From mboxrd@z Thu Jan 1 00:00:00 1970 From: lehmann@ans-netz.de (Oliver Lehmann) Date: Sat, 07 Nov 2015 23:13:38 +0100 Subject: [TUHS] System III - TCP/IP In-Reply-To: References: <20151107200358.Horde.M1lYZZyTC4t0Qb8KrygKzhy@avocado.salatschuessel.net> <20151107192043.GA11895@mcvoy.com> Message-ID: <20151107231338.Horde.d8EIgIMMf9VYCExfovc2K_M@avocado.salatschuessel.net> Clem Cole wrote: > 0.) adding pty's is pretty easy. Just grab the BSD code and start > hacking. Grab the original version of Steven's book - he will talk to you > through it in detail. It's a two drivers, the control (ptyX) and the > slave (ttypX) . Having them in your system is handy anyway for a number > of things besides networking (like the expect(1) program). Yeah.. having pty would be good anyway but the C compiler is so limited that it would be just some sort of rewrite instead of "porting"... It is basically a pcc as of 1981 with whatever Zilog hacked into it additionally. I once tried to get a current pcc onto the system but... yeah... I guess I lack skill ;) - the new PCC would need to create Z8001 ASM code.... something I lack skill. - an optimizer... haha... no way I could even optimize Z8001 ASM code by hand ;) - if a new linker is needed - how to create Zilogs s.out format... - I guess I would need to recompile the whole kernel with this new compiler to have every object work "together" - but I still lack some sources (most of them I "retranslated" from disassembled object files to C code - but 2 or 3 are just are too hard to retranslate) I also had a look at the C-Compiler which comes with Plexis SYSIII (which is available as source somewhere in the WWW) but this is a compiler capable of non-segmented executables (one 64K segment adressable) only but I need a Compiler creating segmented executables (128 64k segments accessable = 8MB address space) So I'm stuck with this '81ish C compiler - ok - but who said that it would be an easy task ;) > 4.) It strikes me that it would be easiest of all to just take something > like a $20-$30 "Wiznet" 5100 board from Sparkfun and interface it to the > system either via serial or better if you have a parallel port. Then run > the IP/TCP stack externally. This is what we do for Arduino's and other > small micro controllers - the board is well documents and easy to talk > too. They have both RJ45 style and wireless too. I have a hand full of > them here and am pretty happy with them for what they provide. I was thinking about the same. I have connectors on the board where I have access to the System complete bus. They where used for things like a RTC, system memory, Tape controller and so on. So I could build up a board using Wiznet and thats it. So no need of a PIO or SIO which would things kinda slow anyway... Wiznet would be of course faster as the Z8001 would do in any way. But what comes to my mind with that solution... I would need a kernel part which interacts with the wiznet chip (driver). I can not directly access the wiznet from userland which would be non privileged code which is not allowed to execute privileged instructions (raises a trap). On the userland side... I probably have to "rewrite" all the stuff which comes with BSD and uses libnet. Pr is there something like a compatibility layer which provides a library looking like a libnet API but internally "redirects" everything to the wiznet (via the to-be-written-kernel-driver) so I wouldn't need to reinvent all the userland tools as well? Oliver