]> xenbits.xensource.com Git - qemu-xen.git/commitdiff
backup: Copy only dirty areas
authorMax Reitz <mreitz@redhat.com>
Thu, 1 Aug 2019 17:38:59 +0000 (19:38 +0200)
committerMax Reitz <mreitz@redhat.com>
Tue, 6 Aug 2019 11:17:01 +0000 (13:17 +0200)
The backup job must only copy areas that the copy_bitmap reports as
dirty.  This is always the case when using traditional non-offloading
backup, because it copies each cluster separately.  When offloading the
copy operation, we sometimes copy more than one cluster at a time, but
we only check whether the first one is dirty.

Therefore, whenever copy offloading is possible, the backup job
currently produces wrong output when the guest writes to an area of
which an inner part has already been backed up, because that inner part
will be re-copied.

Fixes: 9ded4a0114968e98b41494fc035ba14f84cdf700
Signed-off-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-id: 20190801173900.23851-2-mreitz@redhat.com
Cc: qemu-stable@nongnu.org
Signed-off-by: Max Reitz <mreitz@redhat.com>
block/backup.c

index 715e1d3be88488cba2746fb387cf7bc316175bbf..1ee271f9f17848cc053f36a113996a871a466932 100644 (file)
@@ -202,22 +202,31 @@ static int coroutine_fn backup_do_cow(BackupBlockJob *job,
     cow_request_begin(&cow_request, job, start, end);
 
     while (start < end) {
+        int64_t dirty_end;
+
         if (!hbitmap_get(job->copy_bitmap, start)) {
             trace_backup_do_cow_skip(job, start);
             start += job->cluster_size;
             continue; /* already copied */
         }
 
+        dirty_end = hbitmap_next_zero(job->copy_bitmap, start, (end - start));
+        if (dirty_end < 0) {
+            dirty_end = end;
+        }
+
         trace_backup_do_cow_process(job, start);
 
         if (job->use_copy_range) {
-            ret = backup_cow_with_offload(job, start, end, is_write_notifier);
+            ret = backup_cow_with_offload(job, start, dirty_end,
+                                          is_write_notifier);
             if (ret < 0) {
                 job->use_copy_range = false;
             }
         }
         if (!job->use_copy_range) {
-            ret = backup_cow_with_bounce_buffer(job, start, end, is_write_notifier,
+            ret = backup_cow_with_bounce_buffer(job, start, dirty_end,
+                                                is_write_notifier,
                                                 error_is_read, &bounce_buffer);
         }
         if (ret < 0) {