Re: trouble generating a core file

[Stefan Krüger <skrueger@xxxxxxxxxxxxxxxx>]

Chuck Tuffli wrote:
> I have a user who is being sadistic to my driver and ultimately causes a
> panic. Once ddb pops up, I tried typing 'panic' to generate a core file,
> but the system prints out the following:
> 
> LWKT_WAIT_IPIQ WARNING! 1 wait 0 (-1)
> panic: LWKT_WAIT_IPIQ
> mp_lock = 00000001; cpuid = 1; lapic.id = 01000000
> boot() called on cpu #1
> 
> The (-1) number cycles through the values from -1 to -9 and then seems
> to hang without dumping a core. The kernel is based on the _Preview tag
> from today using a GENERIC conf file with DDB, SMP, and IOAPIC enabled.
> The system is based on an Intel chipset (SE7221BK1-E) and has a P4 with
> hyperthreading enabled.
> 
> The above message doesn't print out with HT disabled and it appears that
> the core is being written out. But, then the following is displayed at
> boot:
> 
> Checking for core dump...
> savecore: reboot after panic: from debugger
> savecore: system went down at Wed Dec 31 16:00:56 1969
> dump time is unreasonable
> 
> When the system comes up, the date is
> 
> Sun Feb  5 22:40:48 PST 2006
> 
> Prior to the machine crashing, the date got set with 'rdate -n
> us.pool.ntp.org' and was correct, but the system has come up with that
> particular date from time to time today.
> 
> For now, I can live without HT enabled, but I'd like to get a core file.
> Is the proceedure I'm using to generate the core wrong? Also why is the
> system date so wacky? TIA!

could you
$ ident src/sys/kern/lwkt_ipiq.c
and check if you're using at least 1.10? I had a similar LWKT_WAIT_IPIQ
panic which was fixed with 1.10...

take care,
SK