Series comparison

-[PULL 0/5] Block patches
+[Qemu-devel] [PULL v2 00/24] Block patches
-The following changes since commit d5c75ec500d96f1d93447f990cd5a4ef5ba27fae:
+The following changes since commit 56f9e46b841c7be478ca038d8d4085d776ab4b0d:
-  Merge remote-tracking branch 'remotes/stefanberger/tags/pull-tpm-2020-05-08-1' into staging (2020-05-12 17:00:10 +0100)
+  Merge remote-tracking branch 'remotes/armbru/tags/pull-qapi-2017-02-20' into staging (2017-02-20 17:42:47 +0000)
-are available in the Git repository at:
+are available in the git repository at:
-  https://github.com/XanClic/qemu.git tags/pull-block-2020-05-13
+  git://github.com/stefanha/qemu.git tags/block-pull-request
-for you to fetch changes up to fc9aefc8c0d3c6392656ea661ce72c1583b70bbd:
+for you to fetch changes up to a7b91d35bab97a2d3e779d0c64c9b837b52a6cf7:
-  block/block-copy: fix use-after-free of task pointer (2020-05-13 14:20:31 +0200)
+  coroutine-lock: make CoRwlock thread-safe and fair (2017-02-21 11:39:40 +0000)
 ----------------------------------------------------------------
-Block patches:
+Pull request
-- zstd compression for qcow2
-- Fix use-after-free
+v2:
  * Rebased to resolve scsi conflicts
 ----------------------------------------------------------------
-Denis Plotnikov (4):
-  qcow2: introduce compression type feature
-  qcow2: rework the cluster compression routine
-  qcow2: add zstd cluster compression
-  iotests: 287: add qcow2 compression type test
-Vladimir Sementsov-Ogievskiy (1):
+Paolo Bonzini (24):
-  block/block-copy: fix use-after-free of task pointer
+  block: move AioContext, QEMUTimer, main-loop to libqemuutil
   aio: introduce aio_co_schedule and aio_co_wake
   block-backend: allow blk_prw from coroutine context
   test-thread-pool: use generic AioContext infrastructure
   io: add methods to set I/O handlers on AioContext
   io: make qio_channel_yield aware of AioContexts
   nbd: convert to use qio_channel_yield
   coroutine-lock: reschedule coroutine on the AioContext it was running
     on
   blkdebug: reschedule coroutine on the AioContext it is running on
   qed: introduce qed_aio_start_io and qed_aio_next_io_cb
   aio: push aio_context_acquire/release down to dispatching
   block: explicitly acquire aiocontext in timers that need it
   block: explicitly acquire aiocontext in callbacks that need it
   block: explicitly acquire aiocontext in bottom halves that need it
   block: explicitly acquire aiocontext in aio callbacks that need it
   aio-posix: partially inline aio_dispatch into aio_poll
   async: remove unnecessary inc/dec pairs
   block: document fields protected by AioContext lock
   coroutine-lock: make CoMutex thread-safe
   coroutine-lock: add limited spinning to CoMutex
   test-aio-multithread: add performance comparison with thread-based
     mutexes
   coroutine-lock: place CoMutex before CoQueue in header
   coroutine-lock: add mutex argument to CoQueue APIs
   coroutine-lock: make CoRwlock thread-safe and fair
- docs/interop/qcow2.txt           |   1 +
+ Makefile.objs                       |   4 -
- configure                        |   2 +-
+ stubs/Makefile.objs                 |   1 +
- qapi/block-core.json             |  23 ++-
+ tests/Makefile.include              |  19 +-
- block/qcow2.h                    |  20 ++-
+ util/Makefile.objs                  |   6 +-
- include/block/block_int.h        |   1 +
+ block/nbd-client.h                  |   2 +-
- block/block-copy.c               |   2 +-
+ block/qed.h                         |   3 +
- block/qcow2-threads.c            | 240 +++++++++++++++++++++++++++++--
+ include/block/aio.h                 |  38 ++-
- block/qcow2.c                    | 120 ++++++++++++++++
+ include/block/block_int.h           |  64 +++--
- tests/qemu-iotests/031.out       |  14 +-
+ include/io/channel.h                |  72 +++++-
- tests/qemu-iotests/036.out       |   4 +-
+ include/qemu/coroutine.h            |  84 ++++---
- tests/qemu-iotests/049.out       | 102 ++++++-------
+ include/qemu/coroutine_int.h        |  11 +-
- tests/qemu-iotests/060.out       |   1 +
+ include/sysemu/block-backend.h      |  14 +-
- tests/qemu-iotests/061.out       |  34 +++--
+ tests/iothread.h                    |  25 ++
- tests/qemu-iotests/065           |  28 ++--
+ block/backup.c                      |   2 +-
- tests/qemu-iotests/080           |   2 +-
+ block/blkdebug.c                    |   9 +-
- tests/qemu-iotests/082.out       |  48 ++++++-
+ block/blkreplay.c                   |   2 +-
- tests/qemu-iotests/085.out       |  38 ++---
+ block/block-backend.c               |  13 +-
- tests/qemu-iotests/144.out       |   4 +-
+ block/curl.c                        |  44 +++-
- tests/qemu-iotests/182.out       |   2 +-
+ block/gluster.c                     |   9 +-
- tests/qemu-iotests/185.out       |   8 +-
+ block/io.c                          |  42 +---
- tests/qemu-iotests/198.out       |   2 +
+ block/iscsi.c                       |  15 +-
- tests/qemu-iotests/206.out       |   5 +
+ block/linux-aio.c                   |  10 +-
- tests/qemu-iotests/242.out       |   5 +
+ block/mirror.c                      |  12 +-
- tests/qemu-iotests/255.out       |   8 +-
+ block/nbd-client.c                  | 119 +++++----
- tests/qemu-iotests/274.out       |  49 ++++---
+ block/nfs.c                         |   9 +-
- tests/qemu-iotests/280.out       |   2 +-
+ block/qcow2-cluster.c               |   4 +-
- tests/qemu-iotests/287           | 152 ++++++++++++++++++++
+ block/qed-cluster.c                 |   2 +
- tests/qemu-iotests/287.out       |  67 +++++++++
+ block/qed-table.c                   |  12 +-
- tests/qemu-iotests/common.filter |   3 +-
+ block/qed.c                         |  58 +++--
- tests/qemu-iotests/group         |   1 +
+ block/sheepdog.c                    |  31 +--
-files changed, 825 insertions(+), 163 deletions(-)
+ block/ssh.c                         |  29 +--
- create mode 100755 tests/qemu-iotests/287
+ block/throttle-groups.c             |   4 +-
- create mode 100644 tests/qemu-iotests/287.out
+ block/win32-aio.c                   |   9 +-
  dma-helpers.c                       |   2 +
  hw/9pfs/9p.c                        |   2 +-
  hw/block/virtio-blk.c               |  19 +-
  hw/scsi/scsi-bus.c                  |   2 +
  hw/scsi/scsi-disk.c                 |  15 ++
  hw/scsi/scsi-generic.c              |  20 +-
  hw/scsi/virtio-scsi.c               |   7 +
  io/channel-command.c                |  13 +
  io/channel-file.c                   |  11 +
  io/channel-socket.c                 |  16 +-
  io/channel-tls.c                    |  12 +
  io/channel-watch.c                  |   6 +
  io/channel.c                        |  97 ++++++--
  nbd/client.c                        |   2 +-
  nbd/common.c                        |   9 +-
  nbd/server.c                        |  94 +++-----
  stubs/linux-aio.c                   |  32 +++
  stubs/set-fd-handler.c              |  11 -
  tests/iothread.c                    |  91 +++++++
  tests/test-aio-multithread.c        | 463 ++++++++++++++++++++++++++++++++++++
  tests/test-thread-pool.c            |  12 +-
  aio-posix.c => util/aio-posix.c     |  62 ++---
  aio-win32.c => util/aio-win32.c     |  30 +--
  util/aiocb.c                        |  55 +++++
  async.c => util/async.c             |  84 ++++++-
  iohandler.c => util/iohandler.c     |   0
  main-loop.c => util/main-loop.c     |   0
  util/qemu-coroutine-lock.c          | 254 ++++++++++++++++++--
  util/qemu-coroutine-sleep.c         |   2 +-
  util/qemu-coroutine.c               |   8 +
  qemu-timer.c => util/qemu-timer.c   |   0
  thread-pool.c => util/thread-pool.c |   8 +-
  trace-events                        |  11 -
  util/trace-events                   |  17 +-
 files changed, 1712 insertions(+), 533 deletions(-)
  create mode 100644 tests/iothread.h
  create mode 100644 stubs/linux-aio.c
  create mode 100644 tests/iothread.c
  create mode 100644 tests/test-aio-multithread.c
  rename aio-posix.c => util/aio-posix.c (94%)
  rename aio-win32.c => util/aio-win32.c (95%)
  create mode 100644 util/aiocb.c
  rename async.c => util/async.c (82%)
  rename iohandler.c => util/iohandler.c (100%)
  rename main-loop.c => util/main-loop.c (100%)
  rename qemu-timer.c => util/qemu-timer.c (100%)
  rename thread-pool.c => util/thread-pool.c (97%)
 --
-.26.2
+.9.3

-New patch
+[Qemu-devel] [PULL v2 01/24] block: move AioContext, QEMUTimer, main-loop to libqemuutil
+From: Paolo Bonzini <pbonzini@redhat.com>
 AioContext is fairly self contained, the only dependency is QEMUTimer but
 that in turn doesn't need anything else.  So move them out of block-obj-y
 to avoid introducing a dependency from io/ to block-obj-y.
 main-loop and its dependency iohandler also need to be moved, because
 later in this series io/ will call iohandler_get_aio_context.
 [Changed copyright "the QEMU team" to "other QEMU contributors" as
 suggested by Daniel Berrange and agreed by Paolo.
 --Stefan]
 Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
 Reviewed-by: Fam Zheng <famz@redhat.com>
 Message-id: 20170213135235.12274-2-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
  Makefile.objs                       |  4 ---
  stubs/Makefile.objs                 |  1 +
  tests/Makefile.include              | 11 ++++----
  util/Makefile.objs                  |  6 +++-
  block/io.c                          | 29 -------------------
  stubs/linux-aio.c                   | 32 +++++++++++++++++++++
  stubs/set-fd-handler.c              | 11 --------
  aio-posix.c => util/aio-posix.c     |  2 +-
  aio-win32.c => util/aio-win32.c     |  0
  util/aiocb.c                        | 55 +++++++++++++++++++++++++++++++++++++
  async.c => util/async.c             |  3 +-
  iohandler.c => util/iohandler.c     |  0
  main-loop.c => util/main-loop.c     |  0
  qemu-timer.c => util/qemu-timer.c   |  0
  thread-pool.c => util/thread-pool.c |  2 +-
  trace-events                        | 11 --------
  util/trace-events                   | 11 ++++++++
 files changed, 114 insertions(+), 64 deletions(-)
  create mode 100644 stubs/linux-aio.c
  rename aio-posix.c => util/aio-posix.c (99%)
  rename aio-win32.c => util/aio-win32.c (100%)
  create mode 100644 util/aiocb.c
  rename async.c => util/async.c (99%)
  rename iohandler.c => util/iohandler.c (100%)
  rename main-loop.c => util/main-loop.c (100%)
  rename qemu-timer.c => util/qemu-timer.c (100%)
  rename thread-pool.c => util/thread-pool.c (99%)
 diff --git a/Makefile.objs b/Makefile.objs
 index XXXXXXX..XXXXXXX 100644
 --- a/Makefile.objs
 +++ b/Makefile.objs
@@ -XXX,XX +XXX,XX @@ chardev-obj-y = chardev/
  #######################################################################
  # block-obj-y is code used by both qemu system emulation and qemu-img
 -block-obj-y = async.o thread-pool.o
  block-obj-y += nbd/
  block-obj-y += block.o blockjob.o
 -block-obj-y += main-loop.o iohandler.o qemu-timer.o
 -block-obj-$(CONFIG_POSIX) += aio-posix.o
 -block-obj-$(CONFIG_WIN32) += aio-win32.o
  block-obj-y += block/
  block-obj-y += qemu-io-cmds.o
  block-obj-$(CONFIG_REPLICATION) += replication.o
 diff --git a/stubs/Makefile.objs b/stubs/Makefile.objs
 index XXXXXXX..XXXXXXX 100644
 --- a/stubs/Makefile.objs
 +++ b/stubs/Makefile.objs
@@ -XXX,XX +XXX,XX @@ stub-obj-y += get-vm-name.o
  stub-obj-y += iothread.o
  stub-obj-y += iothread-lock.o
  stub-obj-y += is-daemonized.o
 +stub-obj-$(CONFIG_LINUX_AIO) += linux-aio.o
  stub-obj-y += machine-init-done.o
  stub-obj-y += migr-blocker.o
  stub-obj-y += monitor.o
 diff --git a/tests/Makefile.include b/tests/Makefile.include
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/Makefile.include
 +++ b/tests/Makefile.include
@@ -XXX,XX +XXX,XX @@ check-unit-y += tests/test-visitor-serialization$(EXESUF)
  check-unit-y += tests/test-iov$(EXESUF)
  gcov-files-test-iov-y = util/iov.c
  check-unit-y += tests/test-aio$(EXESUF)
 +gcov-files-test-aio-y = util/async.c util/qemu-timer.o
 +gcov-files-test-aio-$(CONFIG_WIN32) += util/aio-win32.c
 +gcov-files-test-aio-$(CONFIG_POSIX) += util/aio-posix.c
  check-unit-y += tests/test-throttle$(EXESUF)
  gcov-files-test-aio-$(CONFIG_WIN32) = aio-win32.c
  gcov-files-test-aio-$(CONFIG_POSIX) = aio-posix.c
@@ -XXX,XX +XXX,XX @@ tests/check-qjson$(EXESUF): tests/check-qjson.o $(test-util-obj-y)
  tests/check-qom-interface$(EXESUF): tests/check-qom-interface.o $(test-qom-obj-y)
  tests/check-qom-proplist$(EXESUF): tests/check-qom-proplist.o $(test-qom-obj-y)
 -tests/test-char$(EXESUF): tests/test-char.o qemu-timer.o \
 -    $(test-util-obj-y) $(qtest-obj-y) $(test-block-obj-y) $(chardev-obj-y)
 +tests/test-char$(EXESUF): tests/test-char.o $(test-util-obj-y) $(qtest-obj-y) $(test-io-obj-y) $(chardev-obj-y)
  tests/test-coroutine$(EXESUF): tests/test-coroutine.o $(test-block-obj-y)
  tests/test-aio$(EXESUF): tests/test-aio.o $(test-block-obj-y)
  tests/test-throttle$(EXESUF): tests/test-throttle.o $(test-block-obj-y)
@@ -XXX,XX +XXX,XX @@ tests/test-vmstate$(EXESUF): tests/test-vmstate.o \
      migration/vmstate.o migration/qemu-file.o \
          migration/qemu-file-channel.o migration/qjson.o \
      $(test-io-obj-y)
 -tests/test-timed-average$(EXESUF): tests/test-timed-average.o qemu-timer.o \
 -    $(test-util-obj-y)
 +tests/test-timed-average$(EXESUF): tests/test-timed-average.o $(test-util-obj-y)
  tests/test-base64$(EXESUF): tests/test-base64.o \
      libqemuutil.a libqemustub.a
  tests/ptimer-test$(EXESUF): tests/ptimer-test.o tests/ptimer-test-stubs.o hw/core/ptimer.o libqemustub.a
@@ -XXX,XX +XXX,XX @@ tests/usb-hcd-ehci-test$(EXESUF): tests/usb-hcd-ehci-test.o $(libqos-usb-obj-y)
  tests/usb-hcd-xhci-test$(EXESUF): tests/usb-hcd-xhci-test.o $(libqos-usb-obj-y)
  tests/pc-cpu-test$(EXESUF): tests/pc-cpu-test.o
  tests/postcopy-test$(EXESUF): tests/postcopy-test.o
 -tests/vhost-user-test$(EXESUF): tests/vhost-user-test.o qemu-timer.o \
 +tests/vhost-user-test$(EXESUF): tests/vhost-user-test.o $(test-util-obj-y) \
      $(qtest-obj-y) $(test-io-obj-y) $(libqos-virtio-obj-y) $(libqos-pc-obj-y) \
      $(chardev-obj-y)
  tests/qemu-iotests/socket_scm_helper$(EXESUF): tests/qemu-iotests/socket_scm_helper.o
 diff --git a/util/Makefile.objs b/util/Makefile.objs
 index XXXXXXX..XXXXXXX 100644
 --- a/util/Makefile.objs
 +++ b/util/Makefile.objs
@@ -XXX,XX +XXX,XX @@
  util-obj-y = osdep.o cutils.o unicode.o qemu-timer-common.o
  util-obj-y += bufferiszero.o
  util-obj-y += lockcnt.o
 +util-obj-y += aiocb.o async.o thread-pool.o qemu-timer.o
 +util-obj-y += main-loop.o iohandler.o
 +util-obj-$(CONFIG_POSIX) += aio-posix.o
  util-obj-$(CONFIG_POSIX) += compatfd.o
  util-obj-$(CONFIG_POSIX) += event_notifier-posix.o
  util-obj-$(CONFIG_POSIX) += mmap-alloc.o
  util-obj-$(CONFIG_POSIX) += oslib-posix.o
  util-obj-$(CONFIG_POSIX) += qemu-openpty.o
  util-obj-$(CONFIG_POSIX) += qemu-thread-posix.o
 -util-obj-$(CONFIG_WIN32) += event_notifier-win32.o
  util-obj-$(CONFIG_POSIX) += memfd.o
 +util-obj-$(CONFIG_WIN32) += aio-win32.o
 +util-obj-$(CONFIG_WIN32) += event_notifier-win32.o
  util-obj-$(CONFIG_WIN32) += oslib-win32.o
  util-obj-$(CONFIG_WIN32) += qemu-thread-win32.o
  util-obj-y += envlist.o path.o module.o
 diff --git a/block/io.c b/block/io.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/io.c
 +++ b/block/io.c
@@ -XXX,XX +XXX,XX @@ BlockAIOCB *bdrv_aio_flush(BlockDriverState *bs,
      return &acb->common;
  }
 -void *qemu_aio_get(const AIOCBInfo *aiocb_info, BlockDriverState *bs,
 -                   BlockCompletionFunc *cb, void *opaque)
 -{
 -    BlockAIOCB *acb;
 -
 -    acb = g_malloc(aiocb_info->aiocb_size);
 -    acb->aiocb_info = aiocb_info;
 -    acb->bs = bs;
 -    acb->cb = cb;
 -    acb->opaque = opaque;
 -    acb->refcnt = 1;
 -    return acb;
 -}
 -
 -void qemu_aio_ref(void *p)
 -{
 -    BlockAIOCB *acb = p;
 -    acb->refcnt++;
 -}
 -
 -void qemu_aio_unref(void *p)
 -{
 -    BlockAIOCB *acb = p;
 -    assert(acb->refcnt > 0);
 -    if (--acb->refcnt == 0) {
 -        g_free(acb);
 -    }
 -}
 -
  /**************************************************************/
  /* Coroutine block device emulation */
 diff --git a/stubs/linux-aio.c b/stubs/linux-aio.c
 new file mode 100644
 index XXXXXXX..XXXXXXX
 --- /dev/null
 +++ b/stubs/linux-aio.c
@@ -XXX,XX +XXX,XX @@
 +/*
 + * Linux native AIO support.
