libvirt/src/qemu/qemu_process.c

3795 lines
116 KiB
C
Raw Normal View History

/*
* qemu_process.h: QEMU process management
*
* Copyright (C) 2006-2011 Red Hat, Inc.
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*
*/
#include <config.h>
#include <fcntl.h>
#include <unistd.h>
#include <signal.h>
#include <sys/stat.h>
#include <sys/time.h>
#include <sys/resource.h>
#include "qemu_process.h"
#include "qemu_domain.h"
#include "qemu_cgroup.h"
#include "qemu_capabilities.h"
#include "qemu_monitor.h"
#include "qemu_command.h"
#include "qemu_hostdev.h"
#include "qemu_hotplug.h"
#include "qemu_bridge_filter.h"
#include "qemu_migration.h"
#if HAVE_NUMACTL
# define NUMA_VERSION1_COMPATIBILITY 1
# include <numa.h>
#endif
#include "datatypes.h"
#include "logging.h"
#include "virterror_internal.h"
#include "memory.h"
#include "hooks.h"
#include "virfile.h"
#include "virpidfile.h"
#include "util.h"
#include "c-ctype.h"
#include "nodeinfo.h"
#include "processinfo.h"
#include "domain_audit.h"
#include "domain_nwfilter.h"
#include "locking/domain_lock.h"
#include "network/bridge_driver.h"
#include "uuid.h"
#define VIR_FROM_THIS VIR_FROM_QEMU
#define START_POSTFIX ": starting up\n"
#define ATTACH_POSTFIX ": attaching\n"
#define SHUTDOWN_POSTFIX ": shutting down\n"
/**
* qemudRemoveDomainStatus
*
* remove all state files of a domain from statedir
*
* Returns 0 on success
*/
static int
qemuProcessRemoveDomainStatus(struct qemud_driver *driver,
virDomainObjPtr vm)
{
char ebuf[1024];
char *file = NULL;
qemuDomainObjPrivatePtr priv = vm->privateData;
if (virAsprintf(&file, "%s/%s.xml", driver->stateDir, vm->def->name) < 0) {
virReportOOMError();
return(-1);
}
if (unlink(file) < 0 && errno != ENOENT && errno != ENOTDIR)
VIR_WARN("Failed to remove domain XML for %s: %s",
vm->def->name, virStrerror(errno, ebuf, sizeof(ebuf)));
VIR_FREE(file);
if (priv->pidfile &&
unlink(priv->pidfile) < 0 &&
errno != ENOENT)
VIR_WARN("Failed to remove PID file for %s: %s",
vm->def->name, virStrerror(errno, ebuf, sizeof(ebuf)));
return 0;
}
/* XXX figure out how to remove this */
extern struct qemud_driver *qemu_driver;
/*
* This is a callback registered with a qemuMonitorPtr instance,
* and to be invoked when the monitor console hits an end of file
* condition, or error, thus indicating VM shutdown should be
* performed
*/
static void
qemuProcessHandleMonitorEOF(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
virDomainObjPtr vm)
{
struct qemud_driver *driver = qemu_driver;
virDomainEventPtr event = NULL;
qemuDomainObjPrivatePtr priv;
int eventReason = VIR_DOMAIN_EVENT_STOPPED_SHUTDOWN;
int stopReason = VIR_DOMAIN_SHUTOFF_SHUTDOWN;
const char *auditReason = "shutdown";
VIR_DEBUG("Received EOF on %p '%s'", vm, vm->def->name);
qemuDriverLock(driver);
virDomainObjLock(vm);
if (!virDomainObjIsActive(vm)) {
VIR_DEBUG("Domain %p is not active, ignoring EOF", vm);
virDomainObjUnlock(vm);
qemuDriverUnlock(driver);
return;
}
priv = vm->privateData;
if (priv->monJSON && !priv->gotShutdown) {
VIR_DEBUG("Monitor connection to '%s' closed without SHUTDOWN event; "
"assuming the domain crashed", vm->def->name);
eventReason = VIR_DOMAIN_EVENT_STOPPED_FAILED;
stopReason = VIR_DOMAIN_SHUTOFF_CRASHED;
auditReason = "failed";
}
event = virDomainEventNewFromObj(vm,
VIR_DOMAIN_EVENT_STOPPED,
eventReason);
qemuProcessStop(driver, vm, 0, stopReason);
virDomainAuditStop(vm, auditReason);
if (!vm->persistent)
qemuDomainRemoveInactive(driver, vm);
else
virDomainObjUnlock(vm);
if (event)
qemuDomainEventQueue(driver, event);
qemuDriverUnlock(driver);
}
/*
* This is invoked when there is some kind of error
* parsing data to/from the monitor. The VM can continue
* to run, but no further monitor commands will be
* allowed
*/
static void
qemuProcessHandleMonitorError(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
virDomainObjPtr vm)
{
struct qemud_driver *driver = qemu_driver;
virDomainEventPtr event = NULL;
VIR_DEBUG("Received error on %p '%s'", vm, vm->def->name);
qemuDriverLock(driver);
virDomainObjLock(vm);
((qemuDomainObjPrivatePtr) vm->privateData)->monError = true;
event = virDomainEventControlErrorNewFromObj(vm);
if (event)
qemuDomainEventQueue(driver, event);
virDomainObjUnlock(vm);
qemuDriverUnlock(driver);
}
static virDomainDiskDefPtr
qemuProcessFindDomainDiskByPath(virDomainObjPtr vm,
const char *path)
{
snapshot: also support disks by path I got confused when 'virsh domblkinfo dom disk' required the path to a disk (which can be ambiguous, since a single file can back multiple disks), rather than the unambiguous target device name that I was using in disk snapshots. So, in true developer fashion, I went for the best of both worlds - all interfaces that operate on a disk (aka block) now accept either the target name or the unambiguous path to the backing file used by the disk. * src/conf/domain_conf.h (virDomainDiskIndexByName): Add parameter. (virDomainDiskPathByName): New prototype. * src/libvirt_private.syms (domain_conf.h): Export it. * src/conf/domain_conf.c (virDomainDiskIndexByName): Also allow searching by path, and decide whether ambiguity is okay. (virDomainDiskPathByName): New function. (virDomainDiskRemoveByName, virDomainSnapshotAlignDisks): Update callers. * src/qemu/qemu_driver.c (qemudDomainBlockPeek) (qemuDomainAttachDeviceConfig, qemuDomainUpdateDeviceConfig) (qemuDomainGetBlockInfo, qemuDiskPathToAlias): Likewise. * src/qemu/qemu_process.c (qemuProcessFindDomainDiskByPath): Likewise. * src/libxl/libxl_driver.c (libxlDomainAttachDeviceDiskLive) (libxlDomainDetachDeviceDiskLive, libxlDomainAttachDeviceConfig) (libxlDomainUpdateDeviceConfig): Likewise. * src/uml/uml_driver.c (umlDomainBlockPeek): Likewise. * src/xen/xend_internal.c (xenDaemonDomainBlockPeek): Likewise. * docs/formatsnapshot.html.in: Update documentation. * tools/virsh.pod (domblkstat, domblkinfo): Likewise. * docs/schemas/domaincommon.rng (diskTarget): Tighten pattern on disk targets. * docs/schemas/domainsnapshot.rng (disksnapshot): Update to match. * tests/domainsnapshotxml2xmlin/disk_snapshot.xml: Update test.
2011-08-20 02:38:36 +00:00
int i = virDomainDiskIndexByName(vm->def, path, true);
snapshot: also support disks by path I got confused when 'virsh domblkinfo dom disk' required the path to a disk (which can be ambiguous, since a single file can back multiple disks), rather than the unambiguous target device name that I was using in disk snapshots. So, in true developer fashion, I went for the best of both worlds - all interfaces that operate on a disk (aka block) now accept either the target name or the unambiguous path to the backing file used by the disk. * src/conf/domain_conf.h (virDomainDiskIndexByName): Add parameter. (virDomainDiskPathByName): New prototype. * src/libvirt_private.syms (domain_conf.h): Export it. * src/conf/domain_conf.c (virDomainDiskIndexByName): Also allow searching by path, and decide whether ambiguity is okay. (virDomainDiskPathByName): New function. (virDomainDiskRemoveByName, virDomainSnapshotAlignDisks): Update callers. * src/qemu/qemu_driver.c (qemudDomainBlockPeek) (qemuDomainAttachDeviceConfig, qemuDomainUpdateDeviceConfig) (qemuDomainGetBlockInfo, qemuDiskPathToAlias): Likewise. * src/qemu/qemu_process.c (qemuProcessFindDomainDiskByPath): Likewise. * src/libxl/libxl_driver.c (libxlDomainAttachDeviceDiskLive) (libxlDomainDetachDeviceDiskLive, libxlDomainAttachDeviceConfig) (libxlDomainUpdateDeviceConfig): Likewise. * src/uml/uml_driver.c (umlDomainBlockPeek): Likewise. * src/xen/xend_internal.c (xenDaemonDomainBlockPeek): Likewise. * docs/formatsnapshot.html.in: Update documentation. * tools/virsh.pod (domblkstat, domblkinfo): Likewise. * docs/schemas/domaincommon.rng (diskTarget): Tighten pattern on disk targets. * docs/schemas/domainsnapshot.rng (disksnapshot): Update to match. * tests/domainsnapshotxml2xmlin/disk_snapshot.xml: Update test.
2011-08-20 02:38:36 +00:00
if (i >= 0)
return vm->def->disks[i];
qemuReportError(VIR_ERR_INTERNAL_ERROR,
_("no disk found with path %s"),
path);
return NULL;
}
static virDomainDiskDefPtr
qemuProcessFindDomainDiskByAlias(virDomainObjPtr vm,
const char *alias)
{
int i;
if (STRPREFIX(alias, QEMU_DRIVE_HOST_PREFIX))
alias += strlen(QEMU_DRIVE_HOST_PREFIX);
for (i = 0; i < vm->def->ndisks; i++) {
virDomainDiskDefPtr disk;
disk = vm->def->disks[i];
if (disk->info.alias != NULL && STREQ(disk->info.alias, alias))
return disk;
}
qemuReportError(VIR_ERR_INTERNAL_ERROR,
_("no disk found with alias %s"),
alias);
return NULL;
}
static int
qemuProcessGetVolumeQcowPassphrase(virConnectPtr conn,
virDomainDiskDefPtr disk,
char **secretRet,
size_t *secretLen)
{
virSecretPtr secret;
char *passphrase;
unsigned char *data;
size_t size;
int ret = -1;
virStorageEncryptionPtr enc;
if (!disk->encryption) {
qemuReportError(VIR_ERR_INTERNAL_ERROR,
_("disk %s does not have any encryption information"),
disk->src);
return -1;
}
enc = disk->encryption;
if (!conn) {
qemuReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("cannot find secrets without a connection"));
goto cleanup;
}
if (conn->secretDriver == NULL ||
conn->secretDriver->lookupByUUID == NULL ||
conn->secretDriver->getValue == NULL) {
qemuReportError(VIR_ERR_OPERATION_INVALID, "%s",
_("secret storage not supported"));
goto cleanup;
}
if (enc->format != VIR_STORAGE_ENCRYPTION_FORMAT_QCOW ||
enc->nsecrets != 1 ||
enc->secrets[0]->type !=
VIR_STORAGE_ENCRYPTION_SECRET_TYPE_PASSPHRASE) {
qemuReportError(VIR_ERR_XML_ERROR,
_("invalid <encryption> for volume %s"), disk->src);
goto cleanup;
}
secret = conn->secretDriver->lookupByUUID(conn,
enc->secrets[0]->uuid);
if (secret == NULL)
goto cleanup;
libvirt: do not mix internal flags into public API There were two API in driver.c that were silently masking flags bits prior to calling out to the drivers, and several others that were explicitly masking flags bits. This is not forward-compatible - if we ever have that many flags in the future, then talking to an old server that masks out the flags would be indistinguishable from talking to a new server that can honor the flag. In general, libvirt.c should forward _all_ flags on to drivers, and only the drivers should reject unknown flags. In the case of virDrvSecretGetValue, the solution is to separate the internal driver callback function to have two parameters instead of one, with only one parameter affected by the public API. In the case of virDomainGetXMLDesc, it turns out that no one was ever mixing VIR_DOMAIN_XML_INTERNAL_STATUS with the dumpxml path in the first place; that internal flag was only used in saving and restoring state files, which happened to be in functions internal to a single file, so there is no mixing of the internal flag with a public flags argument. Additionally, virDomainMemoryStats passed a flags argument over RPC, but not to the driver. * src/driver.h (VIR_DOMAIN_XML_FLAGS_MASK) (VIR_SECRET_GET_VALUE_FLAGS_MASK): Delete. (virDrvSecretGetValue): Separate out internal flags. (virDrvDomainMemoryStats): Provide missing flags argument. * src/driver.c (verify): Drop unused check. * src/conf/domain_conf.h (virDomainObjParseFile): Delete declaration. (virDomainXMLInternalFlags): Move... * src/conf/domain_conf.c: ...here. Delete redundant include. (virDomainObjParseFile): Make static. * src/libvirt.c (virDomainGetXMLDesc, virSecretGetValue): Update clients. (virDomainMemoryPeek, virInterfaceGetXMLDesc) (virDomainMemoryStats, virDomainBlockPeek, virNetworkGetXMLDesc) (virStoragePoolGetXMLDesc, virStorageVolGetXMLDesc) (virNodeNumOfDevices, virNodeListDevices, virNWFilterGetXMLDesc): Don't mask unknown flags. * src/interface/netcf_driver.c (interfaceGetXMLDesc): Reject unknown flags. * src/secret/secret_driver.c (secretGetValue): Update clients. * src/remote/remote_driver.c (remoteSecretGetValue) (remoteDomainMemoryStats): Likewise. * src/qemu/qemu_process.c (qemuProcessGetVolumeQcowPassphrase): Likewise. * src/qemu/qemu_driver.c (qemudDomainMemoryStats): Likewise. * daemon/remote.c (remoteDispatchDomainMemoryStats): Likewise.
