with 'operations' definitions in SBC-3
r35j
Ralph,
I did a real quick look at this. My only comment, based on this very quick
look, is that the sections titled Read-ahead caches and Write-back caches
should be changed to Read-ahead caching and Write-back caching. This is
because read-ahead and write-back are not types of caches but rather ways in
which caches may be used.
Bye for now,
George Penokie
LSI Corporation
3033 41 St NW
Rochester , MN 55901
507-328-9017
george.penokie at lsi.com
From: owner-t10 at t10.org [mailto:owner-t10 at t10.org] On Behalf Of Ralph Weber
Sent: Thursday, October 10, 2013 4:29 PM
To: t10 at t10.org
Subject: Western Digital issues with 'operations' definitions in SBC-3 r35j
I wish to begin this note with a Thank You to the numerous contributors who
commented upon the previous WD postings on the confusion raised by the
caching changes SBC-3 r35j.
A great deal of floundering about was required to gain true insight regarding
the problems to be resolved, which have proven to be:
  *   an over use of the glossary as coordination point for the true meaning
and usage of "operations", specifically read operations, unmap operations,
verify operations, and write operations;
  *   fuzzy glossary logic promoted as a necessary evil to avoid writing
"read operations and verify operations" in hundreds of places (necessary,
yes, evil, not so much ... as long as enough words, including a 'shall', can
be thrown at the problem, all of which is not the province of the glossary);
and
  *   uncertainty about how completely the SYNCHRONIZE CACHE aspects of the
VERIFY command had been rendered in the new text.
A proposal that includes detailed text changes based the fruits of the
tedious investigations of these issues is now available:
http://www.t10.org/cgi-bin/ac.pl?t=d&f=13-257r0.pdf
The changes in this document are at once more wordy and less technical than
anything previously proposed. Reviewing it should provide plenty of
"interesting times".
It goes without saying that WD sees some form of resolution to the issues
described in 13-257r0 as a necessary precursor to publication of SBC-3.
All the best,
.Ralph



More information about the T10 mailing list