libxl: wait for the ack when issuing power control requests

Currently only suspend power control requests wait for an ack from the
domain, while power off or reboot requests simply write the command to
xenstore and exit.

Introduce a 1 minute wait for the domain to acknowledge the request, or
else return an error. The suspend code is slightly modified to use the
new infrastructure added, but shouldn't have any functional change.

Fix the ocaml bindings and also provide a backwards compatible
interface for the reboot and poweroff libxl API functions.

Reported-by: Ross Lagerwall <ross.lagerwall@citrix.com>
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Anthony PERARD <anthony.perard@citrix.com>
Release-acked-by: Juergen Gross <jgross@suse.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
[ wei: change ret to rc to fix build ]
Signed-off-by: Wei Liu <wl@xen.org>

tools/libxl/libxl.h

@@ -1239,6 +1239,8 @@ void libxl_mac_copy(libxl_ctx *ctx, libxl_mac *dst, const libxl_mac *src);
  *   libxl_set_vcpuonline()
  *   libxl_retrieve_domain_configuration()
  *   libxl_qemu_monitor_command()
+ *   libxl_domain_shutdown()
+ *   libxl_domain_reboot()
  */
 #define LIBXL_HAVE_FN_USING_QMP_ASYNC 1
 
@@ -1628,8 +1630,27 @@ int libxl_domain_remus_start(libxl_ctx *ctx, libxl_domain_remus_info *info,
                              const libxl_asyncop_how *ao_how)
                              LIBXL_EXTERNAL_CALLERS_ONLY;
 
-int libxl_domain_shutdown(libxl_ctx *ctx, uint32_t domid);
-int libxl_domain_reboot(libxl_ctx *ctx, uint32_t domid);
+int libxl_domain_shutdown(libxl_ctx *ctx, uint32_t domid,
+                          const libxl_asyncop_how *ao_how)
+                          LIBXL_EXTERNAL_CALLERS_ONLY;
+int libxl_domain_reboot(libxl_ctx *ctx, uint32_t domid,
+                        const libxl_asyncop_how *ao_how)
+                        LIBXL_EXTERNAL_CALLERS_ONLY;
+#if defined(LIBXL_API_VERSION) && LIBXL_API_VERSION < 0x041300
+static inline int libxl_domain_shutdown_0x041200(libxl_ctx *ctx,
+                                                 uint32_t domid)
+{
+    return libxl_domain_shutdown(ctx, domid, NULL);
+}
+#define libxl_domain_shutdown libxl_domain_shutdown_0x041200
+static inline int libxl_domain_reboot_0x041200(libxl_ctx *ctx,
+                                               uint32_t domid)
+{
+    return libxl_domain_reboot(ctx, domid, NULL);
+}
+#define libxl_domain_reboot libxl_domain_reboot_0x041200
+#endif
+
 int libxl_domain_destroy(libxl_ctx *ctx, uint32_t domid,
                          const libxl_asyncop_how *ao_how)
                          LIBXL_EXTERNAL_CALLERS_ONLY;

tools/libxl/libxl_dom_suspend.c

@@ -193,16 +193,11 @@ static void domain_suspend_callback_common(libxl__egc *egc,
     LOGD(DEBUG, domid, "issuing %s suspend request via XenBus control node",
         dsps->type != LIBXL_DOMAIN_TYPE_PV ? "PVH/HVM" : "PV");
 
-    libxl__domain_pvcontrol_write(gc, XBT_NULL, domid, "suspend");
-
-    dsps->pvcontrol.path = libxl__domain_pvcontrol_xspath(gc, domid);
-    if (!dsps->pvcontrol.path) { rc = ERROR_FAIL; goto err; }
-
     dsps->pvcontrol.ao = ao;
-    dsps->pvcontrol.what = "guest acknowledgement of suspend request";
-    dsps->pvcontrol.timeout_ms = 60 * 1000;
     dsps->pvcontrol.callback = domain_suspend_common_pvcontrol_suspending;
-    libxl__xswait_start(gc, &dsps->pvcontrol);
+    rc = libxl__domain_pvcontrol(egc, &dsps->pvcontrol, domid, "suspend");
+    if (rc) goto err;
+
     return;
 
  err:

tools/libxl/libxl_domain.c

@@ -763,49 +763,91 @@ char * libxl__domain_pvcontrol_read(libxl__gc *gc, xs_transaction_t t,
     return libxl__xs_read(gc, t, shutdown_path);
 }
 
-int libxl__domain_pvcontrol_write(libxl__gc *gc, xs_transaction_t t,
-                                  uint32_t domid, const char *cmd)
+int libxl__domain_pvcontrol(libxl__egc *egc, libxl__xswait_state *pvcontrol,
+                            domid_t domid, const char *cmd)
 {
+    STATE_AO_GC(pvcontrol->ao);
     const char *shutdown_path;
+    int rc;
+
+    rc = libxl__domain_pvcontrol_available(gc, domid);
+    if (rc < 0)
+        return rc;
 
     shutdown_path = libxl__domain_pvcontrol_xspath(gc, domid);
     if (!shutdown_path)
         return ERROR_FAIL;
 