 + *
 + * Copyright (C) 2009 IBM, Corp.
 + * Copyright (C) 2009 Red Hat, Inc.
 + *
 + * This work is licensed under the terms of the GNU GPL, version 2 or later.
 + * See the COPYING file in the top-level directory.
 + */
 +#include "qemu/osdep.h"
 +#include "block/aio.h"
 +#include "block/raw-aio.h"
 +
 +void laio_detach_aio_context(LinuxAioState *s, AioContext *old_context)
 +{
 +    abort();
 +}
 +
 +void laio_attach_aio_context(LinuxAioState *s, AioContext *new_context)
 +{
 +    abort();
 +}
 +
 +LinuxAioState *laio_init(void)
 +{
 +    abort();
 +}
 +
 +void laio_cleanup(LinuxAioState *s)
 +{
 +    abort();
 +}
 diff --git a/stubs/set-fd-handler.c b/stubs/set-fd-handler.c
 index XXXXXXX..XXXXXXX 100644
 --- a/stubs/set-fd-handler.c
 +++ b/stubs/set-fd-handler.c
@@ -XXX,XX +XXX,XX @@ void qemu_set_fd_handler(int fd,
  {
      abort();
  }
 -
 -void aio_set_fd_handler(AioContext *ctx,
 -                        int fd,
 -                        bool is_external,
 -                        IOHandler *io_read,
 -                        IOHandler *io_write,
 -                        AioPollFn *io_poll,
 -                        void *opaque)
 -{
 -    abort();
 -}
 diff --git a/aio-posix.c b/util/aio-posix.c
 similarity index 99%
 rename from aio-posix.c
 rename to util/aio-posix.c
 index XXXXXXX..XXXXXXX 100644
 --- a/aio-posix.c
 +++ b/util/aio-posix.c
@@ -XXX,XX +XXX,XX @@
  #include "qemu/rcu_queue.h"
  #include "qemu/sockets.h"
  #include "qemu/cutils.h"
 -#include "trace-root.h"
 +#include "trace.h"
  #ifdef CONFIG_EPOLL_CREATE1
  #include <sys/epoll.h>
  #endif
 diff --git a/aio-win32.c b/util/aio-win32.c
 similarity index 100%
 rename from aio-win32.c
 rename to util/aio-win32.c
 diff --git a/util/aiocb.c b/util/aiocb.c
 new file mode 100644
 index XXXXXXX..XXXXXXX
 --- /dev/null
 +++ b/util/aiocb.c
@@ -XXX,XX +XXX,XX @@
 +/*
 + * BlockAIOCB allocation
 + *
 + * Copyright (c) 2003-2017 Fabrice Bellard and other QEMU contributors
 + *
 + * Permission is hereby granted, free of charge, to any person obtaining a copy
 + * of this software and associated documentation files (the "Software"), to deal
 + * in the Software without restriction, including without limitation the rights
 + * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
 + * copies of the Software, and to permit persons to whom the Software is
 + * furnished to do so, subject to the following conditions:
 + *
 + * The above copyright notice and this permission notice shall be included in
 + * all copies or substantial portions of the Software.
 + *
 + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
 + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
 + * THE SOFTWARE.
 + */
 +
 +#include "qemu/osdep.h"
 +#include "block/aio.h"
 +
 +void *qemu_aio_get(const AIOCBInfo *aiocb_info, BlockDriverState *bs,
 +                   BlockCompletionFunc *cb, void *opaque)
 +{
 +    BlockAIOCB *acb;
 +
 +    acb = g_malloc(aiocb_info->aiocb_size);
 +    acb->aiocb_info = aiocb_info;
 +    acb->bs = bs;
 +    acb->cb = cb;
 +    acb->opaque = opaque;
 +    acb->refcnt = 1;
 +    return acb;
 +}
 +
 +void qemu_aio_ref(void *p)
 +{
 +    BlockAIOCB *acb = p;
 +    acb->refcnt++;
 +}
 +
 +void qemu_aio_unref(void *p)
 +{
 +    BlockAIOCB *acb = p;
 +    assert(acb->refcnt > 0);
 +    if (--acb->refcnt == 0) {
 +        g_free(acb);
 +    }
 +}
 diff --git a/async.c b/util/async.c
 similarity index 99%
 rename from async.c
 rename to util/async.c
 index XXXXXXX..XXXXXXX 100644
 --- a/async.c
 +++ b/util/async.c
@@ -XXX,XX +XXX,XX @@
  /*
 - * QEMU System Emulator
 + * Data plane event loop
   *
   * Copyright (c) 2003-2008 Fabrice Bellard
 + * Copyright (c) 2009-2017 QEMU contributors
   *
   * Permission is hereby granted, free of charge, to any person obtaining a copy
   * of this software and associated documentation files (the "Software"), to deal
 diff --git a/iohandler.c b/util/iohandler.c
 similarity index 100%
 rename from iohandler.c
 rename to util/iohandler.c
 diff --git a/main-loop.c b/util/main-loop.c
 similarity index 100%
 rename from main-loop.c
 rename to util/main-loop.c
 diff --git a/qemu-timer.c b/util/qemu-timer.c
 similarity index 100%
 rename from qemu-timer.c
 rename to util/qemu-timer.c
 diff --git a/thread-pool.c b/util/thread-pool.c
 similarity index 99%
 rename from thread-pool.c
 rename to util/thread-pool.c
 index XXXXXXX..XXXXXXX 100644
 --- a/thread-pool.c
 +++ b/util/thread-pool.c
@@ -XXX,XX +XXX,XX @@
  #include "qemu/queue.h"
  #include "qemu/thread.h"
  #include "qemu/coroutine.h"
 -#include "trace-root.h"
 +#include "trace.h"
  #include "block/thread-pool.h"
  #include "qemu/main-loop.h"
 diff --git a/trace-events b/trace-events
 index XXXXXXX..XXXXXXX 100644
 --- a/trace-events
 +++ b/trace-events
@@ -XXX,XX +XXX,XX @@
  #
  # The <format-string> should be a sprintf()-compatible format string.
 -# aio-posix.c
 -run_poll_handlers_begin(void *ctx, int64_t max_ns) "ctx %p max_ns %"PRId64
 -run_poll_handlers_end(void *ctx, bool progress) "ctx %p progress %d"
 -poll_shrink(void *ctx, int64_t old, int64_t new) "ctx %p old %"PRId64" new %"PRId64
 -poll_grow(void *ctx, int64_t old, int64_t new) "ctx %p old %"PRId64" new %"PRId64
 -
 -# thread-pool.c
 -thread_pool_submit(void *pool, void *req, void *opaque) "pool %p req %p opaque %p"
 -thread_pool_complete(void *pool, void *req, void *opaque, int ret) "pool %p req %p opaque %p ret %d"
 -thread_pool_cancel(void *req, void *opaque) "req %p opaque %p"
 -
  # ioport.c
  cpu_in(unsigned int addr, char size, unsigned int val) "addr %#x(%c) value %u"
  cpu_out(unsigned int addr, char size, unsigned int val) "addr %#x(%c) value %u"
 diff --git a/util/trace-events b/util/trace-events
 index XXXXXXX..XXXXXXX 100644
 --- a/util/trace-events
 +++ b/util/trace-events
@@ -XXX,XX +XXX,XX @@
  # See docs/tracing.txt for syntax documentation.
 +# util/aio-posix.c
 +run_poll_handlers_begin(void *ctx, int64_t max_ns) "ctx %p max_ns %"PRId64
 +run_poll_handlers_end(void *ctx, bool progress) "ctx %p progress %d"
 +poll_shrink(void *ctx, int64_t old, int64_t new) "ctx %p old %"PRId64" new %"PRId64
 +poll_grow(void *ctx, int64_t old, int64_t new) "ctx %p old %"PRId64" new %"PRId64
 +
 +# util/thread-pool.c
 +thread_pool_submit(void *pool, void *req, void *opaque) "pool %p req %p opaque %p"
 +thread_pool_complete(void *pool, void *req, void *opaque, int ret) "pool %p req %p opaque %p ret %d"
 +thread_pool_cancel(void *req, void *opaque) "req %p opaque %p"
 +
  # util/buffer.c
  buffer_resize(const char *buf, size_t olen, size_t len) "%s: old %zd, new %zd"
  buffer_move_empty(const char *buf, size_t len, const char *from) "%s: %zd bytes from %s"
 --
 .9.3

-New patch
+[Qemu-devel] [PULL v2 02/24] aio: introduce aio_co_schedule and aio_co_wake
+From: Paolo Bonzini <pbonzini@redhat.com>
 aio_co_wake provides the infrastructure to start a coroutine on a "home"
 AioContext.  It will be used by CoMutex and CoQueue, so that coroutines
 don't jump from one context to another when they go to sleep on a
 mutex or waitqueue.  However, it can also be used as a more efficient
 alternative to one-shot bottom halves, and saves the effort of tracking
 which AioContext a coroutine is running on.
 aio_co_schedule is the part of aio_co_wake that starts a coroutine
 on a remove AioContext, but it is also useful to implement e.g.
 bdrv_set_aio_context callbacks.
 The implementation of aio_co_schedule is based on a lock-free
 multiple-producer, single-consumer queue.  The multiple producers use
 cmpxchg to add to a LIFO stack.  The consumer (a per-AioContext bottom
 half) grabs all items added so far, inverts the list to make it FIFO,
 and goes through it one item at a time until it's empty.  The data
 structure was inspired by OSv, which uses it in the very code we'll
 "port" to QEMU for the thread-safe CoMutex.
 Most of the new code is really tests.
 Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
 Reviewed-by: Fam Zheng <famz@redhat.com>
 Message-id: 20170213135235.12274-3-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
  tests/Makefile.include       |   8 +-
  include/block/aio.h          |  32 +++++++
  include/qemu/coroutine_int.h |  11 ++-
  tests/iothread.h             |  25 +++++
  tests/iothread.c             |  91 ++++++++++++++++++
  tests/test-aio-multithread.c | 213 +++++++++++++++++++++++++++++++++++++++++++
  util/async.c                 |  65 +++++++++++++
  util/qemu-coroutine.c        |   8 ++
  util/trace-events            |   4 +
 files changed, 453 insertions(+), 4 deletions(-)
  create mode 100644 tests/iothread.h
  create mode 100644 tests/iothread.c
  create mode 100644 tests/test-aio-multithread.c
 diff --git a/tests/Makefile.include b/tests/Makefile.include
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/Makefile.include
 +++ b/tests/Makefile.include
@@ -XXX,XX +XXX,XX @@ check-unit-y += tests/test-aio$(EXESUF)
  gcov-files-test-aio-y = util/async.c util/qemu-timer.o
  gcov-files-test-aio-$(CONFIG_WIN32) += util/aio-win32.c
  gcov-files-test-aio-$(CONFIG_POSIX) += util/aio-posix.c
 +check-unit-y += tests/test-aio-multithread$(EXESUF)
 +gcov-files-test-aio-multithread-y = $(gcov-files-test-aio-y)
 +gcov-files-test-aio-multithread-y += util/qemu-coroutine.c tests/iothread.c
  check-unit-y += tests/test-throttle$(EXESUF)
 -gcov-files-test-aio-$(CONFIG_WIN32) = aio-win32.c
 -gcov-files-test-aio-$(CONFIG_POSIX) = aio-posix.c
  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
@@ -XXX,XX +XXX,XX @@ test-qapi-obj-y = tests/test-qapi-visit.o tests/test-qapi-types.o \
      $(test-qom-obj-y)
  test-crypto-obj-y = $(crypto-obj-y) $(test-qom-obj-y)
  test-io-obj-y = $(io-obj-y) $(test-crypto-obj-y)
 -test-block-obj-y = $(block-obj-y) $(test-io-obj-y)
 +test-block-obj-y = $(block-obj-y) $(test-io-obj-y) tests/iothread.o
  tests/check-qint$(EXESUF): tests/check-qint.o $(test-util-obj-y)
  tests/check-qstring$(EXESUF): tests/check-qstring.o $(test-util-obj-y)
@@ -XXX,XX +XXX,XX @@ tests/check-qom-proplist$(EXESUF): tests/check-qom-proplist.o $(test-qom-obj-y)
  tests/test-char$(EXESUF): tests/test-char.o $(test-util-obj-y) $(qtest-obj-y) $(test-io-obj-y) $(chardev-obj-y)
  tests/test-coroutine$(EXESUF): tests/test-coroutine.o $(test-block-obj-y)
  tests/test-aio$(EXESUF): tests/test-aio.o $(test-block-obj-y)
 +tests/test-aio-multithread$(EXESUF): tests/test-aio-multithread.o $(test-block-obj-y)
  tests/test-throttle$(EXESUF): tests/test-throttle.o $(test-block-obj-y)
  tests/test-blockjob$(EXESUF): tests/test-blockjob.o $(test-block-obj-y) $(test-util-obj-y)
  tests/test-blockjob-txn$(EXESUF): tests/test-blockjob-txn.o $(test-block-obj-y) $(test-util-obj-y)
 diff --git a/include/block/aio.h b/include/block/aio.h
 index XXXXXXX..XXXXXXX 100644
 --- a/include/block/aio.h
 +++ b/include/block/aio.h
@@ -XXX,XX +XXX,XX @@ typedef void QEMUBHFunc(void *opaque);
  typedef bool AioPollFn(void *opaque);
  typedef void IOHandler(void *opaque);
 +struct Coroutine;
  struct ThreadPool;
  struct LinuxAioState;
@@ -XXX,XX +XXX,XX @@ struct AioContext {
      bool notified;
      EventNotifier notifier;
 +    QSLIST_HEAD(, Coroutine) scheduled_coroutines;
 +    QEMUBH *co_schedule_bh;
 +
      /* Thread pool for performing work and receiving completion callbacks.
