[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[f-cpu] Re: F-ROFS

On Wed, Mar 13, 2002 at 12:14:42AM +0100, Yann Guidon wrote:
>  - We have to simulate our system using a VHDL simulator
> which is by definition very limited from the I/O point of
> view : file open, close, read and write, that's all.
> not even a file seek. All I/O will have to go through that,
> before we're able to simulate more than a character input/output
> (a virtual console). That's very poor : no HDD or video frame
> buffer, almost nothing.

I fail to see the problem of not having a video frame buffer.  Unless
you want to run X11 in the simulation, which I doubt.  What's so
important about video?

All you need is a simulated serial port and then put the Linux console
on that.  Serial consoles are well supported and VHDL I/O is sufficient
for that.  Hook up an xterm or Linux text console to that I/O and you
can even run programs needing a powerful terminal.

HDD also isn't important, they aren't anything special and there are ram

> So i started thinking about a really simple "Read-Only File System"
> (not exactly a RomFS but you get the idea). This must be handled
> with simple structures and algorithms but allow the direct
> execution of binaries, allow the loading of drivers and giver some
> functionalities that allow a kernel to be configured and started.

Isn't this way too complicated at this stage?  You need to define the
startup protocol for Linux (where are the args, where is the initrd,
what is the CPU state), then you should write a simple loader that just
implements that and boots one kernel from a fixed ROM address with fixed
args (like none).

That should do for some months, until you get a kernel to boot to the
point where it wants to mount a root fs and exec init.

Andreas Bombe <bombe@informatik.tu-muenchen.de>    DSA key 0x04880A44
To unsubscribe, send an e-mail to majordomo@seul.org with
unsubscribe f-cpu       in the body. http://f-cpu.seul.org/