[PATCH] AMCC Crypto4xx Device Driver v2]

[James Hsiao]

Hi Kim,

Thanks for comments.
Please see inline.

James
On Wed, 2008-10-29 at 10:54 -0500, Kim Phillips wrote:
> On Tue, 28 Oct 2008 19:18:20 -0700
> James Hsiao <jhsiao at amcc.com> wrote:
> 
> > On Tue, 2008-10-28 at 20:51 -0400, Josh Boyer wrote:
> > > - The question on ABLKCIPHER kconfig was ignored
> > Is ABLKCIPHER a sub set of BLKCIPHER? So, if BLKCIPHER is selected then
> > if ABLKCIPHER is present, it will use ABLKCIPHER otherwise using
> > BLKCIPHER algorithm? Correct?
> 
> this is my bad, commit 653ebd9c8510a7d647ed23e66e1338f848ebdbab
> "blkcipher: Merge ablkcipher and blkcipher into one option/module"
> renders CRYPTO_ABLKCIPHER obsolete in favour of CRYPTO_BLKCIPHER.
> 
Ok. 

> > > - Marking functions static
> > We have more than one file, that is why some of the function are not
> > static.
> 
> have one file then?

This is our first round of patch. We will support many more algorithms
and it will have much more files. I am afraid one file will be way too
big.

> 
> > > - Global lsec_core variable which doesn't allow for more than one
> > > device
> > We only support single incidence of device.
> 
> it's also a matter of unnecessarily polluting global namespace.
> 

I look into modify this.

> > > - Complete lack of locking code, how do you enforce mutually exclusive
> > > access to the device?
> > 
> > The crypto engine have couple bits 'command ready' and 'packet done',
> > which servers as semaphore here. So, software don't need extra locking.
> 
> know if it's applicable at all, but I imagine it's not SMP safe.

No, it's not SMP safe. This driver only support single core processors.
> 
> Kim