On Tue, Jul 04, 2017 at 07:49:15PM +0100, Dr. David Alan Gilbert (git) wrote:
> From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
>
> When we issue a cancel and clean up the RDMA channel
> send a CONTROL_ERROR to get the destination to quit.
>
> The rdma_cleanup code waits for the event to come back
> from the rdma_disconnect; but that wont happen until the
> destination quits and there's currently nothing to force
> it.
>
> Note this makes the case of a cancel work while the destination
> is alive, and it already works if the destination is
> truly dead. Note it doesn't fix the case where the destination
> is hung (we get stuck waiting for the rdma_disconnect event).
>
> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Looks like we'll print this as well when we cancel the migration
(before sending the RDMA_CONTROL_ERROR):
error_report("Early error. Sending error.");
But I don't think it really matters. So:
Reviewed-by: Peter Xu <peterx@redhat.com>
> ---
> migration/rdma.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/migration/rdma.c b/migration/rdma.c
> index bfb0a43740..3d17db3a23 100644
> --- a/migration/rdma.c
> +++ b/migration/rdma.c
> @@ -2260,7 +2260,9 @@ static void qemu_rdma_cleanup(RDMAContext *rdma)
> int ret, idx;
>
> if (rdma->cm_id && rdma->connected) {
> - if (rdma->error_state && !rdma->received_error) {
> + if ((rdma->error_state ||
> + migrate_get_current()->state == MIGRATION_STATUS_CANCELLING) &&
> + !rdma->received_error) {
> RDMAControlHeader head = { .len = 0,
> .type = RDMA_CONTROL_ERROR,
> .repeat = 1,
> --
> 2.13.0
>
--
Peter Xu