[Tinyos-devel] TEP116 HIL/HAL discussion

[Janos Sallai]

>> Why do you want to specify from  
>> above what the HAL for the radio components should look like and  
>> even what names they should have?
>
>To make platform definition simpler.

Does this imply then that TEP 105, as well as the time stamping/sync
TEPs, should prescribe that the chip specific XActiveMessageC components
MUST provide the respective interfaces?

Another question: Where (by which component) should the respective HIL
interfaces provided? If the chip specific XActiveMessageC components are
not HIL but HAL, then probably somewhere else. Should we require that
the platform's ActiveMessageC MUST provide LowPowerListening (in TEP
105), PacketTimeStamp (timestamping TEP), TimeSyncAMSend and
TimeSyncPacket (in the packet time sync TEP)? This would override TEP
116's statement on the required signature of ActiveMessageC.

Janos

-----Original Message-----
From: tinyos-devel-bounces at millennium.berkeley.edu
[mailto:tinyos-devel-bounces at millennium.berkeley.edu] On Behalf Of
Philip Levis
Sent: Thursday, May 15, 2008 1:35 PM
To: Vlado Handziski
Cc: TinyOS Development
Subject: Re: [Tinyos-devel] TEP116 HIL/HAL discussion


On May 15, 2008, at 11:27 AM, Vlado Handziski wrote:

> The additional interfaces in the component are irrelevant for the  
> discussion. Section 5 in the TEP talks about the same interfaces as  
> in ActiveMessageC (btw. there seems to be two sections numbered 5 in  
> the document). What I am saying is that the requirements listed in  
> Section 5 are implementation detail for the HIL and have nothing to  
> do with the HAL level interface. Why do you want to specify from  
> above what the HAL for the radio components should look like and  
> even what names they should have?

To make platform definition simpler.

> The HAL interface can be completely different for each chip/platform  
> and not contain any of the interfaces exported in ActiveMessageC. It  
> can happen that a particular chip/platform HAL is so well aligned  
> with the current HIL that they share those interfaces making the  
> component be both HIL/HAL implementation, but in general this is not  
> a requirement.

But this has come up before. There can be more than one HAL interface.  
TEP 116 states that for a radio chip implementation to be compliant  
with 116, it MUST provide an AM abstraction.

If CC2420ActiveMessageC is not a HAL, and it is also not a HIL, what  
is it?

Phil
_______________________________________________
Tinyos-devel mailing list
Tinyos-devel at millennium.berkeley.edu
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-deve
l