Re: [Qemu-devel] [PATCH 07/20] block: allow customizing the granularity of the dirty bitmap

[Eric Blake]

On 12/12/2012 06:46 AM, Paolo Bonzini wrote:
> Signed-off-by: Paolo Bonzini <address@hidden>
> ---
>  block-migration.c |  5 +++--
>  block.c           | 17 ++++++++++-------
>  block.h           |  5 +----
>  block/mirror.c    | 14 ++++----------
>  qapi-schema.json  |  4 +++-
>  5 files changed, 21 insertions(+), 24 deletions(-)
> 

> @@ -4218,16 +4220,17 @@ void *qemu_blockalign(BlockDriverState *bs, size_t 
> size)
>      return qemu_memalign((bs && bs->buffer_alignment) ? bs->buffer_alignment 
> : 512, size);
>  }
>  
> -void bdrv_set_dirty_tracking(BlockDriverState *bs, int enable)
> +void bdrv_set_dirty_tracking(BlockDriverState *bs, int granularity)
>  {
>      int64_t bitmap_size;
>  
> -    if (enable) {
> -        if (!bs->dirty_bitmap) {
> -            bitmap_size = (bdrv_getlength(bs) >> BDRV_SECTOR_BITS);
> -            bs->dirty_bitmap = hbitmap_alloc(bitmap_size,
> -                                             
> BDRV_LOG_SECTORS_PER_DIRTY_CHUNK);
> -        }
> +    assert((granularity & (granularity - 1)) == 0);
> +
> +    if (granularity) {
> +        granularity >>= BDRV_SECTOR_BITS;

Given that granularity is specified in bytes, does it make sense for a
user to want a granularity of 4G?  I guess another way of wording this
question is: Are you sure 'int granularity' is the right type?

> +++ b/qapi-schema.json
> @@ -667,10 +667,12 @@
>  #
>  # @count: number of dirty bytes according to the dirty bitmap
>  #
> +# @granularity: granularity of the dirty bitmap in bytes

Mention that this field was added in 1.4.

> +#
>  # Since: 1.3
>  ##
>  { 'type': 'BlockDirtyInfo',
> -  'data': {'count': 'int'} }
> +  'data': {'count': 'int', 'granularity': 'int'} }
>  
>  ##
>  # @BlockInfo:
> 

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org

signature.asc
Description: OpenPGP digital signature