libvirt/src/libvirt.c

12851 lines
340 KiB
C
Raw Normal View History

2005-11-02 12:50:21 +00:00
/*
* libvirt.c: Main interfaces for the libvirt library to handle virtualization
2005-11-02 12:50:21 +00:00
* domains from a process running in domain 0
*
build: consistently use C99 varargs macros Prior to this patch, there was an inconsistent mix between GNU and C99. For consistency, and potential portability to other compilers, stick with the C99 vararg macro syntax. * src/conf/cpu_conf.c (virCPUReportError): Use C99 rather than GNU vararg macro syntax. * src/conf/domain_conf.c (virDomainReportError): Likewise. * src/conf/domain_event.c (eventReportError): Likewise. * src/conf/interface_conf.c (virInterfaceReportError): Likewise. * src/conf/network_conf.c (virNetworkReportError): Likewise. * src/conf/node_device_conf.h (virNodeDeviceReportError): Likewise. * src/conf/secret_conf.h (virSecretReportError): Likewise. * src/conf/storage_conf.h (virStorageReportError): Likewise. * src/esx/esx_device_monitor.c (ESX_ERROR): Use C99 rather than GNU vararg macro syntax. * src/esx/esx_driver.c (ESX_ERROR): Likewise. * src/esx/esx_interface_driver.c (ESX_ERROR): Likewise. * src/esx/esx_network_driver.c (ESX_ERROR): Likewise. * src/esx/esx_secret_driver.c (ESX_ERROR): Likewise. * src/esx/esx_storage_driver.c (ESX_ERROR): Likewise. * src/esx/esx_util.c (ESX_ERROR): Likewise. * src/esx/esx_vi.c (ESX_VI_ERROR): Likewise. * src/esx/esx_vi_methods.c (ESX_VI_ERROR): Likewise. * src/esx/esx_vi_types.c (ESX_VI_ERROR): Likewise. * src/esx/esx_vmx.c (ESX_ERROR): Likewise. * src/util/hostusb.c (usbReportError): Use C99 rather than GNU vararg macro syntax. * src/util/json.c (virJSONError): Likewise. * src/util/macvtap.c (ReportError): Likewise. * src/util/pci.c (pciReportError): Likewise. * src/util/stats_linux.c (virStatsError): Likewise. * src/util/util.c (virUtilError): Likewise. * src/util/xml.c (virXMLError): Likewise. * src/xen/proxy_internal.c (virProxyError): Use C99 rather than GNU vararg macro syntax. * src/xen/sexpr.c (virSexprError): Likewise. * src/xen/xen_driver.c (xenUnifiedError): Likewise. * src/xen/xen_hypervisor.c (virXenError): Likewise. * src/xen/xen_inotify.c (virXenInotifyError): Likewise. * src/xen/xend_internal.c (virXendError): Likewise. * src/xen/xm_internal.c (xenXMError): Likewise. * src/xen/xs_internal.c (virXenStoreError): Likewise. * src/cpu/cpu.h (virCPUReportError): Use C99 rather than GNU vararg macro syntax. * src/datatypes.c (virLibConnError): Likewise. * src/interface/netcf_driver.c (interfaceReportError): Likewise. * src/libvirt.c (virLibStreamError): Likewise. * src/lxc/lxc_conf.h (lxcError): Likewise. * src/network/bridge_driver.c (networkReportError): Likewise. * src/nodeinfo.c (nodeReportError): Likewise. * src/opennebula/one_conf.h (oneError): Likewise. * src/openvz/openvz_conf.h (openvzError): Likewise. * src/phyp/phyp_driver.c (PHYP_ERROR): Likewise. * src/qemu/qemu_conf.h (qemuReportError): Likewise. * src/remote/remote_driver.c (errorf): Likewise. * src/security/security_driver.h (virSecurityReportError): Likewise. * src/test/test_driver.c (testError): Likewise. * src/uml/uml_conf.h (umlReportError): Likewise. * src/vbox/vbox_driver.c (vboxError): Likewise. * src/vbox/vbox_tmpl.c (vboxError): Likewise.
2010-03-01 23:38:28 +00:00
* Copyright (C) 2005-2006, 2008-2010 Red Hat, Inc.
2005-11-02 12:50:21 +00:00
*
* See COPYING.LIB for the License of this software
*
* Daniel Veillard <veillard@redhat.com>
*/
#include <config.h>
2005-11-02 12:50:21 +00:00
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <unistd.h>
#include <assert.h>
#include <sys/wait.h>
2009-01-22 20:27:01 +00:00
#include <time.h>
#include <gcrypt.h>
#include <libxml/parser.h>
#include <libxml/xpath.h>
#include <libxml/uri.h>
#include "getpass.h"
#ifdef HAVE_WINSOCK2_H
# include <winsock2.h>
#endif
#include "virterror_internal.h"
#include "logging.h"
#include "datatypes.h"
#include "driver.h"
2007-08-09 20:19:12 +00:00
#include "uuid.h"
#include "util.h"
#include "memory.h"
2008-11-21 12:16:08 +00:00
#ifndef WITH_DRIVER_MODULES
# ifdef WITH_TEST
# include "test/test_driver.h"
# endif
# ifdef WITH_XEN
# include "xen/xen_driver.h"
# endif
# ifdef WITH_REMOTE
# include "remote/remote_driver.h"
# endif
# ifdef WITH_OPENVZ
# include "openvz/openvz_driver.h"
# endif
# ifdef WITH_PHYP
# include "phyp/phyp_driver.h"
# endif
# ifdef WITH_VBOX
# include "vbox/vbox_driver.h"
# endif
# ifdef WITH_ESX
# include "esx/esx_driver.h"
# endif
# ifdef WITH_XENAPI
# include "xenapi/xenapi_driver.h"
# endif
2008-11-21 12:16:08 +00:00
#endif
#define VIR_FROM_THIS VIR_FROM_NONE
2005-11-02 12:50:21 +00:00
/*
* TODO:
* - use lock to protect against concurrent accesses ?
* - use reference counting to guarantee coherent pointer state ?
2005-11-02 12:50:21 +00:00
*/
#define MAX_DRIVERS 10
static virDriverPtr virDriverTab[MAX_DRIVERS];
static int virDriverTabCount = 0;
static virNetworkDriverPtr virNetworkDriverTab[MAX_DRIVERS];
static int virNetworkDriverTabCount = 0;
static virInterfaceDriverPtr virInterfaceDriverTab[MAX_DRIVERS];
static int virInterfaceDriverTabCount = 0;
static virStorageDriverPtr virStorageDriverTab[MAX_DRIVERS];
static int virStorageDriverTabCount = 0;
static virDeviceMonitorPtr virDeviceMonitorTab[MAX_DRIVERS];
static int virDeviceMonitorTabCount = 0;
static virSecretDriverPtr virSecretDriverTab[MAX_DRIVERS];
static int virSecretDriverTabCount = 0;
static virNWFilterDriverPtr virNWFilterDriverTab[MAX_DRIVERS];
static int virNWFilterDriverTabCount = 0;
2008-06-26 09:37:51 +00:00
#ifdef WITH_LIBVIRTD
static virStateDriverPtr virStateDriverTab[MAX_DRIVERS];
static int virStateDriverTabCount = 0;
2008-06-26 09:37:51 +00:00
#endif
static int initialized = 0;
#if defined(POLKIT_AUTH)
static int virConnectAuthGainPolkit(const char *privilege) {
const char *const args[] = {
POLKIT_AUTH, "--obtain", privilege, NULL
};
int childpid, status, ret;
/* Root has all rights */
if (getuid() == 0)
return 0;
if ((childpid = fork()) < 0)
return -1;
if (!childpid) {
execvp(args[0], (char **)args);
_exit(-1);
}
while ((ret = waitpid(childpid, &status, 0) == -1) && errno == EINTR);
if (ret == -1) {
return -1;
}
if (!WIFEXITED(status) ||
(WEXITSTATUS(status) != 0 && WEXITSTATUS(status) != 1)) {
return -1;
}
return 0;
}
#endif
static int virConnectAuthCallbackDefault(virConnectCredentialPtr cred,
unsigned int ncred,
void *cbdata ATTRIBUTE_UNUSED) {
int i;
for (i = 0 ; i < ncred ; i++) {
char buf[1024];
char *bufptr = buf;
size_t len;
switch (cred[i].type) {
case VIR_CRED_EXTERNAL: {
if (STRNEQ(cred[i].challenge, "PolicyKit"))
return -1;
2008-04-04 15:09:19 +00:00
#if defined(POLKIT_AUTH)
if (virConnectAuthGainPolkit(cred[i].prompt) < 0)
return -1;
2008-04-04 15:09:19 +00:00
#else
/*
* Ignore & carry on. Although we can't auth
* directly, the user may have authenticated
* themselves already outside context of libvirt
*/
#endif
break;
}
2008-04-04 15:09:19 +00:00
case VIR_CRED_USERNAME:
case VIR_CRED_AUTHNAME:
case VIR_CRED_ECHOPROMPT:
case VIR_CRED_REALM:
if (printf("%s: ", cred[i].prompt) < 0)
return -1;
if (fflush(stdout) != 0)
return -1;
if (!fgets(buf, sizeof(buf), stdin)) {
if (feof(stdin)) { /* Treat EOF as "" */
buf[0] = '\0';
break;
}
return -1;
}
len = strlen(buf);
if (len != 0 && buf[len-1] == '\n')
buf[len-1] = '\0';
break;
case VIR_CRED_PASSPHRASE:
case VIR_CRED_NOECHOPROMPT:
if (printf("%s: ", cred[i].prompt) < 0)
return -1;
if (fflush(stdout) != 0)
return -1;
bufptr = getpass("");
if (!bufptr)
return -1;
break;
default:
return -1;
}
2008-05-22 15:12:25 +00:00
if (cred[i].type != VIR_CRED_EXTERNAL) {
if (STREQ(bufptr, "") && cred[i].defresult)
cred[i].result = strdup(cred[i].defresult);
else
cred[i].result = strdup(bufptr);
if (!cred[i].result)
return -1;
cred[i].resultlen = strlen(cred[i].result);
}
}
return 0;
}
/* Don't typically want VIR_CRED_USERNAME. It enables you to authenticate
* as one user, and act as another. It just results in annoying
* prompts for the username twice & is very rarely what you want
*/
static int virConnectCredTypeDefault[] = {
VIR_CRED_AUTHNAME,
VIR_CRED_ECHOPROMPT,
VIR_CRED_REALM,
VIR_CRED_PASSPHRASE,
VIR_CRED_NOECHOPROMPT,
VIR_CRED_EXTERNAL,
};
static virConnectAuth virConnectAuthDefault = {
virConnectCredTypeDefault,
sizeof(virConnectCredTypeDefault)/sizeof(int),
virConnectAuthCallbackDefault,
NULL,
};
/*
* virConnectAuthPtrDefault
*
* A default implementation of the authentication callbacks. This
* implementation is suitable for command line based tools. It will
* prompt for username, passwords, realm and one time keys as needed.
* It will print on STDOUT, and read from STDIN. If this is not
* suitable for the application's needs an alternative implementation
* should be provided.
*/
virConnectAuthPtr virConnectAuthPtrDefault = &virConnectAuthDefault;
#if HAVE_WINSOCK2_H
static int
winsock_init (void)
{
WORD winsock_version, err;
WSADATA winsock_data;
/* http://msdn2.microsoft.com/en-us/library/ms742213.aspx */
winsock_version = MAKEWORD (2, 2);
err = WSAStartup (winsock_version, &winsock_data);
return err == 0 ? 0 : -1;
}
#endif
static int virTLSMutexInit (void **priv)
{ \
virMutexPtr lock = NULL;
if (VIR_ALLOC(lock) < 0)
return ENOMEM;
if (virMutexInit(lock) < 0) {
VIR_FREE(lock);
return errno;
}
*priv = lock;
return 0;
}
static int virTLSMutexDestroy(void **priv)
{
virMutexPtr lock = *priv;
virMutexDestroy(lock);
VIR_FREE(lock);
return 0;
}
static int virTLSMutexLock(void **priv)
{
virMutexPtr lock = *priv;
virMutexLock(lock);
return 0;
}
static int virTLSMutexUnlock(void **priv)
{
virMutexPtr lock = *priv;
virMutexUnlock(lock);
return 0;
}
static struct gcry_thread_cbs virTLSThreadImpl = {
/* GCRY_THREAD_OPTION_VERSION was added in gcrypt 1.4.2 */
#ifdef GCRY_THREAD_OPTION_VERSION
(GCRY_THREAD_OPTION_PTHREAD | (GCRY_THREAD_OPTION_VERSION << 8)),
#else
GCRY_THREAD_OPTION_PTHREAD,
#endif
NULL,
virTLSMutexInit,
virTLSMutexDestroy,
virTLSMutexLock,
virTLSMutexUnlock,
NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL
};
/**
* virInitialize:
*
* Initialize the library. It's better to call this routine at startup
* in multithreaded applications to avoid potential race when initializing
* the library.
*
* Returns 0 in case of success, -1 in case of error
*/
int
virInitialize(void)
{
if (initialized)
return(0);
2009-01-15 19:56:05 +00:00
initialized = 1;
2009-01-20 12:01:45 +00:00
if (virThreadInitialize() < 0 ||
2009-01-22 20:27:01 +00:00
virErrorInitialize() < 0 ||
virRandomInitialize(time(NULL) ^ getpid()))
2009-01-15 19:56:05 +00:00
return -1;
gcry_control(GCRYCTL_SET_THREAD_CBS, &virTLSThreadImpl);
gcry_check_version(NULL);
virLogSetFromEnv();
DEBUG0("register drivers");
#if HAVE_WINSOCK2_H
if (winsock_init () == -1) return -1;
#endif
if (!bindtextdomain(GETTEXT_PACKAGE, LOCALEBASEDIR))
return (-1);
/*
* Note that the order is important: the first ones have a higher
* priority when calling virConnectOpen.
*/
2008-11-21 12:16:08 +00:00
#ifdef WITH_DRIVER_MODULES
/* We don't care if any of these fail, because the whole point
* is to allow users to only install modules they want to use.
2009-10-26 23:02:46 +00:00
* If they try to open a connection for a module that
2008-11-21 12:16:08 +00:00
* is not loaded they'll get a suitable error at that point
*/
virDriverLoadModule("test");
virDriverLoadModule("xen");
virDriverLoadModule("openvz");
virDriverLoadModule("vbox");
virDriverLoadModule("esx");
virDriverLoadModule("xenapi");
2008-11-21 12:16:08 +00:00
virDriverLoadModule("remote");
#else
# ifdef WITH_TEST
if (testRegister() == -1) return -1;
# endif
# ifdef WITH_XEN
2008-11-21 12:16:08 +00:00
if (xenRegister () == -1) return -1;
# endif
# ifdef WITH_OPENVZ
if (openvzRegister() == -1) return -1;
# endif
# ifdef WITH_PHYP
if (phypRegister() == -1) return -1;
# endif
# ifdef WITH_VBOX
if (vboxRegister() == -1) return -1;
# endif
# ifdef WITH_ESX
if (esxRegister() == -1) return -1;
# endif
# ifdef WITH_XENAPI
if (xenapiRegister() == -1) return -1;
# endif
# ifdef WITH_REMOTE
if (remoteRegister () == -1) return -1;
# endif
#endif
2007-02-14 01:40:09 +00:00
return(0);
}
2009-01-15 19:56:05 +00:00
#ifdef WIN32
BOOL WINAPI
DllMain (HINSTANCE instance, DWORD reason, LPVOID ignore);
BOOL WINAPI
DllMain (HINSTANCE instance ATTRIBUTE_UNUSED,
DWORD reason,
LPVOID ignore ATTRIBUTE_UNUSED)
{
switch (reason) {
case DLL_PROCESS_ATTACH:
virInitialize();
break;
case DLL_THREAD_ATTACH:
/* Nothing todo in libvirt yet */
break;
case DLL_THREAD_DETACH:
/* Release per-thread local data */
virThreadOnExit();
break;
case DLL_PROCESS_DETACH:
/* Don't bother releasing per-thread data
since (hopefully) windows cleans up
everything on process exit */
break;
}
return TRUE;
}
#endif
/**
* virLibConnError:
* @conn: the connection if available
* @error: the error number
* @info: extra information string
*
* Handle an error at the connection level
*/
static void
virLibConnError(virConnectPtr conn, virErrorNumber error, const char *info)
{
const char *errmsg;
if (error == VIR_ERR_OK)
return;
errmsg = virErrorMsg(error, info);
virRaiseError(conn, NULL, NULL, VIR_FROM_NONE, error, VIR_ERR_ERROR,
errmsg, info, NULL, 0, 0, errmsg, info);
}
/**
* virLibConnWarning:
* @conn: the connection if available
* @error: the error number
* @info: extra information string
*
* Handle an error at the connection level
*/
static void
virLibConnWarning(virConnectPtr conn, virErrorNumber error, const char *info)
{
const char *errmsg;
if (error == VIR_ERR_OK)
return;
errmsg = virErrorMsg(error, info);
virRaiseError(conn, NULL, NULL, VIR_FROM_NONE, error, VIR_ERR_WARNING,
errmsg, info, NULL, 0, 0, errmsg, info);
}
/**
* virLibDomainError:
* @domain: the domain if available
* @error: the error number
* @info: extra information string
*
* Handle an error at the connection level
*/
static void
virLibDomainError(virDomainPtr domain, virErrorNumber error,
const char *info)
{
virConnectPtr conn = NULL;
const char *errmsg;
if (error == VIR_ERR_OK)
return;
errmsg = virErrorMsg(error, info);
if (error != VIR_ERR_INVALID_DOMAIN) {
conn = domain->conn;
}
virRaiseError(conn, domain, NULL, VIR_FROM_DOM, error, VIR_ERR_ERROR,
errmsg, info, NULL, 0, 0, errmsg, info);
}
/**
* virLibNetworkError:
* @conn: the connection if available
* @error: the error number
* @info: extra information string
*
* Handle an error at the connection level
*/
static void
virLibNetworkError(virNetworkPtr network, virErrorNumber error,
const char *info)
{
virConnectPtr conn = NULL;
const char *errmsg;
if (error == VIR_ERR_OK)
return;
errmsg = virErrorMsg(error, info);
if (error != VIR_ERR_INVALID_NETWORK) {
conn = network->conn;
}
virRaiseError(conn, NULL, network, VIR_FROM_NET, error, VIR_ERR_ERROR,
errmsg, info, NULL, 0, 0, errmsg, info);
}
/**
* virLibInterfaceError:
* @conn: the connection if available
* @error: the error number
* @info: extra information string
*
* Handle an error at the connection level
*/
static void
virLibInterfaceError(virInterfacePtr iface, virErrorNumber error,
const char *info)
{
virConnectPtr conn = NULL;
const char *errmsg;
if (error == VIR_ERR_OK)
return;
errmsg = virErrorMsg(error, info);
if (error != VIR_ERR_INVALID_INTERFACE) {
conn = iface->conn;
}
virRaiseError(conn, NULL, NULL, VIR_FROM_INTERFACE, error, VIR_ERR_ERROR,
errmsg, info, NULL, 0, 0, errmsg, info);
}
/**
* virLibStoragePoolError:
* @conn: the connection if available
* @error: the error number
* @info: extra information string
*
* Handle an error at the connection level
*/
static void
virLibStoragePoolError(virStoragePoolPtr pool, virErrorNumber error,
const char *info)
{
virConnectPtr conn = NULL;
const char *errmsg;
if (error == VIR_ERR_OK)
return;
errmsg = virErrorMsg(error, info);
if (error != VIR_ERR_INVALID_STORAGE_POOL)
conn = pool->conn;
virRaiseError(conn, NULL, NULL, VIR_FROM_STORAGE, error, VIR_ERR_ERROR,
errmsg, info, NULL, 0, 0, errmsg, info);
}
/**
* virLibStorageVolError:
* @conn: the connection if available
* @error: the error number
* @info: extra information string
*
* Handle an error at the connection level
*/
static void
virLibStorageVolError(virStorageVolPtr vol, virErrorNumber error,
const char *info)
{
virConnectPtr conn = NULL;
const char *errmsg;
if (error == VIR_ERR_OK)
return;
errmsg = virErrorMsg(error, info);
if (error != VIR_ERR_INVALID_STORAGE_VOL)
conn = vol->conn;
virRaiseError(conn, NULL, NULL, VIR_FROM_STORAGE, error, VIR_ERR_ERROR,
errmsg, info, NULL, 0, 0, errmsg, info);
}
/**
* virLibNodeDeviceError:
* @dev: the device if available
* @error: the error number
* @info: extra information string
*
* Handle an error at the node device level
*/
static void
virLibNodeDeviceError(virNodeDevicePtr dev, virErrorNumber error,
const char *info)
{
virConnectPtr conn = NULL;
const char *errmsg;
if (error == VIR_ERR_OK)
return;
errmsg = virErrorMsg(error, info);
if (error != VIR_ERR_INVALID_NODE_DEVICE)
conn = dev->conn;
virRaiseError(conn, NULL, NULL, VIR_FROM_NODEDEV, error, VIR_ERR_ERROR,
errmsg, info, NULL, 0, 0, errmsg, info);
}
build: consistently use C99 varargs macros Prior to this patch, there was an inconsistent mix between GNU and C99. For consistency, and potential portability to other compilers, stick with the C99 vararg macro syntax. * src/conf/cpu_conf.c (virCPUReportError): Use C99 rather than GNU vararg macro syntax. * src/conf/domain_conf.c (virDomainReportError): Likewise. * src/conf/domain_event.c (eventReportError): Likewise. * src/conf/interface_conf.c (virInterfaceReportError): Likewise. * src/conf/network_conf.c (virNetworkReportError): Likewise. * src/conf/node_device_conf.h (virNodeDeviceReportError): Likewise. * src/conf/secret_conf.h (virSecretReportError): Likewise. * src/conf/storage_conf.h (virStorageReportError): Likewise. * src/esx/esx_device_monitor.c (ESX_ERROR): Use C99 rather than GNU vararg macro syntax. * src/esx/esx_driver.c (ESX_ERROR): Likewise. * src/esx/esx_interface_driver.c (ESX_ERROR): Likewise. * src/esx/esx_network_driver.c (ESX_ERROR): Likewise. * src/esx/esx_secret_driver.c (ESX_ERROR): Likewise. * src/esx/esx_storage_driver.c (ESX_ERROR): Likewise. * src/esx/esx_util.c (ESX_ERROR): Likewise. * src/esx/esx_vi.c (ESX_VI_ERROR): Likewise. * src/esx/esx_vi_methods.c (ESX_VI_ERROR): Likewise. * src/esx/esx_vi_types.c (ESX_VI_ERROR): Likewise. * src/esx/esx_vmx.c (ESX_ERROR): Likewise. * src/util/hostusb.c (usbReportError): Use C99 rather than GNU vararg macro syntax. * src/util/json.c (virJSONError): Likewise. * src/util/macvtap.c (ReportError): Likewise. * src/util/pci.c (pciReportError): Likewise. * src/util/stats_linux.c (virStatsError): Likewise. * src/util/util.c (virUtilError): Likewise. * src/util/xml.c (virXMLError): Likewise. * src/xen/proxy_internal.c (virProxyError): Use C99 rather than GNU vararg macro syntax. * src/xen/sexpr.c (virSexprError): Likewise. * src/xen/xen_driver.c (xenUnifiedError): Likewise. * src/xen/xen_hypervisor.c (virXenError): Likewise. * src/xen/xen_inotify.c (virXenInotifyError): Likewise. * src/xen/xend_internal.c (virXendError): Likewise. * src/xen/xm_internal.c (xenXMError): Likewise. * src/xen/xs_internal.c (virXenStoreError): Likewise. * src/cpu/cpu.h (virCPUReportError): Use C99 rather than GNU vararg macro syntax. * src/datatypes.c (virLibConnError): Likewise. * src/interface/netcf_driver.c (interfaceReportError): Likewise. * src/libvirt.c (virLibStreamError): Likewise. * src/lxc/lxc_conf.h (lxcError): Likewise. * src/network/bridge_driver.c (networkReportError): Likewise. * src/nodeinfo.c (nodeReportError): Likewise. * src/opennebula/one_conf.h (oneError): Likewise. * src/openvz/openvz_conf.h (openvzError): Likewise. * src/phyp/phyp_driver.c (PHYP_ERROR): Likewise. * src/qemu/qemu_conf.h (qemuReportError): Likewise. * src/remote/remote_driver.c (errorf): Likewise. * src/security/security_driver.h (virSecurityReportError): Likewise. * src/test/test_driver.c (testError): Likewise. * src/uml/uml_conf.h (umlReportError): Likewise. * src/vbox/vbox_driver.c (vboxError): Likewise. * src/vbox/vbox_tmpl.c (vboxError): Likewise.
2010-03-01 23:38:28 +00:00
#define virLibStreamError(conn, code, ...) \
virReportErrorHelper(conn, VIR_FROM_NONE, code, __FILE__, \
build: consistently use C99 varargs macros Prior to this patch, there was an inconsistent mix between GNU and C99. For consistency, and potential portability to other compilers, stick with the C99 vararg macro syntax. * src/conf/cpu_conf.c (virCPUReportError): Use C99 rather than GNU vararg macro syntax. * src/conf/domain_conf.c (virDomainReportError): Likewise. * src/conf/domain_event.c (eventReportError): Likewise. * src/conf/interface_conf.c (virInterfaceReportError): Likewise. * src/conf/network_conf.c (virNetworkReportError): Likewise. * src/conf/node_device_conf.h (virNodeDeviceReportError): Likewise. * src/conf/secret_conf.h (virSecretReportError): Likewise. * src/conf/storage_conf.h (virStorageReportError): Likewise. * src/esx/esx_device_monitor.c (ESX_ERROR): Use C99 rather than GNU vararg macro syntax. * src/esx/esx_driver.c (ESX_ERROR): Likewise. * src/esx/esx_interface_driver.c (ESX_ERROR): Likewise. * src/esx/esx_network_driver.c (ESX_ERROR): Likewise. * src/esx/esx_secret_driver.c (ESX_ERROR): Likewise. * src/esx/esx_storage_driver.c (ESX_ERROR): Likewise. * src/esx/esx_util.c (ESX_ERROR): Likewise. * src/esx/esx_vi.c (ESX_VI_ERROR): Likewise. * src/esx/esx_vi_methods.c (ESX_VI_ERROR): Likewise. * src/esx/esx_vi_types.c (ESX_VI_ERROR): Likewise. * src/esx/esx_vmx.c (ESX_ERROR): Likewise. * src/util/hostusb.c (usbReportError): Use C99 rather than GNU vararg macro syntax. * src/util/json.c (virJSONError): Likewise. * src/util/macvtap.c (ReportError): Likewise. * src/util/pci.c (pciReportError): Likewise. * src/util/stats_linux.c (virStatsError): Likewise. * src/util/util.c (virUtilError): Likewise. * src/util/xml.c (virXMLError): Likewise. * src/xen/proxy_internal.c (virProxyError): Use C99 rather than GNU vararg macro syntax. * src/xen/sexpr.c (virSexprError): Likewise. * src/xen/xen_driver.c (xenUnifiedError): Likewise. * src/xen/xen_hypervisor.c (virXenError): Likewise. * src/xen/xen_inotify.c (virXenInotifyError): Likewise. * src/xen/xend_internal.c (virXendError): Likewise. * src/xen/xm_internal.c (xenXMError): Likewise. * src/xen/xs_internal.c (virXenStoreError): Likewise. * src/cpu/cpu.h (virCPUReportError): Use C99 rather than GNU vararg macro syntax. * src/datatypes.c (virLibConnError): Likewise. * src/interface/netcf_driver.c (interfaceReportError): Likewise. * src/libvirt.c (virLibStreamError): Likewise. * src/lxc/lxc_conf.h (lxcError): Likewise. * src/network/bridge_driver.c (networkReportError): Likewise. * src/nodeinfo.c (nodeReportError): Likewise. * src/opennebula/one_conf.h (oneError): Likewise. * src/openvz/openvz_conf.h (openvzError): Likewise. * src/phyp/phyp_driver.c (PHYP_ERROR): Likewise. * src/qemu/qemu_conf.h (qemuReportError): Likewise. * src/remote/remote_driver.c (errorf): Likewise. * src/security/security_driver.h (virSecurityReportError): Likewise. * src/test/test_driver.c (testError): Likewise. * src/uml/uml_conf.h (umlReportError): Likewise. * src/vbox/vbox_driver.c (vboxError): Likewise. * src/vbox/vbox_tmpl.c (vboxError): Likewise.
2010-03-01 23:38:28 +00:00
__FUNCTION__, __LINE__, __VA_ARGS__)
/**
* virLibSecretError:
* @secret: the secret if available
* @error: the error number
* @info: extra information string
*
* Handle an error at the secret level
*/
static void
virLibSecretError(virSecretPtr secret, virErrorNumber error, const char *info)
{
virConnectPtr conn = NULL;
const char *errmsg;
if (error == VIR_ERR_OK)
return;
errmsg = virErrorMsg(error, info);
if (error != VIR_ERR_INVALID_SECRET)
conn = secret->conn;
virRaiseError(conn, NULL, NULL, VIR_FROM_SECRET, error, VIR_ERR_ERROR,
errmsg, info, NULL, 0, 0, errmsg, info);
}
/**
* virLibNWFilterError:
* @conn: the connection if available
* @error: the error number
* @info: extra information string
*
* Handle an error at the connection level
*/
static void
virLibNWFilterError(virNWFilterPtr pool, virErrorNumber error,
const char *info)
{
virConnectPtr conn = NULL;
const char *errmsg;
if (error == VIR_ERR_OK)
return;
errmsg = virErrorMsg(error, info);
if (error != VIR_ERR_INVALID_NWFILTER)
conn = pool->conn;
virRaiseError(conn, NULL, NULL, VIR_FROM_NWFILTER, error, VIR_ERR_ERROR,
errmsg, info, NULL, 0, 0, errmsg, info);
}
/**
* virLibDomainSnapshotError:
* @snapshot: the snapshot if available
* @error: the error number
* @info: extra information string
*
* Handle an error at the domain snapshot level
*/
static void
virLibDomainSnapshotError(virDomainSnapshotPtr snapshot, virErrorNumber error, const char *info)
{
virConnectPtr conn = NULL;
const char *errmsg;
if (error == VIR_ERR_OK)
return;
errmsg = virErrorMsg(error, info);
if (error != VIR_ERR_INVALID_DOMAIN_SNAPSHOT)
conn = snapshot->domain->conn;
virRaiseError(conn, NULL, NULL, VIR_FROM_DOMAIN_SNAPSHOT, error, VIR_ERR_ERROR,
errmsg, info, NULL, 0, 0, errmsg, info);
}
/**
* virRegisterNetworkDriver:
* @driver: pointer to a network driver block
*
* Register a network virtualization driver
*
* Returns the driver priority or -1 in case of error.
*/
int
virRegisterNetworkDriver(virNetworkDriverPtr driver)
{
if (virInitialize() < 0)
return -1;
if (driver == NULL) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
return(-1);
}
if (virNetworkDriverTabCount >= MAX_DRIVERS) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
return(-1);
}
2008-11-21 12:16:08 +00:00
DEBUG ("registering %s as network driver %d",
driver->name, virNetworkDriverTabCount);
virNetworkDriverTab[virNetworkDriverTabCount] = driver;
return virNetworkDriverTabCount++;
}
/**
* virRegisterInterfaceDriver:
* @driver: pointer to an interface driver block
*
* Register an interface virtualization driver
*
* Returns the driver priority or -1 in case of error.
*/
int
virRegisterInterfaceDriver(virInterfaceDriverPtr driver)
{
if (virInitialize() < 0)
return -1;
if (driver == NULL) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
return(-1);
}
if (virInterfaceDriverTabCount >= MAX_DRIVERS) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
return(-1);
}
DEBUG ("registering %s as interface driver %d",
driver->name, virInterfaceDriverTabCount);
virInterfaceDriverTab[virInterfaceDriverTabCount] = driver;
return virInterfaceDriverTabCount++;
}
/**
* virRegisterStorageDriver:
* @driver: pointer to a storage driver block
*
* Register a storage virtualization driver
*
* Returns the driver priority or -1 in case of error.
*/
int
virRegisterStorageDriver(virStorageDriverPtr driver)
{
if (virInitialize() < 0)
return -1;
if (driver == NULL) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
return(-1);
}
if (virStorageDriverTabCount >= MAX_DRIVERS) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
return(-1);
}
2008-11-21 12:16:08 +00:00
DEBUG ("registering %s as storage driver %d",
driver->name, virStorageDriverTabCount);
virStorageDriverTab[virStorageDriverTabCount] = driver;
return virStorageDriverTabCount++;
}
/**
* virRegisterDeviceMonitor:
* @driver: pointer to a device monitor block
*
* Register a device monitor
*
* Returns the driver priority or -1 in case of error.
*/
int
virRegisterDeviceMonitor(virDeviceMonitorPtr driver)
{
if (virInitialize() < 0)
return -1;
if (driver == NULL) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
return(-1);
}
if (virDeviceMonitorTabCount >= MAX_DRIVERS) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
return(-1);
}
2008-11-21 12:46:39 +00:00
DEBUG ("registering %s as device driver %d",
driver->name, virDeviceMonitorTabCount);
virDeviceMonitorTab[virDeviceMonitorTabCount] = driver;
return virDeviceMonitorTabCount++;
}
/**
* virRegisterSecretDriver:
* @driver: pointer to a secret driver block
*
* Register a secret driver
*
* Returns the driver priority or -1 in case of error.
*/
int
virRegisterSecretDriver(virSecretDriverPtr driver)
{
if (virInitialize() < 0)
return -1;
if (driver == NULL) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
return(-1);
}
if (virSecretDriverTabCount >= MAX_DRIVERS) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
return(-1);
}
DEBUG ("registering %s as secret driver %d",
driver->name, virSecretDriverTabCount);
virSecretDriverTab[virSecretDriverTabCount] = driver;
return virSecretDriverTabCount++;
}
/**
* virRegisterNWFilterDriver:
* @driver: pointer to a network filter driver block
*
* Register a network filter virtualization driver
*
* Returns the driver priority or -1 in case of error.
*/
int
virRegisterNWFilterDriver(virNWFilterDriverPtr driver)
{
if (virInitialize() < 0)
return -1;
if (driver == NULL) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
return -1;
}
if (virNWFilterDriverTabCount >= MAX_DRIVERS) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
return -1;
}
DEBUG ("registering %s as network filter driver %d",
driver->name, virNWFilterDriverTabCount);
virNWFilterDriverTab[virNWFilterDriverTabCount] = driver;
return virNWFilterDriverTabCount++;
}
/**
* virRegisterDriver:
* @driver: pointer to a driver block
*
* Register a virtualization driver
*
* Returns the driver priority or -1 in case of error.
*/
int
virRegisterDriver(virDriverPtr driver)
{
if (virInitialize() < 0)
return -1;
if (driver == NULL) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
return(-1);
}
if (virDriverTabCount >= MAX_DRIVERS) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
return(-1);
}
if (driver->no < 0) {
virLibConnError
(NULL, VIR_ERR_INVALID_ARG,
"virRegisterDriver: tried to register an internal Xen driver");
return -1;
}
DEBUG ("registering %s as driver %d",
driver->name, virDriverTabCount);
virDriverTab[virDriverTabCount] = driver;
return virDriverTabCount++;
}
2008-06-26 09:37:51 +00:00
#ifdef WITH_LIBVIRTD
/**
* virRegisterStateDriver:
* @driver: pointer to a driver block
*
* Register a virtualization driver
*
* Returns the driver priority or -1 in case of error.
*/
int
virRegisterStateDriver(virStateDriverPtr driver)
{
if (virInitialize() < 0)
return -1;
if (driver == NULL) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
return(-1);
}
if (virStateDriverTabCount >= MAX_DRIVERS) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
return(-1);
}
virStateDriverTab[virStateDriverTabCount] = driver;
return virStateDriverTabCount++;
}
/**
* virStateInitialize:
* @privileged: set to 1 if running with root priviledge, 0 otherwise
*
* Initialize all virtualization drivers.
*
* Returns 0 if all succeed, -1 upon any failure.
