On Mon, Jan 29, 2018 at 11:32:00 -0500, John Ferlan wrote:
Add data to qemuDomainJobObj in order to store the dump completion event information. Once the event has been received future code waiting on the event will be able to process the stats and error buffer. If there's no async job, we can just ignore.
Signed-off-by: John Ferlan <jferlan@redhat.com> --- src/qemu/qemu_domain.c | 5 +++++ src/qemu/qemu_domain.h | 3 +++ src/qemu/qemu_process.c | 32 ++++++++++++++++++++++++++++++++ 3 files changed, 40 insertions(+)
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c index 6b4bd3cca..d8b2b3067 100644 --- a/src/qemu/qemu_domain.c +++ b/src/qemu/qemu_domain.c @@ -334,6 +334,11 @@ qemuDomainObjResetAsyncJob(qemuDomainObjPrivatePtr priv) job->spiceMigration = false; job->spiceMigrated = false; job->postcopyEnabled = false; + job->dumpCompleted = false;
OK
+ qemuMonitorEventDumpStatsFree(job->dumpCompletedStats); + job->dumpCompletedStats = NULL;
VIR_FREE would be a bit shorter, but... (see below).
+ VIR_FREE(job->dumpCompletedError);
OK
+ job->dumpCompletedError = NULL;
This is redundant.
VIR_FREE(job->current); }
diff --git a/src/qemu/qemu_domain.h b/src/qemu/qemu_domain.h index ddfc46dcd..7dab758fb 100644 --- a/src/qemu/qemu_domain.h +++ b/src/qemu/qemu_domain.h @@ -164,6 +164,9 @@ struct qemuDomainJobObj { * should wait for it to finish */ bool spiceMigrated; /* spice migration completed */ bool postcopyEnabled; /* post-copy migration was enabled */ + bool dumpCompleted; /* dump completed */
OK
+ qemuMonitorDumpStatsPtr dumpCompletedStats; /* dump completion stats */
Heh, why do we need to structures for storing statistics? Just use the qemuDomainJobInfoPtr structure we already point to from current.
+ char *dumpCompletedError; /* dump completion event error */
Since there can only be one job at a time and it should fail at most once, I think we could just have a generic char *error and possibly reuse it for other jobs in the future.
};
typedef void (*qemuDomainCleanupCallback)(virQEMUDriverPtr driver, diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 3a697de03..de43f6ac0 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -1690,6 +1690,37 @@ qemuProcessHandleMigrationPass(qemuMonitorPtr mon ATTRIBUTE_UNUSED, }
+static int +qemuProcessHandleDumpCompleted(qemuMonitorPtr mon ATTRIBUTE_UNUSED, + virDomainObjPtr vm, + qemuMonitorDumpStatsPtr stats, + const char *error, + void *opaque ATTRIBUTE_UNUSED) +{ + qemuDomainObjPrivatePtr priv; + + virObjectLock(vm); + + VIR_DEBUG("Dump completed for domain %p %s with stats=%p error='%s'", + vm, vm->def->name, stats, NULLSTR(error)); + + priv = vm->privateData; + if (priv->job.asyncJob == QEMU_ASYNC_JOB_NONE) { + VIR_DEBUG("got DUMP_COMPLETED event without a dump_completed job"); + goto cleanup; + } + priv->job.dumpCompleted = true; + VIR_STEAL_PTR(priv->job.dumpCompletedStats, stats);
priv->jobs.current->... = *stats;
+ ignore_value(VIR_STRDUP_QUIET(priv->job.dumpCompletedError, error)); + virDomainObjBroadcast(vm); + + cleanup: + qemuMonitorEventDumpStatsFree(stats);
Ah so in fact there wouldn't be any memory leak in qemuMonitorJSONHandleDumpCompleted since QEMU driver will always register this handler, but it's awkward anyway.
+ virObjectUnlock(vm); + return 0; +} + + static qemuMonitorCallbacks monitorCallbacks = { .eofNotify = qemuProcessHandleMonitorEOF, .errorNotify = qemuProcessHandleMonitorError, @@ -1718,6 +1749,7 @@ static qemuMonitorCallbacks monitorCallbacks = { .domainMigrationPass = qemuProcessHandleMigrationPass, .domainAcpiOstInfo = qemuProcessHandleAcpiOstInfo, .domainBlockThreshold = qemuProcessHandleBlockThreshold, + .domainDumpCompleted = qemuProcessHandleDumpCompleted, };
Jirka