2011-07-13 21:31:56 +00:00
data = conn->secretDriver->getValue(secret, &size, 0,
VIR_SECRET_GET_VALUE_INTERNAL_CALL);
virUnrefSecret(secret);
if (data == NULL)
goto cleanup;
if (memchr(data, '\0', size) != NULL) {
memset(data, 0, size);
VIR_FREE(data);
qemuReportError(VIR_ERR_XML_ERROR,
_("format='qcow' passphrase for %s must not contain a "
"'\\0'"), disk->src);
goto cleanup;
}
if (VIR_ALLOC_N(passphrase, size + 1) < 0) {
memset(data, 0, size);
VIR_FREE(data);
virReportOOMError();
goto cleanup;
}
memcpy(passphrase, data, size);
passphrase[size] = '\0';
memset(data, 0, size);
VIR_FREE(data);
*secretRet = passphrase;
*secretLen = size;
ret = 0;
cleanup:
return ret;
}
static int
qemuProcessFindVolumeQcowPassphrase(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
virConnectPtr conn,
virDomainObjPtr vm,
const char *path,
char **secretRet,
size_t *secretLen)
{
virDomainDiskDefPtr disk;
int ret = -1;
virDomainObjLock(vm);
disk = qemuProcessFindDomainDiskByPath(vm, path);
if (!disk)
goto cleanup;
ret = qemuProcessGetVolumeQcowPassphrase(conn, disk, secretRet, secretLen);
cleanup:
virDomainObjUnlock(vm);
return ret;
}
static int
qemuProcessHandleReset(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
virDomainObjPtr vm)
{
struct qemud_driver *driver = qemu_driver;
virDomainEventPtr event;
virDomainObjLock(vm);
event = virDomainEventRebootNewFromObj(vm);
virDomainObjUnlock(vm);
if (event) {
qemuDriverLock(driver);
qemuDomainEventQueue(driver, event);
qemuDriverUnlock(driver);
}
return 0;
}
/*
* Since we have the '-no-shutdown' flag set, the
* QEMU process will currently have guest OS shutdown
* and the CPUS stopped. To fake the reboot, we thus
* want todo a reset of the virtual hardware, followed
* by restart of the CPUs. This should result in the
* guest OS booting up again
*/
static void
qemuProcessFakeReboot(void *opaque)
{
struct qemud_driver *driver = qemu_driver;
virDomainObjPtr vm = opaque;
qemuDomainObjPrivatePtr priv = vm->privateData;
virDomainEventPtr event = NULL;
int ret = -1;
VIR_DEBUG("vm=%p", vm);
qemuDriverLock(driver);
virDomainObjLock(vm);
if (qemuDomainObjBeginJob(driver, vm, QEMU_JOB_MODIFY) < 0)
goto cleanup;
if (!virDomainObjIsActive(vm)) {
qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("guest unexpectedly quit"));
goto endjob;
}
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
qemuDomainObjEnterMonitorWithDriver(driver, vm);
if (qemuMonitorSystemReset(priv->mon) < 0) {
qemuDomainObjExitMonitorWithDriver(driver, vm);
goto endjob;
}
qemuDomainObjExitMonitorWithDriver(driver, vm);
if (!virDomainObjIsActive(vm)) {
qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("guest unexpectedly quit"));
goto endjob;
}
if (qemuProcessStartCPUs(driver, vm, NULL,
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
VIR_DOMAIN_RUNNING_BOOTED,
QEMU_ASYNC_JOB_NONE) < 0) {
if (virGetLastError() == NULL)
qemuReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("resume operation failed"));
goto endjob;
}
priv->gotShutdown = false;
event = virDomainEventNewFromObj(vm,
VIR_DOMAIN_EVENT_RESUMED,
VIR_DOMAIN_EVENT_RESUMED_UNPAUSED);
ret = 0;
endjob:
if (qemuDomainObjEndJob(driver, vm) == 0)
vm = NULL;
cleanup:
if (vm) {
if (ret == -1)
qemuProcessKill(vm, false);
if (virDomainObjUnref(vm) > 0)
virDomainObjUnlock(vm);
}
if (event)
qemuDomainEventQueue(driver, event);
qemuDriverUnlock(driver);
}
static void
qemuProcessShutdownOrReboot(virDomainObjPtr vm)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
if (priv->gotShutdown) {
VIR_DEBUG("Ignoring repeated SHUTDOWN event from domain %s",
vm->def->name);
return;
}
priv->gotShutdown = true;
if (priv->fakeReboot) {
qemuDomainSetFakeReboot(qemu_driver, vm, false);
virDomainObjRef(vm);
virThread th;
if (virThreadCreate(&th,
false,
qemuProcessFakeReboot,
vm) < 0) {
VIR_ERROR(_("Failed to create reboot thread, killing domain"));
qemuProcessKill(vm, true);
/* Safe to ignore value since ref count was incremented above */
ignore_value(virDomainObjUnref(vm));
}
} else {
qemuProcessKill(vm, true);
}
}
static int
qemuProcessHandleShutdown(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
virDomainObjPtr vm)
{
VIR_DEBUG("vm=%p", vm);
virDomainObjLock(vm);
qemuProcessShutdownOrReboot(vm);
virDomainObjUnlock(vm);
return 0;
}
static int
qemuProcessHandleStop(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
virDomainObjPtr vm)
{
struct qemud_driver *driver = qemu_driver;
virDomainEventPtr event = NULL;
virDomainPausedReason reason = VIR_DOMAIN_PAUSED_UNKNOWN;
virDomainObjLock(vm);
if (virDomainObjGetState(vm, NULL) == VIR_DOMAIN_RUNNING) {
qemuDomainObjPrivatePtr priv = vm->privateData;
if (priv->gotShutdown) {
VIR_DEBUG("Got STOP event after SHUTDOWN, assuming we are stopping"
" for shutdown");
reason = VIR_DOMAIN_PAUSED_SHUTTING_DOWN;
}
VIR_DEBUG("Transitioned guest %s to paused state, reason=%s",
vm->def->name, virDomainPausedReasonTypeToString(reason));
virDomainObjSetState(vm, VIR_DOMAIN_PAUSED, reason);
event = virDomainEventNewFromObj(vm,
VIR_DOMAIN_EVENT_SUSPENDED,
VIR_DOMAIN_EVENT_SUSPENDED_PAUSED);
VIR_FREE(priv->lockState);
if (virDomainLockProcessPause(driver->lockManager, vm, &priv->lockState) < 0)
VIR_WARN("Unable to release lease on %s", vm->def->name);
VIR_DEBUG("Preserving lock state '%s'", NULLSTR(priv->lockState));
if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0) {
VIR_WARN("Unable to save status on vm %s after state change",
vm->def->name);
}
}
virDomainObjUnlock(vm);
if (event) {
qemuDriverLock(driver);
if (event)
qemuDomainEventQueue(driver, event);
qemuDriverUnlock(driver);
}
return 0;
}
static int
qemuProcessHandleRTCChange(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
virDomainObjPtr vm,
long long offset)
{
struct qemud_driver *driver = qemu_driver;
virDomainEventPtr event;
virDomainObjLock(vm);
event = virDomainEventRTCChangeNewFromObj(vm, offset);
if (vm->def->clock.offset == VIR_DOMAIN_CLOCK_OFFSET_VARIABLE)
vm->def->clock.data.adjustment = offset;
if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0)
VIR_WARN("unable to save domain status with RTC change");
virDomainObjUnlock(vm);
if (event) {
qemuDriverLock(driver);
qemuDomainEventQueue(driver, event);
qemuDriverUnlock(driver);
}
return 0;
}
static int
qemuProcessHandleWatchdog(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
virDomainObjPtr vm,
int action)
{
struct qemud_driver *driver = qemu_driver;
virDomainEventPtr watchdogEvent = NULL;
virDomainEventPtr lifecycleEvent = NULL;
virDomainObjLock(vm);
watchdogEvent = virDomainEventWatchdogNewFromObj(vm, action);
if (action == VIR_DOMAIN_EVENT_WATCHDOG_PAUSE &&
virDomainObjGetState(vm, NULL) == VIR_DOMAIN_RUNNING) {
qemuDomainObjPrivatePtr priv = vm->privateData;
VIR_DEBUG("Transitioned guest %s to paused state due to watchdog", vm->def->name);
virDomainObjSetState(vm, VIR_DOMAIN_PAUSED, VIR_DOMAIN_PAUSED_WATCHDOG);
lifecycleEvent = virDomainEventNewFromObj(vm,
VIR_DOMAIN_EVENT_SUSPENDED,
VIR_DOMAIN_EVENT_SUSPENDED_WATCHDOG);
VIR_FREE(priv->lockState);
if (virDomainLockProcessPause(driver->lockManager, vm, &priv->lockState) < 0)
VIR_WARN("Unable to release lease on %s", vm->def->name);
VIR_DEBUG("Preserving lock state '%s'", NULLSTR(priv->lockState));
if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0) {
VIR_WARN("Unable to save status on vm %s after watchdog event",
vm->def->name);
}
}
if (vm->def->watchdog->action == VIR_DOMAIN_WATCHDOG_ACTION_DUMP) {
struct qemuDomainWatchdogEvent *wdEvent;
if (VIR_ALLOC(wdEvent) == 0) {
wdEvent->action = VIR_DOMAIN_WATCHDOG_ACTION_DUMP;
wdEvent->vm = vm;
/* Hold an extra reference because we can't allow 'vm' to be
* deleted before handling watchdog event is finished.
*/
virDomainObjRef(vm);
if (virThreadPoolSendJob(driver->workerPool, 0, wdEvent) < 0) {
if (virDomainObjUnref(vm) == 0)
vm = NULL;
VIR_FREE(wdEvent);
}
} else {
virReportOOMError();
}
}
if (vm)
virDomainObjUnlock(vm);
if (watchdogEvent || lifecycleEvent) {
qemuDriverLock(driver);
if (watchdogEvent)
qemuDomainEventQueue(driver, watchdogEvent);
if (lifecycleEvent)
qemuDomainEventQueue(driver, lifecycleEvent);
qemuDriverUnlock(driver);
}
return 0;
}
static int
qemuProcessHandleIOError(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
virDomainObjPtr vm,
const char *diskAlias,
int action,
const char *reason)
{
struct qemud_driver *driver = qemu_driver;
virDomainEventPtr ioErrorEvent = NULL;
virDomainEventPtr ioErrorEvent2 = NULL;
virDomainEventPtr lifecycleEvent = NULL;
const char *srcPath;
const char *devAlias;
virDomainDiskDefPtr disk;
virDomainObjLock(vm);
disk = qemuProcessFindDomainDiskByAlias(vm, diskAlias);
if (disk) {
srcPath = disk->src;
devAlias = disk->info.alias;
} else {
srcPath = "";
devAlias = "";
}
ioErrorEvent = virDomainEventIOErrorNewFromObj(vm, srcPath, devAlias, action);
ioErrorEvent2 = virDomainEventIOErrorReasonNewFromObj(vm, srcPath, devAlias, action, reason);
if (action == VIR_DOMAIN_EVENT_IO_ERROR_PAUSE &&
virDomainObjGetState(vm, NULL) == VIR_DOMAIN_RUNNING) {
qemuDomainObjPrivatePtr priv = vm->privateData;
VIR_DEBUG("Transitioned guest %s to paused state due to IO error", vm->def->name);
virDomainObjSetState(vm, VIR_DOMAIN_PAUSED, VIR_DOMAIN_PAUSED_IOERROR);
lifecycleEvent = virDomainEventNewFromObj(vm,
VIR_DOMAIN_EVENT_SUSPENDED,
VIR_DOMAIN_EVENT_SUSPENDED_IOERROR);
VIR_FREE(priv->lockState);
if (virDomainLockProcessPause(driver->lockManager, vm, &priv->lockState) < 0)
VIR_WARN("Unable to release lease on %s", vm->def->name);
VIR_DEBUG("Preserving lock state '%s'", NULLSTR(priv->lockState));
if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0)
VIR_WARN("Unable to save status on vm %s after IO error", vm->def->name);
}
virDomainObjUnlock(vm);
if (ioErrorEvent || ioErrorEvent2 || lifecycleEvent) {
qemuDriverLock(driver);
if (ioErrorEvent)
qemuDomainEventQueue(driver, ioErrorEvent);
if (ioErrorEvent2)
qemuDomainEventQueue(driver, ioErrorEvent2);
if (lifecycleEvent)
qemuDomainEventQueue(driver, lifecycleEvent);
qemuDriverUnlock(driver);
}
return 0;
}
static int
qemuProcessHandleBlockJob(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
virDomainObjPtr vm,
const char *diskAlias,
int type,
int status)
{
struct qemud_driver *driver = qemu_driver;
virDomainEventPtr event = NULL;
const char *path;
virDomainDiskDefPtr disk;
virDomainObjLock(vm);
disk = qemuProcessFindDomainDiskByAlias(vm, diskAlias);
if (disk) {
path = disk->src;
event = virDomainEventBlockJobNewFromObj(vm, path, type, status);
}
virDomainObjUnlock(vm);
if (event) {
qemuDriverLock(driver);
qemuDomainEventQueue(driver, event);
qemuDriverUnlock(driver);
}
return 0;
}
static int
qemuProcessHandleGraphics(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
virDomainObjPtr vm,
int phase,
int localFamily,
const char *localNode,
const char *localService,
int remoteFamily,
const char *remoteNode,
const char *remoteService,
const char *authScheme,
const char *x509dname,
const char *saslUsername)
{
struct qemud_driver *driver = qemu_driver;
virDomainEventPtr event;
virDomainEventGraphicsAddressPtr localAddr = NULL;
virDomainEventGraphicsAddressPtr remoteAddr = NULL;
virDomainEventGraphicsSubjectPtr subject = NULL;
int i;
if (VIR_ALLOC(localAddr) < 0)
goto no_memory;
localAddr->family = localFamily;
if (!(localAddr->service = strdup(localService)) ||
!(localAddr->node = strdup(localNode)))
goto no_memory;
if (VIR_ALLOC(remoteAddr) < 0)
goto no_memory;
remoteAddr->family = remoteFamily;
if (!(remoteAddr->service = strdup(remoteService)) ||
!(remoteAddr->node = strdup(remoteNode)))
goto no_memory;
if (VIR_ALLOC(subject) < 0)
goto no_memory;
if (x509dname) {
if (VIR_REALLOC_N(subject->identities, subject->nidentity+1) < 0)
goto no_memory;
subject->nidentity++;
if (!(subject->identities[subject->nidentity-1].type = strdup("x509dname")) ||
!(subject->identities[subject->nidentity-1].name = strdup(x509dname)))
goto no_memory;
}
if (saslUsername) {
if (VIR_REALLOC_N(subject->identities, subject->nidentity+1) < 0)
goto no_memory;
subject->nidentity++;
if (!(subject->identities[subject->nidentity-1].type = strdup("saslUsername")) ||
!(subject->identities[subject->nidentity-1].name = strdup(saslUsername)))
goto no_memory;
}
virDomainObjLock(vm);
event = virDomainEventGraphicsNewFromObj(vm, phase, localAddr, remoteAddr, authScheme, subject);
virDomainObjUnlock(vm);
if (event) {
qemuDriverLock(driver);
qemuDomainEventQueue(driver, event);
qemuDriverUnlock(driver);
}
return 0;
no_memory:
virReportOOMError();
if (localAddr) {
VIR_FREE(localAddr->service);
VIR_FREE(localAddr->node);
VIR_FREE(localAddr);
}
if (remoteAddr) {
VIR_FREE(remoteAddr->service);
VIR_FREE(remoteAddr->node);
VIR_FREE(remoteAddr);
}
if (subject) {
for (i = 0 ; i < subject->nidentity ; i++) {
VIR_FREE(subject->identities[i].type);
VIR_FREE(subject->identities[i].name);
}
VIR_FREE(subject->identities);
VIR_FREE(subject);
}
return -1;
}
static void qemuProcessHandleMonitorDestroy(qemuMonitorPtr mon,
virDomainObjPtr vm)
{
qemuDomainObjPrivatePtr priv;
virDomainObjLock(vm);
priv = vm->privateData;
if (priv->mon == mon)
priv->mon = NULL;
if (virDomainObjUnref(vm) > 0)
virDomainObjUnlock(vm);
}
static qemuMonitorCallbacks monitorCallbacks = {
.destroy = qemuProcessHandleMonitorDestroy,
.eofNotify = qemuProcessHandleMonitorEOF,
.errorNotify = qemuProcessHandleMonitorError,
.diskSecretLookup = qemuProcessFindVolumeQcowPassphrase,
.domainShutdown = qemuProcessHandleShutdown,
.domainStop = qemuProcessHandleStop,
.domainReset = qemuProcessHandleReset,
.domainRTCChange = qemuProcessHandleRTCChange,
.domainWatchdog = qemuProcessHandleWatchdog,
.domainIOError = qemuProcessHandleIOError,
.domainGraphics = qemuProcessHandleGraphics,
.domainBlockJob = qemuProcessHandleBlockJob,
};
static int
qemuConnectMonitor(struct qemud_driver *driver, virDomainObjPtr vm)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
int ret = -1;
qemuMonitorPtr mon = NULL;
if (virSecurityManagerSetDaemonSocketLabel(driver->securityManager,
vm) < 0) {
VIR_ERROR(_("Failed to set security context for monitor for %s"),
vm->def->name);
goto error;
}
/* Hold an extra reference because we can't allow 'vm' to be
* deleted while the monitor is active */
virDomainObjRef(vm);
ignore_value(virTimeMs(&priv->monStart));
virDomainObjUnlock(vm);
qemuDriverUnlock(driver);
mon = qemuMonitorOpen(vm,
priv->monConfig,
priv->monJSON,
&monitorCallbacks);
qemuDriverLock(driver);
virDomainObjLock(vm);
priv->monStart = 0;
/* Safe to ignore value since ref count was incremented above */
if (mon == NULL)
ignore_value(virDomainObjUnref(vm));
if (!virDomainObjIsActive(vm)) {
qemuMonitorClose(mon);
goto error;
}
priv->mon = mon;
if (virSecurityManagerClearSocketLabel(driver->securityManager, vm) < 0) {
VIR_ERROR(_("Failed to clear security context for monitor for %s"),
vm->def->name);
goto error;
}
if (priv->mon == NULL) {
VIR_INFO("Failed to connect monitor for %s", vm->def->name);
goto error;
}
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
qemuDomainObjEnterMonitorWithDriver(driver, vm);
ret = qemuMonitorSetCapabilities(priv->mon);
qemuDomainObjExitMonitorWithDriver(driver, vm);
error:
return ret;
}
typedef int qemuProcessLogHandleOutput(virDomainObjPtr vm,
const char *output,
int fd);
/*
* Returns -1 for error, 0 on success
*/
static int
qemuProcessReadLogOutput(virDomainObjPtr vm,
int fd,
char *buf,
size_t buflen,
qemuProcessLogHandleOutput func,
const char *what,
int timeout)
{
int retries = (timeout*10);
int got = 0;
char *debug = NULL;
int ret = -1;
char *filter_next = buf;
