Re: 1000000000 as irq count init (was: Procinfo behaving strange

[Adrian Bunk]

On Sun, 23 Jun 2002, Axel Thimm wrote:

> On Fri, Jun 21, 2002 at 08:57:25AM +0200, Rasmus Bøg Hansen wrote:
> > I upgraded a short time ago from kernel 2.4.18 to 2.4.19-pre10, but now
> > procinfo reports interrupts in a strange way.
> >
> > 2.4.19-pre10:
> >
> > # procinfo
> > [...]
> > irq  0:1000207681 timer                 irq  8:1000000003
> > irq  1:1000004868 keyboard              irq  9:1000000000 acpi
> > irq  2:1000000000 cascade [4]           irq 10:1000007854 eth0
> > irq  3:1000000000                       irq 11:1000114199 nvidia
> > irq  4:1000000000                       irq 12:1000026199 PS/2 Mouse
> > irq  5:1000003195 es1370                irq 13:1000000000
> > irq  6:1000000000                       irq 14:1000016806 ide0
> > irq  7:1000000000                       irq 15:1000000000
> > [...]
>
> I can second this (while I switched from 2.4.19-pre6 to 2.4.18 with RedHat/SGI
> patches, but obviously the same patch hit me also).
>...

This is a known issue. The following patch to fix it is already in
Marcelos BK reporsitory (which means that the fix will be in -pre11/-rc1):

# --------------------------------------------
# 02/06/07	[e-mail]	1.537.1.44
# [PATCH] Remove debug code from /proc/stat
#
# This patch removes a piece of debug code which crept into the kernel
# by accident.
# --------------------------------------------
#
diff -Nru a/fs/proc/proc_misc.c b/fs/proc/proc_misc.c
--- a/fs/proc/proc_misc.c	Wed Jun 12 17:08:55 2002
+++ b/fs/proc/proc_misc.c	Wed Jun 12 17:08:55 2002
@@ -336,7 +336,7 @@
 #if !defined(CONFIG_ARCH_S390)
 	for (i = 0 ; i < NR_IRQS ; i++)
 		proc_sprintf(page, &off, &len,
-			     " %u", kstat_irqs(i) + 1000000000);
+			     " %u", kstat_irqs(i));
 #endif

 	proc_sprintf(page, &off, &len, "\ndisk_io: ");

cu
Adrian

-- 

You only think this is a free country. Like the US the UK spends a lot of
time explaining its a free country because its a police state.
								Alan Cox