*/
int virStateInitialize(int privileged) {
int i, ret = 0;
if (virInitialize() < 0)
return -1;
for (i = 0 ; i < virStateDriverTabCount ; i++) {
if (virStateDriverTab[i]->initialize &&
virStateDriverTab[i]->initialize(privileged) < 0) {
VIR_ERROR(_("Initialization of %s state driver failed"),
virStateDriverTab[i]->name);
ret = -1;
}
}
return ret;
}
/**
* virStateCleanup:
*
* Run each virtualization driver's cleanup method.
*
* Returns 0 if all succeed, -1 upon any failure.
*/
2008-11-17 11:03:25 +00:00
int virStateCleanup(void) {
int i, ret = 0;
for (i = 0 ; i < virStateDriverTabCount ; i++) {
if (virStateDriverTab[i]->cleanup &&
virStateDriverTab[i]->cleanup() < 0)
ret = -1;
}
return ret;
}
/**
* virStateReload:
*
* Run each virtualization driver's reload method.
*
* Returns 0 if all succeed, -1 upon any failure.
*/
2008-11-17 11:03:25 +00:00
int virStateReload(void) {
int i, ret = 0;
for (i = 0 ; i < virStateDriverTabCount ; i++) {
if (virStateDriverTab[i]->reload &&
virStateDriverTab[i]->reload() < 0)
ret = -1;
}
return ret;
}
/**
* virStateActive:
*
* Run each virtualization driver's "active" method.
*
* Returns 0 if none are active, 1 if at least one is.
*/
2008-11-17 11:03:25 +00:00
int virStateActive(void) {
int i, ret = 0;
for (i = 0 ; i < virStateDriverTabCount ; i++) {
if (virStateDriverTab[i]->active &&
virStateDriverTab[i]->active())
ret = 1;
}
return ret;
}
2008-06-26 09:37:51 +00:00
#endif
/**
* virGetVersion:
* @libVer: return value for the library version (OUT)
* @type: the type of connection/driver looked at
* @typeVer: return value for the version of the hypervisor (OUT)
*
* Provides two information back, @libVer is the version of the library
* while @typeVer will be the version of the hypervisor type @type against
* which the library was compiled. If @type is NULL, "Xen" is assumed, if
* @type is unknown or not available, an error code will be returned and
* @typeVer will be 0.
*
* Returns -1 in case of failure, 0 otherwise, and values for @libVer and
* @typeVer have the format major * 1,000,000 + minor * 1,000 + release.
*/
int
virGetVersion(unsigned long *libVer, const char *type,
unsigned long *typeVer)
{
DEBUG("libVir=%p, type=%s, typeVer=%p", libVer, type, typeVer);
if (!initialized)
if (virInitialize() < 0)
goto error;
if (libVer == NULL)
goto error;
*libVer = LIBVIR_VERSION_NUMBER;
if (typeVer != NULL) {
if (type == NULL)
type = "Xen";
/* FIXME: Add _proper_ type version handling for loadable driver modules... */
#ifdef WITH_DRIVER_MODULES
*typeVer = LIBVIR_VERSION_NUMBER;
#else
*typeVer = 0;
# if WITH_XEN
if (STRCASEEQ(type, "Xen"))
*typeVer = xenUnifiedVersion();
# endif
# if WITH_TEST
if (STRCASEEQ(type, "Test"))
*typeVer = LIBVIR_VERSION_NUMBER;
# endif
# if WITH_QEMU
if (STRCASEEQ(type, "QEMU"))
*typeVer = LIBVIR_VERSION_NUMBER;
# endif
# if WITH_LXC
if (STRCASEEQ(type, "LXC"))
*typeVer = LIBVIR_VERSION_NUMBER;
# endif
# if WITH_PHYP
if (STRCASEEQ(type, "phyp"))
*typeVer = LIBVIR_VERSION_NUMBER;
# endif
# if WITH_OPENVZ
if (STRCASEEQ(type, "OpenVZ"))
*typeVer = LIBVIR_VERSION_NUMBER;
# endif
# if WITH_VBOX
if (STRCASEEQ(type, "VBox"))
*typeVer = LIBVIR_VERSION_NUMBER;
# endif
# if WITH_UML
if (STRCASEEQ(type, "UML"))
*typeVer = LIBVIR_VERSION_NUMBER;
# endif
# if WITH_ONE
if (STRCASEEQ(type, "ONE"))
*typeVer = LIBVIR_VERSION_NUMBER;
# endif
# if WITH_ESX
if (STRCASEEQ(type, "ESX"))
*typeVer = LIBVIR_VERSION_NUMBER;
# endif
# if WITH_XENAPI
if (STRCASEEQ(type, "XenAPI"))
*typeVer = LIBVIR_VERSION_NUMBER;
# endif
# if WITH_REMOTE
if (STRCASEEQ(type, "Remote"))
*typeVer = remoteVersion();
# endif
if (*typeVer == 0) {
virLibConnError(NULL, VIR_ERR_NO_SUPPORT, type);
goto error;
}
#endif /* WITH_DRIVER_MODULES */
}
return (0);
error:
virDispatchError(NULL);
return -1;
}
static virConnectPtr
do_open (const char *name,
virConnectAuthPtr auth,
int flags)
{
int i, res;
virConnectPtr ret;
2009-01-20 12:01:45 +00:00
virResetLastError();
ret = virGetConnect();
if (ret == NULL)
return NULL;
/*
* If no URI is passed, then check for an environment string if not
* available probe the compiled in drivers to find a default hypervisor
* if detectable.
*/
if (!name || name[0] == '\0') {
char *defname = getenv("LIBVIRT_DEFAULT_URI");
if (defname && *defname) {
DEBUG("Using LIBVIRT_DEFAULT_URI %s", defname);
name = defname;
} else {
name = NULL;
}
}
if (name) {
/* Convert xen -> xen:/// for back compat */
if (STRCASEEQ(name, "xen"))
name = "xen:///";
/* Convert xen:// -> xen:/// because xmlParseURI cannot parse the
* former. This allows URIs such as xen://localhost to work.
*/
if (STREQ (name, "xen://"))
name = "xen:///";
ret->uri = xmlParseURI (name);
if (!ret->uri) {
virLibConnError (ret, VIR_ERR_INVALID_ARG,
_("could not parse connection URI"));
goto failed;
}
DEBUG("name \"%s\" to URI components:\n"
" scheme %s\n"
" opaque %s\n"
" authority %s\n"
" server %s\n"
" user %s\n"
" port %d\n"
" path %s\n",
name,
NULLSTR(ret->uri->scheme), NULLSTR(ret->uri->opaque),
NULLSTR(ret->uri->authority), NULLSTR(ret->uri->server),
NULLSTR(ret->uri->user), ret->uri->port,
NULLSTR(ret->uri->path));
} else {
DEBUG0("no name, allowing driver auto-select");
}
/* Cleansing flags */
ret->flags = flags & VIR_CONNECT_RO;
for (i = 0; i < virDriverTabCount; i++) {
/* We're going to probe the remote driver next. So we have already
* probed all other client-side-only driver before, but none of them
* accepted the URI.
* If the scheme corresponds to a known but disabled client-side-only
* driver then report a useful error, instead of a cryptic one about
* not being able to connect to libvirtd or not being able to find
* certificates. */
if (virDriverTab[i]->no == VIR_DRV_REMOTE &&
ret->uri != NULL && ret->uri->scheme != NULL &&
(
#ifndef WITH_PHYP
STRCASEEQ(ret->uri->scheme, "phyp") ||
#endif
#ifndef WITH_ESX
STRCASEEQ(ret->uri->scheme, "esx") ||
STRCASEEQ(ret->uri->scheme, "gsx") ||
#endif
#ifndef WITH_XENAPI
STRCASEEQ(ret->uri->scheme, "xenapi") ||
#endif
false)) {
virReportErrorHelper(NULL, VIR_FROM_NONE, VIR_ERR_INVALID_ARG,
__FILE__, __FUNCTION__, __LINE__,
_("libvirt was built without the '%s' driver"),
ret->uri->scheme);
goto failed;
}
DEBUG("trying driver %d (%s) ...",
i, virDriverTab[i]->name);
res = virDriverTab[i]->open (ret, auth, flags);
DEBUG("driver %d %s returned %s",
i, virDriverTab[i]->name,
res == VIR_DRV_OPEN_SUCCESS ? "SUCCESS" :
(res == VIR_DRV_OPEN_DECLINED ? "DECLINED" :
(res == VIR_DRV_OPEN_ERROR ? "ERROR" : "unknown status")));
if (res == VIR_DRV_OPEN_ERROR) goto failed;
else if (res == VIR_DRV_OPEN_SUCCESS) {
ret->driver = virDriverTab[i];
break;
}
}
if (!ret->driver) {
/* If we reach here, then all drivers declined the connection. */
virLibConnError (NULL, VIR_ERR_NO_CONNECT, name);
goto failed;
}
for (i = 0; i < virNetworkDriverTabCount; i++) {
res = virNetworkDriverTab[i]->open (ret, auth, flags);
DEBUG("network driver %d %s returned %s",
i, virNetworkDriverTab[i]->name,
res == VIR_DRV_OPEN_SUCCESS ? "SUCCESS" :
(res == VIR_DRV_OPEN_DECLINED ? "DECLINED" :
(res == VIR_DRV_OPEN_ERROR ? "ERROR" : "unknown status")));
if (res == VIR_DRV_OPEN_ERROR) {
if (STREQ(virNetworkDriverTab[i]->name, "remote")) {
virLibConnWarning (NULL, VIR_WAR_NO_NETWORK,
"Is the daemon running ?");
}
break;
} else if (res == VIR_DRV_OPEN_SUCCESS) {
ret->networkDriver = virNetworkDriverTab[i];
break;
}
}
for (i = 0; i < virInterfaceDriverTabCount; i++) {
res = virInterfaceDriverTab[i]->open (ret, auth, flags);
DEBUG("interface driver %d %s returned %s",
i, virInterfaceDriverTab[i]->name,
res == VIR_DRV_OPEN_SUCCESS ? "SUCCESS" :
(res == VIR_DRV_OPEN_DECLINED ? "DECLINED" :
(res == VIR_DRV_OPEN_ERROR ? "ERROR" : "unknown status")));
if (res == VIR_DRV_OPEN_ERROR) {
if (STREQ(virInterfaceDriverTab[i]->name, "remote")) {
virLibConnWarning (NULL, VIR_WAR_NO_INTERFACE,
"Is the daemon running ?");
}
break;
} else if (res == VIR_DRV_OPEN_SUCCESS) {
ret->interfaceDriver = virInterfaceDriverTab[i];
break;
}
}
/* Secondary driver for storage. Optional */
for (i = 0; i < virStorageDriverTabCount; i++) {
res = virStorageDriverTab[i]->open (ret, auth, flags);
DEBUG("storage driver %d %s returned %s",
i, virStorageDriverTab[i]->name,
res == VIR_DRV_OPEN_SUCCESS ? "SUCCESS" :
(res == VIR_DRV_OPEN_DECLINED ? "DECLINED" :
(res == VIR_DRV_OPEN_ERROR ? "ERROR" : "unknown status")));
if (res == VIR_DRV_OPEN_ERROR) {
if (STREQ(virStorageDriverTab[i]->name, "remote")) {
virLibConnWarning (NULL, VIR_WAR_NO_STORAGE,
"Is the daemon running ?");
}
break;
} else if (res == VIR_DRV_OPEN_SUCCESS) {
ret->storageDriver = virStorageDriverTab[i];
break;
}
}
/* Node driver (optional) */
for (i = 0; i < virDeviceMonitorTabCount; i++) {
res = virDeviceMonitorTab[i]->open (ret, auth, flags);
DEBUG("node driver %d %s returned %s",
i, virDeviceMonitorTab[i]->name,
res == VIR_DRV_OPEN_SUCCESS ? "SUCCESS" :
(res == VIR_DRV_OPEN_DECLINED ? "DECLINED" :
(res == VIR_DRV_OPEN_ERROR ? "ERROR" : "unknown status")));
if (res == VIR_DRV_OPEN_ERROR) {
if (STREQ(virDeviceMonitorTab[i]->name, "remote")) {
virLibConnWarning (NULL, VIR_WAR_NO_NODE,
"Is the libvirtd daemon running ?");
} else {
char *msg;
2008-12-23 13:03:29 +00:00
if (virAsprintf(&msg, "Is the %s daemon running?",
virDeviceMonitorTab[i]->name) > 0) {
virLibConnWarning (NULL, VIR_WAR_NO_NODE, msg);
VIR_FREE(msg);
}
}
break;
} else if (res == VIR_DRV_OPEN_SUCCESS) {
ret->deviceMonitor = virDeviceMonitorTab[i];
break;
}
}
/* Secret manipulation driver. Optional */
for (i = 0; i < virSecretDriverTabCount; i++) {
res = virSecretDriverTab[i]->open (ret, auth, flags);
DEBUG("secret driver %d %s returned %s",
i, virSecretDriverTab[i]->name,
res == VIR_DRV_OPEN_SUCCESS ? "SUCCESS" :
(res == VIR_DRV_OPEN_DECLINED ? "DECLINED" :
(res == VIR_DRV_OPEN_ERROR ? "ERROR" : "unknown status")));
if (res == VIR_DRV_OPEN_ERROR) {
if (STREQ(virSecretDriverTab[i]->name, "remote")) {
virLibConnWarning (NULL, VIR_WAR_NO_SECRET,
"Is the daemon running ?");
}
break;
} else if (res == VIR_DRV_OPEN_SUCCESS) {
ret->secretDriver = virSecretDriverTab[i];
break;
}
}
/* Network filter driver. Optional */
for (i = 0; i < virNWFilterDriverTabCount; i++) {
res = virNWFilterDriverTab[i]->open (ret, auth, flags);
DEBUG("nwfilter driver %d %s returned %s",
i, virNWFilterDriverTab[i]->name,
res == VIR_DRV_OPEN_SUCCESS ? "SUCCESS" :
(res == VIR_DRV_OPEN_DECLINED ? "DECLINED" :
(res == VIR_DRV_OPEN_ERROR ? "ERROR" : "unknown status")));
if (res == VIR_DRV_OPEN_ERROR) {
if (STREQ(virNWFilterDriverTab[i]->name, "remote")) {
virLibConnWarning (NULL, VIR_WAR_NO_NWFILTER,
_("Is the daemon running ?"));
}
break;
} else if (res == VIR_DRV_OPEN_SUCCESS) {
ret->nwfilterDriver = virNWFilterDriverTab[i];
break;
}
}
return ret;
failed:
virUnrefConnect(ret);
return NULL;
}
/**
* virConnectOpen:
* @name: URI of the hypervisor
*
* This function should be called first to get a connection to the
* Hypervisor and xen store
*
* Returns a pointer to the hypervisor connection or NULL in case of error
*
* If @name is NULL then probing will be done to determine a suitable
* default driver to activate. This involves trying each hypervisor
* in turn until one successfully opens. If the LIBVIRT_DEFAULT_URI
* environment variable is set, then it will be used in preference
* to probing for a driver.
*
* If connecting to an unprivileged hypervisor driver which requires
* the libvirtd daemon to be active, it will automatically be launched
* if not already running. This can be prevented by setting the
* environment variable LIBVIRT_AUTOSTART=0
*
* URIs are documented at http://libvirt.org/uri.html
*/
virConnectPtr
virConnectOpen (const char *name)
{
virConnectPtr ret = NULL;
if (!initialized)
if (virInitialize() < 0)
goto error;
DEBUG("name=%s", name);
ret = do_open (name, NULL, 0);
if (!ret)
goto error;
return ret;
error:
virDispatchError(NULL);
return NULL;
}
/**
* virConnectOpenReadOnly:
* @name: URI of the hypervisor
*
* This function should be called first to get a restricted connection to the
* library functionalities. The set of APIs usable are then restricted
* on the available methods to control the domains.
*
* See virConnectOpen for notes about environment variables which can
* have an effect on opening drivers
*
* Returns a pointer to the hypervisor connection or NULL in case of error
*
* URIs are documented at http://libvirt.org/uri.html
*/
virConnectPtr
virConnectOpenReadOnly(const char *name)
{
virConnectPtr ret = NULL;
if (!initialized)
if (virInitialize() < 0)
goto error;
DEBUG("name=%s", name);
ret = do_open (name, NULL, VIR_CONNECT_RO);
if (!ret)
goto error;
return ret;
error:
virDispatchError(NULL);
return NULL;
}
/**
* virConnectOpenAuth:
* @name: URI of the hypervisor
* @auth: Authenticate callback parameters
* @flags: Open flags
*
* This function should be called first to get a connection to the
* Hypervisor. If necessary, authentication will be performed fetching
* credentials via the callback
*
* See virConnectOpen for notes about environment variables which can
* have an effect on opening drivers
*
* Returns a pointer to the hypervisor connection or NULL in case of error
*
* URIs are documented at http://libvirt.org/uri.html
*/
virConnectPtr
virConnectOpenAuth(const char *name,
virConnectAuthPtr auth,
int flags)
{
virConnectPtr ret = NULL;
if (!initialized)
if (virInitialize() < 0)
goto error;
DEBUG("name=%s, auth=%p, flags=%d", NULLSTR(name), auth, flags);
ret = do_open (name, auth, flags);
if (!ret)
goto error;
return ret;
error:
virDispatchError(NULL);
return NULL;
2005-11-02 12:50:21 +00:00
}
/**
* virConnectClose:
2005-11-02 12:50:21 +00:00
* @conn: pointer to the hypervisor connection
*
* This function closes the connection to the Hypervisor. This should
* not be called if further interaction with the Hypervisor are needed
* especially if there is running domain which need further monitoring by
* the application.
*
* Returns 0 in case of success or -1 in case of error.
*/
int
virConnectClose(virConnectPtr conn)
{
int ret = -1;
DEBUG("conn=%p", conn);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
goto error;
2009-01-20 12:01:45 +00:00
}
ret = virUnrefConnect(conn);
if (ret < 0)
goto error;
return ret;
error:
virDispatchError(NULL);
return ret;
2005-11-02 12:50:21 +00:00
}
/**
* virConnectRef:
* @conn: the connection to hold a reference on
*
* Increment the reference count on the connection. For each
* additional call to this method, there shall be a corresponding
* call to virConnectClose to release the reference count, once
* the caller no longer needs the reference to this object.
*
* This method is typically useful for applications where multiple
* threads are using a connection, and it is required that the
* connection remain open until all threads have finished using
* it. ie, each new thread using a connection would increment
* the reference count.
*
* Returns 0 in case of success, -1 in case of failure
*/
int
virConnectRef(virConnectPtr conn)
{
if ((!VIR_IS_CONNECT(conn))) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
virDispatchError(NULL);
return(-1);
}
virMutexLock(&conn->lock);
DEBUG("conn=%p refs=%d", conn, conn->refs);
conn->refs++;
virMutexUnlock(&conn->lock);
return 0;
}
/*
* Not for public use. This function is part of the internal
* implementation of driver features in the remote case.
*/
int
2008-11-17 11:03:25 +00:00
virDrvSupportsFeature (virConnectPtr conn, int feature)
{
2009-01-20 12:01:45 +00:00
int ret;
DEBUG("conn=%p, feature=%d", conn, feature);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
2009-01-20 12:01:45 +00:00
}
2009-01-20 12:01:45 +00:00
ret = VIR_DRV_SUPPORTS_FEATURE (conn->driver, conn, feature);
if (ret < 0)
virDispatchError(conn);
2009-01-20 12:01:45 +00:00
return ret;
}
/**
* virConnectGetType:
* @conn: pointer to the hypervisor connection
*
* Get the name of the Hypervisor software used.
*
* Returns NULL in case of error, a static zero terminated string otherwise.
*
* See also:
* http://www.redhat.com/archives/libvir-list/2007-February/msg00096.html
*/
const char *
virConnectGetType(virConnectPtr conn)
{
const char *ret;
DEBUG("conn=%p", conn);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (conn->driver->type) {
ret = conn->driver->type (conn);
if (ret) return ret;
}
return conn->driver->name;
}
2005-11-02 12:50:21 +00:00
/**
* virConnectGetVersion:
2005-11-02 12:50:21 +00:00
* @conn: pointer to the hypervisor connection
* @hvVer: return value for the version of the running hypervisor (OUT)
2005-11-02 12:50:21 +00:00
*
* Get the version level of the Hypervisor running. This may work only with
* hypervisor call, i.e. with privileged access to the hypervisor, not
* with a Read-Only connection.
*
* Returns -1 in case of error, 0 otherwise. if the version can't be
* extracted by lack of capacities returns 0 and @hvVer is 0, otherwise
* @hvVer value is major * 1,000,000 + minor * 1,000 + release
*/
int
virConnectGetVersion(virConnectPtr conn, unsigned long *hvVer)
{
DEBUG("conn=%p, hvVer=%p", conn, hvVer);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
2009-01-20 12:01:45 +00:00
return -1;
}
if (hvVer == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->version) {
int ret = conn->driver->version (conn, hvVer);
if (ret < 0)
goto error;
return ret;
}
2007-02-14 01:40:09 +00:00
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return -1;
}
/**
* virConnectGetLibVersion:
* @conn: pointer to the hypervisor connection
* @libVer: returns the libvirt library version used on the connection (OUT)
*
* Provides @libVer, which is the version of libvirt used by the
* daemon running on the @conn host
*
* Returns -1 in case of failure, 0 otherwise, and values for @libVer have
* the format major * 1,000,000 + minor * 1,000 + release.
*/
int
virConnectGetLibVersion(virConnectPtr conn, unsigned long *libVer)
{
int ret = -1;
DEBUG("conn=%p, libVir=%p", conn, libVer);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
if (libVer == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->driver->libvirtVersion) {
ret = conn->driver->libvirtVersion(conn, libVer);
if (ret < 0)
goto error;
return ret;
}
*libVer = LIBVIR_VERSION_NUMBER;
return 0;
error:
virDispatchError(conn);
return ret;
}
/**
* virConnectGetHostname:
* @conn: pointer to a hypervisor connection
*
* This returns the system hostname on which the hypervisor is
* running (the result of the gethostname system call). If
* we are connected to a remote system, then this returns the
* hostname of the remote system.
*
* Returns the hostname which must be freed by the caller, or
* NULL if there was an error.
*/
char *
virConnectGetHostname (virConnectPtr conn)
{
DEBUG("conn=%p", conn);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->getHostname) {
char *ret = conn->driver->getHostname (conn);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return NULL;
}
/**
* virConnectGetURI:
* @conn: pointer to a hypervisor connection
*
* This returns the URI (name) of the hypervisor connection.
* Normally this is the same as or similar to the string passed
* to the virConnectOpen/virConnectOpenReadOnly call, but
* the driver may make the URI canonical. If name == NULL
* was passed to virConnectOpen, then the driver will return
* a non-NULL URI which can be used to connect to the same
* hypervisor later.
*
* Returns the URI string which must be freed by the caller, or
* NULL if there was an error.
*/
char *
virConnectGetURI (virConnectPtr conn)
{
char *name;
DEBUG("conn=%p", conn);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
name = (char *)xmlSaveUri(conn->uri);
if (!name) {
virReportOOMError();
2009-01-20 12:01:45 +00:00
goto error;
}
return name;
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
2009-01-20 12:01:45 +00:00
return NULL;
}
/**
* virConnectGetMaxVcpus:
* @conn: pointer to the hypervisor connection
* @type: value of the 'type' attribute in the <domain> element
*
* Provides the maximum number of virtual CPUs supported for a guest VM of a
* specific type. The 'type' parameter here corresponds to the 'type'
* attribute in the <domain> element of the XML.
*
* Returns the maximum of virtual CPU or -1 in case of error.
*/
int
virConnectGetMaxVcpus(virConnectPtr conn,
const char *type)
{
DEBUG("conn=%p, type=%s", conn, type);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
2009-01-20 12:01:45 +00:00
return -1;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->getMaxVcpus) {
int ret = conn->driver->getMaxVcpus (conn, type);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return -1;
}
/**
* virConnectListDomains:
* @conn: pointer to the hypervisor connection
* @ids: array to collect the list of IDs of active domains
* @maxids: size of @ids
*
* Collect the list of active domains, and store their ID in @maxids
*
* Returns the number of domain found or -1 in case of error
*/
int
virConnectListDomains(virConnectPtr conn, int *ids, int maxids)
{
DEBUG("conn=%p, ids=%p, maxids=%d", conn, ids, maxids);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
2009-01-20 12:01:45 +00:00
return -1;
}
if ((ids == NULL) || (maxids < 0)) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->listDomains) {
int ret = conn->driver->listDomains (conn, ids, maxids);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return -1;
}
2005-12-07 11:03:24 +00:00
/**
* virConnectNumOfDomains:
* @conn: pointer to the hypervisor connection
*
* Provides the number of active domains.
*
2005-12-07 11:03:24 +00:00
* Returns the number of domain found or -1 in case of error
*/
int
virConnectNumOfDomains(virConnectPtr conn)
{
DEBUG("conn=%p", conn);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
2005-12-16 18:41:46 +00:00
2009-01-20 12:01:45 +00:00
if (conn->driver->numOfDomains) {
int ret = conn->driver->numOfDomains (conn);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return -1;
2005-12-07 11:03:24 +00:00
}
/**
* virDomainGetConnect:
* @dom: pointer to a domain
*
* Provides the connection pointer associated with a domain. The
* reference counter on the connection is not increased by this
* call.
*
* WARNING: When writing libvirt bindings in other languages, do
* not use this function. Instead, store the connection and
* the domain object together.
*
* Returns the virConnectPtr or NULL in case of failure.
*/
virConnectPtr
virDomainGetConnect (virDomainPtr dom)
{
DEBUG("dom=%p", dom);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN (dom)) {
virLibDomainError (NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
return dom->conn;
}
/**
* virDomainCreateXML:
* @conn: pointer to the hypervisor connection
* @xmlDesc: string containing an XML description of the domain
* @flags: bitwise-or of supported virDomainCreateFlags
*
* Launch a new guest domain, based on an XML description similar
* to the one returned by virDomainGetXMLDesc()
* This function may requires privileged access to the hypervisor.
* The domain is not persistent, so its definition will disappear when it
* is destroyed, or if the host is restarted (see virDomainDefineXML() to
* define persistent domains).
*
* Returns a new domain object or NULL in case of failure
*/
virDomainPtr
virDomainCreateXML(virConnectPtr conn, const char *xmlDesc,
unsigned int flags)
{
DEBUG("conn=%p, xmlDesc=%s, flags=%d", conn, xmlDesc, flags);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (xmlDesc == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
if (conn->flags & VIR_CONNECT_RO) {
virLibConnError(conn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->domainCreateXML) {
virDomainPtr ret;
ret = conn->driver->domainCreateXML (conn, xmlDesc, flags);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return NULL;
}
/**
* virDomainCreateLinux:
* @conn: pointer to the hypervisor connection
* @xmlDesc: string containing an XML description of the domain
* @flags: callers should always pass 0
*
* Deprecated after 0.4.6.
* Renamed to virDomainCreateXML() providing identical functionality.
* This existing name will left indefinitely for API compatibility.
*
* Returns a new domain object or NULL in case of failure
*/
virDomainPtr
virDomainCreateLinux(virConnectPtr conn, const char *xmlDesc,
unsigned int flags)
{
return(virDomainCreateXML(conn, xmlDesc, flags));
}
/**
* virDomainLookupByID:
* @conn: pointer to the hypervisor connection
* @id: the domain ID number
*
* Try to find a domain based on the hypervisor ID number
* Note that this won't work for inactive domains which have an ID of -1,
* in that case a lookup based on the Name or UUId need to be done instead.
*
* Returns a new domain object or NULL in case of failure. If the
* domain cannot be found, then VIR_ERR_NO_DOMAIN error is raised.
*/
virDomainPtr
virDomainLookupByID(virConnectPtr conn, int id)
{
DEBUG("conn=%p, id=%d", conn, id);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (id < 0) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->domainLookupByID) {
virDomainPtr ret;
ret = conn->driver->domainLookupByID (conn, id);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return NULL;
}
/**
* virDomainLookupByUUID:
* @conn: pointer to the hypervisor connection
2006-05-22 14:38:33 +00:00
* @uuid: the raw UUID for the domain
*
* Try to lookup a domain on the given hypervisor based on its UUID.
*
* Returns a new domain object or NULL in case of failure. If the
* domain cannot be found, then VIR_ERR_NO_DOMAIN error is raised.
*/
virDomainPtr
virDomainLookupByUUID(virConnectPtr conn, const unsigned char *uuid)
{
DEBUG("conn=%p, uuid=%s", conn, uuid);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (uuid == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->domainLookupByUUID) {
virDomainPtr ret;
ret = conn->driver->domainLookupByUUID (conn, uuid);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return NULL;
}
2006-05-22 14:38:33 +00:00
/**
* virDomainLookupByUUIDString:
* @conn: pointer to the hypervisor connection
* @uuidstr: the string UUID for the domain
*
* Try to lookup a domain on the given hypervisor based on its UUID.
*
* Returns a new domain object or NULL in case of failure. If the
* domain cannot be found, then VIR_ERR_NO_DOMAIN error is raised.
2006-05-22 14:38:33 +00:00
*/
virDomainPtr
virDomainLookupByUUIDString(virConnectPtr conn, const char *uuidstr)
{
unsigned char uuid[VIR_UUID_BUFLEN];
DEBUG("conn=%p, uuidstr=%s", conn, uuidstr);
2009-01-20 12:01:45 +00:00
virResetLastError();
2006-05-22 14:38:33 +00:00
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
2006-05-22 14:38:33 +00:00
return (NULL);
}
if (uuidstr == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
2006-05-22 14:38:33 +00:00
}
if (virUUIDParse(uuidstr, uuid) < 0) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
2006-05-22 14:38:33 +00:00
}
2006-05-22 14:38:33 +00:00
return virDomainLookupByUUID(conn, &uuid[0]);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
2009-01-20 12:01:45 +00:00
return NULL;
2006-05-22 14:38:33 +00:00
}
/**
* virDomainLookupByName:
* @conn: pointer to the hypervisor connection
* @name: name for the domain
*
* Try to lookup a domain on the given hypervisor based on its name.
*
* Returns a new domain object or NULL in case of failure. If the
* domain cannot be found, then VIR_ERR_NO_DOMAIN error is raised.
*/
virDomainPtr
virDomainLookupByName(virConnectPtr conn, const char *name)
{
DEBUG("conn=%p, name=%s", conn, name);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (name == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->domainLookupByName) {
virDomainPtr dom;
dom = conn->driver->domainLookupByName (conn, name);
if (!dom)
goto error;
return dom;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return NULL;
}
/**
* virDomainDestroy:
* @domain: a domain object
*
* Destroy the domain object. The running instance is shutdown if not down
* already and all resources used by it are given back to the hypervisor. This
* does not free the associated virDomainPtr object.
* This function may require privileged access
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virDomainDestroy(virDomainPtr domain)
{
virConnectPtr conn;
DEBUG("domain=%p", domain);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = domain->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->domainDestroy) {
int ret;
ret = conn->driver->domainDestroy (domain);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return -1;
}
/**
* virDomainFree:
* @domain: a domain object
*
* Free the domain object. The running instance is kept alive.
* The data structure is freed and should not be used thereafter.
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virDomainFree(virDomainPtr domain)
{
DEBUG("domain=%p", domain);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (virUnrefDomain(domain) < 0) {
virDispatchError(NULL);
2009-01-20 12:01:45 +00:00
return -1;
}
return(0);
}
/**
* virDomainRef:
* @domain: the domain to hold a reference on
*
* Increment the reference count on the domain. For each
* additional call to this method, there shall be a corresponding
* call to virDomainFree to release the reference count, once
* the caller no longer needs the reference to this object.
*
* This method is typically useful for applications where multiple
* threads are using a connection, and it is required that the
* connection remain open until all threads have finished using
* it. ie, each new thread using a domain would increment
* the reference count.
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virDomainRef(virDomainPtr domain)
{
if ((!VIR_IS_CONNECTED_DOMAIN(domain))) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
virDispatchError(NULL);
return(-1);
}
virMutexLock(&domain->conn->lock);
DEBUG("domain=%p refs=%d", domain, domain->refs);
domain->refs++;
virMutexUnlock(&domain->conn->lock);
return 0;
}
/**
* virDomainSuspend:
* @domain: a domain object
*
* Suspends an active domain, the process is frozen without further access
* to CPU resources and I/O but the memory used by the domain at the
* hypervisor level will stay allocated. Use virDomainResume() to reactivate
* the domain.
* This function may requires privileged access.
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virDomainSuspend(virDomainPtr domain)
{
virConnectPtr conn;
DEBUG("domain=%p", domain);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = domain->conn;
2009-01-20 12:01:45 +00:00
if (conn->driver->domainSuspend) {
int ret;
ret = conn->driver->domainSuspend (domain);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainResume:
* @domain: a domain object
*
2009-10-26 23:02:46 +00:00
* Resume a suspended domain, the process is restarted from the state where
* it was frozen by calling virSuspendDomain().
* This function may requires privileged access
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virDomainResume(virDomainPtr domain)
{
virConnectPtr conn;
DEBUG("domain=%p", domain);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = domain->conn;
2009-01-20 12:01:45 +00:00
if (conn->driver->domainResume) {
int ret;
ret = conn->driver->domainResume (domain);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainSave:
* @domain: a domain object
* @to: path for the output file
*
* This method will suspend a domain and save its memory contents to
* a file on disk. After the call, if successful, the domain is not
* listed as running anymore (this may be a problem).
* Use virDomainRestore() to restore a domain after saving.
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virDomainSave(virDomainPtr domain, const char *to)
{
char filepath[4096];
virConnectPtr conn;
DEBUG("domain=%p, to=%s", domain, to);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = domain->conn;
if (to == NULL) {
virLibDomainError(domain, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
/*
* We must absolutize the file path as the save is done out of process
* TODO: check for URI when libxml2 is linked in.
*/
if (to[0] != '/') {
unsigned int len, t;
t = strlen(to);
if (getcwd(filepath, sizeof(filepath) - (t + 3)) == NULL)
return (-1);
len = strlen(filepath);
/* that should be covered by getcwd() semantic, but be 100% sure */
if (len > sizeof(filepath) - (t + 3))
return (-1);
filepath[len] = '/';
strcpy(&filepath[len + 1], to);
to = &filepath[0];
}
2009-01-20 12:01:45 +00:00
if (conn->driver->domainSave) {
int ret;
ret = conn->driver->domainSave (domain, to);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainRestore:
* @conn: pointer to the hypervisor connection
* @from: path to the
*
* This method will restore a domain saved to disk by virDomainSave().
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virDomainRestore(virConnectPtr conn, const char *from)
{
char filepath[4096];
DEBUG("conn=%p, from=%s", conn, from);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (conn->flags & VIR_CONNECT_RO) {
virLibConnError(conn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
if (from == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
/*
* We must absolutize the file path as the restore is done out of process
* TODO: check for URI when libxml2 is linked in.
*/
if (from[0] != '/') {
unsigned int len, t;
t = strlen(from);
2009-01-20 12:01:45 +00:00
if (getcwd(filepath, sizeof(filepath) - (t + 3)) == NULL) {
virLibConnError(conn, VIR_ERR_SYSTEM_ERROR,
_("cannot get working directory"));
goto error;
}
len = strlen(filepath);
/* that should be covered by getcwd() semantic, but be 100% sure */
2009-01-20 12:01:45 +00:00
if (len > sizeof(filepath) - (t + 3)) {
virLibConnError(conn, VIR_ERR_INTERNAL_ERROR,
_("path too long"));
goto error;
}
filepath[len] = '/';
strcpy(&filepath[len + 1], from);
from = &filepath[0];
}
2009-01-20 12:01:45 +00:00
if (conn->driver->domainRestore) {
int ret;
ret = conn->driver->domainRestore (conn, from);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return -1;
}
/**
* virDomainCoreDump:
* @domain: a domain object
* @to: path for the core file
* @flags: extra flags, currently unused
*
* This method will dump the core of a domain on a given file for analysis.
* Note that for remote Xen Daemon the file path will be interpreted in
* the remote host.
