Re: recent changes prohibit vinum swap.

On Friday, 26 September 2003 at 19:28:45 -0400, David Gilbert wrote:
>>>>>> "Robert" == Robert Watson <rwatson_at_freebsd.org> writes:
>
> Robert> On Fri, 26 Sep 2003, David Gilbert wrote:
>
>>> Recent changes to -CURRENT prohibit vinum swap:
>>>
>>> [1:6:306]root_at_mu:~> swapon /dev/vinum/swapmu swapon:
>>> /dev/vinum/swapmu: Operation not supported by device
>
> Robert> In order to support swapping, Vinum will need to be modified
> Robert> to use struct disk and the disk(9) API, rather than exposing
> Robert> its storage devices directly via struct cdevsw and
> Robert> make_dev(9).  I.e., Vinum probably needs to start approaching
> Robert> things as "disks" rather than "devices", a distinction that's
> Robert> becoming more mature in -CURRENT.
>
>>> From a quick read of vinumconfig.c, I'm guessing this wouldn't be
>>> hard to
> Robert> implement.  Some subset of struct sd, struct plex, and struct
> Robert> volume will need to start holding a struct disk instance which
> Robert> would be passed to disk_create() instead of a call to
> Robert> make_dev().  Much of the remainder will just consist of a bit
> Robert> of tweaking to make Vinum extract its data from
> bp-> bio_disk->d_drv1 instead of bp->b_dev, replacing the ioctl dev_t
> Robert> argument with a disk argument, etc.
>
> Is this something that someone can help me with quickly, or should I
> downgrade the machine until it's been done? 

Don't hold your breath.  This will probably happen in the course of
migrating Vinum functionality to GEOM.

> Is there a quick hack to make it work for now?

None that I know of.

> If I must downgrade, what date would be appropriate?

Sorry, I can't help there.  Maybe phk can give you some indication.

> Robert> I also noticed that the vinum commandline tool is a bit
> Robert> devfs-unfriendly, or at least, it gets pretty verbose about
> Robert> how all the files/directories it wants to create are already
> Robert> present.  It could be that a test for devfs conditionally
> Robert> causing a test for EEXIST would go a long way in muffling the
> Robert> somewhat loud complaining :-).
>
> Well... vinum is fragile in a whole bunch of ways.  vinum rm often
> leaves things in an inconsistant state.  I almost always reboot now
> after using it.  vinum rename doesn't change the devfs vinum directory
> ... which then also requires a reboot to correct.

Hmm.  That's another one to look at.

> Another thing that's very fragile is resetconfig.  It blanks memory,
> but not disk.

It should do.  It leaves the device names, though.  That's arguably a
bug.

Greg
--
See complete headers for address and phone numbers.
NOTE: Due to the currently active Microsoft-based worms, I am limiting
all incoming mail to 131,072 bytes.  This is enough for normal mail,
but not for large attachments.  Please send these as URLs.