From owner-cvs-sys  Thu May  1 13:27:07 1997
Return-Path: <owner-cvs-sys>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.5/8.8.5) id NAA14298
          for cvs-sys-outgoing; Thu, 1 May 1997 13:27:07 -0700 (PDT)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by hub.freebsd.org (8.8.5/8.8.5) with ESMTP id NAA14292;
          Thu, 1 May 1997 13:27:02 -0700 (PDT)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.8.5/8.6.9) id GAA24197; Fri, 2 May 1997 06:23:21 +1000
Date: Fri, 2 May 1997 06:23:21 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199705012023.GAA24197@godzilla.zeta.org.au>
To: cvs-all@freebsd.org, CVS-committers@freebsd.org, cvs-sys@freebsd.org,
        sos@freefall.FreeBSD.ORG
Subject: Re: cvs commit:  src/sys/scsi sd.c
Sender: owner-cvs-sys@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

>  Modified:    sys/scsi  sd.c
>  Log:
>  Remove old code in #if 0, leftovers from the days before != 512byte
>  sector support.

Please put it back.  It shows more or less the correct way to do things,
which is to let dscheck() do all the checking for all disks.

The != 512byte sector support breaks at least EOF handling.  (dscheck()
rewrites both bp-b_resid and bp->b_bcount for transfers that cross the
end of the partition, but the != 512byte sector support does extra work
to prevent the change to bp->b_bcount.  I think it does this prevent
truncation giving a count that isn't a multiple of sec_blk_ratio.  I think
the count is always a multiple except for misconfigured partitions.
Partition sizes currently need to be multiples of sec_blk_ratio to
prevent this.  Partition sizes should really be in units of sectors.)

Bruce