Jared D. McNeill wrote:
Michael Lorenz wrote:I guess you may see further messages after more minutes. If so the problem is caused by pmf merge and there might be something wrong in scsipi or ahc(4) detach.ahc at pci doesn't have a detach method so I added one - it's never called.I think tsutsui meant suspend, not detach.
I also noticed on one of the other lists that someone mentioned a potential vnode leak -- they weren't getting the unmount for the root FS and hence the disk caches weren't being flushed on last close but instead late during shutdown after interrupts had been disabled.
--rafal