Initial Linux-wireless contribution

[Eugene Krasnikov]

Hi Kalle,

My understanding is that wcnss_core IS that glue layer between
platform and wcn36xx. So my idea was to move "struct ieee80211_hw
*private_hw" somehow to the wcnss_core? Or at least store it there.
What do you think about that idea?


2013/8/4 Kalle Valo <kvalo at qca.qualcomm.com>:
> Eugene Krasnikov <k.eugene.e at gmail.com> writes:
>
>> wcn36xx is getting better and better every day and now we are
>> approaching the stage when we can make a first push to linux wireless.
>> What else do you think must be added before initial push? Please
>> review the code and come back if something crucial is missing.
>
> I did a quick review now. I think most important is to come up with a
> platform struct for abstracting out all msm out-of-tree code which is
> not in upstream.
>
> I'm thinking that we should split the driver into two modules:
> wcn36xx.ko (the actual driver) and wcn36xx-msm.ko (a small glue layer
> between wcn36xx and msm core). The latter will contain all msm specific
> code and the two modules would communicate through the platform struct.
> And the code for wcn36xx.ko we can send upstream but wcn36xx-msm.ko need
> to stay as an out-of-tree module.
>
> This way we wcn36xx would be acceptable in upstream.
>
> --
> Kalle Valo



-- 
Best regards,
Eugene