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virDomainCoreDump(virDomainPtr domain, const char *to, int flags)
{
char filepath[4096];
virConnectPtr conn;
DEBUG("domain=%p, to=%s, flags=%d", domain, to, flags);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = domain->conn;
if (to == NULL) {
virLibDomainError(domain, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
/*
* We must absolutize the file path as the save is done out of process
* TODO: check for URI when libxml2 is linked in.
*/
if (to[0] != '/') {
unsigned int len, t;
t = strlen(to);
2009-01-20 12:01:45 +00:00
if (getcwd(filepath, sizeof(filepath) - (t + 3)) == NULL) {
virLibDomainError(domain, VIR_ERR_SYSTEM_ERROR,
_("cannot get current directory"));
goto error;
}
len = strlen(filepath);
/* that should be covered by getcwd() semantic, but be 100% sure */
2009-01-20 12:01:45 +00:00
if (len > sizeof(filepath) - (t + 3)) {
virLibDomainError(domain, VIR_ERR_INTERNAL_ERROR,
_("path too long"));
goto error;
}
filepath[len] = '/';
strcpy(&filepath[len + 1], to);
to = &filepath[0];
}
2009-01-20 12:01:45 +00:00
if (conn->driver->domainCoreDump) {
int ret;
ret = conn->driver->domainCoreDump (domain, to, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainShutdown:
* @domain: a domain object
*
* Shutdown a domain, the domain object is still usable there after but
* the domain OS is being stopped. Note that the guest OS may ignore the
* request.
*
* TODO: should we add an option for reboot, knowing it may not be doable
* in the general case ?
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virDomainShutdown(virDomainPtr domain)
{
virConnectPtr conn;
DEBUG("domain=%p", domain);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = domain->conn;
2009-01-20 12:01:45 +00:00
if (conn->driver->domainShutdown) {
int ret;
ret = conn->driver->domainShutdown (domain);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainReboot:
* @domain: a domain object
* @flags: extra flags for the reboot operation, not used yet
*
* Reboot a domain, the domain object is still usable there after but
* the domain OS is being stopped for a restart.
* Note that the guest OS may ignore the request.
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virDomainReboot(virDomainPtr domain, unsigned int flags)
{
virConnectPtr conn;
DEBUG("domain=%p, flags=%u", domain, flags);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = domain->conn;
2009-01-20 12:01:45 +00:00
if (conn->driver->domainReboot) {
int ret;
ret = conn->driver->domainReboot (domain, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainGetName:
* @domain: a domain object
*
* Get the public name for that domain
*
* Returns a pointer to the name or NULL, the string need not be deallocated
* its lifetime will be the same as the domain object.
*/
const char *
virDomainGetName(virDomainPtr domain)
{
DEBUG("domain=%p", domain);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
return (domain->name);
}
/**
* virDomainGetUUID:
* @domain: a domain object
* @uuid: pointer to a VIR_UUID_BUFLEN bytes array
*
* Get the UUID for a domain
*
* Returns -1 in case of error, 0 in case of success
*/
int
virDomainGetUUID(virDomainPtr domain, unsigned char *uuid)
{
DEBUG("domain=%p, uuid=%p", domain, uuid);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (uuid == NULL) {
virLibDomainError(domain, VIR_ERR_INVALID_ARG, __FUNCTION__);
virDispatchError(domain->conn);
return (-1);
}
memcpy(uuid, &domain->uuid[0], VIR_UUID_BUFLEN);
return (0);
}
2006-05-22 14:38:33 +00:00
/**
* virDomainGetUUIDString:
* @domain: a domain object
* @buf: pointer to a VIR_UUID_STRING_BUFLEN bytes array
2006-05-22 14:38:33 +00:00
*
* Get the UUID for a domain as string. For more information about
2006-05-22 14:38:33 +00:00
* UUID see RFC4122.
*
* Returns -1 in case of error, 0 in case of success
*/
int
virDomainGetUUIDString(virDomainPtr domain, char *buf)
{
unsigned char uuid[VIR_UUID_BUFLEN];
DEBUG("domain=%p, buf=%p", domain, buf);
2009-01-20 12:01:45 +00:00
virResetLastError();
2006-05-22 14:38:33 +00:00
if (!VIR_IS_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
2006-05-22 14:38:33 +00:00
return (-1);
}
if (buf == NULL) {
virLibDomainError(domain, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
2006-05-22 14:38:33 +00:00
}
2006-05-22 14:38:33 +00:00
if (virDomainGetUUID(domain, &uuid[0]))
2009-01-20 12:01:45 +00:00
goto error;
2006-05-22 14:38:33 +00:00
2007-08-09 20:19:12 +00:00
virUUIDFormat(uuid, buf);
2006-05-22 14:38:33 +00:00
return (0);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
2009-01-20 12:01:45 +00:00
return -1;
2006-05-22 14:38:33 +00:00
}
/**
* virDomainGetID:
* @domain: a domain object
*
* Get the hypervisor ID number for the domain
*
* Returns the domain ID number or (unsigned int) -1 in case of error
*/
unsigned int
virDomainGetID(virDomainPtr domain)
{
DEBUG("domain=%p", domain);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return ((unsigned int) -1);
}
return (domain->id);
}
/**
* virDomainGetOSType:
* @domain: a domain object
*
* Get the type of domain operation system.
*
* Returns the new string or NULL in case of error, the string must be
* freed by the caller.
*/
char *
virDomainGetOSType(virDomainPtr domain)
{
virConnectPtr conn;
DEBUG("domain=%p", domain);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
conn = domain->conn;
2009-01-20 12:01:45 +00:00
if (conn->driver->domainGetOSType) {
char *ret;
ret = conn->driver->domainGetOSType (domain);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return NULL;
}
/**
* virDomainGetMaxMemory:
* @domain: a domain object or NULL
*
* Retrieve the maximum amount of physical memory allocated to a
* domain. If domain is NULL, then this get the amount of memory reserved
* to Domain0 i.e. the domain where the application runs.
*
* Returns the memory size in kilobytes or 0 in case of error.
*/
unsigned long
virDomainGetMaxMemory(virDomainPtr domain)
{
virConnectPtr conn;
DEBUG("domain=%p", domain);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (0);
}
conn = domain->conn;
2009-01-20 12:01:45 +00:00
if (conn->driver->domainGetMaxMemory) {
unsigned long ret;
ret = conn->driver->domainGetMaxMemory (domain);
if (ret == 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return 0;
}
/**
* virDomainSetMaxMemory:
* @domain: a domain object or NULL
* @memory: the memory size in kilobytes
*
* Dynamically change the maximum amount of physical memory allocated to a
* domain. If domain is NULL, then this change the amount of memory reserved
* to Domain0 i.e. the domain where the application runs.
* This function requires privileged access to the hypervisor.
*
* This command only changes the runtime configuration of the domain,
* so can only be called on an active domain.
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virDomainSetMaxMemory(virDomainPtr domain, unsigned long memory)
{
virConnectPtr conn;
DEBUG("domain=%p, memory=%lu", domain, memory);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
if (memory < 4096) {
virLibDomainError(domain, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = domain->conn;
2009-01-20 12:01:45 +00:00
if (conn->driver->domainSetMaxMemory) {
int ret;
ret = conn->driver->domainSetMaxMemory (domain, memory);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainSetMemory:
* @domain: a domain object or NULL
* @memory: the memory size in kilobytes
*
* Dynamically change the target amount of physical memory allocated to a
* domain. If domain is NULL, then this change the amount of memory reserved
* to Domain0 i.e. the domain where the application runs.
* This function may requires privileged access to the hypervisor.
*
* This command only changes the runtime configuration of the domain,
* so can only be called on an active domain.
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virDomainSetMemory(virDomainPtr domain, unsigned long memory)
{
virConnectPtr conn;
DEBUG("domain=%p, memory=%lu", domain, memory);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
if (memory < 4096) {
virLibDomainError(domain, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = domain->conn;
2009-01-20 12:01:45 +00:00
if (conn->driver->domainSetMemory) {
int ret;
ret = conn->driver->domainSetMemory (domain, memory);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainGetInfo:
* @domain: a domain object
* @info: pointer to a virDomainInfo structure allocated by the user
*
* Extract information about a domain. Note that if the connection
* used to get the domain is limited only a partial set of the information
* can be extracted.
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virDomainGetInfo(virDomainPtr domain, virDomainInfoPtr info)
{
virConnectPtr conn;
DEBUG("domain=%p, info=%p", domain, info);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (info == NULL) {
virLibDomainError(domain, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
memset(info, 0, sizeof(virDomainInfo));
conn = domain->conn;
2009-01-20 12:01:45 +00:00
if (conn->driver->domainGetInfo) {
int ret;
ret = conn->driver->domainGetInfo (domain, info);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainGetXMLDesc:
* @domain: a domain object
* @flags: an OR'ed set of virDomainXMLFlags
*
* Provide an XML description of the domain. The description may be reused
* later to relaunch the domain with virDomainCreateXML().
*
* Returns a 0 terminated UTF-8 encoded XML instance, or NULL in case of error.
* the caller must free() the returned value.
*/
char *
virDomainGetXMLDesc(virDomainPtr domain, int flags)
{
virConnectPtr conn;
DEBUG("domain=%p, flags=%d", domain, flags);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
conn = domain->conn;
if ((conn->flags & VIR_CONNECT_RO) && (flags & VIR_DOMAIN_XML_SECURE)) {
virLibConnError(conn, VIR_ERR_OPERATION_DENIED,
_("virDomainGetXMLDesc with secure flag"));
goto error;
}
flags &= VIR_DOMAIN_XML_FLAGS_MASK;
2009-01-20 12:01:45 +00:00
if (conn->driver->domainDumpXML) {
char *ret;
ret = conn->driver->domainDumpXML (domain, flags);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return NULL;
}
2009-05-21 13:46:35 +00:00
/**
* virConnectDomainXMLFromNative:
* @conn: a connection object
* @nativeFormat: configuration format importing from
* @nativeConfig: the configuration data to import
* @flags: currently unused, pass 0
*
* Reads native configuration data describing a domain, and
* generates libvirt domain XML. The format of the native
* data is hypervisor dependant.
*
* Returns a 0 terminated UTF-8 encoded XML instance, or NULL in case of error.
* the caller must free() the returned value.
*/
char *virConnectDomainXMLFromNative(virConnectPtr conn,
const char *nativeFormat,
const char *nativeConfig,
unsigned int flags)
{
DEBUG("conn=%p, format=%s config=%s flags=%u", conn, nativeFormat, nativeConfig, flags);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
2009-05-21 13:46:35 +00:00
return (NULL);
}
if (nativeFormat == NULL || nativeConfig == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
2009-05-21 13:46:35 +00:00
}
if (conn->driver->domainXMLFromNative) {
char *ret;
ret = conn->driver->domainXMLFromNative (conn,
nativeFormat,
nativeConfig,
flags);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
2009-05-21 13:46:35 +00:00
return NULL;
}
/**
* virConnectDomainXMLToNative:
* @conn: a connection object
* @nativeFormat: configuration format exporting to
* @domainXml: the domain configuration to export
* @flags: currently unused, pass 0
*
* Reads a domain XML configuration document, and generates
* generates a native configuration file describing the domain.
* The format of the native data is hypervisor dependant.
*
* Returns a 0 terminated UTF-8 encoded native config datafile, or NULL in case of error.
* the caller must free() the returned value.
*/
char *virConnectDomainXMLToNative(virConnectPtr conn,
const char *nativeFormat,
const char *domainXml,
unsigned int flags)
{
DEBUG("conn=%p, format=%s xml=%s flags=%u", conn, nativeFormat, domainXml, flags);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
2009-05-21 13:46:35 +00:00
return (NULL);
}
if (nativeFormat == NULL || domainXml == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
2009-05-21 13:46:35 +00:00
}
if (conn->driver->domainXMLToNative) {
char *ret;
ret = conn->driver->domainXMLToNative(conn,
nativeFormat,
domainXml,
flags);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
2009-05-21 13:46:35 +00:00
return NULL;
}
static virDomainPtr
virDomainMigrateVersion1 (virDomainPtr domain,
virConnectPtr dconn,
unsigned long flags,
const char *dname,
const char *uri,
unsigned long bandwidth)
{
virDomainPtr ddomain = NULL;
char *uri_out = NULL;
char *cookie = NULL;
int cookielen = 0, ret;
virDomainInfo info;
ret = virDomainGetInfo (domain, &info);
if (ret == 0 && info.state == VIR_DOMAIN_PAUSED) {
flags |= VIR_MIGRATE_PAUSED;
}
/* Prepare the migration.
*
* The destination host may return a cookie, or leave cookie as
* NULL.
*
* The destination host MUST set uri_out if uri_in is NULL.
*
* If uri_in is non-NULL, then the destination host may modify
* the URI by setting uri_out. If it does not wish to modify
* the URI, it should leave uri_out as NULL.
*/
if (dconn->driver->domainMigratePrepare
(dconn, &cookie, &cookielen, uri, &uri_out, flags, dname,
bandwidth) == -1)
goto done;
if (uri == NULL && uri_out == NULL) {
virLibConnError (domain->conn, VIR_ERR_INTERNAL_ERROR,
_("domainMigratePrepare did not set uri"));
goto done;
}
if (uri_out)
uri = uri_out; /* Did domainMigratePrepare change URI? */
assert (uri != NULL);
/* Perform the migration. The driver isn't supposed to return
* until the migration is complete.
*/
if (domain->conn->driver->domainMigratePerform
(domain, cookie, cookielen, uri, flags, dname, bandwidth) == -1)
goto done;
/* Get the destination domain and return it or error.
* 'domain' no longer actually exists at this point
* (or so we hope), but we still use the object in memory
* in order to get the name.
*/
dname = dname ? dname : domain->name;
if (dconn->driver->domainMigrateFinish)
ddomain = dconn->driver->domainMigrateFinish
(dconn, dname, cookie, cookielen, uri, flags);
else
ddomain = virDomainLookupByName (dconn, dname);
done:
VIR_FREE (uri_out);
VIR_FREE (cookie);
return ddomain;
}
static virDomainPtr
virDomainMigrateVersion2 (virDomainPtr domain,
virConnectPtr dconn,
unsigned long flags,
const char *dname,
const char *uri,
unsigned long bandwidth)
{
virDomainPtr ddomain = NULL;
char *uri_out = NULL;
char *cookie = NULL;
char *dom_xml = NULL;
int cookielen = 0, ret;
virDomainInfo info;
virErrorPtr orig_err = NULL;
/* Prepare the migration.
*
* The destination host may return a cookie, or leave cookie as
* NULL.
*
* The destination host MUST set uri_out if uri_in is NULL.
*
* If uri_in is non-NULL, then the destination host may modify
* the URI by setting uri_out. If it does not wish to modify
* the URI, it should leave uri_out as NULL.
*/
/* In version 2 of the protocol, the prepare step is slightly
* different. We fetch the domain XML of the source domain
* and pass it to Prepare2.
*/
if (!domain->conn->driver->domainDumpXML) {
virLibConnError (domain->conn, VIR_ERR_INTERNAL_ERROR, __FUNCTION__);
virDispatchError(domain->conn);
return NULL;
}
dom_xml = domain->conn->driver->domainDumpXML (domain,
VIR_DOMAIN_XML_SECURE |
VIR_DOMAIN_XML_UPDATE_CPU);
if (!dom_xml)
return NULL;
ret = virDomainGetInfo (domain, &info);
if (ret == 0 && info.state == VIR_DOMAIN_PAUSED) {
flags |= VIR_MIGRATE_PAUSED;
}
ret = dconn->driver->domainMigratePrepare2
(dconn, &cookie, &cookielen, uri, &uri_out, flags, dname,
bandwidth, dom_xml);
VIR_FREE (dom_xml);
if (ret == -1)
goto done;
if (uri == NULL && uri_out == NULL) {
virLibConnError (domain->conn, VIR_ERR_INTERNAL_ERROR,
_("domainMigratePrepare2 did not set uri"));
virDispatchError(domain->conn);
goto done;
}
if (uri_out)
uri = uri_out; /* Did domainMigratePrepare2 change URI? */
assert (uri != NULL);
/* Perform the migration. The driver isn't supposed to return
* until the migration is complete.
*/
ret = domain->conn->driver->domainMigratePerform
(domain, cookie, cookielen, uri, flags, dname, bandwidth);
/* Perform failed. Make sure Finish doesn't overwrite the error */
if (ret < 0)
orig_err = virSaveLastError();
/* In version 2 of the migration protocol, we pass the
* status code from the sender to the destination host,
* so it can do any cleanup if the migration failed.
*/
dname = dname ? dname : domain->name;
ddomain = dconn->driver->domainMigrateFinish2
(dconn, dname, cookie, cookielen, uri, flags, ret);
done:
if (orig_err) {
virSetError(orig_err);
virFreeError(orig_err);
}
VIR_FREE (uri_out);
VIR_FREE (cookie);
return ddomain;
}
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
/*
* This is sort of a migration v3
*
* In this version, the client does not talk to the destination
* libvirtd. The source libvirtd will still try to talk to the
* destination libvirtd though, and will do the prepare/perform/finish
* steps.
*/
static int
virDomainMigratePeer2Peer (virDomainPtr domain,
unsigned long flags,
const char *dname,
const char *uri,
unsigned long bandwidth)
{
if (!domain->conn->driver->domainMigratePerform) {
virLibConnError (domain->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
virDispatchError(domain->conn);
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
return -1;
}
/* Perform the migration. The driver isn't supposed to return
* until the migration is complete.
*/
return domain->conn->driver->domainMigratePerform(domain,
NULL, /* cookie */
0, /* cookielen */
uri,
flags,
dname,
bandwidth);
}
/*
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
* This is a variation on v1 & 2 migration
*
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
* This is for hypervisors which can directly handshake
* without any libvirtd involvement on destination either
* from client, or source libvirt.
*
* eg, XenD can talk direct to XenD, so libvirtd on dest
* does not need to be involved at all, or even running
*/
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
static int
virDomainMigrateDirect (virDomainPtr domain,
unsigned long flags,
const char *dname,
const char *uri,
unsigned long bandwidth)
{
if (!domain->conn->driver->domainMigratePerform) {
virLibConnError (domain->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
virDispatchError(domain->conn);
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
return -1;
}
/* Perform the migration. The driver isn't supposed to return
* until the migration is complete.
*/
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
return domain->conn->driver->domainMigratePerform(domain,
NULL, /* cookie */
0, /* cookielen */
uri,
flags,
dname,
bandwidth);
}
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
/**
* virDomainMigrate:
* @domain: a domain object
* @dconn: destination host (a connection object)
* @flags: flags
* @dname: (optional) rename domain to this at destination
* @uri: (optional) dest hostname/URI as seen from the source host
* @bandwidth: (optional) specify migration bandwidth limit in Mbps
*
* Migrate the domain object from its current host to the destination
* host given by dconn (a connection to the destination host).
*
* Flags may be one of more of the following:
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
* VIR_MIGRATE_LIVE Do not pause the VM during migration
* VIR_MIGRATE_PEER2PEER Direct connection between source & destination hosts
* VIR_MIGRATE_TUNNELLED Tunnel migration data over the libvirt RPC channel
* VIR_MIGRATE_PERSIST_DEST If the migration is successful, persist the domain
* on the destination host.
* VIR_MIGRATE_UNDEFINE_SOURCE If the migration is successful, undefine the
* domain on the source host.
* VIR_MIGRATE_PAUSED Leave the domain suspended on the remote side.
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
*
* VIR_MIGRATE_TUNNELLED requires that VIR_MIGRATE_PEER2PEER be set.
* Applications using the VIR_MIGRATE_PEER2PEER flag will probably
* prefer to invoke virDomainMigrateToURI, avoiding the need to
* open connection to the destination host themselves.
*
* If a hypervisor supports renaming domains during migration,
* then you may set the dname parameter to the new name (otherwise
* it keeps the same name). If this is not supported by the
* hypervisor, dname must be NULL or else you will get an error.
*
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
* If the VIR_MIGRATE_PEER2PEER flag is set, the uri parameter
* must be a valid libvirt connection URI, by which the source
* libvirt driver can connect to the destination libvirt. If
* omitted, the dconn connection object will be queried for its
* current URI.
*
* If the VIR_MIGRATE_PEER2PEER flag is NOT set, the URI parameter
* takes a hypervisor specific format. The hypervisor capabilities
* XML includes details of the support URI schemes. If omitted
* the dconn will be asked for a default URI.
*
* In either case it is typically only necessary to specify a
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
* URI if the destination host has multiple interfaces and a
* specific interface is required to transmit migration data.
*
* The maximum bandwidth (in Mbps) that will be used to do migration
* can be specified with the bandwidth parameter. If set to 0,
* libvirt will choose a suitable default. Some hypervisors do
* not support this feature and will return an error if bandwidth
* is not 0.
*
* To see which features are supported by the current hypervisor,
* see virConnectGetCapabilities, /capabilities/host/migration_features.
*
* There are many limitations on migration imposed by the underlying
* technology - for example it may not be possible to migrate between
* different processors even with the same architecture, or between
* different types of hypervisor.
*
* Returns the new domain object if the migration was successful,
* or NULL in case of error. Note that the new domain object
* exists in the scope of the destination connection (dconn).
*/
virDomainPtr
virDomainMigrate (virDomainPtr domain,
virConnectPtr dconn,
unsigned long flags,
const char *dname,
const char *uri,
unsigned long bandwidth)
{
virDomainPtr ddomain = NULL;
DEBUG("domain=%p, dconn=%p, flags=%lu, dname=%s, uri=%s, bandwidth=%lu",
domain, dconn, flags, NULLSTR(dname), NULLSTR(uri), bandwidth);
2009-01-20 12:01:45 +00:00
virResetLastError();
/* First checkout the source */
2009-01-20 12:01:45 +00:00
if (!VIR_IS_CONNECTED_DOMAIN (domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
/* Now checkout the destination */
if (!VIR_IS_CONNECT(dconn)) {
virLibConnError(domain->conn, VIR_ERR_INVALID_CONN, __FUNCTION__);
goto error;
}
if (dconn->flags & VIR_CONNECT_RO) {
/* NB, deliberately report error against source object, not dest */
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
if (flags & VIR_MIGRATE_PEER2PEER) {
if (VIR_DRV_SUPPORTS_FEATURE (domain->conn->driver, domain->conn,
VIR_DRV_FEATURE_MIGRATION_P2P)) {
char *dstURI = NULL;
if (uri == NULL) {
dstURI = virConnectGetURI(dconn);
if (!dstURI)
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
return NULL;
}
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
if (virDomainMigratePeer2Peer(domain, flags, dname, uri ? uri : dstURI, bandwidth) < 0) {
VIR_FREE(dstURI);
goto error;
}
VIR_FREE(dstURI);
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
ddomain = virDomainLookupByName (dconn, dname ? dname : domain->name);
} else {
/* This driver does not support peer to peer migration */
virLibConnError (domain->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
goto error;
}
} else {
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
if (flags & VIR_MIGRATE_TUNNELLED) {
virLibConnError(domain->conn, VIR_ERR_OPERATION_INVALID,
_("cannot perform tunnelled migration without using peer2peer flag"));
goto error;
}
/* Check that migration is supported by both drivers. */
if (VIR_DRV_SUPPORTS_FEATURE(domain->conn->driver, domain->conn,
VIR_DRV_FEATURE_MIGRATION_V1) &&
VIR_DRV_SUPPORTS_FEATURE(dconn->driver, dconn,
VIR_DRV_FEATURE_MIGRATION_V1))
ddomain = virDomainMigrateVersion1(domain, dconn, flags, dname, uri, bandwidth);
else if (VIR_DRV_SUPPORTS_FEATURE(domain->conn->driver, domain->conn,
VIR_DRV_FEATURE_MIGRATION_V2) &&
VIR_DRV_SUPPORTS_FEATURE(dconn->driver, dconn,
VIR_DRV_FEATURE_MIGRATION_V2))
ddomain = virDomainMigrateVersion2(domain, dconn, flags, dname, uri, bandwidth);
else {
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
/* This driver does not support any migration method */
virLibConnError(domain->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
goto error;
}
}
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
if (ddomain == NULL)
goto error;
return ddomain;
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
2009-01-20 12:01:45 +00:00
return NULL;
}
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
/**
* virDomainMigrateToURI:
* @domain: a domain object
* @duri: mandatory URI for the destination host
* @flags: flags
* @dname: (optional) rename domain to this at destination
* @bandwidth: (optional) specify migration bandwidth limit in Mbps
*
* Migrate the domain object from its current host to the destination
* host given by duri.
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
*
* Flags may be one of more of the following:
* VIR_MIGRATE_LIVE Do not pause the VM during migration
* VIR_MIGRATE_PEER2PEER Direct connection between source & destination hosts
* VIR_MIGRATE_TUNNELLED Tunnel migration data over the libvirt RPC channel
* VIR_MIGRATE_PERSIST_DEST If the migration is successful, persist the domain
* on the destination host.
* VIR_MIGRATE_UNDEFINE_SOURCE If the migration is successful, undefine the
* domain on the source host.
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
*
* The operation of this API hinges on the VIR_MIGRATE_PEER2PEER flag.
* If the VIR_MIGRATE_PEER2PEER flag is NOT set, the duri parameter
* takes a hypervisor specific format. The uri_transports element of the
* hypervisor capabilities XML includes details of the supported URI
* schemes. Not all hypervisors will support this mode of migration, so
* if the VIR_MIGRATE_PEER2PEER flag is not set, then it may be necessary
* to use the alternative virDomainMigrate API providing and explicit
* virConnectPtr for the destination host.
*
* If the VIR_MIGRATE_PEER2PEER flag IS set, the duri parameter
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
* must be a valid libvirt connection URI, by which the source
* libvirt driver can connect to the destination libvirt.
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
*
* VIR_MIGRATE_TUNNELLED requires that VIR_MIGRATE_PEER2PEER be set.
*
* If a hypervisor supports renaming domains during migration,
* the dname parameter specifies the new name for the domain.
* Setting dname to NULL keeps the domain name the same. If domain
* renaming is not supported by the hypervisor, dname must be NULL or
* else an error will be returned.
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
*
* The maximum bandwidth (in Mbps) that will be used to do migration
* can be specified with the bandwidth parameter. If set to 0,
* libvirt will choose a suitable default. Some hypervisors do
* not support this feature and will return an error if bandwidth
* is not 0.
*
* To see which features are supported by the current hypervisor,
* see virConnectGetCapabilities, /capabilities/host/migration_features.
*
* There are many limitations on migration imposed by the underlying
* technology - for example it may not be possible to migrate between
* different processors even with the same architecture, or between
* different types of hypervisor.
*
* Returns 0 if the migration succeeded, -1 upon error.
*/
int
virDomainMigrateToURI (virDomainPtr domain,
const char *duri,
unsigned long flags,
const char *dname,
unsigned long bandwidth)
{
DEBUG("domain=%p, duri=%p, flags=%lu, dname=%s, bandwidth=%lu",
domain, NULLSTR(duri), flags, NULLSTR(dname), bandwidth);
virResetLastError();
/* First checkout the source */
if (!VIR_IS_CONNECTED_DOMAIN (domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
return -1;
}
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (duri == NULL) {
virLibConnError (domain->conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (flags & VIR_MIGRATE_PEER2PEER) {
if (VIR_DRV_SUPPORTS_FEATURE (domain->conn->driver, domain->conn,
VIR_DRV_FEATURE_MIGRATION_P2P)) {
if (virDomainMigratePeer2Peer (domain, flags, dname, duri, bandwidth) < 0)
goto error;
} else {
/* No peer to peer migration supported */
virLibConnError (domain->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
goto error;
}
} else {
if (VIR_DRV_SUPPORTS_FEATURE (domain->conn->driver, domain->conn,
VIR_DRV_FEATURE_MIGRATION_DIRECT)) {
if (virDomainMigrateDirect (domain, flags, dname, duri, bandwidth) < 0)
goto error;
} else {
/* Cannot do a migration with only the perform step */
virLibConnError (domain->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
goto error;
}
}
return 0;
error:
virDispatchError(domain->conn);
Support a new peer-to-peer migration mode & public API Introduces several new public API options for migration - VIR_MIGRATE_PEER2PEER: With this flag the client only invokes the virDomainMigratePerform method, expecting the source host driver to do whatever is required to complete the entire migration process. - VIR_MIGRATE_TUNNELLED: With this flag the actual data for migration will be tunnelled over the libvirtd RPC channel. This requires that VIR_MIGRATE_PEER2PEER is also set. - virDomainMigrateToURI: This is variant of the existing virDomainMigrate method which does not require any virConnectPtr for the destination host. Given suitable driver support, this allows for all the same modes as virDomainMigrate() The URI for VIR_MIGRATE_PEER2PEER must be a valid libvirt URI. For non-p2p migration a hypervisor specific migration URI is used. virDomainMigrateToURI without a PEER2PEER flag is only support for Xen currently, and it involves XenD talking directly to XenD, no libvirtd involved at all. * include/libvirt/libvirt.h.in: Add VIR_MIGRATE_PEER2PEER flag for migration * src/libvirt_internal.h: Add feature flags for peer to peer migration (VIR_FEATURE_MIGRATE_P2P) and direct migration (VIR_MIGRATE_PEER2PEER mode) * src/libvirt.c: Implement support for VIR_MIGRATE_PEER2PEER and virDomainMigrateToURI APIs. * src/xen/xen_driver.c: Advertise support for DIRECT migration * src/xen/xend_internal.c: Add TODO item for p2p migration * src/libvirt_public.syms: Export virDomainMigrateToURI method * src/qemu/qemu_driver.c: Add support for PEER2PEER and migration, and adapt TUNNELLED migration. * tools/virsh.c: Add --p2p and --direct args and use the new virDomainMigrateToURI method where possible.
2009-09-17 17:10:04 +00:00
return -1;
}
/*
* Not for public use. This function is part of the internal
* implementation of migration in the remote case.
*/
int
2008-11-17 11:03:25 +00:00
virDomainMigratePrepare (virConnectPtr dconn,
char **cookie,
int *cookielen,
const char *uri_in,
char **uri_out,
unsigned long flags,
const char *dname,
unsigned long bandwidth)
{
2009-02-02 18:41:56 +00:00
VIR_DEBUG("dconn=%p, cookie=%p, cookielen=%p, uri_in=%s, uri_out=%p, "
"flags=%lu, dname=%s, bandwidth=%lu", dconn, cookie, cookielen,
NULLSTR(uri_in), uri_out, flags, NULLSTR(dname), bandwidth);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT (dconn)) {
virLibConnError (NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
if (dconn->flags & VIR_CONNECT_RO) {
virLibConnError(dconn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (dconn->driver->domainMigratePrepare) {
int ret;
ret = dconn->driver->domainMigratePrepare (dconn, cookie, cookielen,
uri_in, uri_out,
flags, dname, bandwidth);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (dconn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(dconn);
return -1;
}
/*
* Not for public use. This function is part of the internal
* implementation of migration in the remote case.
*/
int
2008-11-17 11:03:25 +00:00
virDomainMigratePerform (virDomainPtr domain,
const char *cookie,
int cookielen,
const char *uri,
unsigned long flags,
const char *dname,
unsigned long bandwidth)
{
virConnectPtr conn;
2009-02-02 18:41:56 +00:00
VIR_DEBUG("domain=%p, cookie=%p, cookielen=%d, uri=%s, flags=%lu, "
"dname=%s, bandwidth=%lu", domain, cookie, cookielen, uri, flags,
NULLSTR(dname), bandwidth);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN (domain)) {
virLibDomainError (NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
conn = domain->conn;
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->domainMigratePerform) {
int ret;
ret = conn->driver->domainMigratePerform (domain, cookie, cookielen,
uri,
flags, dname, bandwidth);
if (ret < 0)
goto error;
return ret;
}
virLibDomainError (domain, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/*
* Not for public use. This function is part of the internal
* implementation of migration in the remote case.
*/
virDomainPtr
2008-11-17 11:03:25 +00:00
virDomainMigrateFinish (virConnectPtr dconn,
const char *dname,
const char *cookie,
int cookielen,
const char *uri,
unsigned long flags)
{
2009-02-02 18:41:56 +00:00
VIR_DEBUG("dconn=%p, dname=%s, cookie=%p, cookielen=%d, uri=%s, "
"flags=%lu", dconn, NULLSTR(dname), cookie, cookielen,
uri, flags);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT (dconn)) {
virLibConnError (NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
if (dconn->flags & VIR_CONNECT_RO) {
virLibConnError(dconn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (dconn->driver->domainMigrateFinish) {
virDomainPtr ret;
ret = dconn->driver->domainMigrateFinish (dconn, dname,
cookie, cookielen,
uri, flags);
if (!ret)
goto error;
return ret;
}
virLibConnError (dconn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(dconn);
return NULL;
}
/*
* Not for public use. This function is part of the internal
* implementation of migration in the remote case.
*/
int
2008-11-17 11:03:25 +00:00
virDomainMigratePrepare2 (virConnectPtr dconn,
char **cookie,
int *cookielen,
const char *uri_in,
char **uri_out,
unsigned long flags,
const char *dname,
unsigned long bandwidth,
const char *dom_xml)
{
2009-02-02 18:41:56 +00:00
VIR_DEBUG("dconn=%p, cookie=%p, cookielen=%p, uri_in=%s, uri_out=%p,"
"flags=%lu, dname=%s, bandwidth=%lu, dom_xml=%s", dconn,
cookie, cookielen, uri_in, uri_out, flags, NULLSTR(dname),
bandwidth, dom_xml);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT (dconn)) {
virLibConnError (NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
if (dconn->flags & VIR_CONNECT_RO) {
virLibConnError(dconn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (dconn->driver->domainMigratePrepare2) {
int ret;
ret = dconn->driver->domainMigratePrepare2 (dconn, cookie, cookielen,
uri_in, uri_out,
flags, dname, bandwidth,
dom_xml);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (dconn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(dconn);
return -1;
}
/*
* Not for public use. This function is part of the internal
* implementation of migration in the remote case.
*/
virDomainPtr
2008-11-17 11:03:25 +00:00
virDomainMigrateFinish2 (virConnectPtr dconn,
const char *dname,
const char *cookie,
int cookielen,
const char *uri,
unsigned long flags,
int retcode)
{
2009-02-02 18:41:56 +00:00
VIR_DEBUG("dconn=%p, dname=%s, cookie=%p, cookielen=%d, uri=%s, "
"flags=%lu, retcode=%d", dconn, NULLSTR(dname), cookie,
cookielen, uri, flags, retcode);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT (dconn)) {
virLibConnError (NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
if (dconn->flags & VIR_CONNECT_RO) {
virLibConnError(dconn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (dconn->driver->domainMigrateFinish2) {
virDomainPtr ret;
ret = dconn->driver->domainMigrateFinish2 (dconn, dname,
cookie, cookielen,
uri, flags,
retcode);
if (!ret)
goto error;
return ret;
}
virLibConnError (dconn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(dconn);
return NULL;
}
/*
* Not for public use. This function is part of the internal
* implementation of migration in the remote case.
*/
int
virDomainMigratePrepareTunnel(virConnectPtr conn,
virStreamPtr st,
unsigned long flags,
const char *dname,
unsigned long bandwidth,
const char *dom_xml)
{
VIR_DEBUG("conn=%p, stream=%p, flags=%lu, dname=%s, "
"bandwidth=%lu, dom_xml=%s", conn, st, flags,
NULLSTR(dname), bandwidth, dom_xml);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
if (conn->flags & VIR_CONNECT_RO) {
virLibConnError(conn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (conn != st->conn) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->driver->domainMigratePrepareTunnel) {
int rv = conn->driver->domainMigratePrepareTunnel(conn, st,
flags, dname,
bandwidth, dom_xml);
if (rv < 0)
goto error;
return rv;
}
virLibConnError(conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virNodeGetInfo:
* @conn: pointer to the hypervisor connection
* @info: pointer to a virNodeInfo structure allocated by the user
*
* Extract hardware information about the node.
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virNodeGetInfo(virConnectPtr conn, virNodeInfoPtr info)
{
DEBUG("conn=%p, info=%p", conn, info);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (info == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->nodeGetInfo) {
int ret;
ret = conn->driver->nodeGetInfo (conn, info);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return -1;
}
/**
* virConnectGetCapabilities:
* @conn: pointer to the hypervisor connection
*
* Provides capabilities of the hypervisor / driver.
*
* Returns NULL in case of error, or an XML string
* defining the capabilities.
