serial 18250: too much work for irq 17

[TBH]

Hey guys, I have found that the cause of this problem is the 56k modem, remove the modem, and it should take care of it

[KermitTensmeyer]

I have this same problem.

No Modem.

Modem-Manager installed on fresh (brand-new install) on both live-cd install and full dvd install.

apt-get purge modemmanager.

however

`cat /proc/interrupts` shows irq17 associated with radeon (ie video card)

why does the kernel think that serial8250 is associated with irq17?


tensmeyer.kermit@consultant.com

[ well of course it's going to have problems if the serial8250 driver is responding to irq17 and irq17 is being driven by the video card. that is much to many interrupts for a tty driver to handle!]

[ph0852]

Hi, I'm getting system crashes. The messages "serial8250: too much work for irq17" showed up toward the bottom of my dmesg output after restarting.

I just completed a new install of 10.04 Lucid with Kernal 2.6.32-22-generic, Gnome 2.30.0 on a Dell Dimension Intel-Celeron 2.4 GHz.

The system seems to run fine up until it crashes. Performance is good. We typically are running firefox with 4-6 tabs. I've had terminal up also. Not normally much else running, although I was looking at some pictures earlier today and had gimp up.

I have 10+ years experience on Unix doing software development and support and used Linux for about the last year. I'm happy to look in the logs if that is helpful.

cat /proc/interrupts
CPU0
0: 56 IO-APIC-edge timer
1: 1792 IO-APIC-edge i8042
4: 2 IO-APIC-edge
6: 2 IO-APIC-edge floppy
7: 0 IO-APIC-edge parport0
8: 0 IO-APIC-edge rtc0
9: 0 IO-APIC-fasteoi acpi
14: 12940 IO-APIC-edge ata_piix
15: 28733 IO-APIC-edge ata_piix
16: 46698 IO-APIC-fasteoi uhci_hcd:usb2, i915@pci:0000:00:02.0
17: 244 IO-APIC-fasteoi Intel 82801DB-ICH4, eth0
18: 0 IO-APIC-fasteoi uhci_hcd:usb4
19: 0 IO-APIC-fasteoi uhci_hcd:usb3
23: 140874 IO-APIC-fasteoi ehci_hcd:usb1
NMI: 0 Non-maskable interrupts
LOC: 381706 Local timer interrupts
SPU: 0 Spurious interrupts
PMI: 0 Performance monitoring interrupts
PND: 0 Performance pending work
RES: 0 Rescheduling interrupts
CAL: 0 Function call interrupts
TLB: 0 TLB shootdowns
TRM: 0 Thermal event interrupts
THR: 0 Threshold APIC interrupts
MCE: 0 Machine check exceptions
MCP: 10 Machine check polls
ERR: 0
MIS: 0

[littlepeon]

hey there,
got a dell 2400, and now i have the same issues
did NOT have this with jaunty, but upgraded to karmic and now have this
only thing added to this system is a KVM sharing the keyboard/mouse/monitor with another older compaq system.
originally i thought it was the KVM causing this irq situation, but have done some digging and found this thread.

go figure
-Peon

[bdk6m2007]

Not that this is much help, but I have a dell 2400 with the same symptoms listed above. After my most recent update the problem is gone. So it appears something was fix somewhere that at least made the problem go away.

I guess word to the wise is to be sure you've got the latest updates and retest.

[ph0852]

Having the latest updates is good advice. I wish that was the solution, but I doubt it since Update Manager is running every day and I'm accepting the updates. I just ran Update Manager manually and loaded the latest. Maybe today's fixes will resolve....

My Dell is a Dimension 2400. I suspect the hardware is the issue. Since I'm getting the crash a couple of times a day I should know tomorrow if the system is now stable. If the hardware is the issue I'm hoping there is a configuration change to address the issue. Some of the older messages in the thread mentioned more interrupts than a serial port can accept. If I'm reading my configuration correctly (see above) IRQ17 is associated with ETH0 on my system.

This is a summary of the history of this issue. The system was running windows up until last weekend. Over the weekend I installed Lucid with the basic install pulled from the default download site. The issue started showing up ever since. Update Manager has run daily. The crashes occur once or twice daily.

[rustiguzzi]

When I first saw these messages on my Dell Dimension 2400, recently converted to Ubuntu 10.04.1, I thought they began with 'serial 18250' but on looking closely I found that they actually began with 'serial8250'. It makes a difference, as I will explain . . .

At boot-up time there were dozens of them in the syslog, interspersed with messages of the form

[ 75.252661] ttyS2: 163328 input overrun(s)

but they went unnoticed until one day I wanted to boot in recovery mode, when they came up on-screen and the system froze in mid-boot. I found irq17 to be associated with the ethernet connection:

17: 1779 IO-APIC-fasteoi Intel 82801DB-ICH4, eth0

so didn't think it could be anything to do with the PCI modem card, which had solved it for others. Then (eventually) light dawned, in that '8250' is a type of UART. See this thread:

http://readlist.com/lists/redhat.com...38/190954.html

for some more info. The simple solution was to remove the PCI modem, whereupon the error messages ceased. Why the modem, unused, should be causing interrrupts on irq17 is a mystery, but at least they've gone away now.

BTW, the chief sufferers with this seem to be folk using Dell Dimension 2400s, who have recently upgraded their OS (and not only Ubuntu).