Merge remote-tracking branch 'remotes/juanquintela/tags/migration-pull-request' into staging

Migration pull request

This series fixes problems with migration-cancel while using multifd.
In some cases it can hang waiting in a semaphore.

Please apply.

# gpg: Signature made Thu 25 Jul 2019 11:56:57 BST
# gpg:                using RSA key 1899FF8EDEBF58CCEE034B82F487EF185872D723
# gpg: Good signature from "Juan Quintela <quintela@redhat.com>" [full]
# gpg:                 aka "Juan Quintela <quintela@trasno.org>" [full]
# Primary key fingerprint: 1899 FF8E DEBF 58CC EE03  4B82 F487 EF18 5872 D723

* remotes/juanquintela/tags/migration-pull-request:
  migration: fix migrate_cancel multifd migration leads destination hung forever
  migration: Make explicit that we are quitting multifd
  migration: fix migrate_cancel leads live_migration thread hung forever
  migration: fix migrate_cancel leads live_migration thread endless loop

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

1 files changed, 57 insertions, 9 deletions

diff --git a/migration/ram.c b/migration/ram.c
index 2b0774c..889148d 100644
--- a/migration/ram.c
+++ b/migration/ram.c
@@ -677,6 +677,8 @@ typedef struct {
     QemuMutex mutex;
     /* is this channel thread running */
     bool running;
+    /* should this thread finish */
+    bool quit;
     /* array of pages to receive */
     MultiFDPages_t *pages;
     /* packet allocated len */
@@ -920,7 +922,7 @@ struct {
  * false.
  */
 
-static void multifd_send_pages(void)
+static int multifd_send_pages(void)
 {
     int i;
     static int next_channel;
@@ -933,6 +935,11 @@ static void multifd_send_pages(void)
         p = &multifd_send_state->params[i];
 
         qemu_mutex_lock(&p->mutex);
+        if (p->quit) {
+            error_report("%s: channel %d has already quit!", __func__, i);
+            qemu_mutex_unlock(&p->mutex);
+            return -1;
+        }
         if (!p->pending_job) {
             p->pending_job++;
             next_channel = (i + 1) % migrate_multifd_channels();
@@ -951,9 +958,11 @@ static void multifd_send_pages(void)
     ram_counters.transferred += transferred;;
     qemu_mutex_unlock(&p->mutex);
     qemu_sem_post(&p->sem);
+
+    return 1;
 }
 
-static void multifd_queue_page(RAMBlock *block, ram_addr_t offset)
+static int multifd_queue_page(RAMBlock *block, ram_addr_t offset)
 {
     MultiFDPages_t *pages = multifd_send_state->pages;
 
@@ -968,15 +977,19 @@ static void multifd_queue_page(RAMBlock *block, ram_addr_t offset)
         pages->used++;
 
         if (pages->used < pages->allocated) {
-            return;
+            return 1;
         }
     }
 
-    multifd_send_pages();
+    if (multifd_send_pages() < 0) {
+        return -1;
+    }
 
     if (pages->block != block) {
-        multifd_queue_page(block, offset);
+        return  multifd_queue_page(block, offset);
     }
+
+    return 1;
 }
 
 static void multifd_send_terminate_threads(Error *err)
@@ -1049,7 +1062,10 @@ static void multifd_send_sync_main(void)
         return;
     }
     if (multifd_send_state->pages->used) {
-        multifd_send_pages();
+        if (multifd_send_pages() < 0) {
+            error_report("%s: multifd_send_pages fail", __func__);
+            return;
+        }
     }
     for (i = 0; i < migrate_multifd_channels(); i++) {
         MultiFDSendParams *p = &multifd_send_state->params[i];
@@ -1058,6 +1074,12 @@ static void multifd_send_sync_main(void)
 
         qemu_mutex_lock(&p->mutex);
 
+        if (p->quit) {
+            error_report("%s: channel %d has already quit", __func__, i);
+            qemu_mutex_unlock(&p->mutex);
+            return;
+        }
+
         p->packet_num = multifd_send_state->packet_num++;
         p->flags |= MULTIFD_FLAG_SYNC;
         p->pending_job++;
@@ -1077,7 +1099,8 @@ static void *multifd_send_thread(void *opaque)
 {
     MultiFDSendParams *p = opaque;
     Error *local_err = NULL;
-    int ret;
+    int ret = 0;
+    uint32_t flags = 0;
 
     trace_multifd_send_thread_start(p->id);
     rcu_register_thread();
@@ -1095,7 +1118,7 @@ static void *multifd_send_thread(void *opaque)
         if (p->pending_job) {
             uint32_t used = p->pages->used;
             uint64_t packet_num = p->packet_num;
-            uint32_t flags = p->flags;
+            flags = p->flags;
 
             p->next_packet_size = used * qemu_target_page_size();
             multifd_send_fill_packet(p);
@@ -1144,6 +1167,17 @@ out:
         multifd_send_terminate_threads(local_err);
     }
 
+    /*
+     * Error happen, I will exit, but I can't just leave, tell
+     * who pay attention to me.
+     */
+    if (ret != 0) {
+        if (flags & MULTIFD_FLAG_SYNC) {
+            qemu_sem_post(&multifd_send_state->sem_sync);
+        }
+        qemu_sem_post(&multifd_send_state->channels_ready);
+    }
+
     qemu_mutex_lock(&p->mutex);
     p->running = false;
     qemu_mutex_unlock(&p->mutex);
@@ -1234,6 +1268,7 @@ static void multifd_recv_terminate_threads(Error *err)
         MultiFDRecvParams *p = &multifd_recv_state->params[i];
 
         qemu_mutex_lock(&p->mutex);
+        p->quit = true;
         /* We could arrive here for two reasons:
            - normal quit, i.e. everything went fine, just finished
            - error quit: We close the channels so the channel threads
@@ -1256,6 +1291,12 @@ int multifd_load_cleanup(Error **errp)
         MultiFDRecvParams *p = &multifd_recv_state->params[i];
 
         if (p->running) {
+            p->quit = true;
+            /*
+             * multifd_recv_thread may hung at MULTIFD_FLAG_SYNC handle code,
+             * however try to wakeup it without harm in cleanup phase.
+             */
+            qemu_sem_post(&p->sem_sync);
             qemu_thread_join(&p->thread);
         }
         object_unref(OBJECT(p->c));
@@ -1319,6 +1360,10 @@ static void *multifd_recv_thread(void *opaque)
         uint32_t used;
         uint32_t flags;
 
+        if (p->quit) {
+            break;
+        }
+
         ret = qio_channel_read_all_eof(p->c, (void *)p->packet,
                                        p->packet_len, &local_err);
         if (ret == 0) {   /* EOF */
@@ -1390,6 +1435,7 @@ int multifd_load_setup(void)
 
         qemu_mutex_init(&p->mutex);
         qemu_sem_init(&p->sem_sync, 0);
+        p->quit = false;
         p->id = i;
         p->pages = multifd_pages_init(page_count);
         p->packet_len = sizeof(MultiFDPacket_t)
@@ -2033,7 +2079,9 @@ static int ram_save_page(RAMState *rs, PageSearchStatus *pss, bool last_stage)
 static int ram_save_multifd_page(RAMState *rs, RAMBlock *block,
                                  ram_addr_t offset)
 {
-    multifd_queue_page(block, offset);
+    if (multifd_queue_page(block, offset) < 0) {
+        return -1;
+    }
     ram_counters.normal++;
 
     return 1;