Merge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging

Block layer patches (rebased Stefan's pull request)

# gpg: Signature made Thu 12 Nov 2015 15:34:16 GMT using RSA key ID C88F2FD6
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>"

* remotes/kevin/tags/for-upstream: (43 commits)
  block: Update copyright of the accounting code
  scsi-disk: Account for failed operations
  macio: Account for failed operations
  ide: Account for failed and invalid operations
  atapi: Account for failed and invalid operations
  xen_disk: Account for failed and invalid operations
  virtio-blk: Account for failed and invalid operations
  nvme: Account for failed and invalid operations
  iotests: Add test for the block device statistics
  block: Use QEMU_CLOCK_VIRTUAL for the accounting code in qtest mode
  qemu-io: Account for failed, invalid and flush operations
  block: New option to define the intervals for collecting I/O statistics
  block: Add average I/O queue depth to BlockDeviceTimedStats
  block: Compute minimum, maximum and average I/O latencies
  block: Allow configuring whether to account failed and invalid ops
  block: Add statistics for failed and invalid I/O operations
  block: Add idle_time_ns to BlockDeviceStats
  util: Infrastructure for computing recent averages
  block: define 'clock_type' for the accounting code
  ide: Account for write operations correctly
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

8 files changed, 885 insertions, 3 deletions

diff --git a/tests/Makefile b/tests/Makefile
index 92969e8288..90c4141ac5 100644
--- a/tests/Makefile
+++ b/tests/Makefile
@@ -47,6 +47,8 @@ check-unit-y += tests/test-thread-pool$(EXESUF)
 gcov-files-test-thread-pool-y = thread-pool.c
 gcov-files-test-hbitmap-y = util/hbitmap.c
 check-unit-y += tests/test-hbitmap$(EXESUF)
+gcov-files-test-hbitmap-y = blockjob.c
+check-unit-y += tests/test-blockjob-txn$(EXESUF)
 check-unit-y += tests/test-x86-cpuid$(EXESUF)
 # all code tested by test-x86-cpuid is inside topology.h
 gcov-files-test-x86-cpuid-y =
@@ -81,6 +83,7 @@ check-unit-y += tests/test-crypto-cipher$(EXESUF)
 check-unit-$(CONFIG_GNUTLS) += tests/test-crypto-tlscredsx509$(EXESUF)
 check-unit-$(CONFIG_GNUTLS) += tests/test-crypto-tlssession$(EXESUF)
 check-unit-$(CONFIG_LINUX) += tests/test-qga$(EXESUF)
+check-unit-y += tests/test-timed-average$(EXESUF)
 
 check-block-$(CONFIG_POSIX) += tests/qemu-iotests-quick.sh
 
@@ -390,6 +393,7 @@ tests/test-coroutine$(EXESUF): tests/test-coroutine.o $(test-block-obj-y)
 tests/test-aio$(EXESUF): tests/test-aio.o $(test-block-obj-y)
 tests/test-rfifolock$(EXESUF): tests/test-rfifolock.o $(test-util-obj-y)
 tests/test-throttle$(EXESUF): tests/test-throttle.o $(test-block-obj-y)
+tests/test-blockjob-txn$(EXESUF): tests/test-blockjob-txn.o $(test-block-obj-y) $(test-util-obj-y)
 tests/test-thread-pool$(EXESUF): tests/test-thread-pool.o $(test-block-obj-y)
 tests/test-iov$(EXESUF): tests/test-iov.o $(test-util-obj-y)
 tests/test-hbitmap$(EXESUF): tests/test-hbitmap.o $(test-util-obj-y)
@@ -409,6 +413,9 @@ tests/test-vmstate$(EXESUF): tests/test-vmstate.o \
 	migration/vmstate.o migration/qemu-file.o migration/qemu-file-buf.o \
         migration/qemu-file-unix.o qjson.o \
 	$(test-qom-obj-y)
+tests/test-timed-average$(EXESUF): tests/test-timed-average.o qemu-timer.o \
+	libqemuutil.a stubs/clock-warp.o stubs/cpu-get-icount.o \
+	stubs/notify-event.o stubs/replay.o
 
 tests/test-qapi-types.c tests/test-qapi-types.h :\
 $(SRC_PATH)/tests/qapi-schema/qapi-schema-test.json $(SRC_PATH)/scripts/qapi-types.py $(qapi-py)
diff --git a/tests/qemu-iotests/124 b/tests/qemu-iotests/124
index 9ccd11809f..c928f0101b 100644
--- a/tests/qemu-iotests/124
+++ b/tests/qemu-iotests/124
@@ -36,6 +36,23 @@ def try_remove(img):
         pass
 
