[PATCH 00/12] migration: Yet another round of atomic counters

[PATCH 00/12] migration: Yet another round of atomic counters

[Juan Quintela]

Hi

Goal of the whole series was to be able to move rate_limit logic to
not use qemu_file.  Goal achieved.

Removal of trasnferred atomic counter.

After this series, we have three atomic counters:
- multifd_bytes
- rdma_bytes
- qemu_file_trasferred

And we only need to setup one (and only one) of these each time that
we sent anything.

Please review.

Later, Juan.

Juan Quintela (12):
  qemu-file: We only call qemu_file_transferred_* on the sending side
  qemu_file: Use a stat64 for qemu_file_transferred
  qemu_file: total_transferred is not used anymore
  migration: Use the number of transferred bytes directly
  qemu_file: Remove unused qemu_file_transferred()
  qemu-file: Remove _noflush from qemu_file_transferred_noflush()
  migration: migration_transferred_bytes() don't need the QEMUFile
  migration: migration_rate_limit_reset() don't need the QEMUFile
  qemu-file: Simplify qemu_file_get_error()
  migration: Use migration_transferred_bytes()
  migration: Remove transferred atomic counter
  qemu-file: Make qemu_fflush() return errors

 migration/migration-stats.h | 16 ++++++--------
 migration/qemu-file.h       | 27 ++++-------------------
 migration/block.c           |  4 ++--
 migration/colo.c            | 11 +++-------
 migration/migration-stats.c | 10 ++++-----
 migration/migration.c       | 17 ++++++---------
 migration/multifd.c         |  3 ---
 migration/qemu-file.c       | 43 +++++++++++--------------------------
 migration/ram.c             | 29 +++++++++----------------
 migration/rdma.c            |  4 +---
 migration/savevm.c          |  9 ++++----
 migration/vmstate.c         |  4 ++--
 12 files changed, 56 insertions(+), 121 deletions(-)

-- 
2.41.0

[PATCH 01/12] qemu-file: We only call qemu_file_transferred_* on the sending side

[Juan Quintela]

Remove the increase in qemu_file_fill_buffer() and add asserts to
qemu_file_transferred* functions.

Signed-off-by: Juan Quintela <quintela@redhat.com>
---
 migration/qemu-file.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/migration/qemu-file.c b/migration/qemu-file.c
index 3fb25148d1..6814c562e6 100644
--- a/migration/qemu-file.c
+++ b/migration/qemu-file.c
@@ -337,7 +337,6 @@ static ssize_t coroutine_mixed_fn qemu_fill_buffer(QEMUFile *f)
 
     if (len > 0) {
         f->buf_size += len;
-        f->total_transferred += len;
     } else if (len == 0) {
         qemu_file_set_error_obj(f, -EIO, local_error);
     } else {
@@ -627,6 +626,8 @@ uint64_t qemu_file_transferred_noflush(QEMUFile *f)
     uint64_t ret = f->total_transferred;
     int i;
 
+    g_assert(qemu_file_is_writable(f));
+
     for (i = 0; i < f->iovcnt; i++) {
         ret += f->iov[i].iov_len;
     }
@@ -636,6 +637,7 @@ uint64_t qemu_file_transferred_noflush(QEMUFile *f)
 
 uint64_t qemu_file_transferred(QEMUFile *f)
 {
+    g_assert(qemu_file_is_writable(f));
     qemu_fflush(f);
     return f->total_transferred;
 }
-- 
2.41.0

[PATCH 02/12] qemu_file: Use a stat64 for qemu_file_transferred

[Juan Quintela]

This way we can read it from any thread.
I checked that it gives the same value than the current one.  We never
use to qemu_files at the same time.

Signed-off-by: Juan Quintela <quintela@redhat.com>
---
 migration/migration-stats.h | 4 ++++
 migration/qemu-file.c       | 5 +++--
 2 files changed, 7 insertions(+), 2 deletions(-)

diff --git a/migration/migration-stats.h b/migration/migration-stats.h
index 2358caad63..b7795e7914 100644
--- a/migration/migration-stats.h
+++ b/migration/migration-stats.h
@@ -81,6 +81,10 @@ typedef struct {
      * Number of bytes sent during precopy stage.
      */
     Stat64 precopy_bytes;
+    /*
+     * Number of bytes transferred with QEMUFile.
+     */
+    Stat64 qemu_file_transferred;
     /*
      * Amount of transferred data at the start of current cycle.
      */
diff --git a/migration/qemu-file.c b/migration/qemu-file.c
index 6814c562e6..384985f534 100644
--- a/migration/qemu-file.c
+++ b/migration/qemu-file.c
@@ -283,6 +283,7 @@ void qemu_fflush(QEMUFile *f)
         } else {
             uint64_t size = iov_size(f->iov, f->iovcnt);
             f->total_transferred += size;
+            stat64_add(&mig_stats.qemu_file_transferred, size);
         }
 
         qemu_iovec_release_ram(f);
@@ -623,7 +624,7 @@ int coroutine_mixed_fn qemu_get_byte(QEMUFile *f)
 
 uint64_t qemu_file_transferred_noflush(QEMUFile *f)
 {
-    uint64_t ret = f->total_transferred;
+    uint64_t ret = stat64_get(&mig_stats.qemu_file_transferred);
     int i;
 
     g_assert(qemu_file_is_writable(f));
@@ -639,7 +640,7 @@ uint64_t qemu_file_transferred(QEMUFile *f)
 {
     g_assert(qemu_file_is_writable(f));
     qemu_fflush(f);
-    return f->total_transferred;
+    return stat64_get(&mig_stats.qemu_file_transferred);
 }
 
 void qemu_put_be16(QEMUFile *f, unsigned int v)
-- 
2.41.0

[PATCH 03/12] qemu_file: total_transferred is not used anymore

[Juan Quintela]

Signed-off-by: Juan Quintela <quintela@redhat.com>
---
 migration/qemu-file.c | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/migration/qemu-file.c b/migration/qemu-file.c
index 384985f534..641ab703cc 100644
--- a/migration/qemu-file.c
+++ b/migration/qemu-file.c
@@ -41,9 +41,6 @@ struct QEMUFile {
     QIOChannel *ioc;
     bool is_writable;
 
-    /* The sum of bytes transferred on the wire */
-    uint64_t total_transferred;
-
     int buf_index;
     int buf_size; /* 0 when writing */
     uint8_t buf[IO_BUF_SIZE];
@@ -282,7 +279,6 @@ void qemu_fflush(QEMUFile *f)
             qemu_file_set_error_obj(f, -EIO, local_error);
         } else {
             uint64_t size = iov_size(f->iov, f->iovcnt);
-            f->total_transferred += size;
             stat64_add(&mig_stats.qemu_file_transferred, size);
         }
 
-- 
2.41.0

[PATCH 04/12] migration: Use the number of transferred bytes directly

[Juan Quintela]

We only use migration_transferred_bytes() to calculate the rate_limit,
for that we don't need to flush whatever is on the qemu_file buffer.
Remember that the buffer is really small (normal case is 32K if we use
iov's can be 64 * TARGET_PAGE_SIZE), so this is not relevant to
calculations.