* The client must free the returned string after use.
*/
char *
virConnectGetCapabilities (virConnectPtr conn)
{
DEBUG("conn=%p", conn);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT (conn)) {
virLibConnError (NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->getCapabilities) {
char *ret;
ret = conn->driver->getCapabilities (conn);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return NULL;
}
/**
* virNodeGetFreeMemory:
* @conn: pointer to the hypervisor connection
*
* provides the free memory available on the Node
* Note: most libvirt APIs provide memory sizes in kilobytes, but in this
* function the returned value is in bytes. Divide by 1024 as necessary.
*
* Returns the available free memory in bytes or 0 in case of error
*/
unsigned long long
virNodeGetFreeMemory(virConnectPtr conn)
{
DEBUG("conn=%p", conn);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT (conn)) {
virLibConnError (NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return 0;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->getFreeMemory) {
unsigned long long ret;
ret = conn->driver->getFreeMemory (conn);
if (ret == 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return 0;
}
/**
* virDomainGetSchedulerType:
* @domain: pointer to domain object
* @nparams: number of scheduler parameters(return value)
*
* Get the scheduler type.
*
* Returns NULL in case of error. The caller must free the returned string.
*/
char *
virDomainGetSchedulerType(virDomainPtr domain, int *nparams)
{
virConnectPtr conn;
char *schedtype;
DEBUG("domain=%p, nparams=%p", domain, nparams);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
conn = domain->conn;
if (conn->driver->domainGetSchedulerType){
schedtype = conn->driver->domainGetSchedulerType (domain, nparams);
2009-01-20 12:01:45 +00:00
if (!schedtype)
goto error;
return schedtype;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return NULL;
}
/**
* virDomainGetSchedulerParameters:
* @domain: pointer to domain object
* @params: pointer to scheduler parameter object
* (return value)
* @nparams: pointer to number of scheduler parameter
* (this value should be same than the returned value
* nparams of virDomainGetSchedulerType)
*
* Get the scheduler parameters, the @params array will be filled with the
* values.
*
* Returns -1 in case of error, 0 in case of success.
*/
int
virDomainGetSchedulerParameters(virDomainPtr domain,
virSchedParameterPtr params, int *nparams)
{
virConnectPtr conn;
DEBUG("domain=%p, params=%p, nparams=%p", domain, params, nparams);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
conn = domain->conn;
2009-01-20 12:01:45 +00:00
if (conn->driver->domainGetSchedulerParameters) {
int ret;
ret = conn->driver->domainGetSchedulerParameters (domain, params, nparams);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainSetSchedulerParameters:
* @domain: pointer to domain object
* @params: pointer to scheduler parameter objects
* @nparams: number of scheduler parameter
* (this value should be same or less than the returned value
* nparams of virDomainGetSchedulerType)
*
* Change the scheduler parameters
*
* Returns -1 in case of error, 0 in case of success.
*/
int
virDomainSetSchedulerParameters(virDomainPtr domain,
virSchedParameterPtr params, int nparams)
{
virConnectPtr conn;
DEBUG("domain=%p, params=%p, nparams=%d", domain, params, nparams);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = domain->conn;
2009-01-20 12:01:45 +00:00
if (conn->driver->domainSetSchedulerParameters) {
int ret;
ret = conn->driver->domainSetSchedulerParameters (domain, params, nparams);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainBlockStats:
* @dom: pointer to the domain object
* @path: path to the block device
* @stats: block device stats (returned)
* @size: size of stats structure
*
* This function returns block device (disk) stats for block
* devices attached to the domain.
*
* The path parameter is the name of the block device. Get this
* by calling virDomainGetXMLDesc and finding the <target dev='...'>
* attribute within //domain/devices/disk. (For example, "xvda").
*
* Domains may have more than one block device. To get stats for
* each you should make multiple calls to this function.
*
* Individual fields within the stats structure may be returned
* as -1, which indicates that the hypervisor does not support
* that particular statistic.
*
* Returns: 0 in case of success or -1 in case of failure.
*/
int
virDomainBlockStats (virDomainPtr dom, const char *path,
virDomainBlockStatsPtr stats, size_t size)
{
virConnectPtr conn;
struct _virDomainBlockStats stats2 = { -1, -1, -1, -1, -1 };
DEBUG("domain=%p, path=%s, stats=%p, size=%zi", dom, path, stats, size);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN (dom)) {
virLibDomainError (NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
if (!path || !stats || size > sizeof stats2) {
2009-01-20 12:01:45 +00:00
virLibDomainError (dom, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
conn = dom->conn;
if (conn->driver->domainBlockStats) {
if (conn->driver->domainBlockStats (dom, path, &stats2) == -1)
2009-01-20 12:01:45 +00:00
goto error;
memcpy (stats, &stats2, size);
return 0;
}
virLibDomainError (dom, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(dom->conn);
return -1;
}
/**
* virDomainInterfaceStats:
* @dom: pointer to the domain object
* @path: path to the interface
* @stats: network interface stats (returned)
* @size: size of stats structure
*
* This function returns network interface stats for interfaces
* attached to the domain.
*
* The path parameter is the name of the network interface.
*
2009-10-26 23:02:46 +00:00
* Domains may have more than one network interface. To get stats for
* each you should make multiple calls to this function.
*
* Individual fields within the stats structure may be returned
* as -1, which indicates that the hypervisor does not support
* that particular statistic.
*
* Returns: 0 in case of success or -1 in case of failure.
*/
int
virDomainInterfaceStats (virDomainPtr dom, const char *path,
virDomainInterfaceStatsPtr stats, size_t size)
{
virConnectPtr conn;
struct _virDomainInterfaceStats stats2 = { -1, -1, -1, -1,
-1, -1, -1, -1 };
DEBUG("domain=%p, path=%s, stats=%p, size=%zi", dom, path, stats, size);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN (dom)) {
virLibDomainError (NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
if (!path || !stats || size > sizeof stats2) {
2009-01-20 12:01:45 +00:00
virLibDomainError (dom, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
conn = dom->conn;
if (conn->driver->domainInterfaceStats) {
if (conn->driver->domainInterfaceStats (dom, path, &stats2) == -1)
2009-01-20 12:01:45 +00:00
goto error;
memcpy (stats, &stats2, size);
return 0;
}
virLibDomainError (dom, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(dom->conn);
return -1;
}
/**
* virDomainMemoryStats:
* @dom: pointer to the domain object
* @stats: nr_stats-sized array of stat structures (returned)
* @nr_stats: number of memory statistics requested
* @flags: unused, always pass 0
*
* This function provides memory statistics for the domain.
*
* Up to 'nr_stats' elements of 'stats' will be populated with memory statistics
* from the domain. Only statistics supported by the domain, the driver, and
* this version of libvirt will be returned.
*
* Memory Statistics:
*
* VIR_DOMAIN_MEMORY_STAT_SWAP_IN:
* The total amount of data read from swap space (in kb).
* VIR_DOMAIN_MEMORY_STAT_SWAP_OUT:
* The total amount of memory written out to swap space (in kb).
* VIR_DOMAIN_MEMORY_STAT_MAJOR_FAULT:
* The number of page faults that required disk IO to service.
* VIR_DOMAIN_MEMORY_STAT_MINOR_FAULT:
* The number of page faults serviced without disk IO.
* VIR_DOMAIN_MEMORY_STAT_UNUSED:
* The amount of memory which is not being used for any purpose (in kb).
* VIR_DOMAIN_MEMORY_STAT_AVAILABLE:
* The total amount of memory available to the domain's OS (in kb).
*
* Returns: The number of stats provided or -1 in case of failure.
*/
int virDomainMemoryStats (virDomainPtr dom, virDomainMemoryStatPtr stats,
unsigned int nr_stats, unsigned int flags)
{
virConnectPtr conn;
unsigned long nr_stats_ret = 0;
DEBUG("domain=%p, stats=%p, nr_stats=%u", dom, stats, nr_stats);
if (flags != 0) {
virLibDomainError (dom, VIR_ERR_INVALID_ARG,
_("flags must be zero"));
goto error;
}
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN (dom)) {
virLibDomainError (NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
if (!stats || nr_stats == 0)
return 0;
if (nr_stats > VIR_DOMAIN_MEMORY_STAT_NR)
nr_stats = VIR_DOMAIN_MEMORY_STAT_NR;
conn = dom->conn;
if (conn->driver->domainMemoryStats) {
nr_stats_ret = conn->driver->domainMemoryStats (dom, stats, nr_stats);
if (nr_stats_ret == -1)
goto error;
return nr_stats_ret;
}
virLibDomainError (dom, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(dom->conn);
return -1;
}
/**
* virDomainBlockPeek:
* @dom: pointer to the domain object
* @path: path to the block device
* @offset: offset within block device
* @size: size to read
* @buffer: return buffer (must be at least size bytes)
* @flags: unused, always pass 0
*
* This function allows you to read the contents of a domain's
* disk device.
*
* Typical uses for this are to determine if the domain has
* written a Master Boot Record (indicating that the domain
* has completed installation), or to try to work out the state
* of the domain's filesystems.
*
* (Note that in the local case you might try to open the
* block device or file directly, but that won't work in the
* remote case, nor if you don't have sufficient permission.
* Hence the need for this call).
*
* 'path' must be a device or file corresponding to the domain.
* In other words it must be the precise string returned in
* a <disk><source dev='...'/></disk> from
* virDomainGetXMLDesc.
*
* 'offset' and 'size' represent an area which must lie entirely
* within the device or file. 'size' may be 0 to test if the
* call would succeed.
*
* 'buffer' is the return buffer and must be at least 'size' bytes.
*
* NB. The remote driver imposes a 64K byte limit on 'size'.
* For your program to be able to work reliably over a remote
* connection you should split large requests to <= 65536 bytes.
*
* Returns: 0 in case of success or -1 in case of failure.
*/
int
virDomainBlockPeek (virDomainPtr dom,
const char *path,
unsigned long long offset /* really 64 bits */,
size_t size,
void *buffer,
unsigned int flags)
{
virConnectPtr conn;
DEBUG("domain=%p, path=%s, offset=%lld, size=%zi, buffer=%p",
dom, path, offset, size, buffer);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN (dom)) {
virLibDomainError (NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
conn = dom->conn;
if (dom->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(dom, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
if (!path) {
virLibDomainError (dom, VIR_ERR_INVALID_ARG,
_("path is NULL"));
2009-01-20 12:01:45 +00:00
goto error;
}
if (flags != 0) {
virLibDomainError (dom, VIR_ERR_INVALID_ARG,
_("flags must be zero"));
2009-01-20 12:01:45 +00:00
goto error;
}
/* Allow size == 0 as an access test. */
if (size > 0 && !buffer) {
virLibDomainError (dom, VIR_ERR_INVALID_ARG,
_("buffer is NULL"));
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->domainBlockPeek) {
int ret;
ret =conn->driver->domainBlockPeek (dom, path, offset, size,
buffer, flags);
if (ret < 0)
goto error;
return ret;
}
virLibDomainError (dom, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(dom->conn);
return -1;
}
/**
* virDomainMemoryPeek:
* @dom: pointer to the domain object
* @start: start of memory to peek
* @size: size of memory to peek
* @buffer: return buffer (must be at least size bytes)
* @flags: flags, see below
*
* This function allows you to read the contents of a domain's
* memory.
*
* The memory which is read is controlled by the 'start', 'size'
* and 'flags' parameters.
*
* If 'flags' is VIR_MEMORY_VIRTUAL then the 'start' and 'size'
* parameters are interpreted as virtual memory addresses for
* whichever task happens to be running on the domain at the
* moment. Although this sounds haphazard it is in fact what
* you want in order to read Linux kernel state, because it
* ensures that pointers in the kernel image can be interpreted
* coherently.
*
* 'buffer' is the return buffer and must be at least 'size' bytes.
* 'size' may be 0 to test if the call would succeed.
*
* NB. The remote driver imposes a 64K byte limit on 'size'.
* For your program to be able to work reliably over a remote
* connection you should split large requests to <= 65536 bytes.
*
* Returns: 0 in case of success or -1 in case of failure.
*/
int
virDomainMemoryPeek (virDomainPtr dom,
unsigned long long start /* really 64 bits */,
size_t size,
void *buffer,
unsigned int flags)
{
virConnectPtr conn;
DEBUG ("domain=%p, start=%lld, size=%zi, buffer=%p, flags=%d",
dom, start, size, buffer, flags);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN (dom)) {
virLibDomainError (NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
conn = dom->conn;
if (dom->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(dom, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
/* Note on access to physical memory: A VIR_MEMORY_PHYSICAL flag is
* a possibility. However it isn't really useful unless the caller
* can also access registers, particularly CR3 on x86 in order to
* get the Page Table Directory. Since registers are different on
* every architecture, that would imply another call to get the
* machine registers.
*
* The QEMU driver handles VIR_MEMORY_VIRTUAL, mapping it
* to the qemu 'memsave' command which does the virtual to physical
* mapping inside qemu.
*
* The QEMU driver also handles VIR_MEMORY_PHYSICAL, mapping it
* to the qemu 'pmemsave' command.
*
* At time of writing there is no Xen driver. However the Xen
* hypervisor only lets you map physical pages from other domains,
* and so the Xen driver would have to do the virtual to physical
* mapping by chasing 2, 3 or 4-level page tables from the PTD.
* There is example code in libxc (xc_translate_foreign_address)
* which does this, although we cannot copy this code directly
* because of incompatible licensing.
*/
if (flags != VIR_MEMORY_VIRTUAL && flags != VIR_MEMORY_PHYSICAL) {
virLibDomainError (dom, VIR_ERR_INVALID_ARG,
_("flags parameter must be VIR_MEMORY_VIRTUAL or VIR_MEMORY_PHYSICAL"));
2009-01-20 12:01:45 +00:00
goto error;
}
/* Allow size == 0 as an access test. */
if (size > 0 && !buffer) {
virLibDomainError (dom, VIR_ERR_INVALID_ARG,
_("buffer is NULL but size is non-zero"));
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->domainMemoryPeek) {
int ret;
ret = conn->driver->domainMemoryPeek (dom, start, size,
buffer, flags);
if (ret < 0)
goto error;
return ret;
}
virLibDomainError (dom, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(dom->conn);
return -1;
}
/**
* virDomainGetBlockInfo:
* @domain: a domain object
* @path: path to the block device or file
* @info: pointer to a virDomainBlockInfo structure allocated by the user
* @flags: currently unused, pass zero
*
* Extract information about a domain's block device.
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virDomainGetBlockInfo(virDomainPtr domain, const char *path, virDomainBlockInfoPtr info, unsigned int flags)
{
virConnectPtr conn;
DEBUG("domain=%p, info=%p flags=%u", domain, info, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (info == NULL) {
virLibDomainError(domain, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
memset(info, 0, sizeof(virDomainBlockInfo));
conn = domain->conn;
if (conn->driver->domainGetBlockInfo) {
int ret;
ret = conn->driver->domainGetBlockInfo (domain, path, info, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(domain->conn);
return -1;
}
/************************************************************************
* *
* Handling of defined but not running domains *
* *
************************************************************************/
/**
* virDomainDefineXML:
* @conn: pointer to the hypervisor connection
* @xml: the XML description for the domain, preferably in UTF-8
*
* Define a domain, but does not start it.
* This definition is persistent, until explicitly undefined with
* virDomainUndefine(). A previous definition for this domain would be
* overriden if it already exists.
*
* Returns NULL in case of error, a pointer to the domain otherwise
*/
virDomainPtr
virDomainDefineXML(virConnectPtr conn, const char *xml) {
DEBUG("conn=%p, xml=%s", conn, xml);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (conn->flags & VIR_CONNECT_RO) {
virLibConnError(conn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
if (xml == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->domainDefineXML) {
virDomainPtr ret;
ret = conn->driver->domainDefineXML (conn, xml);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return NULL;
}
/**
* virDomainUndefine:
* @domain: pointer to a defined domain
*
* Undefine a domain but does not stop it if it is running
*
* Returns 0 in case of success, -1 in case of error
*/
int
virDomainUndefine(virDomainPtr domain) {
virConnectPtr conn;
DEBUG("domain=%p", domain);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = domain->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->domainUndefine) {
int ret;
ret = conn->driver->domainUndefine (domain);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virConnectNumOfDefinedDomains:
* @conn: pointer to the hypervisor connection
*
* Provides the number of defined but inactive domains.
*
* Returns the number of domain found or -1 in case of error
*/
int
virConnectNumOfDefinedDomains(virConnectPtr conn)
{
DEBUG("conn=%p", conn);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
2009-01-20 12:01:45 +00:00
if (conn->driver->numOfDefinedDomains) {
int ret;
ret = conn->driver->numOfDefinedDomains (conn);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return -1;
}
/**
* virConnectListDefinedDomains:
* @conn: pointer to the hypervisor connection
* @names: pointer to an array to store the names
* @maxnames: size of the array
*
* list the defined but inactive domains, stores the pointers to the names
* in @names
*
* Returns the number of names provided in the array or -1 in case of error
*/
int
2007-03-06 21:55:44 +00:00
virConnectListDefinedDomains(virConnectPtr conn, char **const names,
int maxnames) {
DEBUG("conn=%p, names=%p, maxnames=%d", conn, names, maxnames);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if ((names == NULL) || (maxnames < 0)) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->listDefinedDomains) {
int ret;
ret = conn->driver->listDefinedDomains (conn, names, maxnames);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return -1;
}
/**
* virDomainCreate:
* @domain: pointer to a defined domain
*
* Launch a defined domain. If the call succeeds the domain moves from the
* defined to the running domains pools.
*
* Returns 0 in case of success, -1 in case of error
*/
int
virDomainCreate(virDomainPtr domain) {
virConnectPtr conn;
DEBUG("domain=%p", domain);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = domain->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->domainCreate) {
int ret;
ret = conn->driver->domainCreate (domain);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainCreateWithFlags:
* @domain: pointer to a defined domain
* @flags: bitwise-or of supported virDomainCreateFlags
*
* Launch a defined domain. If the call succeeds the domain moves from the
* defined to the running domains pools.
*
* Returns 0 in case of success, -1 in case of error
*/
int
virDomainCreateWithFlags(virDomainPtr domain, unsigned int flags) {
virConnectPtr conn;
DEBUG("domain=%p, flags=%d", domain, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = domain->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (conn->driver->domainCreateWithFlags) {
int ret;
ret = conn->driver->domainCreateWithFlags (domain, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainGetAutostart:
* @domain: a domain object
* @autostart: the value returned
*
* Provides a boolean value indicating whether the domain
* configured to be automatically started when the host
* machine boots.
*
* Returns -1 in case of error, 0 in case of success
*/
int
virDomainGetAutostart(virDomainPtr domain,
int *autostart)
{
virConnectPtr conn;
DEBUG("domain=%p, autostart=%p", domain, autostart);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (!autostart) {
virLibDomainError(domain, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = domain->conn;
2009-01-20 12:01:45 +00:00
if (conn->driver->domainGetAutostart) {
int ret;
ret = conn->driver->domainGetAutostart (domain, autostart);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainSetAutostart:
* @domain: a domain object
* @autostart: whether the domain should be automatically started 0 or 1
*
* Configure the domain to be automatically started
* when the host machine boots.
*
* Returns -1 in case of error, 0 in case of success
*/
int
virDomainSetAutostart(virDomainPtr domain,
int autostart)
{
virConnectPtr conn;
DEBUG("domain=%p, autostart=%d", domain, autostart);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = domain->conn;
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->domainSetAutostart) {
int ret;
ret = conn->driver->domainSetAutostart (domain, autostart);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainSetVcpus:
* @domain: pointer to domain object, or NULL for Domain0
* @nvcpus: the new number of virtual CPUs for this domain
*
* Dynamically change the number of virtual CPUs used by the domain.
* Note that this call may fail if the underlying virtualization hypervisor
* does not support it or if growing the number is arbitrary limited.
* This function requires privileged access to the hypervisor.
*
* This command only changes the runtime configuration of the domain,
* so can only be called on an active domain.
*
* Returns 0 in case of success, -1 in case of failure.
*/
int
virDomainSetVcpus(virDomainPtr domain, unsigned int nvcpus)
{
virConnectPtr conn;
DEBUG("domain=%p, nvcpus=%u", domain, nvcpus);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
if (nvcpus < 1) {
virLibDomainError(domain, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = domain->conn;
2009-01-20 12:01:45 +00:00
if (conn->driver->domainSetVcpus) {
int ret;
ret = conn->driver->domainSetVcpus (domain, nvcpus);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainPinVcpu:
* @domain: pointer to domain object, or NULL for Domain0
* @vcpu: virtual CPU number
* @cpumap: pointer to a bit map of real CPUs (in 8-bit bytes) (IN)
* Each bit set to 1 means that corresponding CPU is usable.
* Bytes are stored in little-endian order: CPU0-7, 8-15...
* In each byte, lowest CPU number is least significant bit.
* @maplen: number of bytes in cpumap, from 1 up to size of CPU map in
* underlying virtualization system (Xen...).
* If maplen < size, missing bytes are set to zero.
* If maplen > size, failure code is returned.
*
* Dynamically change the real CPUs which can be allocated to a virtual CPU.
* This function requires privileged access to the hypervisor.
*
* This command only changes the runtime configuration of the domain,
* so can only be called on an active domain.
*
* Returns 0 in case of success, -1 in case of failure.
*/
int
virDomainPinVcpu(virDomainPtr domain, unsigned int vcpu,
unsigned char *cpumap, int maplen)
{
virConnectPtr conn;
DEBUG("domain=%p, vcpu=%u, cpumap=%p, maplen=%d", domain, vcpu, cpumap, maplen);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
if ((vcpu > 32000) || (cpumap == NULL) || (maplen < 1)) {
virLibDomainError(domain, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = domain->conn;
2009-01-20 12:01:45 +00:00
if (conn->driver->domainPinVcpu) {
int ret;
ret = conn->driver->domainPinVcpu (domain, vcpu, cpumap, maplen);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainGetVcpus:
* @domain: pointer to domain object, or NULL for Domain0
* @info: pointer to an array of virVcpuInfo structures (OUT)
* @maxinfo: number of structures in info array
2009-10-26 23:02:46 +00:00
* @cpumaps: pointer to a bit map of real CPUs for all vcpus of this
* domain (in 8-bit bytes) (OUT)
* If cpumaps is NULL, then no cpumap information is returned by the API.
* It's assumed there is <maxinfo> cpumap in cpumaps array.
* The memory allocated to cpumaps must be (maxinfo * maplen) bytes
* (ie: calloc(maxinfo, maplen)).
* One cpumap inside cpumaps has the format described in
* virDomainPinVcpu() API.
* @maplen: number of bytes in one cpumap, from 1 up to size of CPU map in
* underlying virtualization system (Xen...).
* Must be zero when cpumaps is NULL and positive when it is non-NULL.
*
* Extract information about virtual CPUs of domain, store it in info array
* and also in cpumaps if this pointer isn't NULL.
*
* Returns the number of info filled in case of success, -1 in case of failure.
*/
int
virDomainGetVcpus(virDomainPtr domain, virVcpuInfoPtr info, int maxinfo,
unsigned char *cpumaps, int maplen)
{
virConnectPtr conn;
DEBUG("domain=%p, info=%p, maxinfo=%d, cpumaps=%p, maplen=%d", domain, info, maxinfo, cpumaps, maplen);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if ((info == NULL) || (maxinfo < 1)) {
virLibDomainError(domain, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
/* Ensure that domainGetVcpus (aka remoteDomainGetVcpus) does not
try to memcpy anything into a NULL pointer. */
if ((cpumaps == NULL && maplen != 0)
|| (cpumaps && maplen <= 0)) {
virLibDomainError(domain, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = domain->conn;
2009-01-20 12:01:45 +00:00
if (conn->driver->domainGetVcpus) {
int ret;
ret = conn->driver->domainGetVcpus (domain, info, maxinfo,
cpumaps, maplen);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainGetMaxVcpus:
* @domain: pointer to domain object
*
* Provides the maximum number of virtual CPUs supported for
* the guest VM. If the guest is inactive, this is basically
* the same as virConnectGetMaxVcpus. If the guest is running
* this will reflect the maximum number of virtual CPUs the
* guest was booted with.
*
* Returns the maximum of virtual CPU or -1 in case of error.
*/
int
virDomainGetMaxVcpus(virDomainPtr domain)
{
virConnectPtr conn;
DEBUG("domain=%p", domain);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = domain->conn;
2009-01-20 12:01:45 +00:00
if (conn->driver->domainGetMaxVcpus) {
int ret;
ret = conn->driver->domainGetMaxVcpus (domain);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainGetSecurityLabel:
* @domain: a domain object
* @seclabel: pointer to a virSecurityLabel structure
*
* Extract security label of an active domain. The 'label' field
* in the @seclabel argument will be initialized to the empty
* string if the domain is not running under a security model.
*
* Returns 0 in case of success, -1 in case of failure
*/
int
virDomainGetSecurityLabel(virDomainPtr domain, virSecurityLabelPtr seclabel)
{
virConnectPtr conn;
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
if (seclabel == NULL) {
virLibDomainError(domain, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
conn = domain->conn;
if (conn->driver->domainGetSecurityLabel) {
int ret;
ret = conn->driver->domainGetSecurityLabel(domain, seclabel);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virNodeGetSecurityModel:
* @conn: a connection object
* @secmodel: pointer to a virSecurityModel structure
*
* Extract the security model of a hypervisor. The 'model' field
* in the @secmodel argument may be initialized to the empty
* string if the driver has not activated a security model.
*
* Returns 0 in case of success, -1 in case of failure
*/
int
virNodeGetSecurityModel(virConnectPtr conn, virSecurityModelPtr secmodel)
{
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
if (secmodel == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->driver->nodeGetSecurityModel) {
int ret;
ret = conn->driver->nodeGetSecurityModel(conn, secmodel);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virDomainAttachDevice:
* @domain: pointer to domain object
* @xml: pointer to XML description of one device
*
* Create a virtual device attachment to backend. This function,
* having hotplug semantics, is only allowed on an active domain.
*
* For compatibility, this method can also be used to change the media
* in an existing CDROM/Floppy device, however, applications are
* recommended to use the virDomainUpdateDeviceFlag method instead.
*
* Returns 0 in case of success, -1 in case of failure.
*/
int
virDomainAttachDevice(virDomainPtr domain, const char *xml)
{
virConnectPtr conn;
DEBUG("domain=%p, xml=%s", domain, xml);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = domain->conn;
2009-01-20 12:01:45 +00:00
if (conn->driver->domainAttachDevice) {
int ret;
ret = conn->driver->domainAttachDevice (domain, xml);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainAttachDeviceFlags:
* @domain: pointer to domain object
* @xml: pointer to XML description of one device
* @flags: an OR'ed set of virDomainDeviceModifyFlags
*
* Attach a virtual device to a domain, using the flags parameter
* to control how the device is attached. VIR_DOMAIN_DEVICE_MODIFY_CURRENT
* specifies that the device allocation is made based on current domain
* state. VIR_DOMAIN_DEVICE_MODIFY_LIVE specifies that the device shall be
* allocated to the active domain instance only and is not added to the
* persisted domain configuration. VIR_DOMAIN_DEVICE_MODIFY_CONFIG
* specifies that the device shall be allocated to the persisted domain
* configuration only. Note that the target hypervisor must return an
* error if unable to satisfy flags. E.g. the hypervisor driver will
* return failure if LIVE is specified but it only supports modifying the
* persisted device allocation.
*
* For compatibility, this method can also be used to change the media
* in an existing CDROM/Floppy device, however, applications are
* recommended to use the virDomainUpdateDeviceFlag method instead.
*
* Returns 0 in case of success, -1 in case of failure.
*/
int
virDomainAttachDeviceFlags(virDomainPtr domain,
const char *xml, unsigned int flags)
{
virConnectPtr conn;
DEBUG("domain=%p, xml=%s, flags=%d", domain, xml, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
return (-1);
}
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
conn = domain->conn;
if (conn->driver->domainAttachDeviceFlags) {
2009-01-20 12:01:45 +00:00
int ret;
ret = conn->driver->domainAttachDeviceFlags(domain, xml, flags);
2009-01-20 12:01:45 +00:00
if (ret < 0)
goto error;
return ret;
}
virLibConnError(conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainDetachDevice:
* @domain: pointer to domain object
* @xml: pointer to XML description of one device
*
* Destroy a virtual device attachment to backend. This function,
* having hot-unplug semantics, is only allowed on an active domain.
*
* Returns 0 in case of success, -1 in case of failure.
*/
int
virDomainDetachDevice(virDomainPtr domain, const char *xml)
{
virConnectPtr conn;
DEBUG("domain=%p, xml=%s", domain, xml);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = domain->conn;
2009-01-20 12:01:45 +00:00
if (conn->driver->domainDetachDevice) {
int ret;
ret = conn->driver->domainDetachDevice (domain, xml);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainDetachDeviceFlags:
* @domain: pointer to domain object
* @xml: pointer to XML description of one device
* @flags: an OR'ed set of virDomainDeviceModifyFlags
*
* Detach a virtual device from a domain, using the flags parameter
* to control how the device is detached. VIR_DOMAIN_DEVICE_MODIFY_CURRENT
* specifies that the device allocation is removed based on current domain
* state. VIR_DOMAIN_DEVICE_MODIFY_LIVE specifies that the device shall be
* deallocated from the active domain instance only and is not from the
* persisted domain configuration. VIR_DOMAIN_DEVICE_MODIFY_CONFIG
* specifies that the device shall be deallocated from the persisted domain
* configuration only. Note that the target hypervisor must return an
* error if unable to satisfy flags. E.g. the hypervisor driver will
* return failure if LIVE is specified but it only supports removing the
* persisted device allocation.
*
* Returns 0 in case of success, -1 in case of failure.
*/
int
virDomainDetachDeviceFlags(virDomainPtr domain,
const char *xml, unsigned int flags)
{
virConnectPtr conn;
DEBUG("domain=%p, xml=%s, flags=%d", domain, xml, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
return (-1);
}
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
conn = domain->conn;
if (conn->driver->domainDetachDeviceFlags) {
int ret;
ret = conn->driver->domainDetachDeviceFlags(domain, xml, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainUpdateDeviceFlags:
* @domain: pointer to domain object
* @xml: pointer to XML description of one device
* @flags: an OR'ed set of virDomainDeviceModifyFlags
*
* Change a virtual device on a domain, using the flags parameter
* to control how the device is changed. VIR_DOMAIN_DEVICE_MODIFY_CURRENT
* specifies that the device change is made based on current domain
* state. VIR_DOMAIN_DEVICE_MODIFY_LIVE specifies that the device shall be
* changed on the active domain instance only and is not added to the
* persisted domain configuration. VIR_DOMAIN_DEVICE_MODIFY_CONFIG
* specifies that the device shall be changed on the persisted domain
* configuration only. Note that the target hypervisor must return an
* error if unable to satisfy flags. E.g. the hypervisor driver will
* return failure if LIVE is specified but it only supports modifying the
* persisted device allocation.
*
* This method is used for actions such changing CDROM/Floppy device
* media, altering the graphics configuration such as password,
* reconfiguring the NIC device backend connectivity, etc.
*
* Returns 0 in case of success, -1 in case of failure.
*/
int
virDomainUpdateDeviceFlags(virDomainPtr domain,
const char *xml, unsigned int flags)
{
virConnectPtr conn;
DEBUG("domain=%p, xml=%s, flags=%d", domain, xml, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
return (-1);
}
if (domain->conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
conn = domain->conn;
if (conn->driver->domainUpdateDeviceFlags) {
int ret;
ret = conn->driver->domainUpdateDeviceFlags(domain, xml, flags);
2009-01-20 12:01:45 +00:00
if (ret < 0)
goto error;
return ret;
}
virLibConnError(conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virNodeGetCellsFreeMemory:
* @conn: pointer to the hypervisor connection
* @freeMems: pointer to the array of unsigned long long
* @startCell: index of first cell to return freeMems info on.
* @maxCells: Maximum number of cells for which freeMems information can
* be returned.
*
* This call returns the amount of free memory in one or more NUMA cells.
* The @freeMems array must be allocated by the caller and will be filled
* with the amount of free memory in bytes for each cell requested,
* starting with startCell (in freeMems[0]), up to either
* (startCell + maxCells), or the number of additional cells in the node,
* whichever is smaller.
*
* Returns the number of entries filled in freeMems, or -1 in case of error.
*/
int
virNodeGetCellsFreeMemory(virConnectPtr conn, unsigned long long *freeMems,
int startCell, int maxCells)
{
DEBUG("conn=%p, freeMems=%p, startCell=%d, maxCells=%d",
conn, freeMems, startCell, maxCells);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
2007-09-30 13:17:45 +00:00
if ((freeMems == NULL) || (maxCells <= 0) || (startCell < 0)) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->driver->nodeGetCellsFreeMemory) {
int ret;
ret = conn->driver->nodeGetCellsFreeMemory (conn, freeMems, startCell, maxCells);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return -1;
}
/**
* virNetworkGetConnect:
* @net: pointer to a network
*
* Provides the connection pointer associated with a network. The
* reference counter on the connection is not increased by this
* call.
*
* WARNING: When writing libvirt bindings in other languages, do
* not use this function. Instead, store the connection and
* the network object together.
*
* Returns the virConnectPtr or NULL in case of failure.
*/
virConnectPtr
virNetworkGetConnect (virNetworkPtr net)
{
DEBUG("net=%p", net);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_NETWORK (net)) {
virLibNetworkError (NULL, VIR_ERR_INVALID_NETWORK, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
return net->conn;
}
/**
* virConnectNumOfNetworks:
* @conn: pointer to the hypervisor connection
*
* Provides the number of active networks.
*
* Returns the number of network found or -1 in case of error
*/
int
virConnectNumOfNetworks(virConnectPtr conn)
{
DEBUG("conn=%p", conn);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
2009-01-20 12:01:45 +00:00
if (conn->networkDriver && conn->networkDriver->numOfNetworks) {
int ret;
ret = conn->networkDriver->numOfNetworks (conn);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return -1;
}
/**
* virConnectListNetworks:
* @conn: pointer to the hypervisor connection
* @names: array to collect the list of names of active networks
* @maxnames: size of @names
*
* Collect the list of active networks, and store their names in @names
*
* Returns the number of networks found or -1 in case of error
*/
int
2007-03-06 21:55:44 +00:00
virConnectListNetworks(virConnectPtr conn, char **const names, int maxnames)
{
DEBUG("conn=%p, names=%p, maxnames=%d", conn, names, maxnames);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if ((names == NULL) || (maxnames < 0)) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->networkDriver && conn->networkDriver->listNetworks) {
int ret;
ret = conn->networkDriver->listNetworks (conn, names, maxnames);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return -1;
}
/**
* virConnectNumOfDefinedNetworks:
* @conn: pointer to the hypervisor connection
*
* Provides the number of inactive networks.
*
* Returns the number of networks found or -1 in case of error
*/
int
virConnectNumOfDefinedNetworks(virConnectPtr conn)
{
DEBUG("conn=%p", conn);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
2009-01-20 12:01:45 +00:00
if (conn->networkDriver && conn->networkDriver->numOfDefinedNetworks) {
int ret;
ret = conn->networkDriver->numOfDefinedNetworks (conn);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return -1;
}
/**
* virConnectListDefinedNetworks:
* @conn: pointer to the hypervisor connection
* @names: pointer to an array to store the names
* @maxnames: size of the array
*
* list the inactive networks, stores the pointers to the names in @names
*
* Returns the number of names provided in the array or -1 in case of error
*/
int
2007-03-06 21:55:44 +00:00
virConnectListDefinedNetworks(virConnectPtr conn, char **const names,
int maxnames)
{
DEBUG("conn=%p, names=%p, maxnames=%d", conn, names, maxnames);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if ((names == NULL) || (maxnames < 0)) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->networkDriver && conn->networkDriver->listDefinedNetworks) {
int ret;
ret = conn->networkDriver->listDefinedNetworks (conn,
names, maxnames);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return -1;
}
/**
* virNetworkLookupByName:
* @conn: pointer to the hypervisor connection
* @name: name for the network
*
* Try to lookup a network on the given hypervisor based on its name.
*
* Returns a new network object or NULL in case of failure. If the
* network cannot be found, then VIR_ERR_NO_NETWORK error is raised.