 
+def transaction_action(action, **kwargs):
+    return {
+        'type': action,
+        'data': dict((k.replace('_', '-'), v) for k, v in kwargs.iteritems())
+    }
+
+
+def transaction_bitmap_clear(node, name, **kwargs):
+    return transaction_action('block-dirty-bitmap-clear',
+                              node=node, name=name, **kwargs)
+
+
+def transaction_drive_backup(device, target, **kwargs):
+    return transaction_action('drive-backup', device=device, target=target,
+                              **kwargs)
+
+
 class Bitmap:
     def __init__(self, name, drive):
         self.name = name
@@ -122,9 +139,12 @@ class TestIncrementalBackup(iotests.QMPTestCase):
     def do_qmp_backup(self, error='Input/output error', **kwargs):
         res = self.vm.qmp('drive-backup', **kwargs)
         self.assert_qmp(res, 'return', {})
+        return self.wait_qmp_backup(kwargs['device'], error)
+
 
+    def wait_qmp_backup(self, device, error='Input/output error'):
         event = self.vm.event_wait(name="BLOCK_JOB_COMPLETED",
-                                   match={'data': {'device': kwargs['device']}})
+                                   match={'data': {'device': device}})
         self.assertNotEqual(event, None)
 
         try:
@@ -139,6 +159,12 @@ class TestIncrementalBackup(iotests.QMPTestCase):
             return False
 
 
+    def wait_qmp_backup_cancelled(self, device):
+        event = self.vm.event_wait(name='BLOCK_JOB_CANCELLED',
+                                   match={'data': {'device': device}})
+        self.assertNotEqual(event, None)
+
+
     def create_anchor_backup(self, drive=None):
         if drive is None:
             drive = self.drives[-1]
@@ -264,6 +290,43 @@ class TestIncrementalBackup(iotests.QMPTestCase):
         return self.do_incremental_simple(granularity=131072)
 
 
+    def test_incremental_transaction(self):
+        '''Test: Verify backups made from transactionally created bitmaps.
+
+        Create a bitmap "before" VM execution begins, then create a second
+        bitmap AFTER writes have already occurred. Use transactions to create
+        a full backup and synchronize both bitmaps to this backup.
+        Create an incremental backup through both bitmaps and verify that
+        both backups match the current drive0 image.
+        '''
+
+        drive0 = self.drives[0]
+        bitmap0 = self.add_bitmap('bitmap0', drive0)
+        self.hmp_io_writes(drive0['id'], (('0xab', 0, 512),
+                                          ('0xfe', '16M', '256k'),
+                                          ('0x64', '32736k', '64k')))
+        bitmap1 = self.add_bitmap('bitmap1', drive0)
+
+        result = self.vm.qmp('transaction', actions=[
+            transaction_bitmap_clear(bitmap0.drive['id'], bitmap0.name),
+            transaction_bitmap_clear(bitmap1.drive['id'], bitmap1.name),
+            transaction_drive_backup(drive0['id'], drive0['backup'],
+                                     sync='full', format=drive0['fmt'])
+        ])
+        self.assert_qmp(result, 'return', {})
+        self.wait_until_completed(drive0['id'])
+        self.files.append(drive0['backup'])
+
+        self.hmp_io_writes(drive0['id'], (('0x9a', 0, 512),
+                                          ('0x55', '8M', '352k'),
+                                          ('0x78', '15872k', '1M')))
+        # Both bitmaps should be correctly in sync.
+        self.create_incremental(bitmap0)
+        self.create_incremental(bitmap1)
+        self.vm.shutdown()
+        self.check_backups()
+
+
     def test_incremental_failure(self):
         '''Test: Verify backups made after a failure are correct.
 
@@ -321,6 +384,123 @@ class TestIncrementalBackup(iotests.QMPTestCase):
         self.check_backups()
 
