On 21.01.22 18:05, Emanuele Giuseppe Esposito wrote:
Signed-off-by: Emanuele Giuseppe Esposito <eespo...@redhat.com>
---
include/block/block_int-common.h | 67 +++++++++++++++++++-------------
1 file changed, 40 insertions(+), 27 deletions(-)
diff --git a/include/block/block_int-common.h b/include/block/block_int-common.h
index e007dbf768..cc8c8835ba 100644
--- a/include/block/block_int-common.h
+++ b/include/block/block_int-common.h
@@ -815,12 +815,16 @@ struct BdrvChildClass {
*/
bool parent_is_bds;
+ /*
+ * Global state (GS) API. These functions run under the BQL lock.
+ *
+ * See include/block/block-global-state.h for more information about
+ * the GS API.
+ */
void (*inherit_options)(BdrvChildRole role, bool parent_is_format,
int *child_flags, QDict *child_options,
int parent_flags, QDict *parent_options);
-
void (*change_media)(BdrvChild *child, bool load);
- void (*resize)(BdrvChild *child);
/*
* Returns a name that is supposedly more useful for human users than the
The method this comment belongs to is `.get_name()`. It’s exposed
through `bdrv_get_parent_name()`, which is called by
`bdrv_get_device_name()` and `bdrv_get_device_or_node_name()` – so I
think it should be classified as I/O.
@@ -837,6 +841,40 @@ struct BdrvChildClass {
*/
char *(*get_parent_desc)(BdrvChild *child);
This function is very similar, so we might also want to reconsider
classifying it as I/O. There’s no need, because all of its callers do
run in the main thread, but at the same time I don’t believe there’s
anything stopping us (and it starts to sound to me like all functions of
the “get name” kind perhaps should ideally be I/O, in that they
shouldn’t require the GS context).
Up to you. O:)
(Rest of this patch looks good!)
Hanna