       * Has its own locking.
       */
@@ -XXX,XX +XXX,XX @@ static inline bool aio_node_check(AioContext *ctx, bool is_external)
  }
  /**
 + * aio_co_schedule:
 + * @ctx: the aio context
 + * @co: the coroutine
 + *
 + * Start a coroutine on a remote AioContext.
 + *
 + * The coroutine must not be entered by anyone else while aio_co_schedule()
 + * is active.  In addition the coroutine must have yielded unless ctx
 + * is the context in which the coroutine is running (i.e. the value of
 + * qemu_get_current_aio_context() from the coroutine itself).
 + */
 +void aio_co_schedule(AioContext *ctx, struct Coroutine *co);
 +
 +/**
 + * aio_co_wake:
 + * @co: the coroutine
 + *
 + * Restart a coroutine on the AioContext where it was running last, thus
 + * preventing coroutines from jumping from one context to another when they
 + * go to sleep.
 + *
 + * aio_co_wake may be executed either in coroutine or non-coroutine
 + * context.  The coroutine must not be entered by anyone else while
 + * aio_co_wake() is active.
 + */
 +void aio_co_wake(struct Coroutine *co);
 +
 +/**
   * Return the AioContext whose event loop runs in the current thread.
   *
   * If called from an IOThread this will be the IOThread's AioContext.  If
 diff --git a/include/qemu/coroutine_int.h b/include/qemu/coroutine_int.h
 index XXXXXXX..XXXXXXX 100644
 --- a/include/qemu/coroutine_int.h
 +++ b/include/qemu/coroutine_int.h
@@ -XXX,XX +XXX,XX @@ struct Coroutine {
      CoroutineEntry *entry;
      void *entry_arg;
      Coroutine *caller;
 +
 +    /* Only used when the coroutine has terminated.  */
      QSLIST_ENTRY(Coroutine) pool_next;
 +
      size_t locks_held;
 -    /* Coroutines that should be woken up when we yield or terminate */
 +    /* Coroutines that should be woken up when we yield or terminate.
 +     * Only used when the coroutine is running.
 +     */
      QSIMPLEQ_HEAD(, Coroutine) co_queue_wakeup;
 +
 +    /* Only used when the coroutine has yielded.  */
 +    AioContext *ctx;
      QSIMPLEQ_ENTRY(Coroutine) co_queue_next;
 +    QSLIST_ENTRY(Coroutine) co_scheduled_next;
  };
  Coroutine *qemu_coroutine_new(void);
 diff --git a/tests/iothread.h b/tests/iothread.h
 new file mode 100644
 index XXXXXXX..XXXXXXX
 --- /dev/null
 +++ b/tests/iothread.h
@@ -XXX,XX +XXX,XX @@
 +/*
 + * Event loop thread implementation for unit tests
 + *
 + * Copyright Red Hat Inc., 2013, 2016
 + *
 + * Authors:
 + *  Stefan Hajnoczi   <stefanha@redhat.com>
 + *  Paolo Bonzini     <pbonzini@redhat.com>
 + *
 + * This work is licensed under the terms of the GNU GPL, version 2 or later.
 + * See the COPYING file in the top-level directory.
 + */
 +#ifndef TEST_IOTHREAD_H
 +#define TEST_IOTHREAD_H
 +
 +#include "block/aio.h"
 +#include "qemu/thread.h"
 +
 +typedef struct IOThread IOThread;
 +
 +IOThread *iothread_new(void);
 +void iothread_join(IOThread *iothread);
 +AioContext *iothread_get_aio_context(IOThread *iothread);
 +
 +#endif
 diff --git a/tests/iothread.c b/tests/iothread.c
 new file mode 100644
 index XXXXXXX..XXXXXXX
 --- /dev/null
 +++ b/tests/iothread.c
@@ -XXX,XX +XXX,XX @@
 +/*
 + * Event loop thread implementation for unit tests
 + *
 + * Copyright Red Hat Inc., 2013, 2016
 + *
 + * Authors:
 + *  Stefan Hajnoczi   <stefanha@redhat.com>
 + *  Paolo Bonzini     <pbonzini@redhat.com>
 + *
 + * This work is licensed under the terms of the GNU GPL, version 2 or later.
 + * See the COPYING file in the top-level directory.
 + *
 + */
 +
 +#include "qemu/osdep.h"
 +#include "qapi/error.h"
 +#include "block/aio.h"
 +#include "qemu/main-loop.h"
 +#include "qemu/rcu.h"
 +#include "iothread.h"
 +
 +struct IOThread {
 +    AioContext *ctx;
 +
 +    QemuThread thread;
 +    QemuMutex init_done_lock;
 +    QemuCond init_done_cond;    /* is thread initialization done? */
 +    bool stopping;
 +};
 +
 +static __thread IOThread *my_iothread;
 +
 +AioContext *qemu_get_current_aio_context(void)
 +{
 +    return my_iothread ? my_iothread->ctx : qemu_get_aio_context();
 +}
 +
 +static void *iothread_run(void *opaque)
 +{
 +    IOThread *iothread = opaque;
 +
 +    rcu_register_thread();
 +
 +    my_iothread = iothread;
 +    qemu_mutex_lock(&iothread->init_done_lock);
 +    iothread->ctx = aio_context_new(&error_abort);
 +    qemu_cond_signal(&iothread->init_done_cond);
 +    qemu_mutex_unlock(&iothread->init_done_lock);
 +
 +    while (!atomic_read(&iothread->stopping)) {
 +        aio_poll(iothread->ctx, true);
 +    }
 +
 +    rcu_unregister_thread();
 +    return NULL;
 +}
 +
 +void iothread_join(IOThread *iothread)
 +{
 +    iothread->stopping = true;
 +    aio_notify(iothread->ctx);
 +    qemu_thread_join(&iothread->thread);
 +    qemu_cond_destroy(&iothread->init_done_cond);
 +    qemu_mutex_destroy(&iothread->init_done_lock);
 +    aio_context_unref(iothread->ctx);
 +    g_free(iothread);
 +}
 +
 +IOThread *iothread_new(void)
 +{
 +    IOThread *iothread = g_new0(IOThread, 1);
 +
 +    qemu_mutex_init(&iothread->init_done_lock);
 +    qemu_cond_init(&iothread->init_done_cond);
 +    qemu_thread_create(&iothread->thread, NULL, iothread_run,
 +                       iothread, QEMU_THREAD_JOINABLE);
 +
 +    /* Wait for initialization to complete */
 +    qemu_mutex_lock(&iothread->init_done_lock);
 +    while (iothread->ctx == NULL) {
 +        qemu_cond_wait(&iothread->init_done_cond,
 +                       &iothread->init_done_lock);
 +    }
 +    qemu_mutex_unlock(&iothread->init_done_lock);
 +    return iothread;
 +}
 +
 +AioContext *iothread_get_aio_context(IOThread *iothread)
 +{
 +    return iothread->ctx;
 +}
 diff --git a/tests/test-aio-multithread.c b/tests/test-aio-multithread.c
 new file mode 100644
 index XXXXXXX..XXXXXXX
 --- /dev/null
 +++ b/tests/test-aio-multithread.c
@@ -XXX,XX +XXX,XX @@
 +/*
 + * AioContext multithreading tests
 + *
 + * Copyright Red Hat, Inc. 2016
 + *
 + * Authors:
 + *  Paolo Bonzini    <pbonzini@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 "qemu/osdep.h"
 +#include <glib.h>
 +#include "block/aio.h"
 +#include "qapi/error.h"
 +#include "qemu/coroutine.h"
 +#include "qemu/thread.h"
 +#include "qemu/error-report.h"
 +#include "iothread.h"
 +
 +/* AioContext management */
 +
 +#define NUM_CONTEXTS 5
 +
 +static IOThread *threads[NUM_CONTEXTS];
 +static AioContext *ctx[NUM_CONTEXTS];
 +static __thread int id = -1;
 +
 +static QemuEvent done_event;
 +
 +/* Run a function synchronously on a remote iothread. */
 +
 +typedef struct CtxRunData {
 +    QEMUBHFunc *cb;
 +    void *arg;
 +} CtxRunData;
 +
 +static void ctx_run_bh_cb(void *opaque)
 +{
 +    CtxRunData *data = opaque;
 +
 +    data->cb(data->arg);
 +    qemu_event_set(&done_event);
 +}
 +
 +static void ctx_run(int i, QEMUBHFunc *cb, void *opaque)
 +{
 +    CtxRunData data = {
 +        .cb = cb,
 +        .arg = opaque
 +    };
 +
 +    qemu_event_reset(&done_event);
 +    aio_bh_schedule_oneshot(ctx[i], ctx_run_bh_cb, &data);
 +    qemu_event_wait(&done_event);
 +}
 +
 +/* Starting the iothreads. */
 +
 +static void set_id_cb(void *opaque)
 +{
 +    int *i = opaque;
 +
 +    id = *i;
 +}
 +
 +static void create_aio_contexts(void)
 +{
 +    int i;
 +
 +    for (i = 0; i < NUM_CONTEXTS; i++) {
 +        threads[i] = iothread_new();
 +        ctx[i] = iothread_get_aio_context(threads[i]);
 +    }
 +
 +    qemu_event_init(&done_event, false);
 +    for (i = 0; i < NUM_CONTEXTS; i++) {
 +        ctx_run(i, set_id_cb, &i);
 +    }
 +}
 +
 +/* Stopping the iothreads. */
 +
 +static void join_aio_contexts(void)
 +{
 +    int i;
 +
 +    for (i = 0; i < NUM_CONTEXTS; i++) {
 +        aio_context_ref(ctx[i]);
 +    }
 +    for (i = 0; i < NUM_CONTEXTS; i++) {
 +        iothread_join(threads[i]);
 +    }
 +    for (i = 0; i < NUM_CONTEXTS; i++) {
 +        aio_context_unref(ctx[i]);
 +    }
 +    qemu_event_destroy(&done_event);
 +}
 +
 +/* Basic test for the stuff above. */
 +
 +static void test_lifecycle(void)
 +{
 +    create_aio_contexts();
 +    join_aio_contexts();
 +}
 +
 +/* aio_co_schedule test.  */
 +
 +static Coroutine *to_schedule[NUM_CONTEXTS];
 +
 +static bool now_stopping;
 +
 +static int count_retry;
 +static int count_here;
 +static int count_other;
 +
 +static bool schedule_next(int n)
 +{
 +    Coroutine *co;
 +
 +    co = atomic_xchg(&to_schedule[n], NULL);
 +    if (!co) {
 +        atomic_inc(&count_retry);
 +        return false;
 +    }
 +
 +    if (n == id) {
 +        atomic_inc(&count_here);
 +    } else {
 +        atomic_inc(&count_other);
 +    }
 +
 +    aio_co_schedule(ctx[n], co);
 +    return true;
 +}
 +
 +static void finish_cb(void *opaque)
 +{
 +    schedule_next(id);
 +}
 +
 +static coroutine_fn void test_multi_co_schedule_entry(void *opaque)
 +{
 +    g_assert(to_schedule[id] == NULL);
 +    atomic_mb_set(&to_schedule[id], qemu_coroutine_self());
 +
 +    while (!atomic_mb_read(&now_stopping)) {
 +        int n;
 +
 +        n = g_test_rand_int_range(0, NUM_CONTEXTS);
 +        schedule_next(n);
 +        qemu_coroutine_yield();
 +
 +        g_assert(to_schedule[id] == NULL);
 +        atomic_mb_set(&to_schedule[id], qemu_coroutine_self());
 +    }
 +}
 +
 +
 +static void test_multi_co_schedule(int seconds)
 +{
 +    int i;
 +
 +    count_here = count_other = count_retry = 0;
 +    now_stopping = false;
 +
 +    create_aio_contexts();
 +    for (i = 0; i < NUM_CONTEXTS; i++) {
 +        Coroutine *co1 = qemu_coroutine_create(test_multi_co_schedule_entry, NULL);
 +        aio_co_schedule(ctx[i], co1);
 +    }
 +
 +    g_usleep(seconds * 1000000);
 +
 +    atomic_mb_set(&now_stopping, true);
 +    for (i = 0; i < NUM_CONTEXTS; i++) {
 +        ctx_run(i, finish_cb, NULL);
 +        to_schedule[i] = NULL;
 +    }
 +
 +    join_aio_contexts();
 +    g_test_message("scheduled %d, queued %d, retry %d, total %d\n",
 +                  count_other, count_here, count_retry,
 +                  count_here + count_other + count_retry);
 +}
 +
 +static void test_multi_co_schedule_1(void)
 +{
 +    test_multi_co_schedule(1);
 +}
 +
 +static void test_multi_co_schedule_10(void)
 +{
 +    test_multi_co_schedule(10);
 +}
 +
 +/* End of tests.  */
 +
 +int main(int argc, char **argv)
 +{
 +    init_clocks();
 +
 +    g_test_init(&argc, &argv, NULL);
 +    g_test_add_func("/aio/multi/lifecycle", test_lifecycle);
 +    if (g_test_quick()) {
 +        g_test_add_func("/aio/multi/schedule", test_multi_co_schedule_1);
 +    } else {
 +        g_test_add_func("/aio/multi/schedule", test_multi_co_schedule_10);
 +    }
 +    return g_test_run();
 +}
 diff --git a/util/async.c b/util/async.c
 index XXXXXXX..XXXXXXX 100644
 --- a/util/async.c
 +++ b/util/async.c
@@ -XXX,XX +XXX,XX @@
  #include "qemu/main-loop.h"
  #include "qemu/atomic.h"
  #include "block/raw-aio.h"
 +#include "qemu/coroutine_int.h"
 +#include "trace.h"
  /***********************************************************/
  /* bottom halves (can be seen as timers which expire ASAP) */
@@ -XXX,XX +XXX,XX @@ aio_ctx_finalize(GSource     *source)
      }
  #endif
 +    assert(QSLIST_EMPTY(&ctx->scheduled_coroutines));
 +    qemu_bh_delete(ctx->co_schedule_bh);
 +
      qemu_lockcnt_lock(&ctx->list_lock);
      assert(!qemu_lockcnt_count(&ctx->list_lock));
      while (ctx->first_bh) {
@@ -XXX,XX +XXX,XX @@ static bool event_notifier_poll(void *opaque)
      return atomic_read(&ctx->notified);
  }
 +static void co_schedule_bh_cb(void *opaque)
 +{
 +    AioContext *ctx = opaque;
 +    QSLIST_HEAD(, Coroutine) straight, reversed;
 +
 +    QSLIST_MOVE_ATOMIC(&reversed, &ctx->scheduled_coroutines);
 +    QSLIST_INIT(&straight);
 +
 +    while (!QSLIST_EMPTY(&reversed)) {
 +        Coroutine *co = QSLIST_FIRST(&reversed);
 +        QSLIST_REMOVE_HEAD(&reversed, co_scheduled_next);
 +        QSLIST_INSERT_HEAD(&straight, co, co_scheduled_next);
 +    }
 +
 +    while (!QSLIST_EMPTY(&straight)) {
 +        Coroutine *co = QSLIST_FIRST(&straight);
 +        QSLIST_REMOVE_HEAD(&straight, co_scheduled_next);
 +        trace_aio_co_schedule_bh_cb(ctx, co);
 +        qemu_coroutine_enter(co);
 +    }
 +}
 +
  AioContext *aio_context_new(Error **errp)
  {
      int ret;
@@ -XXX,XX +XXX,XX @@ AioContext *aio_context_new(Error **errp)
      }
      g_source_set_can_recurse(&ctx->source, true);
      qemu_lockcnt_init(&ctx->list_lock);
 +
 +    ctx->co_schedule_bh = aio_bh_new(ctx, co_schedule_bh_cb, ctx);
 +    QSLIST_INIT(&ctx->scheduled_coroutines);
 +
      aio_set_event_notifier(ctx, &ctx->notifier,
                             false,
                             (EventNotifierHandler *)
@@ -XXX,XX +XXX,XX @@ fail:
      return NULL;
  }
 +void aio_co_schedule(AioContext *ctx, Coroutine *co)
 +{
 +    trace_aio_co_schedule(ctx, co);
 +    QSLIST_INSERT_HEAD_ATOMIC(&ctx->scheduled_coroutines,
 +                              co, co_scheduled_next);
 +    qemu_bh_schedule(ctx->co_schedule_bh);
 +}
 +
 +void aio_co_wake(struct Coroutine *co)
 +{
 +    AioContext *ctx;
 +
 +    /* Read coroutine before co->ctx.  Matches smp_wmb in
 +     * qemu_coroutine_enter.
