Re: [Qemu-devel] [PATCH v4 3/8] mirror: Do zero write on target if sectors not allocated

[Paolo Bonzini <pbonzini@redhat.com>]

On 25/05/2015 16:36, Paolo Bonzini wrote:
> 
> 
> On 22/05/2015 05:40, Fam Zheng wrote:
>> +    ret = bdrv_get_block_status(source, NULL, sector_num, nb_sectors, &pnum);
>> +    if (ret < 0 || pnum < nb_sectors ||
>> +            (ret & BDRV_BLOCK_ALLOCATED && !(ret & BDRV_BLOCK_ZERO))) {
>> +        bdrv_aio_readv(source, sector_num, &op->qiov, nb_sectors,
>> +                       mirror_read_complete, op);
>> +    } else if (ret & BDRV_BLOCK_ZERO) {
>> +        bdrv_aio_write_zeroes(s->target, sector_num, op->nb_sectors,
>> +                              s->unmap ? BDRV_REQ_MAY_UNMAP : 0,
>> +                              mirror_write_complete, op);
>> +    } else {
>> +        assert(!(ret & BDRV_BLOCK_ALLOCATED));
>> +        bdrv_aio_discard(s->target, sector_num, op->nb_sectors,
>> +                         mirror_write_complete, op);
>> +    }
> 
> This doesn't work if you have a backing file.  You want to test
> BDRV_BLOCK_DATA, not BDRV_BLOCK_ALLOCATED.
> 
> On the other hand, if BDRV_BLOCK_ALLOCATED is nonzero, you need to
> recurse on bs->backing_hd.  The logic is very similar to
> bdrv_is_allocated_above, but you need to write bdrv_get_block_status_above.

Oops, I totally missed the "NULL" in the first line.  Still, I think
BDRV_BLOCK_DATA is a better check than BDRV_BLOCK_ALLOCATED.

Paolo