[PATCH v5 1/4] x86/monitor: add get_capabilities to monitor_op domctl

* [PATCH v5 1/4] x86/monitor: add get_capabilities to monitor_op domctl
@ 2015-07-09 13:14 Tamas K Lengyel
  2015-07-09 13:14 ` [PATCH v5 2/4] x86/vm_event: toggle singlestep from vm_event response Tamas K Lengyel
                   ` (2 more replies)
  0 siblings, 3 replies; 10+ messages in thread
From: Tamas K Lengyel @ 2015-07-09 13:14 UTC (permalink / raw)
  To: xen-devel
  Cc: kevin.tian, wei.liu2, ian.campbell, rcojocaru,
	stefano.stabellini, ian.jackson, eddie.dong, jbeulich,
	jun.nakajima, andrew.cooper3, Tamas K Lengyel, keir

Add option to monitor_op domctl to determine the monitor capabilities of the
system.

Signed-off-by: Tamas K Lengyel <tlengyel@novetta.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Acked-by: Razvan Cojocaru <rcojocaru@bitdefender.com>
---
v5: fix libxc error return format
v4: add inline function wrapper for is_singlestep_supported to hvm.h
v3: move is_singlestep_supported into vmx
    sanity check capabilities for each monitor_op enable/disable
    fix comment typo
v2: skipped
---
 tools/libxc/include/xenctrl.h |  6 ++++++
 tools/libxc/xc_monitor.c      | 24 ++++++++++++++++++++++++
 xen/arch/x86/hvm/vmx/vmx.c    |  6 ++++++
 xen/arch/x86/monitor.c        | 33 +++++++++++++++++++++++++++++++--
 xen/common/domctl.c           |  2 ++
 xen/include/asm-x86/hvm/hvm.h |  6 ++++++
 xen/include/public/domctl.h   | 18 +++++++++++++++---
 7 files changed, 90 insertions(+), 5 deletions(-)

diff --git a/tools/libxc/include/xenctrl.h b/tools/libxc/include/xenctrl.h
index 71539a7..4fe347b 100644
--- a/tools/libxc/include/xenctrl.h
+++ b/tools/libxc/include/xenctrl.h
@@ -2364,6 +2364,12 @@ int xc_mem_access_disable_emulate(xc_interface *xch, domid_t domain_id);
 void *xc_monitor_enable(xc_interface *xch, domid_t domain_id, uint32_t *port);
 int xc_monitor_disable(xc_interface *xch, domid_t domain_id);
 int xc_monitor_resume(xc_interface *xch, domid_t domain_id);
+/*
+ * Get a bitmap of supported monitor events in the form
+ * (1 << XEN_DOMCTL_MONITOR_EVENT_*).
+ */
+int xc_monitor_get_capabilities(xc_interface *xch, domid_t domain_id,
+                                uint32_t *capabilities);
 int xc_monitor_write_ctrlreg(xc_interface *xch, domid_t domain_id,
                              uint16_t index, bool enable, bool sync,
                              bool onchangeonly);
diff --git a/tools/libxc/xc_monitor.c b/tools/libxc/xc_monitor.c
index 63013de..b64bce3 100644
--- a/tools/libxc/xc_monitor.c
+++ b/tools/libxc/xc_monitor.c
@@ -45,6 +45,30 @@ int xc_monitor_resume(xc_interface *xch, domid_t domain_id)
                                NULL);
 }
 
