x86/vmx: add hvm functions to get/set non-register state

During VM forking and resetting a failed vmentry has been observed due
to the guest non-register state going out-of-sync with the guest register
state. For example, a VM fork reset right after a STI instruction can trigger
the failed entry. This is due to the guest non-register state not being saved
from the parent VM, thus the reset operation only copies the register state.

Fix this by adding a new pair of hvm functions to get/set the guest
non-register state so that the overall vCPU state remains in sync.

Signed-off-by: Tamas K Lengyel <tamas.lengyel@intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Acked-by: Jan Beulich <jbeulich@suse.com>

diff --git a/xen/arch/x86/hvm/vmx/vmx.c b/xen/arch/x86/hvm/vmx/vmx.c
index cc8c4e9f044a2c986e9b02e05174be5de4f9bcdb..d03e78bf0de4a7f047e9a8c296807ab6cdb40171 100644 (file)
--- a/xen/arch/x86/hvm/vmx/vmx.c
+++ b/xen/arch/x86/hvm/vmx/vmx.c
@@ -1334,6 +1334,36 @@ static void cf_check vmx_set_interrupt_shadow(
     __vmwrite(GUEST_INTERRUPTIBILITY_INFO, intr_shadow);
 }
 
+static void cf_check vmx_get_nonreg_state(struct vcpu *v,
+    struct hvm_vcpu_nonreg_state *nrs)
+{
+    vmx_vmcs_enter(v);
+
+    __vmread(GUEST_ACTIVITY_STATE, &nrs->vmx.activity_state);
+    __vmread(GUEST_INTERRUPTIBILITY_INFO, &nrs->vmx.interruptibility_info);
+    __vmread(GUEST_PENDING_DBG_EXCEPTIONS, &nrs->vmx.pending_dbg);
+
+    if ( cpu_has_vmx_virtual_intr_delivery )
+        __vmread(GUEST_INTR_STATUS, &nrs->vmx.interrupt_status);
+
+    vmx_vmcs_exit(v);
+}
+
+static void cf_check vmx_set_nonreg_state(struct vcpu *v,
+    struct hvm_vcpu_nonreg_state *nrs)
+{
+    vmx_vmcs_enter(v);
+
+    __vmwrite(GUEST_ACTIVITY_STATE, nrs->vmx.activity_state);
+    __vmwrite(GUEST_INTERRUPTIBILITY_INFO, nrs->vmx.interruptibility_info);
+    __vmwrite(GUEST_PENDING_DBG_EXCEPTIONS, nrs->vmx.pending_dbg);
+
+    if ( cpu_has_vmx_virtual_intr_delivery )
+        __vmwrite(GUEST_INTR_STATUS, nrs->vmx.interrupt_status);
+
+    vmx_vmcs_exit(v);
+}
+
 static void vmx_load_pdptrs(struct vcpu *v)
 {
     uint32_t cr3 = v->arch.hvm.guest_cr[3];
@@ -2487,6 +2517,8 @@ static struct hvm_function_table __initdata_cf_clobber vmx_function_table = {
     .load_cpu_ctxt        = vmx_load_vmcs_ctxt,
     .get_interrupt_shadow = vmx_get_interrupt_shadow,
     .set_interrupt_shadow = vmx_set_interrupt_shadow,
+    .get_nonreg_state     = vmx_get_nonreg_state,
+    .set_nonreg_state     = vmx_set_nonreg_state,
     .guest_x86_mode       = vmx_guest_x86_mode,
     .get_cpl              = _vmx_get_cpl,
     .get_segment_register = vmx_get_segment_register,

diff --git a/xen/arch/x86/include/asm/hvm/hvm.h b/xen/arch/x86/include/asm/hvm/hvm.h
index 5b7ec0cf69261b23e90ba1df0ccdf35953f39ddf..caaeacabc7fd62dd732e65a08c65c70c633f6f27 100644 (file)
--- a/xen/arch/x86/include/asm/hvm/hvm.h
+++ b/xen/arch/x86/include/asm/hvm/hvm.h
@@ -84,6 +84,17 @@ enum hvm_intblk {
 /* update_guest_cr() flags. */
 #define HVM_UPDATE_GUEST_CR3_NOFLUSH 0x00000001
 