buf[0] = '\0';
/* This relies on log message format generated by virLogFormatString() and
* might need to be modified when message format changes. */
if (virAsprintf(&debug, ": %d: debug : ", vm->pid) < 0) {
virReportOOMError();
return -1;
}
while (retries) {
ssize_t func_ret, bytes;
int isdead = 0;
char *eol;
func_ret = func(vm, buf, fd);
if (kill(vm->pid, 0) == -1 && errno == ESRCH)
isdead = 1;
/* Any failures should be detected before we read the log, so we
* always have something useful to report on failure. */
bytes = saferead(fd, buf+got, buflen-got-1);
if (bytes < 0) {
virReportSystemError(errno,
_("Failure while reading %s log output"),
what);
goto cleanup;
}
got += bytes;
buf[got] = '\0';
/* Filter out debug messages from intermediate libvirt process */
while ((eol = strchr(filter_next, '\n'))) {
*eol = '\0';
if (strstr(filter_next, debug)) {
memmove(filter_next, eol + 1, got - (eol - buf));
got -= eol + 1 - filter_next;
} else {
filter_next = eol + 1;
*eol = '\n';
}
}
if (got == buflen-1) {
qemuReportError(VIR_ERR_INTERNAL_ERROR,
_("Out of space while reading %s log output: %s"),
what, buf);
goto cleanup;
}
if (isdead) {
qemuReportError(VIR_ERR_INTERNAL_ERROR,
_("Process exited while reading %s log output: %s"),
what, buf);
goto cleanup;
}
if (func_ret <= 0) {
ret = func_ret;
goto cleanup;
}
usleep(100*1000);
retries--;
}
qemuReportError(VIR_ERR_INTERNAL_ERROR,
_("Timed out while reading %s log output: %s"),
what, buf);
cleanup:
VIR_FREE(debug);
return ret;
}
/*
* Look at a chunk of data from the QEMU stdout logs and try to
* find a TTY device, as indicated by a line like
*
* char device redirected to /dev/pts/3
*
* Returns -1 for error, 0 success, 1 continue reading
*/
static int
qemuProcessExtractTTYPath(const char *haystack,
size_t *offset,
char **path)
{
static const char needle[] = "char device redirected to";
char *tmp, *dev;
VIR_FREE(*path);
/* First look for our magic string */
if (!(tmp = strstr(haystack + *offset, needle))) {
return 1;
}
tmp += sizeof(needle);
dev = tmp;
/*
* And look for first whitespace character and nul terminate
* to mark end of the pty path
*/
while (*tmp) {
if (c_isspace(*tmp)) {
*path = strndup(dev, tmp-dev);
if (*path == NULL) {
virReportOOMError();
return -1;
}
/* ... now further update offset till we get EOL */
*offset = tmp - haystack;
return 0;
}
tmp++;
}
/*
* We found a path, but didn't find any whitespace,
* so it must be still incomplete - we should at
* least see a \n - indicate that we want to carry
* on trying again
*/
return 1;
}
static int
qemuProcessLookupPTYs(virDomainChrDefPtr *devices,
int count,
virHashTablePtr paths,
bool chardevfmt)
{
int i;
const char *prefix = chardevfmt ? "char" : "";
for (i = 0 ; i < count ; i++) {
virDomainChrDefPtr chr = devices[i];
if (chr->source.type == VIR_DOMAIN_CHR_TYPE_PTY) {
char id[32];
const char *path;
if (snprintf(id, sizeof(id), "%s%s",
prefix, chr->info.alias) >= sizeof(id))
return -1;
path = (const char *) virHashLookup(paths, id);
if (path == NULL) {
if (chr->source.data.file.path == NULL) {
/* neither the log output nor 'info chardev' had a
* pty path for this chardev, report an error
*/
qemuReportError(VIR_ERR_INTERNAL_ERROR,
_("no assigned pty for device %s"), id);
return -1;
} else {
/* 'info chardev' had no pty path for this chardev,
* but the log output had, so we're fine
*/
continue;
}
}
VIR_FREE(chr->source.data.file.path);
chr->source.data.file.path = strdup(path);
if (chr->source.data.file.path == NULL) {
virReportOOMError();
return -1;
}
}
}
return 0;
}
static int
qemuProcessFindCharDevicePTYsMonitor(virDomainObjPtr vm,
virBitmapPtr qemuCaps,
virHashTablePtr paths)
{
bool chardevfmt = qemuCapsGet(qemuCaps, QEMU_CAPS_CHARDEV);
if (qemuProcessLookupPTYs(vm->def->serials, vm->def->nserials,
paths, chardevfmt) < 0)
return -1;
if (qemuProcessLookupPTYs(vm->def->parallels, vm->def->nparallels,
paths, chardevfmt) < 0)
return -1;
if (qemuProcessLookupPTYs(vm->def->channels, vm->def->nchannels,
paths, chardevfmt) < 0)
return -1;
if (qemuProcessLookupPTYs(vm->def->consoles, vm->def->nconsoles,
paths, chardevfmt) < 0)
return -1;
return 0;
}
static int
qemuProcessFindCharDevicePTYs(virDomainObjPtr vm,
const char *output,
int fd ATTRIBUTE_UNUSED)
{
size_t offset = 0;
int ret, i;
/* The order in which QEMU prints out the PTY paths is
the order in which it procsses its serial and parallel
device args. This code must match that ordering.... */
/* first comes the serial devices */
for (i = 0 ; i < vm->def->nserials ; i++) {
virDomainChrDefPtr chr = vm->def->serials[i];
if (chr->source.type == VIR_DOMAIN_CHR_TYPE_PTY) {
if ((ret = qemuProcessExtractTTYPath(output, &offset,
&chr->source.data.file.path)) != 0)
return ret;
}
}
/* then the parallel devices */
for (i = 0 ; i < vm->def->nparallels ; i++) {
virDomainChrDefPtr chr = vm->def->parallels[i];
if (chr->source.type == VIR_DOMAIN_CHR_TYPE_PTY) {
if ((ret = qemuProcessExtractTTYPath(output, &offset,
&chr->source.data.file.path)) != 0)
return ret;
}
}
/* then the channel devices */
for (i = 0 ; i < vm->def->nchannels ; i++) {
virDomainChrDefPtr chr = vm->def->channels[i];
if (chr->source.type == VIR_DOMAIN_CHR_TYPE_PTY) {
if ((ret = qemuProcessExtractTTYPath(output, &offset,
&chr->source.data.file.path)) != 0)
return ret;
}
}
for (i = 0 ; i < vm->def->nconsoles ; i++) {
virDomainChrDefPtr chr = vm->def->consoles[i];
if (chr->source.type == VIR_DOMAIN_CHR_TYPE_PTY &&
chr->targetType == VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_VIRTIO) {
if ((ret = qemuProcessExtractTTYPath(output, &offset,
&chr->source.data.file.path)) != 0)
return ret;
}
}
return 0;
}
static void qemuProcessFreePtyPath(void *payload, const void *name ATTRIBUTE_UNUSED)
{
VIR_FREE(payload);
}
static void
qemuProcessReadLogFD(int logfd, char *buf, int maxlen, int off)
{
int ret;
char *tmpbuf = buf + off;
ret = saferead(logfd, tmpbuf, maxlen - off - 1);
if (ret < 0) {
ret = 0;
}
tmpbuf[ret] = '\0';
}
static int
qemuProcessWaitForMonitor(struct qemud_driver* driver,
virDomainObjPtr vm,
virBitmapPtr qemuCaps,
off_t pos)
{
char *buf = NULL;
size_t buf_size = 4096; /* Plenty of space to get startup greeting */
int logfd = -1;
int ret = -1;
virHashTablePtr paths = NULL;
build: detect potentential uninitialized variables Even with -Wuninitialized (which is part of autobuild.sh --enable-compile-warnings=error), gcc does NOT catch this use of an uninitialized variable: { if (cond) goto error; int a = 1; error: printf("%d", a); } which prints 0 (supposing the stack started life wiped) if cond was true. Clang will catch it, but we don't use clang as often. Using gcc -Wjump-misses-init catches it, but also gives false positives: { if (cond) goto error; int a = 1; return a; error: return 0; } Here, a was never used in the scope of the error block, so declaring it after goto is technically fine (and clang agrees). However, given that our HACKING already documents a preference to C89 decl-before-statement, the false positive warning is enough of a prod to comply with HACKING. [Personally, I'd _really_ rather use C99 decl-after-statement to minimize scope, but until gcc can efficiently and reliably catch scoping and uninitialized usage bugs, I'll settle with the compromise of enforcing a coding standard that happens to reject false positives if it can also detect real bugs.] * acinclude.m4 (LIBVIRT_COMPILE_WARNINGS): Add -Wjump-misses-init. * src/util/util.c (__virExec): Adjust offenders. * src/conf/domain_conf.c (virDomainTimerDefParseXML): Likewise. * src/remote/remote_driver.c (doRemoteOpen): Likewise. * src/phyp/phyp_driver.c (phypGetLparNAME, phypGetLparProfile) (phypGetVIOSFreeSCSIAdapter, phypVolumeGetKey) (phypGetStoragePoolDevice) (phypVolumeGetPhysicalVolumeByStoragePool) (phypVolumeGetPath): Likewise. * src/vbox/vbox_tmpl.c (vboxNetworkUndefineDestroy) (vboxNetworkCreate, vboxNetworkDumpXML) (vboxNetworkDefineCreateXML): Likewise. * src/xenapi/xenapi_driver.c (getCapsObject) (xenapiDomainDumpXML): Likewise. * src/xenapi/xenapi_utils.c (createVMRecordFromXml): Likewise. * src/security/security_selinux.c (SELinuxGenNewContext): Likewise. * src/qemu/qemu_command.c (qemuBuildCommandLine): Likewise. * src/qemu/qemu_hotplug.c (qemuDomainChangeEjectableMedia): Likewise. * src/qemu/qemu_process.c (qemuProcessWaitForMonitor): Likewise. * src/qemu/qemu_monitor_text.c (qemuMonitorTextGetPtyPaths): Likewise. * src/qemu/qemu_driver.c (qemudDomainShutdown) (qemudDomainBlockStats, qemudDomainMemoryPeek): Likewise. * src/storage/storage_backend_iscsi.c (virStorageBackendCreateIfaceIQN): Likewise. * src/node_device/node_device_udev.c (udevProcessPCI): Likewise.
2011-04-01 15:41:45 +00:00
qemuDomainObjPrivatePtr priv;
if (pos != -1) {
if ((logfd = qemuDomainOpenLog(driver, vm, pos)) < 0)
return -1;
if (VIR_ALLOC_N(buf, buf_size) < 0) {
virReportOOMError();
return -1;
}
if (qemuProcessReadLogOutput(vm, logfd, buf, buf_size,
qemuProcessFindCharDevicePTYs,
"console", 30) < 0)
goto closelog;
}
VIR_DEBUG("Connect monitor to %p '%s'", vm, vm->def->name);
if (qemuConnectMonitor(driver, vm) < 0) {
goto cleanup;
}
/* Try to get the pty path mappings again via the monitor. This is much more
* reliable if it's available.
* Note that the monitor itself can be on a pty, so we still need to try the
* log output method. */
paths = virHashCreate(0, qemuProcessFreePtyPath);
if (paths == NULL)
goto cleanup;
build: detect potentential uninitialized variables Even with -Wuninitialized (which is part of autobuild.sh --enable-compile-warnings=error), gcc does NOT catch this use of an uninitialized variable: { if (cond) goto error; int a = 1; error: printf("%d", a); } which prints 0 (supposing the stack started life wiped) if cond was true. Clang will catch it, but we don't use clang as often. Using gcc -Wjump-misses-init catches it, but also gives false positives: { if (cond) goto error; int a = 1; return a; error: return 0; } Here, a was never used in the scope of the error block, so declaring it after goto is technically fine (and clang agrees). However, given that our HACKING already documents a preference to C89 decl-before-statement, the false positive warning is enough of a prod to comply with HACKING. [Personally, I'd _really_ rather use C99 decl-after-statement to minimize scope, but until gcc can efficiently and reliably catch scoping and uninitialized usage bugs, I'll settle with the compromise of enforcing a coding standard that happens to reject false positives if it can also detect real bugs.] * acinclude.m4 (LIBVIRT_COMPILE_WARNINGS): Add -Wjump-misses-init. * src/util/util.c (__virExec): Adjust offenders. * src/conf/domain_conf.c (virDomainTimerDefParseXML): Likewise. * src/remote/remote_driver.c (doRemoteOpen): Likewise. * src/phyp/phyp_driver.c (phypGetLparNAME, phypGetLparProfile) (phypGetVIOSFreeSCSIAdapter, phypVolumeGetKey) (phypGetStoragePoolDevice) (phypVolumeGetPhysicalVolumeByStoragePool) (phypVolumeGetPath): Likewise. * src/vbox/vbox_tmpl.c (vboxNetworkUndefineDestroy) (vboxNetworkCreate, vboxNetworkDumpXML) (vboxNetworkDefineCreateXML): Likewise. * src/xenapi/xenapi_driver.c (getCapsObject) (xenapiDomainDumpXML): Likewise. * src/xenapi/xenapi_utils.c (createVMRecordFromXml): Likewise. * src/security/security_selinux.c (SELinuxGenNewContext): Likewise. * src/qemu/qemu_command.c (qemuBuildCommandLine): Likewise. * src/qemu/qemu_hotplug.c (qemuDomainChangeEjectableMedia): Likewise. * src/qemu/qemu_process.c (qemuProcessWaitForMonitor): Likewise. * src/qemu/qemu_monitor_text.c (qemuMonitorTextGetPtyPaths): Likewise. * src/qemu/qemu_driver.c (qemudDomainShutdown) (qemudDomainBlockStats, qemudDomainMemoryPeek): Likewise. * src/storage/storage_backend_iscsi.c (virStorageBackendCreateIfaceIQN): Likewise. * src/node_device/node_device_udev.c (udevProcessPCI): Likewise.
2011-04-01 15:41:45 +00:00
priv = vm->privateData;
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
qemuDomainObjEnterMonitorWithDriver(driver, vm);
ret = qemuMonitorGetPtyPaths(priv->mon, paths);
qemuDomainObjExitMonitorWithDriver(driver, vm);
VIR_DEBUG("qemuMonitorGetPtyPaths returned %i", ret);
if (ret == 0)
ret = qemuProcessFindCharDevicePTYsMonitor(vm, qemuCaps, paths);
cleanup:
virHashFree(paths);
if (pos != -1 && kill(vm->pid, 0) == -1 && errno == ESRCH) {
/* VM is dead, any other error raised in the interim is probably
* not as important as the qemu cmdline output */
qemuProcessReadLogFD(logfd, buf, buf_size, strlen(buf));
qemuReportError(VIR_ERR_INTERNAL_ERROR,
_("process exited while connecting to monitor: %s"),
buf);
ret = -1;
}
closelog:
if (VIR_CLOSE(logfd) < 0) {
char ebuf[1024];
VIR_WARN("Unable to close logfile: %s",
virStrerror(errno, ebuf, sizeof ebuf));
}
VIR_FREE(buf);
return ret;
}
static int
qemuProcessDetectVcpuPIDs(struct qemud_driver *driver,
virDomainObjPtr vm)
{
pid_t *cpupids = NULL;
int ncpupids;
qemuDomainObjPrivatePtr priv = vm->privateData;
if (vm->def->virtType != VIR_DOMAIN_VIRT_KVM) {
priv->nvcpupids = 1;
if (VIR_ALLOC_N(priv->vcpupids, priv->nvcpupids) < 0) {
virReportOOMError();
return -1;
}
priv->vcpupids[0] = vm->pid;
return 0;
}
/* What follows is now all KVM specific */
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
qemuDomainObjEnterMonitorWithDriver(driver, vm);
if ((ncpupids = qemuMonitorGetCPUInfo(priv->mon, &cpupids)) < 0) {
qemuDomainObjExitMonitorWithDriver(driver, vm);
return -1;
}
qemuDomainObjExitMonitorWithDriver(driver, vm);
/* Treat failure to get VCPU<->PID mapping as non-fatal */
if (ncpupids == 0)
return 0;
if (ncpupids != vm->def->vcpus) {
qemuReportError(VIR_ERR_INTERNAL_ERROR,
_("got wrong number of vCPU pids from QEMU monitor. "
"got %d, wanted %d"),
ncpupids, vm->def->vcpus);
VIR_FREE(cpupids);
return -1;
}
priv->nvcpupids = ncpupids;
priv->vcpupids = cpupids;
return 0;
}
/*
* Set NUMA memory policy for qemu process, to be run between
* fork/exec of QEMU only.
*/
#if HAVE_NUMACTL
static int
qemuProcessInitNumaMemoryPolicy(virDomainObjPtr vm)
{
nodemask_t mask;
int mode = -1;
int node = -1;
int ret = -1;
int i = 0;
int maxnode = 0;
bool warned = false;
if (!vm->def->numatune.memory.nodemask)
return 0;
VIR_DEBUG("Setting NUMA memory policy");
if (numa_available() < 0) {
qemuReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("Host kernel is not aware of NUMA."));
return -1;
}
maxnode = numa_max_node() + 1;
/* Convert nodemask to NUMA bitmask. */
nodemask_zero(&mask);
for (i = 0; i < VIR_DOMAIN_CPUMASK_LEN; i++) {
if (vm->def->numatune.memory.nodemask[i]) {
if (i > NUMA_NUM_NODES) {
qemuReportError(VIR_ERR_INTERNAL_ERROR,
_("Host cannot support NUMA node %d"), i);
return -1;
}
if (i > maxnode && !warned) {
VIR_WARN("nodeset is out of range, there is only %d NUMA "
"nodes on host", maxnode);
warned = true;
}
nodemask_set(&mask, i);
}
}
mode = vm->def->numatune.memory.mode;
if (mode == VIR_DOMAIN_NUMATUNE_MEM_STRICT) {
numa_set_bind_policy(1);
numa_set_membind(&mask);
numa_set_bind_policy(0);
} else if (mode == VIR_DOMAIN_NUMATUNE_MEM_PREFERRED) {
int nnodes = 0;
for (i = 0; i < NUMA_NUM_NODES; i++) {
if (nodemask_isset(&mask, i)) {
node = i;
nnodes++;
}
}
if (nnodes != 1) {
qemuReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("NUMA memory tuning in 'preferred' mode "
"only supports single node"));
goto cleanup;
}
numa_set_bind_policy(0);
numa_set_preferred(node);
} else if (mode == VIR_DOMAIN_NUMATUNE_MEM_INTERLEAVE) {
numa_set_interleave_mask(&mask);
} else {
/* XXX: Shouldn't go here, as we already do checking when
* parsing domain XML.
