I'll have to start my document about a virtual hardware to make breaking the
bochs plug-in happen over again from scratch then.
By the way, what is the current method used here to pass data between the
various relatively independent parts of the VM right now (I don't have my
UNIX box where I am right now, and probably didn't look close enough when I
did have it.....)? I'd hope to God that we aren't using globals for all of
this too.... In any case, this would help in my final (I hope) reworking of
my plug-in/software machine structure proposal.
Also, while its on my mind, have we implemented any PCI/ISA/etc. handling
yet (or does our machine just pretend it is bus-less right now)? (I had
planned my proposal around a pseudo PCI bus structure/specification--and
that is why I'd like to know.)
Drew Northup, N1XIM
> -----Original Message-----
> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf
> Of Kevin Lawton
> Sent: Wednesday, January 03, 2001 8:17 AM
> To: [EMAIL PROTECTED]
> Subject: [plex86] C++ use/performance in plex86
>
>
> Charles Levert wrote:
> >
> > The Linux Kernel Mailing List FAQ already addresses C vs C++. Linux
> > and Plex86 are two, but then again, both are system programming.
> > Reading that part of the FAQ is instructive in the context of this
> > discussion.
> >
> > http://www.tux.org/lkml/index.html#s1-4
>
> I added an entry in the FAQ, that C is the language to program
> all new plex86 code. I also added this URL.
>
> Not only do I fully agree with the decision to keep the Linux kernel
> in C, but it is important that we keep at least certain parts of our
> virtualization in C, too to be compatible with the Linux kernel and
> other C-based kernels. This is because, for purposes of improving
> performance, it is quite conceivable we may move parts of certain
> functionality into monitor space. This will prevent the long distance
> context switches back to user space and then back again to the monitor.
>
> The C++ plugin code that we have, is OK for now as it has accelerated
> the coding progress (bochs + write-cache). But from now on, all
> new code should be in C, for the above reasons.
>
> To be stated another way, no new C++ code is going to be committed
> to CVS. Thus further discussions would be futile.
>
> -Kevin
>
>
>
> --
> =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
> Kevin Lawton [EMAIL PROTECTED]
> MandrakeSoft, Inc. Plex86 developer
> http://www.linux-mandrake.com/ http://www.plex86.org/
>
