[RFC,v4,03/36] target/i386: Implement mc->kvm_type() to get VM type

Commit Message

Xiaoyao Li May 12, 2022, 3:17 a.m. UTC

TDX VM requires VM type KVM_X86_TDX_VM to be passed to
kvm_ioctl(KVM_CREATE_VM). Hence implement mc->kvm_type() for i386
architecture.

If tdx-guest object is specified to confidential-guest-support, like,

  qemu -machine ...,confidential-guest-support=tdx0 \
       -object tdx-guest,id=tdx0,...

it parses VM type as KVM_X86_TDX_VM. Otherwise, it's KVM_X86_DEFAULT_VM.

Signed-off-by: Xiaoyao Li <[e-mail]>
---
 hw/i386/x86.c              |  6 ++++++
 target/i386/kvm/kvm.c      | 30 ++++++++++++++++++++++++++++++
 target/i386/kvm/kvm_i386.h |  1 +
 3 files changed, 37 insertions(+)

Comments

Gerd Hoffmann May 23, 2022, 8:36 a.m. UTC | #1

Hi,

> +    if (!(kvm_check_extension(KVM_STATE(ms->accelerator), KVM_CAP_VM_TYPES) & BIT(kvm_type))) {
> +        error_report("vm-type %s not supported by KVM", vm_type_name[kvm_type]);
> +        exit(1);
> +    }

Not sure why TDX needs a new vm type whereas sev doesn't.  But that's up
for debate in the kernel tdx patches, not here.  Assuming the kernel
interface actually merged will look like this the patch makes sense.

Acked-by: Gerd Hoffmann <[e-mail]>

take care,
  Gerd

Isaku Yamahata May 23, 2022, 2:55 p.m. UTC | #2

On Mon, May 23, 2022 at 10:36:16AM +0200,
Gerd Hoffmann <[e-mail]> wrote:

>   Hi,
> 
> > +    if (!(kvm_check_extension(KVM_STATE(ms->accelerator), KVM_CAP_VM_TYPES) & BIT(kvm_type))) {
> > +        error_report("vm-type %s not supported by KVM", vm_type_name[kvm_type]);
> > +        exit(1);
> > +    }
> 
> Not sure why TDX needs a new vm type whereas sev doesn't.  But that's up
> for debate in the kernel tdx patches, not here.  Assuming the kernel
> interface actually merged will look like this the patch makes sense.

Because VM operations, e.g. KVM_CREATE_VCPU, require TDX specific one in KVM
side, we need to tell this VM is TD.
Also it's for consistency.  It's common pattern to specify vm type with
KVM_CREATE_VM when among other archs.  S390, PPC, MIPS, and ARM64.  Only SEV is
an exception.  It makes default VM into confidential VM after KVM_CREATE_VM.

Thanks,

> 
> Acked-by: Gerd Hoffmann <[e-mail]>
> 
> take care,
>   Gerd
> 
>

Patch

diff --git a/hw/i386/x86.c b/hw/i386/x86.c
index fba790f7b49c..4d0b0047627d 100644
--- a/hw/i386/x86.c
+++ b/hw/i386/x86.c
@@ -1345,6 +1345,11 @@  static void machine_set_sgx_epc(Object *obj, Visitor *v, const char *name,
     qapi_free_SgxEPCList(list);
 }
 
+static int x86_kvm_type(MachineState *ms, const char *vm_type)
+{
+    return kvm_get_vm_type(ms, vm_type);
+}
+
 static void x86_machine_initfn(Object *obj)
 {
     X86MachineState *x86ms = X86_MACHINE(obj);
@@ -1368,6 +1373,7 @@  static void x86_machine_class_init(ObjectClass *oc, void *data)
     mc->cpu_index_to_instance_props = x86_cpu_index_to_props;
     mc->get_default_cpu_node_id = x86_get_default_cpu_node_id;
     mc->possible_cpu_arch_ids = x86_possible_cpu_arch_ids;
+    mc->kvm_type = x86_kvm_type;
     x86mc->save_tsc_khz = true;
     x86mc->fwcfg_dma_enabled = true;
     nc->nmi_monitor_handler = x86_nmi;
diff --git a/target/i386/kvm/kvm.c b/target/i386/kvm/kvm.c
index c885763a5bb5..3f8a9183fa9b 100644
--- a/target/i386/kvm/kvm.c
+++ b/target/i386/kvm/kvm.c
@@ -30,6 +30,7 @@ 
 #include "sysemu/runstate.h"
 #include "kvm_i386.h"
 #include "sev.h"
+#include "tdx.h"
 #include "hyperv.h"
 #include "hyperv-proto.h"
 
@@ -142,6 +143,35 @@  static struct kvm_msr_list *kvm_feature_msrs;
 #define BUS_LOCK_SLICE_TIME 1000000000ULL /* ns */
 static RateLimit bus_lock_ratelimit_ctrl;
 
+static const char* vm_type_name[] = {
+    [KVM_X86_DEFAULT_VM] = "X86_DEFAULT_VM",
+    [KVM_X86_TDX_VM] = "X86_TDX_VM",
+};
+
+int kvm_get_vm_type(MachineState *ms, const char *vm_type)
+{
+    int kvm_type = KVM_X86_DEFAULT_VM;
+
+    if (ms->cgs && object_dynamic_cast(OBJECT(ms->cgs), TYPE_TDX_GUEST)) {
+        kvm_type = KVM_X86_TDX_VM;
+    }
+
+    /*
+     * old KVM doesn't support KVM_CAP_VM_TYPES and KVM_X86_DEFAULT_VM
+     * is always supported
+     */
+    if (kvm_type == KVM_X86_DEFAULT_VM) {
+        return kvm_type;
+    }
+
+    if (!(kvm_check_extension(KVM_STATE(ms->accelerator), KVM_CAP_VM_TYPES) & BIT(kvm_type))) {
+        error_report("vm-type %s not supported by KVM", vm_type_name[kvm_type]);
+        exit(1);
+    }
+
+    return kvm_type;
+}
+
 int kvm_has_pit_state2(void)
 {
     return has_pit_state2;
diff --git a/target/i386/kvm/kvm_i386.h b/target/i386/kvm/kvm_i386.h
index 4124912c202e..b434feaa6b1d 100644
--- a/target/i386/kvm/kvm_i386.h
+++ b/target/i386/kvm/kvm_i386.h
@@ -37,6 +37,7 @@  bool kvm_has_adjust_clock(void);
 bool kvm_has_adjust_clock_stable(void);
 bool kvm_has_exception_payload(void);
 void kvm_synchronize_all_tsc(void);
+int kvm_get_vm_type(MachineState *ms, const char *vm_type);
 void kvm_arch_reset_vcpu(X86CPU *cs);
 void kvm_arch_do_init_vcpu(X86CPU *cs);