*/
qemuReportError(VIR_ERR_XML_ERROR,
"%s", _("Invalid mode for memory NUMA tuning."));
goto cleanup;
}
ret = 0;
cleanup:
return ret;
}
#else
static int
qemuProcessInitNumaMemoryPolicy(virDomainObjPtr vm)
{
if (vm->def->numatune.memory.nodemask) {
qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("libvirt is compiled without NUMA tuning support"));
return -1;
}
return 0;
}
#endif
/*
* To be run between fork/exec of QEMU only
*/
static int
qemuProcessInitCpuAffinity(virDomainObjPtr vm)
{
int i, hostcpus, maxcpu = QEMUD_CPUMASK_LEN;
virNodeInfo nodeinfo;
unsigned char *cpumap;
int cpumaplen;
VIR_DEBUG("Setting CPU affinity");
if (nodeGetInfo(NULL, &nodeinfo) < 0)
return -1;
/* setaffinity fails if you set bits for CPUs which
* aren't present, so we have to limit ourselves */
hostcpus = VIR_NODEINFO_MAXCPUS(nodeinfo);
if (maxcpu > hostcpus)
maxcpu = hostcpus;
cpumaplen = VIR_CPU_MAPLEN(maxcpu);
if (VIR_ALLOC_N(cpumap, cpumaplen) < 0) {
virReportOOMError();
return -1;
}
if (vm->def->cpumask) {
/* XXX why don't we keep 'cpumask' in the libvirt cpumap
* format to start with ?!?! */
for (i = 0 ; i < maxcpu && i < vm->def->cpumasklen ; i++)
if (vm->def->cpumask[i])
VIR_USE_CPU(cpumap, i);
} else {
/* You may think this is redundant, but we can't assume libvirtd
* itself is running on all pCPUs, so we need to explicitly set
* the spawned QEMU instance to all pCPUs if no map is given in
* its config file */
for (i = 0 ; i < maxcpu ; i++)
VIR_USE_CPU(cpumap, i);
}
/* We are pressuming we are running between fork/exec of QEMU
* so use '0' to indicate our own process ID. No threads are
* running at this point
*/
if (virProcessInfoSetAffinity(0, /* Self */
cpumap, cpumaplen, maxcpu) < 0) {
VIR_FREE(cpumap);
return -1;
}
VIR_FREE(cpumap);
return 0;
}
/* set link states to down on interfaces at qemu start */
static int
qemuProcessSetLinkStates(virDomainObjPtr vm)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
virDomainDefPtr def = vm->def;
int i;
int ret = 0;
for (i = 0; i < def->nnets; i++) {
if (def->nets[i]->linkstate == VIR_DOMAIN_NET_INTERFACE_LINK_STATE_DOWN) {
VIR_DEBUG("Setting link state: %s", def->nets[i]->info.alias);
if (!qemuCapsGet(priv->qemuCaps, QEMU_CAPS_NETDEV)) {
qemuReportError(VIR_ERR_NO_SUPPORT,
_("Setting of link state is not supported by this qemu"));
return -1;
}
ret = qemuMonitorSetLink(priv->mon,
def->nets[i]->info.alias,
VIR_DOMAIN_NET_INTERFACE_LINK_STATE_DOWN);
if (ret != 0) {
qemuReportError(VIR_ERR_OPERATION_FAILED,
_("Couldn't set link state on interface: %s"), def->nets[i]->info.alias);
break;
}
}
}
return ret;
}
/* Set CPU affinities for vcpus if vcpupin xml provided. */
static int
qemuProcessSetVcpuAffinites(virConnectPtr conn,
virDomainObjPtr vm)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
virDomainDefPtr def = vm->def;
virNodeInfo nodeinfo;
pid_t vcpupid;
unsigned char *cpumask;
int vcpu, cpumaplen, hostcpus, maxcpu, n;
unsigned char *cpumap = NULL;
int ret = -1;
if (virNodeGetInfo(conn, &nodeinfo) != 0) {
return -1;
}
if (!def->cputune.nvcpupin)
return 0;
if (priv->vcpupids == NULL) {
qemuReportError(VIR_ERR_OPERATION_INVALID,
"%s", _("cpu affinity is not supported"));
return -1;
}
hostcpus = VIR_NODEINFO_MAXCPUS(nodeinfo);
cpumaplen = VIR_CPU_MAPLEN(hostcpus);
maxcpu = cpumaplen * 8;
if (maxcpu > hostcpus)
maxcpu = hostcpus;
if (VIR_ALLOC_N(cpumap, cpumaplen) < 0) {
virReportOOMError();
return -1;
}
for (n = 0; n < def->cputune.nvcpupin; n++) {
int i;
vcpu = def->cputune.vcpupin[n]->vcpuid;
memset(cpumap, 0, cpumaplen);
cpumask = (unsigned char *)def->cputune.vcpupin[n]->cpumask;
vcpupid = priv->vcpupids[vcpu];
2011-06-07 08:51:51 +00:00
for (i = 0 ; i < VIR_DOMAIN_CPUMASK_LEN ; i++) {
if (cpumask[i])
2011-06-07 08:51:51 +00:00
VIR_USE_CPU(cpumap, i);
}
if (virProcessInfoSetAffinity(vcpupid,
cpumap,
cpumaplen,
maxcpu) < 0) {
goto cleanup;
}
}
ret = 0;
cleanup:
VIR_FREE(cpumap);
return ret;
}
static int
qemuProcessInitPasswords(virConnectPtr conn,
struct qemud_driver *driver,
virDomainObjPtr vm)
{
int ret = 0;
qemuDomainObjPrivatePtr priv = vm->privateData;
if (vm->def->ngraphics == 1) {
if (vm->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC) {
ret = qemuDomainChangeGraphicsPasswords(driver, vm,
VIR_DOMAIN_GRAPHICS_TYPE_VNC,
&vm->def->graphics[0]->data.vnc.auth,
driver->vncPassword);
} else if (vm->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_SPICE) {
ret = qemuDomainChangeGraphicsPasswords(driver, vm,
VIR_DOMAIN_GRAPHICS_TYPE_SPICE,
&vm->def->graphics[0]->data.spice.auth,
driver->spicePassword);
}
}
if (ret < 0)
goto cleanup;
if (qemuCapsGet(priv->qemuCaps, QEMU_CAPS_DEVICE)) {
int i;
for (i = 0 ; i < vm->def->ndisks ; i++) {
char *secret;
size_t secretLen;
const char *alias;
if (!vm->def->disks[i]->encryption ||
!vm->def->disks[i]->src)
continue;
if (qemuProcessGetVolumeQcowPassphrase(conn,
vm->def->disks[i],
&secret, &secretLen) < 0)
goto cleanup;
alias = vm->def->disks[i]->info.alias;
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
qemuDomainObjEnterMonitorWithDriver(driver, vm);
ret = qemuMonitorSetDrivePassphrase(priv->mon, alias, secret);
VIR_FREE(secret);
qemuDomainObjExitMonitorWithDriver(driver, vm);
if (ret < 0)
goto cleanup;
}
}
cleanup:
return ret;
}
#define QEMU_PCI_VENDOR_INTEL 0x8086
#define QEMU_PCI_VENDOR_LSI_LOGIC 0x1000
#define QEMU_PCI_VENDOR_REDHAT 0x1af4
#define QEMU_PCI_VENDOR_CIRRUS 0x1013
#define QEMU_PCI_VENDOR_REALTEK 0x10ec
#define QEMU_PCI_VENDOR_AMD 0x1022
#define QEMU_PCI_VENDOR_ENSONIQ 0x1274
#define QEMU_PCI_VENDOR_VMWARE 0x15ad
#define QEMU_PCI_VENDOR_QEMU 0x1234
#define QEMU_PCI_PRODUCT_DISK_VIRTIO 0x1001
#define QEMU_PCI_PRODUCT_BALLOON_VIRTIO 0x1002
#define QEMU_PCI_PRODUCT_NIC_NE2K 0x8029
#define QEMU_PCI_PRODUCT_NIC_PCNET 0x2000
#define QEMU_PCI_PRODUCT_NIC_RTL8139 0x8139
#define QEMU_PCI_PRODUCT_NIC_E1000 0x100E
#define QEMU_PCI_PRODUCT_NIC_VIRTIO 0x1000
#define QEMU_PCI_PRODUCT_VGA_CIRRUS 0x00b8
#define QEMU_PCI_PRODUCT_VGA_VMWARE 0x0405
#define QEMU_PCI_PRODUCT_VGA_STDVGA 0x1111
#define QEMU_PCI_PRODUCT_AUDIO_AC97 0x2415
#define QEMU_PCI_PRODUCT_AUDIO_ES1370 0x5000
#define QEMU_PCI_PRODUCT_CONTROLLER_PIIX 0x7010
#define QEMU_PCI_PRODUCT_CONTROLLER_LSI 0x0012
#define QEMU_PCI_PRODUCT_WATCHDOG_I63000ESB 0x25ab
static int
qemuProcessAssignNextPCIAddress(virDomainDeviceInfo *info,
int vendor,
int product,
qemuMonitorPCIAddress *addrs,
int naddrs)
{
int found = 0;
int i;
VIR_DEBUG("Look for %x:%x out of %d", vendor, product, naddrs);
for (i = 0 ; (i < naddrs) && !found; i++) {
VIR_DEBUG("Maybe %x:%x", addrs[i].vendor, addrs[i].product);
if (addrs[i].vendor == vendor &&
addrs[i].product == product) {
VIR_DEBUG("Match %d", i);
found = 1;
break;
}
}
if (!found) {
return -1;
}
/* Blank it out so this device isn't matched again */
addrs[i].vendor = 0;
addrs[i].product = 0;
if (info->type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE)
info->type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI;
if (info->type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI) {
info->addr.pci.domain = addrs[i].addr.domain;
info->addr.pci.bus = addrs[i].addr.bus;
info->addr.pci.slot = addrs[i].addr.slot;
info->addr.pci.function = addrs[i].addr.function;
}
return 0;
}
static int
qemuProcessGetPCIDiskVendorProduct(virDomainDiskDefPtr def,
unsigned *vendor,
unsigned *product)
{
switch (def->bus) {
case VIR_DOMAIN_DISK_BUS_VIRTIO:
*vendor = QEMU_PCI_VENDOR_REDHAT;
*product = QEMU_PCI_PRODUCT_DISK_VIRTIO;
break;
default:
return -1;
}
return 0;
}
static int
qemuProcessGetPCINetVendorProduct(virDomainNetDefPtr def,
unsigned *vendor,
unsigned *product)
{
if (!def->model)
return -1;
if (STREQ(def->model, "ne2k_pci")) {
*vendor = QEMU_PCI_VENDOR_REALTEK;
*product = QEMU_PCI_PRODUCT_NIC_NE2K;
} else if (STREQ(def->model, "pcnet")) {
*vendor = QEMU_PCI_VENDOR_AMD;
*product = QEMU_PCI_PRODUCT_NIC_PCNET;
} else if (STREQ(def->model, "rtl8139")) {
*vendor = QEMU_PCI_VENDOR_REALTEK;
*product = QEMU_PCI_PRODUCT_NIC_RTL8139;
} else if (STREQ(def->model, "e1000")) {
*vendor = QEMU_PCI_VENDOR_INTEL;
*product = QEMU_PCI_PRODUCT_NIC_E1000;
} else if (STREQ(def->model, "virtio")) {
*vendor = QEMU_PCI_VENDOR_REDHAT;
*product = QEMU_PCI_PRODUCT_NIC_VIRTIO;
} else {
VIR_INFO("Unexpected NIC model %s, cannot get PCI address",
def->model);
return -1;
}
return 0;
}
static int
qemuProcessGetPCIControllerVendorProduct(virDomainControllerDefPtr def,
unsigned *vendor,
unsigned *product)
{
switch (def->type) {
case VIR_DOMAIN_CONTROLLER_TYPE_SCSI:
*vendor = QEMU_PCI_VENDOR_LSI_LOGIC;
*product = QEMU_PCI_PRODUCT_CONTROLLER_LSI;
break;
case VIR_DOMAIN_CONTROLLER_TYPE_FDC:
/* XXX we could put in the ISA bridge address, but
that's not technically the FDC's address */
return -1;
case VIR_DOMAIN_CONTROLLER_TYPE_IDE:
*vendor = QEMU_PCI_VENDOR_INTEL;
*product = QEMU_PCI_PRODUCT_CONTROLLER_PIIX;
break;
default:
VIR_INFO("Unexpected controller type %s, cannot get PCI address",
virDomainControllerTypeToString(def->type));
return -1;
}
return 0;
}
static int
qemuProcessGetPCIVideoVendorProduct(virDomainVideoDefPtr def,
unsigned *vendor,
unsigned *product)
{
switch (def->type) {
case VIR_DOMAIN_VIDEO_TYPE_CIRRUS:
*vendor = QEMU_PCI_VENDOR_CIRRUS;
*product = QEMU_PCI_PRODUCT_VGA_CIRRUS;
break;
case VIR_DOMAIN_VIDEO_TYPE_VGA:
*vendor = QEMU_PCI_VENDOR_QEMU;
*product = QEMU_PCI_PRODUCT_VGA_STDVGA;
break;
case VIR_DOMAIN_VIDEO_TYPE_VMVGA:
*vendor = QEMU_PCI_VENDOR_VMWARE;
*product = QEMU_PCI_PRODUCT_VGA_VMWARE;
break;
default:
return -1;
}
return 0;
}
static int
qemuProcessGetPCISoundVendorProduct(virDomainSoundDefPtr def,
unsigned *vendor,
unsigned *product)
{
switch (def->model) {
case VIR_DOMAIN_SOUND_MODEL_ES1370:
*vendor = QEMU_PCI_VENDOR_ENSONIQ;
*product = QEMU_PCI_PRODUCT_AUDIO_ES1370;
break;
case VIR_DOMAIN_SOUND_MODEL_AC97:
*vendor = QEMU_PCI_VENDOR_INTEL;
*product = QEMU_PCI_PRODUCT_AUDIO_AC97;
break;
default:
return -1;
}
return 0;
}
static int
qemuProcessGetPCIWatchdogVendorProduct(virDomainWatchdogDefPtr def,
unsigned *vendor,
unsigned *product)
{
switch (def->model) {
case VIR_DOMAIN_WATCHDOG_MODEL_I6300ESB:
*vendor = QEMU_PCI_VENDOR_INTEL;
*product = QEMU_PCI_PRODUCT_WATCHDOG_I63000ESB;
break;
default:
return -1;
}
return 0;
}
static int
qemuProcessGetPCIMemballoonVendorProduct(virDomainMemballoonDefPtr def,
unsigned *vendor,
unsigned *product)
{
switch (def->model) {
case VIR_DOMAIN_MEMBALLOON_MODEL_VIRTIO:
*vendor = QEMU_PCI_VENDOR_REDHAT;
*product = QEMU_PCI_PRODUCT_BALLOON_VIRTIO;
break;
default:
return -1;
}
return 0;
}
/*
* This entire method assumes that PCI devices in 'info pci'
* match ordering of devices specified on the command line
* wrt to devices of matching vendor+product
*
* XXXX this might not be a valid assumption if we assign
* some static addrs on CLI. Have to check that...
*/
static int
qemuProcessDetectPCIAddresses(virDomainObjPtr vm,
qemuMonitorPCIAddress *addrs,
int naddrs)
{
unsigned int vendor = 0, product = 0;
int i;
/* XXX should all these vendor/product IDs be kept in the
* actual device data structure instead ?
