Re: [Qemu-devel] [PATCH v7 06/22] migration: Improve migration thread error handling

["Dr. David Alan Gilbert" <dgilbert@redhat.com>]

* Juan Quintela (quintela@redhat.com) wrote:
> We now report errors also when we finish migration, not only on info
> migrate.  We plan to use this error from several places, and we want
> the first error to happen to win, so we add an mutex to order it.
> 
> Signed-off-by: Juan Quintela <quintela@redhat.com>

I think this is OK for errors from the migration code itself;
I'd prefer not to lost any multiplelayers of errors we get
from devices (e.g.we see an error about a particular bit, and then
see it's from a PCI device etc)

> diff --git a/migration/ram.c b/migration/ram.c
> index e18b3e2d4f..e0179fc838 100644
> --- a/migration/ram.c
> +++ b/migration/ram.c
> @@ -1789,7 +1789,7 @@ int ram_discard_range(const char *rbname, uint64_t start, size_t length)
>      RAMBlock *rb = qemu_ram_block_by_name(rbname);
>  
>      if (!rb) {
> -        error_report("ram_discard_range: Failed to find block '%s'", rbname);
> +        error_report("ram_discard_rang0e: Failed to find block '%s'", rbname);

Typo!

Dave

>          goto err;
>      }
>  
> diff --git a/migration/tls.c b/migration/tls.c
> index 596e8790bd..026a008667 100644
> --- a/migration/tls.c
> +++ b/migration/tls.c
> @@ -119,7 +119,6 @@ static void migration_tls_outgoing_handshake(QIOTask *task,
>      if (qio_task_propagate_error(task, &err)) {
>          trace_migration_tls_outgoing_handshake_error(error_get_pretty(err));
>          migrate_fd_error(s, err);
> -        error_free(err);
>      } else {
>          trace_migration_tls_outgoing_handshake_complete();
>          migration_channel_connect(s, ioc, NULL);
> -- 
> 2.13.5
> 
--
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK