On 05/27/14 16:53, Jiri Denemark wrote:
In general, we should only remove a backend after seeing
DEVICE_DELETED
event for a corresponding frontend. This doesn't make any difference for
disks attached using -drive or drive_add since QEMU automatically
removes their frontends but it's still better to make our code
s/frontends/backends/ ?
consistent. And it may start making difference in case we switch to
attaching disks using -blockdev.
Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com>
---
src/qemu/qemu_hotplug.c | 37 +++++++++++++------------------------
1 file changed, 13 insertions(+), 24 deletions(-)
diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c
index fd1f002..43c52bf 100644
--- a/src/qemu/qemu_hotplug.c
+++ b/src/qemu/qemu_hotplug.c
@@ -2472,10 +2472,23 @@ qemuDomainRemoveDiskDevice(virQEMUDriverPtr driver,
virObjectEventPtr event;
size_t i;
const char *src = virDomainDiskGetSource(disk);
+ qemuDomainObjPrivatePtr priv = vm->privateData;
+ char *drivestr;
VIR_DEBUG("Removing disk %s from domain %p %s",
disk->info.alias, vm, vm->def->name);
+ /* build the actual drive id string as the disk->info.alias doesn't
+ * contain the QEMU_DRIVE_HOST_PREFIX that is passed to qemu */
+ if (virAsprintf(&drivestr, "%s%s",
+ QEMU_DRIVE_HOST_PREFIX, disk->info.alias) < 0)
+ return;
Should this be treated the same way for propagating errors as in the
previous patch? I'm not going to enforce this here as the allocaion is
unlikely to fail.
+
+ qemuDomainObjEnterMonitor(driver, vm);
+ qemuMonitorDriveDel(priv->mon, drivestr);
+ qemuDomainObjExitMonitor(driver, vm);
+ VIR_FREE(drivestr);
+
virDomainAuditDisk(vm, src, NULL, "detach", true);
event = virDomainEventDeviceRemovedNewFromObj(vm, disk->info.alias);
ACK if you go with this patch as-is, v2 if you upgrade the error reporting.
Peter