Revert "iothread: release iothread around aio_poll"
This reverts commit a0710f7995f914e3044e5899bd8ff6c43c62f916.
In qemu-devel email message <556DBF87.2020908@de.ibm.com>, Christian
Borntraeger writes:
Having many guests all with a kernel/ramdisk (via -kernel) and
several null block devices will result in hangs. All hanging
guests are in partition detection code waiting for an I/O to return
so very early maybe even the first I/O.
Reverting that commit "fixes" the hangs.
Reverting this commit for the 2.4 release. More time is needed to
investigate and correct this patch.
Reported-by: Christian Borntraeger <borntraeger@de.ibm.com>
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
diff --git a/async.c b/async.c
index 46d9e63..77d080d 100644
--- a/async.c
+++ b/async.c
@@ -280,6 +280,12 @@
aio_notify(opaque);
}
+static void aio_rfifolock_cb(void *opaque)
+{
+ /* Kick owner thread in case they are blocked in aio_poll() */
+ aio_notify(opaque);
+}
+
AioContext *aio_context_new(Error **errp)
{
int ret;
@@ -297,7 +303,7 @@
event_notifier_test_and_clear);
ctx->thread_pool = NULL;
qemu_mutex_init(&ctx->bh_lock);
- rfifolock_init(&ctx->lock, NULL, NULL);
+ rfifolock_init(&ctx->lock, aio_rfifolock_cb, ctx);
timerlistgroup_init(&ctx->tlg, aio_timerlist_notify, ctx);
return ctx;
diff --git a/iothread.c b/iothread.c
index 0416fc4..878a594 100644
--- a/iothread.c
+++ b/iothread.c
@@ -31,14 +31,21 @@
static void *iothread_run(void *opaque)
{
IOThread *iothread = opaque;
+ bool blocking;
qemu_mutex_lock(&iothread->init_done_lock);
iothread->thread_id = qemu_get_thread_id();
qemu_cond_signal(&iothread->init_done_cond);
qemu_mutex_unlock(&iothread->init_done_lock);
- while (!atomic_read(&iothread->stopping)) {
- aio_poll(iothread->ctx, true);
+ while (!iothread->stopping) {
+ aio_context_acquire(iothread->ctx);
+ blocking = true;
+ while (!iothread->stopping && aio_poll(iothread->ctx, blocking)) {
+ /* Progress was made, keep going */
+ blocking = false;
+ }
+ aio_context_release(iothread->ctx);
}
return NULL;
}
diff --git a/tests/test-aio.c b/tests/test-aio.c
index 4b0cb45..a7cb5c9 100644
--- a/tests/test-aio.c
+++ b/tests/test-aio.c
@@ -107,7 +107,6 @@
typedef struct {
QemuMutex start_lock;
- EventNotifier notifier;
bool thread_acquired;
} AcquireTestData;
@@ -119,8 +118,6 @@
qemu_mutex_lock(&data->start_lock);
qemu_mutex_unlock(&data->start_lock);
- g_usleep(500000);
- event_notifier_set(&data->notifier);
aio_context_acquire(ctx);
aio_context_release(ctx);
@@ -129,19 +126,20 @@
return NULL;
}
-static void dummy_notifier_read(EventNotifier *n)
+static void dummy_notifier_read(EventNotifier *unused)
{
- event_notifier_test_and_clear(n);
+ g_assert(false); /* should never be invoked */
}
static void test_acquire(void)
{
QemuThread thread;
+ EventNotifier notifier;
AcquireTestData data;
/* Dummy event notifier ensures aio_poll() will block */
- event_notifier_init(&data.notifier, false);
- aio_set_event_notifier(ctx, &data.notifier, dummy_notifier_read);
+ event_notifier_init(¬ifier, false);
+ aio_set_event_notifier(ctx, ¬ifier, dummy_notifier_read);
g_assert(!aio_poll(ctx, false)); /* consume aio_notify() */
qemu_mutex_init(&data.start_lock);
@@ -155,13 +153,12 @@
/* Block in aio_poll(), let other thread kick us and acquire context */
aio_context_acquire(ctx);
qemu_mutex_unlock(&data.start_lock); /* let the thread run */
- g_assert(aio_poll(ctx, true));
- g_assert(!data.thread_acquired);
+ g_assert(!aio_poll(ctx, true));
aio_context_release(ctx);
qemu_thread_join(&thread);
- aio_set_event_notifier(ctx, &data.notifier, NULL);
- event_notifier_cleanup(&data.notifier);
+ aio_set_event_notifier(ctx, ¬ifier, NULL);
+ event_notifier_cleanup(¬ifier);
g_assert(data.thread_acquired);
}