On 09/15/2018 04:08 PM, Johannes Erdfelt wrote:
The bool type is 8-bits, but some assembly code in vmx_vcpu_run compares the __launched member as 32-bits. While the current layout of the structure leaves 24-bits of padding after __launched, adding members with small types after __launched or rearranging members could result in a "KVM: entry failed, hardware error 0x5" error. Signed-off-by: Johannes Erdfelt <johannes@xxxxxxxxxxx> --- arch/x86/kvm/vmx.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 1d26f3c..c802b25 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -10714,7 +10714,7 @@ static void __noclone vmx_vcpu_run(struct kvm_vcpu *vcpu) "mov %%" _ASM_AX", %%cr2 \n\t" "3: \n\t" /* Check if vmlaunch of vmresume is needed */ - "cmpl $0, %c[launched](%0) \n\t" + "cmpb $0, %c[launched](%0) \n\t" /* Load guest registers. Don't clobber flags. */ "mov %c[rax](%0), %%" _ASM_AX " \n\t" "mov %c[rbx](%0), %%" _ASM_BX " \n\t"
Reviewed-by: Krish Sadhukhan <krish.sadhukhan@xxxxxxxxxx>
