On 16/01/07 12:46, Eva Myers wrote:
> Dear All,
> Thank you very much for all your help with setting up my new SL4
> machine.  I have just one more minor problem: the following sequence
> of messages appears in the system logs every half hour.
> 
> Jan 16 11:29:35 shanghai kernel: irq 217: nobody cared! (screaming interrupt?)
> Jan 16 11:29:35 shanghai kernel: irq 217: Please try booting with acpi=off and report a bug
> Jan 16 11:29:35 shanghai kernel:  [<c0107532>] __report_bad_irq+0x3a/0x77
> Jan 16 11:29:35 shanghai kernel:  [<c01077a9>] note_interrupt+0xea/0x115
> Jan 16 11:29:35 shanghai kernel:  [<c0107a55>] do_IRQ+0x143/0x1ae
> Jan 16 11:29:35 shanghai kernel:  [<c02d51a4>] common_interrupt+0x18/0x20
> Jan 16 11:29:35 shanghai kernel:  [<c01040e8>] mwait_idle+0x33/0x42
> Jan 16 11:29:35 shanghai kernel:  [<c01040a0>] cpu_idle+0x26/0x3b
> Jan 16 11:29:35 shanghai kernel: handlers:
> Jan 16 11:29:35 shanghai kernel: [<c0241ae2>] (ide_intr+0x0/0x11e)
> Jan 16 11:29:35 shanghai kernel: [<c025a6c0>] (usb_hcd_irq+0x0/0x4b)
> Jan 16 11:29:35 shanghai kernel: Disabling IRQ #217

The output from
# cat /proc/interrupts
# dmesg
# lspci -vvv
could help to at least identify what is going on. Also, you could
perhaps identify the process that causes this (cronjob? calendar alarm?)

> This machine has an Intel D965SSCK Core 2 Duo 6300 2MB cache chipset,
> and I'll post the rest of the specs if you think they are relevant.
> Is the fix just to add acpi=off to the kernel options?

acpi=noirq (on the kernel .. line in /boot.grub/grub.conf) is a bit
softer. But this isn't guaranteed to make the problem go away, if you
really have a device that tries to send an interrupt.

Regards
Jan