*/
for (i = 0 ; i < vm->def->ndisks ; i++) {
if (qemuProcessGetPCIDiskVendorProduct(vm->def->disks[i], &vendor, &product) < 0)
continue;
if (qemuProcessAssignNextPCIAddress(&(vm->def->disks[i]->info),
vendor, product,
addrs, naddrs) < 0) {
qemuReportError(VIR_ERR_INTERNAL_ERROR,
_("cannot find PCI address for VirtIO disk %s"),
vm->def->disks[i]->dst);
return -1;
}
}
for (i = 0 ; i < vm->def->nnets ; i++) {
if (qemuProcessGetPCINetVendorProduct(vm->def->nets[i], &vendor, &product) < 0)
continue;
if (qemuProcessAssignNextPCIAddress(&(vm->def->nets[i]->info),
vendor, product,
addrs, naddrs) < 0) {
qemuReportError(VIR_ERR_INTERNAL_ERROR,
_("cannot find PCI address for %s NIC"),
vm->def->nets[i]->model);
return -1;
}
}
for (i = 0 ; i < vm->def->ncontrollers ; i++) {
if (qemuProcessGetPCIControllerVendorProduct(vm->def->controllers[i], &vendor, &product) < 0)
continue;
if (qemuProcessAssignNextPCIAddress(&(vm->def->controllers[i]->info),
vendor, product,
addrs, naddrs) < 0) {
qemuReportError(VIR_ERR_INTERNAL_ERROR,
_("cannot find PCI address for controller %s"),
virDomainControllerTypeToString(vm->def->controllers[i]->type));
return -1;
}
}
for (i = 0 ; i < vm->def->nvideos ; i++) {
if (qemuProcessGetPCIVideoVendorProduct(vm->def->videos[i], &vendor, &product) < 0)
continue;
if (qemuProcessAssignNextPCIAddress(&(vm->def->videos[i]->info),
vendor, product,
addrs, naddrs) < 0) {
qemuReportError(VIR_ERR_INTERNAL_ERROR,
_("cannot find PCI address for video adapter %s"),
virDomainVideoTypeToString(vm->def->videos[i]->type));
return -1;
}
}
for (i = 0 ; i < vm->def->nsounds ; i++) {
if (qemuProcessGetPCISoundVendorProduct(vm->def->sounds[i], &vendor, &product) < 0)
continue;
if (qemuProcessAssignNextPCIAddress(&(vm->def->sounds[i]->info),
vendor, product,
addrs, naddrs) < 0) {
qemuReportError(VIR_ERR_INTERNAL_ERROR,
_("cannot find PCI address for sound adapter %s"),
virDomainSoundModelTypeToString(vm->def->sounds[i]->model));
return -1;
}
}
if (vm->def->watchdog &&
qemuProcessGetPCIWatchdogVendorProduct(vm->def->watchdog, &vendor, &product) == 0) {
if (qemuProcessAssignNextPCIAddress(&(vm->def->watchdog->info),
vendor, product,
addrs, naddrs) < 0) {
qemuReportError(VIR_ERR_INTERNAL_ERROR,
_("cannot find PCI address for watchdog %s"),
virDomainWatchdogModelTypeToString(vm->def->watchdog->model));
return -1;
}
}
if (vm->def->memballoon &&
qemuProcessGetPCIMemballoonVendorProduct(vm->def->memballoon, &vendor, &product) == 0) {
if (qemuProcessAssignNextPCIAddress(&(vm->def->memballoon->info),
vendor, product,
addrs, naddrs) < 0) {
qemuReportError(VIR_ERR_INTERNAL_ERROR,
_("cannot find PCI address for balloon %s"),
virDomainMemballoonModelTypeToString(vm->def->memballoon->model));
return -1;
}
}
/* XXX console (virtio) */
/* ... and now things we don't have in our xml */
/* XXX USB controller ? */
/* XXX what about other PCI devices (ie bridges) */
return 0;
}
static int
qemuProcessInitPCIAddresses(struct qemud_driver *driver,
virDomainObjPtr vm)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
int naddrs;
int ret;
qemuMonitorPCIAddress *addrs = NULL;
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
qemuDomainObjEnterMonitorWithDriver(driver, vm);
naddrs = qemuMonitorGetAllPCIAddresses(priv->mon,
&addrs);
qemuDomainObjExitMonitorWithDriver(driver, vm);
ret = qemuProcessDetectPCIAddresses(vm, addrs, naddrs);
VIR_FREE(addrs);
return ret;
}
static int qemuProcessNextFreePort(struct qemud_driver *driver,
int startPort)
{
int i;
for (i = startPort ; i < QEMU_VNC_PORT_MAX; i++) {
int fd;
int reuse = 1;
struct sockaddr_in addr;
bool used = false;
if (virBitmapGetBit(driver->reservedVNCPorts,
i - QEMU_VNC_PORT_MIN, &used) < 0)
VIR_DEBUG("virBitmapGetBit failed on bit %d", i - QEMU_VNC_PORT_MIN);
if (used)
continue;
addr.sin_family = AF_INET;
addr.sin_port = htons(i);
addr.sin_addr.s_addr = htonl(INADDR_ANY);
fd = socket(PF_INET, SOCK_STREAM, 0);
if (fd < 0)
return -1;
if (setsockopt(fd, SOL_SOCKET, SO_REUSEADDR, (void*)&reuse, sizeof(reuse)) < 0) {
VIR_FORCE_CLOSE(fd);
break;
}
if (bind(fd, (struct sockaddr*)&addr, sizeof(addr)) == 0) {
/* Not in use, lets grab it */
VIR_FORCE_CLOSE(fd);
/* Add port to bitmap of reserved ports */
if (virBitmapSetBit(driver->reservedVNCPorts,
i - QEMU_VNC_PORT_MIN) < 0) {
VIR_DEBUG("virBitmapSetBit failed on bit %d",
i - QEMU_VNC_PORT_MIN);
}
return i;
}
VIR_FORCE_CLOSE(fd);
if (errno == EADDRINUSE) {
/* In use, try next */
continue;
}
/* Some other bad failure, get out.. */
break;
}
return -1;
}
static void
qemuProcessReturnPort(struct qemud_driver *driver,
int port)
{
if (port < QEMU_VNC_PORT_MIN)
return;
if (virBitmapClearBit(driver->reservedVNCPorts,
port - QEMU_VNC_PORT_MIN) < 0)
VIR_DEBUG("Could not mark port %d as unused", port);
}
static int
qemuProcessPrepareChardevDevice(virDomainDefPtr def ATTRIBUTE_UNUSED,
virDomainChrDefPtr dev,
void *opaque ATTRIBUTE_UNUSED)
{
int fd;
if (dev->source.type != VIR_DOMAIN_CHR_TYPE_FILE)
return 0;
if ((fd = open(dev->source.data.file.path,
O_CREAT | O_APPEND, S_IRUSR|S_IWUSR)) < 0) {
virReportSystemError(errno,
_("Unable to pre-create chardev file '%s'"),
dev->source.data.file.path);
return -1;
}
VIR_FORCE_CLOSE(fd);
return 0;
}
static int
qemuProcessLimits(struct qemud_driver *driver)
{
if (driver->maxProcesses > 0) {
struct rlimit rlim;
rlim.rlim_cur = rlim.rlim_max = driver->maxProcesses;
if (setrlimit(RLIMIT_NPROC, &rlim) < 0) {
virReportSystemError(errno,
_("cannot limit number of processes to %d"),
driver->maxProcesses);
return -1;
}
}
return 0;
}
struct qemuProcessHookData {
virConnectPtr conn;
virDomainObjPtr vm;
struct qemud_driver *driver;
};
static int qemuProcessHook(void *data)
{
struct qemuProcessHookData *h = data;
int ret = -1;
int fd;
/* Some later calls want pid present */
h->vm->pid = getpid();
VIR_DEBUG("Obtaining domain lock");
/*
* Since we're going to leak the returned FD to QEMU,
* we need to make sure it gets a sensible label.
* This mildly sucks, because there could be other
* sockets the lock driver opens that we don't want
* labelled. So far we're ok though.
*/
if (virSecurityManagerSetSocketLabel(h->driver->securityManager, h->vm) < 0)
goto cleanup;
if (virDomainLockProcessStart(h->driver->lockManager,
h->vm,
/* QEMU is always pased initially */
true,
&fd) < 0)
goto cleanup;
if (virSecurityManagerClearSocketLabel(h->driver->securityManager, h->vm) < 0)
goto cleanup;
if (qemuProcessLimits(h->driver) < 0)
goto cleanup;
/* This must take place before exec(), so that all QEMU
* memory allocation is on the correct NUMA node
*/
VIR_DEBUG("Moving procss to cgroup");
if (qemuAddToCgroup(h->driver, h->vm->def) < 0)
goto cleanup;
/* This must be done after cgroup placement to avoid resetting CPU
* affinity */
VIR_DEBUG("Setup CPU affinity");
if (qemuProcessInitCpuAffinity(h->vm) < 0)
goto cleanup;
if (qemuProcessInitNumaMemoryPolicy(h->vm) < 0)
return -1;
VIR_DEBUG("Setting up security labelling");
if (virSecurityManagerSetProcessLabel(h->driver->securityManager, h->vm) < 0)
goto cleanup;
ret = 0;
cleanup:
VIR_DEBUG("Hook complete ret=%d", ret);
return ret;
}
int
qemuProcessPrepareMonitorChr(struct qemud_driver *driver,
virDomainChrSourceDefPtr monConfig,
const char *vm)
{
monConfig->type = VIR_DOMAIN_CHR_TYPE_UNIX;
monConfig->data.nix.listen = true;
if (virAsprintf(&monConfig->data.nix.path, "%s/%s.monitor",
driver->libDir, vm) < 0) {
virReportOOMError();
return -1;
}
return 0;
}
/*
* Precondition: Both driver and vm must be locked,
* and a job must be active. This method will call
* {Enter,Exit}MonitorWithDriver
*/
int
qemuProcessStartCPUs(struct qemud_driver *driver, virDomainObjPtr vm,
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
virConnectPtr conn, virDomainRunningReason reason,
enum qemuDomainAsyncJob asyncJob)
{
int ret;
qemuDomainObjPrivatePtr priv = vm->privateData;
VIR_DEBUG("Using lock state '%s'", NULLSTR(priv->lockState));
if (virDomainLockProcessResume(driver->lockManager, vm, priv->lockState) < 0) {
/* Don't free priv->lockState on error, because we need
* to make sure we have state still present if the user
* tries to resume again
*/
return -1;
}
VIR_FREE(priv->lockState);
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
ret = qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob);
if (ret == 0) {
ret = qemuMonitorStartCPUs(priv->mon, conn);
qemuDomainObjExitMonitorWithDriver(driver, vm);
}
if (ret == 0) {
virDomainObjSetState(vm, VIR_DOMAIN_RUNNING, reason);
} else {
if (virDomainLockProcessPause(driver->lockManager, vm, &priv->lockState) < 0)
VIR_WARN("Unable to release lease on %s", vm->def->name);
VIR_DEBUG("Preserving lock state '%s'", NULLSTR(priv->lockState));
}
return ret;
}
int qemuProcessStopCPUs(struct qemud_driver *driver, virDomainObjPtr vm,
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
virDomainPausedReason reason,
enum qemuDomainAsyncJob asyncJob)
{
int ret;
int oldState;
int oldReason;
qemuDomainObjPrivatePtr priv = vm->privateData;
VIR_FREE(priv->lockState);
oldState = virDomainObjGetState(vm, &oldReason);
virDomainObjSetState(vm, VIR_DOMAIN_PAUSED, reason);
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
ret = qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob);
qemu: Allow all query commands to be run during long jobs Query commands are safe to be called during long running jobs (such as migration). This patch makes them all work without the need to special-case every single one of them. The patch introduces new job.asyncCond condition and associated job.asyncJob which are dedicated to asynchronous (from qemu monitor point of view) jobs that can take arbitrarily long time to finish while qemu monitor is still usable for other commands. The existing job.active (and job.cond condition) is used all other synchronous jobs (including the commands run during async job). Locking schema is changed to use these two conditions. While asyncJob is active, only allowed set of synchronous jobs is allowed (the set can be different according to a particular asyncJob) so any method that communicates to qemu monitor needs to check if it is allowed to be executed during current asyncJob (if any). Once the check passes, the method needs to normally acquire job.cond to ensure no other command is running. Since domain object lock is released during that time, asyncJob could have been started in the meantime so the method needs to recheck the first condition. Then, normal jobs set job.active and asynchronous jobs set job.asyncJob and optionally change the list of allowed job groups. Since asynchronous jobs only set job.asyncJob, other allowed commands can still be run when domain object is unlocked (when communicating to remote libvirtd or sleeping). To protect its own internal synchronous commands, the asynchronous job needs to start a special nested job before entering qemu monitor. The nested job doesn't check asyncJob, it only acquires job.cond and sets job.active to block other jobs.
2011-06-30 09:23:50 +00:00
if (ret == 0) {
ret = qemuMonitorStopCPUs(priv->mon);
qemuDomainObjExitMonitorWithDriver(driver, vm);
}
if (ret == 0) {
if (virDomainLockProcessPause(driver->lockManager, vm, &priv->lockState) < 0)
VIR_WARN("Unable to release lease on %s", vm->def->name);
VIR_DEBUG("Preserving lock state '%s'", NULLSTR(priv->lockState));
} else {
virDomainObjSetState(vm, oldState, oldReason);
}
return ret;
}
static int
qemuProcessNotifyNets(virDomainDefPtr def)
{
int ii;
for (ii = 0 ; ii < def->nnets ; ii++) {
virDomainNetDefPtr net = def->nets[ii];
if (networkNotifyActualDevice(net) < 0)
return -1;
}
return 0;
}
static int
qemuProcessFiltersInstantiate(virConnectPtr conn,
virDomainDefPtr def)
{
int err = 0;
int i;
if (!conn)
return 1;
for (i = 0 ; i < def->nnets ; i++) {
virDomainNetDefPtr net = def->nets[i];
if ((net->filter) && (net->ifname)) {
if (virDomainConfNWFilterInstantiate(conn, net)) {
err = 1;
break;
}
}
}
return err;
}
static int
qemuProcessUpdateState(struct qemud_driver *driver, virDomainObjPtr vm)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
virDomainState state;
virDomainPausedReason reason;
bool running;
int ret;
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
qemuDomainObjEnterMonitorWithDriver(driver, vm);
ret = qemuMonitorGetStatus(priv->mon, &running, &reason);
qemuDomainObjExitMonitorWithDriver(driver, vm);
if (ret < 0 || !virDomainObjIsActive(vm))
return -1;
state = virDomainObjGetState(vm, NULL);
if (state == VIR_DOMAIN_PAUSED && running) {
VIR_DEBUG("Domain %s was unpaused while its monitor was disconnected;"
" changing state to running", vm->def->name);
virDomainObjSetState(vm, VIR_DOMAIN_RUNNING,
VIR_DOMAIN_RUNNING_UNPAUSED);
} else if (state == VIR_DOMAIN_RUNNING && !running) {
VIR_DEBUG("Domain %s was paused (%s) while its monitor was"
" disconnected; changing state to paused",
vm->def->name, virDomainPausedReasonTypeToString(reason));
virDomainObjSetState(vm, VIR_DOMAIN_PAUSED, reason);
} else if (state == VIR_DOMAIN_SHUTOFF && running) {
VIR_DEBUG("Domain %s finished booting; changing state to running",
vm->def->name);
virDomainObjSetState(vm, VIR_DOMAIN_RUNNING, VIR_DOMAIN_RUNNING_BOOTED);
}
return 0;
}
static int
qemuProcessRecoverMigration(struct qemud_driver *driver,
virDomainObjPtr vm,
virConnectPtr conn,
enum qemuDomainAsyncJob job,
enum qemuMigrationJobPhase phase,
virDomainState state,
int reason)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
if (job == QEMU_ASYNC_JOB_MIGRATION_IN) {
switch (phase) {
case QEMU_MIGRATION_PHASE_NONE:
case QEMU_MIGRATION_PHASE_PERFORM2:
case QEMU_MIGRATION_PHASE_BEGIN3:
case QEMU_MIGRATION_PHASE_PERFORM3:
case QEMU_MIGRATION_PHASE_PERFORM3_DONE:
case QEMU_MIGRATION_PHASE_CONFIRM3_CANCELLED:
case QEMU_MIGRATION_PHASE_CONFIRM3:
case QEMU_MIGRATION_PHASE_LAST:
break;
case QEMU_MIGRATION_PHASE_PREPARE:
VIR_DEBUG("Killing unfinished incoming migration for domain %s",
vm->def->name);
return -1;
case QEMU_MIGRATION_PHASE_FINISH2:
/* source domain is already killed so let's just resume the domain
* and hope we are all set */
VIR_DEBUG("Incoming migration finished, resuming domain %s",
vm->def->name);
if (qemuProcessStartCPUs(driver, vm, conn,
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
VIR_DOMAIN_RUNNING_UNPAUSED,
QEMU_ASYNC_JOB_NONE) < 0) {
VIR_WARN("Could not resume domain %s", vm->def->name);
}
break;
case QEMU_MIGRATION_PHASE_FINISH3:
/* migration finished, we started resuming the domain but didn't
* confirm success or failure yet; killing it seems safest */
VIR_DEBUG("Killing migrated domain %s", vm->def->name);
return -1;
}
} else if (job == QEMU_ASYNC_JOB_MIGRATION_OUT) {
switch (phase) {
case QEMU_MIGRATION_PHASE_NONE:
case QEMU_MIGRATION_PHASE_PREPARE:
case QEMU_MIGRATION_PHASE_FINISH2:
case QEMU_MIGRATION_PHASE_FINISH3:
case QEMU_MIGRATION_PHASE_LAST:
break;
case QEMU_MIGRATION_PHASE_BEGIN3:
/* nothing happen so far, just forget we were about to migrate the
* domain */
break;
case QEMU_MIGRATION_PHASE_PERFORM2:
case QEMU_MIGRATION_PHASE_PERFORM3:
/* migration is still in progress, let's cancel it and resume the
* domain */
VIR_DEBUG("Canceling unfinished outgoing migration of domain %s",
vm->def->name);
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
qemuDomainObjEnterMonitor(driver, vm);
ignore_value(qemuMonitorMigrateCancel(priv->mon));
qemuDomainObjExitMonitor(driver, vm);
/* resume the domain but only if it was paused as a result of
* migration */
if (state == VIR_DOMAIN_PAUSED &&
(reason == VIR_DOMAIN_PAUSED_MIGRATION ||
reason == VIR_DOMAIN_PAUSED_UNKNOWN)) {
if (qemuProcessStartCPUs(driver, vm, conn,
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
VIR_DOMAIN_RUNNING_UNPAUSED,
QEMU_ASYNC_JOB_NONE) < 0) {
VIR_WARN("Could not resume domain %s", vm->def->name);
}
}
break;
case QEMU_MIGRATION_PHASE_PERFORM3_DONE:
/* migration finished but we didn't have a chance to get the result
* of Finish3 step; third party needs to check what to do next
*/
break;
case QEMU_MIGRATION_PHASE_CONFIRM3_CANCELLED:
/* Finish3 failed, we need to resume the domain */
VIR_DEBUG("Resuming domain %s after failed migration",
vm->def->name);
if (state == VIR_DOMAIN_PAUSED &&
(reason == VIR_DOMAIN_PAUSED_MIGRATION ||
reason == VIR_DOMAIN_PAUSED_UNKNOWN)) {
if (qemuProcessStartCPUs(driver, vm, conn,
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
VIR_DOMAIN_RUNNING_UNPAUSED,
QEMU_ASYNC_JOB_NONE) < 0) {
VIR_WARN("Could not resume domain %s", vm->def->name);
}
}
break;
case QEMU_MIGRATION_PHASE_CONFIRM3:
/* migration completed, we need to kill the domain here */
return -1;
}
}
return 0;
}
static int
qemuProcessRecoverJob(struct qemud_driver *driver,
virDomainObjPtr vm,
virConnectPtr conn,
const struct qemuDomainJobObj *job)
{
qemuDomainObjPrivatePtr priv = vm->privateData;
virDomainState state;
int reason;
state = virDomainObjGetState(vm, &reason);
switch (job->asyncJob) {
case QEMU_ASYNC_JOB_MIGRATION_OUT:
case QEMU_ASYNC_JOB_MIGRATION_IN:
if (qemuProcessRecoverMigration(driver, vm, conn, job->asyncJob,
job->phase, state, reason) < 0)
return -1;
break;
case QEMU_ASYNC_JOB_SAVE:
case QEMU_ASYNC_JOB_DUMP:
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
qemuDomainObjEnterMonitor(driver, vm);
ignore_value(qemuMonitorMigrateCancel(priv->mon));
qemuDomainObjExitMonitor(driver, vm);
/* resume the domain but only if it was paused as a result of
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
* running save/dump operation. Although we are recovering an
* async job, this function is run at startup and must resume
* things using sync monitor connections. */
if (state == VIR_DOMAIN_PAUSED &&
((job->asyncJob == QEMU_ASYNC_JOB_DUMP &&
reason == VIR_DOMAIN_PAUSED_DUMP) ||
(job->asyncJob == QEMU_ASYNC_JOB_SAVE &&
reason == VIR_DOMAIN_PAUSED_SAVE) ||
reason == VIR_DOMAIN_PAUSED_UNKNOWN)) {
if (qemuProcessStartCPUs(driver, vm, conn,
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
VIR_DOMAIN_RUNNING_UNPAUSED,
QEMU_ASYNC_JOB_NONE) < 0) {
VIR_WARN("Could not resume domain %s after", vm->def->name);
}
}
break;
case QEMU_ASYNC_JOB_NONE:
case QEMU_ASYNC_JOB_LAST:
break;
}
if (!virDomainObjIsActive(vm))
return -1;
switch (job->active) {
case QEMU_JOB_QUERY:
/* harmless */
break;
case QEMU_JOB_DESTROY:
VIR_DEBUG("Domain %s should have already been destroyed",
vm->def->name);
return -1;
case QEMU_JOB_SUSPEND:
/* mostly harmless */
break;
case QEMU_JOB_MODIFY:
/* XXX depending on the command we may be in an inconsistent state and
* we should probably fall back to "monitor error" state and refuse to
*/
break;
case QEMU_JOB_MIGRATION_OP:
case QEMU_JOB_ABORT:
case QEMU_JOB_ASYNC:
case QEMU_JOB_ASYNC_NESTED:
/* async job was already handled above */
case QEMU_JOB_NONE:
case QEMU_JOB_LAST:
break;
}
return 0;
}
struct qemuProcessReconnectData {
virConnectPtr conn;
struct qemud_driver *driver;
void *payload;
struct qemuDomainJobObj oldjob;
};
/*
* Open an existing VM's monitor, re-detect VCPU threads
* and re-reserve the security labels in use
*
* We own the virConnectPtr we are passed here - whoever started
* this thread function has increased the reference counter to it
* so that we now have to close it.