Signed-off-by: Juan Quintela <quintela@redhat.com>
---
 migration/migration-stats.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/migration/migration-stats.c b/migration/migration-stats.c
index 4cc989d975..1d9197b4c3 100644
--- a/migration/migration-stats.c
+++ b/migration/migration-stats.c
@@ -63,7 +63,7 @@ uint64_t migration_transferred_bytes(QEMUFile *f)
 {
     uint64_t multifd = stat64_get(&mig_stats.multifd_bytes);
     uint64_t rdma = stat64_get(&mig_stats.rdma_bytes);
-    uint64_t qemu_file = qemu_file_transferred(f);
+    uint64_t qemu_file = stat64_get(&mig_stats.qemu_file_transferred);
 
     trace_migration_transferred_bytes(qemu_file, multifd, rdma);
     return qemu_file + multifd + rdma;
-- 
2.41.0

[PATCH 05/12] qemu_file: Remove unused qemu_file_transferred()

[Juan Quintela]

Signed-off-by: Juan Quintela <quintela@redhat.com>
---
 migration/qemu-file.h | 18 ------------------
 migration/qemu-file.c |  7 -------
 2 files changed, 25 deletions(-)

diff --git a/migration/qemu-file.h b/migration/qemu-file.h
index a29c37b0d0..8b71152754 100644
--- a/migration/qemu-file.h
+++ b/migration/qemu-file.h
@@ -33,24 +33,6 @@ QEMUFile *qemu_file_new_input(QIOChannel *ioc);
 QEMUFile *qemu_file_new_output(QIOChannel *ioc);
 int qemu_fclose(QEMUFile *f);
 
-/*
- * qemu_file_transferred:
- *
- * Report the total number of bytes transferred with
- * this file.
- *
- * For writable files, any pending buffers will be
- * flushed, so the reported value will be equal to
- * the number of bytes transferred on the wire.
- *
- * For readable files, the reported value will be
- * equal to the number of bytes transferred on the
- * wire.
- *
- * Returns: the total bytes transferred
- */
-uint64_t qemu_file_transferred(QEMUFile *f);
-
 /*
  * qemu_file_transferred_noflush:
  *
diff --git a/migration/qemu-file.c b/migration/qemu-file.c
index 641ab703cc..efa5f11033 100644
--- a/migration/qemu-file.c
+++ b/migration/qemu-file.c
@@ -632,13 +632,6 @@ uint64_t qemu_file_transferred_noflush(QEMUFile *f)
     return ret;
 }
 
-uint64_t qemu_file_transferred(QEMUFile *f)
-{
-    g_assert(qemu_file_is_writable(f));
-    qemu_fflush(f);
-    return stat64_get(&mig_stats.qemu_file_transferred);
-}
-
 void qemu_put_be16(QEMUFile *f, unsigned int v)
 {
     qemu_put_byte(f, v >> 8);
-- 
2.41.0

[PATCH 06/12] qemu-file: Remove _noflush from qemu_file_transferred_noflush()

[Juan Quintela]

qemu_file_transferred() don't exist anymore, so we can reuse the name.

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

---

v2: Update the documentation (thanks fabiano)
---
 migration/qemu-file.h | 9 ++++-----
 migration/block.c     | 4 ++--
 migration/qemu-file.c | 2 +-
 migration/savevm.c    | 6 +++---
 migration/vmstate.c   | 4 ++--
 5 files changed, 12 insertions(+), 13 deletions(-)

diff --git a/migration/qemu-file.h b/migration/qemu-file.h
index 8b71152754..1b2f6b8d8f 100644
--- a/migration/qemu-file.h
+++ b/migration/qemu-file.h
@@ -34,15 +34,14 @@ QEMUFile *qemu_file_new_output(QIOChannel *ioc);
 int qemu_fclose(QEMUFile *f);
 
 /*
- * qemu_file_transferred_noflush:
+ * qemu_file_transferred:
  *
- * As qemu_file_transferred except for writable files, where no flush
- * is performed and the reported amount will include the size of any
- * queued buffers, on top of the amount actually transferred.
+ * No flush is performed and the reported amount will include the size
+ * of any queued buffers, on top of the amount actually transferred.
  *
  * Returns: the total bytes transferred and queued
  */
-uint64_t qemu_file_transferred_noflush(QEMUFile *f);
+uint64_t qemu_file_transferred(QEMUFile *f);
 
 /*
  * put_buffer without copying the buffer.
diff --git a/migration/block.c b/migration/block.c
index b60698d6e2..47f11d0e4f 100644
--- a/migration/block.c
+++ b/migration/block.c
@@ -752,7 +752,7 @@ static int block_save_setup(QEMUFile *f, void *opaque)
 static int block_save_iterate(QEMUFile *f, void *opaque)
 {
     int ret;
-    uint64_t last_bytes = qemu_file_transferred_noflush(f);
+    uint64_t last_bytes = qemu_file_transferred(f);
 
     trace_migration_block_save("iterate", block_mig_state.submitted,
                                block_mig_state.transferred);
@@ -804,7 +804,7 @@ static int block_save_iterate(QEMUFile *f, void *opaque)
     }
 
     qemu_put_be64(f, BLK_MIG_FLAG_EOS);
-    uint64_t delta_bytes = qemu_file_transferred_noflush(f) - last_bytes;
+    uint64_t delta_bytes = qemu_file_transferred(f) - last_bytes;
     return (delta_bytes > 0);
 }
 
diff --git a/migration/qemu-file.c b/migration/qemu-file.c
index efa5f11033..0158db2a54 100644
--- a/migration/qemu-file.c
+++ b/migration/qemu-file.c
@@ -618,7 +618,7 @@ int coroutine_mixed_fn qemu_get_byte(QEMUFile *f)
     return result;
 }
 
