05.05.2021 13:10, Stefan Hajnoczi wrote:
On Thu, Apr 22, 2021 at 01:09:50AM +0300, Vladimir Sementsov-Ogievskiy wrote:
@@ -1981,8 +1985,15 @@ bdrv_co_write_req_prepare(BdrvChild *child, int64_t
offset, int64_t bytes,
} else {
assert(child->perm & BLK_PERM_WRITE);
On Thu, Apr 22, 2021 at 01:09:50AM +0300, Vladimir Sementsov-Ogievskiy wrote:
> @@ -1981,8 +1985,15 @@ bdrv_co_write_req_prepare(BdrvChild *child, int64_t
> offset, int64_t bytes,
> } else {
> assert(child->perm & BLK_PERM_WRITE);
> }
> -return
30.04.2021 13:04, Max Reitz wrote:
On 22.04.21 00:09, Vladimir Sementsov-Ogievskiy wrote:
write-notifiers are used only for write-threshold. New code for such
purpose should create filters.
Let's handle write-threshold simply in generic code and drop write
notifiers at all.
Also move part of
On 22.04.21 00:09, Vladimir Sementsov-Ogievskiy wrote:
write-notifiers are used only for write-threshold. New code for such
purpose should create filters.
Let's handle write-threshold simply in generic code and drop write
notifiers at all.
Also move part of write-threshold API that is used
this patch, is something of your patches 1-4 needed? Probably you may just
resend your series not touching write-threshold, and just note in cover-letter that
write-threshold is covered by "[PATCH] block: simplify write-threshold and drop
write notifiers"
--
Best regards,
Vladimir
On 22/04/2021 00:09, Vladimir Sementsov-Ogievskiy wrote:
write-notifiers are used only for write-threshold. New code for such
purpose should create filters.
Let's handle write-threshold simply in generic code and drop write
notifiers at all.
Also move part of write-threshold API that is
write-notifiers are used only for write-threshold. New code for such
purpose should create filters.
Let's handle write-threshold simply in generic code and drop write
notifiers at all.
Also move part of write-threshold API that is used only for testing to
the test.
Signed-off-by: Vladimir