*/
static void
qemuProcessReconnect(void *opaque)
{
struct qemuProcessReconnectData *data = opaque;
struct qemud_driver *driver = data->driver;
virDomainObjPtr obj = data->payload;
qemuDomainObjPrivatePtr priv;
virConnectPtr conn = data->conn;
struct qemuDomainJobObj oldjob;
int state;
int reason;
memcpy(&oldjob, &data->oldjob, sizeof(oldjob));
VIR_FREE(data);
qemuDriverLock(driver);
virDomainObjLock(obj);
VIR_DEBUG("Reconnect monitor to %p '%s'", obj, obj->def->name);
priv = obj->privateData;
qemu: Allow all query commands to be run during long jobs Query commands are safe to be called during long running jobs (such as migration). This patch makes them all work without the need to special-case every single one of them. The patch introduces new job.asyncCond condition and associated job.asyncJob which are dedicated to asynchronous (from qemu monitor point of view) jobs that can take arbitrarily long time to finish while qemu monitor is still usable for other commands. The existing job.active (and job.cond condition) is used all other synchronous jobs (including the commands run during async job). Locking schema is changed to use these two conditions. While asyncJob is active, only allowed set of synchronous jobs is allowed (the set can be different according to a particular asyncJob) so any method that communicates to qemu monitor needs to check if it is allowed to be executed during current asyncJob (if any). Once the check passes, the method needs to normally acquire job.cond to ensure no other command is running. Since domain object lock is released during that time, asyncJob could have been started in the meantime so the method needs to recheck the first condition. Then, normal jobs set job.active and asynchronous jobs set job.asyncJob and optionally change the list of allowed job groups. Since asynchronous jobs only set job.asyncJob, other allowed commands can still be run when domain object is unlocked (when communicating to remote libvirtd or sleeping). To protect its own internal synchronous commands, the asynchronous job needs to start a special nested job before entering qemu monitor. The nested job doesn't check asyncJob, it only acquires job.cond and sets job.active to block other jobs.
2011-06-30 09:23:50 +00:00
/* Set fake job so that EnterMonitor* doesn't want to start a new one */
priv->job.active = QEMU_JOB_MODIFY;
/* Hold an extra reference because we can't allow 'vm' to be
* deleted if qemuConnectMonitor() failed */
virDomainObjRef(obj);
/* XXX check PID liveliness & EXE path */
if (qemuConnectMonitor(driver, obj) < 0)
goto error;
if (qemuUpdateActivePciHostdevs(driver, obj->def) < 0) {
goto error;
}
if (qemuProcessUpdateState(driver, obj) < 0)
goto error;
state = virDomainObjGetState(obj, &reason);
if (state == VIR_DOMAIN_SHUTOFF) {
VIR_DEBUG("Domain '%s' wasn't fully started yet, killing it",
obj->def->name);
goto error;
}
/* If upgrading from old libvirtd we won't have found any
* caps in the domain status, so re-query them
*/
if (!priv->qemuCaps &&
qemuCapsExtractVersionInfo(obj->def->emulator, obj->def->os.arch,
NULL,
&priv->qemuCaps) < 0)
goto error;
/* In case the domain was paused for shutdown while we were not running,
* we need to finish the shutdown process. And we need to do it after
* we have qemuCaps filled in.
*/
if (state == VIR_DOMAIN_PAUSED
&& reason == VIR_DOMAIN_PAUSED_SHUTTING_DOWN) {
VIR_DEBUG("Domain %s shut down while we were not running;"
" finishing shutdown sequence", obj->def->name);
qemuProcessShutdownOrReboot(obj);
goto endjob;
}
if (qemuCapsGet(priv->qemuCaps, QEMU_CAPS_DEVICE)) {
priv->persistentAddrs = 1;
if (!(priv->pciaddrs = qemuDomainPCIAddressSetCreate(obj->def)) ||
qemuAssignDevicePCISlots(obj->def, priv->pciaddrs) < 0)
goto error;
}
if (virSecurityManagerReserveLabel(driver->securityManager, obj) < 0)
goto error;
if (qemuProcessNotifyNets(obj->def) < 0)
goto error;
if (qemuProcessFiltersInstantiate(conn, obj->def))
goto error;
if (qemuDomainCheckEjectableMedia(driver, obj) < 0)
goto error;
if (qemuProcessRecoverJob(driver, obj, conn, &oldjob) < 0)
goto error;
qemu: Allow all query commands to be run during long jobs Query commands are safe to be called during long running jobs (such as migration). This patch makes them all work without the need to special-case every single one of them. The patch introduces new job.asyncCond condition and associated job.asyncJob which are dedicated to asynchronous (from qemu monitor point of view) jobs that can take arbitrarily long time to finish while qemu monitor is still usable for other commands. The existing job.active (and job.cond condition) is used all other synchronous jobs (including the commands run during async job). Locking schema is changed to use these two conditions. While asyncJob is active, only allowed set of synchronous jobs is allowed (the set can be different according to a particular asyncJob) so any method that communicates to qemu monitor needs to check if it is allowed to be executed during current asyncJob (if any). Once the check passes, the method needs to normally acquire job.cond to ensure no other command is running. Since domain object lock is released during that time, asyncJob could have been started in the meantime so the method needs to recheck the first condition. Then, normal jobs set job.active and asynchronous jobs set job.asyncJob and optionally change the list of allowed job groups. Since asynchronous jobs only set job.asyncJob, other allowed commands can still be run when domain object is unlocked (when communicating to remote libvirtd or sleeping). To protect its own internal synchronous commands, the asynchronous job needs to start a special nested job before entering qemu monitor. The nested job doesn't check asyncJob, it only acquires job.cond and sets job.active to block other jobs.
2011-06-30 09:23:50 +00:00
priv->job.active = QEMU_JOB_NONE;
/* update domain state XML with possibly updated state in virDomainObj */
if (virDomainSaveStatus(driver->caps, driver->stateDir, obj) < 0)
goto error;
if (obj->def->id >= driver->nextvmid)
driver->nextvmid = obj->def->id + 1;
endjob:
if (qemuDomainObjEndJob(driver, obj) == 0)
obj = NULL;
if (obj && virDomainObjUnref(obj) > 0)
virDomainObjUnlock(obj);
qemuDriverUnlock(driver);
virConnectClose(conn);
return;
error:
if (qemuDomainObjEndJob(driver, obj) == 0)
obj = NULL;
if (obj) {
if (!virDomainObjIsActive(obj)) {
if (virDomainObjUnref(obj) > 0)
virDomainObjUnlock(obj);
qemuDriverUnlock(driver);
return;
}
if (virDomainObjUnref(obj) > 0) {
/* We can't get the monitor back, so must kill the VM
* to remove danger of it ending up running twice if
* user tries to start it again later
*/
qemuProcessStop(driver, obj, 0, VIR_DOMAIN_SHUTOFF_FAILED);
if (!obj->persistent)
qemuDomainRemoveInactive(driver, obj);
else
virDomainObjUnlock(obj);
}
}
qemuDriverUnlock(driver);
virConnectClose(conn);
}
static void
qemuProcessReconnectHelper(void *payload,
const void *name ATTRIBUTE_UNUSED,
void *opaque)
{
virThread thread;
struct qemuProcessReconnectData *src = opaque;
struct qemuProcessReconnectData *data;
virDomainObjPtr obj = payload;
if (VIR_ALLOC(data) < 0) {
virReportOOMError();
return;
}
memcpy(data, src, sizeof(*data));
data->payload = payload;
/* This iterator is called with driver being locked.
* We create a separate thread to run qemuProcessReconnect in it.
* However, qemuProcessReconnect needs to:
* 1. lock driver
* 2. just before monitor reconnect do lightweight MonitorEnter
* (increase VM refcount, unlock VM & driver)
* 3. reconnect to monitor
* 4. do lightweight MonitorExit (lock driver & VM)
* 5. continue reconnect process
* 6. EndJob
* 7. unlock driver
*
* It is necessary to NOT hold driver lock for the entire run
* of reconnect, otherwise we will get blocked if there is
* unresponsive qemu.
* However, iterating over hash table MUST be done on locked
* driver.
*
* NB, we can't do normal MonitorEnter & MonitorExit because
* these two lock the monitor lock, which does not exists in
* this early phase.
*/
virDomainObjLock(obj);
qemuDomainObjRestoreJob(obj, &data->oldjob);
if (qemuDomainObjBeginJobWithDriver(src->driver, obj, QEMU_JOB_MODIFY) < 0)
goto error;
/* Since we close the connection later on, we have to make sure
* that the threads we start see a valid connection throughout their
* lifetime. We simply increase the reference counter here.
*/
virConnectRef(data->conn);
if (virThreadCreate(&thread, true, qemuProcessReconnect, data) < 0) {
virConnectClose(data->conn);
qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("Could not create thread. QEMU initialization "
"might be incomplete"));
if (qemuDomainObjEndJob(src->driver, obj) == 0) {
obj = NULL;
} else if (virDomainObjUnref(obj) > 0) {
/* We can't spawn a thread and thus connect to monitor.
* Kill qemu */
qemuProcessStop(src->driver, obj, 0, VIR_DOMAIN_SHUTOFF_FAILED);
if (!obj->persistent)
qemuDomainRemoveInactive(src->driver, obj);
else
virDomainObjUnlock(obj);
}
goto error;
}
virDomainObjUnlock(obj);
return;
error:
VIR_FREE(data);
}
/**
* qemuProcessReconnectAll
*
* Try to re-open the resources for live VMs that we care
* about.
