Re: [Qemu-devel] [PATCH 5/6] block: Add backing passthrough implementations for copy_range

[Stefan Hajnoczi]

On Fri, Jun 08, 2018 at 02:04:16PM +0800, Fam Zheng wrote:
> +int coroutine_fn bdrv_co_copy_range_from_backing(BlockDriverState *bs,
> +                                                 BdrvChild *src, uint64_t 
> src_offset,
> +                                                 BdrvChild *dst, uint64_t 
> dst_offset,
> +                                                 uint64_t bytes, 
> BdrvRequestFlags flags)
> +{
> +    if (!src->bs) {
> +        return -ENOMEDIUM;
> +    }
> +    return bdrv_co_copy_range_from(src->bs->backing, src_offset, dst,
> +                                   dst_offset, bytes, flags);
> +}
> +
> +int coroutine_fn bdrv_co_copy_range_to_backing(BlockDriverState *bs,
> +                                               BdrvChild *src, uint64_t 
> src_offset,
> +                                               BdrvChild *dst, uint64_t 
> dst_offset,
> +                                               uint64_t bytes, 
> BdrvRequestFlags flags)
> +{
> +    if (!dst->bs) {
> +        return -ENOMEDIUM;
> +    }
> +    return bdrv_co_copy_range_to(src, src_offset, dst->bs->backing,
> +                                 dst_offset, bytes, flags);
> +}

If src->bs or dst->bs were NULL, then bdrv_co_copy_range() would have
already crashed in bdrv_inc_in_flight(src/dst_bs).  Should
.bdrv_co_copy_range_to/from() implementations really check for
ENOMEDIUM?

signature.asc
Description: PGP signature