The bytes_xfer_now/prev counters are only used by the auto convergence
logic. However, they are used alongside the dirty_pages_rate counter,
which is calculated (and required) outside of this logic. The problem
with this approach is that if the auto convergence capability is changed
while a migration is ongoing, the relationship of the counters will be
broken.

This moves the management of bytes_xfer_now/prev counters outside of the
auto convergence logic to address this issue.

Signed-off-by: Felipe Franciosi <felipe@nutanix.com>
---
 migration/ram.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/migration/ram.c b/migration/ram.c
index 495ecbe..1a3d9e6 100644
--- a/migration/ram.c
+++ b/migration/ram.c
@@ -697,6 +697,7 @@ static void migration_bitmap_sync(RAMState *rs)
         /* calculate period counters */
         rs->dirty_pages_rate = rs->num_dirty_pages_period * 1000
             / (end_time - rs->time_last_bitmap_sync);
+        bytes_xfer_now = ram_bytes_transferred();
 
         if (migrate_auto_converge()) {
             /* The following detection logic can be refined later. For now:
@@ -704,7 +705,6 @@ static void migration_bitmap_sync(RAMState *rs)
                amount of bytes that just got transferred since the last time we
                were in this routine. If that happens twice, start or increase
                throttling */
-            bytes_xfer_now = ram_bytes_transferred();
 
             if ((rs->num_dirty_pages_period * TARGET_PAGE_SIZE >
                    (bytes_xfer_now - rs->bytes_xfer_prev) / 2) &&
@@ -713,7 +713,6 @@ static void migration_bitmap_sync(RAMState *rs)
                     rs->dirty_rate_high_cnt = 0;
                     mig_throttle_guest_down();
             }
-            rs->bytes_xfer_prev = bytes_xfer_now;
         }
 
         if (migrate_use_xbzrle()) {
@@ -730,6 +729,7 @@ static void migration_bitmap_sync(RAMState *rs)
         /* reset period counters */
         rs->time_last_bitmap_sync = end_time;
         rs->num_dirty_pages_period = 0;
+        rs->bytes_xfer_prev = bytes_xfer_now;
     }
     if (migrate_use_events()) {
         qapi_event_send_migration_pass(rs->bitmap_sync_count, NULL);
-- 
1.9.5

On Wed, May 24, 2017 at 05:10:02PM +0100, Felipe Franciosi wrote:
> The bytes_xfer_now/prev counters are only used by the auto convergence
> logic. However, they are used alongside the dirty_pages_rate counter,
> which is calculated (and required) outside of this logic. The problem
> with this approach is that if the auto convergence capability is changed
> while a migration is ongoing, the relationship of the counters will be
> broken.
> 
> This moves the management of bytes_xfer_now/prev counters outside of the
> auto convergence logic to address this issue.
> 
> Signed-off-by: Felipe Franciosi <felipe@nutanix.com>

Reviewed-by: Peter Xu <peterx@redhat.com>

> ---
>  migration/ram.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/migration/ram.c b/migration/ram.c
> index 495ecbe..1a3d9e6 100644
> --- a/migration/ram.c
> +++ b/migration/ram.c
> @@ -697,6 +697,7 @@ static void migration_bitmap_sync(RAMState *rs)
>          /* calculate period counters */
>          rs->dirty_pages_rate = rs->num_dirty_pages_period * 1000
>              / (end_time - rs->time_last_bitmap_sync);
> +        bytes_xfer_now = ram_bytes_transferred();
>  
>          if (migrate_auto_converge()) {
>              /* The following detection logic can be refined later. For now:
> @@ -704,7 +705,6 @@ static void migration_bitmap_sync(RAMState *rs)
>                 amount of bytes that just got transferred since the last time we
>                 were in this routine. If that happens twice, start or increase
>                 throttling */
> -            bytes_xfer_now = ram_bytes_transferred();
>  
>              if ((rs->num_dirty_pages_period * TARGET_PAGE_SIZE >
>                     (bytes_xfer_now - rs->bytes_xfer_prev) / 2) &&
> @@ -713,7 +713,6 @@ static void migration_bitmap_sync(RAMState *rs)
>                      rs->dirty_rate_high_cnt = 0;
>                      mig_throttle_guest_down();
>              }
> -            rs->bytes_xfer_prev = bytes_xfer_now;
>          }
>  
>          if (migrate_use_xbzrle()) {
> @@ -730,6 +729,7 @@ static void migration_bitmap_sync(RAMState *rs)
>          /* reset period counters */
>          rs->time_last_bitmap_sync = end_time;
>          rs->num_dirty_pages_period = 0;
> +        rs->bytes_xfer_prev = bytes_xfer_now;
>      }
>      if (migrate_use_events()) {
>          qapi_event_send_migration_pass(rs->bitmap_sync_count, NULL);
> -- 
> 1.9.5
> 

-- 
Peter Xu

Felipe Franciosi <felipe@nutanix.com> wrote:
> The bytes_xfer_now/prev counters are only used by the auto convergence
> logic. However, they are used alongside the dirty_pages_rate counter,
> which is calculated (and required) outside of this logic. The problem
> with this approach is that if the auto convergence capability is changed
> while a migration is ongoing, the relationship of the counters will be
> broken.
>
> This moves the management of bytes_xfer_now/prev counters outside of the
> auto convergence logic to address this issue.
>
> Signed-off-by: Felipe Franciosi <felipe@nutanix.com>

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