SBC-3 comment for Next T10 meeting

Gerry Houlder gerry.houlder at seagate.com
Wed Apr 24 13:19:50 PDT 2013


Formatted message: <a href="http://www.t10.org/cgi-bin/ac.pl?t=r&f=r1304241_f.htm">HTML-formatted message</a>

Yes, i agree that Fred's proposed wording retains the original meaning of
this sentence, while the change made in the last CAP meeting does not.
On Wed, Apr 24, 2013 at 1:42 PM, Knight, Frederick <
Frederick.Knight at netapp.com> wrote:
>  ** **
>
> During LB comment resolution in Vancouver, it appears that the following
> change was made:****
>
> ** **
>
> *4.15 Caches*
>
> <…>****
>
> The power, if any, that allows a non-volatile cache to remain non-volatile
> may become low enough to prevent****
>
> the non-volatile cache from remaining non-becoming volatile for a vendor
> specific minimum time (e.g., the****
>
> battery voltage becomes too low too sustain cache contents beyond a vendor
> specific time). If this occurs and****
>
> ** **
>
> Old text: … prevent the NV cache from remaining non-volatile .. [ means
> allowing it to become volatile ]****
>
> New text: … prevent the NV cache from becoming volatile … [ means
> preventing it from becoming volatile ]****
>
> ** **
>
> The original text describes low power preventing the cache from staying
> non-volatile [i.e., allowing the cache to become volatile].  NOW, the new
> text describes that low power will prevent the cache from becoming
volatile.
> ****
>
> ** **
>
> The new text is not correct.	The text should be restored or changed to
> something like:****
>
> ** **
>
> The power, if any, that allows a non-volatile cache to remain non-volatile
> may become low enough to prevent****
>
> the non-volatile cache from remaining non-becoming allow the non-volatile
> cache to become volatile for a vendor specific minimum time (e.g., the
> battery voltage becomes too low too sustain cache contents beyond a
> vendor specific time).****
>
> ** **
>
> Since CAP made this change in the first place, the wording to fix it needs
> to go before CAP too.****
>
> ** **
>
>		  Fred****
>
> ** **
>



More information about the T10 mailing list