On 5. 3. 2020 13:51, Daniel P. Berrangé wrote:
This converts the QEMU monitor APIs to use the per-VM event loop, which involves switching from virEvent APIs to GMainContext / GSource APIs.
A GSocket is used as a convenient way to create a GSource for a socket, but is not yet used for actual I/O.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- src/qemu/qemu_monitor.c | 145 ++++++++++++++++------------------- src/qemu/qemu_monitor.h | 3 +- src/qemu/qemu_process.c | 6 +- tests/qemumonitortestutils.c | 1 + 4 files changed, 71 insertions(+), 84 deletions(-)
@@ -831,25 +804,32 @@ qemuMonitorOpen(virDomainObjPtr vm, * * Registers the monitor in the event loop. The caller has to hold the * lock for @mon. - * - * Returns true in case of success, false otherwise */ -bool +void qemuMonitorRegister(qemuMonitorPtr mon) { - virObjectRef(mon); - if ((mon->watch = virEventAddHandle(mon->fd, - VIR_EVENT_HANDLE_HANGUP | - VIR_EVENT_HANDLE_ERROR | - VIR_EVENT_HANDLE_READABLE, - qemuMonitorIO, - mon, - virObjectFreeCallback)) < 0) { - virObjectUnref(mon); - return false; + GIOCondition cond = 0; + + if (mon->lastError.code == VIR_ERR_OK) { + cond |= G_IO_IN; + + if ((mon->msg && mon->msg->txOffset < mon->msg->txLength) && + !mon->waitGreeting) + cond |= G_IO_OUT; }
- return true; + mon->watch = g_socket_create_source(mon->socket, + cond, + NULL); + + virObjectRef(mon); + g_source_set_callback(mon->watch, + (GSourceFunc)qemuMonitorIO, + mon, + NULL);
So previously, we passed virObjectFreeCallback, so that @mon is unrefed when removing the watch. But now you're passing NULL which will lead to a memleak. Replace it with "(GDestroyNotify) virObjectUnref". Michal