Subject: Re: The obvious inherent superiority of SVR4 device naming...
To: Peter Seebach <firstname.lastname@example.org>
From: Ted Lemon <email@example.com>
Date: 12/21/1994 23:54:26
Let me see if I can spell this out. In the berkeley naming scheme, a
disk device would be referred to as xxyz, where xx is the device type
(e.g., sd), y is the unit number (e.g., 0) and z is the partition
(i.e., a). The partition numbering is completely unambiguous.
Partition a is the first partition in the table. Partition b is the
second. Partition c is the third, and is traditionally sized to
cover the whole disk. Some implementations enforce this as a rule.
Partition d is the fourth, e the fifth, f the sixth and g the seventh
entry in the partition table..
What I don't understand is how this materially differs from the
naming scheme you're describing. The only difference I see is that
we're using letters, not numbers. The name of a partition does not
depend on its function. Where are you getting this notion that root
and swap always have specific names? Sure, traditionally root is
partition a, but I've used systems where I had a spare root on b which
I could boot.
> p.s.: I also liked that SCSI device 5 was device 5. Always. I don't like
> having sd0, cd0, and st0 - in fact, this confused me and made it difficult
> for me to install, because unit 5 was sd1. ???
This I agree with. I don't understand the motivation for the
numbering scheme that I hear is being used in the 386 port. To me it
seems fraught with danger. I'll probably hack it out of my kernels at
some point. Fortunately for me, the pmax port doesn't currently
operate this way.
Ted Lemon firstname.lastname@example.org
+1 415 477 5045
Fight to preserve your freedom to program: Join the League for
Programming Freedom! For info, contact email@example.com.