Re: MSP430 in gcc4.9 ... enable interrupts?

[David Brown <david dot brown at hesbynett dot no>]

On 18/02/14 00:12, DJ Delorie wrote:
>> I presume these will be part of the headers for the library
>> distributed for msp430 gcc by TI/Redhat?
> 
> I can't speak for TI's or Red Hat's plans.  GNU's typical non-custom
> embedded runtime is newlib/libgloss, which usually doesn't have that
> much in the way of chip-specific headers or library functions.

Fair enough.  I don't know if Red Hat will be distributing anything
themselves, but I'm confident that TI will distribute chip-specific
headers with pre-built msp430 gcc packages (their aim, after all, is to
get more people to buy msp430 chips - and making the tools as easy and
powerful as possible is part of that).  I was just wondering where these
headers would fit in.

> 
>> is that for the "critical" attribute that exists in the old msp430
>> port (which disables interrupts for the duration of the function)?
> 
> Yes, for things like that.  They're documented under "Function
> Attributes" in the "Extensions to the C Language Family" chapter of
> the current GCC manual.
> 

Ah yes, I missed it when I first looked - the documentation makes
"critical" look like an option to the "interrupt" attribute, rather than
a stand-alone attribute.  It seems to me a rather strange idea to have
both "interrupt" and "critical" on the same function - an "interrupt"
function is inherently "critical" (and "reentrant") in that interrupts
are disabled before it enters, and restored or re-enabled on exit.  It
would make a difference on processors like ARMs with several interrupt
levels, but not on an msp430 with its single level.

David