On Sun, Apr 07, 2024 at 06:52:44PM +0800,
Binbin Wu <binbin.wu@xxxxxxxxxxxxxxx> wrote:
>
>
> On 2/26/2024 4:26 PM, isaku.yamahata@xxxxxxxxx wrote:
> > From: Isaku Yamahata <isaku.yamahata@xxxxxxxxx>
> >
> > Add a flag, KVM_DEBUGREG_AUTO_SWITCHED_GUEST, to skip saving/restoring DRs
> > irrespective of any other flags.
>
> Here "irrespective of any other flags" sounds like other flags will be
> ignored if KVM_DEBUGREG_AUTO_SWITCHED_GUEST is set.
> But the code below doesn't align with it.
Sure, let's update the commit message.
> > TDX-SEAM unconditionally saves and
> > restores guest DRs and reset to architectural INIT state on TD exit.
> > So, KVM needs to save host DRs before TD enter without restoring guest DRs
> > and restore host DRs after TD exit.
> >
> > Opportunistically convert the KVM_DEBUGREG_* definitions to use BIT().
> >
> > Reported-by: Xiaoyao Li <xiaoyao.li@xxxxxxxxx>
> > Signed-off-by: Sean Christopherson <sean.j.christopherson@xxxxxxxxx>
> > Co-developed-by: Chao Gao <chao.gao@xxxxxxxxx>
> > Signed-off-by: Chao Gao <chao.gao@xxxxxxxxx>
> > Signed-off-by: Isaku Yamahata <isaku.yamahata@xxxxxxxxx>
> > ---
> > arch/x86/include/asm/kvm_host.h | 10 ++++++++--
> > arch/x86/kvm/vmx/tdx.c | 1 +
> > arch/x86/kvm/x86.c | 11 ++++++++---
> > 3 files changed, 17 insertions(+), 5 deletions(-)
> >
> > diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
> > index 3ab85c3d86ee..a9df898c6fbd 100644
> > --- a/arch/x86/include/asm/kvm_host.h
> > +++ b/arch/x86/include/asm/kvm_host.h
> > @@ -610,8 +610,14 @@ struct kvm_pmu {
> > struct kvm_pmu_ops;
> > enum {
> > - KVM_DEBUGREG_BP_ENABLED = 1,
> > - KVM_DEBUGREG_WONT_EXIT = 2,
> > + KVM_DEBUGREG_BP_ENABLED = BIT(0),
> > + KVM_DEBUGREG_WONT_EXIT = BIT(1),
> > + /*
> > + * Guest debug registers (DR0-3 and DR6) are saved/restored by hardware
> > + * on exit from or enter to guest. KVM needn't switch them. Because DR7
> > + * is cleared on exit from guest, DR7 need to be saved/restored.
> > + */
> > + KVM_DEBUGREG_AUTO_SWITCH = BIT(2),
> > };
> > struct kvm_mtrr_range {
> > diff --git a/arch/x86/kvm/vmx/tdx.c b/arch/x86/kvm/vmx/tdx.c
> > index 7aa9188f384d..ab7403a19c5d 100644
> > --- a/arch/x86/kvm/vmx/tdx.c
> > +++ b/arch/x86/kvm/vmx/tdx.c
> > @@ -586,6 +586,7 @@ int tdx_vcpu_create(struct kvm_vcpu *vcpu)
> > vcpu->arch.efer = EFER_SCE | EFER_LME | EFER_LMA | EFER_NX;
> > + vcpu->arch.switch_db_regs = KVM_DEBUGREG_AUTO_SWITCH;
> > vcpu->arch.cr0_guest_owned_bits = -1ul;
> > vcpu->arch.cr4_guest_owned_bits = -1ul;
> > 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)) {
> > set_debugreg(0, 7);
> > set_debugreg(vcpu->arch.eff_db[0], 0);
> > set_debugreg(vcpu->arch.eff_db[1], 1);
> > @@ -11059,6 +11059,7 @@ static int vcpu_enter_guest(struct kvm_vcpu *vcpu)
> > */
> > if (unlikely(vcpu->arch.switch_db_regs & KVM_DEBUGREG_WONT_EXIT)) {
> > WARN_ON(vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP);
> > + WARN_ON(vcpu->arch.switch_db_regs & KVM_DEBUGREG_AUTO_SWITCH);
> > static_call(kvm_x86_sync_dirty_debug_regs)(vcpu);
> > kvm_update_dr0123(vcpu);
> > kvm_update_dr7(vcpu);
> > @@ -11071,8 +11072,12 @@ static int vcpu_enter_guest(struct kvm_vcpu *vcpu)
> > * care about the messed up debug address registers. But if
> > * we have some of them active, restore the old state.
> > */
> > - if (hw_breakpoint_active())
> > - hw_breakpoint_restore();
> > + if (hw_breakpoint_active()) {
> > + if (!(vcpu->arch.switch_db_regs & KVM_DEBUGREG_AUTO_SWITCH))
> > + hw_breakpoint_restore();
> > + else
> > + set_debugreg(__this_cpu_read(cpu_dr7), 7);
>
> According to TDX module 1.5 ABI spec:
> DR0-3, DR6 and DR7 are set to their architectural INIT value, why is only
> DR7 restored?
This hunk should be dropped. Thank you for finding this.
I checked the base SPEC, the ABI spec, and the TDX module code. It seems the
documentation bug of the TDX module 1.5 base architecture specification.
The TDX module code:
- restores guest DR<N> on TD Entry to guest.
- saves guest DR<N> on TD Exit from guest TD
- initializes DR<N> on TD Exit to host VMM
TDX module 1.5 base architecture specification:
15.1.2.1 Context Switch
By design, the Intel TDX module context-switches all debug/tracing state that
the guest TD is allowed to use.
DR0-3, DR6 and IA32_DS_AREA MSR are context-switched in TDH.VP.ENTER and
TD exit flows
RFLAGS, IA32_DEBUGCTL MSR and DR7 are saved and cleared on VM exits from
the guest TD and restored on VM entry to the guest TD.
TDX module 1.5 ABI specification:
5.3.65. TDH.VP.ENTER Leaf
CPU State Preservation Following a Successful TD Entry and a TD Exit
Following a successful TD entry and a TD exit, some CPU state is modified:
Registers DR0, DR1, DR2, DR3, DR6 and DR7 are set to their architectural
INIT value.
--
Isaku Yamahata <isaku.yamahata@xxxxxxxxx>
