Re: [ECOS] context for creating interrupts on edb7312

[Gary Thomas <gary@mlbassoc.com>]

On Wed, 2003-11-19 at 09:34, Aaron Case wrote:
> Hello,
> 
> My question pertains to the context for which an interrupt would be created
> and attached.
> 
> The Massa text shows an example where an interrupt is created, attached, and
> unmasked in cyg_user_start(), see code listing 3.2. Yet 3 pages later, see
> Item list 3.7, it gives thread(and not init as in the code listing) as the
> only viable context for creating a interrupt(via cyg_interrupt_create).
> This seems to be in conflict.
> 
> Intuitively, creation of interrupts would be in cyg_user_start, init
> context, like the code listing example BUT, I was never able to get this to
> work. I have however been able to get work in the thread context, which
> agrees with the item list but seems counter intuitive.
> 
> The eCos reference manual confirms my intuition by stating the following in
> the valid context section of the ref manual...
> 
> "In a typical configuration interrupt handlers are created and attached
> during system initialization, and never detached or deleted. However it is
> possible to perform these operations at thread level, if desired..."
> 
> I, however, have not be able to get interrupts to work in the init context,
> only in the thread context.
> 
> ------Is creating interrupts within the init context possible?-------
> 

What do you mean "creating" interrupts?  Do you mean actually fielding 
them (when they occur)?  In that case, no, since interrupts are not
turned on (enabled) until you call the scheduler.

> I am using the EDB7312 processor with the arm720t core.
> 
> I would prefer to locate all interrupt creation and configuration in
> cyg_user_start, the init context, but have had little success in doing so.
> 

Can you explain what you mean by this?

> I would also like to add, Im not criticizing the online docs, they appear
> correct, just a bit generic(so as to apply to all platforms, I assume).

As is should be (IMO) since the notions of interrupts and handlers are
platform independent (while what causes the interrupts, which ones are
hooked to which devices, etc, are target specific)

-- 
Gary Thomas <gary@mlbassoc.com>
MLB Associates


-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss