[RFC,5/9] migration: check for dup pages only in bulk stage

Submitted by Peter Lieven on March 12, 2013, 3:51 p.m.

Details

Message ID 513F4EEC.9050903@dlhnet.de
State New
Headers show

Commit Message

Peter Lieven March 12, 2013, 3:51 p.m.
pages are dup (most likely zero) from the very beginning.
unless there is an application running inside the VM
zeroing out large areas of memory its wasted time
for checking for dup pages after the bulk stage.

Signed-off-by: Peter Lieven <pl@kamp.de>
---
  arch_init.c |    2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)

Patch hide | download patch | download mbox

diff --git a/arch_init.c b/arch_init.c
index d48b914..3d09327 100644
--- a/arch_init.c
+++ b/arch_init.c
@@ -445,7 +445,7 @@  static int ram_save_block(QEMUFile *f, bool last_stage)

              /* In doubt sent page as normal */
              bytes_sent = -1;
-            if (is_dup_page(p)) {
+            if (ram_bulk_stage && is_dup_page(p)) {
                  acct_info.dup_pages++;
                  bytes_sent = save_block_hdr(f, block, offset, cont,
                                              RAM_SAVE_FLAG_COMPRESS);