Port-sparc archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Magma bizarrity?



On Wed, Feb 25, 2009 at 04:01:51AM -0500, der Mouse wrote:
Thanks to a fellow Sun fan's generosity, I just acquired a Magma
16-port serial card.  It's got a breakout box which mounts directly on
the back of the card, with 16 8P8C jacks on it.  I've worked out the
pinout of ecah port, and it looks a little odd, but that's not a
problem since I'm going to be doing my own cabling anyway.

Magma pinouts:
http://www.magma.com/support/faqs/sbus/sun.htm

But there's a stranger surprise lurking.  The ports on the breakout box
are arranged in a 4x4 grid and marked with the numbers 0 through 15 in
an obvious way:

0  1  2  3
4  5  6  7
8  9 10 11
12 13 14 15

But the arrangement of the ports in terms of minor device number,
rather than of the printing on the breakout box, is

14  5  0 10
13  4  8  9
12  7  3 11
15  6  1  2

This looks totally bizarre.  I looked at the code, thinking maybe it
was some kind of race which port got which tty unit, but the code is
saner than that; port numbering is based on chip and channel numbers,
not something quasi-random like scheduling of kernel threads.  (Also,
the arrangement does not change with a reboot, though that's a fairly
weak test.)

I assume you used MAKEDEV under netbsd to make the device special
files? How has that file changed since 1.4T and -current?

Cheers,
Jonathan.


Home | Main Index | Thread Index | Old Index