-uint64_t qemu_file_transferred_noflush(QEMUFile *f)
+uint64_t qemu_file_transferred(QEMUFile *f)
 {
     uint64_t ret = stat64_get(&mig_stats.qemu_file_transferred);
     int i;
diff --git a/migration/savevm.c b/migration/savevm.c
index 8622f229e5..9c90499609 100644
--- a/migration/savevm.c
+++ b/migration/savevm.c
@@ -927,9 +927,9 @@ static int vmstate_load(QEMUFile *f, SaveStateEntry *se)
 static void vmstate_save_old_style(QEMUFile *f, SaveStateEntry *se,
                                    JSONWriter *vmdesc)
 {
-    uint64_t old_offset = qemu_file_transferred_noflush(f);
+    uint64_t old_offset = qemu_file_transferred(f);
     se->ops->save_state(f, se->opaque);
-    uint64_t size = qemu_file_transferred_noflush(f) - old_offset;
+    uint64_t size = qemu_file_transferred(f) - old_offset;
 
     if (vmdesc) {
         json_writer_int64(vmdesc, "size", size);
@@ -3053,7 +3053,7 @@ bool save_snapshot(const char *name, bool overwrite, const char *vmstate,
         goto the_end;
     }
     ret = qemu_savevm_state(f, errp);
-    vm_state_size = qemu_file_transferred_noflush(f);
+    vm_state_size = qemu_file_transferred(f);
     ret2 = qemu_fclose(f);
     if (ret < 0) {
         goto the_end;
diff --git a/migration/vmstate.c b/migration/vmstate.c
index 1cf9e45b85..16420fa9a3 100644
--- a/migration/vmstate.c
+++ b/migration/vmstate.c
@@ -386,7 +386,7 @@ int vmstate_save_state_v(QEMUFile *f, const VMStateDescription *vmsd,
                 void *curr_elem = first_elem + size * i;
 
                 vmsd_desc_field_start(vmsd, vmdesc_loop, field, i, n_elems);
-                old_offset = qemu_file_transferred_noflush(f);
+                old_offset = qemu_file_transferred(f);
                 if (field->flags & VMS_ARRAY_OF_POINTER) {
                     assert(curr_elem);
                     curr_elem = *(void **)curr_elem;
@@ -416,7 +416,7 @@ int vmstate_save_state_v(QEMUFile *f, const VMStateDescription *vmsd,
                     return ret;
                 }
 
-                written_bytes = qemu_file_transferred_noflush(f) - old_offset;
+                written_bytes = qemu_file_transferred(f) - old_offset;
                 vmsd_desc_field_end(vmsd, vmdesc_loop, field, written_bytes, i);
 
                 /* Compressed arrays only care about the first element */
-- 
2.41.0

[PATCH 07/12] migration: migration_transferred_bytes() don't need the QEMUFile

[Juan Quintela]

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Juan Quintela <quintela@redhat.com>
---
 migration/migration-stats.h | 4 +---
 migration/migration-stats.c | 6 +++---
 migration/migration.c       | 6 +++---
 3 files changed, 7 insertions(+), 9 deletions(-)

diff --git a/migration/migration-stats.h b/migration/migration-stats.h
index b7795e7914..e3863bf9bb 100644
--- a/migration/migration-stats.h
+++ b/migration/migration-stats.h
@@ -137,11 +137,9 @@ void migration_rate_set(uint64_t new_rate);
 /**
  * migration_transferred_bytes: Return number of bytes transferred
  *
- * @f: QEMUFile used for main migration channel
- *
  * Returns how many bytes have we transferred since the beginning of
  * the migration.  It accounts for bytes sent through any migration
  * channel, multifd, qemu_file, rdma, ....
  */
-uint64_t migration_transferred_bytes(QEMUFile *f);
+uint64_t migration_transferred_bytes(void);
 #endif
diff --git a/migration/migration-stats.c b/migration/migration-stats.c
index 1d9197b4c3..4ae8c0c722 100644
--- a/migration/migration-stats.c
+++ b/migration/migration-stats.c
@@ -30,7 +30,7 @@ bool migration_rate_exceeded(QEMUFile *f)
     }
 
     uint64_t rate_limit_start = stat64_get(&mig_stats.rate_limit_start);
-    uint64_t rate_limit_current = migration_transferred_bytes(f);
+    uint64_t rate_limit_current = migration_transferred_bytes();
     uint64_t rate_limit_used = rate_limit_current - rate_limit_start;
 
     if (rate_limit_max > 0 && rate_limit_used > rate_limit_max) {
@@ -56,10 +56,10 @@ void migration_rate_set(uint64_t limit)
 
 void migration_rate_reset(QEMUFile *f)
 {
-    stat64_set(&mig_stats.rate_limit_start, migration_transferred_bytes(f));
+    stat64_set(&mig_stats.rate_limit_start, migration_transferred_bytes());
 }
 
-uint64_t migration_transferred_bytes(QEMUFile *f)
+uint64_t migration_transferred_bytes(void)
 {
     uint64_t multifd = stat64_get(&mig_stats.multifd_bytes);
     uint64_t rdma = stat64_get(&mig_stats.rdma_bytes);
diff --git a/migration/migration.c b/migration/migration.c
index 67547eb6a1..e199d2f50d 100644
--- a/migration/migration.c
+++ b/migration/migration.c
@@ -2694,7 +2694,7 @@ static MigThrError migration_detect_error(MigrationState *s)
 
 static void migration_calculate_complete(MigrationState *s)
 {
-    uint64_t bytes = migration_transferred_bytes(s->to_dst_file);
+    uint64_t bytes = migration_transferred_bytes();
     int64_t end_time = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
     int64_t transfer_time;
 
@@ -2720,7 +2720,7 @@ static void update_iteration_initial_status(MigrationState *s)
      * wrong speed calculation.
      */
     s->iteration_start_time = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
-    s->iteration_initial_bytes = migration_transferred_bytes(s->to_dst_file);
+    s->iteration_initial_bytes = migration_transferred_bytes();
     s->iteration_initial_pages = ram_get_total_transferred_pages();
 }
 
@@ -2739,7 +2739,7 @@ static void migration_update_counters(MigrationState *s,
     }
 
     switchover_bw = migrate_avail_switchover_bandwidth();
-    current_bytes = migration_transferred_bytes(s->to_dst_file);
+    current_bytes = migration_transferred_bytes();
     transferred = current_bytes - s->iteration_initial_bytes;
     time_spent = current_time - s->iteration_start_time;
     bandwidth = (double)transferred / time_spent;
-- 
2.41.0

[PATCH 08/12] migration: migration_rate_limit_reset() don't need the QEMUFile

[Juan Quintela]

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Juan Quintela <quintela@redhat.com>
---
 migration/migration-stats.h | 4 +---
 migration/migration-stats.c | 2 +-
 migration/migration.c       | 2 +-
 3 files changed, 3 insertions(+), 5 deletions(-)

diff --git a/migration/migration-stats.h b/migration/migration-stats.h
index e3863bf9bb..68f3939188 100644
--- a/migration/migration-stats.h
+++ b/migration/migration-stats.h
@@ -120,10 +120,8 @@ uint64_t migration_rate_get(void);
  * migration_rate_reset: Reset the rate limit counter.
  *
  * This is called when we know we start a new transfer cycle.
- *
- * @f: QEMUFile used for main migration channel
  */
-void migration_rate_reset(QEMUFile *f);
+void migration_rate_reset(void);
 
