Denis, thanks a lot for the reminder and the analysis here. I knew about this issue at one point -- see https://bugzilla.redhat.com/show_bug.cgi?id=1185253 -- but by now I've completely forgotten that HyperV enlightenments and UEFI SMP Win7 don't mix.
Also, for your analysis in comment #7 -- thanks for that too; I've never dug into it this deep. In the RHBZ I referenced above, there's a link to MSDN -- https://technet.microsoft.com/en-us/library/dn282285.aspx -- which indicates that the UEFI Win7 family was never meant to be run as HyperV guests. Those docs were enough explanation to me.
I don't think hacking on OVMF's VBE shim would be smart at this point -- the VBE shim is already an ugly hack to trick Win7 into working. I think the best course of action here is to disable HyperV enlightenments for Win7 UEFI guests. That's what virt-manager does as well:
Denis, thanks a lot for the reminder and the analysis here. I knew about this issue at one point -- see https:/ /bugzilla. redhat. com/show_ bug.cgi? id=1185253 -- but by now I've completely forgotten that HyperV enlightenments and UEFI SMP Win7 don't mix.
Also, for your analysis in comment #7 -- thanks for that too; I've never dug into it this deep. In the RHBZ I referenced above, there's a link to MSDN -- https:/ /technet. microsoft. com/en- us/library/ dn282285. aspx -- which indicates that the UEFI Win7 family was never meant to be run as HyperV guests. Those docs were enough explanation to me.
I don't think hacking on OVMF's VBE shim would be smart at this point -- the VBE shim is already an ugly hack to trick Win7 into working. I think the best course of action here is to disable HyperV enlightenments for Win7 UEFI guests. That's what virt-manager does as well:
https:/ /github. com/virt- manager/ virt-manager/ commit/ cbba1c4dd381
Given that this is not a QEMU issue, I'm closing the report (again).