 +     */
 +    smp_read_barrier_depends();
 +    ctx = atomic_read(&co->ctx);
 +
 +    if (ctx != qemu_get_current_aio_context()) {
 +        aio_co_schedule(ctx, co);
 +        return;
 +    }
 +
 +    if (qemu_in_coroutine()) {
 +        Coroutine *self = qemu_coroutine_self();
 +        assert(self != co);
 +        QSIMPLEQ_INSERT_TAIL(&self->co_queue_wakeup, co, co_queue_next);
 +    } else {
 +        aio_context_acquire(ctx);
 +        qemu_coroutine_enter(co);
 +        aio_context_release(ctx);
 +    }
 +}
 +
  void aio_context_ref(AioContext *ctx)
  {
      g_source_ref(&ctx->source);
 diff --git a/util/qemu-coroutine.c b/util/qemu-coroutine.c
 index XXXXXXX..XXXXXXX 100644
 --- a/util/qemu-coroutine.c
 +++ b/util/qemu-coroutine.c
@@ -XXX,XX +XXX,XX @@
  #include "qemu/atomic.h"
  #include "qemu/coroutine.h"
  #include "qemu/coroutine_int.h"
 +#include "block/aio.h"
  enum {
      POOL_BATCH_SIZE = 64,
@@ -XXX,XX +XXX,XX @@ void qemu_coroutine_enter(Coroutine *co)
      }
      co->caller = self;
 +    co->ctx = qemu_get_current_aio_context();
 +
 +    /* Store co->ctx before anything that stores co.  Matches
 +     * barrier in aio_co_wake.
 +     */
 +    smp_wmb();
 +
      ret = qemu_coroutine_switch(self, co, COROUTINE_ENTER);
      qemu_co_queue_run_restart(co);
 diff --git a/util/trace-events b/util/trace-events
 index XXXXXXX..XXXXXXX 100644
 --- a/util/trace-events
 +++ b/util/trace-events
@@ -XXX,XX +XXX,XX @@ run_poll_handlers_end(void *ctx, bool progress) "ctx %p progress %d"
  poll_shrink(void *ctx, int64_t old, int64_t new) "ctx %p old %"PRId64" new %"PRId64
  poll_grow(void *ctx, int64_t old, int64_t new) "ctx %p old %"PRId64" new %"PRId64
 +# util/async.c
 +aio_co_schedule(void *ctx, void *co) "ctx %p co %p"
 +aio_co_schedule_bh_cb(void *ctx, void *co) "ctx %p co %p"
 +
  # util/thread-pool.c
  thread_pool_submit(void *pool, void *req, void *opaque) "pool %p req %p opaque %p"
  thread_pool_complete(void *pool, void *req, void *opaque, int ret) "pool %p req %p opaque %p ret %d"
 --
 .9.3

-New patch
+[Qemu-devel] [PULL v2 03/24] block-backend: allow blk_prw from coroutine context
+From: Paolo Bonzini <pbonzini@redhat.com>
+qcow2_create2 calls this.  Do not run a nested event loop, as that
+breaks when aio_co_wake tries to queue the coroutine on the co_queue_wakeup
+list of the currently running one.
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
+Reviewed-by: Fam Zheng <famz@redhat.com>
+Message-id: 20170213135235.12274-4-pbonzini@redhat.com
+Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
+---
+ block/block-backend.c | 12 ++++++++----
+file changed, 8 insertions(+), 4 deletions(-)
+diff --git a/block/block-backend.c b/block/block-backend.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/block-backend.c
++++ b/block/block-backend.c
+@@ -XXX,XX +XXX,XX @@ static int blk_prw(BlockBackend *blk, int64_t offset, uint8_t *buf,
+ {
+     QEMUIOVector qiov;
+     struct iovec iov;
+-    Coroutine *co;
+     BlkRwCo rwco;
+     iov = (struct iovec) {
+@@ -XXX,XX +XXX,XX @@ static int blk_prw(BlockBackend *blk, int64_t offset, uint8_t *buf,
+         .ret    = NOT_DONE,
+     };
+-    co = qemu_coroutine_create(co_entry, &rwco);
+-    qemu_coroutine_enter(co);
+-    BDRV_POLL_WHILE(blk_bs(blk), rwco.ret == NOT_DONE);
++    if (qemu_in_coroutine()) {
++        /* Fast-path if already in coroutine context */
++        co_entry(&rwco);
++    } else {
++        Coroutine *co = qemu_coroutine_create(co_entry, &rwco);
++        qemu_coroutine_enter(co);
++        BDRV_POLL_WHILE(blk_bs(blk), rwco.ret == NOT_DONE);
++    }
+     return rwco.ret;
+ }
+--
+.9.3

-New patch
+[Qemu-devel] [PULL v2 04/24] test-thread-pool: use generic AioContext infrastructure
+From: Paolo Bonzini <pbonzini@redhat.com>
+Once the thread pool starts using aio_co_wake, it will also need
+qemu_get_current_aio_context().  Make test-thread-pool create
+an AioContext with qemu_init_main_loop, so that stubs/iothread.c
+and tests/iothread.c can provide the rest.
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
+Reviewed-by: Fam Zheng <famz@redhat.com>
+Message-id: 20170213135235.12274-5-pbonzini@redhat.com
+Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
+---
+ tests/test-thread-pool.c | 12 +++---------
+file changed, 3 insertions(+), 9 deletions(-)
+diff --git a/tests/test-thread-pool.c b/tests/test-thread-pool.c
+index XXXXXXX..XXXXXXX 100644
+--- a/tests/test-thread-pool.c
++++ b/tests/test-thread-pool.c
+@@ -XXX,XX +XXX,XX @@
+ #include "qapi/error.h"
+ #include "qemu/timer.h"
+ #include "qemu/error-report.h"
++#include "qemu/main-loop.h"
+ static AioContext *ctx;
+ static ThreadPool *pool;
+@@ -XXX,XX +XXX,XX @@ static void test_cancel_async(void)
+ int main(int argc, char **argv)
+ {
+     int ret;
+-    Error *local_error = NULL;
+-    init_clocks();
+-
+-    ctx = aio_context_new(&local_error);
+-    if (!ctx) {
+-        error_reportf_err(local_error, "Failed to create AIO Context: ");
+-        exit(1);
+-    }
++    qemu_init_main_loop(&error_abort);
++    ctx = qemu_get_current_aio_context();
+     pool = aio_get_thread_pool(ctx);
+     g_test_init(&argc, &argv, NULL);
+@@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv)
+     ret = g_test_run();
+-    aio_context_unref(ctx);
+     return ret;
+ }
+--
+.9.3

-New patch
+[Qemu-devel] [PULL v2 05/24] io: add methods to set I/O handlers on AioContext
+From: Paolo Bonzini <pbonzini@redhat.com>
 This is in preparation for making qio_channel_yield work on
 AioContexts other than the main one.
 Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
 Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
 Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
 Reviewed-by: Fam Zheng <famz@redhat.com>
 Message-id: 20170213135235.12274-6-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
  include/io/channel.h | 25 +++++++++++++++++++++++++
  io/channel-command.c | 13 +++++++++++++
  io/channel-file.c    | 11 +++++++++++
  io/channel-socket.c  | 16 +++++++++++-----
  io/channel-tls.c     | 12 ++++++++++++
  io/channel-watch.c   |  6 ++++++
  io/channel.c         | 11 +++++++++++
 files changed, 89 insertions(+), 5 deletions(-)
 diff --git a/include/io/channel.h b/include/io/channel.h
 index XXXXXXX..XXXXXXX 100644
 --- a/include/io/channel.h
 +++ b/include/io/channel.h
@@ -XXX,XX +XXX,XX @@
  #include "qemu-common.h"
  #include "qom/object.h"
 +#include "block/aio.h"
  #define TYPE_QIO_CHANNEL "qio-channel"
  #define QIO_CHANNEL(obj)                                    \
@@ -XXX,XX +XXX,XX @@ struct QIOChannelClass {
                       off_t offset,
                       int whence,
                       Error **errp);
 +    void (*io_set_aio_fd_handler)(QIOChannel *ioc,
 +                                  AioContext *ctx,
 +                                  IOHandler *io_read,
 +                                  IOHandler *io_write,
 +                                  void *opaque);
  };
  /* General I/O handling functions */
@@ -XXX,XX +XXX,XX @@ void qio_channel_yield(QIOChannel *ioc,
  void qio_channel_wait(QIOChannel *ioc,
                        GIOCondition condition);
 +/**
 + * qio_channel_set_aio_fd_handler:
 + * @ioc: the channel object
 + * @ctx: the AioContext to set the handlers on
 + * @io_read: the read handler
 + * @io_write: the write handler
 + * @opaque: the opaque value passed to the handler
 + *
 + * This is used internally by qio_channel_yield().  It can
 + * be used by channel implementations to forward the handlers
 + * to another channel (e.g. from #QIOChannelTLS to the
 + * underlying socket).
 + */
 +void qio_channel_set_aio_fd_handler(QIOChannel *ioc,
 +                                    AioContext *ctx,
 +                                    IOHandler *io_read,
 +                                    IOHandler *io_write,
 +                                    void *opaque);
 +
  #endif /* QIO_CHANNEL_H */
 diff --git a/io/channel-command.c b/io/channel-command.c
 index XXXXXXX..XXXXXXX 100644
 --- a/io/channel-command.c
 +++ b/io/channel-command.c
@@ -XXX,XX +XXX,XX @@ static int qio_channel_command_close(QIOChannel *ioc,
  }
 +static void qio_channel_command_set_aio_fd_handler(QIOChannel *ioc,
 +                                                   AioContext *ctx,
 +                                                   IOHandler *io_read,
 +                                                   IOHandler *io_write,
 +                                                   void *opaque)
 +{
 +    QIOChannelCommand *cioc = QIO_CHANNEL_COMMAND(ioc);
 +    aio_set_fd_handler(ctx, cioc->readfd, false, io_read, NULL, NULL, opaque);
 +    aio_set_fd_handler(ctx, cioc->writefd, false, NULL, io_write, NULL, opaque);
 +}
 +
 +
  static GSource *qio_channel_command_create_watch(QIOChannel *ioc,
                                                   GIOCondition condition)
  {
@@ -XXX,XX +XXX,XX @@ static void qio_channel_command_class_init(ObjectClass *klass,
      ioc_klass->io_set_blocking = qio_channel_command_set_blocking;
      ioc_klass->io_close = qio_channel_command_close;
      ioc_klass->io_create_watch = qio_channel_command_create_watch;
 +    ioc_klass->io_set_aio_fd_handler = qio_channel_command_set_aio_fd_handler;
  }
  static const TypeInfo qio_channel_command_info = {
 diff --git a/io/channel-file.c b/io/channel-file.c
 index XXXXXXX..XXXXXXX 100644
 --- a/io/channel-file.c
 +++ b/io/channel-file.c
@@ -XXX,XX +XXX,XX @@ static int qio_channel_file_close(QIOChannel *ioc,
  }
 +static void qio_channel_file_set_aio_fd_handler(QIOChannel *ioc,
 +                                                AioContext *ctx,
 +                                                IOHandler *io_read,
 +                                                IOHandler *io_write,
 +                                                void *opaque)
 +{
 +    QIOChannelFile *fioc = QIO_CHANNEL_FILE(ioc);
 +    aio_set_fd_handler(ctx, fioc->fd, false, io_read, io_write, NULL, opaque);
 +}
 +
  static GSource *qio_channel_file_create_watch(QIOChannel *ioc,
                                                GIOCondition condition)
  {
@@ -XXX,XX +XXX,XX @@ static void qio_channel_file_class_init(ObjectClass *klass,
      ioc_klass->io_seek = qio_channel_file_seek;
      ioc_klass->io_close = qio_channel_file_close;
      ioc_klass->io_create_watch = qio_channel_file_create_watch;
 +    ioc_klass->io_set_aio_fd_handler = qio_channel_file_set_aio_fd_handler;
  }
  static const TypeInfo qio_channel_file_info = {
 diff --git a/io/channel-socket.c b/io/channel-socket.c
 index XXXXXXX..XXXXXXX 100644
 --- a/io/channel-socket.c
 +++ b/io/channel-socket.c
@@ -XXX,XX +XXX,XX @@ qio_channel_socket_set_blocking(QIOChannel *ioc,
          qemu_set_block(sioc->fd);
      } else {
          qemu_set_nonblock(sioc->fd);
 -#ifdef WIN32
 -        WSAEventSelect(sioc->fd, ioc->event,
 -                       FD_READ | FD_ACCEPT | FD_CLOSE |
 -                       FD_CONNECT | FD_WRITE | FD_OOB);
 -#endif
      }
      return 0;
  }
@@ -XXX,XX +XXX,XX @@ qio_channel_socket_shutdown(QIOChannel *ioc,
      return 0;
  }
 +static void qio_channel_socket_set_aio_fd_handler(QIOChannel *ioc,
 +                                                  AioContext *ctx,
 +                                                  IOHandler *io_read,
 +                                                  IOHandler *io_write,
 +                                                  void *opaque)
 +{
 +    QIOChannelSocket *sioc = QIO_CHANNEL_SOCKET(ioc);
 +    aio_set_fd_handler(ctx, sioc->fd, false, io_read, io_write, NULL, opaque);
 +}
 +
  static GSource *qio_channel_socket_create_watch(QIOChannel *ioc,
                                                  GIOCondition condition)
  {
@@ -XXX,XX +XXX,XX @@ static void qio_channel_socket_class_init(ObjectClass *klass,
      ioc_klass->io_set_cork = qio_channel_socket_set_cork;
      ioc_klass->io_set_delay = qio_channel_socket_set_delay;
      ioc_klass->io_create_watch = qio_channel_socket_create_watch;
 +    ioc_klass->io_set_aio_fd_handler = qio_channel_socket_set_aio_fd_handler;
  }
  static const TypeInfo qio_channel_socket_info = {
 diff --git a/io/channel-tls.c b/io/channel-tls.c
 index XXXXXXX..XXXXXXX 100644
 --- a/io/channel-tls.c
 +++ b/io/channel-tls.c
@@ -XXX,XX +XXX,XX @@ static int qio_channel_tls_close(QIOChannel *ioc,
      return qio_channel_close(tioc->master, errp);
  }
 +static void qio_channel_tls_set_aio_fd_handler(QIOChannel *ioc,
 +                                               AioContext *ctx,
 +                                               IOHandler *io_read,
 +                                               IOHandler *io_write,
 +                                               void *opaque)
 +{
 +    QIOChannelTLS *tioc = QIO_CHANNEL_TLS(ioc);
 +
 +    qio_channel_set_aio_fd_handler(tioc->master, ctx, io_read, io_write, opaque);
 +}
 +
  static GSource *qio_channel_tls_create_watch(QIOChannel *ioc,
                                               GIOCondition condition)
  {
@@ -XXX,XX +XXX,XX @@ static void qio_channel_tls_class_init(ObjectClass *klass,
      ioc_klass->io_close = qio_channel_tls_close;
      ioc_klass->io_shutdown = qio_channel_tls_shutdown;
      ioc_klass->io_create_watch = qio_channel_tls_create_watch;
 +    ioc_klass->io_set_aio_fd_handler = qio_channel_tls_set_aio_fd_handler;
  }
  static const TypeInfo qio_channel_tls_info = {
 diff --git a/io/channel-watch.c b/io/channel-watch.c
 index XXXXXXX..XXXXXXX 100644
 --- a/io/channel-watch.c
 +++ b/io/channel-watch.c
@@ -XXX,XX +XXX,XX @@ GSource *qio_channel_create_socket_watch(QIOChannel *ioc,
      GSource *source;
      QIOChannelSocketSource *ssource;
 +#ifdef WIN32
 +    WSAEventSelect(socket, ioc->event,
 +                   FD_READ | FD_ACCEPT | FD_CLOSE |
 +                   FD_CONNECT | FD_WRITE | FD_OOB);
 +#endif
 +
      source = g_source_new(&qio_channel_socket_source_funcs,
                            sizeof(QIOChannelSocketSource));
      ssource = (QIOChannelSocketSource *)source;
 diff --git a/io/channel.c b/io/channel.c
 index XXXXXXX..XXXXXXX 100644
 --- a/io/channel.c
 +++ b/io/channel.c
@@ -XXX,XX +XXX,XX @@ GSource *qio_channel_create_watch(QIOChannel *ioc,
  }
 +void qio_channel_set_aio_fd_handler(QIOChannel *ioc,
 +                                    AioContext *ctx,
 +                                    IOHandler *io_read,
 +                                    IOHandler *io_write,
 +                                    void *opaque)
 +{
 +    QIOChannelClass *klass = QIO_CHANNEL_GET_CLASS(ioc);
 +
 +    klass->io_set_aio_fd_handler(ioc, ctx, io_read, io_write, opaque);
 +}
 +
  guint qio_channel_add_watch(QIOChannel *ioc,
                              GIOCondition condition,
                              QIOChannelFunc func,
 --
 .9.3

-New patch
+[Qemu-devel] [PULL v2 06/24] io: make qio_channel_yield aware of AioContexts
+From: Paolo Bonzini <pbonzini@redhat.com>
 Support separate coroutines for reading and writing, and place the
 read/write handlers on the AioContext that the QIOChannel is registered
 with.
 Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
 Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
 Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
 Reviewed-by: Fam Zheng <famz@redhat.com>
 Message-id: 20170213135235.12274-7-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
  include/io/channel.h | 47 ++++++++++++++++++++++++++--
  io/channel.c         | 86 +++++++++++++++++++++++++++++++++++++++-------------
 files changed, 109 insertions(+), 24 deletions(-)
 diff --git a/include/io/channel.h b/include/io/channel.h
 index XXXXXXX..XXXXXXX 100644
 --- a/include/io/channel.h
 +++ b/include/io/channel.h
@@ -XXX,XX +XXX,XX @@
  #include "qemu-common.h"
  #include "qom/object.h"
 +#include "qemu/coroutine.h"
  #include "block/aio.h"
  #define TYPE_QIO_CHANNEL "qio-channel"
@@ -XXX,XX +XXX,XX @@ struct QIOChannel {
      Object parent;
      unsigned int features; /* bitmask of QIOChannelFeatures */
      char *name;
 +    AioContext *ctx;
 +    Coroutine *read_coroutine;
 +    Coroutine *write_coroutine;
  #ifdef _WIN32
      HANDLE event; /* For use with GSource on Win32 */
  #endif
@@ -XXX,XX +XXX,XX @@ guint qio_channel_add_watch(QIOChannel *ioc,
  /**
 + * qio_channel_attach_aio_context:
 + * @ioc: the channel object
 + * @ctx: the #AioContext to set the handlers on
 + *
 + * Request that qio_channel_yield() sets I/O handlers on
 + * the given #AioContext.  If @ctx is %NULL, qio_channel_yield()
 + * uses QEMU's main thread event loop.
 + *
 + * You can move a #QIOChannel from one #AioContext to another even if
 + * I/O handlers are set for a coroutine.  However, #QIOChannel provides
 + * no synchronization between the calls to qio_channel_yield() and
 + * qio_channel_attach_aio_context().
 + *
 + * Therefore you should first call qio_channel_detach_aio_context()
 + * to ensure that the coroutine is not entered concurrently.  Then,
 + * while the coroutine has yielded, call qio_channel_attach_aio_context(),
 + * and then aio_co_schedule() to place the coroutine on the new
 + * #AioContext.  The calls to qio_channel_detach_aio_context()
 + * and qio_channel_attach_aio_context() should be protected with
 + * aio_context_acquire() and aio_context_release().
 + */
 +void qio_channel_attach_aio_context(QIOChannel *ioc,
 +                                    AioContext *ctx);
 +
 +/**
 + * qio_channel_detach_aio_context:
 + * @ioc: the channel object
 + *
 + * Disable any I/O handlers set by qio_channel_yield().  With the
 + * help of aio_co_schedule(), this allows moving a coroutine that was
 + * paused by qio_channel_yield() to another context.
 + */
 +void qio_channel_detach_aio_context(QIOChannel *ioc);
 +
 +/**
   * qio_channel_yield:
   * @ioc: the channel object
   * @condition: the I/O condition to wait for
   *
 - * Yields execution from the current coroutine until
 - * the condition indicated by @condition becomes
 - * available.
 + * Yields execution from the current coroutine until the condition
 + * indicated by @condition becomes available.  @condition must
 + * be either %G_IO_IN or %G_IO_OUT; it cannot contain both.  In
 + * addition, no two coroutine can be waiting on the same condition
 + * and channel at the same time.
   *
   * This must only be called from coroutine context
   */
 diff --git a/io/channel.c b/io/channel.c
 index XXXXXXX..XXXXXXX 100644
 --- a/io/channel.c
 +++ b/io/channel.c
@@ -XXX,XX +XXX,XX @@
  #include "qemu/osdep.h"
  #include "io/channel.h"
  #include "qapi/error.h"
 -#include "qemu/coroutine.h"
 +#include "qemu/main-loop.h"
  bool qio_channel_has_feature(QIOChannel *ioc,
                               QIOChannelFeature feature)
@@ -XXX,XX +XXX,XX @@ off_t qio_channel_io_seek(QIOChannel *ioc,
  }
 -typedef struct QIOChannelYieldData QIOChannelYieldData;
 -struct QIOChannelYieldData {
 -    QIOChannel *ioc;
 -    Coroutine *co;
 -};
 +static void qio_channel_set_aio_fd_handlers(QIOChannel *ioc);
 +static void qio_channel_restart_read(void *opaque)
 +{
 +    QIOChannel *ioc = opaque;
 +    Coroutine *co = ioc->read_coroutine;
 +
 +    ioc->read_coroutine = NULL;
 +    qio_channel_set_aio_fd_handlers(ioc);
 +    aio_co_wake(co);
 +}
 -static gboolean qio_channel_yield_enter(QIOChannel *ioc,
 -                                        GIOCondition condition,
 -                                        gpointer opaque)
 +static void qio_channel_restart_write(void *opaque)
  {
 -    QIOChannelYieldData *data = opaque;
 -    qemu_coroutine_enter(data->co);
 -    return FALSE;
 +    QIOChannel *ioc = opaque;
 +    Coroutine *co = ioc->write_coroutine;
 +
 +    ioc->write_coroutine = NULL;
 +    qio_channel_set_aio_fd_handlers(ioc);
 +    aio_co_wake(co);
  }
 +static void qio_channel_set_aio_fd_handlers(QIOChannel *ioc)
 +{
 +    IOHandler *rd_handler = NULL, *wr_handler = NULL;
 +    AioContext *ctx;
 +
 +    if (ioc->read_coroutine) {
 +        rd_handler = qio_channel_restart_read;
 +    }
 +    if (ioc->write_coroutine) {
 +        wr_handler = qio_channel_restart_write;
 +    }
 +
 +    ctx = ioc->ctx ? ioc->ctx : iohandler_get_aio_context();
 +    qio_channel_set_aio_fd_handler(ioc, ctx, rd_handler, wr_handler, ioc);
 +}
 +
 +void qio_channel_attach_aio_context(QIOChannel *ioc,
 +                                    AioContext *ctx)
 +{
 +    AioContext *old_ctx;
 +    if (ioc->ctx == ctx) {
 +        return;
 +    }
 +
 +    old_ctx = ioc->ctx ? ioc->ctx : iohandler_get_aio_context();
 +    qio_channel_set_aio_fd_handler(ioc, old_ctx, NULL, NULL, NULL);
 +    ioc->ctx = ctx;
 +    qio_channel_set_aio_fd_handlers(ioc);
 +}
 +
 +void qio_channel_detach_aio_context(QIOChannel *ioc)
 +{
 +    ioc->read_coroutine = NULL;
 +    ioc->write_coroutine = NULL;
 +    qio_channel_set_aio_fd_handlers(ioc);
 +    ioc->ctx = NULL;
 +}
  void coroutine_fn qio_channel_yield(QIOChannel *ioc,
                                      GIOCondition condition)
  {
 -    QIOChannelYieldData data;
 -
      assert(qemu_in_coroutine());
 -    data.ioc = ioc;
 -    data.co = qemu_coroutine_self();
 -    qio_channel_add_watch(ioc,
 -                          condition,
 -                          qio_channel_yield_enter,
 -                          &data,
 -                          NULL);
 +    if (condition == G_IO_IN) {
 +        assert(!ioc->read_coroutine);
 +        ioc->read_coroutine = qemu_coroutine_self();
 +    } else if (condition == G_IO_OUT) {
 +        assert(!ioc->write_coroutine);
 +        ioc->write_coroutine = qemu_coroutine_self();
 +    } else {
 +        abort();
 +    }
 +    qio_channel_set_aio_fd_handlers(ioc);
      qemu_coroutine_yield();
  }
 --
 .9.3

-New patch
+[Qemu-devel] [PULL v2 07/24] nbd: convert to use qio_channel_yield
+From: Paolo Bonzini <pbonzini@redhat.com>
+In the client, read the reply headers from a coroutine, switching the
+read side between the "read header" coroutine and the I/O coroutine that
+reads the body of the reply.
+In the server, if the server can read more requests it will create a new
+"read request" coroutine as soon as a request has been read.  Otherwise,
+the new coroutine is created in nbd_request_put.
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
+Reviewed-by: Fam Zheng <famz@redhat.com>
+Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
+Message-id: 20170213135235.12274-8-pbonzini@redhat.com
+Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
+---
+ block/nbd-client.h |   2 +-
+ block/nbd-client.c | 117 ++++++++++++++++++++++++-----------------------------
+ nbd/client.c       |   2 +-
+ nbd/common.c       |   9 +----
+ nbd/server.c       |  94 +++++++++++++-----------------------------
+files changed, 83 insertions(+), 141 deletions(-)
+diff --git a/block/nbd-client.h b/block/nbd-client.h
+index XXXXXXX..XXXXXXX 100644
+--- a/block/nbd-client.h
++++ b/block/nbd-client.h
+@@ -XXX,XX +XXX,XX @@ typedef struct NBDClientSession {
+     CoMutex send_mutex;
+     CoQueue free_sema;
+-    Coroutine *send_coroutine;
++    Coroutine *read_reply_co;
+     int in_flight;
+     Coroutine *recv_coroutine[MAX_NBD_REQUESTS];
+diff --git a/block/nbd-client.c b/block/nbd-client.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/nbd-client.c
++++ b/block/nbd-client.c
+@@ -XXX,XX +XXX,XX @@
+ #define HANDLE_TO_INDEX(bs, handle) ((handle) ^ ((uint64_t)(intptr_t)bs))
+ #define INDEX_TO_HANDLE(bs, index)  ((index)  ^ ((uint64_t)(intptr_t)bs))
+-static void nbd_recv_coroutines_enter_all(NBDClientSession *s)
++static void nbd_recv_coroutines_enter_all(BlockDriverState *bs)
+ {
++    NBDClientSession *s = nbd_get_client_session(bs);
+     int i;
+     for (i = 0; i < MAX_NBD_REQUESTS; i++) {
+@@ -XXX,XX +XXX,XX @@ static void nbd_recv_coroutines_enter_all(NBDClientSession *s)
+             qemu_coroutine_enter(s->recv_coroutine[i]);
+         }
+     }
++    BDRV_POLL_WHILE(bs, s->read_reply_co);
+ }
+ static void nbd_teardown_connection(BlockDriverState *bs)
+@@ -XXX,XX +XXX,XX @@ static void nbd_teardown_connection(BlockDriverState *bs)
+     qio_channel_shutdown(client->ioc,
+                          QIO_CHANNEL_SHUTDOWN_BOTH,
+                          NULL);
+-    nbd_recv_coroutines_enter_all(client);
++    nbd_recv_coroutines_enter_all(bs);
+     nbd_client_detach_aio_context(bs);
+     object_unref(OBJECT(client->sioc));
+@@ -XXX,XX +XXX,XX @@ static void nbd_teardown_connection(BlockDriverState *bs)
+     client->ioc = NULL;
+ }
+-static void nbd_reply_ready(void *opaque)
++static coroutine_fn void nbd_read_reply_entry(void *opaque)
+ {
+-    BlockDriverState *bs = opaque;
+-    NBDClientSession *s = nbd_get_client_session(bs);
++    NBDClientSession *s = opaque;
+     uint64_t i;
+     int ret;
+-    if (!s->ioc) { /* Already closed */
+-        return;
+-    }
+-
+-    if (s->reply.handle == 0) {
+-        /* No reply already in flight.  Fetch a header.  It is possible
+-         * that another thread has done the same thing in parallel, so
+-         * the socket is not readable anymore.
+-         */
++    for (;;) {
++        assert(s->reply.handle == 0);
+         ret = nbd_receive_reply(s->ioc, &s->reply);
+-        if (ret == -EAGAIN) {
+-            return;
+-        }
+         if (ret < 0) {
+-            s->reply.handle = 0;
+-            goto fail;
++            break;
+         }
+-    }
+-    /* There's no need for a mutex on the receive side, because the
+-     * handler acts as a synchronization point and ensures that only
+-     * one coroutine is called until the reply finishes.  */
+-    i = HANDLE_TO_INDEX(s, s->reply.handle);
+-    if (i >= MAX_NBD_REQUESTS) {
+-        goto fail;
+-    }
++        /* There's no need for a mutex on the receive side, because the
++         * handler acts as a synchronization point and ensures that only
++         * one coroutine is called until the reply finishes.