*/
virNetworkPtr
virNetworkLookupByName(virConnectPtr conn, const char *name)
{
DEBUG("conn=%p, name=%s", conn, name);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (name == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->networkDriver && conn->networkDriver->networkLookupByName) {
virNetworkPtr ret;
ret = conn->networkDriver->networkLookupByName (conn, name);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return NULL;
}
/**
* virNetworkLookupByUUID:
* @conn: pointer to the hypervisor connection
* @uuid: the raw UUID for the network
*
* Try to lookup a network on the given hypervisor based on its UUID.
*
* Returns a new network object or NULL in case of failure. If the
* network cannot be found, then VIR_ERR_NO_NETWORK error is raised.
*/
virNetworkPtr
virNetworkLookupByUUID(virConnectPtr conn, const unsigned char *uuid)
{
DEBUG("conn=%p, uuid=%s", conn, uuid);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (uuid == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->networkDriver && conn->networkDriver->networkLookupByUUID){
virNetworkPtr ret;
ret = conn->networkDriver->networkLookupByUUID (conn, uuid);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return NULL;
}
/**
* virNetworkLookupByUUIDString:
* @conn: pointer to the hypervisor connection
* @uuidstr: the string UUID for the network
*
* Try to lookup a network on the given hypervisor based on its UUID.
*
* Returns a new network object or NULL in case of failure. If the
* network cannot be found, then VIR_ERR_NO_NETWORK error is raised.
*/
virNetworkPtr
virNetworkLookupByUUIDString(virConnectPtr conn, const char *uuidstr)
{
unsigned char uuid[VIR_UUID_BUFLEN];
DEBUG("conn=%p, uuidstr=%s", conn, uuidstr);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (uuidstr == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
if (virUUIDParse(uuidstr, uuid) < 0) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
return virNetworkLookupByUUID(conn, &uuid[0]);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
2009-01-20 12:01:45 +00:00
return NULL;
}
/**
* virNetworkCreateXML:
* @conn: pointer to the hypervisor connection
* @xmlDesc: an XML description of the network
*
* Create and start a new virtual network, based on an XML description
* similar to the one returned by virNetworkGetXMLDesc()
*
* Returns a new network object or NULL in case of failure
*/
virNetworkPtr
virNetworkCreateXML(virConnectPtr conn, const char *xmlDesc)
{
DEBUG("conn=%p, xmlDesc=%s", conn, xmlDesc);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (xmlDesc == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
if (conn->flags & VIR_CONNECT_RO) {
virLibConnError(conn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->networkDriver && conn->networkDriver->networkCreateXML) {
virNetworkPtr ret;
ret = conn->networkDriver->networkCreateXML (conn, xmlDesc);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return NULL;
}
/**
* virNetworkDefineXML:
* @conn: pointer to the hypervisor connection
* @xml: the XML description for the network, preferably in UTF-8
*
* Define a network, but does not create it
*
* Returns NULL in case of error, a pointer to the network otherwise
*/
virNetworkPtr
virNetworkDefineXML(virConnectPtr conn, const char *xml)
{
DEBUG("conn=%p, xml=%s", conn, xml);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (conn->flags & VIR_CONNECT_RO) {
virLibConnError(conn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
if (xml == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->networkDriver && conn->networkDriver->networkDefineXML) {
virNetworkPtr ret;
ret = conn->networkDriver->networkDefineXML (conn, xml);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return NULL;
}
/**
* virNetworkUndefine:
* @network: pointer to a defined network
*
* Undefine a network but does not stop it if it is running
*
* Returns 0 in case of success, -1 in case of error
*/
int
virNetworkUndefine(virNetworkPtr network) {
virConnectPtr conn;
DEBUG("network=%p", network);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_NETWORK(network)) {
virLibNetworkError(NULL, VIR_ERR_INVALID_NETWORK, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = network->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibNetworkError(network, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->networkDriver && conn->networkDriver->networkUndefine) {
int ret;
ret = conn->networkDriver->networkUndefine (network);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(network->conn);
return -1;
}
/**
* virNetworkCreate:
* @network: pointer to a defined network
*
* Create and start a defined network. If the call succeed the network
* moves from the defined to the running networks pools.
*
* Returns 0 in case of success, -1 in case of error
*/
int
virNetworkCreate(virNetworkPtr network)
{
virConnectPtr conn;
DEBUG("network=%p", network);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_NETWORK(network)) {
virLibNetworkError(NULL, VIR_ERR_INVALID_NETWORK, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = network->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibNetworkError(network, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->networkDriver && conn->networkDriver->networkCreate) {
int ret;
ret = conn->networkDriver->networkCreate (network);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(network->conn);
return -1;
}
/**
* virNetworkDestroy:
* @network: a network object
*
* Destroy the network object. The running instance is shutdown if not down
* already and all resources used by it are given back to the hypervisor. This
* does not free the associated virNetworkPtr object.
* This function may require privileged access
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virNetworkDestroy(virNetworkPtr network)
{
virConnectPtr conn;
DEBUG("network=%p", network);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_NETWORK(network)) {
virLibNetworkError(NULL, VIR_ERR_INVALID_NETWORK, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = network->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibNetworkError(network, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->networkDriver && conn->networkDriver->networkDestroy) {
int ret;
ret = conn->networkDriver->networkDestroy (network);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(network->conn);
return -1;
}
/**
* virNetworkFree:
* @network: a network object
*
* Free the network object. The running instance is kept alive.
* The data structure is freed and should not be used thereafter.
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virNetworkFree(virNetworkPtr network)
{
DEBUG("network=%p", network);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_NETWORK(network)) {
virLibNetworkError(NULL, VIR_ERR_INVALID_NETWORK, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (virUnrefNetwork(network) < 0) {
virDispatchError(NULL);
return (-1);
}
return(0);
}
/**
* virNetworkRef:
* @network: the network to hold a reference on
*
* Increment the reference count on the network. For each
* additional call to this method, there shall be a corresponding
* call to virNetworkFree to release the reference count, once
* the caller no longer needs the reference to this object.
*
* This method is typically useful for applications where multiple
* threads are using a connection, and it is required that the
* connection remain open until all threads have finished using
* it. ie, each new thread using a network would increment
* the reference count.
*
* Returns 0 in case of success, -1 in case of failure.
*/
int
virNetworkRef(virNetworkPtr network)
{
if ((!VIR_IS_CONNECTED_NETWORK(network))) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
virDispatchError(NULL);
return(-1);
}
virMutexLock(&network->conn->lock);
DEBUG("network=%p refs=%d", network, network->refs);
network->refs++;
virMutexUnlock(&network->conn->lock);
return 0;
}
/**
* virNetworkGetName:
* @network: a network object
*
* Get the public name for that network
*
* Returns a pointer to the name or NULL, the string need not be deallocated
* its lifetime will be the same as the network object.
*/
const char *
virNetworkGetName(virNetworkPtr network)
{
DEBUG("network=%p", network);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_NETWORK(network)) {
virLibNetworkError(NULL, VIR_ERR_INVALID_NETWORK, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
return (network->name);
}
/**
* virNetworkGetUUID:
* @network: a network object
* @uuid: pointer to a VIR_UUID_BUFLEN bytes array
*
* Get the UUID for a network
*
* Returns -1 in case of error, 0 in case of success
*/
int
virNetworkGetUUID(virNetworkPtr network, unsigned char *uuid)
{
DEBUG("network=%p, uuid=%p", network, uuid);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_NETWORK(network)) {
virLibNetworkError(NULL, VIR_ERR_INVALID_NETWORK, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (uuid == NULL) {
virLibNetworkError(network, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
memcpy(uuid, &network->uuid[0], VIR_UUID_BUFLEN);
return (0);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(network->conn);
2009-01-20 12:01:45 +00:00
return -1;
}
/**
* virNetworkGetUUIDString:
* @network: a network object
* @buf: pointer to a VIR_UUID_STRING_BUFLEN bytes array
*
* Get the UUID for a network as string. For more information about
* UUID see RFC4122.
*
* Returns -1 in case of error, 0 in case of success
*/
int
virNetworkGetUUIDString(virNetworkPtr network, char *buf)
{
unsigned char uuid[VIR_UUID_BUFLEN];
DEBUG("network=%p, buf=%p", network, buf);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_NETWORK(network)) {
virLibNetworkError(NULL, VIR_ERR_INVALID_NETWORK, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (buf == NULL) {
virLibNetworkError(network, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
if (virNetworkGetUUID(network, &uuid[0]))
goto error;
2007-08-09 20:19:12 +00:00
virUUIDFormat(uuid, buf);
return (0);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(network->conn);
2009-01-20 12:01:45 +00:00
return -1;
}
/**
* virNetworkGetXMLDesc:
* @network: a network object
* @flags: an OR'ed set of extraction flags, not used yet
*
* Provide an XML description of the network. The description may be reused
* later to relaunch the network with virNetworkCreateXML().
*
* Returns a 0 terminated UTF-8 encoded XML instance, or NULL in case of error.
* the caller must free() the returned value.
*/
char *
virNetworkGetXMLDesc(virNetworkPtr network, int flags)
{
virConnectPtr conn;
DEBUG("network=%p, flags=%d", network, flags);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_NETWORK(network)) {
virLibNetworkError(NULL, VIR_ERR_INVALID_NETWORK, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (flags != 0) {
virLibNetworkError(network, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = network->conn;
2009-01-20 12:01:45 +00:00
if (conn->networkDriver && conn->networkDriver->networkDumpXML) {
char *ret;
ret = conn->networkDriver->networkDumpXML (network, flags);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(network->conn);
return NULL;
}
/**
* virNetworkGetBridgeName:
* @network: a network object
*
* Provides a bridge interface name to which a domain may connect
* a network interface in order to join the network.
*
* Returns a 0 terminated interface name, or NULL in case of error.
* the caller must free() the returned value.
*/
char *
virNetworkGetBridgeName(virNetworkPtr network)
{
virConnectPtr conn;
DEBUG("network=%p", network);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_NETWORK(network)) {
virLibNetworkError(NULL, VIR_ERR_INVALID_NETWORK, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
conn = network->conn;
2009-01-20 12:01:45 +00:00
if (conn->networkDriver && conn->networkDriver->networkGetBridgeName) {
char *ret;
ret = conn->networkDriver->networkGetBridgeName (network);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(network->conn);
return NULL;
}
/**
* virNetworkGetAutostart:
* @network: a network object
* @autostart: the value returned
*
* Provides a boolean value indicating whether the network
* configured to be automatically started when the host
* machine boots.
*
* Returns -1 in case of error, 0 in case of success
*/
int
virNetworkGetAutostart(virNetworkPtr network,
int *autostart)
{
virConnectPtr conn;
DEBUG("network=%p, autostart=%p", network, autostart);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_NETWORK(network)) {
virLibNetworkError(NULL, VIR_ERR_INVALID_NETWORK, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (!autostart) {
virLibNetworkError(network, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = network->conn;
2009-01-20 12:01:45 +00:00
if (conn->networkDriver && conn->networkDriver->networkGetAutostart) {
int ret;
ret = conn->networkDriver->networkGetAutostart (network, autostart);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(network->conn);
return -1;
}
/**
* virNetworkSetAutostart:
* @network: a network object
* @autostart: whether the network should be automatically started 0 or 1
*
* Configure the network to be automatically started
* when the host machine boots.
*
* Returns -1 in case of error, 0 in case of success
*/
int
virNetworkSetAutostart(virNetworkPtr network,
int autostart)
{
virConnectPtr conn;
DEBUG("network=%p, autostart=%d", network, autostart);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_NETWORK(network)) {
virLibNetworkError(NULL, VIR_ERR_INVALID_NETWORK, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (network->conn->flags & VIR_CONNECT_RO) {
virLibNetworkError(network, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = network->conn;
2009-01-20 12:01:45 +00:00
if (conn->networkDriver && conn->networkDriver->networkSetAutostart) {
int ret;
ret = conn->networkDriver->networkSetAutostart (network, autostart);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(network->conn);
return -1;
}
/**
* virInterfaceGetConnect:
* @iface: pointer to an interface
*
* Provides the connection pointer associated with an interface. The
* reference counter on the connection is not increased by this
* call.
*
* WARNING: When writing libvirt bindings in other languages, do
* not use this function. Instead, store the connection and
* the interface object together.
*
* Returns the virConnectPtr or NULL in case of failure.
*/
virConnectPtr
virInterfaceGetConnect (virInterfacePtr iface)
{
DEBUG("iface=%p", iface);
virResetLastError();
if (!VIR_IS_CONNECTED_INTERFACE (iface)) {
virLibInterfaceError (NULL, VIR_ERR_INVALID_INTERFACE, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
return iface->conn;
}
/**
* virConnectNumOfInterfaces:
* @conn: pointer to the hypervisor connection
*
* Provides the number of active interfaces on the physical host.
*
* Returns the number of active interfaces found or -1 in case of error
*/
int
virConnectNumOfInterfaces(virConnectPtr conn)
{
DEBUG("conn=%p", conn);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (conn->interfaceDriver && conn->interfaceDriver->numOfInterfaces) {
int ret;
ret = conn->interfaceDriver->numOfInterfaces (conn);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virConnectListInterfaces:
* @conn: pointer to the hypervisor connection
* @names: array to collect the list of names of interfaces
* @maxnames: size of @names
*
* Collect the list of active physical host interfaces,
* and store their names in @names
*
* Returns the number of interfaces found or -1 in case of error
*/
int
virConnectListInterfaces(virConnectPtr conn, char **const names, int maxnames)
{
DEBUG("conn=%p, names=%p, maxnames=%d", conn, names, maxnames);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if ((names == NULL) || (maxnames < 0)) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->interfaceDriver && conn->interfaceDriver->listInterfaces) {
int ret;
ret = conn->interfaceDriver->listInterfaces (conn, names, maxnames);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virConnectNumOfDefinedInterfaces:
* @conn: pointer to the hypervisor connection
*
* Provides the number of defined (inactive) interfaces on the physical host.
*
* Returns the number of defined interface found or -1 in case of error
*/
int
virConnectNumOfDefinedInterfaces(virConnectPtr conn)
{
DEBUG("conn=%p", conn);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (conn->interfaceDriver && conn->interfaceDriver->numOfDefinedInterfaces) {
int ret;
ret = conn->interfaceDriver->numOfDefinedInterfaces (conn);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virConnectListDefinedInterfaces:
* @conn: pointer to the hypervisor connection
* @names: array to collect the list of names of interfaces
* @maxnames: size of @names
*
* Collect the list of defined (inactive) physical host interfaces,
* and store their names in @names.
*
* Returns the number of interfaces found or -1 in case of error
*/
int
virConnectListDefinedInterfaces(virConnectPtr conn,
char **const names,
int maxnames)
{
DEBUG("conn=%p, names=%p, maxnames=%d", conn, names, maxnames);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if ((names == NULL) || (maxnames < 0)) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->interfaceDriver && conn->interfaceDriver->listDefinedInterfaces) {
int ret;
ret = conn->interfaceDriver->listDefinedInterfaces (conn, names, maxnames);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virInterfaceLookupByName:
* @conn: pointer to the hypervisor connection
* @name: name for the interface
*
* Try to lookup an interface on the given hypervisor based on its name.
*
* Returns a new interface object or NULL in case of failure. If the
* interface cannot be found, then VIR_ERR_NO_INTERFACE error is raised.
*/
virInterfacePtr
virInterfaceLookupByName(virConnectPtr conn, const char *name)
{
DEBUG("conn=%p, name=%s", conn, name);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (name == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->interfaceDriver && conn->interfaceDriver->interfaceLookupByName) {
virInterfacePtr ret;
ret = conn->interfaceDriver->interfaceLookupByName (conn, name);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return NULL;
}
/**
* virInterfaceLookupByMACString:
* @conn: pointer to the hypervisor connection
* @macstr: the MAC for the interface (null-terminated ASCII format)
*
* Try to lookup an interface on the given hypervisor based on its MAC.
*
* Returns a new interface object or NULL in case of failure. If the
* interface cannot be found, then VIR_ERR_NO_INTERFACE error is raised.
*/
virInterfacePtr
virInterfaceLookupByMACString(virConnectPtr conn, const char *macstr)
{
DEBUG("conn=%p, macstr=%s", conn, macstr);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (macstr == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->interfaceDriver && conn->interfaceDriver->interfaceLookupByMACString) {
virInterfacePtr ret;
ret = conn->interfaceDriver->interfaceLookupByMACString (conn, macstr);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return NULL;
}
/**
* virInterfaceGetName:
* @iface: an interface object
*
* Get the public name for that interface
*
* Returns a pointer to the name or NULL, the string need not be deallocated
* its lifetime will be the same as the interface object.
*/
const char *
virInterfaceGetName(virInterfacePtr iface)
{
DEBUG("iface=%p", iface);
virResetLastError();
if (!VIR_IS_INTERFACE(iface)) {
virLibInterfaceError(NULL, VIR_ERR_INVALID_INTERFACE, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
return (iface->name);
}
/**
* virInterfaceGetMACString:
* @iface: an interface object
*
* Get the MAC for an interface as string. For more information about
* MAC see RFC4122.
*
* Returns a pointer to the MAC address (in null-terminated ASCII
* format) or NULL, the string need not be deallocated its lifetime
* will be the same as the interface object.
*/
const char *
virInterfaceGetMACString(virInterfacePtr iface)
{
DEBUG("iface=%p", iface);
virResetLastError();
if (!VIR_IS_INTERFACE(iface)) {
virLibInterfaceError(NULL, VIR_ERR_INVALID_INTERFACE, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
return (iface->mac);
}
/**
* virInterfaceGetXMLDesc:
* @iface: an interface object
* @flags: an OR'ed set of extraction flags. Current valid bits:
*
* VIR_INTERFACE_XML_INACTIVE - return the static configuration,
* suitable for use redefining the
* interface via virInterfaceDefineXML()
*
* Provide an XML description of the interface. If
* VIR_INTERFACE_XML_INACTIVE is set, the description may be reused
* later to redefine the interface with virInterfaceDefineXML(). If it
* is not set, the ip address and netmask will be the current live
* setting of the interface, not the settings from the config files.
*
* Returns a 0 terminated UTF-8 encoded XML instance, or NULL in case of error.
* the caller must free() the returned value.
*/
char *
virInterfaceGetXMLDesc(virInterfacePtr iface, unsigned int flags)
{
virConnectPtr conn;
DEBUG("iface=%p, flags=%d", iface, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_INTERFACE(iface)) {
virLibInterfaceError(NULL, VIR_ERR_INVALID_INTERFACE, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if ((flags & ~VIR_INTERFACE_XML_INACTIVE) != 0) {
virLibInterfaceError(iface, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
conn = iface->conn;
if (conn->interfaceDriver && conn->interfaceDriver->interfaceGetXMLDesc) {
char *ret;
ret = conn->interfaceDriver->interfaceGetXMLDesc (iface, flags);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(iface->conn);
return NULL;
}
/**
* virInterfaceDefineXML:
* @conn: pointer to the hypervisor connection
* @xml: the XML description for the interface, preferably in UTF-8
* @flags: and OR'ed set of extraction flags, not used yet
*
* Define an interface (or modify existing interface configuration)
*
* Returns NULL in case of error, a pointer to the interface otherwise
*/
virInterfacePtr
virInterfaceDefineXML(virConnectPtr conn, const char *xml, unsigned int flags)
{
DEBUG("conn=%p, xml=%s, flags=%d", conn, xml, flags);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (conn->flags & VIR_CONNECT_RO) {
virLibConnError(conn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (xml == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->interfaceDriver && conn->interfaceDriver->interfaceDefineXML) {
virInterfacePtr ret;
ret = conn->interfaceDriver->interfaceDefineXML (conn, xml, flags);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return NULL;
}
/**
* virInterfaceUndefine:
* @iface: pointer to a defined interface
*
* Undefine an interface, ie remove it from the config.
* This does not free the associated virInterfacePtr object.
*
* Returns 0 in case of success, -1 in case of error
*/
int
virInterfaceUndefine(virInterfacePtr iface) {
virConnectPtr conn;
DEBUG("iface=%p", iface);
virResetLastError();
if (!VIR_IS_CONNECTED_INTERFACE(iface)) {
virLibInterfaceError(NULL, VIR_ERR_INVALID_INTERFACE, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = iface->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibInterfaceError(iface, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (conn->interfaceDriver && conn->interfaceDriver->interfaceUndefine) {
int ret;
ret = conn->interfaceDriver->interfaceUndefine (iface);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(iface->conn);
return -1;
}
/**
* virInterfaceCreate:
* @iface: pointer to a defined interface
* @flags: and OR'ed set of extraction flags, not used yet
*
* Activate an interface (ie call "ifup")
*
* Returns 0 in case of success, -1 in case of error
*/
int
virInterfaceCreate(virInterfacePtr iface, unsigned int flags)
{
virConnectPtr conn;
DEBUG("iface=%p, flags=%d", iface, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_INTERFACE(iface)) {
virLibInterfaceError(NULL, VIR_ERR_INVALID_INTERFACE, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = iface->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibInterfaceError(iface, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (conn->interfaceDriver && conn->interfaceDriver->interfaceCreate) {
int ret;
ret = conn->interfaceDriver->interfaceCreate (iface, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(iface->conn);
return -1;
}
/**
* virInterfaceDestroy:
* @iface: an interface object
* @flags: and OR'ed set of extraction flags, not used yet
*
* deactivate an interface (ie call "ifdown")
* This does not remove the interface from the config, and
* does not free the associated virInterfacePtr object.
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virInterfaceDestroy(virInterfacePtr iface, unsigned int flags)
{
virConnectPtr conn;
DEBUG("iface=%p, flags=%d", iface, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_INTERFACE(iface)) {
virLibInterfaceError(NULL, VIR_ERR_INVALID_INTERFACE, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = iface->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibInterfaceError(iface, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (conn->interfaceDriver && conn->interfaceDriver->interfaceDestroy) {
int ret;
ret = conn->interfaceDriver->interfaceDestroy (iface, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(iface->conn);
return -1;
}
/**
* virInterfaceRef:
* @iface: the interface to hold a reference on
*
* Increment the reference count on the interface. For each
* additional call to this method, there shall be a corresponding
* call to virInterfaceFree to release the reference count, once
* the caller no longer needs the reference to this object.
*
* This method is typically useful for applications where multiple
* threads are using a connection, and it is required that the
* connection remain open until all threads have finished using
* it. ie, each new thread using an interface would increment
* the reference count.
*
* Returns 0 in case of success, -1 in case of failure.
*/
int
virInterfaceRef(virInterfacePtr iface)
{
if ((!VIR_IS_CONNECTED_INTERFACE(iface))) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
virDispatchError(NULL);
return(-1);
}
virMutexLock(&iface->conn->lock);
DEBUG("iface=%p refs=%d", iface, iface->refs);
iface->refs++;
virMutexUnlock(&iface->conn->lock);
return 0;
}
/**
* virInterfaceFree:
* @iface: an interface object
*
* Free the interface object. The interface itself is unaltered.
* The data structure is freed and should not be used thereafter.
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virInterfaceFree(virInterfacePtr iface)
{
DEBUG("iface=%p", iface);
virResetLastError();
if (!VIR_IS_CONNECTED_INTERFACE(iface)) {
virLibInterfaceError(NULL, VIR_ERR_INVALID_INTERFACE, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (virUnrefInterface(iface) < 0) {
virDispatchError(NULL);
return (-1);
}
return(0);
}
/**
* virStoragePoolGetConnect:
* @pool: pointer to a pool
*
* Provides the connection pointer associated with a storage pool. The
* reference counter on the connection is not increased by this
* call.
*
* WARNING: When writing libvirt bindings in other languages, do
* not use this function. Instead, store the connection and
* the pool object together.
*
* Returns the virConnectPtr or NULL in case of failure.
*/
virConnectPtr
virStoragePoolGetConnect (virStoragePoolPtr pool)
{
DEBUG("pool=%p", pool);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_STORAGE_POOL (pool)) {
virLibStoragePoolError (NULL, VIR_ERR_INVALID_STORAGE_POOL, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
return pool->conn;
}
/**
* virConnectNumOfStoragePools:
* @conn: pointer to hypervisor connection
*
* Provides the number of active storage pools
*
* Returns the number of pools found, or -1 on error
*/
int
virConnectNumOfStoragePools (virConnectPtr conn)
{
DEBUG("conn=%p", conn);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->numOfPools) {
int ret;
ret = conn->storageDriver->numOfPools (conn);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return -1;
}
/**
* virConnectListStoragePools:
* @conn: pointer to hypervisor connection
* @names: array of char * to fill with pool names (allocated by caller)
* @maxnames: size of the names array
*
* Provides the list of names of active storage pools
* upto maxnames. If there are more than maxnames, the
* remaining names will be silently ignored.
*
* Returns 0 on success, -1 on error
*/
int
virConnectListStoragePools (virConnectPtr conn,
char **const names,
int maxnames)
{
DEBUG("conn=%p, names=%p, maxnames=%d", conn, names, maxnames);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if ((names == NULL) || (maxnames < 0)) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->listPools) {
int ret;
ret = conn->storageDriver->listPools (conn, names, maxnames);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
2009-01-20 12:01:45 +00:00
return -1;
}
/**
* virConnectNumOfDefinedStoragePools:
* @conn: pointer to hypervisor connection
*
* Provides the number of inactive storage pools
*
* Returns the number of pools found, or -1 on error
*/
int
virConnectNumOfDefinedStoragePools(virConnectPtr conn)
{
DEBUG("conn=%p", conn);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->numOfDefinedPools) {
int ret;
ret = conn->storageDriver->numOfDefinedPools (conn);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return -1;
}
/**
* virConnectListDefinedStoragePools:
* @conn: pointer to hypervisor connection
* @names: array of char * to fill with pool names (allocated by caller)
* @maxnames: size of the names array
*
* Provides the list of names of inactive storage pools
* upto maxnames. If there are more than maxnames, the
* remaining names will be silently ignored.
*
* Returns 0 on success, -1 on error
*/
int
virConnectListDefinedStoragePools(virConnectPtr conn,
char **const names,
int maxnames)
{
DEBUG("conn=%p, names=%p, maxnames=%d", conn, names, maxnames);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if ((names == NULL) || (maxnames < 0)) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->listDefinedPools) {
int ret;
ret = conn->storageDriver->listDefinedPools (conn, names, maxnames);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return -1;
}
/**
* virConnectFindStoragePoolSources:
* @conn: pointer to hypervisor connection
* @type: type of storage pool sources to discover
* @srcSpec: XML document specifying discovery source
* @flags: flags for discovery (unused, pass 0)
*
* Talks to a storage backend and attempts to auto-discover the set of
* available storage pool sources. e.g. For iSCSI this would be a set of
* iSCSI targets. For NFS this would be a list of exported paths. The
* srcSpec (optional for some storage pool types, e.g. local ones) is
* an instance of the storage pool's source element specifying where
* to look for the pools.
*
* srcSpec is not required for some types (e.g., those querying
* local storage resources only)
*
* Returns an xml document consisting of a SourceList element
* containing a source document appropriate to the given pool
* type for each discovered source.
*/
char *
virConnectFindStoragePoolSources(virConnectPtr conn,
const char *type,
const char *srcSpec,
unsigned int flags)
{
2009-01-20 12:01:45 +00:00
DEBUG("conn=%p, type=%s, src=%s, flags=%u", conn, type ? type : "", srcSpec ? srcSpec : "", flags);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
if (type == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
if (conn->flags & VIR_CONNECT_RO) {
virLibConnError(conn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->findPoolSources) {
char *ret;
ret = conn->storageDriver->findPoolSources(conn, type, srcSpec, flags);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return NULL;
}
/**
* virStoragePoolLookupByName:
* @conn: pointer to hypervisor connection
* @name: name of pool to fetch
*
* Fetch a storage pool based on its unique name
*
* Returns a virStoragePoolPtr object, or NULL if no matching pool is found
*/
virStoragePoolPtr
virStoragePoolLookupByName(virConnectPtr conn,
const char *name)
{
DEBUG("conn=%p, name=%s", conn, name);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (name == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->poolLookupByName) {
virStoragePoolPtr ret;
ret = conn->storageDriver->poolLookupByName (conn, name);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return NULL;
}
/**
* virStoragePoolLookupByUUID:
* @conn: pointer to hypervisor connection
* @uuid: globally unique id of pool to fetch
*
* Fetch a storage pool based on its globally unique id
*
* Returns a virStoragePoolPtr object, or NULL if no matching pool is found
*/
virStoragePoolPtr
virStoragePoolLookupByUUID(virConnectPtr conn,
const unsigned char *uuid)
{
DEBUG("conn=%p, uuid=%s", conn, uuid);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (uuid == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->poolLookupByUUID) {
virStoragePoolPtr ret;
ret = conn->storageDriver->poolLookupByUUID (conn, uuid);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
2009-01-20 12:01:45 +00:00
return NULL;
}
/**
* virStoragePoolLookupByUUIDString:
* @conn: pointer to hypervisor connection
* @uuidstr: globally unique id of pool to fetch
*
* Fetch a storage pool based on its globally unique id
*
* Returns a virStoragePoolPtr object, or NULL if no matching pool is found
*/
virStoragePoolPtr
virStoragePoolLookupByUUIDString(virConnectPtr conn,
2009-01-20 12:01:45 +00:00
const char *uuidstr)
{
unsigned char uuid[VIR_UUID_BUFLEN];
DEBUG("conn=%p, uuidstr=%s", conn, uuidstr);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (uuidstr == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
if (virUUIDParse(uuidstr, uuid) < 0) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
return virStoragePoolLookupByUUID(conn, uuid);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
2009-01-20 12:01:45 +00:00
return NULL;
}
/**
* virStoragePoolLookupByVolume:
* @vol: pointer to storage volume
*
* Fetch a storage pool which contains a particular volume
*
* Returns a virStoragePoolPtr object, or NULL if no matching pool is found
*/
virStoragePoolPtr
virStoragePoolLookupByVolume(virStorageVolPtr vol)
{
DEBUG("vol=%p", vol);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_STORAGE_VOL(vol)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
2009-01-20 12:01:45 +00:00
if (vol->conn->storageDriver && vol->conn->storageDriver->poolLookupByVolume) {
virStoragePoolPtr ret;
ret = vol->conn->storageDriver->poolLookupByVolume (vol);
if (!ret)
goto error;
return ret;
}
virLibConnError (vol->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(vol->conn);
2009-01-20 12:01:45 +00:00
return NULL;
}
/**
* virStoragePoolCreateXML:
* @conn: pointer to hypervisor connection
* @xmlDesc: XML description for new pool
* @flags: future flags, use 0 for now
*
* Create a new storage based on its XML description. The
* pool is not persistent, so its definition will disappear
* when it is destroyed, or if the host is restarted
*
* Returns a virStoragePoolPtr object, or NULL if creation failed
*/
virStoragePoolPtr
virStoragePoolCreateXML(virConnectPtr conn,
const char *xmlDesc,
unsigned int flags)
{
DEBUG("conn=%p, xmlDesc=%s", conn, xmlDesc);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (xmlDesc == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
if (conn->flags & VIR_CONNECT_RO) {
virLibConnError(conn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->poolCreateXML) {
virStoragePoolPtr ret;
ret = conn->storageDriver->poolCreateXML (conn, xmlDesc, flags);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return NULL;
}
/**
* virStoragePoolDefineXML:
* @conn: pointer to hypervisor connection
* @xml: XML description for new pool
* @flags: future flags, use 0 for now
*
* Define a new inactive storage pool based on its XML description. The
* pool is persistent, until explicitly undefined.
*
* Returns a virStoragePoolPtr object, or NULL if creation failed
*/
virStoragePoolPtr
virStoragePoolDefineXML(virConnectPtr conn,
const char *xml,
unsigned int flags)
{
DEBUG("conn=%p, xml=%s", conn, xml);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (conn->flags & VIR_CONNECT_RO) {
virLibConnError(conn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
if (xml == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->poolDefineXML) {
virStoragePoolPtr ret;
ret = conn->storageDriver->poolDefineXML (conn, xml, flags);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
2009-01-20 12:01:45 +00:00
return NULL;
}
/**
* virStoragePoolBuild:
* @pool: pointer to storage pool
* @flags: future flags, use 0 for now
*
* Build the underlying storage pool
*
* Returns 0 on success, or -1 upon failure
*/
int
virStoragePoolBuild(virStoragePoolPtr pool,
unsigned int flags)
{
virConnectPtr conn;
DEBUG("pool=%p, flags=%u", pool, flags);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_STORAGE_POOL(pool)) {
virLibStoragePoolError(NULL, VIR_ERR_INVALID_NETWORK, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = pool->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibStoragePoolError(pool, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->poolBuild) {
int ret;
ret = conn->storageDriver->poolBuild (pool, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(pool->conn);
2009-01-20 12:01:45 +00:00
return -1;
}
/**
* virStoragePoolUndefine:
* @pool: pointer to storage pool
*
* Undefine an inactive storage pool
*
* Returns 0 on success, -1 on failure
*/
int
virStoragePoolUndefine(virStoragePoolPtr pool)
{
virConnectPtr conn;
DEBUG("pool=%p", pool);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_STORAGE_POOL(pool)) {
virLibStoragePoolError(NULL, VIR_ERR_INVALID_NETWORK, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = pool->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibStoragePoolError(pool, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->poolUndefine) {
int ret;
ret = conn->storageDriver->poolUndefine (pool);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(pool->conn);
2009-01-20 12:01:45 +00:00
return -1;
}
/**
* virStoragePoolCreate:
* @pool: pointer to storage pool
* @flags: future flags, use 0 for now
*
* Starts an inactive storage pool
*
* Returns 0 on success, or -1 if it could not be started
*/
int
virStoragePoolCreate(virStoragePoolPtr pool,
unsigned int flags)
{
virConnectPtr conn;
DEBUG("pool=%p", pool);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_STORAGE_POOL(pool)) {
virLibStoragePoolError(NULL, VIR_ERR_INVALID_STORAGE_POOL, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = pool->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibStoragePoolError(pool, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->poolCreate) {
int ret;
ret = conn->storageDriver->poolCreate (pool, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(pool->conn);
2009-01-20 12:01:45 +00:00
return -1;
}
/**
* virStoragePoolDestroy:
* @pool: pointer to storage pool
*
* Destroy an active storage pool. This will deactivate the
* pool on the host, but keep any persistent config associated
* with it. If it has a persistent config it can later be
* restarted with virStoragePoolCreate(). This does not free
* the associated virStoragePoolPtr object.
*
* Returns 0 on success, or -1 if it could not be destroyed
*/
int
virStoragePoolDestroy(virStoragePoolPtr pool)
{
virConnectPtr conn;
DEBUG("pool=%p", pool);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_STORAGE_POOL(pool)) {
virLibStoragePoolError(NULL, VIR_ERR_INVALID_STORAGE_POOL, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = pool->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibStoragePoolError(pool, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->poolDestroy) {
int ret;
ret = conn->storageDriver->poolDestroy (pool);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(pool->conn);
return -1;
}
/**
* virStoragePoolDelete:
* @pool: pointer to storage pool
* @flags: flags for obliteration process
*
* Delete the underlying pool resources. This is
* a non-recoverable operation. The virStoragePoolPtr object
* itself is not free'd.
*
* Returns 0 on success, or -1 if it could not be obliterate
*/
int
virStoragePoolDelete(virStoragePoolPtr pool,
unsigned int flags)
{
virConnectPtr conn;
DEBUG("pool=%p, flags=%u", pool, flags);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_STORAGE_POOL(pool)) {
virLibStoragePoolError(NULL, VIR_ERR_INVALID_STORAGE_POOL, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = pool->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibStoragePoolError(pool, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->poolDelete) {
int ret;
ret = conn->storageDriver->poolDelete (pool, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(pool->conn);
return -1;
}
/**
* virStoragePoolFree:
* @pool: pointer to storage pool
*
* Free a storage pool object, releasing all memory associated with
* it. Does not change the state of the pool on the host.
*
* Returns 0 on success, or -1 if it could not be free'd.
