Bitmap lock/unlock were added to bdrv_enable_dirty_bitmap in
8b1402ce80d, but some places were not updated correspondingly, which
leads to trying to take this lock twice, which is dead-lock. Fix this.
Actually, iotest 199 (about dirty bitmap postcopy migration) is broken
now, and this fixes it.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-id:
20180625165745.25259-3-vsementsov@virtuozzo.com
Signed-off-by: John Snow <jsnow@redhat.com>
/* Called with BQL taken. */
void bdrv_dirty_bitmap_enable_successor(BdrvDirtyBitmap *bitmap)
{
+ assert(bitmap->mutex == bitmap->successor->mutex);
qemu_mutex_lock(bitmap->mutex);
- bdrv_enable_dirty_bitmap(bitmap->successor);
+ bdrv_enable_dirty_bitmap_locked(bitmap->successor);
qemu_mutex_unlock(bitmap->mutex);
}
DirtyBitmapLoadBitmapState *b = item->data;
if (b->migrated) {
- bdrv_enable_dirty_bitmap(b->bitmap);
+ bdrv_enable_dirty_bitmap_locked(b->bitmap);
} else {
bdrv_dirty_bitmap_enable_successor(b->bitmap);
}
if (enabled_bitmaps == NULL) {
/* in postcopy */
bdrv_reclaim_dirty_bitmap_locked(s->bs, s->bitmap, &error_abort);
- bdrv_enable_dirty_bitmap(s->bitmap);
+ bdrv_enable_dirty_bitmap_locked(s->bitmap);
} else {
/* target not started, successor must be empty */
int64_t count = bdrv_get_dirty_count(s->bitmap);