Re: kern/41725 (uvm fault and problems unmounting file sytems at shutdown)

[David Young <dyoung%pobox.com@localhost>]

On Sun, Jul 26, 2009 at 01:25:45PM +0200, Alan Barrett wrote:
> On Sat, 25 Jul 2009, David Young wrote:
> > Try this patch.  It maintains the openmask as a disk is supposed to.
> 
> Thank you; that does help.  Now I see this:
> 
>       ppb2: detached
>       ppb1: detached
>       audio1: detached
> 
>       unmounting 0xd02e5004 /cgd1a/ (/dev/cgd1a)...
>       unmounting 0xd02d422c / (/dev/md0a)...
>       forcefully unmounting /cgd1a/ (/dev/cgd1a)...forcefully unmounted
>       /dev/cgd1a on /cgd1a/ type ffs
> 
>       unmounting 0xd02d422c / (/dev/md0a)...
>       forcefully unmounting / (/dev/md0a)...forcefully unmounted
>       /dev/md0a on / type ffs
>       md0: detached
>       acpi0: entering state 5
> 
> The system powers off correctly, and the md0 device is detached after,
> rather than before, the md0a file system is unmounted.  These are both
> improvements over the previous situation.
> 
> However, I remain concerned about the fact that the /dev/md0a and
> /dev/cgd1a file systems were unmounted forcefully instead of gracefully,
> and the fact that I don't see any "cgd1: detached" or "wd0: detached"
> messages.

cgd1 is not a pseudo-device, so it cannot be detached.  If it is backed
by wd0, then wd0 will not be detached.

Dave

-- 
David Young             OJC Technologies
dyoung%ojctech.com@localhost      Urbana, IL * (217) 278-3933