*/
int
virStoragePoolFree(virStoragePoolPtr pool)
{
DEBUG("pool=%p", pool);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_STORAGE_POOL(pool)) {
virLibStoragePoolError(NULL, VIR_ERR_INVALID_STORAGE_POOL, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (virUnrefStoragePool(pool) < 0) {
virDispatchError(NULL);
return (-1);
}
return(0);
}
/**
* virStoragePoolRef:
* @pool: the pool to hold a reference on
*
* Increment the reference count on the pool. For each
* additional call to this method, there shall be a corresponding
* call to virStoragePoolFree to release the reference count, once
* the caller no longer needs the reference to this object.
*
* This method is typically useful for applications where multiple
* threads are using a connection, and it is required that the
* connection remain open until all threads have finished using
* it. ie, each new thread using a pool would increment
* the reference count.
*
* Returns 0 in case of success, -1 in case of failure.
*/
int
virStoragePoolRef(virStoragePoolPtr pool)
{
if ((!VIR_IS_CONNECTED_STORAGE_POOL(pool))) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
virDispatchError(NULL);
return(-1);
}
virMutexLock(&pool->conn->lock);
DEBUG("pool=%p refs=%d", pool, pool->refs);
pool->refs++;
virMutexUnlock(&pool->conn->lock);
return 0;
}
/**
* virStoragePoolRefresh:
* @pool: pointer to storage pool
* @flags: flags to control refresh behaviour (currently unused, use 0)
*
* Request that the pool refresh its list of volumes. This may
* involve communicating with a remote server, and/or initializing
* new devices at the OS layer
*
* Returns 0 if the volume list was refreshed, -1 on failure
*/
int
virStoragePoolRefresh(virStoragePoolPtr pool,
unsigned int flags)
{
virConnectPtr conn;
DEBUG("pool=%p flags=%u", pool, flags);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_STORAGE_POOL(pool)) {
virLibStoragePoolError(NULL, VIR_ERR_INVALID_STORAGE_POOL, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = pool->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibStoragePoolError(pool, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->poolRefresh) {
int ret;
ret = conn->storageDriver->poolRefresh (pool, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(pool->conn);
return -1;
}
/**
* virStoragePoolGetName:
* @pool: pointer to storage pool
*
* Fetch the locally unique name of the storage pool
*
* Returns the name of the pool, or NULL on error
*/
const char*
virStoragePoolGetName(virStoragePoolPtr pool)
{
DEBUG("pool=%p", pool);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_STORAGE_POOL(pool)) {
virLibStoragePoolError(NULL, VIR_ERR_INVALID_STORAGE_POOL, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
return (pool->name);
}
/**
* virStoragePoolGetUUID:
* @pool: pointer to storage pool
* @uuid: buffer of VIR_UUID_BUFLEN bytes in size
*
* Fetch the globally unique ID of the storage pool
*
* Returns 0 on success, or -1 on error;
*/
int
virStoragePoolGetUUID(virStoragePoolPtr pool,
unsigned char *uuid)
{
DEBUG("pool=%p, uuid=%p", pool, uuid);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_STORAGE_POOL(pool)) {
virLibStoragePoolError(NULL, VIR_ERR_INVALID_STORAGE_POOL, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (uuid == NULL) {
virLibStoragePoolError(pool, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
memcpy(uuid, &pool->uuid[0], VIR_UUID_BUFLEN);
return (0);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(pool->conn);
2009-01-20 12:01:45 +00:00
return -1;
}
/**
* virStoragePoolGetUUIDString:
* @pool: pointer to storage pool
* @buf: buffer of VIR_UUID_STRING_BUFLEN bytes in size
*
* Fetch the globally unique ID of the storage pool as a string
*
* Returns 0 on success, or -1 on error;
*/
int
virStoragePoolGetUUIDString(virStoragePoolPtr pool,
char *buf)
{
unsigned char uuid[VIR_UUID_BUFLEN];
DEBUG("pool=%p, buf=%p", pool, buf);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_STORAGE_POOL(pool)) {
virLibStoragePoolError(NULL, VIR_ERR_INVALID_STORAGE_POOL, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (buf == NULL) {
virLibStoragePoolError(pool, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
if (virStoragePoolGetUUID(pool, &uuid[0]))
2009-01-20 12:01:45 +00:00
goto error;
virUUIDFormat(uuid, buf);
return (0);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(pool->conn);
2009-01-20 12:01:45 +00:00
return -1;
}
/**
* virStoragePoolGetInfo:
* @pool: pointer to storage pool
* @info: pointer at which to store info
*
* Get volatile information about the storage pool
* such as free space / usage summary
*
* Returns 0 on success, or -1 on failure.
*/
int
virStoragePoolGetInfo(virStoragePoolPtr pool,
virStoragePoolInfoPtr info)
{
virConnectPtr conn;
DEBUG("pool=%p, info=%p", pool, info);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_STORAGE_POOL(pool)) {
virLibStoragePoolError(NULL, VIR_ERR_INVALID_STORAGE_POOL, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (info == NULL) {
virLibStoragePoolError(pool, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
memset(info, 0, sizeof(virStoragePoolInfo));
conn = pool->conn;
2009-01-20 12:01:45 +00:00
if (conn->storageDriver->poolGetInfo) {
int ret;
ret = conn->storageDriver->poolGetInfo (pool, info);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(pool->conn);
2009-01-20 12:01:45 +00:00
return -1;
}
/**
* virStoragePoolGetXMLDesc:
* @pool: pointer to storage pool
* @flags: flags for XML format options (set of virDomainXMLFlags)
*
* Fetch an XML document describing all aspects of the
* storage pool. This is suitable for later feeding back
* into the virStoragePoolCreateXML method.
*
* Returns a XML document, or NULL on error
*/
char *
virStoragePoolGetXMLDesc(virStoragePoolPtr pool,
unsigned int flags)
{
virConnectPtr conn;
DEBUG("pool=%p, flags=%u", pool, flags);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_STORAGE_POOL(pool)) {
virLibStoragePoolError(NULL, VIR_ERR_INVALID_STORAGE_POOL, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (flags != 0) {
virLibStoragePoolError(pool, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = pool->conn;
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->poolGetXMLDesc) {
char *ret;
ret = conn->storageDriver->poolGetXMLDesc (pool, flags);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(pool->conn);
2009-01-20 12:01:45 +00:00
return NULL;
}
/**
* virStoragePoolGetAutostart:
* @pool: pointer to storage pool
* @autostart: location in which to store autostart flag
*
* Fetches the value of the autostart flag, which determines
* whether the pool is automatically started at boot time
*
* Returns 0 on success, -1 on failure
*/
int
virStoragePoolGetAutostart(virStoragePoolPtr pool,
int *autostart)
{
virConnectPtr conn;
DEBUG("pool=%p, autostart=%p", pool, autostart);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_STORAGE_POOL(pool)) {
virLibStoragePoolError(NULL, VIR_ERR_INVALID_STORAGE_POOL, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (!autostart) {
virLibStoragePoolError(pool, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = pool->conn;
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->poolGetAutostart) {
int ret;
ret = conn->storageDriver->poolGetAutostart (pool, autostart);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(pool->conn);
return -1;
}
/**
* virStoragePoolSetAutostart:
* @pool: pointer to storage pool
* @autostart: new flag setting
*
* Sets the autostart flag
*
* Returns 0 on success, -1 on failure
*/
int
virStoragePoolSetAutostart(virStoragePoolPtr pool,
int autostart)
{
virConnectPtr conn;
DEBUG("pool=%p, autostart=%d", pool, autostart);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_STORAGE_POOL(pool)) {
virLibStoragePoolError(NULL, VIR_ERR_INVALID_STORAGE_POOL, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
if (pool->conn->flags & VIR_CONNECT_RO) {
virLibStoragePoolError(pool, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = pool->conn;
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->poolSetAutostart) {
int ret;
ret = conn->storageDriver->poolSetAutostart (pool, autostart);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(pool->conn);
return -1;
}
/**
* virStoragePoolNumOfVolumes:
* @pool: pointer to storage pool
*
* Fetch the number of storage volumes within a pool
*
* Returns the number of storage pools, or -1 on failure
*/
int
virStoragePoolNumOfVolumes(virStoragePoolPtr pool)
{
DEBUG("pool=%p", pool);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_STORAGE_POOL(pool)) {
virLibConnError(NULL, VIR_ERR_INVALID_STORAGE_POOL, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
2009-01-20 12:01:45 +00:00
if (pool->conn->storageDriver && pool->conn->storageDriver->poolNumOfVolumes) {
int ret;
ret = pool->conn->storageDriver->poolNumOfVolumes (pool);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (pool->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(pool->conn);
return -1;
}
/**
* virStoragePoolListVolumes:
* @pool: pointer to storage pool
* @names: array in which to storage volume names
* @maxnames: size of names array
*
* Fetch list of storage volume names, limiting to
* at most maxnames.
*
* Returns the number of names fetched, or -1 on error
*/
int
virStoragePoolListVolumes(virStoragePoolPtr pool,
char **const names,
int maxnames)
{
DEBUG("pool=%p, names=%p, maxnames=%d", pool, names, maxnames);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_STORAGE_POOL(pool)) {
virLibConnError(NULL, VIR_ERR_INVALID_STORAGE_POOL, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if ((names == NULL) || (maxnames < 0)) {
virLibConnError(pool->conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (pool->conn->storageDriver && pool->conn->storageDriver->poolListVolumes) {
int ret;
ret = pool->conn->storageDriver->poolListVolumes (pool, names, maxnames);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (pool->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(pool->conn);
return -1;
}
/**
* virStorageVolGetConnect:
* @vol: pointer to a pool
*
* Provides the connection pointer associated with a storage volume. The
* reference counter on the connection is not increased by this
* call.
*
* WARNING: When writing libvirt bindings in other languages, do
* not use this function. Instead, store the connection and
* the volume object together.
*
* Returns the virConnectPtr or NULL in case of failure.
*/
virConnectPtr
virStorageVolGetConnect (virStorageVolPtr vol)
{
DEBUG("vol=%p", vol);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_STORAGE_VOL (vol)) {
virLibStoragePoolError (NULL, VIR_ERR_INVALID_STORAGE_VOL, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
return vol->conn;
}
/**
* virStorageVolLookupByName:
* @pool: pointer to storage pool
* @name: name of storage volume
*
* Fetch a pointer to a storage volume based on its name
* within a pool
*
* Returns a storage volume, or NULL if not found / error
*/
virStorageVolPtr
virStorageVolLookupByName(virStoragePoolPtr pool,
const char *name)
{
DEBUG("pool=%p, name=%s", pool, name);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_STORAGE_POOL(pool)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (name == NULL) {
virLibConnError(pool->conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (pool->conn->storageDriver && pool->conn->storageDriver->volLookupByName) {
virStorageVolPtr ret;
ret = pool->conn->storageDriver->volLookupByName (pool, name);
if (!ret)
goto error;
return ret;
}
virLibConnError (pool->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(pool->conn);
return NULL;
}
/**
* virStorageVolLookupByKey:
* @conn: pointer to hypervisor connection
* @key: globally unique key
*
* Fetch a pointer to a storage volume based on its
* globally unique key
*
* Returns a storage volume, or NULL if not found / error
*/
virStorageVolPtr
virStorageVolLookupByKey(virConnectPtr conn,
const char *key)
{
DEBUG("conn=%p, key=%s", conn, key);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (key == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->volLookupByKey) {
virStorageVolPtr ret;
ret = conn->storageDriver->volLookupByKey (conn, key);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return NULL;
}
/**
* virStorageVolLookupByPath:
* @conn: pointer to hypervisor connection
* @path: locally unique path
*
* Fetch a pointer to a storage volume based on its
* locally (host) unique path
*
* Returns a storage volume, or NULL if not found / error
*/
virStorageVolPtr
virStorageVolLookupByPath(virConnectPtr conn,
const char *path)
{
DEBUG("conn=%p, path=%s", conn, path);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (path == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->volLookupByPath) {
virStorageVolPtr ret;
ret = conn->storageDriver->volLookupByPath (conn, path);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return NULL;
}
/**
* virStorageVolGetName:
* @vol: pointer to storage volume
*
* Fetch the storage volume name. This is unique
* within the scope of a pool
*
* Returns the volume name, or NULL on error
*/
const char*
virStorageVolGetName(virStorageVolPtr vol)
{
DEBUG("vol=%p", vol);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_STORAGE_VOL(vol)) {
virLibStorageVolError(NULL, VIR_ERR_INVALID_STORAGE_VOL, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
return (vol->name);
}
/**
* virStorageVolGetKey:
* @vol: pointer to storage volume
*
* Fetch the storage volume key. This is globally
* unique, so the same volume will have the same
* key no matter what host it is accessed from
*
* Returns the volume key, or NULL on error
*/
const char*
virStorageVolGetKey(virStorageVolPtr vol)
{
DEBUG("vol=%p", vol);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_STORAGE_VOL(vol)) {
virLibStorageVolError(NULL, VIR_ERR_INVALID_STORAGE_VOL, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
return (vol->key);
}
/**
* virStorageVolCreateXML:
* @pool: pointer to storage pool
* @xmldesc: description of volume to create
* @flags: flags for creation (unused, pass 0)
*
* Create a storage volume within a pool based
* on an XML description. Not all pools support
* creation of volumes
*
* Returns the storage volume, or NULL on error
*/
virStorageVolPtr
virStorageVolCreateXML(virStoragePoolPtr pool,
const char *xmldesc,
unsigned int flags)
{
DEBUG("pool=%p, flags=%u", pool, flags);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_STORAGE_POOL(pool)) {
virLibConnError(NULL, VIR_ERR_INVALID_STORAGE_POOL, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (pool->conn->flags & VIR_CONNECT_RO) {
virLibConnError(pool->conn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (pool->conn->storageDriver && pool->conn->storageDriver->volCreateXML) {
virStorageVolPtr ret;
ret = pool->conn->storageDriver->volCreateXML (pool, xmldesc, flags);
if (!ret)
goto error;
return ret;
}
virLibConnError (pool->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(pool->conn);
return NULL;
}
/**
* virStorageVolCreateXMLFrom:
* @pool: pointer to parent pool for the new volume
* @xmldesc: description of volume to create
* @clonevol: storage volume to use as input
* @flags: flags for creation (unused, pass 0)
*
* Create a storage volume in the parent pool, using the
* 'clonevol' volume as input. Information for the new
* volume (name, perms) are passed via a typical volume
* XML description.
*
* Returns the storage volume, or NULL on error
*/
virStorageVolPtr
virStorageVolCreateXMLFrom(virStoragePoolPtr pool,
const char *xmldesc,
virStorageVolPtr clonevol,
unsigned int flags)
{
DEBUG("pool=%p, flags=%u, clonevol=%p", pool, flags, clonevol);
virResetLastError();
if (!VIR_IS_STORAGE_POOL(pool)) {
virLibConnError(NULL, VIR_ERR_INVALID_STORAGE_POOL, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (!VIR_IS_STORAGE_VOL(clonevol)) {
virLibConnError(NULL, VIR_ERR_INVALID_STORAGE_VOL, __FUNCTION__);
goto error;
}
if (pool->conn->flags & VIR_CONNECT_RO ||
clonevol->conn->flags & VIR_CONNECT_RO) {
virLibConnError(pool->conn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (pool->conn->storageDriver &&
pool->conn->storageDriver->volCreateXMLFrom) {
virStorageVolPtr ret;
ret = pool->conn->storageDriver->volCreateXMLFrom (pool, xmldesc,
clonevol, flags);
if (!ret)
goto error;
return ret;
}
virLibConnError (pool->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(pool->conn);
return NULL;
}
/**
* virStorageVolDelete:
* @vol: pointer to storage volume
* @flags: future flags, use 0 for now
*
* Delete the storage volume from the pool
*
* Returns 0 on success, or -1 on error
*/
int
virStorageVolDelete(virStorageVolPtr vol,
unsigned int flags)
{
virConnectPtr conn;
DEBUG("vol=%p, flags=%u", vol, flags);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_STORAGE_VOL(vol)) {
virLibStorageVolError(NULL, VIR_ERR_INVALID_STORAGE_VOL, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = vol->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibStorageVolError(vol, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->volDelete) {
int ret;
ret = conn->storageDriver->volDelete (vol, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(vol->conn);
return -1;
}
/**
* virStorageVolWipe:
* @vol: pointer to storage volume
* @flags: future flags, use 0 for now
*
* Ensure data previously on a volume is not accessible to future reads
*
* Returns 0 on success, or -1 on error
*/
int
virStorageVolWipe(virStorageVolPtr vol,
unsigned int flags)
{
virConnectPtr conn;
VIR_DEBUG("vol=%p, flags=%u", vol, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_STORAGE_VOL(vol)) {
virLibStorageVolError(NULL, VIR_ERR_INVALID_STORAGE_VOL, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = vol->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibStorageVolError(vol, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (conn->storageDriver && conn->storageDriver->volWipe) {
int ret;
ret = conn->storageDriver->volWipe(vol, flags);
if (ret < 0) {
goto error;
}
return ret;
}
virLibConnError(conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(vol->conn);
return -1;
}
/**
* virStorageVolFree:
* @vol: pointer to storage volume
*
* Release the storage volume handle. The underlying
* storage volume continues to exist.
*
* Returns 0 on success, or -1 on error
*/
int
virStorageVolFree(virStorageVolPtr vol)
{
DEBUG("vol=%p", vol);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_STORAGE_VOL(vol)) {
virLibStorageVolError(NULL, VIR_ERR_INVALID_STORAGE_VOL, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (virUnrefStorageVol(vol) < 0) {
virDispatchError(NULL);
return (-1);
}
return(0);
}
/**
* virStorageVolRef:
* @vol: the vol to hold a reference on
*
* Increment the reference count on the vol. For each
* additional call to this method, there shall be a corresponding
* call to virStorageVolFree to release the reference count, once
* the caller no longer needs the reference to this object.
*
* This method is typically useful for applications where multiple
* threads are using a connection, and it is required that the
* connection remain open until all threads have finished using
* it. ie, each new thread using a vol would increment
* the reference count.
*
* Returns 0 in case of success, -1 in case of failure.
*/
int
virStorageVolRef(virStorageVolPtr vol)
{
if ((!VIR_IS_CONNECTED_STORAGE_VOL(vol))) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
virDispatchError(NULL);
return(-1);
}
virMutexLock(&vol->conn->lock);
DEBUG("vol=%p refs=%d", vol, vol->refs);
vol->refs++;
virMutexUnlock(&vol->conn->lock);
return 0;
}
/**
* virStorageVolGetInfo:
* @vol: pointer to storage volume
* @info: pointer at which to store info
*
* Fetches volatile information about the storage
* volume such as its current allocation
*
* Returns 0 on success, or -1 on failure
*/
int
virStorageVolGetInfo(virStorageVolPtr vol,
virStorageVolInfoPtr info)
{
virConnectPtr conn;
DEBUG("vol=%p, info=%p", vol, info);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_STORAGE_VOL(vol)) {
virLibStorageVolError(NULL, VIR_ERR_INVALID_STORAGE_VOL, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (info == NULL) {
virLibStorageVolError(vol, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
memset(info, 0, sizeof(virStorageVolInfo));
conn = vol->conn;
2009-01-20 12:01:45 +00:00
if (conn->storageDriver->volGetInfo){
int ret;
ret = conn->storageDriver->volGetInfo (vol, info);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(vol->conn);
return -1;
}
/**
* virStorageVolGetXMLDesc:
* @vol: pointer to storage volume
* @flags: flags for XML generation (unused, pass 0)
*
* Fetch an XML document describing all aspects of
* the storage volume
*
* Returns the XML document, or NULL on error
*/
char *
virStorageVolGetXMLDesc(virStorageVolPtr vol,
unsigned int flags)
{
virConnectPtr conn;
DEBUG("vol=%p, flags=%u", vol, flags);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_STORAGE_VOL(vol)) {
virLibStorageVolError(NULL, VIR_ERR_INVALID_STORAGE_VOL, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (flags != 0) {
virLibStorageVolError(vol, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
conn = vol->conn;
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->volGetXMLDesc) {
char *ret;
ret = conn->storageDriver->volGetXMLDesc (vol, flags);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(vol->conn);
2009-01-20 12:01:45 +00:00
return NULL;
}
/**
* virStorageVolGetPath:
* @vol: pointer to storage volume
*
* Fetch the storage volume path. Depending on the pool
* configuration this is either persistent across hosts,
* or dynamically assigned at pool startup. Consult
* pool documentation for information on getting the
* persistent naming
*
* Returns the storage volume path, or NULL on error
*/
char *
virStorageVolGetPath(virStorageVolPtr vol)
{
virConnectPtr conn;
DEBUG("vol=%p", vol);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_STORAGE_VOL(vol)) {
virLibStorageVolError(NULL, VIR_ERR_INVALID_STORAGE_VOL, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
conn = vol->conn;
2009-01-20 12:01:45 +00:00
if (conn->storageDriver && conn->storageDriver->volGetPath) {
char *ret;
ret = conn->storageDriver->volGetPath (vol);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(vol->conn);
return NULL;
}
/**
* virNodeNumOfDevices:
* @conn: pointer to the hypervisor connection
* @cap: capability name
* @flags: flags (unused, pass 0)
*
* Provides the number of node devices.
*
* If the optional 'cap' argument is non-NULL, then the count
* will be restricted to devices with the specified capability
*
* Returns the number of node devices or -1 in case of error
*/
int
virNodeNumOfDevices(virConnectPtr conn, const char *cap, unsigned int flags)
{
2009-01-29 17:18:24 +00:00
DEBUG("conn=%p, cap=%s, flags=%d", conn, NULLSTR(cap), flags);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (flags != 0) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->deviceMonitor && conn->deviceMonitor->numOfDevices) {
int ret;
ret = conn->deviceMonitor->numOfDevices (conn, cap, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return -1;
}
/**
* virNodeListDevices:
* @conn: pointer to the hypervisor connection
* @cap: capability name
* @names: array to collect the list of node device names
* @maxnames: size of @names
* @flags: flags (unused, pass 0)
*
* Collect the list of node devices, and store their names in @names
*
* If the optional 'cap' argument is non-NULL, then the count
* will be restricted to devices with the specified capability
*
* Returns the number of node devices found or -1 in case of error
*/
int
virNodeListDevices(virConnectPtr conn,
const char *cap,
char **const names, int maxnames,
unsigned int flags)
{
DEBUG("conn=%p, cap=%s, names=%p, maxnames=%d, flags=%d",
conn, cap, names, maxnames, flags);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if ((flags != 0) || (names == NULL) || (maxnames < 0)) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->deviceMonitor && conn->deviceMonitor->listDevices) {
int ret;
ret = conn->deviceMonitor->listDevices (conn, cap, names, maxnames, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return -1;
}
/**
* virNodeDeviceLookupByName:
* @conn: pointer to the hypervisor connection
* @name: unique device name
*
* Lookup a node device by its name.
*
* Returns a virNodeDevicePtr if found, NULL otherwise.
*/
virNodeDevicePtr virNodeDeviceLookupByName(virConnectPtr conn, const char *name)
{
DEBUG("conn=%p, name=%p", conn, name);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
if (name == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if (conn->deviceMonitor && conn->deviceMonitor->deviceLookupByName) {
virNodeDevicePtr ret;
ret = conn->deviceMonitor->deviceLookupByName (conn, name);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(conn);
return NULL;
}
/**
* virNodeDeviceGetXMLDesc:
* @dev: pointer to the node device
* @flags: flags for XML generation (unused, pass 0)
*
* Fetch an XML document describing all aspects of
* the device.
*
* Returns the XML document, or NULL on error
*/
char *virNodeDeviceGetXMLDesc(virNodeDevicePtr dev, unsigned int flags)
{
DEBUG("dev=%p, conn=%p", dev, dev ? dev->conn : NULL);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_NODE_DEVICE(dev)) {
virLibNodeDeviceError(NULL, VIR_ERR_INVALID_NODE_DEVICE, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
2009-01-20 12:01:45 +00:00
if (dev->conn->deviceMonitor && dev->conn->deviceMonitor->deviceDumpXML) {
char *ret;
ret = dev->conn->deviceMonitor->deviceDumpXML (dev, flags);
if (!ret)
goto error;
return ret;
}
virLibConnError (dev->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(dev->conn);
return NULL;
}
/**
* virNodeDeviceGetName:
* @dev: the device
*
* Just return the device name
*
* Returns the device name or NULL in case of error
*/
const char *virNodeDeviceGetName(virNodeDevicePtr dev)
{
DEBUG("dev=%p, conn=%p", dev, dev ? dev->conn : NULL);
if (!VIR_IS_CONNECTED_NODE_DEVICE(dev)) {
virLibNodeDeviceError(NULL, VIR_ERR_INVALID_NODE_DEVICE, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
return dev->name;
}
/**
* virNodeDeviceGetParent:
* @dev: the device
*
* Accessor for the parent of the device
*
* Returns the name of the device's parent, or NULL if the
* device has no parent.
*/
const char *virNodeDeviceGetParent(virNodeDevicePtr dev)
{
DEBUG("dev=%p, conn=%p", dev, dev ? dev->conn : NULL);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_NODE_DEVICE(dev)) {
virLibNodeDeviceError(NULL, VIR_ERR_INVALID_NODE_DEVICE, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
if (!dev->parent) {
if (dev->conn->deviceMonitor && dev->conn->deviceMonitor->deviceGetParent) {
dev->parent = dev->conn->deviceMonitor->deviceGetParent (dev);
} else {
virLibConnError (dev->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
virDispatchError(dev->conn);
return NULL;
}
}
return dev->parent;
}
/**
* virNodeDeviceNumOfCaps:
* @dev: the device
*
* Accessor for the number of capabilities supported by the device.
*
* Returns the number of capabilities supported by the device.
*/
int virNodeDeviceNumOfCaps(virNodeDevicePtr dev)
{
DEBUG("dev=%p, conn=%p", dev, dev ? dev->conn : NULL);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_NODE_DEVICE(dev)) {
virLibNodeDeviceError(NULL, VIR_ERR_INVALID_NODE_DEVICE, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
2009-01-20 12:01:45 +00:00
if (dev->conn->deviceMonitor && dev->conn->deviceMonitor->deviceNumOfCaps) {
int ret;
ret = dev->conn->deviceMonitor->deviceNumOfCaps (dev);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (dev->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(dev->conn);
return -1;
}
/**
* virNodeDeviceListCaps:
* @dev: the device
* @names: array to collect the list of capability names
* @maxnames: size of @names
*
* Lists the names of the capabilities supported by the device.
*
* Returns the number of capability names listed in @names.
*/
int virNodeDeviceListCaps(virNodeDevicePtr dev,
char **const names,
int maxnames)
{
DEBUG("dev=%p, conn=%p, names=%p, maxnames=%d",
dev, dev ? dev->conn : NULL, names, maxnames);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_NODE_DEVICE(dev)) {
virLibNodeDeviceError(NULL, VIR_ERR_INVALID_NODE_DEVICE, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
2009-01-20 12:01:45 +00:00
if (dev->conn->deviceMonitor && dev->conn->deviceMonitor->deviceListCaps) {
int ret;
ret = dev->conn->deviceMonitor->deviceListCaps (dev, names, maxnames);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (dev->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
2009-01-20 12:01:45 +00:00
error:
virDispatchError(dev->conn);
return -1;
}
/**
* virNodeDeviceFree:
* @dev: pointer to the node device
*
* Drops a reference to the node device, freeing it if
* this was the last reference.
*
* Returns the 0 for success, -1 for error.
*/
int virNodeDeviceFree(virNodeDevicePtr dev)
{
DEBUG("dev=%p, conn=%p", dev, dev ? dev->conn : NULL);
2009-01-20 12:01:45 +00:00
virResetLastError();
if (!VIR_IS_CONNECTED_NODE_DEVICE(dev)) {
virLibNodeDeviceError(NULL, VIR_ERR_INVALID_NODE_DEVICE, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (virUnrefNodeDevice(dev) < 0) {
virDispatchError(NULL);
return (-1);
}
return(0);
}
/**
* virNodeDeviceRef:
* @dev: the dev to hold a reference on
*
* Increment the reference count on the dev. For each
* additional call to this method, there shall be a corresponding
* call to virNodeDeviceFree to release the reference count, once
* the caller no longer needs the reference to this object.
*
* This method is typically useful for applications where multiple
* threads are using a connection, and it is required that the
* connection remain open until all threads have finished using
* it. ie, each new thread using a dev would increment
* the reference count.
*
* Returns 0 in case of success, -1 in case of failure.
*/
int
virNodeDeviceRef(virNodeDevicePtr dev)
{
if ((!VIR_IS_CONNECTED_NODE_DEVICE(dev))) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
virDispatchError(NULL);
return(-1);
}
virMutexLock(&dev->conn->lock);
DEBUG("dev=%p refs=%d", dev, dev->refs);
dev->refs++;
virMutexUnlock(&dev->conn->lock);
return 0;
}
/**
* virNodeDeviceDettach:
* @dev: pointer to the node device
*
* Dettach the node device from the node itself so that it may be
* assigned to a guest domain.
*
* Depending on the hypervisor, this may involve operations such
* as unbinding any device drivers from the device, binding the
* device to a dummy device driver and resetting the device.
*
* If the device is currently in use by the node, this method may
* fail.
*
* Once the device is not assigned to any guest, it may be re-attached
* to the node using the virNodeDeviceReattach() method.
*
* Returns 0 in case of success, -1 in case of failure.
*/
int
virNodeDeviceDettach(virNodeDevicePtr dev)
{
DEBUG("dev=%p, conn=%p", dev, dev ? dev->conn : NULL);
virResetLastError();
if (!VIR_IS_CONNECTED_NODE_DEVICE(dev)) {
virLibNodeDeviceError(NULL, VIR_ERR_INVALID_NODE_DEVICE, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (dev->conn->driver->nodeDeviceDettach) {
int ret;
ret = dev->conn->driver->nodeDeviceDettach (dev);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(dev->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(dev->conn);
return (-1);
}
/**
* virNodeDeviceReAttach:
* @dev: pointer to the node device
*
* Re-attach a previously dettached node device to the node so that it
* may be used by the node again.
*
* Depending on the hypervisor, this may involve operations such
* as resetting the device, unbinding it from a dummy device driver
* and binding it to its appropriate driver.
*
* If the device is currently in use by a guest, this method may fail.
*
* Returns 0 in case of success, -1 in case of failure.
*/
int
virNodeDeviceReAttach(virNodeDevicePtr dev)
{
DEBUG("dev=%p, conn=%p", dev, dev ? dev->conn : NULL);
virResetLastError();
if (!VIR_IS_CONNECTED_NODE_DEVICE(dev)) {
virLibNodeDeviceError(NULL, VIR_ERR_INVALID_NODE_DEVICE, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (dev->conn->driver->nodeDeviceReAttach) {
int ret;
ret = dev->conn->driver->nodeDeviceReAttach (dev);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(dev->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(dev->conn);
return (-1);
}
/**
* virNodeDeviceReset:
* @dev: pointer to the node device
*
* Reset a previously dettached node device to the node before or
* after assigning it to a guest.
*
* The exact reset semantics depends on the hypervisor and device
* type but, for example, KVM will attempt to reset PCI devices with
* a Function Level Reset, Secondary Bus Reset or a Power Management
* D-State reset.
*
* If the reset will affect other devices which are currently in use,
* this function may fail.
*
* Returns 0 in case of success, -1 in case of failure.
*/
int
virNodeDeviceReset(virNodeDevicePtr dev)
{
DEBUG("dev=%p, conn=%p", dev, dev ? dev->conn : NULL);
virResetLastError();
if (!VIR_IS_CONNECTED_NODE_DEVICE(dev)) {
virLibNodeDeviceError(NULL, VIR_ERR_INVALID_NODE_DEVICE, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (dev->conn->driver->nodeDeviceReset) {
int ret;
ret = dev->conn->driver->nodeDeviceReset (dev);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(dev->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(dev->conn);
return (-1);
}
/**
* virNodeDeviceCreateXML:
* @conn: pointer to the hypervisor connection
* @xmlDesc: string containing an XML description of the device to be created
* @flags: callers should always pass 0
*
* Create a new device on the VM host machine, for example, virtual
* HBAs created using vport_create.
*
* Returns a node device object if successful, NULL in case of failure
*/
virNodeDevicePtr
virNodeDeviceCreateXML(virConnectPtr conn,
const char *xmlDesc,
unsigned int flags)
{
VIR_DEBUG("conn=%p, xmlDesc=%s, flags=%d", conn, xmlDesc, flags);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
if (conn->flags & VIR_CONNECT_RO) {
virLibConnError(conn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (xmlDesc == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->deviceMonitor &&
conn->deviceMonitor->deviceCreateXML) {
virNodeDevicePtr dev = conn->deviceMonitor->deviceCreateXML(conn, xmlDesc, flags);
if (dev == NULL)
goto error;
return dev;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return NULL;
}
/**
* virNodeDeviceDestroy:
* @dev: a device object
*
* Destroy the device object. The virtual device is removed from the host operating system.
* This function may require privileged access
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virNodeDeviceDestroy(virNodeDevicePtr dev)
{
DEBUG("dev=%p", dev);
virResetLastError();
if (!VIR_IS_CONNECTED_NODE_DEVICE(dev)) {
virLibNodeDeviceError(NULL, VIR_ERR_INVALID_NODE_DEVICE, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (dev->conn->flags & VIR_CONNECT_RO) {
virLibConnError(dev->conn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (dev->conn->deviceMonitor &&
dev->conn->deviceMonitor->deviceDestroy) {
int retval = dev->conn->deviceMonitor->deviceDestroy(dev);
if (retval < 0) {
goto error;
}
return 0;
}
virLibConnError (dev->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(dev->conn);
return -1;
}
/*
* Domain Event Notification
*/
/**
* virConnectDomainEventRegister:
* @conn: pointer to the connection
* @cb: callback to the function handling domain events
* @opaque: opaque data to pass on to the callback
* @freecb: optional function to deallocate opaque when not used anymore
*
Introduce a new public API for domain events The current API for domain events has a number of problems - Only allows for domain lifecycle change events - Does not allow the same callback to be registered multiple times - Does not allow filtering of events to a specific domain This introduces a new more general purpose domain events API typedef enum { VIR_DOMAIN_EVENT_ID_LIFECYCLE = 0, /* virConnectDomainEventCallback */ ...more events later.. } int virConnectDomainEventRegisterAny(virConnectPtr conn, virDomainPtr dom, /* Optional, to filter */ int eventID, virConnectDomainEventGenericCallback cb, void *opaque, virFreeCallback freecb); int virConnectDomainEventDeregisterAny(virConnectPtr conn, int callbackID); Since different event types can received different data in the callback, the API is defined with a generic callback. Specific events will each have a custom signature for their callback. Thus when registering an event it is neccessary to cast the callback to the generic signature eg int myDomainEventCallback(virConnectPtr conn, virDomainPtr dom, int event, int detail, void *opaque) { ... } virConnectDomainEventRegisterAny(conn, NULL, VIR_DOMAIN_EVENT_ID_LIFECYCLE, VIR_DOMAIN_EVENT_CALLBACK(myDomainEventCallback) NULL, NULL); The VIR_DOMAIN_EVENT_CALLBACK() macro simply does a "bad" cast to the generic signature * include/libvirt/libvirt.h.in: Define new APIs for registering domain events * src/driver.h: Internal driver entry points for new events APIs * src/libvirt.c: Wire up public API to driver API for events APIs * src/libvirt_public.syms: Export new APIs * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c, src/xenapi/xenapi_driver.c: Stub out new API entries
2010-03-18 13:01:48 +00:00
* Adds a callback to receive notifications of domain lifecycle events
* occurring on a connection
*
* Use of this method is no longer recommended. Instead applications
* should try virConnectDomainEventRegisterAny which has a more flexible
* API contract
*
* The virDomainPtr object handle passed into the callback upon delivery
* of an event is only valid for the duration of execution of the callback.
* If the callback wishes to keep the domain object after the callback
* returns, it shall take a reference to it, by calling virDomainRef.
* The reference can be released once the object is no longer required
* by calling virDomainFree.