-    return libxl__xs_printf(gc, t, shutdown_path, "%s", cmd);
+    rc = libxl__xs_printf(gc, XBT_NULL, shutdown_path, "%s", cmd);
+    if (rc)
+        return rc;
+
+    pvcontrol->path = shutdown_path;
+    pvcontrol->what = GCSPRINTF("guest acknowledgement of %s request", cmd);
+    pvcontrol->timeout_ms = 60 * 1000;
+    rc = libxl__xswait_start(gc, pvcontrol);
+    if (rc)
+        return rc;
+
+    return 0;
 }
 
-static int libxl__domain_pvcontrol(libxl__gc *gc, uint32_t domid,
-                                   const char *cmd)
+static bool pvcontrol_acked(const char *state)
 {
-    int ret;
+    if (!state || !strcmp(state,""))
+        return true;
 
-    ret = libxl__domain_pvcontrol_available(gc, domid);
-    if (ret < 0)
-        return ret;
-
-    if (!ret)
-        return ERROR_NOPARAVIRT;
-
-    return libxl__domain_pvcontrol_write(gc, XBT_NULL, domid, cmd);
+    return false;
 }
 
-int libxl_domain_shutdown(libxl_ctx *ctx, uint32_t domid)
+/* Xenstore watch callback prototype for the reboot/poweroff operations. */
+static void pvcontrol_cb(libxl__egc *egc, libxl__xswait_state *xswa, int rc,
+                         const char *state);
+
+int libxl_domain_shutdown(libxl_ctx *ctx, uint32_t domid,
+                          const libxl_asyncop_how *ao_how)
 {
-    GC_INIT(ctx);
-    int ret;
-    ret = libxl__domain_pvcontrol(gc, domid, "poweroff");
-    GC_FREE;
-    return ret;
+    AO_CREATE(ctx, domid, ao_how);
+    libxl__xswait_state *pvcontrol;
+    int rc;
+
+    GCNEW(pvcontrol);
+    pvcontrol->ao = ao;
+    pvcontrol->callback = pvcontrol_cb;
+    rc = libxl__domain_pvcontrol(egc, pvcontrol, domid, "poweroff");
+
+    return rc ? AO_CREATE_FAIL(rc) : AO_INPROGRESS;
 }
 
-int libxl_domain_reboot(libxl_ctx *ctx, uint32_t domid)
+int libxl_domain_reboot(libxl_ctx *ctx, uint32_t domid,
+                        const libxl_asyncop_how *ao_how)
 {
-    GC_INIT(ctx);
-    int ret;
-    ret = libxl__domain_pvcontrol(gc, domid, "reboot");
-    GC_FREE;
-    return ret;
+    AO_CREATE(ctx, domid, ao_how);
+    libxl__xswait_state *pvcontrol;
+    int rc;
+
+    GCNEW(pvcontrol);
+    pvcontrol->ao = ao;
+    pvcontrol->callback = pvcontrol_cb;
+    rc = libxl__domain_pvcontrol(egc, pvcontrol, domid, "reboot");
+
+    return rc ? AO_CREATE_FAIL(rc) : AO_INPROGRESS;
+}
+
+static void pvcontrol_cb(libxl__egc *egc, libxl__xswait_state *xswa, int rc,
+                         const char *state)
+{
+    STATE_AO_GC(xswa->ao);
+
+    if (!rc && !pvcontrol_acked(state))
+        return;
+
+    libxl__xswait_stop(gc, xswa);
+
+    if (rc)
+        LOG(ERROR, "guest didn't acknowledge control request: %d", rc);
+
+    libxl__ao_complete(egc, ao, rc);
 }
 
 static void domain_death_occurred(libxl__egc *egc,

tools/libxl/libxl_internal.h

@@ -1378,8 +1378,6 @@ _hidden int libxl__domain_pvcontrol_available(libxl__gc *gc, uint32_t domid);
 _hidden const char *libxl__domain_pvcontrol_xspath(libxl__gc*, uint32_t domid);
 _hidden char * libxl__domain_pvcontrol_read(libxl__gc *gc,
                                             xs_transaction_t t, uint32_t domid);
-_hidden int libxl__domain_pvcontrol_write(libxl__gc *gc, xs_transaction_t t,
-                                          uint32_t domid, const char *cmd);
 
 /* from xl_device */
 _hidden char *libxl__device_disk_string_of_backend(libxl_disk_backend backend);
@@ -4719,6 +4717,11 @@ _hidden void libxl__ev_devlock_init(libxl__ev_devlock *);
 _hidden void libxl__ev_devlock_lock(libxl__egc *, libxl__ev_devlock *);
 _hidden void libxl__ev_devlock_unlock(libxl__gc *, libxl__ev_devlock *);
 
+/* Send control commands over xenstore and wait for an Ack. */
+_hidden int libxl__domain_pvcontrol(libxl__egc *egc,
+                                    libxl__xswait_state *pvcontrol,
+                                    domid_t domid, const char *cmd);
+
 #endif
 
