[libvirt] [PATCH v3 4/4] virsh: improve waiting for block job readiness
After a block job hits 100%, we only need to apply a timeout waiting for
a block job event if exactly one of the BLOCK_JOB or BLOCK_JOB_2
callbacks were able to be registered.
If neither callback could be registered, there's clearly no need for a
timeout.
If both callbacks were registered, then we're guaranteed to eventually
get one of the events. The path being used by virsh must be exactly the
source path or target device in the domain's disk definition, and these
are the respective strings sent back in these two events.
Signed-off-by: Michael Chapman <mike(a)very.puzzling.org>
---
tools/virsh-domain.c | 50 ++++++++++++++++++++++++++++----------------------
1 file changed, 28 insertions(+), 22 deletions(-)
diff --git a/tools/virsh-domain.c b/tools/virsh-domain.c
index 750b273..bf65a60 100644
--- a/tools/virsh-domain.c
+++ b/tools/virsh-domain.c
@@ -1810,14 +1810,17 @@ virshBlockJobWaitFree(virshBlockJobWaitDataPtr data)
* virshBlockJobWait:
* @data: private data initialized by virshBlockJobWaitInit
*
- * Waits for the block job to complete. This function prefers to get an event
- * from libvirt but still has fallback means if the device name can't be matched
+ * Waits for the block job to complete. This function prefers to wait for a
+ * matching VIR_DOMAIN_EVENT_ID_BLOCK_JOB or VIR_DOMAIN_EVENT_ID_BLOCK_JOB_2
+ * event from libvirt; however, it has a fallback mode should either of these
+ * events not be available.
*
- * This function returns values from the virConnectDomainEventBlockJobStatus enum
- * or -1 in case of a internal error. Fallback states if a block job vanishes
- * without triggering the event is VIR_DOMAIN_BLOCK_JOB_COMPLETED. For two phase
- * jobs after the retry count for waiting for the event expires is
- * VIR_DOMAIN_BLOCK_JOB_READY.
+ * This function returns values from the virConnectDomainEventBlockJobStatus
+ * enum or -1 in case of an internal error.
+ *
+ * If the fallback mode is activated the returned event is
+ * VIR_DOMAIN_BLOCK_JOB_COMPLETED if the block job vanishes or
+ * VIR_DOMAIN_BLOCK_JOB_READY if the block job reaches 100%.
*/
static int
virshBlockJobWait(virshBlockJobWaitDataPtr data)
@@ -1872,27 +1875,30 @@ virshBlockJobWait(virshBlockJobWaitDataPtr data)
goto cleanup;
}
- /* if we've got an event for the device we are waiting for we can end
- * the waiting loop */
+ /* If either callback could be registered and we've got an event, we can
+ * can end the waiting loop */
if ((data->cb_id >= 0 || data->cb_id2 >= 0) &&
data->status != -1) {
ret = data->status;
break;
}
- /* since virsh can't guarantee that the path provided by the user will
- * later be matched in the event we will need to keep the fallback
- * approach and claim success if the block job finishes or vanishes. */
- if (result == 0) {
- ret = VIR_DOMAIN_BLOCK_JOB_COMPLETED;
- break;
- }
+ /* Fallback behaviour is only needed if one or both callbacks could not
+ * be registered */
+ if (data->cb_id < 0 || data->cb_id2 < 0) {
+ /* If the block job vanishes, synthesize a COMPLETED event */
+ if (result == 0) {
+ ret = VIR_DOMAIN_BLOCK_JOB_COMPLETED;
+ break;
+ }
- /* for two-phase jobs we will try to wait in the synchronized phase
- * for event arrival since 100% completion doesn't necessarily mean that
- * the block job has finished and can be terminated with success */
- if (info.end == info.cur && --retries == 0) {
- ret = VIR_DOMAIN_BLOCK_JOB_READY;
- break;
+ /* If the block job hits 100%, wait a little while for a possible
+ * event from libvirt unless both callbacks could not be registered
+ * in order to synthesize our own READY event */
+ if (info.end == info.cur &&
+ ((data->cb_id < 0 && data->cb_id2 < 0) || --retries
== 0)) {
+ ret = VIR_DOMAIN_BLOCK_JOB_READY;
+ break;
+ }
}
if (data->verbose && (info.cur != last.cur || info.end != last.end))
--
2.4.3