*
* Returns 0 on success, -1 on failure
*/
int
virConnectDomainEventRegister(virConnectPtr conn,
virConnectDomainEventCallback cb,
void *opaque,
virFreeCallback freecb)
{
2009-01-20 12:01:45 +00:00
DEBUG("conn=%p, cb=%p, opaque=%p, freecb=%p", conn, cb, opaque, freecb);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (cb == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
2009-01-20 12:01:45 +00:00
if ((conn->driver) && (conn->driver->domainEventRegister)) {
int ret;
ret = conn->driver->domainEventRegister (conn, cb, opaque, freecb);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virConnectDomainEventDeregister:
* @conn: pointer to the connection
* @cb: callback to the function handling domain events
*
Introduce a new public API for domain events The current API for domain events has a number of problems - Only allows for domain lifecycle change events - Does not allow the same callback to be registered multiple times - Does not allow filtering of events to a specific domain This introduces a new more general purpose domain events API typedef enum { VIR_DOMAIN_EVENT_ID_LIFECYCLE = 0, /* virConnectDomainEventCallback */ ...more events later.. } int virConnectDomainEventRegisterAny(virConnectPtr conn, virDomainPtr dom, /* Optional, to filter */ int eventID, virConnectDomainEventGenericCallback cb, void *opaque, virFreeCallback freecb); int virConnectDomainEventDeregisterAny(virConnectPtr conn, int callbackID); Since different event types can received different data in the callback, the API is defined with a generic callback. Specific events will each have a custom signature for their callback. Thus when registering an event it is neccessary to cast the callback to the generic signature eg int myDomainEventCallback(virConnectPtr conn, virDomainPtr dom, int event, int detail, void *opaque) { ... } virConnectDomainEventRegisterAny(conn, NULL, VIR_DOMAIN_EVENT_ID_LIFECYCLE, VIR_DOMAIN_EVENT_CALLBACK(myDomainEventCallback) NULL, NULL); The VIR_DOMAIN_EVENT_CALLBACK() macro simply does a "bad" cast to the generic signature * include/libvirt/libvirt.h.in: Define new APIs for registering domain events * src/driver.h: Internal driver entry points for new events APIs * src/libvirt.c: Wire up public API to driver API for events APIs * src/libvirt_public.syms: Export new APIs * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c, src/xenapi/xenapi_driver.c: Stub out new API entries
2010-03-18 13:01:48 +00:00
* Removes a callback previously registered with the virConnectDomainEventRegister
* funtion.
*
* Use of this method is no longer recommended. Instead applications
* should try virConnectDomainEventUnregisterAny which has a more flexible
* API contract
*
* Returns 0 on success, -1 on failure
*/
int
virConnectDomainEventDeregister(virConnectPtr conn,
virConnectDomainEventCallback cb)
{
2009-01-20 12:01:45 +00:00
DEBUG("conn=%p, cb=%p", conn, cb);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (cb == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
2009-01-20 12:01:45 +00:00
goto error;
}
if ((conn->driver) && (conn->driver->domainEventDeregister)) {
int ret;
ret = conn->driver->domainEventDeregister (conn, cb);
if (ret < 0)
goto error;
return ret;
}
2009-01-20 12:01:45 +00:00
virLibConnError(conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virSecretGetConnect:
* @secret: A virSecret secret
*
* Provides the connection pointer associated with a secret. The reference
* counter on the connection is not increased by this call.
*
* WARNING: When writing libvirt bindings in other languages, do not use this
* function. Instead, store the connection and the secret object together.
*
* Returns the virConnectPtr or NULL in case of failure.
*/
virConnectPtr
virSecretGetConnect (virSecretPtr secret)
{
DEBUG("secret=%p", secret);
virResetLastError();
if (!VIR_IS_CONNECTED_SECRET (secret)) {
virLibSecretError (NULL, VIR_ERR_INVALID_SECRET, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
return secret->conn;
}
/**
* virConnectNumOfSecrets:
* @conn: virConnect connection
*
* Fetch number of currently defined secrets.
*
* Returns the number currently defined secrets.
*/
int
virConnectNumOfSecrets(virConnectPtr conn)
{
VIR_DEBUG("conn=%p", conn);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
if (conn->secretDriver != NULL &&
conn->secretDriver->numOfSecrets != NULL) {
int ret;
ret = conn->secretDriver->numOfSecrets(conn);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virConnectListSecrets:
* @conn: virConnect connection
* @uuids: Pointer to an array to store the UUIDs
* @maxuuids: size of the array.
*
* List UUIDs of defined secrets, store pointers to names in uuids.
*
* Returns the number of UUIDs provided in the array, or -1 on failure.
*/
int
virConnectListSecrets(virConnectPtr conn, char **uuids, int maxuuids)
{
VIR_DEBUG("conn=%p, uuids=%p, maxuuids=%d", conn, uuids, maxuuids);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
if (uuids == NULL || maxuuids < 0) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->secretDriver != NULL && conn->secretDriver->listSecrets != NULL) {
int ret;
ret = conn->secretDriver->listSecrets(conn, uuids, maxuuids);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
* virSecretLookupByUUID:
* @conn: pointer to the hypervisor connection
* @uuid: the raw UUID for the secret
*
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
* Try to lookup a secret on the given hypervisor based on its UUID.
* Uses the 16 bytes of raw data to describe the UUID
*
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
* Returns a new secret object or NULL in case of failure. If the
* secret cannot be found, then VIR_ERR_NO_SECRET error is raised.
*/
virSecretPtr
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
virSecretLookupByUUID(virConnectPtr conn, const unsigned char *uuid)
{
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
DEBUG("conn=%p, uuid=%s", conn, uuid);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
return (NULL);
}
if (uuid == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
if (conn->secretDriver &&
conn->secretDriver->lookupByUUID) {
virSecretPtr ret;
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
ret = conn->secretDriver->lookupByUUID (conn, uuid);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return NULL;
}
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
/**
* virSecretLookupByUUIDString:
* @conn: pointer to the hypervisor connection
* @uuidstr: the string UUID for the secret
*
* Try to lookup a secret on the given hypervisor based on its UUID.
* Uses the printable string value to describe the UUID
*
* Returns a new secret object or NULL in case of failure. If the
* secret cannot be found, then VIR_ERR_NO_SECRET error is raised.
*/
virSecretPtr
virSecretLookupByUUIDString(virConnectPtr conn, const char *uuidstr)
{
unsigned char uuid[VIR_UUID_BUFLEN];
DEBUG("conn=%p, uuidstr=%s", conn, uuidstr);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
return (NULL);
}
if (uuidstr == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (virUUIDParse(uuidstr, uuid) < 0) {
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
return virSecretLookupByUUID(conn, &uuid[0]);
error:
virDispatchError(conn);
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
return NULL;
}
/**
* virSecretLookupByUsage:
* @conn: pointer to the hypervisor connection
* @usageType: the type of secret usage
* @usageID: identifier of the object using the secret
*
* Try to lookup a secret on the given hypervisor based on its usage
* The usageID is unique within the set of secrets sharing the
* same usageType value.
*
* Returns a new secret object or NULL in case of failure. If the
* secret cannot be found, then VIR_ERR_NO_SECRET error is raised.
*/
virSecretPtr
virSecretLookupByUsage(virConnectPtr conn,
int usageType,
const char *usageID)
{
DEBUG("conn=%p, usageType=%d usageID=%s", conn, usageType, NULLSTR(usageID));
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (usageID == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->secretDriver &&
conn->secretDriver->lookupByUsage) {
virSecretPtr ret;
ret = conn->secretDriver->lookupByUsage (conn, usageType, usageID);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return NULL;
}
/**
* virSecretDefineXML:
* @conn: virConnect connection
* @xml: XML describing the secret.
* @flags: flags, use 0 for now
*
2009-10-26 23:02:46 +00:00
* If XML specifies a UUID, locates the specified secret and replaces all
* attributes of the secret specified by UUID by attributes specified in xml
* (any attributes not specified in xml are discarded).
*
* Otherwise, creates a new secret with an automatically chosen UUID, and
* initializes its attributes from xml.
*
* Returns a the secret on success, NULL on failure.
*/
virSecretPtr
virSecretDefineXML(virConnectPtr conn, const char *xml, unsigned int flags)
{
VIR_DEBUG("conn=%p, xml=%s, flags=%u", conn, xml, flags);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
if (conn->flags & VIR_CONNECT_RO) {
virLibConnError(conn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (xml == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->secretDriver != NULL && conn->secretDriver->defineXML != NULL) {
virSecretPtr ret;
ret = conn->secretDriver->defineXML(conn, xml, flags);
if (ret == NULL)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return NULL;
}
/**
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
* virSecretGetUUID:
* @secret: A virSecret secret
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
* @uuid: buffer of VIR_UUID_BUFLEN bytes in size
*
* Fetches the UUID of the secret.
*
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
* Returns 0 on success with the uuid buffer being filled, or
* -1 upon failure.
*/
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
int
virSecretGetUUID(virSecretPtr secret, unsigned char *uuid)
{
VIR_DEBUG("secret=%p", secret);
virResetLastError();
if (!VIR_IS_CONNECTED_SECRET(secret)) {
virLibSecretError(NULL, VIR_ERR_INVALID_SECRET, __FUNCTION__);
virDispatchError(NULL);
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
return -1;
}
if (uuid == NULL) {
virLibSecretError(secret, VIR_ERR_INVALID_ARG, __FUNCTION__);
virDispatchError(secret->conn);
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
return -1;
}
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
memcpy(uuid, &secret->uuid[0], VIR_UUID_BUFLEN);
return 0;
}
/**
* virSecretGetUUIDString:
* @secret: a secret object
* @buf: pointer to a VIR_UUID_STRING_BUFLEN bytes array
*
* Get the UUID for a secret as string. For more information about
* UUID see RFC4122.
*
* Returns -1 in case of error, 0 in case of success
*/
int
virSecretGetUUIDString(virSecretPtr secret, char *buf)
{
unsigned char uuid[VIR_UUID_BUFLEN];
DEBUG("secret=%p, buf=%p", secret, buf);
virResetLastError();
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
if (!VIR_IS_SECRET(secret)) {
virLibSecretError(NULL, VIR_ERR_INVALID_SECRET, __FUNCTION__);
virDispatchError(NULL);
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
return (-1);
}
if (buf == NULL) {
virLibSecretError(secret, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (virSecretGetUUID(secret, &uuid[0]))
goto error;
virUUIDFormat(uuid, buf);
return (0);
error:
virDispatchError(secret->conn);
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
return -1;
}
/**
* virSecretGetUsageType:
* @secret: a secret object
*
* Get the type of object which uses this secret. The returned
* value is one of the constants defined in the virSecretUsageType
* enumeration. More values may be added to this enumeration in
* the future, so callers should expect to see usage types they
* do not explicitly know about.
*
* Returns a positive integer identifying the type of object,
* or -1 upon error.
*/
int
virSecretGetUsageType(virSecretPtr secret)
{
DEBUG("secret=%p", secret);
virResetLastError();
if (!VIR_IS_SECRET(secret)) {
virLibSecretError(NULL, VIR_ERR_INVALID_SECRET, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
return (secret->usageType);
}
/**
* virSecretGetUsageID:
* @secret: a secret object
*
* Get the unique identifier of the object with which this
* secret is to be used. The format of the identifier is
* dependant on the usage type of the secret. For a secret
* with a usage type of VIR_SECRET_USAGE_TYPE_VOLUME the
* identifier will be a fully qualfied path name. The
* identifiers are intended to be unique within the set of
* all secrets sharing the same usage type. ie, there shall
* only ever be one secret for each volume path.
*
* Returns a string identifying the object using the secret,
* or NULL upon error
*/
const char *
virSecretGetUsageID(virSecretPtr secret)
{
DEBUG("secret=%p", secret);
virResetLastError();
if (!VIR_IS_SECRET(secret)) {
virLibSecretError(NULL, VIR_ERR_INVALID_SECRET, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
return (secret->usageID);
}
Fix UUID handling in secrets/storage encryption APIs Convert all the secret/storage encryption APIs / wire format to handle UUIDs in raw format instead of non-canonical printable format. Guarentees data format correctness. * docs/schemas/storageencryption.rng: Make UUID mandatory for a secret and validate fully * docs/schemas/secret.rng: Fully validate UUID * include/libvirt/libvirt.h, include/libvirt/libvirt.h.in, Add virSecretLookupByUUID and virSecretGetUUID. Make virSecretGetUUIDString follow normal API design pattern * python/generator.py: Skip generation of virSecretGetUUID, virSecretGetUUIDString and virSecretLookupByUUID * python/libvir.c, python/libvirt-python-api.xml: Manual impl of virSecretGetUUID,virSecretGetUUIDString and virSecretLookupByUUID * qemud/remote.c: s/virSecretLookupByUUIDString/virSecretLookupByUUID/ Fix get_nonnull_secret/make_nonnull_secret to use unsigned char * qemud/remote_protocol.x: Fix remote_nonnull_secret to use a remote_uuid instead of remote_nonnull_string for UUID field. Rename REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING to REMOTE_PROC_SECRET_LOOKUP_BY_UUID_STRING and make it take an remote_uuid value * qemud/remote_dispatch_args.h, qemud/remote_dispatch_prototypes.h, qemud/remote_dispatch_ret.h, qemud/remote_dispatch_table.h, qemud/remote_protocol.c, qemud/remote_protocol.h: Re-generate * src/datatypes.h, src/datatypes.c: Store UUID in raw format instead of printable. Change virGetSecret to use raw format UUID * src/driver.h: Rename virDrvSecretLookupByUUIDString to virDrvSecretLookupByUUID and use raw format UUID * src/libvirt.c: Add virSecretLookupByUUID and virSecretGetUUID and re-implement virSecretLookupByUUIDString and virSecretGetUUIDString in terms of those * src/libvirt_public.syms: Add virSecretLookupByUUID and virSecretGetUUID * src/remote_internal.c: Rename remoteSecretLookupByUUIDString to remoteSecretLookupByUUID. Fix typo in args for remoteSecretDefineXML impl. Use raw UUID format for get_nonnull_secret and make_nonnull_secret * src/storage_encryption_conf.c, src/storage_encryption_conf.h: Storage UUID in raw format, and require it to be present in XML. Use UUID parser to validate. * secret_conf.h, secret_conf.c: Generate a UUID if none is provided. Storage UUID in raw format. * src/secret_driver.c: Adjust to deal with raw UUIDs. Save secrets in a filed with printable UUID, instead of base64 UUID. * src/virsh.c: Adjust for changed public API contract of virSecretGetUUIDString. * src/storage_Backend.c: DOn't undefine secret we just generated upon successful volume creation. Fix to handle raw UUIDs. Generate a non-clashing UUID * src/qemu_driver.c: Change to use lookupByUUID instead of lookupByUUIDString
2009-09-10 16:44:12 +00:00
/**
* virSecretGetXMLDesc:
* @secret: A virSecret secret
* @flags: flags, use 0 for now
*
* Fetches an XML document describing attributes of the secret.
*
* Returns the XML document on success, NULL on failure. The caller must
* free() the XML.
*/
char *
virSecretGetXMLDesc(virSecretPtr secret, unsigned int flags)
{
virConnectPtr conn;
VIR_DEBUG("secret=%p, flags=%u", secret, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_SECRET(secret)) {
virLibSecretError(NULL, VIR_ERR_INVALID_SECRET, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
conn = secret->conn;
if (conn->secretDriver != NULL && conn->secretDriver->getXMLDesc != NULL) {
char *ret;
ret = conn->secretDriver->getXMLDesc(secret, flags);
if (ret == NULL)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return NULL;
}
/**
* virSecretSetValue:
* @secret: A virSecret secret
* @value: Value of the secret
* @value_size: Size of the value
* @flags: flags, use 0 for now
*
* Sets the value of a secret.
*
* Returns 0 on success, -1 on failure.
*/
int
virSecretSetValue(virSecretPtr secret, const unsigned char *value,
size_t value_size, unsigned int flags)
{
virConnectPtr conn;
VIR_DEBUG("secret=%p, value=%p, value_size=%zu, flags=%u", secret, value,
value_size, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_SECRET(secret)) {
virLibSecretError(NULL, VIR_ERR_INVALID_SECRET, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
conn = secret->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibSecretError(secret, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (value == NULL) {
virLibSecretError(secret, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->secretDriver != NULL && conn->secretDriver->setValue != NULL) {
int ret;
ret = conn->secretDriver->setValue(secret, value, value_size, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virSecretGetValue:
* @secret: A virSecret connection
* @value_size: Place for storing size of the secret value
* @flags: flags, use 0 for now
*
* Fetches the value of a secret.
*
* Returns the secret value on success, NULL on failure. The caller must
* free() the secret value.
*/
unsigned char *
virSecretGetValue(virSecretPtr secret, size_t *value_size, unsigned int flags)
{
virConnectPtr conn;
VIR_DEBUG("secret=%p, value_size=%p, flags=%u", secret, value_size, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_SECRET(secret)) {
virLibSecretError(NULL, VIR_ERR_INVALID_SECRET, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
conn = secret->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibSecretError(secret, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (value_size == NULL) {
virLibSecretError(secret, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
flags &= VIR_SECRET_GET_VALUE_FLAGS_MASK;
if (conn->secretDriver != NULL && conn->secretDriver->getValue != NULL) {
unsigned char *ret;
ret = conn->secretDriver->getValue(secret, value_size, flags);
if (ret == NULL)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return NULL;
}
/**
* virSecretUndefine:
* @secret: A virSecret secret
*
* Deletes the specified secret. This does not free the associated
* virSecretPtr object.
*
* Returns 0 on success, -1 on failure.
*/
int
virSecretUndefine(virSecretPtr secret)
{
virConnectPtr conn;
VIR_DEBUG("secret=%p", secret);
virResetLastError();
if (!VIR_IS_CONNECTED_SECRET(secret)) {
virLibSecretError(NULL, VIR_ERR_INVALID_SECRET, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
conn = secret->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibSecretError(secret, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (conn->secretDriver != NULL && conn->secretDriver->undefine != NULL) {
int ret;
ret = conn->secretDriver->undefine(secret);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virSecretRef:
* @secret: the secret to hold a reference on
*
* Increment the reference count on the secret. For each additional call to
* this method, there shall be a corresponding call to virSecretFree to release
* the reference count, once the caller no longer needs the reference to this
* object.
*
* This method is typically useful for applications where multiple threads are
* using a connection, and it is required that the connection remain open until
* all threads have finished using it. ie, each new thread using a secret would
* increment the reference count.
*
* Returns 0 in case of success, -1 in case of failure.
*/
int
virSecretRef(virSecretPtr secret)
{
if (!VIR_IS_CONNECTED_SECRET(secret)) {
virLibSecretError(NULL, VIR_ERR_INVALID_SECRET, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
virMutexLock(&secret->conn->lock);
DEBUG("secret=%p refs=%d", secret, secret->refs);
secret->refs++;
virMutexUnlock(&secret->conn->lock);
return 0;
}
/**
* virSecretFree:
* @secret: pointer to a secret
*
* Release the secret handle. The underlying secret continues to exist.
*
* Returns 0 on success, or -1 on error
*/
int
virSecretFree(virSecretPtr secret)
{
DEBUG("secret=%p", secret);
virResetLastError();
if (!VIR_IS_CONNECTED_SECRET(secret)) {
virLibSecretError(NULL, VIR_ERR_INVALID_SECRET, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
if (virUnrefSecret(secret) < 0) {
virDispatchError(NULL);
return -1;
}
return 0;
}
/**
* virStreamNew:
* @conn: pointer to the connection
* @flags: control features of the stream
*
* Creates a new stream object which can be used to perform
* streamed I/O with other public API function.
*
* When no longer needed, a stream object must be released
* with virStreamFree. If a data stream has been used,
* then the application must call virStreamFinish or
* virStreamAbort before free'ing to, in order to notify
* the driver of termination.
*
* If a non-blocking data stream is required passed
* VIR_STREAM_NONBLOCK for flags, otherwise pass 0.
*
* Returns the new stream, or NULL upon error
*/
virStreamPtr
virStreamNew(virConnectPtr conn,
unsigned int flags)
{
virStreamPtr st;
DEBUG("conn=%p, flags=%u", conn, flags);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
st = virGetStream(conn);
if (st)
st->flags = flags;
return st;
}
/**
* virStreamRef:
* @stream: pointer to the stream
*
* Increment the reference count on the stream. For each
* additional call to this method, there shall be a corresponding
* call to virStreamFree to release the reference count, once
* the caller no longer needs the reference to this object.
*
* Returns 0 in case of success, -1 in case of failure
*/
int
virStreamRef(virStreamPtr stream)
{
if ((!VIR_IS_CONNECTED_STREAM(stream))) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
virDispatchError(NULL);
return(-1);
}
virMutexLock(&stream->conn->lock);
DEBUG("stream=%p refs=%d", stream, stream->refs);
stream->refs++;
virMutexUnlock(&stream->conn->lock);
return 0;
}
/**
* virStreamSend:
* @stream: pointer to the stream object
* @data: buffer to write to stream
* @nbytes: size of @data buffer
*
* Write a series of bytes to the stream. This method may
* block the calling application for an arbitrary amount
* of time. Once an application has finished sending data
* it should call virStreamFinish to wait for successful
* confirmation from the driver, or detect any error
*
* This method may not be used if a stream source has been
* registered
*
* Errors are not guaranteed to be reported synchronously
* with the call, but may instead be delayed until a
* subsequent call.
*
2009-10-26 23:02:46 +00:00
* An example using this with a hypothetical file upload
* API looks like
*
* virStreamPtr st = virStreamNew(conn, 0);
* int fd = open("demo.iso", O_RDONLY)
*
* virConnectUploadFile(conn, "demo.iso", st);
*
* while (1) {
* char buf[1024];
* int got = read(fd, buf, 1024);
* if (got < 0) {
* virStreamAbort(st);
* break;
* }
* if (got == 0) {
* virStreamFinish(st);
* break;
* }
* int offset = 0;
* while (offset < got) {
* int sent = virStreamSend(st, buf+offset, got-offset)
* if (sent < 0) {
* virStreamAbort(st);
* goto done;
* }
* offset += sent;
* }
* }
* if (virStreamFinish(st) < 0)
* ... report an error ....
* done:
* virStreamFree(st);
* close(fd);
*
* Returns the number of bytes written, which may be less
* than requested.
*
* Returns -1 upon error, at which time the stream will
* be marked as aborted, and the caller should now release
* the stream with virStreamFree.
*
* Returns -2 if the outgoing transmit buffers are full &
* the stream is marked as non-blocking.
*/
int virStreamSend(virStreamPtr stream,
const char *data,
size_t nbytes)
{
DEBUG("stream=%p, data=%p, nbytes=%zi", stream, data, nbytes);
virResetLastError();
if (!VIR_IS_CONNECTED_STREAM(stream)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (stream->driver &&
stream->driver->streamSend) {
int ret;
ret = (stream->driver->streamSend)(stream, data, nbytes);
if (ret == -2)
return -2;
if (ret < 0)
goto error;
return ret;
}
virLibConnError(stream->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(stream->conn);
return -1;
}
/**
* virStreamRecv:
* @stream: pointer to the stream object
* @data: buffer to write to stream
* @nbytes: size of @data buffer
*
* Write a series of bytes to the stream. This method may
* block the calling application for an arbitrary amount
* of time.
*
* Errors are not guaranteed to be reported synchronously
* with the call, but may instead be delayed until a
* subsequent call.
*
2009-10-26 23:02:46 +00:00
* An example using this with a hypothetical file download
* API looks like
*
* virStreamPtr st = virStreamNew(conn, 0);
* int fd = open("demo.iso", O_WRONLY, 0600)
*
* virConnectDownloadFile(conn, "demo.iso", st);
*
* while (1) {
* char buf[1024];
* int got = virStreamRecv(st, buf, 1024);
* if (got < 0)
* break;
* if (got == 0) {
* virStreamFinish(st);
* break;
* }
* int offset = 0;
* while (offset < got) {
* int sent = write(fd, buf+offset, got-offset)
* if (sent < 0) {
* virStreamAbort(st);
* goto done;
* }
* offset += sent;
* }
* }
* if (virStreamFinish(st) < 0)
* ... report an error ....
* done:
* virStreamFree(st);
* close(fd);
*
*
* Returns the number of bytes read, which may be less
* than requested.
*
* Returns 0 when the end of the stream is reached, at
* which time the caller should invoke virStreamFinish()
* to get confirmation of stream completion.
*
* Returns -1 upon error, at which time the stream will
* be marked as aborted, and the caller should now release
* the stream with virStreamFree.
*
* Returns -2 if there is no data pending to be read & the
* stream is marked as non-blocking.
*/
int virStreamRecv(virStreamPtr stream,
char *data,
size_t nbytes)
{
DEBUG("stream=%p, data=%p, nbytes=%zi", stream, data, nbytes);
virResetLastError();
if (!VIR_IS_CONNECTED_STREAM(stream)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (stream->driver &&
stream->driver->streamRecv) {
int ret;
ret = (stream->driver->streamRecv)(stream, data, nbytes);
if (ret == -2)
return -2;
if (ret < 0)
goto error;
return ret;
}
virLibConnError(stream->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(stream->conn);
return -1;
}
/**
* virStreamSendAll:
* @stream: pointer to the stream object
* @handler: source callback for reading data from application
* @opaque: application defined data
*
* Send the entire data stream, reading the data from the
* requested data source. This is simply a convenient alternative
* to virStreamSend, for apps that do blocking-I/o.
*
2009-10-26 23:02:46 +00:00
* An example using this with a hypothetical file upload
* API looks like
*
* int mysource(virStreamPtr st, char *buf, int nbytes, void *opaque) {
* int *fd = opaque;
*
* return read(*fd, buf, nbytes);
* }
*
* virStreamPtr st = virStreamNew(conn, 0);
* int fd = open("demo.iso", O_RDONLY)
*
* virConnectUploadFile(conn, st);
* if (virStreamSendAll(st, mysource, &fd) < 0) {
* ...report an error ...
* goto done;
* }
* if (virStreamFinish(st) < 0)
* ...report an error...
* virStreamFree(st);
* close(fd);
*
* Returns 0 if all the data was successfully sent. The caller
* should invoke virStreamFinish(st) to flush the stream upon
* success and then virStreamFree
*
* Returns -1 upon any error, with virStreamAbort() already
* having been called, so the caller need only call
* virStreamFree()
*/
int virStreamSendAll(virStreamPtr stream,
virStreamSourceFunc handler,
void *opaque)
{
char *bytes = NULL;
int want = 1024*64;
int ret = -1;
DEBUG("stream=%p, handler=%p, opaque=%p", stream, handler, opaque);
virResetLastError();
if (!VIR_IS_CONNECTED_STREAM(stream)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (stream->flags & VIR_STREAM_NONBLOCK) {
virLibConnError(NULL, VIR_ERR_OPERATION_INVALID,
_("data sources cannot be used for non-blocking streams"));
goto cleanup;
}
if (VIR_ALLOC_N(bytes, want) < 0) {
virReportOOMError();
goto cleanup;
}
for (;;) {
int got, offset = 0;
got = (handler)(stream, bytes, want, opaque);
if (got < 0) {
virStreamAbort(stream);
goto cleanup;
}
if (got == 0)
break;
while (offset < got) {
int done;
done = virStreamSend(stream, bytes + offset, got - offset);
if (done < 0)
goto cleanup;
offset += done;
}
}
ret = 0;
cleanup:
VIR_FREE(bytes);
if (ret != 0)
virDispatchError(stream->conn);
return ret;
}
/**
* virStreamRecvAll:
* @stream: pointer to the stream object
* @handler: sink callback for writing data to application
* @opaque: application defined data
*
* Receive the entire data stream, sending the data to the
* requested data sink. This is simply a convenient alternative
* to virStreamRecv, for apps that do blocking-I/o.
*
2009-10-26 23:02:46 +00:00
* An example using this with a hypothetical file download
* API looks like
*
* int mysink(virStreamPtr st, const char *buf, int nbytes, void *opaque) {
* int *fd = opaque;
*
* return write(*fd, buf, nbytes);
* }
*
* virStreamPtr st = virStreamNew(conn, 0);
* int fd = open("demo.iso", O_WRONLY)
*
* virConnectUploadFile(conn, st);
* if (virStreamRecvAll(st, mysink, &fd) < 0) {
* ...report an error ...
* goto done;
* }
* if (virStreamFinish(st) < 0)
* ...report an error...
* virStreamFree(st);
* close(fd);
*
* Returns 0 if all the data was successfully received. The caller
* should invoke virStreamFinish(st) to flush the stream upon
* success and then virStreamFree
*
* Returns -1 upon any error, with virStreamAbort() already
* having been called, so the caller need only call
* virStreamFree()
*/
int virStreamRecvAll(virStreamPtr stream,
virStreamSinkFunc handler,
void *opaque)
{
char *bytes = NULL;
int want = 1024*64;
int ret = -1;
DEBUG("stream=%p, handler=%p, opaque=%p", stream, handler, opaque);
virResetLastError();
if (!VIR_IS_CONNECTED_STREAM(stream)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (stream->flags & VIR_STREAM_NONBLOCK) {
virLibConnError(NULL, VIR_ERR_OPERATION_INVALID,
_("data sinks cannot be used for non-blocking streams"));
goto cleanup;
}
if (VIR_ALLOC_N(bytes, want) < 0) {
virReportOOMError();
goto cleanup;
}
for (;;) {
int got, offset = 0;
got = virStreamRecv(stream, bytes, want);
if (got < 0)
goto cleanup;
if (got == 0)
break;
while (offset < got) {
int done;
done = (handler)(stream, bytes + offset, got - offset, opaque);
if (done < 0) {
virStreamAbort(stream);
goto cleanup;
}
offset += done;
}
}
ret = 0;
cleanup:
VIR_FREE(bytes);
if (ret != 0)
virDispatchError(stream->conn);
return ret;
}
/**
* virStreamEventAddCallback:
* @stream: pointer to the stream object
* @events: set of events to monitor
* @cb: callback to invoke when an event occurs
* @opaque: application defined data
* @ff: callback to free @opaque data
*
* Register a callback to be notified when a stream
* becomes writable, or readable. This is most commonly
* used in conjunction with non-blocking data streams
* to integrate into an event loop
*
* Returns 0 on success, -1 upon error
*/
int virStreamEventAddCallback(virStreamPtr stream,
int events,
virStreamEventCallback cb,
void *opaque,
virFreeCallback ff)
{
DEBUG("stream=%p, events=%d, cb=%p, opaque=%p, ff=%p", stream, events, cb, opaque, ff);
virResetLastError();
if (!VIR_IS_CONNECTED_STREAM(stream)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (stream->driver &&
stream->driver->streamAddCallback) {
int ret;
ret = (stream->driver->streamAddCallback)(stream, events, cb, opaque, ff);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(stream->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(stream->conn);
return -1;
}
/**
* virStreamEventUpdateCallback:
* @stream: pointer to the stream object
* @events: set of events to monitor
*
* Changes the set of events to monitor for a stream. This allows
* for event notification to be changed without having to
* unregister & register the callback completely. This method
* is guarenteed to succeed if a callback is already registered
*
* Returns 0 on success, -1 if no callback is registered
*/
int virStreamEventUpdateCallback(virStreamPtr stream,
int events)
{
DEBUG("stream=%p, events=%d", stream, events);
virResetLastError();
if (!VIR_IS_CONNECTED_STREAM(stream)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (stream->driver &&
stream->driver->streamUpdateCallback) {
int ret;
ret = (stream->driver->streamUpdateCallback)(stream, events);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (stream->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(stream->conn);
return -1;
}
/**
* virStreamEventRemoveCallback:
* @stream: pointer to the stream object
*
2009-10-26 23:02:46 +00:00
* Remove an event callback from the stream
*
* Returns 0 on success, -1 on error
*/
int virStreamEventRemoveCallback(virStreamPtr stream)
{
DEBUG("stream=%p", stream);
virResetLastError();
if (!VIR_IS_CONNECTED_STREAM(stream)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (stream->driver &&
stream->driver->streamRemoveCallback) {
int ret;
ret = (stream->driver->streamRemoveCallback)(stream);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (stream->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(stream->conn);
return -1;
}
/**
* virStreamFinish:
* @stream: pointer to the stream object
*
* Indicate that there is no further data is to be transmitted
* on the stream. For output streams this should be called once
* all data has been written. For input streams this should be
* called once virStreamRecv returns end-of-file.
*
* This method is a synchronization point for all asynchronous
* errors, so if this returns a success code the application can
* be sure that all data has been successfully processed.
*
* Returns 0 on success, -1 upon error
*/
int virStreamFinish(virStreamPtr stream)
{
DEBUG("stream=%p", stream);
virResetLastError();
if (!VIR_IS_CONNECTED_STREAM(stream)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (stream->driver &&
stream->driver->streamFinish) {
int ret;
ret = (stream->driver->streamFinish)(stream);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (stream->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(stream->conn);
return -1;
}
/**
* virStreamAbort:
* @stream: pointer to the stream object
*
* Request that the in progress data transfer be cancelled
* abnormally before the end of the stream has been reached.
* For output streams this can be used to inform the driver
* that the stream is being terminated early. For input
* streams this can be used to inform the driver that it
* should stop sending data.
*
* Returns 0 on success, -1 upon error
*/
int virStreamAbort(virStreamPtr stream)
{
DEBUG("stream=%p", stream);
virResetLastError();
if (!VIR_IS_CONNECTED_STREAM(stream)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (stream->driver &&
stream->driver->streamAbort) {
int ret;
ret = (stream->driver->streamAbort)(stream);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (stream->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(stream->conn);
return -1;
}
/**
* virStreamFree:
* @stream: pointer to the stream object
*
* Decrement the reference count on a stream, releasing
* the stream object if the reference count has hit zero.
*
2009-10-26 23:02:46 +00:00
* There must not be an active data transfer in progress
* when releasing the stream. If a stream needs to be
* disposed of prior to end of stream being reached, then
* the virStreamAbort function should be called first.
*
* Returns 0 upon success, or -1 on error
*/
int virStreamFree(virStreamPtr stream)
{
DEBUG("stream=%p", stream);
virResetLastError();
if (!VIR_IS_CONNECTED_STREAM(stream)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
/* XXX Enforce shutdown before free'ing resources ? */
if (virUnrefStream(stream) < 0) {
virDispatchError(NULL);
return (-1);
}
return (0);
}
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
/**
* virDomainIsActive:
* @dom: pointer to the domain object
*
* Determine if the domain is currently running
*
* Returns 1 if running, 0 if inactive, -1 on error
*/
int virDomainIsActive(virDomainPtr dom)
{
DEBUG("dom=%p", dom);
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(dom)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
return (-1);
}
if (dom->conn->driver->domainIsActive) {
int ret;
ret = dom->conn->driver->domainIsActive(dom);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(dom->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(dom->conn);
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
return -1;
}
/**
* virDomainIsPersistent:
* @dom: pointer to the domain object
*
* Determine if the domain has a persistent configuration
* which means it will still exist after shutting down
*
* Returns 1 if persistent, 0 if transient, -1 on error
*/
int virDomainIsPersistent(virDomainPtr dom)
{
DEBUG("dom=%p", dom);
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(dom)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
return (-1);
}
if (dom->conn->driver->domainIsPersistent) {
int ret;
ret = dom->conn->driver->domainIsPersistent(dom);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(dom->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(dom->conn);
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
return -1;
}
/**
* virNetworkIsActive:
* @net: pointer to the network object
*
* Determine if the network is currently running
*
* Returns 1 if running, 0 if inactive, -1 on error
*/
int virNetworkIsActive(virNetworkPtr net)
{
DEBUG("net=%p", net);
virResetLastError();
if (!VIR_IS_CONNECTED_NETWORK(net)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
return (-1);
}
if (net->conn->networkDriver->networkIsActive) {
int ret;
ret = net->conn->networkDriver->networkIsActive(net);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(net->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(net->conn);
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
return -1;
}
/**
* virNetworkIsPersistent:
* @net: pointer to the network object
*
* Determine if the network has a persistent configuration
* which means it will still exist after shutting down
*
* Returns 1 if persistent, 0 if transient, -1 on error
*/
int virNetworkIsPersistent(virNetworkPtr net)
{
DEBUG("net=%p", net);
virResetLastError();
if (!VIR_IS_CONNECTED_NETWORK(net)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
return (-1);
}
if (net->conn->networkDriver->networkIsPersistent) {
int ret;
ret = net->conn->networkDriver->networkIsPersistent(net);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(net->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(net->conn);
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
return -1;
}
/**
* virStoragePoolIsActive:
* @pool: pointer to the storage pool object
*
* Determine if the storage pool is currently running
*
* Returns 1 if running, 0 if inactive, -1 on error
*/
int virStoragePoolIsActive(virStoragePoolPtr pool)
{
DEBUG("pool=%p", pool);
virResetLastError();
if (!VIR_IS_CONNECTED_STORAGE_POOL(pool)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
return (-1);
}
if (pool->conn->storageDriver->poolIsActive) {
int ret;
ret = pool->conn->storageDriver->poolIsActive(pool);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(pool->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(pool->conn);
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
return -1;
}
/**
* virStoragePoolIsPersistent:
* @pool: pointer to the storage pool object
*
* Determine if the storage pool has a persistent configuration
* which means it will still exist after shutting down
*
* Returns 1 if persistent, 0 if transient, -1 on error
*/
int virStoragePoolIsPersistent(virStoragePoolPtr pool)
{
DEBUG("pool=%p", pool);
virResetLastError();
if (!VIR_IS_CONNECTED_STORAGE_POOL(pool)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
return (-1);
}
if (pool->conn->storageDriver->poolIsPersistent) {
int ret;
ret = pool->conn->storageDriver->poolIsPersistent(pool);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(pool->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(pool->conn);
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
return -1;
}
/**
* virConnectNumOfNWFilters:
* @conn: pointer to the hypervisor connection
*
* Provides the number of nwfilters.
