Ralph Shumaker wrote:
Today I booted up my computer and did "sudo ln -s /dev/ttyS14
/dev/modem". Then I launched "/usr/bin/system-control-network" and
activated the modem (already set to dial my ISP and known to work).
Occasionally, the handshaking freezes or gets confused. I've never
found out why. But on this computer, this is the first time that
attempting to cancel the initiation (or dial-up) messes with ppp.
Subsequent attempts to connect complains that ppp is already active.
But then I got something strange. I had opened a terminal to verify
that /dev/modem didn't exist. Then I launched the icon that I set up to
execute "sudo ln -s /dev/ttyS14 /dev/modem". Then I used the terminal
again to verify that /dev/modem had been created. Then I initiated the
dial-up and it hung. I closed the system-control-network and got a
strange error in the terminal I had left open:
[EMAIL PROTECTED] ~]$
Message from [EMAIL PROTECTED] at Mon Mar 13 09:50:44 2006 ...
localhost kernel: Disabling IRQ #10
I opened another terminal and ran dmesg and got:
[snip]
application mixer_applet2 uses obsolete OSS audio interface
CSLIP: code copyright 1989 Regents of the University of California
PPP generic driver version 2.4.2
irq 10: nobody cared!
[<c01516f4>] __report_bad_irq+0x24/0x7f
[<c01517c6>] note_interrupt+0x59/0x83
[<c0150a8a>] __do_IRQ+0x201/0x367
[<c0105b1d>] do_IRQ+0x4a/0x82
=======================
[<c02d0ed1>] rh_report_status+0x0/0x398
[<c0103c0e>] common_interrupt+0x1a/0x20
[<c01282cc>] __do_softirq+0x2c/0x8a
[<c0105c29>] do_softirq+0x3e/0x42
=======================
[<c0105b24>] do_IRQ+0x51/0x82
[<c0105b24>] do_IRQ+0x51/0x82
[<c0103c0e>] common_interrupt+0x1a/0x20
[<c010101a>] default_idle+0x0/0x29
[<c0101040>] default_idle+0x26/0x29
[<c01010a6>] cpu_idle+0x34/0x4c
[<c042472a>] start_kernel+0x15f/0x1b9
[<c04242fe>] unknown_bootoption+0x0/0x1cd
handlers:
[<de9963ec>] (snd_via82xx_interrupt+0x0/0x322 [snd_via82xx])
[<c02d2f1e>] (usb_hcd_irq+0x0/0x52)
[<c0283b64>] (serial8250_interrupt+0x0/0x34e)
Disabling IRQ #10
Anyone know what happened?
This looks like a problem with shared IRQ. The PCI bus has the ability
to share the IRQ (Interrupt ReQuest line) between multiple devices.
Every once in a while one of the devices gets confused and either
forgets what it's supposed to do, or it starts spewing requests for no
good reason. On my machine I use the following command to see what
interrupts are being used for which devices:
$ cat /proc/interrupts
and I get results like this:
[EMAIL PROTECTED] ~]$ cat /proc/interrupts
CPU0
0: 257957843 XT-PIC timer
1: 19271 XT-PIC i8042
2: 0 XT-PIC cascade
5: 48927287 XT-PIC ohci_hcd:usb2, eth0
7: 88 XT-PIC parport0
8: 370613 XT-PIC rtc
10: 20522753 XT-PIC ide3, ohci_hcd:usb3, nvidia
11: 11530161 XT-PIC acpi, EMU10K1, ehci_hcd:usb1,
uhci_hcd:usb4, uhci_hcd:usb5
12: 2778173 XT-PIC i8042
14: 482201 XT-PIC ide0
15: 2318206 XT-PIC ide1
NMI: 0
ERR: 0
I then move my PCI cards around so that I get one interrupt per card if
possible. Looking at the results for this machine, I obviously haven't
done that, but it hasn't given me problems either. Depending on your
motherboard, you may have a listing of which interrupt goes to which
socket to help you choose. This tends to minimize problems with IRQs.
Gus
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
