[RFC PATCH 3/3] Enable BCM2835 mailbox support

[Stephen Warren]

On 07/02/2014 02:51 AM, Craig McGeachie wrote:
> On 01/07/14 04:06, Lubomir Rintel wrote:
>> I'm wondering if there's a reason you're not getting the channel with
>> ipc_request_channel() (that would be mbox_request_channel() in current
>> version of the framework) at the probe time, but each time you send a
>> message with do_message()?
> 
> It's a good question.  Given that it was 9 months ago, I can't be sure.
>  Still, I think that it was because of the nature of the property
> channel for the bcm2835 mailbox, which is used for a variety of
> purposes.  The two uses that I had in mind were, thermal (get
> temperatures) and DMA (query the DMA channels available for use by the
> ARM core).  I think there are also video buffer messages on the property
> channel as well ,but those might be on a separate channel.  Regardless,
> there are a lot of things done on the property channel and if one driver
> holds onto that channel, other drivers will be hard to get working
> correctly.

I don't recall any details, but surely any mbox driver should expose a
separate channel for each type of communication, and then mux them all
into the single HW mbox mechanism. Then, e.g. the DMA driver or display
driver could exclusively acquire their own channel, but without
conflicting with any drivers using any other channel.