++         */
++        i = HANDLE_TO_INDEX(s, s->reply.handle);
++        if (i >= MAX_NBD_REQUESTS || !s->recv_coroutine[i]) {
++            break;
++        }
+-    if (s->recv_coroutine[i]) {
+-        qemu_coroutine_enter(s->recv_coroutine[i]);
+-        return;
++        /* We're woken up by the recv_coroutine itself.  Note that there
++         * is no race between yielding and reentering read_reply_co.  This
++         * is because:
++         *
++         * - if recv_coroutine[i] runs on the same AioContext, it is only
++         *   entered after we yield
++         *
++         * - if recv_coroutine[i] runs on a different AioContext, reentering
++         *   read_reply_co happens through a bottom half, which can only
++         *   run after we yield.
++         */
++        aio_co_wake(s->recv_coroutine[i]);
++        qemu_coroutine_yield();
+     }
+-
+-fail:
+-    nbd_teardown_connection(bs);
+-}
+-
+-static void nbd_restart_write(void *opaque)
+-{
+-    BlockDriverState *bs = opaque;
+-
+-    qemu_coroutine_enter(nbd_get_client_session(bs)->send_coroutine);
++    s->read_reply_co = NULL;
+ }
+ static int nbd_co_send_request(BlockDriverState *bs,
+@@ -XXX,XX +XXX,XX @@ static int nbd_co_send_request(BlockDriverState *bs,
+                                QEMUIOVector *qiov)
+ {
+     NBDClientSession *s = nbd_get_client_session(bs);
+-    AioContext *aio_context;
+     int rc, ret, i;
+     qemu_co_mutex_lock(&s->send_mutex);
+@@ -XXX,XX +XXX,XX @@ static int nbd_co_send_request(BlockDriverState *bs,
+         return -EPIPE;
+     }
+-    s->send_coroutine = qemu_coroutine_self();
+-    aio_context = bdrv_get_aio_context(bs);
+-
+-    aio_set_fd_handler(aio_context, s->sioc->fd, false,
+-                       nbd_reply_ready, nbd_restart_write, NULL, bs);
+     if (qiov) {
+         qio_channel_set_cork(s->ioc, true);
+         rc = nbd_send_request(s->ioc, request);
+@@ -XXX,XX +XXX,XX @@ static int nbd_co_send_request(BlockDriverState *bs,
+     } else {
+         rc = nbd_send_request(s->ioc, request);
+     }
+-    aio_set_fd_handler(aio_context, s->sioc->fd, false,
+-                       nbd_reply_ready, NULL, NULL, bs);
+-    s->send_coroutine = NULL;
+     qemu_co_mutex_unlock(&s->send_mutex);
+     return rc;
+ }
+@@ -XXX,XX +XXX,XX @@ static void nbd_co_receive_reply(NBDClientSession *s,
+ {
+     int ret;
+-    /* Wait until we're woken up by the read handler.  TODO: perhaps
+-     * peek at the next reply and avoid yielding if it's ours?  */
++    /* Wait until we're woken up by nbd_read_reply_entry.  */
+     qemu_coroutine_yield();
+     *reply = s->reply;
+     if (reply->handle != request->handle ||
+@@ -XXX,XX +XXX,XX @@ static void nbd_coroutine_start(NBDClientSession *s,
+     /* s->recv_coroutine[i] is set as soon as we get the send_lock.  */
+ }
+-static void nbd_coroutine_end(NBDClientSession *s,
++static void nbd_coroutine_end(BlockDriverState *bs,
+                               NBDRequest *request)
+ {
++    NBDClientSession *s = nbd_get_client_session(bs);
+     int i = HANDLE_TO_INDEX(s, request->handle);
++
+     s->recv_coroutine[i] = NULL;
+-    if (s->in_flight-- == MAX_NBD_REQUESTS) {
+-        qemu_co_queue_next(&s->free_sema);
++    s->in_flight--;
++    qemu_co_queue_next(&s->free_sema);
++
++    /* Kick the read_reply_co to get the next reply.  */
++    if (s->read_reply_co) {
++        aio_co_wake(s->read_reply_co);
+     }
+ }
+@@ -XXX,XX +XXX,XX @@ int nbd_client_co_preadv(BlockDriverState *bs, uint64_t offset,
+     } else {
+         nbd_co_receive_reply(client, &request, &reply, qiov);
+     }
+-    nbd_coroutine_end(client, &request);
++    nbd_coroutine_end(bs, &request);
+     return -reply.error;
+ }
+@@ -XXX,XX +XXX,XX @@ int nbd_client_co_pwritev(BlockDriverState *bs, uint64_t offset,
+     } else {
+         nbd_co_receive_reply(client, &request, &reply, NULL);
+     }
+-    nbd_coroutine_end(client, &request);
++    nbd_coroutine_end(bs, &request);
+     return -reply.error;
+ }
+@@ -XXX,XX +XXX,XX @@ int nbd_client_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset,
+     } else {
+         nbd_co_receive_reply(client, &request, &reply, NULL);
+     }
+-    nbd_coroutine_end(client, &request);
++    nbd_coroutine_end(bs, &request);
+     return -reply.error;
+ }
+@@ -XXX,XX +XXX,XX @@ int nbd_client_co_flush(BlockDriverState *bs)
+     } else {
+         nbd_co_receive_reply(client, &request, &reply, NULL);
+     }
+-    nbd_coroutine_end(client, &request);
++    nbd_coroutine_end(bs, &request);
+     return -reply.error;
+ }
+@@ -XXX,XX +XXX,XX @@ int nbd_client_co_pdiscard(BlockDriverState *bs, int64_t offset, int count)
+     } else {
+         nbd_co_receive_reply(client, &request, &reply, NULL);
+     }
+-    nbd_coroutine_end(client, &request);
++    nbd_coroutine_end(bs, &request);
+     return -reply.error;
+ }
+ void nbd_client_detach_aio_context(BlockDriverState *bs)
+ {
+-    aio_set_fd_handler(bdrv_get_aio_context(bs),
+-                       nbd_get_client_session(bs)->sioc->fd,
+-                       false, NULL, NULL, NULL, NULL);
++    NBDClientSession *client = nbd_get_client_session(bs);
++    qio_channel_detach_aio_context(QIO_CHANNEL(client->sioc));
+ }
+ void nbd_client_attach_aio_context(BlockDriverState *bs,
+                                    AioContext *new_context)
+ {
+-    aio_set_fd_handler(new_context, nbd_get_client_session(bs)->sioc->fd,
+-                       false, nbd_reply_ready, NULL, NULL, bs);
++    NBDClientSession *client = nbd_get_client_session(bs);
++    qio_channel_attach_aio_context(QIO_CHANNEL(client->sioc), new_context);
++    aio_co_schedule(new_context, client->read_reply_co);
+ }
+ void nbd_client_close(BlockDriverState *bs)
+@@ -XXX,XX +XXX,XX @@ int nbd_client_init(BlockDriverState *bs,
+     /* Now that we're connected, set the socket to be non-blocking and
+      * kick the reply mechanism.  */
+     qio_channel_set_blocking(QIO_CHANNEL(sioc), false, NULL);
+-
++    client->read_reply_co = qemu_coroutine_create(nbd_read_reply_entry, client);
+     nbd_client_attach_aio_context(bs, bdrv_get_aio_context(bs));
+     logout("Established connection with NBD server\n");
+diff --git a/nbd/client.c b/nbd/client.c
+index XXXXXXX..XXXXXXX 100644
+--- a/nbd/client.c
++++ b/nbd/client.c
+@@ -XXX,XX +XXX,XX @@ ssize_t nbd_receive_reply(QIOChannel *ioc, NBDReply *reply)
+     ssize_t ret;
+     ret = read_sync(ioc, buf, sizeof(buf));
+-    if (ret < 0) {
++    if (ret <= 0) {
+         return ret;
+     }
+diff --git a/nbd/common.c b/nbd/common.c
+index XXXXXXX..XXXXXXX 100644
+--- a/nbd/common.c
++++ b/nbd/common.c
+@@ -XXX,XX +XXX,XX @@ ssize_t nbd_wr_syncv(QIOChannel *ioc,
+         }
+         if (len == QIO_CHANNEL_ERR_BLOCK) {
+             if (qemu_in_coroutine()) {
+-                /* XXX figure out if we can create a variant on
+-                 * qio_channel_yield() that works with AIO contexts
+-                 * and consider using that in this branch */
+-                qemu_coroutine_yield();
+-            } else if (done) {
+-                /* XXX this is needed by nbd_reply_ready.  */
+-                qio_channel_wait(ioc,
+-                                 do_read ? G_IO_IN : G_IO_OUT);
++                qio_channel_yield(ioc, do_read ? G_IO_IN : G_IO_OUT);
+             } else {
+                 return -EAGAIN;
+             }
+diff --git a/nbd/server.c b/nbd/server.c
+index XXXXXXX..XXXXXXX 100644
+--- a/nbd/server.c
++++ b/nbd/server.c
+@@ -XXX,XX +XXX,XX @@ struct NBDClient {
+     CoMutex send_lock;
+     Coroutine *send_coroutine;
+-    bool can_read;
+-
+     QTAILQ_ENTRY(NBDClient) next;
+     int nb_requests;
+     bool closing;
+@@ -XXX,XX +XXX,XX @@ struct NBDClient {
+ /* That's all folks */
+-static void nbd_set_handlers(NBDClient *client);
+-static void nbd_unset_handlers(NBDClient *client);
+-static void nbd_update_can_read(NBDClient *client);
++static void nbd_client_receive_next_request(NBDClient *client);
+ static gboolean nbd_negotiate_continue(QIOChannel *ioc,
+                                        GIOCondition condition,
+@@ -XXX,XX +XXX,XX @@ void nbd_client_put(NBDClient *client)
+          */
+         assert(client->closing);
+-        nbd_unset_handlers(client);
++        qio_channel_detach_aio_context(client->ioc);
+         object_unref(OBJECT(client->sioc));
+         object_unref(OBJECT(client->ioc));
+         if (client->tlscreds) {
+@@ -XXX,XX +XXX,XX @@ static NBDRequestData *nbd_request_get(NBDClient *client)
+     assert(client->nb_requests <= MAX_NBD_REQUESTS - 1);
+     client->nb_requests++;
+-    nbd_update_can_read(client);
+     req = g_new0(NBDRequestData, 1);
+     nbd_client_get(client);
+@@ -XXX,XX +XXX,XX @@ static void nbd_request_put(NBDRequestData *req)
+     g_free(req);
+     client->nb_requests--;
+-    nbd_update_can_read(client);
++    nbd_client_receive_next_request(client);
++
+     nbd_client_put(client);
+ }
+@@ -XXX,XX +XXX,XX @@ static void blk_aio_attached(AioContext *ctx, void *opaque)
+     exp->ctx = ctx;
+     QTAILQ_FOREACH(client, &exp->clients, next) {
+-        nbd_set_handlers(client);
++        qio_channel_attach_aio_context(client->ioc, ctx);
++        if (client->recv_coroutine) {
++            aio_co_schedule(ctx, client->recv_coroutine);
++        }
++        if (client->send_coroutine) {
++            aio_co_schedule(ctx, client->send_coroutine);
++        }
+     }
+ }
+@@ -XXX,XX +XXX,XX @@ static void blk_aio_detach(void *opaque)
+     TRACE("Export %s: Detaching clients from AIO context %p\n", exp->name, exp->ctx);
+     QTAILQ_FOREACH(client, &exp->clients, next) {
+-        nbd_unset_handlers(client);
++        qio_channel_detach_aio_context(client->ioc);
+     }
+     exp->ctx = NULL;
+@@ -XXX,XX +XXX,XX @@ static ssize_t nbd_co_send_reply(NBDRequestData *req, NBDReply *reply,
+     g_assert(qemu_in_coroutine());
+     qemu_co_mutex_lock(&client->send_lock);
+     client->send_coroutine = qemu_coroutine_self();
+-    nbd_set_handlers(client);
+     if (!len) {
+         rc = nbd_send_reply(client->ioc, reply);
+@@ -XXX,XX +XXX,XX @@ static ssize_t nbd_co_send_reply(NBDRequestData *req, NBDReply *reply,
+     }
+     client->send_coroutine = NULL;
+-    nbd_set_handlers(client);
+     qemu_co_mutex_unlock(&client->send_lock);
+     return rc;
+ }
+@@ -XXX,XX +XXX,XX @@ static ssize_t nbd_co_receive_request(NBDRequestData *req,
+     ssize_t rc;
+     g_assert(qemu_in_coroutine());
+-    client->recv_coroutine = qemu_coroutine_self();
+-    nbd_update_can_read(client);
+-
++    assert(client->recv_coroutine == qemu_coroutine_self());
+     rc = nbd_receive_request(client->ioc, request);
+     if (rc < 0) {
+         if (rc != -EAGAIN) {
+@@ -XXX,XX +XXX,XX @@ static ssize_t nbd_co_receive_request(NBDRequestData *req,
+ out:
+     client->recv_coroutine = NULL;
+-    nbd_update_can_read(client);
++    nbd_client_receive_next_request(client);
+     return rc;
+ }
+-static void nbd_trip(void *opaque)
++/* Owns a reference to the NBDClient passed as opaque.  */
++static coroutine_fn void nbd_trip(void *opaque)
+ {
+     NBDClient *client = opaque;
+     NBDExport *exp = client->exp;
+     NBDRequestData *req;
+-    NBDRequest request;
++    NBDRequest request = { 0 };    /* GCC thinks it can be used uninitialized */
+     NBDReply reply;
+     ssize_t ret;
+     int flags;
+     TRACE("Reading request.");
+     if (client->closing) {
++        nbd_client_put(client);
+         return;
+     }
+@@ -XXX,XX +XXX,XX @@ static void nbd_trip(void *opaque)
+ done:
+     nbd_request_put(req);
++    nbd_client_put(client);
+     return;
+ out:
+     nbd_request_put(req);
+     client_close(client);
++    nbd_client_put(client);
+ }
+-static void nbd_read(void *opaque)
++static void nbd_client_receive_next_request(NBDClient *client)
+ {
+-    NBDClient *client = opaque;
+-
+-    if (client->recv_coroutine) {
+-        qemu_coroutine_enter(client->recv_coroutine);
+-    } else {
+-        qemu_coroutine_enter(qemu_coroutine_create(nbd_trip, client));
+-    }
+-}
+-
+-static void nbd_restart_write(void *opaque)
+-{
+-    NBDClient *client = opaque;
+-
+-    qemu_coroutine_enter(client->send_coroutine);
+-}
+-
+-static void nbd_set_handlers(NBDClient *client)
+-{
+-    if (client->exp && client->exp->ctx) {
+-        aio_set_fd_handler(client->exp->ctx, client->sioc->fd, true,
+-                           client->can_read ? nbd_read : NULL,
+-                           client->send_coroutine ? nbd_restart_write : NULL,
+-                           NULL, client);
+-    }
+-}
+-
+-static void nbd_unset_handlers(NBDClient *client)
+-{
+-    if (client->exp && client->exp->ctx) {
+-        aio_set_fd_handler(client->exp->ctx, client->sioc->fd, true, NULL,
+-                           NULL, NULL, NULL);
+-    }
+-}
+-
+-static void nbd_update_can_read(NBDClient *client)
+-{
+-    bool can_read = client->recv_coroutine ||
+-                    client->nb_requests < MAX_NBD_REQUESTS;
+-
+-    if (can_read != client->can_read) {
+-        client->can_read = can_read;
+-        nbd_set_handlers(client);
+-
+-        /* There is no need to invoke aio_notify(), since aio_set_fd_handler()
+-         * in nbd_set_handlers() will have taken care of that */
++    if (!client->recv_coroutine && client->nb_requests < MAX_NBD_REQUESTS) {
++        nbd_client_get(client);
++        client->recv_coroutine = qemu_coroutine_create(nbd_trip, client);
++        aio_co_schedule(client->exp->ctx, client->recv_coroutine);
+     }
+ }
+@@ -XXX,XX +XXX,XX @@ static coroutine_fn void nbd_co_client_start(void *opaque)
+         goto out;
+     }
+     qemu_co_mutex_init(&client->send_lock);
+-    nbd_set_handlers(client);
+     if (exp) {
+         QTAILQ_INSERT_TAIL(&exp->clients, client, next);
+     }
++
++    nbd_client_receive_next_request(client);
++
+ out:
+     g_free(data);
+ }
+@@ -XXX,XX +XXX,XX @@ void nbd_client_new(NBDExport *exp,
+     object_ref(OBJECT(client->sioc));
+     client->ioc = QIO_CHANNEL(sioc);
+     object_ref(OBJECT(client->ioc));
+-    client->can_read = true;
+     client->close = close_fn;
+     data->client = client;
+--
+.9.3

-New patch
+[Qemu-devel] [PULL v2 08/24] coroutine-lock: reschedule coroutine on the AioContext it was running on
+From: Paolo Bonzini <pbonzini@redhat.com>
+As a small step towards the introduction of multiqueue, we want
+coroutines to remain on the same AioContext that started them,
+unless they are moved explicitly with e.g. aio_co_schedule.  This patch
+avoids that coroutines switch AioContext when they use a CoMutex.