 
+    def test_transaction_failure(self):
+        '''Test: Verify backups made from a transaction that partially fails.
+
+        Add a second drive with its own unique pattern, and add a bitmap to each
+        drive. Use blkdebug to interfere with the backup on just one drive and
+        attempt to create a coherent incremental backup across both drives.
+
+        verify a failure in one but not both, then delete the failed stubs and
+        re-run the same transaction.
+
+        verify that both incrementals are created successfully.
+        '''
+
+        # Create a second drive, with pattern:
+        drive1 = self.add_node('drive1')
+        self.img_create(drive1['file'], drive1['fmt'])
+        io_write_patterns(drive1['file'], (('0x14', 0, 512),
+                                           ('0x5d', '1M', '32k'),
+                                           ('0xcd', '32M', '124k')))
+
+        # Create a blkdebug interface to this img as 'drive1'
+        result = self.vm.qmp('blockdev-add', options={
+            'id': drive1['id'],
+            'driver': drive1['fmt'],
+            'file': {
+                'driver': 'blkdebug',
+                'image': {
+                    'driver': 'file',
+                    'filename': drive1['file']
+                },
+                'set-state': [{
+                    'event': 'flush_to_disk',
+                    'state': 1,
+                    'new_state': 2
+                }],
+                'inject-error': [{
+                    'event': 'read_aio',
+                    'errno': 5,
+                    'state': 2,
+                    'immediately': False,
+                    'once': True
+                }],
+            }
+        })
+        self.assert_qmp(result, 'return', {})
+
+        # Create bitmaps and full backups for both drives
+        drive0 = self.drives[0]
+        dr0bm0 = self.add_bitmap('bitmap0', drive0)
+        dr1bm0 = self.add_bitmap('bitmap0', drive1)
+        self.create_anchor_backup(drive0)
+        self.create_anchor_backup(drive1)
+        self.assert_no_active_block_jobs()
+        self.assertFalse(self.vm.get_qmp_events(wait=False))
+
+        # Emulate some writes
+        self.hmp_io_writes(drive0['id'], (('0xab', 0, 512),
+                                          ('0xfe', '16M', '256k'),
+                                          ('0x64', '32736k', '64k')))
+        self.hmp_io_writes(drive1['id'], (('0xba', 0, 512),
+                                          ('0xef', '16M', '256k'),
+                                          ('0x46', '32736k', '64k')))
+
+        # Create incremental backup targets
+        target0 = self.prepare_backup(dr0bm0)
+        target1 = self.prepare_backup(dr1bm0)
+
+        # Ask for a new incremental backup per-each drive,
+        # expecting drive1's backup to fail:
+        transaction = [
+            transaction_drive_backup(drive0['id'], target0, sync='incremental',
+                                     format=drive0['fmt'], mode='existing',
+                                     bitmap=dr0bm0.name),
+            transaction_drive_backup(drive1['id'], target1, sync='incremental',
+                                     format=drive1['fmt'], mode='existing',
+                                     bitmap=dr1bm0.name)
+        ]
+        result = self.vm.qmp('transaction', actions=transaction,
+                             properties={'completion-mode': 'grouped'} )
+        self.assert_qmp(result, 'return', {})
+
+        # Observe that drive0's backup is cancelled and drive1 completes with
+        # an error.
+        self.wait_qmp_backup_cancelled(drive0['id'])
+        self.assertFalse(self.wait_qmp_backup(drive1['id']))
+        error = self.vm.event_wait('BLOCK_JOB_ERROR')
+        self.assert_qmp(error, 'data', {'device': drive1['id'],
+                                        'action': 'report',
+                                        'operation': 'read'})
+        self.assertFalse(self.vm.get_qmp_events(wait=False))
+        self.assert_no_active_block_jobs()
+
+        # Delete drive0's successful target and eliminate our record of the
+        # unsuccessful drive1 target. Then re-run the same transaction.
+        dr0bm0.del_target()
+        dr1bm0.del_target()
+        target0 = self.prepare_backup(dr0bm0)
+        target1 = self.prepare_backup(dr1bm0)
+
+        # Re-run the exact same transaction.
+        result = self.vm.qmp('transaction', actions=transaction,
+                             properties={'completion-mode':'grouped'})
+        self.assert_qmp(result, 'return', {})
+
+        # Both should complete successfully this time.
+        self.assertTrue(self.wait_qmp_backup(drive0['id']))
+        self.assertTrue(self.wait_qmp_backup(drive1['id']))
+        self.make_reference_backup(dr0bm0)
+        self.make_reference_backup(dr1bm0)
+        self.assertFalse(self.vm.get_qmp_events(wait=False))
+        self.assert_no_active_block_jobs()
+
+        # And the images should of course validate.
+        self.vm.shutdown()
+        self.check_backups()
+
+
     def test_sync_dirty_bitmap_missing(self):
         self.assert_no_active_block_jobs()
         self.files.append(self.err_img)
diff --git a/tests/qemu-iotests/124.out b/tests/qemu-iotests/124.out
index 2f7d3902f2..dae404e278 100644
--- a/tests/qemu-iotests/124.out
+++ b/tests/qemu-iotests/124.out
@@ -1,5 +1,5 @@
-.......
+.........
 ----------------------------------------------------------------------
-Ran 7 tests
+Ran 9 tests
 
