[14/17] util/timer: avoid deadlock when shutting down

Commit Message

Alex Bennée Sept. 13, 2024, 5:26 p.m. UTC

When we shut down a guest we disable the timers. However this can
cause deadlock if the guest has queued some async work that is trying
to advance system time and spins forever trying to wind time forward.
Pay attention to the return code and bail early if we can't wind time
forward.

Signed-off-by: Alex Bennée <[e-mail]>
Reported-by: Elisha Hollander <[e-mail]>
---
 util/qemu-timer.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

Patch

diff --git a/util/qemu-timer.c b/util/qemu-timer.c
index 213114be68..6b1533bc2a 100644
--- a/util/qemu-timer.c
+++ b/util/qemu-timer.c
@@ -685,10 +685,17 @@  int64_t qemu_clock_advance_virtual_time(int64_t dest)
 {
     int64_t clock = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL);
     AioContext *aio_context;
+    int64_t deadline;
+
     aio_context = qemu_get_aio_context();
-    while (clock < dest) {
-        int64_t deadline = qemu_clock_deadline_ns_all(QEMU_CLOCK_VIRTUAL,
+
+    deadline = qemu_clock_deadline_ns_all(QEMU_CLOCK_VIRTUAL,
                                                       QEMU_TIMER_ATTR_ALL);
+    /*
+     * A deadline of < 0 indicates this timer is not enabled, so we
+     * won't get far trying to run it forward.
+     */
+    while (deadline >= 0 && clock < dest) {
         int64_t warp = qemu_soonest_timeout(dest - clock, deadline);
 
         qemu_virtual_clock_set_ns(qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL) + warp);
@@ -696,6 +703,9 @@  int64_t qemu_clock_advance_virtual_time(int64_t dest)
         qemu_clock_run_timers(QEMU_CLOCK_VIRTUAL);
         timerlist_run_timers(aio_context->tlg.tl[QEMU_CLOCK_VIRTUAL]);
         clock = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL);
+
+        deadline = qemu_clock_deadline_ns_all(QEMU_CLOCK_VIRTUAL,
+                                              QEMU_TIMER_ATTR_ALL);
     }
     qemu_clock_notify(QEMU_CLOCK_VIRTUAL);