+For now it does not make much of a difference, because the CoMutex
+is not thread-safe and the AioContext itself is used to protect the
+CoMutex from concurrent access.  However, this is going to change.
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
+Reviewed-by: Fam Zheng <famz@redhat.com>
+Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
+Message-id: 20170213135235.12274-9-pbonzini@redhat.com
+Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
+---
+ util/qemu-coroutine-lock.c | 5 ++---
+ util/trace-events          | 1 -
+files changed, 2 insertions(+), 4 deletions(-)
+diff --git a/util/qemu-coroutine-lock.c b/util/qemu-coroutine-lock.c
+index XXXXXXX..XXXXXXX 100644
+--- a/util/qemu-coroutine-lock.c
++++ b/util/qemu-coroutine-lock.c
+@@ -XXX,XX +XXX,XX @@
+ #include "qemu/coroutine.h"
+ #include "qemu/coroutine_int.h"
+ #include "qemu/queue.h"
++#include "block/aio.h"
+ #include "trace.h"
+ void qemu_co_queue_init(CoQueue *queue)
+@@ -XXX,XX +XXX,XX @@ void qemu_co_queue_run_restart(Coroutine *co)
+ static bool qemu_co_queue_do_restart(CoQueue *queue, bool single)
+ {
+-    Coroutine *self = qemu_coroutine_self();
+     Coroutine *next;
+     if (QSIMPLEQ_EMPTY(&queue->entries)) {
+@@ -XXX,XX +XXX,XX @@ static bool qemu_co_queue_do_restart(CoQueue *queue, bool single)
+     while ((next = QSIMPLEQ_FIRST(&queue->entries)) != NULL) {
+         QSIMPLEQ_REMOVE_HEAD(&queue->entries, co_queue_next);
+-        QSIMPLEQ_INSERT_TAIL(&self->co_queue_wakeup, next, co_queue_next);
+-        trace_qemu_co_queue_next(next);
++        aio_co_wake(next);
+         if (single) {
+             break;
+         }
+diff --git a/util/trace-events b/util/trace-events
+index XXXXXXX..XXXXXXX 100644
+--- a/util/trace-events
++++ b/util/trace-events
+@@ -XXX,XX +XXX,XX @@ qemu_coroutine_terminate(void *co) "self %p"
+ # util/qemu-coroutine-lock.c
+ qemu_co_queue_run_restart(void *co) "co %p"
+-qemu_co_queue_next(void *nxt) "next %p"
+ qemu_co_mutex_lock_entry(void *mutex, void *self) "mutex %p self %p"
+ qemu_co_mutex_lock_return(void *mutex, void *self) "mutex %p self %p"
+ qemu_co_mutex_unlock_entry(void *mutex, void *self) "mutex %p self %p"
+--
+.9.3

-New patch
+[Qemu-devel] [PULL v2 09/24] blkdebug: reschedule coroutine on the AioContext it is running on
+From: Paolo Bonzini <pbonzini@redhat.com>
+Keep the coroutine on the same AioContext.  Without this change,
+there would be a race between yielding the coroutine and reentering it.
+While the race cannot happen now, because the code only runs from a single
+AioContext, this will change with multiqueue support in the block layer.
+While doing the change, replace custom bottom half with aio_co_schedule.
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
+Reviewed-by: Fam Zheng <famz@redhat.com>
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
+Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
+Message-id: 20170213135235.12274-10-pbonzini@redhat.com
+Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
+---
+ block/blkdebug.c | 9 +--------
+file changed, 1 insertion(+), 8 deletions(-)
+diff --git a/block/blkdebug.c b/block/blkdebug.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/blkdebug.c
++++ b/block/blkdebug.c
+@@ -XXX,XX +XXX,XX @@ out:
+     return ret;
+ }
+-static void error_callback_bh(void *opaque)
+-{
+-    Coroutine *co = opaque;
+-    qemu_coroutine_enter(co);
+-}
+-
+ static int inject_error(BlockDriverState *bs, BlkdebugRule *rule)
+ {
+     BDRVBlkdebugState *s = bs->opaque;
+@@ -XXX,XX +XXX,XX @@ static int inject_error(BlockDriverState *bs, BlkdebugRule *rule)
+     }
+     if (!immediately) {
+-        aio_bh_schedule_oneshot(bdrv_get_aio_context(bs), error_callback_bh,
+-                                qemu_coroutine_self());
++        aio_co_schedule(qemu_get_current_aio_context(), qemu_coroutine_self());
+         qemu_coroutine_yield();
+     }
+--
+.9.3

-New patch
+[Qemu-devel] [PULL v2 10/24] qed: introduce qed_aio_start_io and qed_aio_next_io_cb
+From: Paolo Bonzini <pbonzini@redhat.com>
+qed_aio_start_io and qed_aio_next_io will not have to acquire/release
+the AioContext, while qed_aio_next_io_cb will.  Split the functionality
+and gain a little type-safety in the process.
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
+Reviewed-by: Fam Zheng <famz@redhat.com>
+Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
+Message-id: 20170213135235.12274-11-pbonzini@redhat.com
+Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
+---
+ block/qed.c | 39 +++++++++++++++++++++++++--------------
+file changed, 25 insertions(+), 14 deletions(-)
+diff --git a/block/qed.c b/block/qed.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/qed.c
++++ b/block/qed.c
+@@ -XXX,XX +XXX,XX @@ static CachedL2Table *qed_new_l2_table(BDRVQEDState *s)
+     return l2_table;
+ }
+-static void qed_aio_next_io(void *opaque, int ret);
++static void qed_aio_next_io(QEDAIOCB *acb, int ret);
++
++static void qed_aio_start_io(QEDAIOCB *acb)
++{
++    qed_aio_next_io(acb, 0);
++}
++
++static void qed_aio_next_io_cb(void *opaque, int ret)
++{
++    QEDAIOCB *acb = opaque;
++
++    qed_aio_next_io(acb, ret);
++}
+ static void qed_plug_allocating_write_reqs(BDRVQEDState *s)
+ {
+@@ -XXX,XX +XXX,XX @@ static void qed_unplug_allocating_write_reqs(BDRVQEDState *s)
+     acb = QSIMPLEQ_FIRST(&s->allocating_write_reqs);
+     if (acb) {
+-        qed_aio_next_io(acb, 0);
++        qed_aio_start_io(acb);
+     }
+ }
+@@ -XXX,XX +XXX,XX @@ static void qed_aio_complete(QEDAIOCB *acb, int ret)
+         QSIMPLEQ_REMOVE_HEAD(&s->allocating_write_reqs, next);
+         acb = QSIMPLEQ_FIRST(&s->allocating_write_reqs);
+         if (acb) {
+-            qed_aio_next_io(acb, 0);
++            qed_aio_start_io(acb);
+         } else if (s->header.features & QED_F_NEED_CHECK) {
+             qed_start_need_check_timer(s);
+         }
+@@ -XXX,XX +XXX,XX @@ static void qed_commit_l2_update(void *opaque, int ret)
+     acb->request.l2_table = qed_find_l2_cache_entry(&s->l2_cache, l2_offset);
+     assert(acb->request.l2_table != NULL);
+-    qed_aio_next_io(opaque, ret);
++    qed_aio_next_io(acb, ret);
+ }
+ /**
+@@ -XXX,XX +XXX,XX @@ static void qed_aio_write_l2_update(QEDAIOCB *acb, int ret, uint64_t offset)
+     if (need_alloc) {
+         /* Write out the whole new L2 table */
+         qed_write_l2_table(s, &acb->request, 0, s->table_nelems, true,
+-                            qed_aio_write_l1_update, acb);
++                           qed_aio_write_l1_update, acb);
+     } else {
+         /* Write out only the updated part of the L2 table */
+         qed_write_l2_table(s, &acb->request, index, acb->cur_nclusters, false,
+-                            qed_aio_next_io, acb);
++                           qed_aio_next_io_cb, acb);
+     }
+     return;
+@@ -XXX,XX +XXX,XX @@ static void qed_aio_write_main(void *opaque, int ret)
+     }
+     if (acb->find_cluster_ret == QED_CLUSTER_FOUND) {
+-        next_fn = qed_aio_next_io;
++        next_fn = qed_aio_next_io_cb;
+     } else {
+         if (s->bs->backing) {
+             next_fn = qed_aio_write_flush_before_l2_update;
+@@ -XXX,XX +XXX,XX @@ static void qed_aio_write_alloc(QEDAIOCB *acb, size_t len)
+     if (acb->flags & QED_AIOCB_ZERO) {
+         /* Skip ahead if the clusters are already zero */
+         if (acb->find_cluster_ret == QED_CLUSTER_ZERO) {
+-            qed_aio_next_io(acb, 0);
++            qed_aio_start_io(acb);
+             return;
+         }
+@@ -XXX,XX +XXX,XX @@ static void qed_aio_read_data(void *opaque, int ret,
+     /* Handle zero cluster and backing file reads */
+     if (ret == QED_CLUSTER_ZERO) {
+         qemu_iovec_memset(&acb->cur_qiov, 0, 0, acb->cur_qiov.size);
+-        qed_aio_next_io(acb, 0);
++        qed_aio_start_io(acb);
+         return;
+     } else if (ret != QED_CLUSTER_FOUND) {
+         qed_read_backing_file(s, acb->cur_pos, &acb->cur_qiov,
+-                              &acb->backing_qiov, qed_aio_next_io, acb);
++                              &acb->backing_qiov, qed_aio_next_io_cb, acb);
+         return;
+     }
+     BLKDBG_EVENT(bs->file, BLKDBG_READ_AIO);
+     bdrv_aio_readv(bs->file, offset / BDRV_SECTOR_SIZE,
+                    &acb->cur_qiov, acb->cur_qiov.size / BDRV_SECTOR_SIZE,
+-                   qed_aio_next_io, acb);
++                   qed_aio_next_io_cb, acb);
+     return;
+ err:
+@@ -XXX,XX +XXX,XX @@ err:
+ /**
+  * Begin next I/O or complete the request
+  */
+-static void qed_aio_next_io(void *opaque, int ret)
++static void qed_aio_next_io(QEDAIOCB *acb, int ret)
+ {
+-    QEDAIOCB *acb = opaque;
+     BDRVQEDState *s = acb_to_s(acb);
+     QEDFindClusterFunc *io_fn = (acb->flags & QED_AIOCB_WRITE) ?
+                                 qed_aio_write_data : qed_aio_read_data;
+@@ -XXX,XX +XXX,XX @@ static BlockAIOCB *qed_aio_setup(BlockDriverState *bs,
+     qemu_iovec_init(&acb->cur_qiov, qiov->niov);
+     /* Start request */
+-    qed_aio_next_io(acb, 0);
++    qed_aio_start_io(acb);
+     return &acb->common;
+ }
+--
+.9.3

-New patch
+[Qemu-devel] [PULL v2 11/24] aio: push aio_context_acquire/release down to dispatching
+From: Paolo Bonzini <pbonzini@redhat.com>
+The AioContext data structures are now protected by list_lock and/or
+they are walked with FOREACH_RCU primitives.  There is no need anymore
+to acquire the AioContext for the entire duration of aio_dispatch.
+Instead, just acquire it before and after invoking the callbacks.
+The next step is then to push it further down.
