14 Jan
2015
14 Jan
'15
2:08 p.m.
On 2015-01-14 17:55, Pontus Pihlgren wrote:
On Wed, Jan 14, 2015 at 09:21:32AM -0500, Noel Chiappa
wrote:
Do you mean to say that bytes are 8-bit wide on PDP-11? Why
would PDP-10 be so much harder?
There are lots of things that becomes complex and/or odd when you are on
a machine with a word size that is not evenly divisible by 8.
You need to manipulate bitfields inside a word, and depending on how
things work, you might either span word boundaries, placing half of an
octet in one word and the other half in the second, or else have four
extra bits sitting around that you should ignore in every word.
All arithmetic are done on 16- or 32-bit quantities, which you must take
into consideration. The IP checksum is an odd beast that do ones
complement addition. So spillover from the 16-bit additions are added
back at the bottom.
Everything is of course doable, but there will be a whole lot of bit
fiddling and juggling doing TCP/IP on a PDP-10.
Johnny
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt at softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol
From:
Johnny Billquist
But for people who actually wants to read several
hundred K of
MACRO-11... Here is your chance... :-)
Ah. Was going to ask what the TCP/IP itself was written in. Well, at least
it's an 8-bit machine - think about writing a TCP/IP in PDP-10 MACRO... :-)
Noel