Commit b606bbb41 reminded me that any time we drop locks to run
back-to-back guest interaction commands, we have to check that
the guest didn't disappear in between the two commands. A quick
audit found a couple of spots that were missing this check.
* src/qemu/qemu_driver.c (qemuDomainShutdownFlags)
(qemuDomainSetVcpusFlags): Check that domain is still up.
Signed-off-by: Eric Blake <eblake@xxxxxxxxxx>
---
I found this by inspection, and did not try to actually reproduce
if it would cause failures. I think a failure could be easier
to reproduce by temporarily sticking a strategic sleep() call just
prior to the qemuDomainObjExtAgent() call, and during the time
while libvirtd is sleeping, manually cause the guest to exit. But
we've definitely fixed issues like this before.
src/qemu/qemu_driver.c | 13 +++++++++++++
1 file changed, 13 insertions(+)
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c
index 02088cc..4ae76e5 100644
--- a/src/qemu/qemu_driver.c
+++ b/src/qemu/qemu_driver.c
@@ -1921,6 +1921,13 @@ static int qemuDomainShutdownFlags(virDomainPtr dom, unsigned int flags)
(ret < 0 && (acpiRequested || !flags))) {
qemuDomainSetFakeReboot(driver, vm, isReboot);
+ /* Even if agent failed, we have to check if guest went away
+ * by itself while our locks were down. */
+ if (useAgent && !virDomainObjIsActive(vm)) {
+ ret = 0;
+ goto endjob;
+ }
+
qemuDomainObjEnterMonitor(driver, vm);
ret = qemuMonitorSystemPowerdown(priv->mon);
qemuDomainObjExitMonitor(driver, vm);
@@ -4360,6 +4367,12 @@ qemuDomainSetVcpusFlags(virDomainPtr dom, unsigned int nvcpus,
if (ncpuinfo < 0)
goto endjob;
+ if (!virDomainObjIsActive(vm)) {
+ virReportError(VIR_ERR_OPERATION_INVALID, "%s",
+ _("domain is not running"));
+ goto endjob;
+ }
+
if (qemuAgentUpdateCPUInfo(nvcpus, cpuinfo, ncpuinfo) < 0)
goto endjob;
--
1.9.3
--
libvir-list mailing list
libvir-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/libvir-list