On 09/08/2017 08:27 AM, Kevin Wolf wrote: > Am 30.08.2017 um 23:05 hat Eric Blake geschrieben: >> Now that we have adjusted the majority of the calls this function >> makes to be byte-based, it is easier to read the code if it makes >> passes over the image using bytes rather than sectors. >> >> Signed-off-by: Eric Blake <ebl...@redhat.com> >> Reviewed-by: John Snow <js...@redhat.com> >> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com> >>
>>
>> - while ((sector = bdrv_dirty_iter_next(dbi) >> BDRV_SECTOR_BITS) != -1) {
>> - uint64_t cluster = sector / sbc;
>> + while ((offset = bdrv_dirty_iter_next(dbi)) != -1) {
>
> Don't you have to multiply both sides of the equation? This would be
> offset != -512, which points out that the previous patch to convert
> bdrv_dirty_iter_next() to byte-based gave it a really awkward interface.
I think what I really need to do is change '!= -1' to '< 0', as that's
much easier to reason about when scaling is present.
>
>> + uint64_t cluster = offset / limit;
>> uint64_t end, write_size;
>> int64_t off;
>>
>> - sector = cluster * sbc;
>> - end = MIN(bm_sectors, sector + sbc);
>> - write_size = bdrv_dirty_bitmap_serialization_size(bitmap,
>> - sector * BDRV_SECTOR_SIZE, (end - sector) * BDRV_SECTOR_SIZE);
>> + offset = cluster * limit;
>
> You just had cluster = offset / limit, so in other words, align down
> offset? If so, this is how it should be written.
Thanks for the close reviews; looks like I have enough things to do a
respin.
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3266
Virtualization: qemu.org | libvirt.org
signature.asc
Description: OpenPGP digital signature
