[meta-intel] [PATCH 1/1 daisy/dizzy] meta-crystalforest: Initial Intel QAT recipe

[Manoharan, Sundara RajaX P]

Hi All 

Since I am moving out  from the Yocto project (KTBR Team ) .
New person Suresh (sureshx.bellampalli at intel.com) will take care of that QAT-core recipe implementation in future and submit again.

Thanks & Regards
Sundar

-----Original Message-----
From: Darren Hart [mailto:dvhart at linux.intel.com] 
Sent: Wednesday, March 18, 2015 3:42 AM
To: Ong, Boon Leong; Manoharan, Sundara RajaX P; Saul Wold; 'paul.eggleton at linux.intel.com'
Cc: meta-intel at yoctoproject.org
Subject: Re: [meta-intel] [PATCH 1/1 daisy/dizzy] meta-crystalforest: Initial Intel QAT recipe

On 3/15/15, 8:04 AM, "Ong, Boon Leong" <boon.leong.ong at intel.com> wrote:

>>
>>Also, is there an effort underway to fix the QAT sources? All of this 
>>looks very fragile and I'm concerned about it's maintainability over 
>>time.
>>
>QAT technology is under-going refactoring whereby kernel driver going 
>into Linux mainline.
>and user-space into 01.org. There is good traction happening there.
>
>As for now, all we have now is QAT source package that is using 
>traditional tarball release with mux layer to switch between 
>coleto-creek and cave-creek chipset variant.
>The recipe
>is not applicable to romley or mohonpeak platform. That is why we are 
>targeting just meta-crystalforest layer which supports both variants.
>
>Ya, the recipe is long because this single tarball QAT Mux  contains 
>two tarballs for the above variants. That is why we split into software 
>packages according to platforms.
>
>It seems to me that maybe the recipe can be restructure into a single 
>inc file and two bb files\ for each of above variants?
>
>Is that reasonable ?


I think that would be a significant improvement.

--
Darren Hart
Intel Open Source Technology Center