question: i/o port 0x61 on x86 archs

question: i/o port 0x61 on x86 archs

[Ravikiran G Thirumalai]

I have a question regdg do_nmi routine in x86; what does location 0x61 
from inb(0x61) do? Is it something very well known in intel archs? 

Kiran

Re: question: i/o port 0x61 on x86 archs

[Randy.Dunlap]

On Sat, 8 Jun 2002, Ravikiran G Thirumalai wrote:

| I have a question regdg do_nmi routine in x86; what does location 0x61
| from inb(0x61) do? Is it something very well known in intel archs?

Yes, see any Intel PIIX or ICH chipset spec.
I'm looking at the 82801BA I/O Controller Hub 2 (ICH2)
spec., section 9.7.1.

Port 0x61 is the NMI status and control register.

-- 
~Randy

Re: question: i/o port 0x61 on x86 archs

[Ingo Oeser]

On Sat, Jun 08, 2002 at 03:01:13PM -0700, Randy Dunlap wrote:
> On Sat, 8 Jun 2002, Ravikiran G Thirumalai wrote:
> | I have a question regdg do_nmi routine in x86; what does location 0x61
> | from inb(0x61) do? Is it something very well known in intel archs?
> Port 0x61 is the NMI status and control register.

So it should exist a '#define' for this somewhere. 

People who tend to disagree here, may try to use *.i files
instead of *.c and *.h files next time.

The IRQ-handling code is full of these beauties and several
drivers are too.

Regards

Ingo Oeser
-- 
Science is what we can tell a computer. Art is everything else. --- D.E.Knuth

Re: question: i/o port 0x61 on x86 archs

[Randy.Dunlap]

On Wed, 12 Jun 2002, Maciej W. Rozycki wrote:

| On Wed, 12 Jun 2002, Ingo Oeser wrote:
|
| > > Port 0x61 is the NMI status and control register.
| >
| > So it should exist a '#define' for this somewhere.
| >
| > People who tend to disagree here, may try to use *.i files
| > instead of *.c and *.h files next time.
|
|  Feel free to submit a patch.

I say "please submit a patch."
I very much dislike such inline constants
(for register offsets and register bits).

-- 
~Randy

Re: question: i/o port 0x61 on x86 archs

[Maciej W. Rozycki]

On Wed, 12 Jun 2002, Ingo Oeser wrote:

> > Port 0x61 is the NMI status and control register.
> 
> So it should exist a '#define' for this somewhere. 
> 
> People who tend to disagree here, may try to use *.i files
> instead of *.c and *.h files next time.

 Feel free to submit a patch.

-- 
+  Maciej W. Rozycki, Technical University of Gdansk, Poland   +
+--------------------------------------------------------------+
+        e-mail: macro@ds2.pg.gda.pl, PGP key available        +