 /**
  * migration_rate_set: Set the maximum amount that can be transferred.
diff --git a/migration/migration-stats.c b/migration/migration-stats.c
index 4ae8c0c722..f690b98a03 100644
--- a/migration/migration-stats.c
+++ b/migration/migration-stats.c
@@ -54,7 +54,7 @@ void migration_rate_set(uint64_t limit)
     stat64_set(&mig_stats.rate_limit_max, limit / XFER_LIMIT_RATIO);
 }
 
-void migration_rate_reset(QEMUFile *f)
+void migration_rate_reset(void)
 {
     stat64_set(&mig_stats.rate_limit_start, migration_transferred_bytes());
 }
diff --git a/migration/migration.c b/migration/migration.c
index e199d2f50d..bb62244288 100644
--- a/migration/migration.c
+++ b/migration/migration.c
@@ -2775,7 +2775,7 @@ static void migration_update_counters(MigrationState *s,
             stat64_get(&mig_stats.dirty_bytes_last_sync) / expected_bw_per_ms;
     }
 
-    migration_rate_reset(s->to_dst_file);
+    migration_rate_reset();
 
     update_iteration_initial_status(s);
 
-- 
2.41.0

[PATCH 09/12] qemu-file: Simplify qemu_file_get_error()

[Juan Quintela]

If we pass a NULL error is the same that returning dirrectly the value.

Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
---
 migration/qemu-file.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/migration/qemu-file.c b/migration/qemu-file.c
index 0158db2a54..7e738743ce 100644
--- a/migration/qemu-file.c
+++ b/migration/qemu-file.c
@@ -204,7 +204,7 @@ void qemu_file_set_error_obj(QEMUFile *f, int ret, Error *err)
  */
 int qemu_file_get_error(QEMUFile *f)
 {
-    return qemu_file_get_error_obj(f, NULL);
+    return f->last_error;
 }
 
 /*
-- 
2.41.0

[PATCH 10/12] migration: Use migration_transferred_bytes()

[Juan Quintela]

There are only two differnces with the old value:

- the amount of QEMUFile that hasn't yet been flushed.  It can be
  discussed what is more exact, the new or the old one.
- the amount of transferred bytes that we forgot to account for (the
  newer is better, i.e. exact).

Notice that this two values are used to:
a - present to the user
b - calculate the rate_limit

So a few KB here and there is not going to make a difference.

Signed-off-by: Juan Quintela <quintela@redhat.com>
---
 migration/migration.c | 2 +-
 migration/ram.c       | 6 +++---
 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/migration/migration.c b/migration/migration.c
index bb62244288..a6cde985a2 100644
--- a/migration/migration.c
+++ b/migration/migration.c
@@ -942,7 +942,7 @@ static void populate_ram_info(MigrationInfo *info, MigrationState *s)
     size_t page_size = qemu_target_page_size();
 
     info->ram = g_malloc0(sizeof(*info->ram));
-    info->ram->transferred = stat64_get(&mig_stats.transferred);
+    info->ram->transferred = migration_transferred_bytes();
     info->ram->total = ram_bytes_total();
     info->ram->duplicate = stat64_get(&mig_stats.zero_pages);
     /* legacy value.  It is not used anymore */
diff --git a/migration/ram.c b/migration/ram.c
index 92769902bb..5ccf70333a 100644
--- a/migration/ram.c
+++ b/migration/ram.c
@@ -564,7 +564,7 @@ void mig_throttle_counter_reset(void)
 
     rs->time_last_bitmap_sync = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
     rs->num_dirty_pages_period = 0;
-    rs->bytes_xfer_prev = stat64_get(&mig_stats.transferred);
+    rs->bytes_xfer_prev = migration_transferred_bytes();
 }
 
 /**
@@ -1030,7 +1030,7 @@ static void migration_trigger_throttle(RAMState *rs)
 {
     uint64_t threshold = migrate_throttle_trigger_threshold();
     uint64_t bytes_xfer_period =
-        stat64_get(&mig_stats.transferred) - rs->bytes_xfer_prev;
+        migration_transferred_bytes() - rs->bytes_xfer_prev;
     uint64_t bytes_dirty_period = rs->num_dirty_pages_period * TARGET_PAGE_SIZE;
     uint64_t bytes_dirty_threshold = bytes_xfer_period * threshold / 100;
 
@@ -1100,7 +1100,7 @@ static void migration_bitmap_sync(RAMState *rs, bool last_stage)
         /* reset period counters */
         rs->time_last_bitmap_sync = end_time;
         rs->num_dirty_pages_period = 0;
-        rs->bytes_xfer_prev = stat64_get(&mig_stats.transferred);
+        rs->bytes_xfer_prev = migration_transferred_bytes();
     }
     if (migrate_events()) {
         uint64_t generation = stat64_get(&mig_stats.dirty_sync_count);
-- 
2.41.0

[PATCH 11/12] migration: Remove transferred atomic counter

[Juan Quintela]

After last commit, it is a write only variable.

Signed-off-by: Juan Quintela <quintela@redhat.com>
---
 migration/migration-stats.h | 4 ----
 migration/multifd.c         | 3 ---
 migration/ram.c             | 1 -
 3 files changed, 8 deletions(-)

diff --git a/migration/migration-stats.h b/migration/migration-stats.h
index 68f3939188..05290ade76 100644
--- a/migration/migration-stats.h
+++ b/migration/migration-stats.h
@@ -97,10 +97,6 @@ typedef struct {
      * Number of bytes sent through RDMA.
      */
     Stat64 rdma_bytes;
-    /*
-     * Total number of bytes transferred.
-     */
-    Stat64 transferred;
     /*
      * Number of pages transferred that were full of zeros.
      */
diff --git a/migration/multifd.c b/migration/multifd.c
index e2a45c667a..ec58c58082 100644
--- a/migration/multifd.c
+++ b/migration/multifd.c
@@ -188,7 +188,6 @@ static int multifd_send_initial_packet(MultiFDSendParams *p, Error **errp)
         return -1;
     }
     stat64_add(&mig_stats.multifd_bytes, size);
-    stat64_add(&mig_stats.transferred, size);
     return 0;
 }
 
@@ -733,8 +732,6 @@ static void *multifd_send_thread(void *opaque)
 
             stat64_add(&mig_stats.multifd_bytes,
                        p->next_packet_size + p->packet_len);
-            stat64_add(&mig_stats.transferred,
-                       p->next_packet_size + p->packet_len);
             p->next_packet_size = 0;
             qemu_mutex_lock(&p->mutex);
             p->pending_job--;
diff --git a/migration/ram.c b/migration/ram.c
index 5ccf70333a..6d2bf50614 100644
--- a/migration/ram.c
+++ b/migration/ram.c
@@ -455,7 +455,6 @@ void ram_transferred_add(uint64_t bytes)
     } else {
         stat64_add(&mig_stats.downtime_bytes, bytes);
     }
-    stat64_add(&mig_stats.transferred, bytes);
 }
 
 struct MigrationOps {
-- 
2.41.0

[PATCH 12/12] qemu-file: Make qemu_fflush() return errors

[Juan Quintela]

This let us simplify code of this shape.

   qemu_fflush(f);
   int ret = qemu_file_get_error(f);
   if (ret) {
      return ret;
   }

into:

   int ret = qemu_fflush(f);
   if (ret) {
      return ret;
   }

I updated all callers where there is any error check.
qemu_fclose() don't need to check for f->last_error because
qemu_fflush() returns it at the beggining of the function.

