On Fri, Feb 19, 2021 at 09:18:56AM -0600, Eric Blake wrote:
Our cow data is locked for proper sharing across connections, so
even
though NBD_CMD_FLUSH is an intentional no-op, a client is guaranteed
that after a flush by connection A, a read by connection B will not
see any stale data, which matches the requirements for always
advertising multi-conn.
---
filters/cow/cow.c | 9 +++++++++
1 file changed, 9 insertions(+)
diff --git a/filters/cow/cow.c b/filters/cow/cow.c
index 93e10f24..1085ee36 100644
--- a/filters/cow/cow.c
+++ b/filters/cow/cow.c
@@ -181,6 +181,14 @@ cow_can_cache (struct nbdkit_next_ops *next_ops, void *nxdata, void
*handle)
return NBDKIT_CACHE_NATIVE;
}
+static int
+cow_can_multi_conn (struct nbdkit_next_ops *next_ops, void *nxdata,
+ void *handle)
+{
+ /* Our cache is consistent between connections. */
+ return 1;
+}
+
/* Override the plugin's .can_fast_zero, because our .zero is not fast */
static int
cow_can_fast_zero (struct nbdkit_next_ops *next_ops, void *nxdata,
@@ -712,6 +720,7 @@ static struct nbdkit_filter filter = {
.can_fua = cow_can_fua,
.can_cache = cow_can_cache,
.can_fast_zero = cow_can_fast_zero,
+ .can_multi_conn = cow_can_multi_conn,
.pread = cow_pread,
.pwrite = cow_pwrite,
.zero = cow_zero,
--
ACK
Rich.
--
Richard Jones, Virtualization Group, Red Hat
http://people.redhat.com/~rjones
Read my programming and virtualization blog:
http://rwmj.wordpress.com
virt-df lists disk usage of guests without needing to install any
software inside the virtual machine. Supports Linux and Windows.
http://people.redhat.com/~rjones/virt-df/