-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello,

On Feb 21, 2007, at 13:12, Bucky Katz wrote:

> Michael Lorenz <macallan@netbsd.org> writes:
>
>> Hello,
>>
>> On Feb 21, 2007, at 09:46, Toru Nishimura wrote:
>>
>>> NetBSD/arm, in general, does smart things with "ipending" global
>>> variable to implement NetBSD-peculiar (other OS camp would call it
>>> strange) spl(9) priority level scheme.  While this approach fits
>>> with 32 way interrupt source, it is rather painful to adapt 64bit
>>> source.
>>
>> Depends wether you really have >30 distinct interrupt sources in
>> use.  Macppc for instance can have 32, 64 or 128 IRQs ( some
>> powerbooks have two Heathrow PICs with 64 IRQ lines each ) but only
>> a fraction is actually connected to anything. So what we do there is
>> to map those interrupts actually in use into a uint32_t. Not that I
>> think this is a particularly nice approach but it does the job and
>> it's fairly painless.
>
> It's not unusual on ARM embedded systems to have > 32 interrupts, as
> GPIOs can be configured to generate interrupts and hardware guys love
> to use GPIOs for signalling.

One IRQ per GPIO? Nice to have, those I've seen so far would group 
GPIOs together ( as in one IRQ per group of pins ) or just pull the 
same IRQ line whenever any pin changes.
Anyway, the interrupt handling is port-specific so you're free to 
change/extend the ipending mechanism if 32bit really isn't enough.

have fun
Michael
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Darwin)

iQEVAwUBRdyUocpnzkX8Yg2nAQIdUwgAqwukoU6lO0GP6/PL61DXUD4A3VbiXNBd
/nzSpx1HVRlwqL4uBRfb9BUZ7Ee2I+1FbYgrWUHKOqUs85jqjOGo1BLjrsGu6Jcs
Bu/pgeypkjT/AECDavECvsdgThFh/s5E7k4hr0Xok25Gwx/jvDSctAiqNr4U+AFj
FJvtFlT8ppBk/Jg676z4vw7bclS1pYC6sfGgvQAyGH4FyZlAl2Z0NSg15ztncwpR
Z4G0Y0HrgA5WJ4u94E2UOx9xV7BIBtoB5jTtCoRcxaiglZ28yGIKjV2pj7JB6lUS
PbKVN3xD8zmeXb1vGcS97/0/et2LyD6PcbZ9oJCq81dNG5VRC0kHYw==
=C7cX
-----END PGP SIGNATURE-----