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Juan Quintela <quintela@redhat.com>

---

In v2: Now that we call always qemu_fflush() for all files, we can
simplify qemu_fclose()
---
 migration/qemu-file.h |  2 +-
 migration/colo.c      | 11 +++--------
 migration/migration.c |  7 +------
 migration/qemu-file.c | 23 +++++++----------------
 migration/ram.c       | 22 +++++++---------------
 migration/rdma.c      |  4 +---
 migration/savevm.c    |  3 +--
 7 files changed, 21 insertions(+), 51 deletions(-)

diff --git a/migration/qemu-file.h b/migration/qemu-file.h
index 1b2f6b8d8f..1774116f79 100644
--- a/migration/qemu-file.h
+++ b/migration/qemu-file.h
@@ -71,7 +71,7 @@ void qemu_file_set_error_obj(QEMUFile *f, int ret, Error *err);
 void qemu_file_set_error(QEMUFile *f, int ret);
 int qemu_file_shutdown(QEMUFile *f);
 QEMUFile *qemu_file_get_return_path(QEMUFile *f);
-void qemu_fflush(QEMUFile *f);
+int qemu_fflush(QEMUFile *f);
 void qemu_file_set_blocking(QEMUFile *f, bool block);
 int qemu_file_get_to_fd(QEMUFile *f, int fd, size_t size);
 
diff --git a/migration/colo.c b/migration/colo.c
index 72f4f7b37e..4447e34914 100644
--- a/migration/colo.c
+++ b/migration/colo.c
@@ -314,9 +314,7 @@ static void colo_send_message(QEMUFile *f, COLOMessage msg,
         return;
     }
     qemu_put_be32(f, msg);
-    qemu_fflush(f);
-
-    ret = qemu_file_get_error(f);
+    ret = qemu_fflush(f);
     if (ret < 0) {
         error_setg_errno(errp, -ret, "Can't send COLO message");
     }
@@ -335,9 +333,7 @@ static void colo_send_message_value(QEMUFile *f, COLOMessage msg,
         return;
     }
     qemu_put_be64(f, value);
-    qemu_fflush(f);
-
-    ret = qemu_file_get_error(f);
+    ret = qemu_fflush(f);
     if (ret < 0) {
         error_setg_errno(errp, -ret, "Failed to send value for message:%s",
                          COLOMessage_str(msg));
@@ -483,8 +479,7 @@ static int colo_do_checkpoint_transaction(MigrationState *s,
     }
 
     qemu_put_buffer(s->to_dst_file, bioc->data, bioc->usage);
-    qemu_fflush(s->to_dst_file);
-    ret = qemu_file_get_error(s->to_dst_file);
+    ret = qemu_fflush(s->to_dst_file);
     if (ret < 0) {
         goto out;
     }
diff --git a/migration/migration.c b/migration/migration.c
index a6cde985a2..ce12fca520 100644
--- a/migration/migration.c
+++ b/migration/migration.c
@@ -305,12 +305,7 @@ static int migrate_send_rp_message(MigrationIncomingState *mis,
     qemu_put_be16(mis->to_src_file, (unsigned int)message_type);
     qemu_put_be16(mis->to_src_file, len);
     qemu_put_buffer(mis->to_src_file, data, len);
-    qemu_fflush(mis->to_src_file);
-
-    /* It's possible that qemu file got error during sending */
-    ret = qemu_file_get_error(mis->to_src_file);
-
-    return ret;
+    return qemu_fflush(mis->to_src_file);
 }
 
 /* Request one page from the source VM at the given start address.
diff --git a/migration/qemu-file.c b/migration/qemu-file.c
index 7e738743ce..d64500310d 100644
--- a/migration/qemu-file.c
+++ b/migration/qemu-file.c
@@ -262,14 +262,14 @@ static void qemu_iovec_release_ram(QEMUFile *f)
  * This will flush all pending data. If data was only partially flushed, it
  * will set an error state.
  */
-void qemu_fflush(QEMUFile *f)
+int qemu_fflush(QEMUFile *f)
 {
     if (!qemu_file_is_writable(f)) {
-        return;
+        return f->last_error;
     }
 
-    if (qemu_file_get_error(f)) {
-        return;
+    if (f->last_error) {
+        return f->last_error;
     }
     if (f->iovcnt > 0) {
         Error *local_error = NULL;
@@ -287,6 +287,7 @@ void qemu_fflush(QEMUFile *f)
 
     f->buf_index = 0;
     f->iovcnt = 0;
+    return f->last_error;
 }
 
 /*
@@ -353,22 +354,12 @@ static ssize_t coroutine_mixed_fn qemu_fill_buffer(QEMUFile *f)
  */
 int qemu_fclose(QEMUFile *f)
 {
-    int ret, ret2;
-    qemu_fflush(f);
-    ret = qemu_file_get_error(f);
-
-    ret2 = qio_channel_close(f->ioc, NULL);
+    int ret = qemu_fflush(f);
+    int ret2 = qio_channel_close(f->ioc, NULL);
     if (ret >= 0) {
         ret = ret2;
     }
     g_clear_pointer(&f->ioc, object_unref);
-
-    /* If any error was spotted before closing, we should report it
-     * instead of the close() return value.
-     */
-    if (f->last_error) {
-        ret = f->last_error;
-    }
     error_free(f->last_error_obj);
     g_free(f);
     trace_qemu_file_fclose();
diff --git a/migration/ram.c b/migration/ram.c
index 6d2bf50614..3a1d9882ce 100644
--- a/migration/ram.c
+++ b/migration/ram.c
@@ -305,17 +305,15 @@ int64_t ramblock_recv_bitmap_send(QEMUFile *file,
 
     qemu_put_be64(file, size);
     qemu_put_buffer(file, (const uint8_t *)le_bitmap, size);
+    g_free(le_bitmap);
     /*
      * Mark as an end, in case the middle part is screwed up due to
      * some "mysterious" reason.
      */
     qemu_put_be64(file, RAMBLOCK_RECV_BITMAP_ENDING);
-    qemu_fflush(file);
-
-    g_free(le_bitmap);
-
-    if (qemu_file_get_error(file)) {
-        return qemu_file_get_error(file);
+    int ret = qemu_fflush(file);
+    if (ret) {
+        return ret;
     }
 
     return size + sizeof(size);
@@ -3055,9 +3053,7 @@ static int ram_save_setup(QEMUFile *f, void *opaque)
     }
 
     qemu_put_be64(f, RAM_SAVE_FLAG_EOS);
-    qemu_fflush(f);
-
-    return 0;
+    return qemu_fflush(f);
 }
 
