Comment 60 for bug 1797581

@Dmitrii - we were at the "junk in ..." in comment #21 already.
The "broken padding" is interesting but might be the same issue with a different message as the newer kernel understands slightly more about it.

That it does only occur with Xenial-HWE but on the same initrd not with the base Xenial kernel is interesting as well. I agree that this is worth the kernel task on this bug (it was the same initrd in this test right)?.

The question is if the initrd really is broken (unlikely since it works with other mem sizes) or that it is broken by qemu/seabios/... when passed to the guest (likely).
Having such a *full set* of affected files would be great.

What we'd need is that you'd copy the kernel+initrd+image+xml used somewhere that we can reach and/or attach it here if it fits - this was asked of all reporters in comment #51 already.

I have seen you have attached the initrd of your case at least in comment #56 - thanks!
I'll taken another look at the issue with that initrd that you attached, will let you know if I need more.