 OK
diff --git a/tests/qemu-iotests/136 b/tests/qemu-iotests/136
new file mode 100644
index 0000000000..f574d83ff7
--- /dev/null
+++ b/tests/qemu-iotests/136
@@ -0,0 +1,349 @@
+#!/usr/bin/env python
+#
+# Tests for block device statistics
+#
+# Copyright (C) 2015 Igalia, S.L.
+# Author: Alberto Garcia <berto@igalia.com>
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+#
+
+import iotests
+import os
+
+interval_length = 10
+nsec_per_sec = 1000000000
+op_latency = nsec_per_sec / 1000 # See qtest_latency_ns in accounting.c
+bad_sector = 8192
+bad_offset = bad_sector * 512
+blkdebug_file = os.path.join(iotests.test_dir, 'blkdebug.conf')
+
+class BlockDeviceStatsTestCase(iotests.QMPTestCase):
+    test_img = "null-aio://"
+    total_rd_bytes = 0
+    total_rd_ops = 0
+    total_wr_bytes = 0
+    total_wr_ops = 0
+    total_wr_merged = 0
+    total_flush_ops = 0
+    failed_rd_ops = 0
+    failed_wr_ops = 0
+    invalid_rd_ops = 0
+    invalid_wr_ops = 0
+    wr_highest_offset = 0
+    account_invalid = False
+    account_failed = False
+
+    def blockstats(self, device):
+        result = self.vm.qmp("query-blockstats")
+        for r in result['return']:
+            if r['device'] == device:
+                return r['stats']
+        raise Exception("Device not found for blockstats: %s" % device)
+
+    def create_blkdebug_file(self):
+        file = open(blkdebug_file, 'w')
+        file.write('''
+[inject-error]
+event = "read_aio"
+errno = "5"
+sector = "%d"
+
+[inject-error]
+event = "write_aio"
+errno = "5"
+sector = "%d"
+''' % (bad_sector, bad_sector))
+        file.close()
+
+    def setUp(self):
+        drive_args = []
+        drive_args.append("stats-intervals=%d" % interval_length)
+        drive_args.append("stats-account-invalid=%s" %
+                          (self.account_invalid and "on" or "off"))
+        drive_args.append("stats-account-failed=%s" %
+                          (self.account_failed and "on" or "off"))
+        self.create_blkdebug_file()
+        self.vm = iotests.VM().add_drive('blkdebug:%s:%s ' %
+                                         (blkdebug_file, self.test_img),
+                                         ','.join(drive_args))
+        self.vm.launch()
+        # Set an initial value for the clock
+        self.vm.qtest("clock_step %d" % nsec_per_sec)
+
+    def tearDown(self):
+        self.vm.shutdown()
+        os.remove(blkdebug_file)
+
+    def accounted_ops(self, read = False, write = False, flush = False):
+        ops = 0
+        if write:
+            ops += self.total_wr_ops
+            if self.account_failed:
+                ops += self.failed_wr_ops
+            if self.account_invalid:
+                ops += self.invalid_wr_ops
+        if read:
+            ops += self.total_rd_ops
+            if self.account_failed:
+                ops += self.failed_rd_ops
+            if self.account_invalid:
+                ops += self.invalid_rd_ops
+        if flush:
+            ops += self.total_flush_ops
+        return ops
+
+    def accounted_latency(self, read = False, write = False, flush = False):
+        latency = 0
+        if write:
+            latency += self.total_wr_ops * op_latency
+            if self.account_failed:
+                latency += self.failed_wr_ops * op_latency
+        if read:
+            latency += self.total_rd_ops * op_latency
+            if self.account_failed:
+                latency += self.failed_rd_ops * op_latency
+        if flush:
+            latency += self.total_flush_ops * op_latency
+        return latency
+
+    def check_values(self):
+        stats = self.blockstats('drive0')
+
+        # Check that the totals match with what we have calculated
+        self.assertEqual(self.total_rd_bytes, stats['rd_bytes'])
+        self.assertEqual(self.total_wr_bytes, stats['wr_bytes'])
+        self.assertEqual(self.total_rd_ops, stats['rd_operations'])
+        self.assertEqual(self.total_wr_ops, stats['wr_operations'])
+        self.assertEqual(self.total_flush_ops, stats['flush_operations'])
+        self.assertEqual(self.wr_highest_offset, stats['wr_highest_offset'])
+        self.assertEqual(self.failed_rd_ops, stats['failed_rd_operations'])
+        self.assertEqual(self.failed_wr_ops, stats['failed_wr_operations'])
+        self.assertEqual(self.invalid_rd_ops, stats['invalid_rd_operations'])
+        self.assertEqual(self.invalid_wr_ops, stats['invalid_wr_operations'])
+        self.assertEqual(self.account_invalid, stats['account_invalid'])
+        self.assertEqual(self.account_failed, stats['account_failed'])
+        self.assertEqual(self.total_wr_merged, stats['wr_merged'])
+
+        # Check that there's exactly one interval with the length we defined
