Deadlock

[Don Lewis]

On  9 Jul, Peter Holm wrote:
> Here's a trace from a deadlock in a kernel from Jul 8 13:51 UTC:
> 
> http://people.freebsd.org/~pho/stress/cons36.html

It sure looks like a mutex implementation problem.  Process 4616 is
waiting on pool mutex c05dd04c, but this process doesn't show up on the
blocked queue for the mutex.  That sounds a lot like the bug fixed in
rev 1.129 of kern_mutex.c.