On Sun, Jan 29, 2012 at 08:07:03AM -0800, Paul Goyette wrote:
Recent amd test runs on both my own and releng's test beds show a huge
increase if failed tests on amd64 - it jumped from less than a handful
to more than 800!
This appears to have started with the malloc(9) to kmem(9) changes that
were committed on 2012-01-27 between 16:50:04 and 22:20:04 UTC.
I was experiencing similar failures (but with RUMP_LOCKDEBUG), and the
following change fixed the problem.
http://mail-index.netbsd.org/source-changes/2012/01/29/msg031213.html