Genode/NOVA+Multiple VMMs Seoul / VBox

[Roger Ferreira]

Hi Alex,
I understand the 2 options.
But I still have more doubt / questions.
In the multiboot scheme, I tried to replace the bzImage with the newer TinyCore 6.3 vmlinuz together with the initrd core.gz.
But that didn´t work.
In thesis, I just replaced the kernel, the bzImage by the vmlinuz.
I even tried to compare bzImage and vmlinuz files using xxd and od -xc. 
vmlinuz begins with MZ.
I get the following error message I found on munich source code:  ERROR(-11, ~mbi->flags & MBI_FLAG_MODS, "module flag missing");
I don´t have any idea how to debug this to find out what is wrong.
I got the munich source code thinking to place some PRINTF stuff, but I don´t know how to make it.
I am using Ubuntu 12.04. ----------
Then I tried the 2nd option.
I create a raw disk using qemu-img, than I use dd to format it from a iso.
But now I don´t know how to configure that in the run script.
I tried in the <multiboot> section without success.
I believe it should be in another place.
Any hint? How to find documentation regarding that?
Thanks for your help and patiente.
Regards,            Roger
> Subject: Re: Genode/NOVA+Multiple VMMs Seoul / VBox
> To: genode-main at lists.sourceforge.net
> From: alexander.boettcher at ...1...
> Date: Sat, 5 Sep 2015 22:25:44 +0200
> 
> Hello,
> 
> On 04.09.2015 15:51, Roger Ferreira wrote:
> > I am not able to capture any log. 
> 
> I fear without any log it will become hard to impossible to get it running.
> 
> > I am using a normal x86 desktop computer.
> 
> You should, for example, obtain a PCI serial card and attach it - if
> your machine hasn't already a serial connector on-board or something
> like Intel AMT SOL (SerialOverLAN).
> 
> > Regarding the seoul multiboot scheme, I saw it does not allow a ISO, correct?
> 
> Yes.
> 
> > It seems to expect a bootloader (munich), them the bzImage, an some aditional g
> > Actually I have prepared a custom remasterized version of TinyCore with some specific libs / apps.
> > It works alone.
> > But when I tried to port to seoul, sitill using munich, I don´t have a bzImage.
> > I have a vmlinuz, core.gz and my own stuff as TCZ extensions.
> > The vmlinuz I customize some files.
> 
> For Seoul you have two ways to boot things - either boot a multiboot
> compliant kernel, which Linux is not, or boot a VM from a raw disk image.
> 
> Munich (as a multiboot kernel) is a small helper to bootstrap a Linux
> kernel. Munich expects as first multiboot image the Linux kernel and the
> second multiboot image has to be the initial ram disk. (see
> http://os.inf.tu-dresden.de/~kauer/oslo/README for pointers to munich)
> 
> For the Seoul/Tinycore demo we had to manually squash the core.gz and
> all the Firefox browser related files into a initial ram disk
> (tc-browser.gz). I fear we didn't documented it well. As far as I
> remember it was no fun. Could be - because we don't create for Seoul VMs
> regularly. So, we have no ready to use work-flow we could share. Setting
> up a VM with Virtualbox is - in that regard - much more user-friendly.
> 
> The other option of course is to install your intended VM setup on a
> disk - or in a VM on a virtual disk, e.g. use Virtualbox on your
> Linux/Windows. Finally use the raw disk image for Seoul - there are ways
> to convert a vdi/vmdk image into a raw disk image. A hybrid iso/usb
> bootable image should also work in principal as raw disk image - however
> never tried.
> 
> Just a note - you may need several iterations of Linux kernel
> configuration tweaking and rebuilds until you may get it running in
> Seoul. Seoul was/is more or less a research VMM and does not support
> everything out-of-a-box what a standard Linux distribution kernel
> enables/expects from the hardware.
> 
> Regards,
> 
> Alex.
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> genode-main mailing list
> genode-main at lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/genode-main
 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.genode.org/pipermail/users/attachments/20150906/2ae17559/attachment.html>