Re: [Qemu-devel] [Qemu-block] [PATCH v10 08/10] Implement new driver for block replication

[Stefan Hajnoczi]

On Fri, Sep 25, 2015 at 02:17:36PM +0800, Wen Congyang wrote:
> +static void replication_start(BlockDriverState *bs, ReplicationMode mode,
> +                              Error **errp)
> +{
> +    BDRVReplicationState *s = bs->opaque;
> +    int64_t active_length, hidden_length, disk_length;
> +    AioContext *aio_context;
> +    Error *local_err = NULL;
> +
> +    if (s->replication_state != BLOCK_REPLICATION_NONE) {
> +        error_setg(errp, "Block replication is running or done");
> +        return;
> +    }
> +
> +    if (s->mode != mode) {
> +        error_setg(errp, "The parameter mode's value is invalid, needs %d,"
> +                   " but receives %d", s->mode, mode);
> +        return;
> +    }
> +
> +    switch (s->mode) {
> +    case REPLICATION_MODE_PRIMARY:
> +        break;
> +    case REPLICATION_MODE_SECONDARY:
> +        s->active_disk = bs->file;
> +        if (!bs->file->backing_hd) {
> +            error_setg(errp, "Active disk doesn't have backing file");
> +            return;
> +        }
> +
> +        s->hidden_disk = s->active_disk->backing_hd;
> +        if (!s->hidden_disk->backing_hd) {
> +            error_setg(errp, "Hidden disk doesn't have backing file");
> +            return;
> +        }
> +
> +        s->secondary_disk = s->hidden_disk->backing_hd;
> +        if (!s->secondary_disk->blk) {
> +            error_setg(errp, "The secondary disk doesn't have block 
> backend");
> +            return;
> +        }
...
> +        aio_context = bdrv_get_aio_context(bs);
> +        aio_context_acquire(aio_context);
> +        bdrv_set_aio_context(s->secondary_disk, aio_context);

Why is this bdrv_set_aio_context() call necessary?

Child BDS nodes are in the same AioContext as their parents.  Other
block jobs need something like this because they operate on a second BDS
which is not bs' backing file chain.  I think you have a different
situation here so it's not needed.