diff --git a/migration/migration.c b/migration/migration.c index d29f7448bd..5302b7b91b 100644 --- a/migration/migration.c +++ b/migration/migration.c @@ -105,7 +105,7 @@ static MigrationIncomingState *current_incoming; static GSList *migration_blockers[MIG_MODE__MAX]; static bool migration_object_check(MigrationState *ms, Error **errp); -static int migration_maybe_pause(MigrationState *s, int new_state); +static bool migration_switchover_start(MigrationState *s); static void migrate_fd_cancel(MigrationState *s); static bool close_return_path_on_source(MigrationState *s); static void migration_completion_end(MigrationState *s); @@ -2657,11 +2657,6 @@ static int postcopy_start(MigrationState *ms, Error **errp) } } - if (!migrate_pause_before_switchover()) { - migrate_set_state(&ms->state, MIGRATION_STATUS_ACTIVE, - MIGRATION_STATUS_POSTCOPY_ACTIVE); - } - trace_postcopy_start(); bql_lock(); trace_postcopy_start_set_run(); @@ -2672,10 +2667,8 @@ static int postcopy_start(MigrationState *ms, Error **errp) goto fail; } - ret = migration_maybe_pause(ms, MIGRATION_STATUS_POSTCOPY_ACTIVE); - if (ret < 0) { - error_setg_errno(errp, -ret, "%s: Failed in migration_maybe_pause()", - __func__); + if (!migration_switchover_start(ms)) { + error_setg(errp, "migration_switchover_start() failed"); goto fail; } @@ -2800,6 +2793,10 @@ static int postcopy_start(MigrationState *ms, Error **errp) */ migration_rate_set(migrate_max_postcopy_bandwidth()); + /* Now, switchover looks all fine, switching to postcopy-active */ + migrate_set_state(&ms->state, MIGRATION_STATUS_DEVICE, + MIGRATION_STATUS_POSTCOPY_ACTIVE); + bql_unlock(); return ret; @@ -2816,14 +2813,39 @@ fail: } /** - * migration_maybe_pause: Pause if required to by - * migrate_pause_before_switchover called with the BQL locked - * Returns: 0 on success + * @migration_switchover_start: Start VM switchover procedure + * + * @s: The migration state object pointer + * + * Prepares for the switchover, depending on "pause-before-switchover" + * capability. + * + * If cap set, state machine goes like: + * [postcopy-]active -> pre-switchover -> device + * + * If cap not set: + * [postcopy-]active -> device + * + * Returns: true on success, false on interruptions. */ -static int migration_maybe_pause(MigrationState *s, int new_state) +static bool migration_switchover_start(MigrationState *s) { + /* Concurrent cancellation? Quit */ + if (s->state == MIGRATION_STATUS_CANCELLING) { + return false; + } + + /* + * No matter precopy or postcopy, since we still hold BQL it must not + * change concurrently to CANCELLING, so it must be either ACTIVE or + * POSTCOPY_ACTIVE. + */ + assert(migration_is_active()); + + /* If the pre stage not requested, directly switch to DEVICE */ if (!migrate_pause_before_switchover()) { - return 0; + migrate_set_state(&s->state, s->state, MIGRATION_STATUS_DEVICE); + return true; } /* Since leaving this state is not atomic with posting the semaphore @@ -2836,23 +2858,22 @@ static int migration_maybe_pause(MigrationState *s, int new_state) /* This block intentionally left blank */ } - /* - * If the migration is cancelled when it is in the completion phase, - * the migration state is set to MIGRATION_STATUS_CANCELLING. - * So we don't need to wait a semaphore, otherwise we would always - * wait for the 'pause_sem' semaphore. - */ - if (s->state != MIGRATION_STATUS_CANCELLING) { - migrate_set_state(&s->state, s->state, - MIGRATION_STATUS_PRE_SWITCHOVER); - bql_unlock(); - qemu_sem_wait(&s->pause_sem); - bql_lock(); - migrate_set_state(&s->state, MIGRATION_STATUS_PRE_SWITCHOVER, - new_state); - } + /* Update [POSTCOPY_]ACTIVE to PRE_SWITCHOVER */ + migrate_set_state(&s->state, s->state, MIGRATION_STATUS_PRE_SWITCHOVER); + bql_unlock(); - return s->state == new_state ? 0 : -EINVAL; + qemu_sem_wait(&s->pause_sem); + + bql_lock(); + /* + * After BQL released and retaken, the state can be CANCELLING if it + * happend during sem_wait().. Only change the state if it's still + * pre-switchover. + */ + migrate_set_state(&s->state, MIGRATION_STATUS_PRE_SWITCHOVER, + MIGRATION_STATUS_DEVICE); + + return s->state == MIGRATION_STATUS_DEVICE; } static int migration_completion_precopy(MigrationState *s) @@ -2868,8 +2889,7 @@ static int migration_completion_precopy(MigrationState *s) } } - ret = migration_maybe_pause(s, MIGRATION_STATUS_DEVICE); - if (ret < 0) { + if (!migration_switchover_start(s)) { goto out_unlock; } diff --git a/qapi/migration.json b/qapi/migration.json index 4679ce9f2a..43babd1df4 100644 --- a/qapi/migration.json +++ b/qapi/migration.json @@ -158,8 +158,11 @@ # # @pre-switchover: Paused before device serialisation. (since 2.11) # -# @device: During device serialisation when pause-before-switchover is -# enabled (since 2.11) +# @device: During device serialisation (also known as switchover phase). +# Before 