+        self.assertEqual(1, len(stats['timed_stats']))
+        timed_stats = stats['timed_stats'][0]
+        self.assertEqual(interval_length, timed_stats['interval_length'])
+
+        total_rd_latency = self.accounted_latency(read = True)
+        if (total_rd_latency != 0):
+            self.assertEqual(total_rd_latency, stats['rd_total_time_ns'])
+            self.assertEqual(op_latency, timed_stats['min_rd_latency_ns'])
+            self.assertEqual(op_latency, timed_stats['max_rd_latency_ns'])
+            self.assertEqual(op_latency, timed_stats['avg_rd_latency_ns'])
+            self.assertLess(0, timed_stats['avg_rd_queue_depth'])
+        else:
+            self.assertEqual(0, stats['rd_total_time_ns'])
+            self.assertEqual(0, timed_stats['min_rd_latency_ns'])
+            self.assertEqual(0, timed_stats['max_rd_latency_ns'])
+            self.assertEqual(0, timed_stats['avg_rd_latency_ns'])
+            self.assertEqual(0, timed_stats['avg_rd_queue_depth'])
+
+        # min read latency <= avg read latency <= max read latency
+        self.assertLessEqual(timed_stats['min_rd_latency_ns'],
+                             timed_stats['avg_rd_latency_ns'])
+        self.assertLessEqual(timed_stats['avg_rd_latency_ns'],
+                             timed_stats['max_rd_latency_ns'])
+
+        total_wr_latency = self.accounted_latency(write = True)
+        if (total_wr_latency != 0):
+            self.assertEqual(total_wr_latency, stats['wr_total_time_ns'])
+            self.assertEqual(op_latency, timed_stats['min_wr_latency_ns'])
+            self.assertEqual(op_latency, timed_stats['max_wr_latency_ns'])
+            self.assertEqual(op_latency, timed_stats['avg_wr_latency_ns'])
+            self.assertLess(0, timed_stats['avg_wr_queue_depth'])
+        else:
+            self.assertEqual(0, stats['wr_total_time_ns'])
+            self.assertEqual(0, timed_stats['min_wr_latency_ns'])
+            self.assertEqual(0, timed_stats['max_wr_latency_ns'])
+            self.assertEqual(0, timed_stats['avg_wr_latency_ns'])
+            self.assertEqual(0, timed_stats['avg_wr_queue_depth'])
+
+        # min write latency <= avg write latency <= max write latency
+        self.assertLessEqual(timed_stats['min_wr_latency_ns'],
+                             timed_stats['avg_wr_latency_ns'])
+        self.assertLessEqual(timed_stats['avg_wr_latency_ns'],
+                             timed_stats['max_wr_latency_ns'])
+
+        total_flush_latency = self.accounted_latency(flush = True)
+        if (total_flush_latency != 0):
+            self.assertEqual(total_flush_latency, stats['flush_total_time_ns'])
+            self.assertEqual(op_latency, timed_stats['min_flush_latency_ns'])
+            self.assertEqual(op_latency, timed_stats['max_flush_latency_ns'])
+            self.assertEqual(op_latency, timed_stats['avg_flush_latency_ns'])
+        else:
+            self.assertEqual(0, stats['flush_total_time_ns'])
+            self.assertEqual(0, timed_stats['min_flush_latency_ns'])
+            self.assertEqual(0, timed_stats['max_flush_latency_ns'])
+            self.assertEqual(0, timed_stats['avg_flush_latency_ns'])
+
+        # min flush latency <= avg flush latency <= max flush latency
+        self.assertLessEqual(timed_stats['min_flush_latency_ns'],
+                             timed_stats['avg_flush_latency_ns'])
+        self.assertLessEqual(timed_stats['avg_flush_latency_ns'],
+                             timed_stats['max_flush_latency_ns'])
+
+        # idle_time_ns must be > 0 if we have performed any operation
+        if (self.accounted_ops(read = True, write = True, flush = True) != 0):
+            self.assertLess(0, stats['idle_time_ns'])
+        else:
+            self.assertFalse(stats.has_key('idle_time_ns'))
+
+        # This test does not alter these, so they must be all 0
+        self.assertEqual(0, stats['rd_merged'])
+        self.assertEqual(0, stats['failed_flush_operations'])
+        self.assertEqual(0, stats['invalid_flush_operations'])
+
+    def do_test_stats(self, rd_size = 0, rd_ops = 0, wr_size = 0, wr_ops = 0,
+                      flush_ops = 0, invalid_rd_ops = 0, invalid_wr_ops = 0,
+                      failed_rd_ops = 0, failed_wr_ops = 0, wr_merged = 0):
+        # The 'ops' list will contain all the requested I/O operations
+        ops = []
+        for i in range(rd_ops):
+            ops.append("aio_read %d %d" % (i * rd_size, rd_size))
+
+        for i in range(wr_ops):
+            ops.append("aio_write %d %d" % (i * wr_size, wr_size))
+
+        for i in range(flush_ops):
+            ops.append("aio_flush")
+
+        highest_offset = wr_ops * wr_size
+
