Boot hangs starting postfix

R. Joseph Wright joseph at
Sat Jun 12 01:06:07 PDT 2004

"Joerg Sonnenberger" <joerg at xxxxxxxxxxxxxxxxx> wrote in message
news:20040611173252.GA1198 at xxxxxxxxxxxxxxxxxxxx
> Can you try uncomment following line (521) in kern/kern_lockf.c and see
> if you still hit that problem?
> do_wakeup:
> #ifdef LOCKF_DEBUG
> if (lf_print_ranges)
> lf_print_lock(lock);
> #endif
> if (wakeup_needed) /* <-- this one! */
> lf_wakeup(lock, start, end);
> error = 0;

I tried this, it did not work.

> It does decrease the performance, but it isn't necessary for operation.
> Also, it could help if you can provide either
> a) ktrace output of all postfix processes or at least the bounce
> b) the dmesg output of a kernel compiled with LOCKF_DEBUG in CFLAGS,
>    use "w lf_print_range 1" from ddb prompt to activate this after the
>    initial boot sequence [beware, heavy]
> c) a way to reproduce this problem, so I can fix it myself
> Joerg

Tomorrow I'll try ktrace.  I don't know how to reproduce the problem, it
started when I built a new world on June 1.  My previous system was from
March.  I also rebuilt the latest postfix-current and got the same results.
I'll send
a dmesg or anything else if you need it.

R. Joseph Wright
Those are my principles.  If you don't like them I have others.
   --Groucho Marx

More information about the Bugs mailing list