[Qemu-devel] [PATCH] Postcopy vs xbzrle: Don't send xbzrle pages once in postcopy [for 2.8]

[Qemu-devel] [PATCH] Postcopy vs xbzrle: Don't send xbzrle pages once in postcopy [for 2.8]

[Dr. David Alan Gilbert (git)]

From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>

xbzrle relies on reading pages that have already been sent
to the destination and then applying the modifications; we can't
do that in postcopy because the destination may well have
modified the page already or the page has been discarded.

I already didn't allow reception of xbzrle pages, but I
forgot to add the test to stop them being sent.

Enabling both xbzrle and postcopy can make some sense;
if you think that your migration might finish if you
have xbzrle, then when it doesn't complete you flick
over to postcopy and stop xbzrle'ing.

This corresponds to RH bug:
https://bugzilla.redhat.com/show_bug.cgi?id=1368422

Symptom is:

Unknown combination of migration flags: 0x60 (postcopy mode)
(either 0x60 or 0x40)

Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
---
 migration/ram.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/migration/ram.c b/migration/ram.c
index a3d70c4..79fdc94 100644
--- a/migration/ram.c
+++ b/migration/ram.c
@@ -771,7 +771,9 @@ static int ram_save_page(QEMUFile *f, PageSearchStatus *pss,
              * page would be stale
              */
             xbzrle_cache_zero_page(current_addr);
-        } else if (!ram_bulk_stage && migrate_use_xbzrle()) {
+        } else if (!ram_bulk_stage &&
+                   !migration_in_postcopy(migrate_get_current()) &&
+                   migrate_use_xbzrle()) {
             pages = save_xbzrle_page(f, &p, current_addr, block,
                                      offset, last_stage, bytes_transferred);
             if (!last_stage) {
-- 
2.7.4

Re: [Qemu-devel] [PATCH] Postcopy vs xbzrle: Don't send xbzrle pages once in postcopy [for 2.8]

[Juan Quintela]

"Dr. David Alan Gilbert (git)" <dgilbert@redhat.com> wrote:
> From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
>
> xbzrle relies on reading pages that have already been sent
> to the destination and then applying the modifications; we can't
> do that in postcopy because the destination may well have
> modified the page already or the page has been discarded.
>
> I already didn't allow reception of xbzrle pages, but I
> forgot to add the test to stop them being sent.
>
> Enabling both xbzrle and postcopy can make some sense;
> if you think that your migration might finish if you
> have xbzrle, then when it doesn't complete you flick
> over to postcopy and stop xbzrle'ing.
>
> This corresponds to RH bug:
> https://bugzilla.redhat.com/show_bug.cgi?id=1368422
>
> Symptom is:
>
> Unknown combination of migration flags: 0x60 (postcopy mode)
> (either 0x60 or 0x40)
>
> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

Reviewed-by: Juan Quintela <quintela@redhat.com>

applied, thanks