On 06/10/20 21:06, Krish Sadhukhan wrote: > v2 -> v3: > Patch# 2: The local variable "nested_vmcb_lma" in > nested_vmcb_check_cr3_cr4() has been removed. > Patch# 3: Commit message has been enhanced to explain what the test > is doing and why, when testing the 1-setting of the > non-MBZ-reserved bits. > Also, the test for legacy-PAE mode has been added. Commit > header reflects this addition. > > > [PATCH 1/4 v3] KVM: nSVM: CR3 MBZ bits are only 63:52 > [PATCH 2/4 v3] KVM: nSVM: Add check for reserved bits for CR3, CR4, DR6, > [PATCH 3/4 v3] nSVM: Test non-MBZ reserved bits in CR3 in long mode and > [PATCH 4/4 v3] KVM: nSVM: nested_vmcb_checks() needs to check all bits > > arch/x86/kvm/svm/nested.c | 52 ++++++++++++++++++++++++++--------------------- > arch/x86/kvm/svm/svm.h | 2 +- > 2 files changed, 30 insertions(+), 24 deletions(-) > > Krish Sadhukhan (3): > KVM: nSVM: CR3 MBZ bits are only 63:52 > KVM: nSVM: Add check for reserved bits for CR3, CR4, DR6, DR7 and EFER to svm_set_nested_state() > KVM: nSVM: nested_vmcb_checks() needs to check all bits of EFER > > x86/svm.h | 4 +++- > x86/svm_tests.c | 66 +++++++++++++++++++++++++++++++++++++++++++++++++++------ > 2 files changed, 63 insertions(+), 7 deletions(-) > > Krish Sadhukhan (1): > nSVM: Test non-MBZ reserved bits in CR3 in long mode and legacy PAE mode > Queued, but I don't really like the duplication in patch 2 so I'll probably punt it to 5.11 and fix it up. Paolo
