[ale] /proc/interrupts

[ari]

no, it's not an error.  it's floating point error reporting, and will be
present in all of your single processor *86 systems >80386.

from the linux kernel source, arch/i386/kernel/irq.c:

-------------
#ifdef __SMP__


/*
 * On SMP boards, irq13 is used for interprocessor interrupts (IPI's).
 */
static struct irqaction irq13 = { smp_message_irq, SA_INTERRUPT, 0, "IPI",
NULL, NULL };

#else

/*
 * Note that on a 486, we don't want to do a SIGFPE on a irq13
 * as the irq is unreliable, and exception 16 works correctly
 * (ie as explained in the intel literature). On a 386, you
 * can't use exception 16 due to bad IBM design, so we have to
 * rely on the less exact irq13.
 *
 * Careful.. Not only is IRQ13 unreliable, but it is also
 * leads to races. IBM designers who came up with it should
 * be shot.
 */

static void math_error_irq(int cpl, void *dev_id, struct pt_regs *regs)
{
   outb(0,0xF0);
   if (ignore_irq13 || !hard_math)
      return;
   math_error();
}  

static struct irqaction irq13 = { math_error_irq, 0, 0, "math error", NULL,
NULL };

#endif
-----------------

you may also want to take a look at Documentation/smp.tex -- also in the linux
kernel source tree.

ari


Dunlap, Randy (randy.dunlap at intel.com) said this stuff:

> I don't think it's an error either.  It's just that IRQ 13
> is associated with floating point exceptions (match coprocessor,
> built-in, or even floating point emulation software).
> 
> ~Randy
> 
> 
> > -----Original Message-----
> > From: Linux Idiot [mailto:esoteric at denali.atlnet.com]
> > Sent: Monday, September 14, 1998 5:52 AM
> > To: Ben Phillips
> > Cc: ale at cc.gatech.edu
> > Subject: Re: [ale] /proc/interrupts
> > 
> > 
> > Ben Phillips wrote:
> > > 
> > > It's not really an error AFAIK, I'm just looking at what I 
> > see when I cat
> > > /proc/interrupts.
> > > 
> > > On Sun, 13 Sep 1998, Dunlap, Randy wrote:
> > > > Is it for IRQ 13 (hex 0d)?  This is an old AT standard for
> > > > floating point exceptions.  What kind of processor are you using?
> > > 
> > > Yes, it is on 13.  It's an old 486SX.  D'you suppose it's 
> > there because of
> > > some real error, or because it's an interrupt set up to 
> > catch errors if they
> > > happen?
> > 
> > Just rebooted my box to win95 and irq 13 is assigned to the math 
> > coprocessor... ??
> > 
> > > 
> > >                       __   _
> > > Ben Phillips         /  '_' )         ,,,    "I wish I 
> > could file suits
> > > pynk at cc.gatech.edu   |  | ()|||||||||[:::)   against 
> > stupidity. The world
> > >                      \__.-._)         '''    would become a jail."
> > >                                                       --  Cybervox
> > 
> > --
> > Until later: Geoffrey		esoteric at denali.atlnet.com
> > 
> > You mean you paid MONEY for Service Pack '98????
> >