Re: [PATCH v2] KVM: VMX: Print VM-instruction error when it may be helpful

[Sean Christopherson <seanjc@google.com>]

On Sat, May 07, 2022, Jim Mattson wrote:
> From: David Matlack <dmatlack@google.com>
> 
> Include the value of the "VM-instruction error" field from the current
> VMCS (if any) in the error message for VMCLEAR and VMPTRLD, since each
> of these instructions may result in more than one VM-instruction
> error. Previously, this field was only reported for VMWRITE errors.
> 
> Signed-off-by: David Matlack <dmatlack@google.com>
> [Rebased and refactored code; dropped the error number for INVVPID and
> INVEPT; reworded commit message.]
> Signed-off-by: Jim Mattson <jmattson@google.com>
> ---
>  arch/x86/kvm/vmx/vmx.c | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
> index d58b763df855..a25da991da07 100644
> --- a/arch/x86/kvm/vmx/vmx.c
> +++ b/arch/x86/kvm/vmx/vmx.c
> @@ -392,12 +392,14 @@ noinline void vmwrite_error(unsigned long field, unsigned long value)
>  
>  noinline void vmclear_error(struct vmcs *vmcs, u64 phys_addr)
>  {
> -	vmx_insn_failed("kvm: vmclear failed: %p/%llx\n", vmcs, phys_addr);
> +	vmx_insn_failed("kvm: vmclear failed: %p/%llx err=%d\n",

I highly doubt it will ever matter, but arguably this should be %u, not %d.

> +			vmcs, phys_addr, vmcs_read32(VM_INSTRUCTION_ERROR));
>  }
>  
>  noinline void vmptrld_error(struct vmcs *vmcs, u64 phys_addr)
>  {
> -	vmx_insn_failed("kvm: vmptrld failed: %p/%llx\n", vmcs, phys_addr);
> +	vmx_insn_failed("kvm: vmptrld failed: %p/%llx err=%d\n",
> +			vmcs, phys_addr, vmcs_read32(VM_INSTRUCTION_ERROR));
>  }
>  
>  noinline void invvpid_error(unsigned long ext, u16 vpid, gva_t gva)
> -- 
> 2.36.0.512.ge40c2bad7a-goog
>