[security-discuss] Re: Design review of context sharing for multi-part requests

[Ferenc Rakoczi]

Krishna,

 the synchronisation among parts is more than just a
software provider issue. On the PKCS#11 level it
is not a problem at all, C_EncryptUpdate() for part n
will not be called before part n-1 finished. On the
lower level, on the other hand, it is a problem even with
a single provider in asynchronous mode. So whoever
dispatches the jobs for the parts has to make sure that
the context is updated before the provider starts
on the next part.

Ferenc


----- Original Message -----
From: Krishna Yenduri <Bhargava.Yenduri at sun.com>
Date: Wednesday, December 6, 2006 11:45 pm
Subject: Design review of context sharing for multi-part requests
To: security-discuss at opensolaris.org

> 
> Please review the attached design document for the RFE -
> 
> 6494834 support check for threshold when using hardware providers even
>                 for multi-part requests
> 
> This RFE allows the sharing of an operation context between
> a Solaris crypto framework hardware provider and a software provider
> for improved performance.
> 
> I would appreciate any review comments by next Tuesday, 12/12/06.
> 
> Thanks,
> -Krishna
>