+        # Two types of invalid operations: unaligned length and unaligned offset
+        for i in range(invalid_rd_ops / 2):
+            ops.append("aio_read 0 511")
+
+        for i in range(invalid_rd_ops / 2, invalid_rd_ops):
+            ops.append("aio_read 13 512")
+
+        for i in range(invalid_wr_ops / 2):
+            ops.append("aio_write 0 511")
+
+        for i in range(invalid_wr_ops / 2, invalid_wr_ops):
+            ops.append("aio_write 13 512")
+
+        for i in range(failed_rd_ops):
+            ops.append("aio_read %d 512" % bad_offset)
+
+        for i in range(failed_wr_ops):
+            ops.append("aio_write %d 512" % bad_offset)
+
+        if failed_wr_ops > 0:
+            highest_offset = max(highest_offset, bad_offset + 512)
+
+        for i in range(wr_merged):
+            first = i * wr_size * 2
+            second = first + wr_size
+            ops.append("multiwrite %d %d ; %d %d" %
+                       (first, wr_size, second, wr_size))
+
+        highest_offset = max(highest_offset, wr_merged * wr_size * 2)
+
+        # Now perform all operations
+        for op in ops:
+            self.vm.hmp_qemu_io("drive0", op)
+
+        # Update the expected totals
+        self.total_rd_bytes += rd_ops * rd_size
+        self.total_rd_ops += rd_ops
+        self.total_wr_bytes += wr_ops * wr_size
+        self.total_wr_ops += wr_ops
+        self.total_wr_merged += wr_merged
+        self.total_flush_ops += flush_ops
+        self.invalid_rd_ops += invalid_rd_ops
+        self.invalid_wr_ops += invalid_wr_ops
+        self.failed_rd_ops += failed_rd_ops
+        self.failed_wr_ops += failed_wr_ops
+
+        self.wr_highest_offset = max(self.wr_highest_offset, highest_offset)
+
+        # Advance the clock so idle_time_ns has a meaningful value
+        self.vm.qtest("clock_step %d" % nsec_per_sec)
+
+        # And check that the actual statistics match the expected ones
+        self.check_values()
+
+    def test_read_only(self):
+        test_values = [[512,    1],
+                       [65536,  1],
+                       [512,   12],
+                       [65536, 12]]
+        for i in test_values:
+            self.do_test_stats(rd_size = i[0], rd_ops = i[1])
+
+    def test_write_only(self):
+        test_values = [[512,    1],
+                       [65536,  1],
+                       [512,   12],
+                       [65536, 12]]
+        for i in test_values:
+            self.do_test_stats(wr_size = i[0], wr_ops = i[1])
+
+    def test_invalid(self):
+        self.do_test_stats(invalid_rd_ops = 7)
+        self.do_test_stats(invalid_wr_ops = 3)
+        self.do_test_stats(invalid_rd_ops = 4, invalid_wr_ops = 5)
+
+    def test_failed(self):
+        self.do_test_stats(failed_rd_ops = 8)
+        self.do_test_stats(failed_wr_ops = 6)
+        self.do_test_stats(failed_rd_ops = 5, failed_wr_ops = 12)
+
+    def test_flush(self):
+        self.do_test_stats(flush_ops = 8)
+
+    def test_merged(self):
+        for i in range(5):
+            self.do_test_stats(wr_merged = i * 3)
+
+    def test_all(self):
+        # rd_size, rd_ops, wr_size, wr_ops, flush_ops
+        # invalid_rd_ops,  invalid_wr_ops,
+        # failed_rd_ops,   failed_wr_ops
+        # wr_merged
+        test_values = [[512,    1, 512,   1, 1, 4, 7, 5, 2, 1],
+                       [65536,  1, 2048, 12, 7, 7, 5, 2, 5, 5],
+                       [32768,  9, 8192,  1, 4, 3, 2, 4, 6, 4],
+                       [16384, 11, 3584, 16, 9, 8, 6, 7, 3, 4]]
+        for i in test_values:
+            self.do_test_stats(*i)
+
+    def test_no_op(self):
+        # All values must be sane before doing any I/O
+        self.check_values()
+
+
+class BlockDeviceStatsTestAccountInvalid(BlockDeviceStatsTestCase):
+    account_invalid = True
+    account_failed = False
+
+class BlockDeviceStatsTestAccountFailed(BlockDeviceStatsTestCase):
+    account_invalid = False
+    account_failed = True
+
+class BlockDeviceStatsTestAccountBoth(BlockDeviceStatsTestCase):
+    account_invalid = True
+    account_failed = True
+
+class BlockDeviceStatsTestCoroutine(BlockDeviceStatsTestCase):
+    test_img = "null-co://"
+
+if __name__ == '__main__':
+    iotests.main(supported_fmts=["raw"])
diff --git a/tests/qemu-iotests/136.out b/tests/qemu-iotests/136.out
new file mode 100644
index 0000000000..0a5e9583a4
--- /dev/null
+++ b/tests/qemu-iotests/136.out
@@ -0,0 +1,5 @@
+........................................
+----------------------------------------------------------------------
+Ran 40 tests
+
+OK
diff --git a/tests/qemu-iotests/group b/tests/qemu-iotests/group
index c69265decd..5a0880893a 100644
--- a/tests/qemu-iotests/group
+++ b/tests/qemu-iotests/group
@@ -136,6 +136,7 @@
