02.10.2017 18:43, Eric Blake wrote:
On 10/02/2017 09:39 AM, Vladimir Sementsov-Ogievskiy wrote:
The function searches for next zero bit.
Also add interface for BdrvDirtyBitmap.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
---
include/block/dirty-bitmap.h | 1 +
include/qemu/hbitmap.h | 8 ++++++++
block/dirty-bitmap.c | 5 +++++
util/hbitmap.c | 29 +++++++++++++++++++++++++++++
4 files changed, 43 insertions(+)
+++ b/block/dirty-bitmap.c
@@ -715,3 +715,8 @@ char *bdrv_dirty_bitmap_sha256(const BdrvDirtyBitmap
*bitmap, Error **errp)
{
return hbitmap_sha256(bitmap->bitmap, errp);
}
+
+int64_t bdrv_dirty_bitmap_next_zero(BdrvDirtyBitmap *bitmap, uint64_t offset)
+{
+ return hbitmap_next_zero(bitmap->bitmap, offset);
+}
Returns an answer in the same scale as the underlying hbitmap; if this
is applied before my byte-based dirty bitmap series, that means offset
is a sector count and the result is likewise a sector number (awkward);
if this is applied after my series, you pass in a byte offset start and
get a byte result (nice).
Oh, I forget to say it: it's based on your series)
I don't see any obvious errors in the implementation, but DO think that
you should include a testsuite enhancement in tests/test-hbitmap.c to
cover the new functionality before we accept this.
--
Best regards,
Vladimir