*/
void
qemuProcessReconnectAll(virConnectPtr conn, struct qemud_driver *driver)
{
struct qemuProcessReconnectData data = {conn, driver};
virHashForEach(driver->domains.objs, qemuProcessReconnectHelper, &data);
}
int qemuProcessStart(virConnectPtr conn,
struct qemud_driver *driver,
virDomainObjPtr vm,
const char *migrateFrom,
bool start_paused,
bool autodestroy,
int stdin_fd,
const char *stdin_path,
virDomainSnapshotObjPtr snapshot,
enum virVMOperationType vmop)
{
int ret;
off_t pos = -1;
char ebuf[1024];
int logfile = -1;
char *timestamp;
qemuDomainObjPrivatePtr priv = vm->privateData;
virCommandPtr cmd = NULL;
struct qemuProcessHookData hookData;
unsigned long cur_balloon;
hookData.conn = conn;
hookData.vm = vm;
hookData.driver = driver;
VIR_DEBUG("Beginning VM startup process");
if (virDomainObjIsActive(vm)) {
qemuReportError(VIR_ERR_OPERATION_INVALID,
"%s", _("VM is already active"));
return -1;
}
/* Do this upfront, so any part of the startup process can add
* runtime state to vm->def that won't be persisted. This let's us
* report implicit runtime defaults in the XML, like vnc listen/socket
*/
VIR_DEBUG("Setting current domain def as transient");
if (virDomainObjSetDefTransient(driver->caps, vm, true) < 0)
goto cleanup;
vm->def->id = driver->nextvmid++;
qemuDomainSetFakeReboot(driver, vm, false);
virDomainObjSetState(vm, VIR_DOMAIN_SHUTOFF, VIR_DOMAIN_SHUTOFF_UNKNOWN);
/* Run an early hook to set-up missing devices */
if (virHookPresent(VIR_HOOK_DRIVER_QEMU)) {
char *xml = virDomainDefFormat(vm->def, 0);
int hookret;
hookret = virHookCall(VIR_HOOK_DRIVER_QEMU, vm->def->name,
VIR_HOOK_QEMU_OP_PREPARE, VIR_HOOK_SUBOP_BEGIN, NULL, xml);
VIR_FREE(xml);
/*
* If the script raised an error abort the launch
*/
if (hookret < 0)
goto cleanup;
}
/* Must be run before security labelling */
VIR_DEBUG("Preparing host devices");
if (qemuPrepareHostDevices(driver, vm->def) < 0)
goto cleanup;
VIR_DEBUG("Preparing chr devices");
if (virDomainChrDefForeach(vm->def,
true,
qemuProcessPrepareChardevDevice,
NULL) < 0)
goto cleanup;
/* If you are using a SecurityDriver with dynamic labelling,
then generate a security label for isolation */
VIR_DEBUG("Generating domain security label (if required)");
if (virSecurityManagerGenLabel(driver->securityManager, vm) < 0) {
virDomainAuditSecurityLabel(vm, false);
goto cleanup;
}
virDomainAuditSecurityLabel(vm, true);
/* Ensure no historical cgroup for this VM is lying around bogus
* settings */
VIR_DEBUG("Ensuring no historical cgroup is lying around");
qemuRemoveCgroup(driver, vm, 1);
if (vm->def->ngraphics == 1) {
if (vm->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC &&
!vm->def->graphics[0]->data.vnc.socket &&
vm->def->graphics[0]->data.vnc.autoport) {
int port = qemuProcessNextFreePort(driver, QEMU_VNC_PORT_MIN);
if (port < 0) {
qemuReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("Unable to find an unused VNC port"));
goto cleanup;
}
vm->def->graphics[0]->data.vnc.port = port;
} else if (vm->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_SPICE &&
vm->def->graphics[0]->data.spice.autoport) {
int port = qemuProcessNextFreePort(driver, QEMU_VNC_PORT_MIN);
int tlsPort = -1;
if (port < 0) {
qemuReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("Unable to find an unused SPICE port"));
goto cleanup;
}
if (driver->spiceTLS) {
tlsPort = qemuProcessNextFreePort(driver, port + 1);
if (tlsPort < 0) {
qemuReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("Unable to find an unused SPICE TLS port"));
qemuProcessReturnPort(driver, port);
goto cleanup;
}
}
vm->def->graphics[0]->data.spice.port = port;
vm->def->graphics[0]->data.spice.tlsPort = tlsPort;
}
}
if (virFileMakePath(driver->logDir) < 0) {
virReportSystemError(errno,
_("cannot create log directory %s"),
driver->logDir);
goto cleanup;
}
VIR_DEBUG("Creating domain log file");
if ((logfile = qemuDomainCreateLog(driver, vm, false)) < 0)
goto cleanup;
VIR_DEBUG("Determining emulator version");
qemuCapsFree(priv->qemuCaps);
priv->qemuCaps = NULL;
if (qemuCapsExtractVersionInfo(vm->def->emulator, vm->def->os.arch,
NULL,
&priv->qemuCaps) < 0)
goto cleanup;
if (qemuAssignDeviceAliases(vm->def, priv->qemuCaps) < 0)
goto cleanup;
VIR_DEBUG("Checking for CDROM and floppy presence");
if (qemuDomainCheckDiskPresence(driver, vm, migrateFrom != NULL) < 0)
goto cleanup;
VIR_DEBUG("Setting up domain cgroup (if required)");
if (qemuSetupCgroup(driver, vm) < 0)
goto cleanup;
if (VIR_ALLOC(priv->monConfig) < 0) {
virReportOOMError();
goto cleanup;
}
VIR_DEBUG("Preparing monitor state");
if (qemuProcessPrepareMonitorChr(driver, priv->monConfig, vm->def->name) < 0)
goto cleanup;
if (qemuCapsGet(priv->qemuCaps, QEMU_CAPS_MONITOR_JSON))
priv->monJSON = 1;
else
priv->monJSON = 0;
priv->monError = false;
priv->monStart = 0;
priv->gotShutdown = false;
VIR_FREE(priv->pidfile);
if (!(priv->pidfile = virPidFileBuildPath(driver->stateDir, vm->def->name))) {
virReportSystemError(errno,
"%s", _("Failed to build pidfile path."));
goto cleanup;
}
if (unlink(priv->pidfile) < 0 &&
errno != ENOENT) {
virReportSystemError(errno,
_("Cannot remove stale PID file %s"),
priv->pidfile);
goto cleanup;
}
/*
* Normally PCI addresses are assigned in the virDomainCreate
* or virDomainDefine methods. We might still need to assign
* some here to cope with the question of upgrades. Regardless
* we also need to populate the PCi address set cache for later
* use in hotplug
*/
if (qemuCapsGet(priv->qemuCaps, QEMU_CAPS_DEVICE)) {
VIR_DEBUG("Assigning domain PCI addresses");
/* Populate cache with current addresses */
if (priv->pciaddrs) {
qemuDomainPCIAddressSetFree(priv->pciaddrs);
priv->pciaddrs = NULL;
}
if (!(priv->pciaddrs = qemuDomainPCIAddressSetCreate(vm->def)))
goto cleanup;
/* Assign any remaining addresses */
if (qemuAssignDevicePCISlots(vm->def, priv->pciaddrs) < 0)
goto cleanup;
priv->persistentAddrs = 1;
} else {
priv->persistentAddrs = 0;
}
VIR_DEBUG("Building emulator command line");
if (!(cmd = qemuBuildCommandLine(conn, driver, vm->def, priv->monConfig,
priv->monJSON != 0, priv->qemuCaps,
migrateFrom, stdin_fd, snapshot, vmop)))
goto cleanup;
/* now that we know it is about to start call the hook if present */
if (virHookPresent(VIR_HOOK_DRIVER_QEMU)) {
char *xml = virDomainDefFormat(vm->def, 0);
int hookret;
hookret = virHookCall(VIR_HOOK_DRIVER_QEMU, vm->def->name,
VIR_HOOK_QEMU_OP_START, VIR_HOOK_SUBOP_BEGIN, NULL, xml);
VIR_FREE(xml);
/*
* If the script raised an error abort the launch
*/
if (hookret < 0)
goto cleanup;
}
if ((timestamp = virTimestamp()) == NULL) {
virReportOOMError();
goto cleanup;
} else {
if (safewrite(logfile, timestamp, strlen(timestamp)) < 0 ||
safewrite(logfile, START_POSTFIX, strlen(START_POSTFIX)) < 0) {
VIR_WARN("Unable to write timestamp to logfile: %s",
virStrerror(errno, ebuf, sizeof ebuf));
}
VIR_FREE(timestamp);
}
virCommandWriteArgLog(cmd, logfile);
qemuDomainObjCheckTaint(driver, vm, logfile);
if ((pos = lseek(logfile, 0, SEEK_END)) < 0)
VIR_WARN("Unable to seek to end of logfile: %s",
virStrerror(errno, ebuf, sizeof ebuf));
VIR_DEBUG("Clear emulator capabilities: %d",
driver->clearEmulatorCapabilities);
if (driver->clearEmulatorCapabilities)
virCommandClearCaps(cmd);
virCommandSetPreExecHook(cmd, qemuProcessHook, &hookData);
virCommandSetOutputFD(cmd, &logfile);
virCommandSetErrorFD(cmd, &logfile);
virCommandNonblockingFDs(cmd);
virCommandSetPidFile(cmd, priv->pidfile);
virCommandDaemonize(cmd);
virCommandRequireHandshake(cmd);
ret = virCommandRun(cmd, NULL);
/* wait for qemu process to show up */
if (ret == 0) {
if (virPidFileReadPath(priv->pidfile, &vm->pid) < 0) {
qemuReportError(VIR_ERR_INTERNAL_ERROR,
_("Domain %s didn't show up"), vm->def->name);
ret = -1;
}
#if 0
} else if (ret == -2) {
/*
* XXX this is bogus. It isn't safe to set vm->pid = child
* because the child no longer exists.
*/
/* The virCommand process that launches the daemon failed. Pending on
* when it failed (we can't determine for sure), there may be
* extra info in the domain log (if the hook failed for example).
*
* Pretend like things succeeded, and let 'WaitForMonitor' report
* the log contents for us.
*/
vm->pid = child;
ret = 0;
#endif
}
VIR_DEBUG("Writing early domain status to disk");
if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0) {
goto cleanup;
}
VIR_DEBUG("Waiting for handshake from child");
if (virCommandHandshakeWait(cmd) < 0) {
goto cleanup;
}
VIR_DEBUG("Setting domain security labels");
if (virSecurityManagerSetAllLabel(driver->securityManager,
vm, stdin_path) < 0)
goto cleanup;
if (stdin_fd != -1) {
/* if there's an fd to migrate from, and it's a pipe, put the
* proper security label on it
*/
struct stat stdin_sb;
VIR_DEBUG("setting security label on pipe used for migration");
if (fstat(stdin_fd, &stdin_sb) < 0) {
virReportSystemError(errno,
_("cannot stat fd %d"), stdin_fd);
goto cleanup;
}
if (S_ISFIFO(stdin_sb.st_mode) &&
virSecurityManagerSetImageFDLabel(driver->securityManager, vm, stdin_fd) < 0)
goto cleanup;
}
VIR_DEBUG("Labelling done, completing handshake to child");
if (virCommandHandshakeNotify(cmd) < 0) {
goto cleanup;
}
VIR_DEBUG("Handshake complete, child running");
if (migrateFrom)
start_paused = true;
if (ret == -1) /* The VM failed to start; tear filters before taps */
virDomainConfVMNWFilterTeardown(vm);
if (ret == -1) /* The VM failed to start */
goto cleanup;
VIR_DEBUG("Waiting for monitor to show up");
if (qemuProcessWaitForMonitor(driver, vm, priv->qemuCaps, pos) < 0)
goto cleanup;
VIR_DEBUG("Detecting VCPU PIDs");
if (qemuProcessDetectVcpuPIDs(driver, vm) < 0)
goto cleanup;
VIR_DEBUG("Setting cgroup for each VCPU(if required)");
if (qemuSetupCgroupForVcpu(driver, vm) < 0)
goto cleanup;
VIR_DEBUG("Setting VCPU affinities");
if (qemuProcessSetVcpuAffinites(conn, vm) < 0)
goto cleanup;
VIR_DEBUG("Setting any required VM passwords");
if (qemuProcessInitPasswords(conn, driver, vm) < 0)
goto cleanup;
/* If we have -device, then addresses are assigned explicitly.
* If not, then we have to detect dynamic ones here */
if (!qemuCapsGet(priv->qemuCaps, QEMU_CAPS_DEVICE)) {
VIR_DEBUG("Determining domain device PCI addresses");
if (qemuProcessInitPCIAddresses(driver, vm) < 0)
goto cleanup;
}
/* set default link states */
/* qemu doesn't support setting this on the command line, so
* enter the monitor */
VIR_DEBUG("Setting network link states");
qemuDomainObjEnterMonitorWithDriver(driver, vm);
if (qemuProcessSetLinkStates(vm) < 0) {
qemuDomainObjExitMonitorWithDriver(driver, vm);
goto cleanup;
}
qemuDomainObjExitMonitorWithDriver(driver, vm);
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
/* Technically, qemuProcessStart can be called from inside
* QEMU_ASYNC_JOB_MIGRATION_IN, but we are okay treating this like
* a sync job since no other job can call into the domain until
* migration completes. */
VIR_DEBUG("Setting initial memory amount");
cur_balloon = vm->def->mem.cur_balloon;
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
qemuDomainObjEnterMonitorWithDriver(driver, vm);
if (qemuMonitorSetBalloon(priv->mon, cur_balloon) < 0) {
qemuDomainObjExitMonitorWithDriver(driver, vm);
goto cleanup;
}
qemuDomainObjExitMonitorWithDriver(driver, vm);
if (!start_paused) {
VIR_DEBUG("Starting domain CPUs");
/* Allow the CPUS to start executing */
if (qemuProcessStartCPUs(driver, vm, conn,
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
VIR_DOMAIN_RUNNING_BOOTED,
QEMU_ASYNC_JOB_NONE) < 0) {
if (virGetLastError() == NULL)
qemuReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("resume operation failed"));
goto cleanup;
}
} else {
virDomainObjSetState(vm, VIR_DOMAIN_PAUSED,
migrateFrom ?
VIR_DOMAIN_PAUSED_MIGRATION :
VIR_DOMAIN_PAUSED_USER);
}
if (autodestroy &&
qemuProcessAutoDestroyAdd(driver, vm, conn) < 0)
goto cleanup;
VIR_DEBUG("Writing domain status to disk");
if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0)
goto cleanup;
virCommandFree(cmd);
VIR_FORCE_CLOSE(logfile);
return 0;
cleanup:
/* We jump here if we failed to start the VM for any reason, or
* if we failed to initialize the now running VM. kill it off and
* pretend we never started it */
virCommandFree(cmd);
VIR_FORCE_CLOSE(logfile);
qemuProcessStop(driver, vm, 0, VIR_DOMAIN_SHUTOFF_FAILED);
return -1;
}
void qemuProcessKill(virDomainObjPtr vm, bool gracefully)
{
int i;
VIR_DEBUG("vm=%s pid=%d gracefully=%d",
vm->def->name, vm->pid, gracefully);
if (!virDomainObjIsActive(vm)) {
VIR_DEBUG("VM '%s' not active", vm->def->name);
return;
}
/* This loop sends SIGTERM, then waits a few iterations
* (1.6 seconds) to see if it dies. If still alive then
* it does SIGKILL, and waits a few more iterations (1.6
* seconds more) to confirm that it has really gone.
*/
for (i = 0 ; i < 15 ; i++) {
int signum;
if (i == 0)
signum = SIGTERM;
else if (i == 8)
signum = SIGKILL;
else
signum = 0; /* Just check for existence */
if (virKillProcess(vm->pid, signum) < 0) {
if (errno != ESRCH) {
char ebuf[1024];
VIR_WARN("Failed to kill process %d %s",
vm->pid, virStrerror(errno, ebuf, sizeof ebuf));
}
break;
}
if (i == 0 && gracefully)
break;
usleep(200 * 1000);
}
}
void qemuProcessStop(struct qemud_driver *driver,
virDomainObjPtr vm,
int migrated,
virDomainShutoffReason reason)
{
int ret;
int retries = 0;
qemuDomainObjPrivatePtr priv = vm->privateData;
virErrorPtr orig_err;
virDomainDefPtr def;
int i;
int logfile = -1;
char *timestamp;
char ebuf[1024];
VIR_DEBUG("Shutting down VM '%s' pid=%d migrated=%d",
vm->def->name, vm->pid, migrated);
if (!virDomainObjIsActive(vm)) {
VIR_DEBUG("VM '%s' not active", vm->def->name);
return;
}
if ((logfile = qemuDomainCreateLog(driver, vm, true)) < 0) {
/* To not break the normal domain shutdown process, skip the
* timestamp log writing if failed on opening log file. */
VIR_WARN("Unable to open logfile: %s",
virStrerror(errno, ebuf, sizeof ebuf));
} else {
if ((timestamp = virTimestamp()) == NULL) {
virReportOOMError();
} else {
if (safewrite(logfile, timestamp, strlen(timestamp)) < 0 ||
safewrite(logfile, SHUTDOWN_POSTFIX,
strlen(SHUTDOWN_POSTFIX)) < 0) {
VIR_WARN("Unable to write timestamp to logfile: %s",
virStrerror(errno, ebuf, sizeof ebuf));
}
VIR_FREE(timestamp);
}
if (VIR_CLOSE(logfile) < 0)
VIR_WARN("Unable to close logfile: %s",
virStrerror(errno, ebuf, sizeof ebuf));
}
/* This method is routinely used in clean up paths. Disable error
* reporting so we don't squash a legit error. */
orig_err = virSaveLastError();
virDomainConfVMNWFilterTeardown(vm);
if (driver->macFilter) {
def = vm->def;
for (i = 0 ; i < def->nnets ; i++) {
virDomainNetDefPtr net = def->nets[i];
if (net->ifname == NULL)
continue;
if ((errno = networkDisallowMacOnPort(driver, net->ifname,
net->mac))) {
virReportSystemError(errno,
_("failed to remove ebtables rule to allow MAC address on '%s'"),
net->ifname);
}
}
}
if (priv->mon)
qemuMonitorClose(priv->mon);
if (priv->monConfig) {
if (priv->monConfig->type == VIR_DOMAIN_CHR_TYPE_UNIX)
unlink(priv->monConfig->data.nix.path);
virDomainChrSourceDefFree(priv->monConfig);
priv->monConfig = NULL;
}
/* shut it off for sure */
qemuProcessKill(vm, false);
/* Stop autodestroy in case guest is restarted */
qemuProcessAutoDestroyRemove(driver, vm);
/* now that we know it's stopped call the hook if present */
if (virHookPresent(VIR_HOOK_DRIVER_QEMU)) {
char *xml = virDomainDefFormat(vm->def, 0);
/* we can't stop the operation even if the script raised an error */
virHookCall(VIR_HOOK_DRIVER_QEMU, vm->def->name,
VIR_HOOK_QEMU_OP_STOPPED, VIR_HOOK_SUBOP_END, NULL, xml);
VIR_FREE(xml);
}
/* Reset Security Labels */
virSecurityManagerRestoreAllLabel(driver->securityManager,
vm, migrated);
virSecurityManagerReleaseLabel(driver->securityManager, vm);
/* Clear out dynamically assigned labels */
if (vm->def->seclabel.type == VIR_DOMAIN_SECLABEL_DYNAMIC) {
