Von meinem iPhone gesendet
Am 21.06.2013 um 22:00 schrieb Paolo Bonzini <pbonz...@redhat.com>:
> Il 20/06/2013 20:20, Peter Lieven ha scritto:
>> + for (lba = 0; lba < iscsilun->num_blocks; lba += 1 << 26) {
>> + nb_sectors = 1 << 26;
>> + if (lba + nb_sectors > iscsilun->num_blocks) {
>> + nb_sectors = iscsilun->num_blocks - lba;
>> + }
>> + nb_sectors *= (iscsilun->block_size / BDRV_SECTOR_SIZE);
>> + n = 0;
>> + ret = iscsi_co_is_allocated(bs, lba, nb_sectors, &n);
>> + if (ret || n != nb_sectors) {
>> + return 0;
>> + }
>
> I would just do lba += n in the for loop, and only exit if n == 0. The
> SCSI spec does not forbid splitting a single allocated area into
> multiple descriptors, or only returning part of an allocated area into
> the last descriptor.
>
> Otherwise looks good, but you may want to cache the result. It would
> not be 1 anymore after the first write.
>
Of course, but after some discussions with Kevin we might have found a better
Solution than extending has_zero_init this far.
Let you know soon :-)
> Paolo
