Re: buildworld panic on ia64

In the last episode (Jul 07), Anton Shterenlikht said:
> On Tue, Jul 07, 2009 at 11:50:58AM +0200, Rink Springer wrote:
> > On Tue, Jul 07, 2009 at 10:48:08AM +0100, Anton Shterenlikht wrote:
> > > panic: mtx_lock_spin() of destroyed mutex _at_ /usr/src/sys/ia64/ia64/trap.c:680
> > > cpuid = 0
> > > KDB: enter: panic
> > > [thread pid 67078 tid 100097 ]
> > > Stopped at      kdb_enter+0x92: [I2]    addl r14=0xffffffffffe2a8e8,gp ;;
> > 
> > Do you have a backtrace ?
> 
> no, sorry, I was too quick to reboot.
> I tried to reproduce the error, got this on the way:

If you add "options KDB_TRACE" to your kernel config, you'll get a stack
trace automatically on every panic.
 
> 	# XXX: bogusly disabled high FP regs
> 
> which is reported from by sys/ia64/ia64/trap.c, and then this error
> (but no panic this time):
> 
> ===> usr.bin/yacc (all)
> cc -O2 -fno-strict-aliasing -pipe  -std=gnu99  -c /usr/src/usr.bin/yacc/closure.c
[...]
> cc -O2 -fno-strict-aliasing -pipe  -std=gnu99   -o ypwhich ypwhich.o
> 1 error
> *** Error code 2
> 1 error
> *** Error code 2
> 1 error

There are no errors in this output; just make reporting that there was an
error.  At -j6, the error message itself may be hundreds of lines back. 
Make lets all remaining parallel jobs finish before exiting.

-- 
	Dan Nelson
	dnelson_at_allantgroup.com