Hi Adam,

I was able to locate the QEMU settings needed and now can run all of the examples except the "L4Linux-Basic" one that reports and error of "Failed to map VGA area but CONFIG_VGA_CONSOLE=y." and stops at that point.

I think that this is a similar message that I get when I try to run any of them on physical hardware Intel i5 NUC7, Intel i7 NUC13, and Dell XPS 8950 (11th Gen Intel(R) Core(TM) i7-11700 @ 2.50GHz   2.50 GHz) which reports that it cannot find a suitable video mode on all 3 systems.

The good news here is that we have made some good progress with getting most of the examples to run under QEMU which is a great start, my friend.

Thanks and have a wonderful day,
Lonnie


On 11/12/2024 5:57 PM, Adam Lackorzynski wrote:
Hi Lonnie,

please first ensure that on your NUC, KVM is properly used with QEMU.
In the conf/Makeconf.boot file there is one commented out
"QEMU_OPTIONS-amd64" line, please use this one instead of the one above.
With -enable-kvm the right virtualization extensions are provided by
QEMU such that VMs work and the error messages you're seeing shall go
away.



Adam

On Mon Nov 11, 2024 at 08:31:03 -0500, Lonnie Cumberland via l4-hackers wrote:
Hello Adam,

Yea, I found the "make help" after sending off the messages and was able to
take the latest "snapshot" (non-git) and build things, but have had only
limited success in running them on my Intel NUC13 Ubuntu 23.10 (x64) system.

I can run the "hello*.efi" versions with the "make qemu" but it seems that
the framebuffer example cannot map a VGA screen and it might be something in
the "dtc" which is the hardware settings for L4Re, if I read the
documentation correctly, but am not sure on this.

Interestingly, is that if I take the "git" version of l4re and build things
with "ham" then the framebuffer example will run under qemu with "make qemu"
to execute it. the l4linux-mag will boot up the graphics but then it just
shows a blank window and I suspect that the same errors (below) are
happening with the errors " terminate called after throwing an instance of
'N2L413Runtime_errorE'"

For bot the git and non-git builds, the VM-multi and VM-multi-p2p examples
l4re boots and I make it through SIGMA0, MOD, and NED while the Console
Server also boots up, but then I get and errors in the VM's starting:

----------------------------------------------------------
Created vcon channel: vm-1 [420010]
cons>
Created vcon channel: vm-2 [421010]
cons>
vm-1    | VMM[]: Created VCPU 0 @ 18000
vm-1    | VMM[]: >>> CAUTION: Support for AMD SVM is experimental, use at
your own risk! <<<
vm-2    | VMM[]: Created VCPU 0 @ 18000
vm-2    | VMM[]: >>> CAUTION: Support for AMD SVM is experimental, use at
your own risk! <<<
vm-2    | terminate called after throwing an instance of
'N2L413Runtime_errorE'
vm-2    |   what: Invalid request: SVM does not support next_rip save.
Aborting!
vm-2    |

Created vcon channel: vm-3 [422010]
cons>
vm-3    | VMM[]: Created VCPU 0 @ 18000
vm-3    | VMM[]: >>> CAUTION: Support for AMD SVM is experimental, use at
your own risk! <<<
vm-1    | terminate called after throwing an instance of
'N2L413Runtime_errorE'
vm-1    |   what: Invalid request: SVM does not support next_rip save.
Aborting!
vm-1    |
vm-3    | terminate called after throwing an instance of
'N2L413Runtime_errorE'
vm-3    |   what: Invalid request: SVM does not support next_rip save.
Aborting!
vm-3    |

----------------------------------------------------------

I think there is some VGA video problem somewhere that is preventing the
VM's from completely booting up for some reason.

Best and have a great day,
Lonnie

On 11/11/2024 1:45 AM, Adam Lackorzynski wrote:
Hi Lonnie,

instead of "make qemu" one can also do "make efiimage" to generate an
(U)EFI image. Please do "make help" to see the other options available.


Adam



      
_______________________________________________
l4-hackers mailing list -- l4-hackers@os.inf.tu-dresden.de
To unsubscribe send an email to l4-hackers-leave@os.inf.tu-dresden.de

Adam