 /**
@@ -3176,10 +3172,8 @@ out:
         }
 
         qemu_put_be64(f, RAM_SAVE_FLAG_EOS);
-        qemu_fflush(f);
         ram_transferred_add(8);
-
-        ret = qemu_file_get_error(f);
+        ret = qemu_fflush(f);
     }
     if (ret < 0) {
         return ret;
@@ -3256,9 +3250,7 @@ static int ram_save_complete(QEMUFile *f, void *opaque)
         qemu_put_be64(f, RAM_SAVE_FLAG_MULTIFD_FLUSH);
     }
     qemu_put_be64(f, RAM_SAVE_FLAG_EOS);
-    qemu_fflush(f);
-
-    return 0;
+    return qemu_fflush(f);
 }
 
 static void ram_state_pending_estimate(void *opaque, uint64_t *must_precopy,
diff --git a/migration/rdma.c b/migration/rdma.c
index 2a1852ec7f..0884333190 100644
--- a/migration/rdma.c
+++ b/migration/rdma.c
@@ -3849,9 +3849,7 @@ int rdma_registration_start(QEMUFile *f, uint64_t flags)
 
     trace_rdma_registration_start(flags);
     qemu_put_be64(f, RAM_SAVE_FLAG_HOOK);
-    qemu_fflush(f);
-
-    return 0;
+    return qemu_fflush(f);
 }
 
 /*
diff --git a/migration/savevm.c b/migration/savevm.c
index 9c90499609..b19d625259 100644
--- a/migration/savevm.c
+++ b/migration/savevm.c
@@ -1583,8 +1583,7 @@ int qemu_savevm_state_complete_precopy(QEMUFile *f, bool iterable_only,
     }
 
 flush:
-    qemu_fflush(f);
-    return 0;
+    return qemu_fflush(f);
 }
 
 /* Give an estimate of the amount left to be transferred,
-- 
2.41.0

Re: [PATCH 01/12] qemu-file: We only call qemu_file_transferred_* on the sending side

[Fabiano Rosas]

Juan Quintela <quintela@redhat.com> writes:

> Remove the increase in qemu_file_fill_buffer() and add asserts to
> qemu_file_transferred* functions.

Patch looks ok, but I would rewrite the whole commit message like this:

Don't increment qemu_file_transferred at qemu_file_fill_buffer

We only call qemu_file_transferred_* on the sending side. Remove the
increment at qemu_file_fill_buffer() and add asserts to
qemu_file_transferred* functions.

Re: [PATCH 02/12] qemu_file: Use a stat64 for qemu_file_transferred

[Fabiano Rosas]

Juan Quintela <quintela@redhat.com> writes:

> This way we can read it from any thread.
> I checked that it gives the same value than the current one.  We never
> use to qemu_files at the same time.
>
> Signed-off-by: Juan Quintela <quintela@redhat.com>

Reviewed-by: Fabiano Rosas <farosas@suse.de>

Re: [PATCH 03/12] qemu_file: total_transferred is not used anymore

[Fabiano Rosas]

Juan Quintela <quintela@redhat.com> writes:

> Signed-off-by: Juan Quintela <quintela@redhat.com>
> ---
>  migration/qemu-file.c | 4 ----
>  1 file changed, 4 deletions(-)
>
> diff --git a/migration/qemu-file.c b/migration/qemu-file.c
> index 384985f534..641ab703cc 100644
> --- a/migration/qemu-file.c
> +++ b/migration/qemu-file.c
> @@ -41,9 +41,6 @@ struct QEMUFile {
>      QIOChannel *ioc;
>      bool is_writable;
>  
> -    /* The sum of bytes transferred on the wire */
> -    uint64_t total_transferred;
> -
>      int buf_index;
>      int buf_size; /* 0 when writing */
>      uint8_t buf[IO_BUF_SIZE];
> @@ -282,7 +279,6 @@ void qemu_fflush(QEMUFile *f)
>              qemu_file_set_error_obj(f, -EIO, local_error);
>          } else {
>              uint64_t size = iov_size(f->iov, f->iovcnt);
> -            f->total_transferred += size;
>              stat64_add(&mig_stats.qemu_file_transferred, size);
>          }

Reviewed-by: Fabiano Rosas <farosas@suse.de>

Re: [PATCH 04/12] migration: Use the number of transferred bytes directly

[Fabiano Rosas]

Juan Quintela <quintela@redhat.com> writes:

> We only use migration_transferred_bytes() to calculate the rate_limit,
> for that we don't need to flush whatever is on the qemu_file buffer.
> Remember that the buffer is really small (normal case is 32K if we use
> iov's can be 64 * TARGET_PAGE_SIZE), so this is not relevant to
> calculations.
>
> Signed-off-by: Juan Quintela <quintela@redhat.com>
> ---
>  migration/migration-stats.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/migration/migration-stats.c b/migration/migration-stats.c
> index 4cc989d975..1d9197b4c3 100644
> --- a/migration/migration-stats.c
> +++ b/migration/migration-stats.c
> @@ -63,7 +63,7 @@ uint64_t migration_transferred_bytes(QEMUFile *f)
>  {
>      uint64_t multifd = stat64_get(&mig_stats.multifd_bytes);
>      uint64_t rdma = stat64_get(&mig_stats.rdma_bytes);
> -    uint64_t qemu_file = qemu_file_transferred(f);
> +    uint64_t qemu_file = stat64_get(&mig_stats.qemu_file_transferred);
>  
>      trace_migration_transferred_bytes(qemu_file, multifd, rdma);
>      return qemu_file + multifd + rdma;

Reviewed-by: Fabiano Rosas <farosas@suse.de>

Re: [PATCH 05/12] qemu_file: Remove unused qemu_file_transferred()

[Fabiano Rosas]

Juan Quintela <quintela@redhat.com> writes:

