On Mon, May 06, 2024 at 11:30:52AM +0800,
Binbin Wu <binbin.wu@xxxxxxxxxxxxxxx> wrote:
> > diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> > index 1b189e86a1f1..fb7597c22f31 100644
> > --- a/arch/x86/kvm/x86.c
> > +++ b/arch/x86/kvm/x86.c
> > @@ -11013,7 +11013,7 @@ static int vcpu_enter_guest(struct kvm_vcpu *vcpu)
> > if (vcpu->arch.guest_fpu.xfd_err)
> > wrmsrl(MSR_IA32_XFD_ERR, vcpu->arch.guest_fpu.xfd_err);
> > - if (unlikely(vcpu->arch.switch_db_regs)) {
> > + if (unlikely(vcpu->arch.switch_db_regs & ~KVM_DEBUGREG_AUTO_SWITCH)) {
>
> As pointed by Paolo in
> https://lore.kernel.org/lkml/ea136ac6-53cf-cdc5-a741-acfb437819b1@xxxxxxxxxx/
> KVM_DEBUGREG_BP_ENABLED could be set in vcpu->arch.switch_db_regs, by
> userspace
> kvm_vcpu_ioctl_x86_set_debugregs() --> kvm_update_dr7()
>
> So it should be fixed as:
>
> - if (unlikely(vcpu->arch.switch_db_regs)) {
> + if (unlikely(vcpu->arch.switch_db_regs &&
> + !(vcpu->arch.switch_db_regs &
> KVM_DEBUGREG_AUTO_SWITCH))) {
Yes, that's true. Thanks for catching this.
--
Isaku Yamahata <isaku.yamahata@xxxxxxxxx>
