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

Activating the block devices causes the locks to be taken on
the backing file.  If we're running with -S and the destination libvirt
hasn't started the destination with 'cont', it's expecting the locks are
still untaken.

Don't activate the block devices if we're not going to autostart the VM;
'cont' already will do that anyway.   This change is tied to the new
migration capability 'late-block-activate' that defaults to off, keeping
the old behaviour by default.

bz: https://bugzilla.redhat.com/show_bug.cgi?id=1560854
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
---
 migration/migration.c | 34 +++++++++++++++++++++++++++-------
 qapi/migration.json   |  6 +++++-
 2 files changed, 32 insertions(+), 8 deletions(-)

diff --git a/migration/migration.c b/migration/migration.c
index 52a5092add..a05c6c893b 100644
--- a/migration/migration.c
+++ b/migration/migration.c
@@ -197,6 +197,16 @@ static void migrate_generate_event(int new_state)
     }
 }
 
+static bool migrate_late_block_activate(void)
+{
+    MigrationState *s;
+
+    s = migrate_get_current();
+
+    return s->enabled_capabilities[
+        MIGRATION_CAPABILITY_LATE_BLOCK_ACTIVATE];
+}
+
 /*
  * Called on -incoming with a defer: uri.
  * The migration can be started later after any parameters have been
@@ -306,13 +316,23 @@ static void process_incoming_migration_bh(void *opaque)
     Error *local_err = NULL;
     MigrationIncomingState *mis = opaque;
 
-    /* Make sure all file formats flush their mutable metadata.
-     * If we get an error here, just don't restart the VM yet. */
-    bdrv_invalidate_cache_all(&local_err);
-    if (local_err) {
-        error_report_err(local_err);
-        local_err = NULL;
-        autostart = false;
+    /* If capability late_block_activate is set:
+     * Only fire up the block code now if we're going to restart the
+     * VM, else 'cont' will do it.
+     * This causes file locking to happen; so we don't want it to happen
+     * unless we really are starting the VM.
+     */
+    if (!migrate_late_block_activate() ||
+         (autostart && (!global_state_received() ||
+            global_state_get_runstate() == RUN_STATE_RUNNING))) {
+        /* Make sure all file formats flush their mutable metadata.
+         * If we get an error here, just don't restart the VM yet. */
+        bdrv_invalidate_cache_all(&local_err);
+        if (local_err) {
+            error_report_err(local_err);
+            local_err = NULL;
+            autostart = false;
+        }
     }
 
     /*
diff --git a/qapi/migration.json b/qapi/migration.json
index 9d0bf82cf4..757302762d 100644
--- a/qapi/migration.json
+++ b/qapi/migration.json
@@ -357,13 +357,17 @@
 # @dirty-bitmaps: If enabled, QEMU will migrate named dirty bitmaps.
 #                 (since 2.12)
 #
+# @late-block-activate: If enabled, the destination will not activate block
+#           devices (and thus take locks) immediately at the end of migration.
+#           (since 2.13)
+#
 # Since: 1.2
 ##
 { 'enum': 'MigrationCapability',
   'data': ['xbzrle', 'rdma-pin-all', 'auto-converge', 'zero-blocks',
            'compress', 'events', 'postcopy-ram', 'x-colo', 'release-ram',
            'block', 'return-path', 'pause-before-switchover', 'x-multifd',
-           'dirty-bitmaps' ] }
+           'dirty-bitmaps', 'late-block-activate' ] }
 
 ##
 # @MigrationCapabilityStatus:
-- 
2.17.0

"Dr. David Alan Gilbert (git)" <dgilbert@redhat.com> wrote:
> From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
>
> Activating the block devices causes the locks to be taken on
> the backing file.  If we're running with -S and the destination libvirt
> hasn't started the destination with 'cont', it's expecting the locks are
> still untaken.

I will try to change the naming, because we have:
- don't activate block device
- and we stop calling bdrv_invalidate_cache_all
  i.e. behind the scenes bdrv_invalidate_cache activates a block device,
  not bad O:-)

> Don't activate the block devices if we're not going to autostart the VM;
> 'cont' already will do that anyway.   This change is tied to the new
> migration capability 'late-block-activate' that defaults to off, keeping
> the old behaviour by default.

I have read the old discussion, and we have:

- current patch (not call bdrv_invalidate_cache() at bottom_handler)
- kevin approach: create two commands
  activate_devices/deactivate_devices and make libvirt use them correctly

I am not really sure which one is better, but as the patch author is
Dave ....

> bz: https://bugzilla.redhat.com/show_bug.cgi?id=1560854
> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
> ---
>  migration/migration.c | 34 +++++++++++++++++++++++++++-------
>  qapi/migration.json   |  6 +++++-
>  2 files changed, 32 insertions(+), 8 deletions(-)

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


> +# @late-block-activate: If enabled, the destination will not activate block
> +#           devices (and thus take locks) immediately at the end of migration.
> +#           (since 2.13)
> +#
>  # Since: 1.2
>  ##
>  { 'enum': 'MigrationCapability',
>    'data': ['xbzrle', 'rdma-pin-all', 'auto-converge', 'zero-blocks',
>             'compress', 'events', 'postcopy-ram', 'x-colo', 'release-ram',
>             'block', 'return-path', 'pause-before-switchover', 'x-multifd',
> -           'dirty-bitmaps' ] }
> +           'dirty-bitmaps', 'late-block-activate' ] }
>  
>  ##
>  # @MigrationCapabilityStatus:

Just wondering, is there a way to enable a capability for 2.13 and
newer, and not activate it for older versions?  I *think* that this
feature should be on in new releases, and off by default with old
machine types.  But I haven't through long enough to know what needs to
be changed.

Later, Juan.

"Dr. David Alan Gilbert (git)" <dgilbert@redhat.com> wrote:
> From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
>
> Activating the block devices causes the locks to be taken on
> the backing file.  If we're running with -S and the destination libvirt
> hasn't started the destination with 'cont', it's expecting the locks are
> still untaken.
>
> Don't activate the block devices if we're not going to autostart the VM;
> 'cont' already will do that anyway.   This change is tied to the new
> migration capability 'late-block-activate' that defaults to off, keeping
> the old behaviour by default.
>
> bz: https://bugzilla.redhat.com/show_bug.cgi?id=1560854
> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

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