On 23/01/17(Mon) 23:53, cheek...@gmx.com wrote:
> I tried Martin's first patch with a newly-cvsup source and have not
> been able to cause another crash after more than 20 reboots, and
> rebooting into the bsd.rd.
Good to know.
> I do still have the panicd vm console.
In that case could you print the trace of all the threads in the bind
process? In ddb it should be:
ddb{3}> tr /p 0t$tidofathread
I don't know a better way to find the tid than using "ps /w" beforehand.
> Is showing one of the isc bind named threads with thrsleep tell you anything?
It helps confirming Alexander's theory.
> ddb{3}> ps
> PID PPID PGRP UID S FLAGS WAIT COMMAND
> 93716 1 93716 0 3 0x100083 ttyin getty
> 29744 1 29744 0 3 0x100083 ttyin getty
> 97017 1 97017 0 3 0x100083 ttyin getty
> 15604 1 15604 0 3 0x100083 ttyin getty
> 82395 1 82395 0 3 0x100083 ttyin ksh
> 16881 1 16881 0 3 0x100098 poll cron
> 82696 1 82696 741 3 0x98 sigwait named
> *82696 1 82696 741 7 0x4000010 named
> 82696 1 82696 741 7 0x4000010 named
> 82696 1 82696 741 7 0x4000010 named
> 82696 1 82696 741 7 0x4000010 named
> 82696 1 82696 741 3 0x4000090 thrsleep named
> 82696 1 82696 741 2 0x4000010 named
> 54311 1 54311 99 3 0x100090 poll sndiod
> 4358 1 4358 110 3 0x100090 poll sndiod
> 7746 64349 64349 95 3 0x100092 kqread smtpd