 132 rw auto quick
 134 rw auto quick
 135 rw auto
+136 rw auto
 137 rw auto
 138 rw auto quick
 139 rw auto quick
diff --git a/tests/test-blockjob-txn.c b/tests/test-blockjob-txn.c
new file mode 100644
index 0000000000..34747e924d
--- /dev/null
+++ b/tests/test-blockjob-txn.c
@@ -0,0 +1,250 @@
+/*
+ * Blockjob transactions tests
+ *
+ * Copyright Red Hat, Inc. 2015
+ *
+ * Authors:
+ *  Stefan Hajnoczi    <stefanha@redhat.com>
+ *
+ * This work is licensed under the terms of the GNU LGPL, version 2 or later.
+ * See the COPYING.LIB file in the top-level directory.
+ */
+
+#include <glib.h>
+#include "qapi/error.h"
+#include "qemu/main-loop.h"
+#include "block/blockjob.h"
+
+typedef struct {
+    BlockJob common;
+    unsigned int iterations;
+    bool use_timer;
+    int rc;
+    int *result;
+} TestBlockJob;
+
+static const BlockJobDriver test_block_job_driver = {
+    .instance_size = sizeof(TestBlockJob),
+};
+
+static void test_block_job_complete(BlockJob *job, void *opaque)
+{
+    BlockDriverState *bs = job->bs;
+    int rc = (intptr_t)opaque;
+
+    if (block_job_is_cancelled(job)) {
+        rc = -ECANCELED;
+    }
+
+    block_job_completed(job, rc);
+    bdrv_unref(bs);
+}
+
+static void coroutine_fn test_block_job_run(void *opaque)
+{
+    TestBlockJob *s = opaque;
+    BlockJob *job = &s->common;
+
+    while (s->iterations--) {
+        if (s->use_timer) {
+            block_job_sleep_ns(job, QEMU_CLOCK_REALTIME, 0);
+        } else {
+            block_job_yield(job);
+        }
+
+        if (block_job_is_cancelled(job)) {
+            break;
+        }
+    }
+
+    block_job_defer_to_main_loop(job, test_block_job_complete,
+                                 (void *)(intptr_t)s->rc);
+}
+
+typedef struct {
+    TestBlockJob *job;
+    int *result;
+} TestBlockJobCBData;
+
+static void test_block_job_cb(void *opaque, int ret)
+{
+    TestBlockJobCBData *data = opaque;
+    if (!ret && block_job_is_cancelled(&data->job->common)) {
+        ret = -ECANCELED;
+    }
+    *data->result = ret;
+    g_free(data);
+}
+
+/* Create a block job that completes with a given return code after a given
+ * number of event loop iterations.  The return code is stored in the given
+ * result pointer.
+ *
+ * The event loop iterations can either be handled automatically with a 0 delay
+ * timer, or they can be stepped manually by entering the coroutine.
+ */
+static BlockJob *test_block_job_start(unsigned int iterations,
+                                      bool use_timer,
+                                      int rc, int *result)
+{
+    BlockDriverState *bs;
+    TestBlockJob *s;
+    TestBlockJobCBData *data;
+
+    data = g_new0(TestBlockJobCBData, 1);
+    bs = bdrv_new();
+    s = block_job_create(&test_block_job_driver, bs, 0, test_block_job_cb,
+                         data, &error_abort);
+    s->iterations = iterations;
+    s->use_timer = use_timer;
+    s->rc = rc;
+    s->result = result;
+    s->common.co = qemu_coroutine_create(test_block_job_run);
+    data->job = s;
+    data->result = result;
+    qemu_coroutine_enter(s->common.co, s);
+    return &s->common;
+}
+
+static void test_single_job(int expected)
+{
+    BlockJob *job;
+    BlockJobTxn *txn;
+    int result = -EINPROGRESS;
+
+    txn = block_job_txn_new();
+    job = test_block_job_start(1, true, expected, &result);
+    block_job_txn_add_job(txn, job);
+
+    if (expected == -ECANCELED) {
+        block_job_cancel(job);
+    }
+
+    while (result == -EINPROGRESS) {
+        aio_poll(qemu_get_aio_context(), true);
+    }
+    g_assert_cmpint(result, ==, expected);
+
+    block_job_txn_unref(txn);
+}
+
+static void test_single_job_success(void)
+{
+    test_single_job(0);
+}
+
+static void test_single_job_failure(void)
+{
+    test_single_job(-EIO);
+}
+
+static void test_single_job_cancel(void)
+{
+    test_single_job(-ECANCELED);
+}
+
+static void test_pair_jobs(int expected1, int expected2)
+{
+    BlockJob *job1;
+    BlockJob *job2;
+    BlockJobTxn *txn;
+    int result1 = -EINPROGRESS;
+    int result2 = -EINPROGRESS;
+
+    txn = block_job_txn_new();
+    job1 = test_block_job_start(1, true, expected1, &result1);
+    block_job_txn_add_job(txn, job1);
+    job2 = test_block_job_start(2, true, expected2, &result2);
+    block_job_txn_add_job(txn, job2);
+
+    if (expected1 == -ECANCELED) {
+        block_job_cancel(job1);
+    }
+    if (expected2 == -ECANCELED) {
+        block_job_cancel(job2);
+    }
+
+    while (result1 == -EINPROGRESS || result2 == -EINPROGRESS) {
+        aio_poll(qemu_get_aio_context(), true);
+    }
+
+    /* Failure or cancellation of one job cancels the other job */
+    if (expected1 != 0) {
+        expected2 = -ECANCELED;
