Re: [PATCH] NMI request/release, version 5 - I think this one's ready

[John Levon]

On Thu, Oct 24, 2002 at 12:43:50PM -0500, Corey Minyard wrote:

> #define NOTIFY_DONE        0x0000        /* Don't care */
> #define NOTIFY_OK        0x0001        /* Suits me */
> #define NOTIFY_STOP_MASK    0x8000        /* Don't call further */
> 
> I'mt taking these to mean that NOTIFY_DONE means you didn't handle it, 
> NOTIFY_OK means you did handle it, and you "or" on NOTIFY_STOP_MASK if 
> you want it to stop.  I'm thinking that stopping is probably a bad idea, 
> if the NMI is really edge triggered.

> There's a comment in do_nmi() that says that the NMI is edge triggered. 

So there is. Darn. You could special case default_do_nmi, only printing
out "Unknown NMI" iff the reason inb() check fails, /and/ no previous
handlers set handled. Theoretically we have to take the cost of the
inb() every time otherwise we can lose one of the NMISC-based things in
default_do_nmi ... I can always see if this makes a noticable practical
difference for oprofile under high interrupt load

(I also need to be able to remove the NMI watchdog handler, but that's
an oprofile-specific problem).

Perhaps things will end being best to leave the current fast-path thing,
but we should make sure that we've explored the possibilities of
removing it first ;)

thanks
john

-- 
"This is playing, not work, therefore it's not a waste of time."
	- Zath