Re: [PATCH] x86/CPU/AMD, mm: Extend with mem_encrypt=sme option

[Brijesh Singh]

On 9/30/17 4:41 PM, Borislav Petkov wrote:

...
>> 3.2 If bit is set, its SEV guest. We set sev_enabled to 'true' and also
>> set 'sme_me_mask'. Return from the function.
>> The SEV state *cannot* be controlled by a command line option.
> So how do you propose to disable SEV? Right now I do:
>
>         if (feature_mask == AMD_SEV_BIT)
>                 sev_enabled = true;

Host OS:
--------
We can control the SEV feature visibility to the hypervisor through the
host OS command line.

mem_encrypt=on -ÂÂ SEV feature is available
mem_encrypt=sme - SEV feature is disabled Â
mem_encrypt =off -Â SEV feature is disabled

When SEV feature is disabled, KVM will not be able to launch any SEV
guests. When SEV support is available, KVM can enable it in a specific
VM by setting SEV bit before executing the VMRUN instruction.

Guest OS:
--------
Checks the MSR_AMD64_SEV to determine if SEV feature is enabled. Please
note that the MSR is a read-only. IOW, MSR is not intercepted by the
hypervisor.

Currently, mem_encrypt=xxx and CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT
is don't care. We can not depend on the command line because when SEV is
enabled in a VM then instruction fetch will be decrypted by the
hardware. If we want then we can perform the comparison between the SEV
state obtained through MSR with user supplied command line and trigger
BUG() if they don't match.


> at the end, when mem_encrypt=sme wasn't supplied on the cmdline. IOW,
> SEV is enabled either when CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT or
> mem_encrypt=on.
>
> Hmmm?
>