+struct hvm_vcpu_nonreg_state {
+    union {
+        struct {
+            uint64_t activity_state;
+            uint64_t interruptibility_info;
+            uint64_t pending_dbg;
+            uint64_t interrupt_status;
+        } vmx;
+    };
+};
+
 /*
  * The hardware virtual machine (HVM) interface abstracts away from the
  * x86/x86_64 CPU virtualization assist specifics. Currently this interface
@@ -122,6 +133,10 @@ struct hvm_function_table {
     /* Examine specifics of the guest state. */
     unsigned int (*get_interrupt_shadow)(struct vcpu *v);
     void (*set_interrupt_shadow)(struct vcpu *v, unsigned int intr_shadow);
+    void (*get_nonreg_state)(struct vcpu *v,
+                             struct hvm_vcpu_nonreg_state *nrs);
+    void (*set_nonreg_state)(struct vcpu *v,
+                             struct hvm_vcpu_nonreg_state *nrs);
     int (*guest_x86_mode)(struct vcpu *v);
     unsigned int (*get_cpl)(struct vcpu *v);
     void (*get_segment_register)(struct vcpu *v, enum x86_segment seg,
@@ -744,6 +759,20 @@ void hvm_set_reg(struct vcpu *v, unsigned int reg, uint64_t val);
         d_->arch.hvm.pi_ops.vcpu_block(v_);                     \
 })
 
+static inline void hvm_get_nonreg_state(struct vcpu *v,
+                                        struct hvm_vcpu_nonreg_state *nrs)
+{
+    if ( hvm_funcs.get_nonreg_state )
+        alternative_vcall(hvm_funcs.get_nonreg_state, v, nrs);
+}
+
+static inline void hvm_set_nonreg_state(struct vcpu *v,
+                                        struct hvm_vcpu_nonreg_state *nrs)
+{
+    if ( hvm_funcs.set_nonreg_state )
+        alternative_vcall(hvm_funcs.set_nonreg_state, v, nrs);
+}
+
 #else  /* CONFIG_HVM */
 
 #define hvm_enabled false

diff --git a/xen/arch/x86/mm/mem_sharing.c b/xen/arch/x86/mm/mem_sharing.c
index a5c16b4429e6dc05dc28b2fb8530f83494e3808a..61538c3fb2485399c37acac98d2e0f7da61ac35a 100644 (file)
--- a/xen/arch/x86/mm/mem_sharing.c
+++ b/xen/arch/x86/mm/mem_sharing.c
@@ -1645,6 +1645,14 @@ static int bring_up_vcpus(struct domain *cd, struct domain *d)
     return 0;
 }
 
+static void copy_vcpu_nonreg_state(struct vcpu *d_vcpu, struct vcpu *cd_vcpu)
+{
+    struct hvm_vcpu_nonreg_state nrs = {};
+
+    hvm_get_nonreg_state(d_vcpu, &nrs);
+    hvm_set_nonreg_state(cd_vcpu, &nrs);
+}
+
 static int copy_vcpu_settings(struct domain *cd, const struct domain *d)
 {
     unsigned int i;
@@ -1653,7 +1661,7 @@ static int copy_vcpu_settings(struct domain *cd, const struct domain *d)
 
     for ( i = 0; i < cd->max_vcpus; i++ )
     {
-        const struct vcpu *d_vcpu = d->vcpu[i];
+        struct vcpu *d_vcpu = d->vcpu[i];
         struct vcpu *cd_vcpu = cd->vcpu[i];
         mfn_t vcpu_info_mfn;
 
@@ -1696,6 +1704,8 @@ static int copy_vcpu_settings(struct domain *cd, const struct domain *d)
 
         hvm_vmtrace_reset(cd_vcpu);
 
+        copy_vcpu_nonreg_state(d_vcpu, cd_vcpu);
+
         /*
          * TODO: to support VMs with PV interfaces copy additional
          * settings here, such as PV timers.