+    } else if (expected2 != 0) {
+        expected1 = -ECANCELED;
+    }
+
+    g_assert_cmpint(result1, ==, expected1);
+    g_assert_cmpint(result2, ==, expected2);
+
+    block_job_txn_unref(txn);
+}
+
+static void test_pair_jobs_success(void)
+{
+    test_pair_jobs(0, 0);
+}
+
+static void test_pair_jobs_failure(void)
+{
+    /* Test both orderings.  The two jobs run for a different number of
+     * iterations so the code path is different depending on which job fails
+     * first.
+     */
+    test_pair_jobs(-EIO, 0);
+    test_pair_jobs(0, -EIO);
+}
+
+static void test_pair_jobs_cancel(void)
+{
+    test_pair_jobs(-ECANCELED, 0);
+    test_pair_jobs(0, -ECANCELED);
+}
+
+static void test_pair_jobs_fail_cancel_race(void)
+{
+    BlockJob *job1;
+    BlockJob *job2;
+    BlockJobTxn *txn;
+    int result1 = -EINPROGRESS;
+    int result2 = -EINPROGRESS;
+
+    txn = block_job_txn_new();
+    job1 = test_block_job_start(1, true, -ECANCELED, &result1);
+    block_job_txn_add_job(txn, job1);
+    job2 = test_block_job_start(2, false, 0, &result2);
+    block_job_txn_add_job(txn, job2);
+
+    block_job_cancel(job1);
+
+    /* Now make job2 finish before the main loop kicks jobs.  This simulates
+     * the race between a pending kick and another job completing.
+     */
+    block_job_enter(job2);
+    block_job_enter(job2);
+
+    while (result1 == -EINPROGRESS || result2 == -EINPROGRESS) {
+        aio_poll(qemu_get_aio_context(), true);
+    }
+
+    g_assert_cmpint(result1, ==, -ECANCELED);
+    g_assert_cmpint(result2, ==, -ECANCELED);
+
+    block_job_txn_unref(txn);
+}
+
+int main(int argc, char **argv)
+{
+    qemu_init_main_loop(&error_abort);
+
+    g_test_init(&argc, &argv, NULL);
+    g_test_add_func("/single/success", test_single_job_success);
+    g_test_add_func("/single/failure", test_single_job_failure);
+    g_test_add_func("/single/cancel", test_single_job_cancel);
+    g_test_add_func("/pair/success", test_pair_jobs_success);
+    g_test_add_func("/pair/failure", test_pair_jobs_failure);
+    g_test_add_func("/pair/cancel", test_pair_jobs_cancel);
+    g_test_add_func("/pair/fail-cancel-race", test_pair_jobs_fail_cancel_race);
+    return g_test_run();
+}
diff --git a/tests/test-timed-average.c b/tests/test-timed-average.c
new file mode 100644
index 0000000000..a049799b80
--- /dev/null
+++ b/tests/test-timed-average.c
@@ -0,0 +1,90 @@
+/*
+ * Timed average computation tests
+ *
+ * Copyright Nodalink, EURL. 2014
+ *
+ * Authors:
+ *  Benoît Canet     <benoit.canet@nodalink.com>
+ *
+ * This work is licensed under the terms of the GNU LGPL, version 2 or later.
+ * See the COPYING.LIB file in the top-level directory.
+ */
+
+#include <glib.h>
+#include <unistd.h>
+
+#include "qemu/timed-average.h"
+
+/* This is the clock for QEMU_CLOCK_VIRTUAL */
+static int64_t my_clock_value;
+
+int64_t cpu_get_clock(void)
+{
+    return my_clock_value;
+}
+
+static void account(TimedAverage *ta)
+{
+    timed_average_account(ta, 1);
+    timed_average_account(ta, 5);
+    timed_average_account(ta, 2);
+    timed_average_account(ta, 4);
+    timed_average_account(ta, 3);
+}
+
+static void test_average(void)
+{
+    TimedAverage ta;
+    uint64_t result;
+    int i;
+
+    /* we will compute some average on a period of 1 second */
+    timed_average_init(&ta, QEMU_CLOCK_VIRTUAL, NANOSECONDS_PER_SECOND);
+
+    result = timed_average_min(&ta);
+    g_assert(result == 0);
+    result = timed_average_avg(&ta);
+    g_assert(result == 0);
+    result = timed_average_max(&ta);
+    g_assert(result == 0);
+
+    for (i = 0; i < 100; i++) {
+        account(&ta);
+        result = timed_average_min(&ta);
+        g_assert(result == 1);
+        result = timed_average_avg(&ta);
+        g_assert(result == 3);
+        result = timed_average_max(&ta);
+        g_assert(result == 5);
+        my_clock_value += NANOSECONDS_PER_SECOND / 10;
+    }
+
+    my_clock_value += NANOSECONDS_PER_SECOND * 100;
+
+    result = timed_average_min(&ta);
+    g_assert(result == 0);
+    result = timed_average_avg(&ta);
+    g_assert(result == 0);
+    result = timed_average_max(&ta);
+    g_assert(result == 0);
+
+    for (i = 0; i < 100; i++) {
+        account(&ta);
+        result = timed_average_min(&ta);
+        g_assert(result == 1);
+        result = timed_average_avg(&ta);
+        g_assert(result == 3);
+        result = timed_average_max(&ta);
+        g_assert(result == 5);
+        my_clock_value += NANOSECONDS_PER_SECOND / 10;
+    }
+}
+
+int main(int argc, char **argv)
+{
+    /* tests in the same order as the header function declarations */
+    g_test_init(&argc, &argv, NULL);
+    g_test_add_func("/timed-average/average", test_average);
+    return g_test_run();
+}
+