x86/sev: Ensure that RMP table fixups are reserved

Bandan Dasrequested to merge
bdas1/centos-stream-9:RHEL-68360 into main

JIRA: https://issues.redhat.com/browse/RHEL-68360

AMD reports that a SEV-SNP host crash can occur with fairly heavy load and repeatedly rebooting multiple k8 pods in a confidential containers setup. The most likely cause is that some PFN ranges can cause violations if they aren't marked reserved and the kernel tries to use them.

Signed-off-by: Bandan Das bsd@redhat.com

Edited by Bandan Das

Merge request reports