On 24.06.21 11:15, Vladimir Sementsov-Ogievskiy wrote:
23.06.2021 18:01, Max Reitz wrote:
.bdrv_co_block_status() implementations are free to return a *pnum that
exceeds @bytes, because bdrv_co_block_status() in block/io.c will clamp
*pnum as necessary.
On the other hand, if drivers' implementations return values for *pnum
that are as large as possible, our recently introduced block-status
cache will become more effective.
So, make a note in block_int.h that @bytes is no upper limit for *pnum.
Suggested-by: Eric Blake <ebl...@redhat.com>
Signed-off-by: Max Reitz <mre...@redhat.com>
---
include/block/block_int.h | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/include/block/block_int.h b/include/block/block_int.h
index fcb599dd1c..f85b96ed99 100644
--- a/include/block/block_int.h
+++ b/include/block/block_int.h
@@ -347,6 +347,11 @@ struct BlockDriver {
* clamped to bdrv_getlength() and aligned to request_alignment,
* as well as non-NULL pnum, map, and file; in turn, the driver
* must return an error or set pnum to an aligned non-zero value.
+ *
+ * Note that @bytes is just a hint on how big of a region the
+ * caller wants to inspect. It is not a limit on *pnum.
+ * Implementations are free to return larger values of *pnum if
+ * doing so does not incur a performance penalty.
Worth mention that the cache will benefit of it?
Oh, right, absolutely. Like so:
"block/io.c's bdrv_co_block_status() will clamp *pnum before returning
it to its caller, but it itself can still make use of the unclamped
*pnum value. Specifically, the block-status cache for protocol nodes
will benefit from storing as large a region as possible."
?
Max