 /*

tools/ocaml/libs/xl/xenlight.ml.in

@@ -37,8 +37,8 @@ module Domain = struct
 	external create_new : ctx -> Domain_config.t -> ?async:'a -> unit -> domid = "stub_libxl_domain_create_new"
 	external create_restore : ctx -> Domain_config.t -> (Unix.file_descr * Domain_restore_params.t) ->
 		?async:'a -> unit -> domid = "stub_libxl_domain_create_restore"
-	external shutdown : ctx -> domid -> unit = "stub_libxl_domain_shutdown"
-	external reboot : ctx -> domid -> unit = "stub_libxl_domain_reboot"
+	external shutdown : ctx -> domid -> ?async:'a -> unit -> unit = "stub_libxl_domain_shutdown"
+	external reboot : ctx -> domid -> ?async:'a -> unit -> unit = "stub_libxl_domain_reboot"
 	external destroy : ctx -> domid -> ?async:'a -> unit -> unit = "stub_libxl_domain_destroy"
 	external suspend : ctx -> domid -> Unix.file_descr -> ?async:'a -> unit -> unit = "stub_libxl_domain_suspend"
 	external pause : ctx -> domid -> ?async:'a -> unit = "stub_libxl_domain_pause"

tools/ocaml/libs/xl/xenlight.mli.in

@@ -39,8 +39,8 @@ module Domain : sig
 	external create_new : ctx -> Domain_config.t -> ?async:'a -> unit -> domid = "stub_libxl_domain_create_new"
 	external create_restore : ctx -> Domain_config.t -> (Unix.file_descr * Domain_restore_params.t) ->
 		?async:'a -> unit -> domid = "stub_libxl_domain_create_restore"
-	external shutdown : ctx -> domid -> unit = "stub_libxl_domain_shutdown"
-	external reboot : ctx -> domid -> unit = "stub_libxl_domain_reboot"
+	external shutdown : ctx -> domid -> ?async:'a -> unit -> unit = "stub_libxl_domain_shutdown"
+	external reboot : ctx -> domid -> ?async:'a -> unit -> unit = "stub_libxl_domain_reboot"
 	external destroy : ctx -> domid -> ?async:'a -> unit -> unit = "stub_libxl_domain_destroy"
 	external suspend : ctx -> domid -> Unix.file_descr -> ?async:'a -> unit -> unit = "stub_libxl_domain_suspend"
 	external pause : ctx -> domid -> ?async:'a -> unit = "stub_libxl_domain_pause"

tools/ocaml/libs/xl/xenlight_stubs.c

@@ -551,32 +551,38 @@ value stub_libxl_domain_create_restore(value ctx, value domain_config, value par
 	CAMLreturn(Val_int(c_domid));
 }
 
-value stub_libxl_domain_shutdown(value ctx, value domid)
+value stub_libxl_domain_shutdown(value ctx, value domid, value async, value unit)
 {
-	CAMLparam2(ctx, domid);
+	CAMLparam4(ctx, domid, async, unit);
 	int ret;
 	uint32_t c_domid = Int_val(domid);
+	libxl_asyncop_how *ao_how = aohow_val(async);
 
 	caml_enter_blocking_section();
-	ret = libxl_domain_shutdown(CTX, c_domid);
+	ret = libxl_domain_shutdown(CTX, c_domid, ao_how);
 	caml_leave_blocking_section();
 
+	free(ao_how);
+
 	if (ret != 0)
 		failwith_xl(ret, "domain_shutdown");
 
 	CAMLreturn(Val_unit);
 }
 
-value stub_libxl_domain_reboot(value ctx, value domid)
+value stub_libxl_domain_reboot(value ctx, value domid, value async, value unit)
 {
-	CAMLparam2(ctx, domid);
+	CAMLparam4(ctx, domid, async, unit);
 	int ret;
 	uint32_t c_domid = Int_val(domid);
+	libxl_asyncop_how *ao_how = aohow_val(async);
 
 	caml_enter_blocking_section();
-	ret = libxl_domain_reboot(CTX, c_domid);
+	ret = libxl_domain_reboot(CTX, c_domid, ao_how);
 	caml_leave_blocking_section();
 
+	free(ao_how);
+
 	if (ret != 0)
 		failwith_xl(ret, "domain_reboot");
 

tools/xl/xl_vmcontrol.c

@@ -103,7 +103,7 @@ static void reboot_domain(uint32_t domid, libxl_evgen_domain_death **deathw,
     int rc;
 
     fprintf(stderr, "Rebooting domain %u\n", domid);
-    rc=libxl_domain_reboot(ctx, domid);
+    rc = libxl_domain_reboot(ctx, domid, NULL);
     if (rc == ERROR_NOPARAVIRT) {
         if (fallback_trigger) {
             fprintf(stderr, "PV control interface not available:"
@@ -136,7 +136,7 @@ static void shutdown_domain(uint32_t domid,
     int rc;
 
     fprintf(stderr, "Shutting down domain %u\n", domid);
-    rc=libxl_domain_shutdown(ctx, domid);
+    rc = libxl_domain_shutdown(ctx, domid, NULL);
     if (rc == ERROR_NOPARAVIRT) {
         if (fallback_trigger) {
             fprintf(stderr, "PV control interface not available:"