+int xc_monitor_get_capabilities(xc_interface *xch, domid_t domain_id,
+                                uint32_t *capabilities)
+{
+    int rc;
+    DECLARE_DOMCTL;
+
+    if ( !capabilities )
+    {
+        errno = EINVAL;
+        return -1;
+    }
+
+    domctl.cmd = XEN_DOMCTL_monitor_op;
+    domctl.domain = domain_id;
+    domctl.u.monitor_op.op = XEN_DOMCTL_MONITOR_OP_GET_CAPABILITIES;
+
+    rc = do_domctl(xch, &domctl);
+    if ( rc )
+        return rc;
+
+    *capabilities = domctl.u.monitor_op.event;
+    return 0;
+}
+
 int xc_monitor_write_ctrlreg(xc_interface *xch, domid_t domain_id,
                              uint16_t index, bool enable, bool sync,
                              bool onchangeonly)
diff --git a/xen/arch/x86/hvm/vmx/vmx.c b/xen/arch/x86/hvm/vmx/vmx.c
index e0e9e75..1f7f9e5 100644
--- a/xen/arch/x86/hvm/vmx/vmx.c
+++ b/xen/arch/x86/hvm/vmx/vmx.c
@@ -1763,6 +1763,11 @@ static void vmx_enable_msr_exit_interception(struct domain *d)
                                          MSR_TYPE_W);
 }
 
+static bool_t vmx_is_singlestep_supported(void)
+{
+    return cpu_has_monitor_trap_flag;
+}
+
 static struct hvm_function_table __initdata vmx_function_table = {
     .name                 = "VMX",
     .cpu_up_prepare       = vmx_cpu_up_prepare,
@@ -1820,6 +1825,7 @@ static struct hvm_function_table __initdata vmx_function_table = {
     .nhvm_hap_walk_L1_p2m = nvmx_hap_walk_L1_p2m,
     .hypervisor_cpuid_leaf = vmx_hypervisor_cpuid_leaf,
     .enable_msr_exit_interception = vmx_enable_msr_exit_interception,
+    .is_singlestep_supported = vmx_is_singlestep_supported,
 };
 
 const struct hvm_function_table * __init start_vmx(void)
diff --git a/xen/arch/x86/monitor.c b/xen/arch/x86/monitor.c
index 896acf7..590a45d 100644
--- a/xen/arch/x86/monitor.c
+++ b/xen/arch/x86/monitor.c
@@ -42,10 +42,29 @@ int status_check(struct xen_domctl_monitor_op *mop, bool_t status)
     return 0;
 }
 
+static inline uint32_t get_capabilities(struct domain *d)
+{
+    uint32_t capabilities = 0;
+
+    if ( !is_hvm_domain(d) || !cpu_has_vmx )
+        return capabilities;
+
+    capabilities = (1 << XEN_DOMCTL_MONITOR_EVENT_WRITE_CTRLREG) |
+                   (1 << XEN_DOMCTL_MONITOR_EVENT_MOV_TO_MSR) |
+                   (1 << XEN_DOMCTL_MONITOR_EVENT_SOFTWARE_BREAKPOINT);
+
+    /* Since we know this is on VMX, we can just call the hvm func */
+    if ( hvm_is_singlestep_supported() )
+        capabilities |= (1 << XEN_DOMCTL_MONITOR_EVENT_SINGLESTEP);
+
+    return capabilities;
+}
+
 int monitor_domctl(struct domain *d, struct xen_domctl_monitor_op *mop)
 {
     int rc;
     struct arch_domain *ad = &d->arch;
+    uint32_t capabilities = get_capabilities(d);
 
     rc = xsm_vm_event_control(XSM_PRIV, d, mop->op, mop->event);
     if ( rc )
@@ -55,8 +74,12 @@ int monitor_domctl(struct domain *d, struct xen_domctl_monitor_op *mop)
      * At the moment only Intel HVM domains are supported. However, event
      * delivery could be extended to AMD and PV domains.
      */
-    if ( !is_hvm_domain(d) || !cpu_has_vmx )
-        return -EOPNOTSUPP;
+
+    if ( mop->op == XEN_DOMCTL_MONITOR_OP_GET_CAPABILITIES )
+    {
+        mop->event = capabilities;
+        return 0;
+    }
 
     /*
      * Sanity check
@@ -65,6 +88,12 @@ int monitor_domctl(struct domain *d, struct xen_domctl_monitor_op *mop)
          mop->op != XEN_DOMCTL_MONITOR_OP_DISABLE )
         return -EOPNOTSUPP;
 
