> Another detail it didn't occur to me to check earlier: the cache.
>
> It turns out that I only see the anomalous behaviour when the processor
> cache is enabled.  *cache off => no corruption.  *cache on => back to
> the glitch.

This is the method for turning on/off most of the Dynarec behaviour at
runtime. RPCEmu uses the cache state to control the Dynarec state, and
with the cache off reverts to code that is very similar to the
interpreter.

It is unfortunately not well documented - I will make sure it gets
added to the manual.

Matthew

_______________________________________________
Rpcemu mailing list
[email protected]
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu

Reply via email to