> Signed-off-by: Juan Quintela <quintela@redhat.com>
> ---
>  migration/qemu-file.h | 18 ------------------
>  migration/qemu-file.c |  7 -------
>  2 files changed, 25 deletions(-)
>
> diff --git a/migration/qemu-file.h b/migration/qemu-file.h
> index a29c37b0d0..8b71152754 100644
> --- a/migration/qemu-file.h
> +++ b/migration/qemu-file.h
> @@ -33,24 +33,6 @@ QEMUFile *qemu_file_new_input(QIOChannel *ioc);
>  QEMUFile *qemu_file_new_output(QIOChannel *ioc);
>  int qemu_fclose(QEMUFile *f);
>  
> -/*
> - * qemu_file_transferred:
> - *
> - * Report the total number of bytes transferred with
> - * this file.
> - *
> - * For writable files, any pending buffers will be
> - * flushed, so the reported value will be equal to
> - * the number of bytes transferred on the wire.
> - *
> - * For readable files, the reported value will be
> - * equal to the number of bytes transferred on the
> - * wire.
> - *
> - * Returns: the total bytes transferred
> - */
> -uint64_t qemu_file_transferred(QEMUFile *f);
> -
>  /*
>   * qemu_file_transferred_noflush:
>   *
> diff --git a/migration/qemu-file.c b/migration/qemu-file.c
> index 641ab703cc..efa5f11033 100644
> --- a/migration/qemu-file.c
> +++ b/migration/qemu-file.c
> @@ -632,13 +632,6 @@ uint64_t qemu_file_transferred_noflush(QEMUFile *f)
>      return ret;
>  }
>  
> -uint64_t qemu_file_transferred(QEMUFile *f)
> -{
> -    g_assert(qemu_file_is_writable(f));
> -    qemu_fflush(f);
> -    return stat64_get(&mig_stats.qemu_file_transferred);
> -}
> -
>  void qemu_put_be16(QEMUFile *f, unsigned int v)
>  {
>      qemu_put_byte(f, v >> 8);

Reviewed-by: Fabiano Rosas <farosas@suse.de>

Re: [PATCH 06/12] qemu-file: Remove _noflush from qemu_file_transferred_noflush()

[Fabiano Rosas]

Juan Quintela <quintela@redhat.com> writes:

> qemu_file_transferred() don't exist anymore, so we can reuse the name.
>
> Signed-off-by: Juan Quintela <quintela@redhat.com>
>
> ---
>
> v2: Update the documentation (thanks fabiano)
> ---
>  migration/qemu-file.h | 9 ++++-----
>  migration/block.c     | 4 ++--
>  migration/qemu-file.c | 2 +-
>  migration/savevm.c    | 6 +++---
>  migration/vmstate.c   | 4 ++--
>  5 files changed, 12 insertions(+), 13 deletions(-)
>
> diff --git a/migration/qemu-file.h b/migration/qemu-file.h
> index 8b71152754..1b2f6b8d8f 100644
> --- a/migration/qemu-file.h
> +++ b/migration/qemu-file.h
> @@ -34,15 +34,14 @@ QEMUFile *qemu_file_new_output(QIOChannel *ioc);
>  int qemu_fclose(QEMUFile *f);
>  
>  /*
> - * qemu_file_transferred_noflush:
> + * qemu_file_transferred:
>   *
> - * As qemu_file_transferred except for writable files, where no flush
> - * is performed and the reported amount will include the size of any
> - * queued buffers, on top of the amount actually transferred.
> + * No flush is performed and the reported amount will include the size
> + * of any queued buffers, on top of the amount actually transferred.
>   *
>   * Returns: the total bytes transferred and queued
>   */
> -uint64_t qemu_file_transferred_noflush(QEMUFile *f);
> +uint64_t qemu_file_transferred(QEMUFile *f);
>  
>  /*
>   * put_buffer without copying the buffer.
> diff --git a/migration/block.c b/migration/block.c
> index b60698d6e2..47f11d0e4f 100644
> --- a/migration/block.c
> +++ b/migration/block.c
> @@ -752,7 +752,7 @@ static int block_save_setup(QEMUFile *f, void *opaque)
>  static int block_save_iterate(QEMUFile *f, void *opaque)
>  {
>      int ret;
> -    uint64_t last_bytes = qemu_file_transferred_noflush(f);
> +    uint64_t last_bytes = qemu_file_transferred(f);
>  
>      trace_migration_block_save("iterate", block_mig_state.submitted,
>                                 block_mig_state.transferred);
> @@ -804,7 +804,7 @@ static int block_save_iterate(QEMUFile *f, void *opaque)
>      }
>  
>      qemu_put_be64(f, BLK_MIG_FLAG_EOS);
> -    uint64_t delta_bytes = qemu_file_transferred_noflush(f) - last_bytes;
> +    uint64_t delta_bytes = qemu_file_transferred(f) - last_bytes;
>      return (delta_bytes > 0);
>  }
>  
> diff --git a/migration/qemu-file.c b/migration/qemu-file.c
> index efa5f11033..0158db2a54 100644
> --- a/migration/qemu-file.c
> +++ b/migration/qemu-file.c
> @@ -618,7 +618,7 @@ int coroutine_mixed_fn qemu_get_byte(QEMUFile *f)
>      return result;
>  }
>  
> -uint64_t qemu_file_transferred_noflush(QEMUFile *f)
> +uint64_t qemu_file_transferred(QEMUFile *f)
>  {
>      uint64_t ret = stat64_get(&mig_stats.qemu_file_transferred);
>      int i;
> diff --git a/migration/savevm.c b/migration/savevm.c
> index 8622f229e5..9c90499609 100644
> --- a/migration/savevm.c
> +++ b/migration/savevm.c
> @@ -927,9 +927,9 @@ static int vmstate_load(QEMUFile *f, SaveStateEntry *se)
>  static void vmstate_save_old_style(QEMUFile *f, SaveStateEntry *se,
>                                     JSONWriter *vmdesc)
>  {
> -    uint64_t old_offset = qemu_file_transferred_noflush(f);
> +    uint64_t old_offset = qemu_file_transferred(f);
>      se->ops->save_state(f, se->opaque);
> -    uint64_t size = qemu_file_transferred_noflush(f) - old_offset;
> +    uint64_t size = qemu_file_transferred(f) - old_offset;
>  
>      if (vmdesc) {
>          json_writer_int64(vmdesc, "size", size);
> @@ -3053,7 +3053,7 @@ bool save_snapshot(const char *name, bool overwrite, const char *vmstate,
>          goto the_end;
>      }
>      ret = qemu_savevm_state(f, errp);
> -    vm_state_size = qemu_file_transferred_noflush(f);
> +    vm_state_size = qemu_file_transferred(f);
>      ret2 = qemu_fclose(f);
>      if (ret < 0) {
>          goto the_end;
> diff --git a/migration/vmstate.c b/migration/vmstate.c
> index 1cf9e45b85..16420fa9a3 100644
> --- a/migration/vmstate.c
> +++ b/migration/vmstate.c
> @@ -386,7 +386,7 @@ int vmstate_save_state_v(QEMUFile *f, const VMStateDescription *vmsd,
>                  void *curr_elem = first_elem + size * i;
>  
>                  vmsd_desc_field_start(vmsd, vmdesc_loop, field, i, n_elems);
> -                old_offset = qemu_file_transferred_noflush(f);
> +                old_offset = qemu_file_transferred(f);
>                  if (field->flags & VMS_ARRAY_OF_POINTER) {
>                      assert(curr_elem);
>                      curr_elem = *(void **)curr_elem;
> @@ -416,7 +416,7 @@ int vmstate_save_state_v(QEMUFile *f, const VMStateDescription *vmsd,
>                      return ret;
>                  }
>  
> -                written_bytes = qemu_file_transferred_noflush(f) - old_offset;
> +                written_bytes = qemu_file_transferred(f) - old_offset;
>                  vmsd_desc_field_end(vmsd, vmdesc_loop, field, written_bytes, i);
>  
>                  /* Compressed arrays only care about the first element */

