Re: Mopd which works with ELF and a.out in Linux and can netboot ULTRIX

[Johnny Billquist <bqt%update.uu.se@localhost>]

Why did you start from the Linux mopd?
The NetBSD mopd is a different thing than whatever you used under Linux.

I think you started from the totally wrong thing here.

  Johnny

On 2020-10-27 04:07, Boris Gjenero wrote:
> My last message was about problems trying to netboot a MicroVAX II. One 
> of the problems was the Linux port of mopd not understanding a.out files 
> and serving them improperly. Back then I did something quick and dirty 
> to get it to work. Afterwards I wanted to do something cleaner. Here it is:
> https://github.com/dreamlayers/netbsd-mopd
>
> It is based on the Linux port which is already in NetBSD CVS, but 
> without ELF and a.out support. For ELF and a.out support, it is using 
> the same strategy, building on Linux with header files from NetBSD. 
> Though in this case it is based on current NetBSD code, and supports 
> ELF, 4k page size a.out and old 1k page size a.out. It also supports 
> a.out files with mid == 0, for Athena-4.3BSD, ULTRIX and possibly other 
> operating systems.
>
> After I got that working I took a look at how to netboot ULTRIX using 
> this mopd. After a few additions, I've successfully netbooted ULTRIX 
> 4.00 on KA630, KA650 and VAXstation 3100 m38. For more information on 
> that see:
> https://github.com/dreamlayers/netbsd-mopd/tree/dreamlayers/ultrix
>
> Yes, that includes the KA630 I was having problems with. One of the 
> problems was really silly: I had the back panel switch in the loopback 
> position, which connects the CPU's console serial output to input. 
> Though even with that fixed, the versions of NetBSD boot loader which 
> output to QDSS don't allow input, either from QDSS or serial. Maybe the 
> wrong firmware routine is being used for input?
>
> Probably I now need to look at how to get some of this work included in 
> NetBSD. At least the fixes to what's already there seem like a good 
> idea. One issue is wrong pfInit() parameter, preventing mopd from 
> working in Linux at all:
> https://github.com/dreamlayers/netbsd-mopd/commit/b94f02f307a3127c20581d8d40bce8ad0ba3b89e 
>
> Also, it seems like mopd was never updated to 4k page size, which is 
> however inconsequential with OMAGIC boot:
> https://github.com/dreamlayers/netbsd-mopd/commit/7b0f60bf5454f2742890f9815ebe56f965f6fe25 
>
>
> One last thing. Be careful about the NiCD batteries in VAXen. So far 3 
> out of 4 were leaking. It's not obvious, because the battery is covered 
> by shrink wrap. The most obvious external sign is corrosion at the 
> battery connector, maybe because the leaking liquid wicks or gets pushed 
> along inside the insulation of the wires. After removing the battery 
> from the holder, more leakage is found on the sides, where gaps in the 
> shrink wrap were covered by the holder. My damage is confined to minor 
> corrosion of the battery connector pins, but others have reported worse. 
> If you need more motivation to take old batteries out of your VAXen, 
> read 
> https://sites.google.com/site/dalbydatormuseum/computers/digital-equipment-corporation/microvax2000 


--
Johnny Billquist                  || "I'm on a bus
                                  ||  on a psychedelic trip
email: bqt%softjar.se@localhost             ||  Reading murder books
pdp is alive!                     ||  tryin' to stay hip" - B. Idol