Re: Kernel panics (in QEMU) when booting with GRUB2 + OVMF + QEMU.

[Jordan Uggla]

On Fri, Aug 14, 2015 at 11:29 PM, Andrei Borzenkov <address@hidden> wrote:
> On 14.08.2015 09:45, Mark Lee wrote:
>>
>> To all,
>>
>> I've encountered a bug when using GRUB2 + OVMF + QEMU on Arch Linux 64
>> bit.
>> The issue can be reproduced when running the latest git code from grub as
>> well.
>>
>> ISSUE: Kernel panics (in QEMU) when booting with GRUB2 + OVMF + QEMU.
>
>
> Once kernel is loaded, GRUB task is over. Why do you think it is GRUB issue?

I worked with Mark in #grub and there were symptoms that Mark has not
described here that lead me to believe that there is either a grub or
OVMF bug, and that is why I asked him to post here.

First Mark demonstrated that the same image, with the same version of
grub and same grub.cfg, was succeeding when booted via BIOS with qemu,
but not when booted via OVMF with qemu, and that the particular error
message before the kernel panic was indicative of the initramfs not
having been loaded. Having largely ruled out misconfiguration I asked
Mark to try running the "linux" and "initrd" commands manually at the
grub shell to see if there were any error messages being displayed.

Mark did this and replied that while the linux command succeeded, the
initrd command failed, and in a way that clearly indicates a bug
somewhere. Specifically, while Mark entered something like "initrd
/boot/initramfs-foo-bar", and everything he had typed was echoed back
to him on screen as expected, the error message he received was (if I
recall correctly) something about the file "/boot/initra" not being
found. In other words, the error message indicated that somewhere the
file name entered had been truncated. I suspect that this happened
when grub was interpreting the grub.cfg as well, and thus the linux
command succeeded but the initramfs command failed, and while grub
still tried to boot, the missing initramfs resulted in the kernel
panic.This, along with other symptoms from previous tries, seems to
indicate to me that there is memory corruption somewhere, either due
to a bug in grub or in OVMF.

I don't have access to the IRC logs of our conversation today (but I
will have access to them Monday) so I'm going by my memory alone.
Hopefully Mark can give the exact error messages he received from
trying to enter the initramfs command directly, or possibly even
reproduce that test and post screenshots or serial output logs of it.
I have not had a chance to try to reproduce this myself, and I don't
know when I will have time to do so.

-- 
Jordan Uggla (Jordan_U on irc.freenode.net)