On Thu, Dec 9, 2010 at 1:40 PM, Stephen Dawson-Haggerty <
[email protected]> wrote:

> +devel
>
> On Thu, Dec 9, 2010 at 12:23 PM, Janos Sallai
> <[email protected]> wrote:
> > We don't have to worry about platform specific word alignment
> > peculiarities as long as we stick to nx_structs and nx_unions. The
> > cc2420x stack I wrote using rfxlink works on the both the msp430 and
> > the atmega128 with the same message_t definition.
>
> Yep, and the metadata is nx_ structs now since they're part of the
> message_t.  It is generally better to avoid these for structures not
> actually sent over the radio since using them results in additional
> code overhead...
>

Does the metadata really need to be nx_structs.   I vagely remember playing
with this at one point and thought it worked to mix and it generates much
better
code.  Worth trying again.


>
> > Regarding Packet.clear(): would it make sense for the new message_t to
> > have a dirty flag in message_t which wouldn't allow the message buffer
> > to be reused without calling Packet.clear() first? Sure, checking it
> > would have a runtime overhead, but we could just have a define to turn
> > that off for production builds.
>
> Hmm, I don't know -- this seems likely to break existing code and also
> will be confusing for new users.  Maybe if we were starting over, but
> it seems like one of those things in the "if it ain't broke, don't fix
> it" at the moment...  Not changing it when it was already zero like
> Miklos proposes seems likely to fix the edge case now, which is that
> when people turn on packet link, things work differently.  Currently,
> the message_t's are cleared by crt0, so as long as we leave it zero we
> shouldn't break much.
>
> I'm fine with this simple patch.
>
> Steve
>
_______________________________________________
Tinyos-help mailing list
[email protected]
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help

Reply via email to