Reduce the time of checkpoint for COLO
authorRao, Lei <lei.rao@intel.com>
Fri, 16 Oct 2020 05:52:01 +0000 (13:52 +0800)
committerJason Wang <jasowang@redhat.com>
Wed, 11 Nov 2020 08:52:23 +0000 (16:52 +0800)
we should set ram_bulk_stage to false after ram_state_init,
otherwise the bitmap will be unused in migration_bitmap_find_dirty.
all pages in ram cache will be flushed to the ram of secondary guest
for each checkpoint.

Signed-off-by: Lei Rao <lei.rao@intel.com>
Signed-off-by: Derek Su <dereksu@qnap.com>
Signed-off-by: Zhang Chen <chen.zhang@intel.com>
Reviewed-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Reviewed-by: Zhang Chen <chen.zhang@intel.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
migration/ram.c

index 2da2b622ab23f84400fddf75b51721d00c935a66..add5396a62f930bc0712fa93d3ad1c685eeb272f 100644 (file)
@@ -3011,6 +3011,18 @@ static void decompress_data_with_multi_threads(QEMUFile *f,
     qemu_mutex_unlock(&decomp_done_lock);
 }
 
+ /*
+  * we must set ram_bulk_stage to false, otherwise in
+  * migation_bitmap_find_dirty the bitmap will be unused and
+  * all the pages in ram cache wil be flushed to the ram of
+  * secondary VM.
+  */
+static void colo_init_ram_state(void)
+{
+    ram_state_init(&ram_state);
+    ram_state->ram_bulk_stage = false;
+}
+
 /*
  * colo cache: this is for secondary VM, we cache the whole
  * memory of the secondary VM, it is need to hold the global lock
@@ -3054,7 +3066,7 @@ int colo_init_ram_cache(void)
         }
     }
 
-    ram_state_init(&ram_state);
+    colo_init_ram_state();
     return 0;
 }