zfs code - May 2006 - Codereview: 6344836 ZFS should use SHA256 from crypto framework

I''d like a codereview of this please:

http://cr.grommit.com/~darrenm/6344836/

Unlike the previous fix that some of you may have seen this does not 
call on the crypto framework APIs but just shares the raw SHA256 
implementation.  This ensures that we use the same API user and kernel 
and avoid any extra "stuff" the crypto framework does (a problem in 
userland only really).

I''m not sure if there is any impact to ZFS boot from this change.

This change is also not suitable for backport to an S10 update as libmd 
(needed for userland) doesn''t exist there yet.  If this change is to be
backported libmd needs to be backported too.

Thanks.

--
Darren J Moffat

Darren J Moffat wrote:
>
> http://cr.grommit.com/~darrenm/6344836/
>
> I''m not sure if there is any impact to ZFS boot from this change.
>
ZFS boot has its own copy of SHA256.

Lin

Lin Ling wrote:
> Darren J Moffat wrote:
> 
>>
>> http://cr.grommit.com/~darrenm/6344836/
>>
>> I''m not sure if there is any impact to ZFS boot from this
change.
>>
> 
> ZFS boot has its own copy of SHA256.
Where in the ON source tree is that ?

Or are you referring to stuff that is run before the Solaris kernel 
starts running ?

--
Darren J Moffat

Darren J Moffat wrote:
>>
>> ZFS boot has its own copy of SHA256.
>
>
> Where in the ON source tree is that ?
>
This is not in the ON source tree yet :).
> Or are you referring to stuff that is run before the Solaris kernel 
> starts running ?
>
Yes.

Lin