Message ID | 20171106094643.14881-11-peterx@redhat.com |
---|---|
State | New |
Headers | show |
Series | [RFC,v3,01/27] char-io: fix possible race on IOWatchPoll | expand |
On Mon, 11/06 17:46, Peter Xu wrote: > +static GMainContext *monitor_io_context_get(void) > +{ > + return iothread_get_g_main_context(mon_global.mon_iothread); > +} > + > +static void monitor_iothread_init(void) > +{ > + mon_global.mon_iothread = iothread_create("monitor_iothread", > + &error_abort); > + /* > + * GContext in iothread is using lazy init - the first time we > + * fetch the context we'll have that initialized. > + */ > + monitor_io_context_get(); Why do you need an eager init here? Fam
On Tue, Nov 07, 2017 at 03:11:31PM +0800, Fam Zheng wrote: > On Mon, 11/06 17:46, Peter Xu wrote: > > +static GMainContext *monitor_io_context_get(void) > > +{ > > + return iothread_get_g_main_context(mon_global.mon_iothread); > > +} > > + > > +static void monitor_iothread_init(void) > > +{ > > + mon_global.mon_iothread = iothread_create("monitor_iothread", > > + &error_abort); > > + /* > > + * GContext in iothread is using lazy init - the first time we > > + * fetch the context we'll have that initialized. > > + */ > > + monitor_io_context_get(); > > Why do you need an eager init here? I thought putting this line with IOThread creation will make it more clear that the IOThread is using GMainContext polling. I am not that strong to keep it, but I would prefer.
On Wed, 11/08 15:25, Peter Xu wrote: > On Tue, Nov 07, 2017 at 03:11:31PM +0800, Fam Zheng wrote: > > On Mon, 11/06 17:46, Peter Xu wrote: > > > +static GMainContext *monitor_io_context_get(void) > > > +{ > > > + return iothread_get_g_main_context(mon_global.mon_iothread); > > > +} > > > + > > > +static void monitor_iothread_init(void) > > > +{ > > > + mon_global.mon_iothread = iothread_create("monitor_iothread", > > > + &error_abort); > > > + /* > > > + * GContext in iothread is using lazy init - the first time we > > > + * fetch the context we'll have that initialized. > > > + */ > > > + monitor_io_context_get(); > > > > Why do you need an eager init here? > > I thought putting this line with IOThread creation will make it more > clear that the IOThread is using GMainContext polling. I am not that > strong to keep it, but I would prefer. The "iothread only starts using GMainContext polling after the lazy init happens" is completely opaque to its user, so I don't think there is anything but confusion to have this. Fam
On Wed, Nov 08, 2017 at 07:18:44PM +0800, Fam Zheng wrote: > On Wed, 11/08 15:25, Peter Xu wrote: > > On Tue, Nov 07, 2017 at 03:11:31PM +0800, Fam Zheng wrote: > > > On Mon, 11/06 17:46, Peter Xu wrote: > > > > +static GMainContext *monitor_io_context_get(void) > > > > +{ > > > > + return iothread_get_g_main_context(mon_global.mon_iothread); > > > > +} > > > > + > > > > +static void monitor_iothread_init(void) > > > > +{ > > > > + mon_global.mon_iothread = iothread_create("monitor_iothread", > > > > + &error_abort); > > > > + /* > > > > + * GContext in iothread is using lazy init - the first time we > > > > + * fetch the context we'll have that initialized. > > > > + */ > > > > + monitor_io_context_get(); > > > > > > Why do you need an eager init here? > > > > I thought putting this line with IOThread creation will make it more > > clear that the IOThread is using GMainContext polling. I am not that > > strong to keep it, but I would prefer. > > The "iothread only starts using GMainContext polling after the lazy init > happens" is completely opaque to its user, so I don't think there is anything > but confusion to have this. Makes sense. Let me remove it.
diff --git a/monitor.c b/monitor.c index a70ab5606b..df1ec8d037 100644 --- a/monitor.c +++ b/monitor.c @@ -76,6 +76,7 @@ #include "qmp-introspect.h" #include "sysemu/qtest.h" #include "sysemu/cpus.h" +#include "sysemu/iothread.h" #include "qemu/cutils.h" #include "qapi/qmp/dispatch.h" @@ -208,6 +209,12 @@ struct Monitor { QTAILQ_ENTRY(Monitor) entry; }; +struct MonitorGlobal { + IOThread *mon_iothread; +}; + +static struct MonitorGlobal mon_global; + /* QMP checker flags */ #define QMP_ACCEPT_UNKNOWNS 1 @@ -4034,12 +4041,29 @@ static void sortcmdlist(void) qsort((void *)info_cmds, array_num, elem_size, compare_mon_cmd); } +static GMainContext *monitor_io_context_get(void) +{ + return iothread_get_g_main_context(mon_global.mon_iothread); +} + +static void monitor_iothread_init(void) +{ + mon_global.mon_iothread = iothread_create("monitor_iothread", + &error_abort); + /* + * GContext in iothread is using lazy init - the first time we + * fetch the context we'll have that initialized. + */ + monitor_io_context_get(); +} + void monitor_init_globals(void) { monitor_init_qmp_commands(); monitor_qapi_event_init(); sortcmdlist(); qemu_mutex_init(&monitor_lock); + monitor_iothread_init(); } /* These functions just adapt the readline interface in a typesafe way. We @@ -4117,6 +4141,13 @@ void monitor_cleanup(void) { Monitor *mon, *next; + /* + * We need to explicitly stop the iothread (but not destroy it), + * cleanup the monitor resources, then destroy the iothread. See + * again on the glib bug mentioned in 2b316774f6 for a reason. + */ + iothread_stop(mon_global.mon_iothread); + qemu_mutex_lock(&monitor_lock); QTAILQ_FOREACH_SAFE(mon, &mon_list, entry, next) { QTAILQ_REMOVE(&mon_list, mon, entry); @@ -4124,6 +4155,9 @@ void monitor_cleanup(void) g_free(mon); } qemu_mutex_unlock(&monitor_lock); + + iothread_destroy(mon_global.mon_iothread); + mon_global.mon_iothread = NULL; } QemuOptsList qemu_mon_opts = {
Create one IOThread for the monitors, prepared to handle all the input/output IOs using existing iothread framework. Signed-off-by: Peter Xu <peterx@redhat.com> --- monitor.c | 34 ++++++++++++++++++++++++++++++++++ 1 file changed, 34 insertions(+)