KVM: nVMX: Do not expose MPX VMX controls when guest MPX disabled
authorLiran Alon <liran.alon@oracle.com>
Fri, 14 Sep 2018 00:25:52 +0000 (03:25 +0300)
committerPaolo Bonzini <pbonzini@redhat.com>
Mon, 1 Oct 2018 13:40:57 +0000 (15:40 +0200)
commit5f76f6f5ff96587af5acd5930f7d9fea81e0d1a8
tree2bf618ad8060f0907b34aa17eabdec37cfc74b7a
parent4679b61f2640e07724507fcd9dc070517bc6e6cb
KVM: nVMX: Do not expose MPX VMX controls when guest MPX disabled

Before this commit, KVM exposes MPX VMX controls to L1 guest only based
on if KVM and host processor supports MPX virtualization.
However, these controls should be exposed to guest only in case guest
vCPU supports MPX.

Without this change, a L1 guest running with kernel which don't have
commit 691bd4340bef ("kvm: vmx: allow host to access guest
MSR_IA32_BNDCFGS") asserts in QEMU on the following:
qemu-kvm: error: failed to set MSR 0xd90 to 0x0
qemu-kvm: .../qemu-2.10.0/target/i386/kvm.c:1801 kvm_put_msrs:
Assertion 'ret == cpu->kvm_msr_buf->nmsrs failed'
This is because L1 KVM kvm_init_msr_list() will see that
vmx_mpx_supported() (As it only checks MPX VMX controls support) and
therefore KVM_GET_MSR_INDEX_LIST IOCTL will include MSR_IA32_BNDCFGS.
However, later when L1 will attempt to set this MSR via KVM_SET_MSRS
IOCTL, it will fail because !guest_cpuid_has_mpx(vcpu).

Therefore, fix the issue by exposing MPX VMX controls to L1 guest only
when vCPU supports MPX.

Fixes: 36be0b9deb23 ("KVM: x86: Add nested virtualization support for MPX")

Reported-by: Eyal Moscovici <eyal.moscovici@oracle.com>
Reviewed-by: Nikita Leshchenko <nikita.leshchenko@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Signed-off-by: Liran Alon <liran.alon@oracle.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
arch/x86/kvm/vmx.c