Re: [PATCH 1/2] Fixed a QEMU hang when guest poweroff in COLO mode

[Juan Quintela <quintela@redhat.com>]

"Rao, Lei" <lei.rao@intel.com> wrote:
> From: "Rao, Lei" <lei.rao@intel.com>
>
> When the PVM guest poweroff, the COLO thread may wait a semaphore
> in colo_process_checkpoint().So, we should wake up the COLO thread
> before migration shutdown.
>
> Signed-off-by: Lei Rao <lei.rao@intel.com>

A couple of notes.

> ---
>  include/migration/colo.h |  1 +
>  migration/colo.c         | 14 ++++++++++++++
>  migration/migration.c    | 10 ++++++++++
>  3 files changed, 25 insertions(+)
>
> diff --git a/include/migration/colo.h b/include/migration/colo.h
> index 768e1f0..525b45a 100644
> --- a/include/migration/colo.h
> +++ b/include/migration/colo.h
> @@ -37,4 +37,5 @@ COLOMode get_colo_mode(void);
>  void colo_do_failover(void);
>  
>  void colo_checkpoint_notify(void *opaque);
> +void colo_shutdown(COLOMode mode);
>  #endif
> diff --git a/migration/colo.c b/migration/colo.c
> index 2415325..385c1d7 100644
> --- a/migration/colo.c
> +++ b/migration/colo.c
> @@ -820,6 +820,20 @@ static void colo_wait_handle_message(MigrationIncomingState *mis,
>      }
>  }
>  
> +void colo_shutdown(COLOMode mode)
> +{
> +    if (mode == COLO_MODE_PRIMARY) {
> +        MigrationState *s = migrate_get_current();
> +
> +        qemu_event_set(&s->colo_checkpoint_event);
> +        qemu_sem_post(&s->colo_exit_sem);
> +    } else {
> +        MigrationIncomingState *mis = migration_incoming_get_current();
> +
> +        qemu_sem_post(&mis->colo_incoming_sem);
> +    }
> +}
> +
>  void *colo_process_incoming_thread(void *opaque)
>  {
>      MigrationIncomingState *mis = opaque;
> diff --git a/migration/migration.c b/migration/migration.c
> index abaf6f9..9df6328 100644
> --- a/migration/migration.c
> +++ b/migration/migration.c
> @@ -225,6 +225,16 @@ void migration_cancel(const Error *error)
>  
>  void migration_shutdown(void)
>  {
> +    COLOMode mode = get_colo_mode();
> +
> +    /*
> +     * When the QEMU main thread exit, the COLO thread
> +     * may wait a semaphore. So, we should wakeup the
> +     * COLO thread before migration shutdown.
> +     */
> +    if (mode != COLO_MODE_NONE) {
> +        colo_shutdown(mode);
> +     }

don't put the code on the main path.

Could you just put all of this inside a colo_shutdown() call?

Thanks, Juan.

>      /*
>       * Cancel the current migration - that will (eventually)
>       * stop the migration using this structure