9.2, this is only used when (1) in precopy, and (2) when +# pre-switchover capability is enabled. After 10.0, this state will +# always be present for every migration procedure as the switchover +# phase. (since 2.11) # # @wait-unplug: wait for device unplug request by guest OS to be # completed. (since 4.2) diff --git a/tests/qemu-iotests/194.out b/tests/qemu-iotests/194.out index 376ed1d2e6..6940e809cd 100644 --- a/tests/qemu-iotests/194.out +++ b/tests/qemu-iotests/194.out @@ -14,6 +14,7 @@ Starting migration... {"return": {}} {"data": {"status": "setup"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} {"data": {"status": "active"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} +{"data": {"status": "device"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} {"data": {"status": "completed"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} Gracefully ending the `drive-mirror` job on source... {"return": {}} diff --git a/tests/qemu-iotests/203.out b/tests/qemu-iotests/203.out index 9d4abba8c5..8e58705e51 100644 --- a/tests/qemu-iotests/203.out +++ b/tests/qemu-iotests/203.out @@ -8,4 +8,5 @@ Starting migration... {"return": {}} {"data": {"status": "setup"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} {"data": {"status": "active"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} +{"data": {"status": "device"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} {"data": {"status": "completed"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} diff --git a/tests/qemu-iotests/234.out b/tests/qemu-iotests/234.out index ac8b64350c..be3e138b58 100644 --- a/tests/qemu-iotests/234.out +++ b/tests/qemu-iotests/234.out @@ -10,6 +10,7 @@ Starting migration to B... {"return": {}} {"data": {"status": "setup"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} {"data": {"status": "active"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} +{"data": {"status": "device"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} {"data": {"status": "completed"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} {"data": {"status": "active"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} {"data": {"status": "completed"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} @@ -27,6 +28,7 @@ Starting migration back to A... {"return": {}} {"data": {"status": "setup"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} {"data": {"status": "active"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} +{"data": {"status": "device"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} {"data": {"status": "completed"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} {"data": {"status": "active"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} {"data": {"status": "completed"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} diff --git a/tests/qemu-iotests/262.out b/tests/qemu-iotests/262.out index b8a2d3598d..bd7706b84b 100644 --- a/tests/qemu-iotests/262.out +++ b/tests/qemu-iotests/262.out @@ -8,6 +8,7 @@ Starting migration to B... {"return": {}} {"data": {"status": "setup"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} {"data": {"status": "active"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} +{"data": {"status": "device"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} {"data": {"status": "completed"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} {"data": {"status": "active"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} {"data": {"status": "completed"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} diff --git a/tests/qemu-iotests/280.out b/tests/qemu-iotests/280.out index 546dbb4a68..37411144ca 100644 --- a/tests/qemu-iotests/280.out +++ b/tests/qemu-iotests/280.out @@ -7,6 +7,7 @@ Enabling migration QMP events on VM... {"return": {}} {"data": {"status": "setup"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} {"data": {"status": "active"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} +{"data": {"status": "device"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} {"data": {"status": "completed"}, "event": "MIGRATION", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} VM is now stopped: diff --git a/tests/qtest/libqos/libqos.c b/tests/qtest/libqos/libqos.c index 5c0fa1f7c5..28a0901a0a 100644 --- a/tests/qtest/libqos/libqos.c +++ b/tests/qtest/libqos/libqos.c @@ -117,13 +117,14 @@ void migrate(QOSState *from, QOSState *to, const char *uri) g_assert(qdict_haskey(sub, "status")); st = qdict_get_str(sub, "status"); - /* "setup", "active", "completed", "failed", "cancelled" */ + /* "setup", "active", "device", "completed", "failed", "cancelled" */ if (strcmp(st, "completed") == 0) { qobject_unref(rsp); break; } if ((strcmp(st, "setup") == 0) || (strcmp(st, "active") == 0) + || (strcmp(st, "device") == 0) || (strcmp(st, "wait-unplug") == 0)) { qobject_unref(rsp); g_usleep(5000);