panic at turnstile_wait (Re: panic at sched_add_internal)

[John Baldwin]

On Friday 16 July 2004 02:11 pm, Norikatsu Shigemura wrote:
> > - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
> > - - [thread 100124]
> > Stopped at      turnstile_wait+0x17a:   movl    %eax,0x10(%edx)
> > db> trace
> > turnstile_wait(c301e340,c0877f40,c36e4420,c0877f40,f9) at
> > turnstile_wait+0x17a _mtx_lock_sleep(c0877f40,0,c07bd610,f9,0) at
> > _mtx_lock_sleep+0x125 _mtx_lock_flags(c0877f40,0,c07bd610,f9,0) at
> > _mtx_lock_flags+0x64
>
> [...]
>
> > -->		LIST_INSERT_HEAD(&ts->ts_free, td->td_turnstile, ts_hash);
> > 		MPASS(owner == ts->ts_owner);
> > 	}
> > 	td->td_turnstile = NULL;
> > [...]
> > - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
> > - -
>
> 	Same panic has happend at same place on new current (+1day).

Do you have INVARIANTS turned on?  Also, can you capture the actual panic 
message?  If you do have INVARIANTS on, then it seems that pointers are 
magically changing to NULL after the assertion check.  I'd be worried about 
bad memory.

-- 
John Baldwin <jhb at FreeBSD.org>  <><  http://www.FreeBSD.org/~jhb/
"Power Users Use the Power to Serve"  =  http://www.FreeBSD.org