rpcemu on Genode (inital rants)

English is not my native language

You’re so good it’s difficult to believe.
Better than a lot of English people I work with.

Were you planning on a simple HowTo set this up document, others might join in the experiment. I’m sure many people have old/unused machines lying around.

Were you planning on a simple HowTo set this up document

I plan to do a demo image, and patches for RPCEMu as soon as I have keyboard working.

A question :

I am trying to use RO5 (i can’t distribute with RO4). I need to get HostFS running.. however despite the roms gets loaded, I only get rpcemuhostfsfiler at the end..

HostFS works, good to keep track of hacks, had swi handler removed :)

But RO5 isn’t working. The suggested stable rom fails directly after RISC OS printout, I can’t get it to use HostFS as default FS.

The newer one on this site , gives me a grey screen and nothing else.

RO5 is now working , however after running !Boot it sometimes stop reacting to keys.
But I managed to run benchmark , and CPU is 200 steps faster than same software in linux :)
~1100% compared to 800%
But really , most of the things I hacked in is very hacky. Timing is not so good.

good to keep track of hacks, had swi handler removed :)

Been there, done that, read the scribbled notes on that bit paper that fell down the back of the desk…

RO5 is now working , however after running !Boot it sometimes stop reacting to keys.

Silly question but can you slow down the CPU in the emulator? Then see how stable it is.

Silly question but can you slow down the CPU in the emulator? Then see how stable it is.

Probably will need to do that , or recode the system emulation to handle fast cpu’s. The goal is to move away from RiscPC architecture,

Probably will need to do that , or recode the system emulation to handle fast cpu’s.

Years back I came across an issue with keyboard events on Windows 3.11 – when the machine was fast enough (CPU) the key/mouse event happened before the handler was expecting it and it then sat there indefinitely waiting for the key/mouse event that had already happened.
A rewrite of the driver software fixed it (after a week or so of MS and Apricot investigation)

I just wondered whether you had hit something similar.

The goal is to move away from RiscPC architecture,

I guessed that – rather like the generic platform I mentioned a little while back, where the underlying hardware capability is passed through to the OS and applications?