On 05.05.21 22:34, Vladimir Sementsov-Ogievskiy wrote:
05.05.2021 19:23, Max Reitz wrote:
On 04.05.21 10:25, Vladimir Sementsov-Ogievskiy wrote:
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsement...@virtuozzo.com>
---
block/write-threshold.c | 3 ---
1 file changed, 3 deletions(-)
diff --git a/block/write-threshold.c b/block/write-threshold.c
index fbf4e6f5c4..db271c5537 100644
--- a/block/write-threshold.c
+++ b/block/write-threshold.c
@@ -12,10 +12,7 @@
*/
#include "qemu/osdep.h"
-#include "block/block_int.h"
We need this (for bs->write_threshold_offset), but it’s included
through block/block_int.h. I’m not sure whether we should drop it
from the includes.
Perhaps we should instead drop block_int.h from write-threshold.h?
Shouldn’t including qemu/osdep.h (which includes qemu/typedefs.h,
which forward-declares BDS) be sufficient there?
-#include "qemu/coroutine.h"
#include "block/write-threshold.h"
-#include "qemu/notify.h"
#include "qapi/error.h"
#include "qapi/qapi-commands-block-core.h"
#include "qapi/qapi-events-block-core.h"
Btw, where does qemu/atomic.h come in? Looks like it comes through
block_int.h. I think we should include it explicitly.
I'm not sure. If something is included through another include, why to
include it explicitly?
Because the other include may change. I’d include something if you need
the feature, and we need atomics here.
And block_int.h doesn’t even provide atomic.h, it comes through various
of its includes (which are probably not included to provide atomics).
So this is already indirect and basically just incidental; block_int.h
doesn’t guarantee atomic.h.
Another thing: I see that other fields in BDS that are to be accessed
with atomic ops have a comment saying so. I think
write_threshold_offset should have, too.
Max
For me, if statement can be removed with no effect, it's an extra
statement.