On Mon 09 Oct 2017 05:07:56 PM CEST, Max Reitz wrote:
> On 2017-10-09 16:48, Alberto Garcia wrote:
>> BDRV_SECTOR_BITS is defined to be 9 in block.h (and BDRV_SECTOR_SIZE
>> is calculated from that), but there are still a few placed where we
>> are using the literal value instead of the macro.
>>
>> Signed-off-by: Alberto Garcia <be...@igalia.com>
>> ---
>> block/qcow2-cluster.c | 6 +++---
>> block/qcow2.c | 4 ++--
>> 2 files changed, 5 insertions(+), 5 deletions(-)
>>
>> diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c
>> index 0e5aec81cb..0a63604af6 100644
>> --- a/block/qcow2-cluster.c
>> +++ b/block/qcow2-cluster.c
>> @@ -748,8 +748,8 @@ uint64_t
>> qcow2_alloc_compressed_cluster_offset(BlockDriverState *bs,
>> return 0;
>> }
>>
>> - nb_csectors = ((cluster_offset + compressed_size - 1) >> 9) -
>> - (cluster_offset >> 9);
>> + nb_csectors = ((cluster_offset + compressed_size - 1) >>
>> BDRV_SECTOR_BITS) -
>> + (cluster_offset >> BDRV_SECTOR_BITS);
>
> I'm not sure about this one, because technically this is not the block
> layer's sector size but actually 512 bytes ("Compressed size of the
> images in sectors of 512 bytes" as per the spec).
You're right, I'll resend the patch leaving this part out for now.
Berto
