2005-04-14 17:11:38

by Lee Revell

[permalink] [raw]
Subject: spurious 8259A interrupt: IRQ7

I get this message occasionally on both my machines. I googled and saw
some references to this message on 2.4 but nothing for 2.6. Some of the
references were to APIC, which I don't have enabled.

Both machines are using VIA chipsets and display the "VIA IRQ fixup"
message on boot. I think this behavior started about the same time that
message started to appear.

On both machines the parallel port is disabled in the BIOS and there's
nothing on IRQ7.

Lee


2005-04-14 20:45:05

by Bjorn Helgaas

[permalink] [raw]
Subject: Re: spurious 8259A interrupt: IRQ7

On Thu, 2005-04-14 at 13:11 -0400, Lee Revell wrote:
> I get this message occasionally on both my machines. I googled and saw
> some references to this message on 2.4 but nothing for 2.6. Some of the
> references were to APIC, which I don't have enabled.
>
> Both machines are using VIA chipsets and display the "VIA IRQ fixup"
> message on boot. I think this behavior started about the same time that
> message started to appear.

The VIA IRQ fixup in 2.6.11 is broken. It works for some, but
not all boxes with VIA hardware.

There's a fix in 2.6.12-rc2-mm3. Actually, I doubt that it will
help you, though -- the 2.6.11 breakage is such that some machines
that need the fixup don't get it (and don't print the "VIA IRQ
fixup message").

2005-04-14 20:56:55

by Lee Revell

[permalink] [raw]
Subject: Re: spurious 8259A interrupt: IRQ7

On Thu, 2005-04-14 at 14:43 -0600, Bjorn Helgaas wrote:
> On Thu, 2005-04-14 at 13:11 -0400, Lee Revell wrote:
> > I get this message occasionally on both my machines. I googled and saw
> > some references to this message on 2.4 but nothing for 2.6. Some of the
> > references were to APIC, which I don't have enabled.
> >
> > Both machines are using VIA chipsets and display the "VIA IRQ fixup"
> > message on boot. I think this behavior started about the same time that
> > message started to appear.
>
> The VIA IRQ fixup in 2.6.11 is broken. It works for some, but
> not all boxes with VIA hardware.
>
> There's a fix in 2.6.12-rc2-mm3. Actually, I doubt that it will
> help you, though -- the 2.6.11 breakage is such that some machines
> that need the fixup don't get it (and don't print the "VIA IRQ
> fixup message").
>

Is the VIA IRQ fixup related to the "spurious interrupts" messages in
any way? Googling the 2.4 threads on the issue gave me the impression
that it's related to broken hardware. I think excessive disk activity
might trigger it.

Anyway it's low priority as the message appears to be completely
harmless.

Lee

2005-04-14 21:04:02

by Bjorn Helgaas

[permalink] [raw]
Subject: Re: spurious 8259A interrupt: IRQ7

On Thu, 2005-04-14 at 16:56 -0400, Lee Revell wrote:
> Is the VIA IRQ fixup related to the "spurious interrupts" messages in
> any way? Googling the 2.4 threads on the issue gave me the impression
> that it's related to broken hardware. I think excessive disk activity
> might trigger it.

If you need the VIA IRQ fixup and don't have it, I would expect
some interrupt to be routed to the wrong IRQ. That might give
you a "spurious interrupt" on the wrong IRQ, but your device would
probably just not work at all.


2005-05-22 22:41:36

by Jesper Juhl

[permalink] [raw]
Subject: Re: spurious 8259A interrupt: IRQ7

On 4/14/05, Bjorn Helgaas <[email protected]> wrote:
> On Thu, 2005-04-14 at 16:56 -0400, Lee Revell wrote:
> > Is the VIA IRQ fixup related to the "spurious interrupts" messages in
> > any way? Googling the 2.4 threads on the issue gave me the impression
> > that it's related to broken hardware. I think excessive disk activity
> > might trigger it.
>
> If you need the VIA IRQ fixup and don't have it, I would expect
> some interrupt to be routed to the wrong IRQ. That might give
> you a "spurious interrupt" on the wrong IRQ, but your device would
> probably just not work at all.
>

For what it's worth I'll chime in with my own info on this subject.

I also see the
spurious 8259A interrupt: IRQ7.
messages from time to time. Actually I got one just a little while ago :
$ dmesg | grep spurious
[18994.222451] spurious 8259A interrupt: IRQ7.
The system seems to be in good shape - it has never caused me any
actual trouble.

As for the via irq fixup I see this :
$ dmesg | grep -i fixup
[ 74.629393] PCI: Via PIC IRQ fixup for 0000:00:09.0, from 255 to 3

The hardware is a ASUS A7M266 mobo with a AMD Athlon (t-bird) 1.4GHz CPU.

The kernel is 2.6.12-rc4-mm2 :
$ uname -a
Linux dragon 2.6.12-rc4-mm2 #2 Mon May 16 18:14:13 CEST 2005 i686
unknown unknown GNU/Linux

I've seen some references to APIC for this issue, so here are the APIC
related settings in my .config :

juhl@dragon:~/download/kernel/linux-2.6.12-rc4-mm2$ grep -i APIC .config
CONFIG_X86_GOOD_APIC=y
# CONFIG_X86_UP_APIC is not set


--
Jesper Juhl <[email protected]>
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please http://www.expita.com/nomime.html