Kernel panic on "cups" startup

[Toralf Lund]

Edward Dekkers wrote:

> Toralf Lund wrote:
>
>> Edward Dekkers wrote:
>>
>>> Toralf Lund wrote:
>>>
>>>> My (Red Hat 9) system has started displaying a "kernel panic" 
>>>> message on the console, then freezing completely, when the "cups" 
>>>> service is started. Everything worked fine a couple of days ago, 
>>>> and as usual, I know of no setup changes made since then. With cups 
>>>> disabled, the system seems to operate correctly, except I can't 
>>>> print, obviously. There is no printer connected directly to this 
>>>> host; cups is used to print across the net. I'm not sure I can give 
>>>> you the exact error message, as it doesn't seem to be stored in the 
>>>> system log or anything.
>>>>
>>>> Any ideas?
>>>>
>>>> - Toralf
>>>>
>>>>
>>> No errors even in /var/log/cups?
>>
>>
>>
>> Nothing there except
>>
>> I [09/Mar/2004:14:19:31 +0100] Listening to 0:631
>> I [09/Mar/2004:14:19:31 +0100] Configured for up to 100 clients.
>> I [09/Mar/2004:14:19:31 +0100] Allowing up to 10 client connections 
>> per host.
>> I [09/Mar/2004:14:19:31 +0100] LoadPPDs: Read "/etc/cups/ppds.dat", 
>> 18 PPDs...
>> I [09/Mar/2004:14:19:32 +0100] LoadPPDs: No new or changed PPDs...
>>
>> - Toralf
>>
>>
>
> Nothing I can see there except the first line - Listening to 0:631.
>
> I have a lot of those lines in my log but they always have a hex value 
> greater than 0, e.g. Listening to 7f000001:631.
>
> May be nothing, but that's all that sprang out at me.

With a little help from the CUPS mailing list, I found a workaround at 
least; I removed the backend program for "serial", and everything works 
now. Conclusion: Something is weird with the serial ports, serial driver 
or similar. When the CUPS backend support scans the ports on startup, 
everything goes wrong.

- Toralf