+    /*
+     * Check if event type is available.
+     */
+    if ( !( capabilities & (1 << mop->event) ) )
+        return -EOPNOTSUPP;
+
     switch ( mop->event )
     {
     case XEN_DOMCTL_MONITOR_EVENT_WRITE_CTRLREG:
diff --git a/xen/common/domctl.c b/xen/common/domctl.c
index a06f15c..7f959f3 100644
--- a/xen/common/domctl.c
+++ b/xen/common/domctl.c
@@ -1167,6 +1167,8 @@ long do_domctl(XEN_GUEST_HANDLE_PARAM(xen_domctl_t) u_domctl)
             break;
 
         ret = monitor_domctl(d, &op->u.monitor_op);
+        if ( !ret )
+            copyback = 1;
         break;
 
     default:
diff --git a/xen/include/asm-x86/hvm/hvm.h b/xen/include/asm-x86/hvm/hvm.h
index 1d1fd35..c9a7184 100644
--- a/xen/include/asm-x86/hvm/hvm.h
+++ b/xen/include/asm-x86/hvm/hvm.h
@@ -202,6 +202,7 @@ struct hvm_function_table {
                                   uint32_t *ecx, uint32_t *edx);
 
     void (*enable_msr_exit_interception)(struct domain *d);
+    bool_t (*is_singlestep_supported)(void);
 };
 
 extern struct hvm_function_table hvm_funcs;
@@ -510,6 +511,11 @@ static inline enum hvm_intblk nhvm_interrupt_blocked(struct vcpu *v)
     return hvm_funcs.nhvm_intr_blocked(v);
 }
 
+static inline bool_t hvm_is_singlestep_supported(void)
+{
+    return (hvm_funcs.is_singlestep_supported &&
+            hvm_funcs.is_singlestep_supported());
+}
 
 #ifndef NDEBUG
 /* Permit use of the Forced Emulation Prefix in HVM guests */
diff --git a/xen/include/public/domctl.h b/xen/include/public/domctl.h
index 2128171..8b1d6b4 100644
--- a/xen/include/public/domctl.h
+++ b/xen/include/public/domctl.h
@@ -994,12 +994,16 @@ DEFINE_XEN_GUEST_HANDLE(xen_domctl_psr_cmt_op_t);
  * via the ring buffer "MONITOR". The ring has to be first enabled
  * with the domctl XEN_DOMCTL_VM_EVENT_OP_MONITOR.
  *
+ * GET_CAPABILITIES can be used to determine which of these features is
+ * available on a given platform.
+ *
  * NOTICE: mem_access events are also delivered via the "MONITOR" ring buffer;
  * however, enabling/disabling those events is performed with the use of
  * memory_op hypercalls!
  */
-#define XEN_DOMCTL_MONITOR_OP_ENABLE   0
-#define XEN_DOMCTL_MONITOR_OP_DISABLE  1
+#define XEN_DOMCTL_MONITOR_OP_ENABLE            0
+#define XEN_DOMCTL_MONITOR_OP_DISABLE           1
+#define XEN_DOMCTL_MONITOR_OP_GET_CAPABILITIES  2
 
 #define XEN_DOMCTL_MONITOR_EVENT_WRITE_CTRLREG         0
 #define XEN_DOMCTL_MONITOR_EVENT_MOV_TO_MSR            1
@@ -1008,7 +1012,15 @@ DEFINE_XEN_GUEST_HANDLE(xen_domctl_psr_cmt_op_t);
 
 struct xen_domctl_monitor_op {
     uint32_t op; /* XEN_DOMCTL_MONITOR_OP_* */
-    uint32_t event; /* XEN_DOMCTL_MONITOR_EVENT_* */
+
+    /*
+     * When used with ENABLE/DISABLE this has to be set to
+     * the requested XEN_DOMCTL_MONITOR_EVENT_* value.
+     * With GET_CAPABILITIES this field returns a bitmap of
+     * events supported by the platform, in the format
+     * (1 << XEN_DOMCTL_MONITOR_EVENT_*).
+     */
+    uint32_t event;
 
     /*
      * Further options when issuing XEN_DOMCTL_MONITOR_OP_ENABLE.
-- 
2.1.4

^ permalink raw reply related	[flat|nested] 10+ messages in thread