Lpc2000

Guille,

Thanks for the observation: it's probably close enough to the mark 
OK. 

I wouldn't quite say I don't think spurious interrupts aren't a 
problem, and it wouldn't be better to avoid them: clearly the answer 
to this is "yes". I think Jaya's work is helpful in documenting how 
this might be done.

However, given that the source of the interrupts is well documented, 
and the fact that they're extremely rare (from memory we saw maybe 
one in 100 million interrupts as spurious), providing an empty 
handler for them seemed a reasonable approach. The handler gets 
called when a spurious interrupt happens (i.e. VIC fails to identify 
the source). If the source of the interrupt is still valid, the 
normal interrupt handler for the event will get called subsequently.

What I'm finding somewhat frustrating is that I'm been told that this 
approach is somehow flawed, without being told why or how.

I also have to say I find it somewhat difficult in framing my points 
without continually provoking an adversarial or abusive reaction from 
Jaya.

Thanks for your attempt to calm the waters, in any case.

Regards
Brendan


--- In lpc2000@yahoogroups.com, "Guillermo Prandi" 
<yahoo.messenger@...> wrote:
>
> I'd like to point out that apparently (correct me if I'm wrong!) 
Jaya 
> is radically against system that tolerate any interrupts whose 
source 
> is unidentifiable. Apparently his products must pass some 
> certification process that rejects this situation. Brendan's 
proposal 
> is a solution to the problem "how to get a system running using 
UART 
> interrupts", and what Jaya is proposing is a solution to the 
> problem "how to get a system running so you never get an interrupt 
> whose source you can't identify". It is clear that Brendan doesn't 
> think that interrupts from unidentifiable sources are not that much 
> of a problem. Hence the misunderstanding (unless I just added a new 
> degree of misunderstanding; if that is case, my apologies).
> 
> Guille
>