*
* Returns the number of nwfilters found or -1 in case of error
*/
int
virConnectNumOfNWFilters(virConnectPtr conn)
{
DEBUG("conn=%p", conn);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
if (conn->nwfilterDriver && conn->nwfilterDriver->numOfNWFilters) {
int ret;
ret = conn->nwfilterDriver->numOfNWFilters (conn);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virConnectListNWFilters:
* @conn: pointer to the hypervisor connection
* @names: array to collect the list of names of network filters
* @maxnames: size of @names
*
* Collect the list of network filters, and store their names in @names
*
* Returns the number of network filters found or -1 in case of error
*/
int
virConnectListNWFilters(virConnectPtr conn, char **const names, int maxnames)
{
DEBUG("conn=%p, names=%p, maxnames=%d", conn, names, maxnames);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
if ((names == NULL) || (maxnames < 0)) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->nwfilterDriver && conn->nwfilterDriver->listNWFilters) {
int ret;
ret = conn->nwfilterDriver->listNWFilters (conn, names, maxnames);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virNWFilterLookupByName:
* @conn: pointer to the hypervisor connection
* @name: name for the network filter
*
* Try to lookup a network filter on the given hypervisor based on its name.
*
* Returns a new nwfilter object or NULL in case of failure. If the
* network filter cannot be found, then VIR_ERR_NO_NWFILTER error is raised.
*/
virNWFilterPtr
virNWFilterLookupByName(virConnectPtr conn, const char *name)
{
DEBUG("conn=%p, name=%s", conn, name);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (name == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->nwfilterDriver && conn->nwfilterDriver->nwfilterLookupByName) {
virNWFilterPtr ret;
ret = conn->nwfilterDriver->nwfilterLookupByName (conn, name);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return NULL;
}
/**
* virNWFilterLookupByUUID:
* @conn: pointer to the hypervisor connection
* @uuid: the raw UUID for the network filter
*
* Try to lookup a network filter on the given hypervisor based on its UUID.
*
* Returns a new nwfilter object or NULL in case of failure. If the
* nwfdilter cannot be found, then VIR_ERR_NO_NWFILTER error is raised.
*/
virNWFilterPtr
virNWFilterLookupByUUID(virConnectPtr conn, const unsigned char *uuid)
{
DEBUG("conn=%p, uuid=%s", conn, uuid);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (uuid == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->nwfilterDriver && conn->nwfilterDriver->nwfilterLookupByUUID){
virNWFilterPtr ret;
ret = conn->nwfilterDriver->nwfilterLookupByUUID (conn, uuid);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return NULL;
}
/**
* virNWFilterLookupByUUIDString:
* @conn: pointer to the hypervisor connection
* @uuidstr: the string UUID for the nwfilter
*
* Try to lookup an nwfilter on the given hypervisor based on its UUID.
*
* Returns a new nwfilter object or NULL in case of failure. If the
* nwfilter cannot be found, then VIR_ERR_NO_NWFILTER error is raised.
*/
virNWFilterPtr
virNWFilterLookupByUUIDString(virConnectPtr conn, const char *uuidstr)
{
unsigned char uuid[VIR_UUID_BUFLEN];
DEBUG("conn=%p, uuidstr=%s", conn, uuidstr);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (uuidstr == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (virUUIDParse(uuidstr, uuid) < 0) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
return virNWFilterLookupByUUID(conn, &uuid[0]);
error:
virDispatchError(conn);
return NULL;
}
/**
* virNWFilterFree:
* @nwfilter: a nwfilter object
*
* Free the nwfilter object. The running instance is kept alive.
* The data structure is freed and should not be used thereafter.
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virNWFilterFree(virNWFilterPtr nwfilter)
{
DEBUG("nwfilter=%p", nwfilter);
virResetLastError();
if (!VIR_IS_CONNECTED_NWFILTER(nwfilter)) {
virLibNWFilterError(NULL, VIR_ERR_INVALID_NWFILTER, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
if (virUnrefNWFilter(nwfilter) < 0) {
virDispatchError(NULL);
return -1;
}
return 0;
}
/**
* virNWFilterGetName:
* @nwfilter: a nwfilter object
*
* Get the public name for the network filter
*
* Returns a pointer to the name or NULL, the string need not be deallocated
* its lifetime will be the same as the nwfilter object.
*/
const char *
virNWFilterGetName(virNWFilterPtr nwfilter)
{
DEBUG("nwfilter=%p", nwfilter);
virResetLastError();
if (!VIR_IS_NWFILTER(nwfilter)) {
virLibNWFilterError(NULL, VIR_ERR_INVALID_NWFILTER, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
return (nwfilter->name);
}
/**
* virNWFilterGetUUID:
* @nwfilter: a nwfilter object
* @uuid: pointer to a VIR_UUID_BUFLEN bytes array
*
* Get the UUID for a network filter
*
* Returns -1 in case of error, 0 in case of success
*/
int
virNWFilterGetUUID(virNWFilterPtr nwfilter, unsigned char *uuid)
{
DEBUG("nwfilter=%p, uuid=%p", nwfilter, uuid);
virResetLastError();
if (!VIR_IS_NWFILTER(nwfilter)) {
virLibNWFilterError(NULL, VIR_ERR_INVALID_NWFILTER, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
if (uuid == NULL) {
virLibNWFilterError(nwfilter, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
memcpy(uuid, &nwfilter->uuid[0], VIR_UUID_BUFLEN);
return 0;
error:
virDispatchError(nwfilter->conn);
return -1;
}
/**
* virNWFilterGetUUIDString:
* @nwfilter: a nwfilter object
* @buf: pointer to a VIR_UUID_STRING_BUFLEN bytes array
*
* Get the UUID for a network filter as string. For more information about
* UUID see RFC4122.
*
* Returns -1 in case of error, 0 in case of success
*/
int
virNWFilterGetUUIDString(virNWFilterPtr nwfilter, char *buf)
{
unsigned char uuid[VIR_UUID_BUFLEN];
DEBUG("nwfilter=%p, buf=%p", nwfilter, buf);
virResetLastError();
if (!VIR_IS_NWFILTER(nwfilter)) {
virLibNWFilterError(NULL, VIR_ERR_INVALID_NWFILTER, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
if (buf == NULL) {
virLibNWFilterError(nwfilter, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (virNWFilterGetUUID(nwfilter, &uuid[0]))
goto error;
virUUIDFormat(uuid, buf);
return 0;
error:
virDispatchError(nwfilter->conn);
return -1;
}
/**
* virNWFilterDefineXML:
* @conn: pointer to the hypervisor connection
* @xmlDesc: an XML description of the nwfilter
*
* Define a new network filter, based on an XML description
* similar to the one returned by virNWFilterGetXMLDesc()
*
* Returns a new nwfilter object or NULL in case of failure
*/
virNWFilterPtr
virNWFilterDefineXML(virConnectPtr conn, const char *xmlDesc)
{
DEBUG("conn=%p, xmlDesc=%s", conn, xmlDesc);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (xmlDesc == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->flags & VIR_CONNECT_RO) {
virLibConnError(conn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (conn->nwfilterDriver && conn->nwfilterDriver->defineXML) {
virNWFilterPtr ret;
ret = conn->nwfilterDriver->defineXML (conn, xmlDesc, 0);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return NULL;
}
/**
* virNWFilterUndefine:
* @nwfilter: a nwfilter object
*
* Undefine the nwfilter object. This call will not succeed if
* a running VM is referencing the filter. This does not free the
* associated virNWFilterPtr object.
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virNWFilterUndefine(virNWFilterPtr nwfilter)
{
virConnectPtr conn;
DEBUG("nwfilter=%p", nwfilter);
virResetLastError();
if (!VIR_IS_CONNECTED_NWFILTER(nwfilter)) {
virLibNWFilterError(NULL, VIR_ERR_INVALID_NWFILTER, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
conn = nwfilter->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibNWFilterError(nwfilter, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (conn->nwfilterDriver && conn->nwfilterDriver->undefine) {
int ret;
ret = conn->nwfilterDriver->undefine (nwfilter);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(nwfilter->conn);
return -1;
}
/**
* virNWFilterGetXMLDesc:
* @nwfilter: a nwfilter object
* @flags: an OR'ed set of extraction flags, not used yet
*
* Provide an XML description of the network filter. The description may be
* reused later to redefine the network filter with virNWFilterCreateXML().
*
* Returns a 0 terminated UTF-8 encoded XML instance, or NULL in case of error.
* the caller must free() the returned value.
*/
char *
virNWFilterGetXMLDesc(virNWFilterPtr nwfilter, int flags)
{
virConnectPtr conn;
DEBUG("nwfilter=%p, flags=%d", nwfilter, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_NWFILTER(nwfilter)) {
virLibNWFilterError(NULL, VIR_ERR_INVALID_NWFILTER, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
if (flags != 0) {
virLibNWFilterError(nwfilter, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
conn = nwfilter->conn;
if (conn->nwfilterDriver && conn->nwfilterDriver->getXMLDesc) {
char *ret;
ret = conn->nwfilterDriver->getXMLDesc (nwfilter, flags);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(nwfilter->conn);
return NULL;
}
/**
* virNWFilterRef:
* @nwfilter: the nwfilter to hold a reference on
*
* Increment the reference count on the nwfilter. For each
* additional call to this method, there shall be a corresponding
* call to virNWFilterFree to release the reference count, once
* the caller no longer needs the reference to this object.
*
* This method is typically useful for applications where multiple
* threads are using a connection, and it is required that the
* connection remain open until all threads have finished using
* it. ie, each new thread using an nwfilter would increment
* the reference count.
*
* Returns 0 in case of success, -1 in case of failure.
*/
int
virNWFilterRef(virNWFilterPtr nwfilter)
{
if ((!VIR_IS_CONNECTED_NWFILTER(nwfilter))) {
virLibConnError(NULL, VIR_ERR_INVALID_ARG, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
virMutexLock(&nwfilter->conn->lock);
DEBUG("nwfilter=%p refs=%d", nwfilter, nwfilter->refs);
nwfilter->refs++;
virMutexUnlock(&nwfilter->conn->lock);
return 0;
}
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
/**
* virInterfaceIsActive:
* @iface: pointer to the interface object
*
* Determine if the interface is currently running
*
* Returns 1 if running, 0 if inactive, -1 on error
*/
int virInterfaceIsActive(virInterfacePtr iface)
{
DEBUG("iface=%p", iface);
virResetLastError();
if (!VIR_IS_CONNECTED_INTERFACE(iface)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
return (-1);
}
if (iface->conn->interfaceDriver->interfaceIsActive) {
int ret;
ret = iface->conn->interfaceDriver->interfaceIsActive(iface);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(iface->conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(iface->conn);
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
return -1;
}
/**
* virConnectIsEncrypted:
* @conn: pointer to the connection object
*
* Determine if the connection to the hypervisor is encrypted
*
* Returns 1 if encrypted, 0 if not encrypted, -1 on error
*/
int virConnectIsEncrypted(virConnectPtr conn)
{
DEBUG("conn=%p", conn);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
return (-1);
}
if (conn->driver->isEncrypted) {
int ret;
ret = conn->driver->isEncrypted(conn);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
return -1;
}
/**
* virConnectIsSecure:
* @conn: pointer to the connection object
*
* Determine if the connection to the hypervisor is secure
*
* A connection will be classed as secure if it is either
* encrypted, or running over a channel which is not exposed
* to eavesdropping (eg a UNIX domain socket, or pipe)
*
* Returns 1 if secure, 0 if secure, -1 on error
*/
int virConnectIsSecure(virConnectPtr conn)
{
DEBUG("conn=%p", conn);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
return (-1);
}
if (conn->driver->isSecure) {
int ret;
ret = conn->driver->isSecure(conn);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
New APIs for checking some object properties Introduce a number of new APIs to expose some boolean properties of objects, which cannot otherwise reliably determined, nor are aspects of the XML configuration. * virDomainIsActive: Checking virDomainGetID is not reliable since it is not possible to distinguish between error condition and inactive domain for ID of -1. * virDomainIsPersistent: Check whether a persistent config exists for the domain * virNetworkIsActive: Check whether the network is active * virNetworkIsPersistent: Check whether a persistent config exists for the network * virStoragePoolIsActive: Check whether the storage pool is active * virStoragePoolIsPersistent: Check whether a persistent config exists for the storage pool * virInterfaceIsActive: Check whether the host interface is active * virConnectIsSecure: whether the communication channel to the hypervisor is secure * virConnectIsEncrypted: whether any network based commnunication channels are encrypted NB, a channel can be secure, even if not encrypted, eg if it does not involve the network, like a UNIX socket, or pipe. * include/libvirt/libvirt.h.in: Define public API * src/driver.h: Define internal driver API * src/libvirt.c: Implement public API entry point * src/libvirt_public.syms: Export API symbols * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/interface/netcf_driver.c, src/network/bridge_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c: Stub out driver tables
2009-10-21 10:49:05 +00:00
return -1;
}
/**
* virConnectCompareCPU:
* @conn: virConnect connection
* @xmlDesc: XML describing the CPU to compare with host CPU
* @flags: currently unused, pass 0
*
* Compares the given CPU description with the host CPU
*
* Returns comparison result according to enum virCPUCompareResult
*/
int
virConnectCompareCPU(virConnectPtr conn,
const char *xmlDesc,
unsigned int flags)
{
VIR_DEBUG("conn=%p, xmlDesc=%s, flags=%u", conn, xmlDesc, flags);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return VIR_CPU_COMPARE_ERROR;
}
if (xmlDesc == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->driver->cpuCompare) {
int ret;
ret = conn->driver->cpuCompare(conn, xmlDesc, flags);
if (ret == VIR_CPU_COMPARE_ERROR)
goto error;
return ret;
}
virLibConnError(conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return VIR_CPU_COMPARE_ERROR;
}
/**
* virConnectBaselineCPU:
*
* @conn: virConnect connection
* @xmlCPUs: array of XML descriptions of host CPUs
* @ncpus: number of CPUs in xmlCPUs
* @flags: fine-tuning flags, currently unused, pass 0.
*
* Computes the most feature-rich CPU which is compatible with all given
* host CPUs.
*
* Returns XML description of the computed CPU or NULL on error.
*/
char *
virConnectBaselineCPU(virConnectPtr conn,
const char **xmlCPUs,
unsigned int ncpus,
unsigned int flags)
{
unsigned int i;
VIR_DEBUG("conn=%p, xmlCPUs=%p, ncpus=%u, flags=%u",
conn, xmlCPUs, ncpus, flags);
if (xmlCPUs) {
for (i = 0; i < ncpus; i++)
VIR_DEBUG("xmlCPUs[%u]=%s", i, NULLSTR(xmlCPUs[i]));
}
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
if (xmlCPUs == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->driver->cpuBaseline) {
char *cpu;
cpu = conn->driver->cpuBaseline(conn, xmlCPUs, ncpus, flags);
if (!cpu)
goto error;
return cpu;
}
virLibConnError(conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return NULL;
}
/**
* virDomainGetJobInfo:
* @domain: a domain object
* @info: pointer to a virDomainJobInfo structure allocated by the user
*
* Extract information about progress of a background job on a domain.
* Will return an error if the domain is not active.
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virDomainGetJobInfo(virDomainPtr domain, virDomainJobInfoPtr info)
{
virConnectPtr conn;
DEBUG("domain=%p, info=%p", domain, info);
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (info == NULL) {
virLibDomainError(domain, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
memset(info, 0, sizeof(virDomainJobInfo));
conn = domain->conn;
if (conn->driver->domainGetJobInfo) {
int ret;
ret = conn->driver->domainGetJobInfo (domain, info);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(domain->conn);
return -1;
}
/**
* virDomainAbortJob:
* @domain: a domain object
*
* Requests that the current background job be aborted at the
* soonest opportunity. This will block until the job has
* either completed, or aborted.
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virDomainAbortJob(virDomainPtr domain)
{
virConnectPtr conn;
DEBUG("domain=%p", domain);
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
conn = domain->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (conn->driver->domainAbortJob) {
int ret;
ret = conn->driver->domainAbortJob(domain);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virDomainMigrateSetMaxDowntime:
* @domain: a domain object
* @downtime: maximum tolerable downtime for live migration, in milliseconds
* @flags: fine-tuning flags, currently unused, use 0
*
* Sets maximum tolerable time for which the domain is allowed to be paused
* at the end of live migration. It's supposed to be called while the domain is
* being live-migrated as a reaction to migration progress.
*
* Returns 0 in case of success, -1 otherwise.
*/
int
virDomainMigrateSetMaxDowntime(virDomainPtr domain,
unsigned long long downtime,
unsigned int flags)
{
virConnectPtr conn;
DEBUG("domain=%p, downtime=%llu, flags=%u", domain, downtime, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
conn = domain->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibDomainError(domain, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (conn->driver->domainMigrateSetMaxDowntime) {
if (conn->driver->domainMigrateSetMaxDowntime(domain, downtime, flags) < 0)
goto error;
return 0;
}
virLibConnError(conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
Introduce a new public API for domain events The current API for domain events has a number of problems - Only allows for domain lifecycle change events - Does not allow the same callback to be registered multiple times - Does not allow filtering of events to a specific domain This introduces a new more general purpose domain events API typedef enum { VIR_DOMAIN_EVENT_ID_LIFECYCLE = 0, /* virConnectDomainEventCallback */ ...more events later.. } int virConnectDomainEventRegisterAny(virConnectPtr conn, virDomainPtr dom, /* Optional, to filter */ int eventID, virConnectDomainEventGenericCallback cb, void *opaque, virFreeCallback freecb); int virConnectDomainEventDeregisterAny(virConnectPtr conn, int callbackID); Since different event types can received different data in the callback, the API is defined with a generic callback. Specific events will each have a custom signature for their callback. Thus when registering an event it is neccessary to cast the callback to the generic signature eg int myDomainEventCallback(virConnectPtr conn, virDomainPtr dom, int event, int detail, void *opaque) { ... } virConnectDomainEventRegisterAny(conn, NULL, VIR_DOMAIN_EVENT_ID_LIFECYCLE, VIR_DOMAIN_EVENT_CALLBACK(myDomainEventCallback) NULL, NULL); The VIR_DOMAIN_EVENT_CALLBACK() macro simply does a "bad" cast to the generic signature * include/libvirt/libvirt.h.in: Define new APIs for registering domain events * src/driver.h: Internal driver entry points for new events APIs * src/libvirt.c: Wire up public API to driver API for events APIs * src/libvirt_public.syms: Export new APIs * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c, src/xenapi/xenapi_driver.c: Stub out new API entries
2010-03-18 13:01:48 +00:00
error:
virDispatchError(conn);
return -1;
}
/**
* virConnectDomainEventRegisterAny:
* @conn: pointer to the connection
* @dom: pointer to the domain
* @eventID: the event type to receive
* @cb: callback to the function handling domain events
* @opaque: opaque data to pass on to the callback
* @freecb: optional function to deallocate opaque when not used anymore
*
* Adds a callback to receive notifications of arbitrary domain events
* occurring on a domain.
*
* If dom is NULL, then events will be monitored for any domain. If dom
* is non-NULL, then only the specific domain will be monitored
*
* Most types of event have a callback providing a custom set of parameters
* for the event. When registering an event, it is thus neccessary to use
* the VIR_DOMAIN_EVENT_CALLBACK() macro to cast the supplied function pointer
* to match the signature of this method.
*
* The virDomainPtr object handle passed into the callback upon delivery
* of an event is only valid for the duration of execution of the callback.
* If the callback wishes to keep the domain object after the callback
* returns, it shall take a reference to it, by calling virDomainRef.
* The reference can be released once the object is no longer required
* by calling virDomainFree.
*
* The return value from this method is a positive integer identifier
* for the callback. To unregister a callback, this callback ID should
* be passed to the virDomainEventUnregisterAny method
*
* Returns a callback identifier on success, -1 on failure
*/
int
virConnectDomainEventRegisterAny(virConnectPtr conn,
virDomainPtr dom,
int eventID,
virConnectDomainEventGenericCallback cb,
void *opaque,
virFreeCallback freecb)
{
DEBUG("conn=%p dom=%p, eventID=%d, cb=%p, opaque=%p, freecb=%p", conn, dom, eventID, cb, opaque, freecb);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (dom != NULL &&
!(VIR_IS_CONNECTED_DOMAIN(dom) && dom->conn == conn)) {
virLibConnError(conn, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(conn);
return (-1);
}
if (eventID < 0 || eventID >= VIR_DOMAIN_EVENT_ID_LAST || cb == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if ((conn->driver) && (conn->driver->domainEventRegisterAny)) {
int ret;
ret = conn->driver->domainEventRegisterAny(conn, dom, eventID, cb, opaque, freecb);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
Introduce a new public API for domain events The current API for domain events has a number of problems - Only allows for domain lifecycle change events - Does not allow the same callback to be registered multiple times - Does not allow filtering of events to a specific domain This introduces a new more general purpose domain events API typedef enum { VIR_DOMAIN_EVENT_ID_LIFECYCLE = 0, /* virConnectDomainEventCallback */ ...more events later.. } int virConnectDomainEventRegisterAny(virConnectPtr conn, virDomainPtr dom, /* Optional, to filter */ int eventID, virConnectDomainEventGenericCallback cb, void *opaque, virFreeCallback freecb); int virConnectDomainEventDeregisterAny(virConnectPtr conn, int callbackID); Since different event types can received different data in the callback, the API is defined with a generic callback. Specific events will each have a custom signature for their callback. Thus when registering an event it is neccessary to cast the callback to the generic signature eg int myDomainEventCallback(virConnectPtr conn, virDomainPtr dom, int event, int detail, void *opaque) { ... } virConnectDomainEventRegisterAny(conn, NULL, VIR_DOMAIN_EVENT_ID_LIFECYCLE, VIR_DOMAIN_EVENT_CALLBACK(myDomainEventCallback) NULL, NULL); The VIR_DOMAIN_EVENT_CALLBACK() macro simply does a "bad" cast to the generic signature * include/libvirt/libvirt.h.in: Define new APIs for registering domain events * src/driver.h: Internal driver entry points for new events APIs * src/libvirt.c: Wire up public API to driver API for events APIs * src/libvirt_public.syms: Export new APIs * src/esx/esx_driver.c, src/lxc/lxc_driver.c, src/opennebula/one_driver.c, src/openvz/openvz_driver.c, src/phyp/phyp_driver.c, src/qemu/qemu_driver.c, src/remote/remote_driver.c, src/test/test_driver.c, src/uml/uml_driver.c, src/vbox/vbox_tmpl.c, src/xen/xen_driver.c, src/xenapi/xenapi_driver.c: Stub out new API entries
2010-03-18 13:01:48 +00:00
/**
* virConnectDomainEventDeregisterAny:
* @conn: pointer to the connection
* @callbackID: the callback identifier
*
* Removes an event callback. The callbackID parameter should be the
* vaule obtained from a previous virDomainEventRegisterAny method.
*
* Returns 0 on success, -1 on failure
*/
int
virConnectDomainEventDeregisterAny(virConnectPtr conn,
int callbackID)
{
DEBUG("conn=%p, callbackID=%d", conn, callbackID);
virResetLastError();
if (!VIR_IS_CONNECT(conn)) {
virLibConnError(NULL, VIR_ERR_INVALID_CONN, __FUNCTION__);
virDispatchError(NULL);
return (-1);
}
if (callbackID < 0) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if ((conn->driver) && (conn->driver->domainEventDeregisterAny)) {
int ret;
ret = conn->driver->domainEventDeregisterAny(conn, callbackID);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virDomainManagedSave:
* @dom: pointer to the domain
* @flags: optional flags currently unused
*
* This method will suspend a domain and save its memory contents to
* a file on disk. After the call, if successful, the domain is not
* listed as running anymore.
* The difference from virDomainSave() is that libvirt is keeping track of
* the saved state itself, and will reuse it once the domain is being
* restarted (automatically or via an explicit libvirt call).
* As a result any running domain is sure to not have a managed saved image.
*
* Returns 0 in case of success or -1 in case of failure
*/
int virDomainManagedSave(virDomainPtr dom, unsigned int flags)
{
virConnectPtr conn;
VIR_DEBUG("dom=%p, flags=%u", dom, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(dom)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
conn = dom->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibDomainError(dom, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (conn->driver->domainManagedSave) {
int ret;
ret = conn->driver->domainManagedSave(dom, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virDomainHasManagedSaveImage:
* @dom: pointer to the domain
* @flags: optional flags currently unused
*
* Check if a domain has a managed save image as created by
* virDomainManagedSave(). Note that any running domain should not have
* such an image, as it should have been removed on restart.
*
* Returns 0 if no image is present, 1 if an image is present, and
* -1 in case of error
*/
int virDomainHasManagedSaveImage(virDomainPtr dom, unsigned int flags)
{
virConnectPtr conn;
VIR_DEBUG("dom=%p, flags=%u", dom, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(dom)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
conn = dom->conn;
if (conn->driver->domainHasManagedSaveImage) {
int ret;
ret = conn->driver->domainHasManagedSaveImage(dom, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virDomainManagedSaveRemove:
* @dom: pointer to the domain
* @flags: optional flags currently unused
*
2010-04-07 08:30:28 +00:00
* Remove any managed save image for this domain.
*
* Returns 0 in case of success, and -1 in case of error
*/
int virDomainManagedSaveRemove(virDomainPtr dom, unsigned int flags)
{
virConnectPtr conn;
VIR_DEBUG("dom=%p, flags=%u", dom, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(dom)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
conn = dom->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibDomainError(dom, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (conn->driver->domainManagedSaveRemove) {
int ret;
ret = conn->driver->domainManagedSaveRemove(dom, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError(conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virDomainSnapshotCreateXML:
* @domain: a domain object
* @xmlDesc: string containing an XML description of the domain
* @flags: unused flag parameters; callers should pass 0
*
* Creates a new snapshot of a domain based on the snapshot xml
* contained in xmlDesc.
*
* Returns an (opaque) virDomainSnapshotPtr on success, NULL on failure.
*/
virDomainSnapshotPtr
virDomainSnapshotCreateXML(virDomainPtr domain,
const char *xmlDesc,
unsigned int flags)
{
virConnectPtr conn;
DEBUG("domain=%p, xmlDesc=%s, flags=%u", domain, xmlDesc, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return NULL;
}
conn = domain->conn;
if (conn->flags & VIR_CONNECT_RO) {
virLibConnError(conn, VIR_ERR_OPERATION_DENIED, __FUNCTION__);
goto error;
}
if (conn->driver->domainSnapshotCreateXML) {
virDomainSnapshotPtr ret;
ret = conn->driver->domainSnapshotCreateXML(domain, xmlDesc, flags);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return NULL;
}
/**
* virDomainSnapshotGetXMLDesc:
* @snapshot: a domain snapshot object
* @flags: unused flag parameters; callers should pass 0
*
* Provide an XML description of the domain snapshot.
*
* Returns a 0 terminated UTF-8 encoded XML instance, or NULL in case of error.
* the caller must free() the returned value.
*/
char *
virDomainSnapshotGetXMLDesc(virDomainSnapshotPtr snapshot,
unsigned int flags)
{
virConnectPtr conn;
DEBUG("snapshot=%p, flags=%d", snapshot, flags);
virResetLastError();
if (!VIR_IS_DOMAIN_SNAPSHOT(snapshot)) {
virLibDomainSnapshotError(NULL, VIR_ERR_INVALID_DOMAIN_SNAPSHOT,
__FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
conn = snapshot->domain->conn;
if ((conn->flags & VIR_CONNECT_RO) && (flags & VIR_DOMAIN_XML_SECURE)) {
virLibConnError(conn, VIR_ERR_OPERATION_DENIED,
_("virDomainSnapshotGetXMLDesc with secure flag"));
goto error;
}
if (conn->driver->domainSnapshotDumpXML) {
char *ret;
ret = conn->driver->domainSnapshotDumpXML(snapshot, flags);
if (!ret)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return NULL;
}
/**
* virDomainSnapshotNum:
* @domain: a domain object
* @flags: unused flag parameters; callers should pass 0
*
* Provides the number of domain snapshots for this domain..
*
* Returns the number of domain snapshost found or -1 in case of error.
*/
int
virDomainSnapshotNum(virDomainPtr domain, unsigned int flags)
{
virConnectPtr conn;
DEBUG("domain=%p", domain);
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
conn = domain->conn;
if (conn->driver->domainSnapshotNum) {
int ret = conn->driver->domainSnapshotNum(domain, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virDomainSnapshotListNames:
* @domain: a domain object
* @names: array to collect the list of names of snapshots
* @nameslen: size of @names
* @flags: unused flag parameters; callers should pass 0
*
* Collect the list of domain snapshots for the given domain, and store
* their names in @names. Caller is responsible for freeing each member
* of the array.
*
* Returns the number of domain snapshots found or -1 in case of error.
*/
int
virDomainSnapshotListNames(virDomainPtr domain, char **names, int nameslen,
unsigned int flags)
{
virConnectPtr conn;
DEBUG("domain=%p, names=%p, nameslen=%d, flags=%u",
domain, names, nameslen, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
conn = domain->conn;
if ((names == NULL) || (nameslen < 0)) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->driver->domainSnapshotListNames) {
int ret = conn->driver->domainSnapshotListNames(domain, names,
nameslen, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virDomainSnapshotLookupByName:
* @domain: a domain object
* @name: name for the domain snapshot
* @flags: unused flag parameters; callers should pass 0
*
* Try to lookup a domain snapshot based on its name.
*
* Returns a domain snapshot object or NULL in case of failure. If the
* domain snapshot cannot be found, then the VIR_ERR_NO_DOMAIN_SNAPSHOT
* error is raised.
*/
virDomainSnapshotPtr
virDomainSnapshotLookupByName(virDomainPtr domain,
const char *name,
unsigned int flags)
{
virConnectPtr conn;
DEBUG("domain=%p, name=%s, flags=%u", domain, name, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
conn = domain->conn;
if (name == NULL) {
virLibConnError(conn, VIR_ERR_INVALID_ARG, __FUNCTION__);
goto error;
}
if (conn->driver->domainSnapshotLookupByName) {
virDomainSnapshotPtr dom;
dom = conn->driver->domainSnapshotLookupByName(domain, name, flags);
if (!dom)
goto error;
return dom;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return NULL;
}
/**
* virDomainHasCurrentSnapshot:
* @domain: pointer to the domain object
* @flags: unused flag parameters; callers should pass 0
*
* Determine if the domain has a current snapshot.
*
* Returns 1 if such snapshot exists, 0 if it doesn't, -1 on error.
*/
int
virDomainHasCurrentSnapshot(virDomainPtr domain, unsigned int flags)
{
virConnectPtr conn;
DEBUG("domain=%p, flags=%u", domain, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return -1;
}
conn = domain->conn;
if (conn->driver->domainHasCurrentSnapshot) {
int ret = conn->driver->domainHasCurrentSnapshot(domain, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virDomainSnapshotCurrent:
* @domain: a domain object
* @flags: unused flag parameters; callers should pass 0
*
* Get the current snapshot for a domain, if any.
*
* Returns a domain snapshot object or NULL in case of failure. If the
* current domain snapshot cannot be found, then the VIR_ERR_NO_DOMAIN_SNAPSHOT
* error is raised.
*/
virDomainSnapshotPtr
virDomainSnapshotCurrent(virDomainPtr domain,
unsigned int flags)
{
virConnectPtr conn;
DEBUG("domain=%p, flags=%u", domain, flags);
virResetLastError();
if (!VIR_IS_CONNECTED_DOMAIN(domain)) {
virLibDomainError(NULL, VIR_ERR_INVALID_DOMAIN, __FUNCTION__);
virDispatchError(NULL);
return (NULL);
}
conn = domain->conn;
if (conn->driver->domainSnapshotCurrent) {
virDomainSnapshotPtr snap;
snap = conn->driver->domainSnapshotCurrent(domain, flags);
if (!snap)
goto error;
return snap;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return NULL;
}
/**
* virDomainRevertToSnapshot:
* @snapshot: a domain snapshot object
* @flags: unused flag parameters; callers should pass 0
*
* Revert the domain to a given snapshot.
*
* Returns 0 if the creation is successful, -1 on error.
*/
int
virDomainRevertToSnapshot(virDomainSnapshotPtr snapshot,
unsigned int flags)
{
virConnectPtr conn;
DEBUG("snapshot=%p, flags=%u", snapshot, flags);
virResetLastError();
if (!VIR_IS_DOMAIN_SNAPSHOT(snapshot)) {
virLibDomainSnapshotError(NULL, VIR_ERR_INVALID_DOMAIN_SNAPSHOT,
__FUNCTION__);
virDispatchError(NULL);
return -1;
}
conn = snapshot->domain->conn;
if (conn->driver->domainRevertToSnapshot) {
int ret = conn->driver->domainRevertToSnapshot(snapshot, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virDomainSnapshotDelete:
* @snapshot: a domain snapshot object
* @flags: flag parameters
*
* Delete the snapshot.
*
* If @flags is 0, then just this snapshot is deleted, and changes from
* this snapshot are automatically merged into children snapshots. If
* flags is VIR_DOMAIN_SNAPSHOT_DELETE_CHILDREN, then this snapshot
* and any children snapshots are deleted.
*
* Returns 0 if the snapshot was successfully deleted, -1 on error.
*/
int
virDomainSnapshotDelete(virDomainSnapshotPtr snapshot,
unsigned int flags)
{
virConnectPtr conn;
DEBUG("snapshot=%p, flags=%u", snapshot, flags);
virResetLastError();
if (!VIR_IS_DOMAIN_SNAPSHOT(snapshot)) {
virLibDomainSnapshotError(NULL, VIR_ERR_INVALID_DOMAIN_SNAPSHOT,
__FUNCTION__);
virDispatchError(NULL);
return -1;
}
conn = snapshot->domain->conn;
if (conn->driver->domainSnapshotDelete) {
int ret = conn->driver->domainSnapshotDelete(snapshot, flags);
if (ret < 0)
goto error;
return ret;
}
virLibConnError (conn, VIR_ERR_NO_SUPPORT, __FUNCTION__);
error:
virDispatchError(conn);
return -1;
}
/**
* virDomainSnapshotFree:
* @snapshot: a domain snapshot object
*
* Free the domain snapshot object. The snapshot itself is not modified.
* The data structure is freed and should not be used thereafter.
*
* Returns 0 in case of success and -1 in case of failure.
*/
int
virDomainSnapshotFree(virDomainSnapshotPtr snapshot)
{
DEBUG("snapshot=%p", snapshot);
virResetLastError();
if (!VIR_IS_DOMAIN_SNAPSHOT(snapshot)) {
virLibDomainSnapshotError(NULL, VIR_ERR_INVALID_DOMAIN_SNAPSHOT,
__FUNCTION__);
virDispatchError(NULL);
return -1;
}
if (virUnrefDomainSnapshot(snapshot) < 0) {
virDispatchError(NULL);
return -1;
}
return 0;
}