Re: [Qemu-block] [Qemu-devel] [PATCH 15/25] qcow2: add .bdrv_can_store_new_dirty_bitmap

[John Snow]

On 02/14/2017 11:59 AM, Vladimir Sementsov-Ogievskiy wrote:
> Realize .bdrv_can_store_new_dirty_bitmap interface.
> 
> Signed-off-by: Vladimir Sementsov-Ogievskiy <address@hidden>
> Reviewed-by: Max Reitz <address@hidden>
> ---
>  block/qcow2-bitmap.c | 40 ++++++++++++++++++++++++++++++++++++++++
>  block/qcow2.c        |  1 +
>  block/qcow2.h        |  4 ++++
>  3 files changed, 45 insertions(+)
> 
> diff --git a/block/qcow2-bitmap.c b/block/qcow2-bitmap.c
> index b177a95..419d51e 100644
> --- a/block/qcow2-bitmap.c
> +++ b/block/qcow2-bitmap.c
> @@ -1182,3 +1182,43 @@ fail:
>  
>      bitmap_list_free(bm_list);
>  }
> +
> +bool qcow2_can_store_new_dirty_bitmap(BlockDriverState *bs,
> +                                      const char *name,
> +                                      uint32_t granularity,
> +                                      Error **errp)
> +{
> +    BDRVQcow2State *s = bs->opaque;
> +    const char *reason = NULL;
> +    bool found;
> +    Qcow2BitmapList *bm_list;
> +
> +    if (check_constraints_on_bitmap(bs, name, granularity) != 0) {
> +        reason = "it doesn't satisfy the constraints";
> +        goto common_errp;
> +    }

"the constraints" are still entirely opaque to the caller, it may not be
obvious that it's because of the name length.

(But maybe you fix this in patch 25.)

> +
> +    if (s->nb_bitmaps == 0) {
> +        return true;
> +    }
> +
> +    bm_list = bitmap_list_load(bs, s->bitmap_directory_offset,
> +                               s->bitmap_directory_size, errp);
> +    if (bm_list == NULL) {
> +        return false;
> +    }

Another opportunity for caching the bm_list.

> +
> +    found = find_bitmap_by_name(bm_list, name);
> +    bitmap_list_free(bm_list);
> +    if (found) {
> +        reason = "bitmap with the same name is already stored";
> +        goto common_errp;
> +    }
> +
> +    return true;
> +
> +common_errp:
> +    error_setg(errp, "Can't make bitmap '%s' persistent in '%s', as %s.",
> +               name, bdrv_get_device_or_node_name(bs), reason);
> +    return false;
> +}

We don't check total numbers of bitmaps here? (I'll keep reading ahead
in the series.)

> diff --git a/block/qcow2.c b/block/qcow2.c
> index d0e41bf..6e1fe53 100644
> --- a/block/qcow2.c
> +++ b/block/qcow2.c
> @@ -3541,6 +3541,7 @@ BlockDriver bdrv_qcow2 = {
>  
>      .bdrv_load_autoloading_dirty_bitmaps = 
> qcow2_load_autoloading_dirty_bitmaps,
>      .bdrv_store_persistent_dirty_bitmaps = 
> qcow2_store_persistent_dirty_bitmaps,
> +    .bdrv_can_store_new_dirty_bitmap = qcow2_can_store_new_dirty_bitmap,
>  };
>  
>  static void bdrv_qcow2_init(void)
> diff --git a/block/qcow2.h b/block/qcow2.h
> index d9a7643..749710d 100644
> --- a/block/qcow2.h
> +++ b/block/qcow2.h
> @@ -616,5 +616,9 @@ void qcow2_cache_put(BlockDriverState *bs, Qcow2Cache *c, 
> void **table);
>  /* qcow2-bitmap.c functions */
>  void qcow2_load_autoloading_dirty_bitmaps(BlockDriverState *bs, Error 
> **errp);
>  void qcow2_store_persistent_dirty_bitmaps(BlockDriverState *bs, Error 
> **errp);
> +bool qcow2_can_store_new_dirty_bitmap(BlockDriverState *bs,
> +                                      const char *name,
> +                                      uint32_t granularity,
> +                                      Error **errp);
>  
>  #endif
> 

Benefit of the doubt:

Reviewed-by: John Snow <address@hidden>