Reviewed-by: Fabiano Rosas <farosas@suse.de>

Re: [PATCH 10/12] migration: Use migration_transferred_bytes()

[Fabiano Rosas]

Juan Quintela <quintela@redhat.com> writes:

> There are only two differnces with the old value:
>
> - the amount of QEMUFile that hasn't yet been flushed.  It can be
>   discussed what is more exact, the new or the old one.
> - the amount of transferred bytes that we forgot to account for (the
>   newer is better, i.e. exact).
>
> Notice that this two values are used to:
> a - present to the user
> b - calculate the rate_limit
>
> So a few KB here and there is not going to make a difference.
>
> Signed-off-by: Juan Quintela <quintela@redhat.com>

Reviewed-by: Fabiano Rosas <farosas@suse.de>

Re: [PATCH 11/12] migration: Remove transferred atomic counter

[Fabiano Rosas]

Juan Quintela <quintela@redhat.com> writes:

> After last commit, it is a write only variable.
>
> Signed-off-by: Juan Quintela <quintela@redhat.com>
> ---
>  migration/migration-stats.h | 4 ----
>  migration/multifd.c         | 3 ---
>  migration/ram.c             | 1 -
>  3 files changed, 8 deletions(-)
>
> diff --git a/migration/migration-stats.h b/migration/migration-stats.h
> index 68f3939188..05290ade76 100644
> --- a/migration/migration-stats.h
> +++ b/migration/migration-stats.h
> @@ -97,10 +97,6 @@ typedef struct {
>       * Number of bytes sent through RDMA.
>       */
>      Stat64 rdma_bytes;
> -    /*
> -     * Total number of bytes transferred.
> -     */
> -    Stat64 transferred;
>      /*
>       * Number of pages transferred that were full of zeros.
>       */
> diff --git a/migration/multifd.c b/migration/multifd.c
> index e2a45c667a..ec58c58082 100644
> --- a/migration/multifd.c
> +++ b/migration/multifd.c
> @@ -188,7 +188,6 @@ static int multifd_send_initial_packet(MultiFDSendParams *p, Error **errp)
>          return -1;
>      }
>      stat64_add(&mig_stats.multifd_bytes, size);
> -    stat64_add(&mig_stats.transferred, size);
>      return 0;
>  }
>  
> @@ -733,8 +732,6 @@ static void *multifd_send_thread(void *opaque)
>  
>              stat64_add(&mig_stats.multifd_bytes,
>                         p->next_packet_size + p->packet_len);
> -            stat64_add(&mig_stats.transferred,
> -                       p->next_packet_size + p->packet_len);
>              p->next_packet_size = 0;
>              qemu_mutex_lock(&p->mutex);
>              p->pending_job--;
> diff --git a/migration/ram.c b/migration/ram.c
> index 5ccf70333a..6d2bf50614 100644
> --- a/migration/ram.c
> +++ b/migration/ram.c
> @@ -455,7 +455,6 @@ void ram_transferred_add(uint64_t bytes)
>      } else {
>          stat64_add(&mig_stats.downtime_bytes, bytes);
>      }
> -    stat64_add(&mig_stats.transferred, bytes);
>  }
>  
>  struct MigrationOps {

Reviewed-by: Fabiano Rosas <farosas@suse.de>

Re: [PATCH 12/12] qemu-file: Make qemu_fflush() return errors

[Fabiano Rosas]

Juan Quintela <quintela@redhat.com> writes:

> This let us simplify code of this shape.
>
>    qemu_fflush(f);
>    int ret = qemu_file_get_error(f);
>    if (ret) {
>       return ret;
>    }
>
> into:
>
>    int ret = qemu_fflush(f);
>    if (ret) {
>       return ret;
>    }
>
> I updated all callers where there is any error check.
> qemu_fclose() don't need to check for f->last_error because
> qemu_fflush() returns it at the beggining of the function.
>
> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
> Signed-off-by: Juan Quintela <quintela@redhat.com>

Reviewed-by: Fabiano Rosas <farosas@suse.de>

Re: [PATCH 02/12] qemu_file: Use a stat64 for qemu_file_transferred

[Eric Blake]

On Tue, Oct 24, 2023 at 05:10:32PM +0200, Juan Quintela wrote:
> This way we can read it from any thread.
> I checked that it gives the same value than the current one.  We never

s/than/as/

> use to qemu_files at the same time.

s/to/two/

> 
> Signed-off-by: Juan Quintela <quintela@redhat.com>
> ---
>  migration/migration-stats.h | 4 ++++
>  migration/qemu-file.c       | 5 +++--
>  2 files changed, 7 insertions(+), 2 deletions(-)
> 

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.
Virtualization:  qemu.org | libguestfs.org

Re: [PATCH 01/12] qemu-file: We only call qemu_file_transferred_* on the sending side

[Juan Quintela]

Fabiano Rosas <farosas@suse.de> wrote:
> Juan Quintela <quintela@redhat.com> writes:
>
>> Remove the increase in qemu_file_fill_buffer() and add asserts to
>> qemu_file_transferred* functions.
>
> Patch looks ok, but I would rewrite the whole commit message like this:
>
> Don't increment qemu_file_transferred at qemu_file_fill_buffer
>
> We only call qemu_file_transferred_* on the sending side. Remove the
> increment at qemu_file_fill_buffer() and add asserts to
> qemu_file_transferred* functions.

Changed the commint text.

Thanks,

Re: [PATCH 02/12] qemu_file: Use a stat64 for qemu_file_transferred

[Juan Quintela]

Eric Blake <eblake@redhat.com> wrote:
> On Tue, Oct 24, 2023 at 05:10:32PM +0200, Juan Quintela wrote:
>> This way we can read it from any thread.
>> I checked that it gives the same value than the current one.  We never
>
> s/than/as/

Done

>> use to qemu_files at the same time.
>
> s/to/two/

Done

Thanks.