On Wed, 23 Aug 1995 15:23:27 +0200 
 Niklas Hallqvist <niklas@appli.se> wrote:

 > Hmm, but doesn't it?  Like fsck, it has to be somewhere in between the
 > layers in order to know block addressing and such.

Ummmm ... not really.  If fsck(8) were like I think you're describing, 
you'd have to have a different one for SCSI disks and SMD disks.  This, 
of course, is not the case, becase the stuff beneath the character-device 
interface translates the offset into a block (and, in some cases, a 
cylinder, track, and sector).

 > Jason>  From my perspective, implementing dynamic resizing of
 > Jason> the ccd device and dynamic re-interleaving is non-trivial.
 > 
 > Probably, but not really *that* hard, or do you think so?  AIX does
 > this in a two step process, first resize of the VG (volume group, not
 > really like NetBSD's ccd, but comparable in this respect) and the
 > resize of the FS.  Couldn't ccdconfig be made to at least *add* disks
 > (and remove the last) to a non-interleaved ccd?  This is useful for
 > things like running out of diskspace on a ccd device.

Note that I said `re-interleaving'.  I have maintained all along that a 
_concatenated_ (i.e. non-interleaved, or, put another way, a ccd with an 
interleave factor of 0) should have no trouble with this.  However, if 
you're using an interleaved ccd, it's just not that easy.  What I mean 
is, not only do you have to move the filesystem blocks around, but you 
have to figure out which component they're on and move the _raw disk 
blocks_ around, too.

In any case, `ccdconfig' doesn't really have to made to do anything.  
Adding a disk to a concatencated ccd is as easy as unconfiguring it, and 
reconfiguring it with the extra component tacked onto the end.  (And, 
doing it that way saves adding the extra ioctls and special case stuff to 
reallocate space for and copy existing configuration information in the 
ccd driver itself.)

--------------------------------------------------------------------------
Jason R. Thorpe                                       thorpej@nas.nasa.gov
NASA Ames Research Center                               Home: 408.866.1912
NAS: M/S 258-6                                          Work: 415.604.0935
Moffett Field, CA 94035                                Pager: 415.428.6939