On Mon, Feb 19, 2024 at 04:44:57PM -0300, Fabiano Rosas wrote:
> The QMP command query_migrate might see incorrect throughput numbers
> if it runs after we've set the migration completion status but before
> migration_calculate_complete() has updated s->total_time and s->mbps.
>
> The migration status would show COMPLETED, but the throughput value
> would be the one from the last iteration and not the one from the
> whole migration. This will usually be a larger value due to the time
> period being smaller (one iteration).
>
> Move migration_calculate_complete() earlier so that the status
> MIGRATION_STATUS_COMPLETED is only emitted after the final counters
> update.
>
> Signed-off-by: Fabiano Rosas <faro...@suse.de>
> ---
> CI run: https://gitlab.com/farosas/qemu/-/pipelines/1182405776
> ---
> migration/migration.c | 10 ++++++----
> 1 file changed, 6 insertions(+), 4 deletions(-)
>
> diff --git a/migration/migration.c b/migration/migration.c
> index ab21de2cad..7486d59da0 100644
> --- a/migration/migration.c
> +++ b/migration/migration.c
> @@ -102,6 +102,7 @@ static int migration_maybe_pause(MigrationState *s,
> int new_state);
> static void migrate_fd_cancel(MigrationState *s);
> static bool close_return_path_on_source(MigrationState *s);
> +static void migration_calculate_complete(MigrationState *s);
>
> static void migration_downtime_start(MigrationState *s)
> {
> @@ -2746,6 +2747,7 @@ static void migration_completion(MigrationState *s)
> migrate_set_state(&s->state, MIGRATION_STATUS_ACTIVE,
> MIGRATION_STATUS_COLO);
> } else {
> + migration_calculate_complete(s);
> migrate_set_state(&s->state, current_active_state,
> MIGRATION_STATUS_COMPLETED);
> }
> @@ -2784,6 +2786,7 @@ static void bg_migration_completion(MigrationState *s)
> goto fail;
> }
>
> + migration_calculate_complete(s);
> migrate_set_state(&s->state, current_active_state,
> MIGRATION_STATUS_COMPLETED);
> return;
> @@ -2993,12 +2996,15 @@ static void
> migration_calculate_complete(MigrationState *s)
> int64_t end_time = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
> int64_t transfer_time;
>
> + /* QMP could read from these concurrently */
> + bql_lock();
> migration_downtime_end(s);
> s->total_time = end_time - s->start_time;
> transfer_time = s->total_time - s->setup_time;
> if (transfer_time) {
> s->mbps = ((double) bytes * 8.0) / transfer_time / 1000;
> }
> + bql_unlock();
The lock is not needed?
AFAIU that was needed because of things like runstate_set() rather than
setting of these fields.
See migration_update_counters() where it also updates mbps without holding
a lock.
Other than that it looks all good.
> }
>
> static void update_iteration_initial_status(MigrationState *s)
> @@ -3145,7 +3151,6 @@ static void migration_iteration_finish(MigrationState
> *s)
> bql_lock();
> switch (s->state) {
> case MIGRATION_STATUS_COMPLETED:
> - migration_calculate_complete(s);
> runstate_set(RUN_STATE_POSTMIGRATE);
> break;
> case MIGRATION_STATUS_COLO:
> @@ -3189,9 +3194,6 @@ static void
> bg_migration_iteration_finish(MigrationState *s)
> bql_lock();
> switch (s->state) {
> case MIGRATION_STATUS_COMPLETED:
> - migration_calculate_complete(s);
> - break;
> -
> case MIGRATION_STATUS_ACTIVE:
> case MIGRATION_STATUS_FAILED:
> case MIGRATION_STATUS_CANCELLED:
> --
> 2.35.3
>
--
Peter Xu
