Re: [Qemu-devel] New machine?

[Peter Maydell]

On 28 January 2011 17:00, Comicinker <address@hidden> wrote:
> I want to boot a embedded linux environment in qemu, namely
> xxsvideo [1]. It is for an arm based fujitsu development board [2].

I'm afraid there is no support in qemu for this hardware.
In general ARM based boards are all very different from one
another -- they have different devices, memory is in different
places and they have different CPUs. You typically have to
build a kernel configured to support the specific device you
are using, and that kernel will then not run on any other hardware.
(This is different from x86, where PCs are sufficiently similar
that a generic x86 kernel will work on all of them.)

> I tried to boot the Kernel with the following command:
>
> qemu-system-arm -kernel uImage-xxsvideo-jadeevalkit-0.8.5.bin -M integratorcp 
> -nographic

This fails because you've tried to boot an image for one ARM
board (XXSvideo) on a model of a completely different board
(Integrator CP). This is as unlikely to work as if you took a
real hardware Integrator and tried to run your image on it.

> Do I have to patch qemu to support a new machine?

Yes, you would have to implement support for this board,
and also implement models of any devices which qemu
doesn't already have (probably most of them). In particular
if you are interested in a working model of the graphics
chipset then this is a very large amount of work and might
even be impossible if the chipset vendor doesn't release
enough documentation on it. As a rough guide, the code
to implement (most of) the OMAP3 chipset support in
qemu is about 29,000 lines of code.

Sorry I don't have a more encouraging answer.

-- PMM