Boot Fiasco at ARM on QEMU

Adam Lackorzynski adam at
Wed Oct 15 22:56:45 CEST 2008

On Tue Oct 14, 2008 at 10:57:07 +0800, Tsai, Tung-Chieh wrote:
> Dear Adam,
> On Fri, Oct 10, 2008 at 5:34 AM, Adam Lackorzynski
> <adam at> wrote:
> > The DEFAULT_RELOC of hello? Was the error an overlap?
> The DEFAULT_RELOC of roottask, an overlap error as follow :
>   #05: loading "hello"
>        from [0206e000-0207b268] to [00140000-00144474]
>   Roottask: cannot load binary because address at 00140000 not free
>     for unkown reason

Ok, thanks, I had a change sitting in this Makefile for some time, I've
committed it now.

> > What do you mean 'log with all test'?
> I build with `E=log`, the entry describe in
> `.../l4/pkg/bootstrap/server/src/modules.list `
> Than I run it on qemu, execute all the test on it.
> It's why I feel strange, since the test on `E=log` passed, why does
> logserver fail to register nameserver ?

Can't say without looking further into it.

> > Does something change if you leave out programs from the end, e.g.
> > loader, bmodfs include its modules, l4io?
> I didn't see any other output about loader, bmodfs, etc ... , is there any
> other way to observe it, e.g using jdb ?

You could enable IPC logging to see why task 6 can't properly talk to
task 5. E.g. if task 5 (names) is called at all, more generally what's
a more detailed error message than what we have now.

> Roottask.
>   Command line found: "roottask task modname "bmodfs" module module
> module module module module"
>        0kB (   0MB) total RAM (reported by bootloader)
>   251332kB ( 246MB) received RAM from Sigma0
>    20696kB (  21MB) reserved RAM for RMGR

How much memory did you configure? (RAM_SIZE_MB)

Adam                 adam at

More information about the l4-hackers mailing list