Re: sector alignment for DataOut PDUs?

[Sajay Selvaraj <sselvara@npd.hcltech.com>]

Hi,

  Now both the options look valid. How about mandating both of them ?

- target should  accept immediate data which is lesser than the
  negotiated value of unsolicited data length
- initiator should not send "less
  than the negotiated unsolicited data length or the total I/O length,
  whichever is smaller".

  I suppose most implementations would work this way even now.

/Sajay

Paul Koning wrote:
> 
> >>>>> "Sajay" == Sajay Selvaraj <sselvara@npd.hcltech.com> writes:
> 
>  Sajay> The sender should definitely avoid sending less data than the
>  Sajay> negotiated value of unsolicited data length when more data is
>  Sajay> available to be sent.
> 
>  Sajay> However, the total data length available to be sent itself
>  Sajay> could be lesser than the negotiated value. This suggests that
>  Sajay> while we should mandate the receiver to accept anything less
>  Sajay> than the negotiated value, we should not force the initiator
>  Sajay> to send exactly the same value as the negotiated value. In
>  Sajay> other words, we cannot "forbid the sender sending less than
>  Sajay> the negotiated value".  I'm for the first option suggested by
>  Sajay> Paul.
> 
> Sajay,
> 
> My comments were meant in the context of the case that the spec
> describes: " when the total data amount to be sent by a command is
> larger than the negotiated amount".  So I abbreviated a bit.  The more
> precise statement would be "one is to forbid the sender sending less
> than the negotiated unsolicited data length or the total I/O length,
> whichever is smaller".
> 
>           paul

-- 
http://san.hcltech.com