Re: In-kernel units for block numbers, etc ...

[Robert Elz <kre%munnari.OZ.AU@localhost>]

    Date:        Sun, 29 Nov 2015 10:47:17 +0000 (UTC)
    From:        mlelstv%serpens.de@localhost (Michael van Elst)
    Message-ID:  <n3el3k$mlr$1%serpens.de@localhost>

  | Maybe this:
  | http://ftp.netbsd.org/pub/NetBSD/misc/mlelstv/vnd.c.diff

Looks reasonable ... I'll see if I can find time to test it in the next
12 hours or so...   If you haven't heard a result by then (maybe 14 hours
from now) then you won't until sometime in Jan (from me anyway).

And while the label behaviour might be expected in some sense, there was
clearly a divergence in operation - 2 of the three vnd tests I set up behaved
exactly as I expected them to.  The only one with the odd effect was the
5121 byte unholy file (on a 4K sector drive).  My guess is that the driver
does as you documented (in all cases), but vndconfig writes a more rational
label before the driver generated version can ever be seen by humans.

Except a label sector write on a 512 byte/sec vnd, on a 4K sector drive,
would fail, as it is just 1 sector.   Hence, the driver version became visible.
My guess is that with the change you provided above, this problem will just
vanish.

kre