if (!vm->def->seclabel.baselabel)
VIR_FREE(vm->def->seclabel.model);
VIR_FREE(vm->def->seclabel.label);
}
VIR_FREE(vm->def->seclabel.imagelabel);
virDomainDefClearDeviceAliases(vm->def);
if (!priv->persistentAddrs) {
virDomainDefClearPCIAddresses(vm->def);
qemuDomainPCIAddressSetFree(priv->pciaddrs);
priv->pciaddrs = NULL;
}
qemuDomainReAttachHostDevices(driver, vm->def);
def = vm->def;
for (i = 0; i < def->nnets; i++) {
virDomainNetDefPtr net = def->nets[i];
#if WITH_MACVTAP
if (virDomainNetGetActualType(net) == VIR_DOMAIN_NET_TYPE_DIRECT) {
delMacvtap(net->ifname, net->mac,
virDomainNetGetActualDirectDev(net),
virDomainNetGetActualDirectMode(net),
virDomainNetGetActualDirectVirtPortProfile(net),
driver->stateDir);
VIR_FREE(net->ifname);
}
#endif
/* release the physical device (or any other resources used by
* this interface in the network driver
*/
networkReleaseActualDevice(net);
}
retry:
if ((ret = qemuRemoveCgroup(driver, vm, 0)) < 0) {
if (ret == -EBUSY && (retries++ < 5)) {
usleep(200*1000);
goto retry;
}
VIR_WARN("Failed to remove cgroup for %s",
vm->def->name);
}
qemuProcessRemoveDomainStatus(driver, vm);
/* Remove VNC port from port reservation bitmap, but only if it was
reserved by the driver (autoport=yes)
*/
if ((vm->def->ngraphics == 1) &&
vm->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC &&
vm->def->graphics[0]->data.vnc.autoport) {
qemuProcessReturnPort(driver, vm->def->graphics[0]->data.vnc.port);
}
if ((vm->def->ngraphics == 1) &&
vm->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_SPICE &&
vm->def->graphics[0]->data.spice.autoport) {
qemuProcessReturnPort(driver, vm->def->graphics[0]->data.spice.port);
qemuProcessReturnPort(driver, vm->def->graphics[0]->data.spice.tlsPort);
}
vm->taint = 0;
vm->pid = -1;
vm->def->id = -1;
virDomainObjSetState(vm, VIR_DOMAIN_SHUTOFF, reason);
VIR_FREE(priv->vcpupids);
priv->nvcpupids = 0;
qemuCapsFree(priv->qemuCaps);
priv->qemuCaps = NULL;
VIR_FREE(priv->pidfile);
/* The "release" hook cleans up additional resources */
if (virHookPresent(VIR_HOOK_DRIVER_QEMU)) {
char *xml = virDomainDefFormat(vm->def, 0);
/* we can't stop the operation even if the script raised an error */
virHookCall(VIR_HOOK_DRIVER_QEMU, vm->def->name,
VIR_HOOK_QEMU_OP_RELEASE, VIR_HOOK_SUBOP_END, NULL, xml);
VIR_FREE(xml);
}
if (vm->newDef) {
virDomainDefFree(vm->def);
vm->def = vm->newDef;
vm->def->id = -1;
vm->newDef = NULL;
}
if (orig_err) {
virSetError(orig_err);
virFreeError(orig_err);
}
}
int qemuProcessAttach(virConnectPtr conn ATTRIBUTE_UNUSED,
struct qemud_driver *driver,
virDomainObjPtr vm,
int pid,
const char *pidfile,
virDomainChrSourceDefPtr monConfig,
bool monJSON)
{
char ebuf[1024];
int logfile = -1;
char *timestamp;
qemuDomainObjPrivatePtr priv = vm->privateData;
bool running = true;
virDomainPausedReason reason;
virSecurityLabelPtr seclabel = NULL;
VIR_DEBUG("Beginning VM attach process");
if (virDomainObjIsActive(vm)) {
qemuReportError(VIR_ERR_OPERATION_INVALID,
"%s", _("VM is already active"));
return -1;
}
/* Do this upfront, so any part of the startup process can add
* runtime state to vm->def that won't be persisted. This let's us
* report implicit runtime defaults in the XML, like vnc listen/socket
*/
VIR_DEBUG("Setting current domain def as transient");
if (virDomainObjSetDefTransient(driver->caps, vm, true) < 0)
goto cleanup;
vm->def->id = driver->nextvmid++;
if (virFileMakePath(driver->logDir) < 0) {
virReportSystemError(errno,
_("cannot create log directory %s"),
driver->logDir);
goto cleanup;
}
VIR_FREE(priv->pidfile);
if (pidfile &&
!(priv->pidfile = strdup(pidfile)))
goto no_memory;
VIR_DEBUG("Detect security driver config");
vm->def->seclabel.type = VIR_DOMAIN_SECLABEL_STATIC;
if (VIR_ALLOC(seclabel) < 0)
goto no_memory;
if (virSecurityManagerGetProcessLabel(driver->securityManager,
vm, seclabel) < 0)
goto cleanup;
if (!(vm->def->seclabel.model = strdup(driver->caps->host.secModel.model)))
goto no_memory;
if (!(vm->def->seclabel.label = strdup(seclabel->label)))
goto no_memory;
VIR_DEBUG("Creating domain log file");
if ((logfile = qemuDomainCreateLog(driver, vm, false)) < 0)
goto cleanup;
VIR_DEBUG("Determining emulator version");
qemuCapsFree(priv->qemuCaps);
priv->qemuCaps = NULL;
if (qemuCapsExtractVersionInfo(vm->def->emulator,
vm->def->os.arch,
NULL,
&priv->qemuCaps) < 0)
goto cleanup;
VIR_DEBUG("Preparing monitor state");
priv->monConfig = monConfig;
monConfig = NULL;
priv->monJSON = monJSON;
priv->gotShutdown = false;
/*
* Normally PCI addresses are assigned in the virDomainCreate
* or virDomainDefine methods. We might still need to assign
* some here to cope with the question of upgrades. Regardless
* we also need to populate the PCi address set cache for later
* use in hotplug
*/
if (qemuCapsGet(priv->qemuCaps, QEMU_CAPS_DEVICE)) {
VIR_DEBUG("Assigning domain PCI addresses");
/* Populate cache with current addresses */
if (priv->pciaddrs) {
qemuDomainPCIAddressSetFree(priv->pciaddrs);
priv->pciaddrs = NULL;
}
if (!(priv->pciaddrs = qemuDomainPCIAddressSetCreate(vm->def)))
goto cleanup;
/* Assign any remaining addresses */
if (qemuAssignDevicePCISlots(vm->def, priv->pciaddrs) < 0)
goto cleanup;
priv->persistentAddrs = 1;
} else {
priv->persistentAddrs = 0;
}
if ((timestamp = virTimestamp()) == NULL) {
virReportOOMError();
goto cleanup;
} else {
if (safewrite(logfile, timestamp, strlen(timestamp)) < 0 ||
safewrite(logfile, ATTACH_POSTFIX, strlen(ATTACH_POSTFIX)) < 0) {
VIR_WARN("Unable to write timestamp to logfile: %s",
virStrerror(errno, ebuf, sizeof ebuf));
}
VIR_FREE(timestamp);
}
qemuDomainObjTaint(driver, vm, VIR_DOMAIN_TAINT_EXTERNAL_LAUNCH, logfile);
vm->pid = pid;
VIR_DEBUG("Waiting for monitor to show up");
if (qemuProcessWaitForMonitor(driver, vm, priv->qemuCaps, -1) < 0)
goto cleanup;
VIR_DEBUG("Detecting VCPU PIDs");
if (qemuProcessDetectVcpuPIDs(driver, vm) < 0)
goto cleanup;
/* If we have -device, then addresses are assigned explicitly.
* If not, then we have to detect dynamic ones here */
if (!qemuCapsGet(priv->qemuCaps, QEMU_CAPS_DEVICE)) {
VIR_DEBUG("Determining domain device PCI addresses");
if (qemuProcessInitPCIAddresses(driver, vm) < 0)
goto cleanup;
}
VIR_DEBUG("Getting initial memory amount");
qemu: fix crash when mixing sync and async monitor jobs Currently, we attempt to run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, and a sync job is also running at the time of the check, then priv->job.active is not QEMU_JOB_NONE. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(), and must instead put the burden on the caller to tell us when an async command wants to do a nested job. Once the burden is on the caller, then only async monitor enters need to worry about whether the VM is still running; for sync monitor enter, the internal return is always 0, so lots of ignore_value can be dropped. * src/qemu/THREADS.txt: Reflect new rules. * src/qemu/qemu_domain.h (qemuDomainObjEnterMonitorAsync): New prototype. * src/qemu/qemu_process.h (qemuProcessStartCPUs) (qemuProcessStopCPUs): Add parameter. * src/qemu/qemu_migration.h (qemuMigrationToFile): Likewise. (qemuMigrationWaitForCompletion): Make static. * src/qemu/qemu_domain.c (qemuDomainObjEnterMonitorInternal): Add parameter. (qemuDomainObjEnterMonitorAsync): New function. (qemuDomainObjEnterMonitor, qemuDomainObjEnterMonitorWithDriver): Update callers. * src/qemu/qemu_driver.c (qemuDomainSaveInternal) (qemudDomainCoreDump, doCoreDump, processWatchdogEvent) (qemudDomainSuspend, qemudDomainResume, qemuDomainSaveImageStartVM) (qemuDomainSnapshotCreateActive, qemuDomainRevertToSnapshot): Likewise. * src/qemu/qemu_process.c (qemuProcessStopCPUs) (qemuProcessFakeReboot, qemuProcessRecoverMigration) (qemuProcessRecoverJob, qemuProcessStart): Likewise. * src/qemu/qemu_migration.c (qemuMigrationToFile) (qemuMigrationWaitForCompletion, qemuMigrationUpdateJobStatus) (qemuMigrationJobStart, qemuDomainMigrateGraphicsRelocate) (doNativeMigrate, doTunnelMigrate, qemuMigrationPerformJob) (qemuMigrationPerformPhase, qemuMigrationFinish) (qemuMigrationConfirm): Likewise. * src/qemu/qemu_hotplug.c: Drop unneeded ignore_value.
2011-07-28 23:18:24 +00:00
qemuDomainObjEnterMonitorWithDriver(driver, vm);
if (qemuMonitorGetBalloonInfo(priv->mon, &vm->def->mem.cur_balloon) < 0) {
qemuDomainObjExitMonitorWithDriver(driver, vm);
goto cleanup;
}
if (qemuMonitorGetStatus(priv->mon, &running, &reason) < 0) {
qemuDomainObjExitMonitorWithDriver(driver, vm);
goto cleanup;
}
if (qemuMonitorGetVirtType(priv->mon, &vm->def->virtType) < 0) {
qemuDomainObjExitMonitorWithDriver(driver, vm);
goto cleanup;
}
qemuDomainObjExitMonitorWithDriver(driver, vm);
if (!virDomainObjIsActive(vm))
goto cleanup;
if (running)
virDomainObjSetState(vm, VIR_DOMAIN_RUNNING,
VIR_DOMAIN_RUNNING_UNPAUSED);
else
virDomainObjSetState(vm, VIR_DOMAIN_PAUSED, reason);
VIR_DEBUG("Writing domain status to disk");
if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0)
goto cleanup;
VIR_FORCE_CLOSE(logfile);
VIR_FREE(seclabel);
return 0;
no_memory:
virReportOOMError();
cleanup:
/* We jump here if we failed to start the VM for any reason, or
* if we failed to initialize the now running VM. kill it off and
* pretend we never started it */
VIR_FORCE_CLOSE(logfile);
VIR_FREE(seclabel);
virDomainChrSourceDefFree(monConfig);
return -1;
}
int qemuProcessAutoDestroyInit(struct qemud_driver *driver)
{
if (!(driver->autodestroy = virHashCreate(5, NULL)))
return -1;
return 0;
}
struct qemuProcessAutoDestroyData {
struct qemud_driver *driver;
virConnectPtr conn;
};
static void qemuProcessAutoDestroyDom(void *payload,
const void *name,
void *opaque)
{
struct qemuProcessAutoDestroyData *data = opaque;
virConnectPtr conn = payload;
const char *uuidstr = name;
unsigned char uuid[VIR_UUID_BUFLEN];
virDomainObjPtr dom;
qemuDomainObjPrivatePtr priv;
virDomainEventPtr event = NULL;
VIR_DEBUG("conn=%p uuidstr=%s thisconn=%p", conn, uuidstr, data->conn);
if (data->conn != conn)
return;
if (virUUIDParse(uuidstr, uuid) < 0) {
VIR_WARN("Failed to parse %s", uuidstr);
return;
}
if (!(dom = virDomainFindByUUID(&data->driver->domains,
uuid))) {
VIR_DEBUG("No domain object to kill");
return;
}
priv = dom->privateData;
qemu: Allow all query commands to be run during long jobs Query commands are safe to be called during long running jobs (such as migration). This patch makes them all work without the need to special-case every single one of them. The patch introduces new job.asyncCond condition and associated job.asyncJob which are dedicated to asynchronous (from qemu monitor point of view) jobs that can take arbitrarily long time to finish while qemu monitor is still usable for other commands. The existing job.active (and job.cond condition) is used all other synchronous jobs (including the commands run during async job). Locking schema is changed to use these two conditions. While asyncJob is active, only allowed set of synchronous jobs is allowed (the set can be different according to a particular asyncJob) so any method that communicates to qemu monitor needs to check if it is allowed to be executed during current asyncJob (if any). Once the check passes, the method needs to normally acquire job.cond to ensure no other command is running. Since domain object lock is released during that time, asyncJob could have been started in the meantime so the method needs to recheck the first condition. Then, normal jobs set job.active and asynchronous jobs set job.asyncJob and optionally change the list of allowed job groups. Since asynchronous jobs only set job.asyncJob, other allowed commands can still be run when domain object is unlocked (when communicating to remote libvirtd or sleeping). To protect its own internal synchronous commands, the asynchronous job needs to start a special nested job before entering qemu monitor. The nested job doesn't check asyncJob, it only acquires job.cond and sets job.active to block other jobs.
2011-06-30 09:23:50 +00:00
if (priv->job.asyncJob) {
VIR_DEBUG("vm=%s has long-term job active, cancelling",
dom->def->name);
qemuDomainObjDiscardAsyncJob(data->driver, dom);
}
qemu: Allow all query commands to be run during long jobs Query commands are safe to be called during long running jobs (such as migration). This patch makes them all work without the need to special-case every single one of them. The patch introduces new job.asyncCond condition and associated job.asyncJob which are dedicated to asynchronous (from qemu monitor point of view) jobs that can take arbitrarily long time to finish while qemu monitor is still usable for other commands. The existing job.active (and job.cond condition) is used all other synchronous jobs (including the commands run during async job). Locking schema is changed to use these two conditions. While asyncJob is active, only allowed set of synchronous jobs is allowed (the set can be different according to a particular asyncJob) so any method that communicates to qemu monitor needs to check if it is allowed to be executed during current asyncJob (if any). Once the check passes, the method needs to normally acquire job.cond to ensure no other command is running. Since domain object lock is released during that time, asyncJob could have been started in the meantime so the method needs to recheck the first condition. Then, normal jobs set job.active and asynchronous jobs set job.asyncJob and optionally change the list of allowed job groups. Since asynchronous jobs only set job.asyncJob, other allowed commands can still be run when domain object is unlocked (when communicating to remote libvirtd or sleeping). To protect its own internal synchronous commands, the asynchronous job needs to start a special nested job before entering qemu monitor. The nested job doesn't check asyncJob, it only acquires job.cond and sets job.active to block other jobs.
2011-06-30 09:23:50 +00:00
if (qemuDomainObjBeginJobWithDriver(data->driver, dom,
QEMU_JOB_DESTROY) < 0)
goto cleanup;
VIR_DEBUG("Killing domain");
qemuProcessStop(data->driver, dom, 1, VIR_DOMAIN_SHUTOFF_DESTROYED);
virDomainAuditStop(dom, "destroyed");
event = virDomainEventNewFromObj(dom,
VIR_DOMAIN_EVENT_STOPPED,
VIR_DOMAIN_EVENT_STOPPED_DESTROYED);
if (qemuDomainObjEndJob(data->driver, dom) == 0)
dom = NULL;
if (dom && !dom->persistent)
qemuDomainRemoveInactive(data->driver, dom);
cleanup:
if (dom)
virDomainObjUnlock(dom);
if (event)
qemuDomainEventQueue(data->driver, event);
virHashRemoveEntry(data->driver->autodestroy, uuidstr);
}
/*
* Precondition: driver is locked
*/
void qemuProcessAutoDestroyRun(struct qemud_driver *driver, virConnectPtr conn)
{
struct qemuProcessAutoDestroyData data = {
driver, conn
};
VIR_DEBUG("conn=%p", conn);
virHashForEach(driver->autodestroy, qemuProcessAutoDestroyDom, &data);
}
void qemuProcessAutoDestroyShutdown(struct qemud_driver *driver)
{
virHashFree(driver->autodestroy);
}
int qemuProcessAutoDestroyAdd(struct qemud_driver *driver,
virDomainObjPtr vm,
virConnectPtr conn)
{
char uuidstr[VIR_UUID_STRING_BUFLEN];
virUUIDFormat(vm->def->uuid, uuidstr);
VIR_DEBUG("vm=%s uuid=%s conn=%p", vm->def->name, uuidstr, conn);
if (virHashAddEntry(driver->autodestroy, uuidstr, conn) < 0)
return -1;
return 0;
}
int qemuProcessAutoDestroyRemove(struct qemud_driver *driver,
virDomainObjPtr vm)
{
char uuidstr[VIR_UUID_STRING_BUFLEN];
virUUIDFormat(vm->def->uuid, uuidstr);
VIR_DEBUG("vm=%s uuid=%s", vm->def->name, uuidstr);
if (virHashRemoveEntry(driver->autodestroy, uuidstr) < 0)
return -1;
return 0;
}
bool qemuProcessAutoDestroyActive(struct qemud_driver *driver,
virDomainObjPtr vm)
{
char uuidstr[VIR_UUID_STRING_BUFLEN];
virUUIDFormat(vm->def->uuid, uuidstr);
VIR_DEBUG("vm=%s uuid=%s", vm->def->name, uuidstr);
if (virHashLookup(driver->autodestroy, uuidstr) != NULL)
return true;
return false;
}