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
+Reviewed-by: Fam Zheng <famz@redhat.com>
+Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
+Message-id: 20170213135235.12274-12-pbonzini@redhat.com
+Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
+---
+ util/aio-posix.c | 25 +++++++++++--------------
+ util/aio-win32.c | 15 +++++++--------
+ util/async.c     |  2 ++
+files changed, 20 insertions(+), 22 deletions(-)
+diff --git a/util/aio-posix.c b/util/aio-posix.c
+index XXXXXXX..XXXXXXX 100644
+--- a/util/aio-posix.c
++++ b/util/aio-posix.c
+@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
+             (revents & (G_IO_IN | G_IO_HUP | G_IO_ERR)) &&
+             aio_node_check(ctx, node->is_external) &&
+             node->io_read) {
++            aio_context_acquire(ctx);
+             node->io_read(node->opaque);
++            aio_context_release(ctx);
+             /* aio_notify() does not count as progress */
+             if (node->opaque != &ctx->notifier) {
+@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
+             (revents & (G_IO_OUT | G_IO_ERR)) &&
+             aio_node_check(ctx, node->is_external) &&
+             node->io_write) {
++            aio_context_acquire(ctx);
+             node->io_write(node->opaque);
++            aio_context_release(ctx);
+             progress = true;
+         }
+@@ -XXX,XX +XXX,XX @@ bool aio_dispatch(AioContext *ctx, bool dispatch_fds)
+     }
+     /* Run our timers */
++    aio_context_acquire(ctx);
+     progress |= timerlistgroup_run_timers(&ctx->tlg);
++    aio_context_release(ctx);
+     return progress;
+ }
+@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
+     int64_t timeout;
+     int64_t start = 0;
+-    aio_context_acquire(ctx);
+-    progress = false;
+-
+     /* aio_notify can avoid the expensive event_notifier_set if
+      * everything (file descriptors, bottom halves, timers) will
+      * be re-evaluated before the next blocking poll().  This is
+@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
+         start = qemu_clock_get_ns(QEMU_CLOCK_REALTIME);
+     }
+-    if (try_poll_mode(ctx, blocking)) {
+-        progress = true;
+-    } else {
++    aio_context_acquire(ctx);
++    progress = try_poll_mode(ctx, blocking);
++    aio_context_release(ctx);
++
++    if (!progress) {
+         assert(npfd == 0);
+         /* fill pollfds */
+@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
+         timeout = blocking ? aio_compute_timeout(ctx) : 0;
+         /* wait until next event */
+-        if (timeout) {
+-            aio_context_release(ctx);
+-        }
+         if (aio_epoll_check_poll(ctx, pollfds, npfd, timeout)) {
+             AioHandler epoll_handler;
+@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
+         } else  {
+             ret = qemu_poll_ns(pollfds, npfd, timeout);
+         }
+-        if (timeout) {
+-            aio_context_acquire(ctx);
+-        }
+     }
+     if (blocking) {
+@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
+         progress = true;
+     }
+-    aio_context_release(ctx);
+-
+     return progress;
+ }
+diff --git a/util/aio-win32.c b/util/aio-win32.c
+index XXXXXXX..XXXXXXX 100644
+--- a/util/aio-win32.c
++++ b/util/aio-win32.c
+@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx, HANDLE event)
+             (revents || event_notifier_get_handle(node->e) == event) &&
+             node->io_notify) {
+             node->pfd.revents = 0;
++            aio_context_acquire(ctx);
+             node->io_notify(node->e);
++            aio_context_release(ctx);
+             /* aio_notify() does not count as progress */
+             if (node->e != &ctx->notifier) {
+@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx, HANDLE event)
+             (node->io_read || node->io_write)) {
+             node->pfd.revents = 0;
+             if ((revents & G_IO_IN) && node->io_read) {
++                aio_context_acquire(ctx);
+                 node->io_read(node->opaque);
++                aio_context_release(ctx);
+                 progress = true;
+             }
+             if ((revents & G_IO_OUT) && node->io_write) {
++                aio_context_acquire(ctx);
+                 node->io_write(node->opaque);
++                aio_context_release(ctx);
+                 progress = true;
+             }
+@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
+     int count;
+     int timeout;
+-    aio_context_acquire(ctx);
+     progress = false;
+     /* aio_notify can avoid the expensive event_notifier_set if
+@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
+         timeout = blocking && !have_select_revents
+             ? qemu_timeout_ns_to_ms(aio_compute_timeout(ctx)) : 0;
+-        if (timeout) {
+-            aio_context_release(ctx);
+-        }
+         ret = WaitForMultipleObjects(count, events, FALSE, timeout);
+         if (blocking) {
+             assert(first);
+             atomic_sub(&ctx->notify_me, 2);
+         }
+-        if (timeout) {
+-            aio_context_acquire(ctx);
+-        }
+         if (first) {
+             aio_notify_accept(ctx);
+@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
+         progress |= aio_dispatch_handlers(ctx, event);
+     } while (count > 0);
++    aio_context_acquire(ctx);
+     progress |= timerlistgroup_run_timers(&ctx->tlg);
+-
+     aio_context_release(ctx);
+     return progress;
+ }
+diff --git a/util/async.c b/util/async.c
+index XXXXXXX..XXXXXXX 100644
+--- a/util/async.c
++++ b/util/async.c
+@@ -XXX,XX +XXX,XX @@ int aio_bh_poll(AioContext *ctx)
+                 ret = 1;
+             }
+             bh->idle = 0;
++            aio_context_acquire(ctx);
+             aio_bh_call(bh);
++            aio_context_release(ctx);
+         }
+         if (bh->deleted) {
+             deleted = true;
+--
+.9.3

-New patch
+[Qemu-devel] [PULL v2 12/24] block: explicitly acquire aiocontext in timers that need it
+From: Paolo Bonzini <pbonzini@redhat.com>
 Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
 Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
 Reviewed-by: Fam Zheng <famz@redhat.com>
 Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
 Message-id: 20170213135235.12274-13-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
  block/qed.h                 |  3 +++
  block/curl.c                |  2 ++
  block/io.c                  |  5 +++++
  block/iscsi.c               |  8 ++++++--
  block/null.c                |  4 ++++
  block/qed.c                 | 12 ++++++++++++
  block/throttle-groups.c     |  2 ++
  util/aio-posix.c            |  2 --
  util/aio-win32.c            |  2 --
  util/qemu-coroutine-sleep.c |  2 +-
 files changed, 35 insertions(+), 7 deletions(-)
 diff --git a/block/qed.h b/block/qed.h
 index XXXXXXX..XXXXXXX 100644
 --- a/block/qed.h
 +++ b/block/qed.h
@@ -XXX,XX +XXX,XX @@ enum {
   */
  typedef void QEDFindClusterFunc(void *opaque, int ret, uint64_t offset, size_t len);
 +void qed_acquire(BDRVQEDState *s);
 +void qed_release(BDRVQEDState *s);
 +
  /**
   * Generic callback for chaining async callbacks
   */
 diff --git a/block/curl.c b/block/curl.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/curl.c
 +++ b/block/curl.c
@@ -XXX,XX +XXX,XX @@ static void curl_multi_timeout_do(void *arg)
          return;
      }
 +    aio_context_acquire(s->aio_context);
      curl_multi_socket_action(s->multi, CURL_SOCKET_TIMEOUT, 0, &running);
      curl_multi_check_completion(s);
 +    aio_context_release(s->aio_context);
  #else
      abort();
  #endif
 diff --git a/block/io.c b/block/io.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/io.c
 +++ b/block/io.c
@@ -XXX,XX +XXX,XX @@ void bdrv_aio_cancel(BlockAIOCB *acb)
          if (acb->aiocb_info->get_aio_context) {
              aio_poll(acb->aiocb_info->get_aio_context(acb), true);
          } else if (acb->bs) {
 +            /* qemu_aio_ref and qemu_aio_unref are not thread-safe, so
 +             * assert that we're not using an I/O thread.  Thread-safe
 +             * code should use bdrv_aio_cancel_async exclusively.
 +             */
 +            assert(bdrv_get_aio_context(acb->bs) == qemu_get_aio_context());
              aio_poll(bdrv_get_aio_context(acb->bs), true);
          } else {
              abort();
 diff --git a/block/iscsi.c b/block/iscsi.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/iscsi.c
 +++ b/block/iscsi.c
@@ -XXX,XX +XXX,XX @@ static void iscsi_retry_timer_expired(void *opaque)
      struct IscsiTask *iTask = opaque;
      iTask->complete = 1;
      if (iTask->co) {
 -        qemu_coroutine_enter(iTask->co);
 +        aio_co_wake(iTask->co);
      }
  }
@@ -XXX,XX +XXX,XX @@ static void iscsi_nop_timed_event(void *opaque)
  {
      IscsiLun *iscsilun = opaque;
 +    aio_context_acquire(iscsilun->aio_context);
      if (iscsi_get_nops_in_flight(iscsilun->iscsi) >= MAX_NOP_FAILURES) {
          error_report("iSCSI: NOP timeout. Reconnecting...");
          iscsilun->request_timed_out = true;
      } else if (iscsi_nop_out_async(iscsilun->iscsi, NULL, NULL, 0, NULL) != 0) {
          error_report("iSCSI: failed to sent NOP-Out. Disabling NOP messages.");
 -        return;
 +        goto out;
      }
      timer_mod(iscsilun->nop_timer, qemu_clock_get_ms(QEMU_CLOCK_REALTIME) + NOP_INTERVAL);
      iscsi_set_events(iscsilun);
 +
 +out:
 +    aio_context_release(iscsilun->aio_context);
  }
  static void iscsi_readcapacity_sync(IscsiLun *iscsilun, Error **errp)
 diff --git a/block/null.c b/block/null.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/null.c
 +++ b/block/null.c
@@ -XXX,XX +XXX,XX @@ static void null_bh_cb(void *opaque)
  static void null_timer_cb(void *opaque)
  {
      NullAIOCB *acb = opaque;
 +    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
 +
 +    aio_context_acquire(ctx);
      acb->common.cb(acb->common.opaque, 0);
 +    aio_context_release(ctx);
      timer_deinit(&acb->timer);
      qemu_aio_unref(acb);
  }
 diff --git a/block/qed.c b/block/qed.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/qed.c
 +++ b/block/qed.c
@@ -XXX,XX +XXX,XX @@ static void qed_need_check_timer_cb(void *opaque)
      trace_qed_need_check_timer_cb(s);
 +    qed_acquire(s);
      qed_plug_allocating_write_reqs(s);
      /* Ensure writes are on disk before clearing flag */
      bdrv_aio_flush(s->bs->file->bs, qed_clear_need_check, s);
 +    qed_release(s);
 +}
 +
 +void qed_acquire(BDRVQEDState *s)
 +{
 +    aio_context_acquire(bdrv_get_aio_context(s->bs));
 +}
 +
 +void qed_release(BDRVQEDState *s)
 +{
 +    aio_context_release(bdrv_get_aio_context(s->bs));
  }
  static void qed_start_need_check_timer(BDRVQEDState *s)
 diff --git a/block/throttle-groups.c b/block/throttle-groups.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/throttle-groups.c
 +++ b/block/throttle-groups.c
@@ -XXX,XX +XXX,XX @@ static void timer_cb(BlockBackend *blk, bool is_write)
      qemu_mutex_unlock(&tg->lock);
      /* Run the request that was waiting for this timer */
 +    aio_context_acquire(blk_get_aio_context(blk));
      empty_queue = !qemu_co_enter_next(&blkp->throttled_reqs[is_write]);
 +    aio_context_release(blk_get_aio_context(blk));
      /* If the request queue was empty then we have to take care of
       * scheduling the next one */
 diff --git a/util/aio-posix.c b/util/aio-posix.c
 index XXXXXXX..XXXXXXX 100644
 --- a/util/aio-posix.c
 +++ b/util/aio-posix.c
@@ -XXX,XX +XXX,XX @@ bool aio_dispatch(AioContext *ctx, bool dispatch_fds)
      }
      /* Run our timers */
 -    aio_context_acquire(ctx);
      progress |= timerlistgroup_run_timers(&ctx->tlg);
 -    aio_context_release(ctx);
      return progress;
  }
 diff --git a/util/aio-win32.c b/util/aio-win32.c
 index XXXXXXX..XXXXXXX 100644
 --- a/util/aio-win32.c
 +++ b/util/aio-win32.c
@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
          progress |= aio_dispatch_handlers(ctx, event);
      } while (count > 0);
 -    aio_context_acquire(ctx);
      progress |= timerlistgroup_run_timers(&ctx->tlg);
 -    aio_context_release(ctx);
      return progress;
  }
 diff --git a/util/qemu-coroutine-sleep.c b/util/qemu-coroutine-sleep.c
 index XXXXXXX..XXXXXXX 100644
 --- a/util/qemu-coroutine-sleep.c
 +++ b/util/qemu-coroutine-sleep.c
@@ -XXX,XX +XXX,XX @@ static void co_sleep_cb(void *opaque)
  {
      CoSleepCB *sleep_cb = opaque;
 -    qemu_coroutine_enter(sleep_cb->co);
 +    aio_co_wake(sleep_cb->co);
  }
  void coroutine_fn co_aio_sleep_ns(AioContext *ctx, QEMUClockType type,
 --
 .9.3

-[PULL 3/5] qcow2: add zstd cluster compression
+[Qemu-devel] [PULL v2 13/24] block: explicitly acquire aiocontext in callbacks that need it
-From: Denis Plotnikov <dplotnikov@virtuozzo.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-zstd significantly reduces cluster compression time.
+This covers both file descriptor callbacks and polling callbacks,
-It provides better compression performance maintaining
+since they execute related code.
 the same level of the compression ratio in comparison with
 zlib, which, at the moment, is the only compression
 method available.
-The performance test results:
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Test compresses and decompresses qemu qcow2 image with just
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-installed rhel-7.6 guest.
+Reviewed-by: Fam Zheng <famz@redhat.com>
-Image cluster size: 64K. Image on disk size: 2.2G
+Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
 Message-id: 20170213135235.12274-14-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
  block/curl.c          | 16 +++++++++++++---
  block/iscsi.c         |  4 ++++
  block/linux-aio.c     |  4 ++++
  block/nfs.c           |  6 ++++++
  block/sheepdog.c      | 29 +++++++++++++++--------------
  block/ssh.c           | 29 +++++++++--------------------
  block/win32-aio.c     | 10 ++++++----
  hw/block/virtio-blk.c |  5 ++++-
  hw/scsi/virtio-scsi.c |  7 +++++++
  util/aio-posix.c      |  7 -------
  util/aio-win32.c      |  6 ------
 files changed, 68 insertions(+), 55 deletions(-)
-The test was conducted with brd disk to reduce the influence
+diff --git a/block/curl.c b/block/curl.c
-of disk subsystem to the test results.
+index XXXXXXX..XXXXXXX 100644
-The results is given in seconds.
+--- a/block/curl.c
++++ b/block/curl.c
-compress cmd:
+@@ -XXX,XX +XXX,XX @@ static void curl_multi_check_completion(BDRVCURLState *s)
-  time ./qemu-img convert -O qcow2 -c -o compression_type=[zlib|zstd]
+     }
-                  src.img [zlib|zstd]_compressed.img
+ }
-decompress cmd
-  time ./qemu-img convert -O qcow2
+-static void curl_multi_do(void *arg)
-                  [zlib|zstd]_compressed.img uncompressed.img
++static void curl_multi_do_locked(CURLState *s)
+ {
-           compression               decompression
+-    CURLState *s = (CURLState *)arg;
-         zlib       zstd           zlib         zstd
+     CURLSocket *socket, *next_socket;
-------------------------------------------------------------
+     int running;
-real     65.5       16.3 (-75 %)    1.9          1.6 (-16 %)
+     int r;
-user     65.0       15.8            5.3          2.5
+@@ -XXX,XX +XXX,XX @@ static void curl_multi_do(void *arg)
-sys       3.3        0.2            2.0          2.0
+     }
+ }
-Both ZLIB and ZSTD gave the same compression ratio: 1.57
-compressed image size in both cases: 1.4G
++static void curl_multi_do(void *arg)
++{
-Signed-off-by: Denis Plotnikov <dplotnikov@virtuozzo.com>
++    CURLState *s = (CURLState *)arg;
-QAPI part:
++
-Acked-by: Markus Armbruster <armbru@redhat.com>
++    aio_context_acquire(s->s->aio_context);
-Message-Id: <20200507082521.29210-4-dplotnikov@virtuozzo.com>
++    curl_multi_do_locked(s);
-Signed-off-by: Max Reitz <mreitz@redhat.com>
++    aio_context_release(s->s->aio_context);
----
++}
- docs/interop/qcow2.txt |   1 +
++
- configure              |   2 +-
+ static void curl_multi_read(void *arg)
- qapi/block-core.json   |   3 +-
+ {
- block/qcow2-threads.c  | 169 +++++++++++++++++++++++++++++++++++++++++
+     CURLState *s = (CURLState *)arg;
- block/qcow2.c          |   7 ++
-files changed, 180 insertions(+), 2 deletions(-)
+-    curl_multi_do(arg);
++    aio_context_acquire(s->s->aio_context);
-diff --git a/docs/interop/qcow2.txt b/docs/interop/qcow2.txt
++    curl_multi_do_locked(s);
-index XXXXXXX..XXXXXXX 100644
+     curl_multi_check_completion(s->s);
---- a/docs/interop/qcow2.txt
++    aio_context_release(s->s->aio_context);
-+++ b/docs/interop/qcow2.txt
+ }
-@@ -XXX,XX +XXX,XX @@ version 2.
+ static void curl_multi_timeout_do(void *arg)
-                     Available compression type values:
+diff --git a/block/iscsi.c b/block/iscsi.c
-: zlib <https://www.zlib.net/>
+index XXXXXXX..XXXXXXX 100644
-+                        1: zstd <http://github.com/facebook/zstd>
+--- a/block/iscsi.c
++++ b/block/iscsi.c
+@@ -XXX,XX +XXX,XX @@ iscsi_process_read(void *arg)
- === Header padding ===
+     IscsiLun *iscsilun = arg;
-diff --git a/configure b/configure
+     struct iscsi_context *iscsi = iscsilun->iscsi;
-index XXXXXXX..XXXXXXX 100755
---- a/configure
++    aio_context_acquire(iscsilun->aio_context);
-+++ b/configure
+     iscsi_service(iscsi, POLLIN);
-@@ -XXX,XX +XXX,XX @@ disabled with --disable-FEATURE, default is enabled if available:
+     iscsi_set_events(iscsilun);
-   lzfse           support of lzfse compression library
++    aio_context_release(iscsilun->aio_context);
-                   (for reading lzfse-compressed dmg images)
+ }
-   zstd            support for zstd compression library
--                  (for migration compression)
+ static void
-+                  (for migration compression and qcow2 cluster compression)
+@@ -XXX,XX +XXX,XX @@ iscsi_process_write(void *arg)
-   seccomp         seccomp support
+     IscsiLun *iscsilun = arg;
-   coroutine-pool  coroutine freelist (better performance)
+     struct iscsi_context *iscsi = iscsilun->iscsi;
-   glusterfs       GlusterFS backend
-diff --git a/qapi/block-core.json b/qapi/block-core.json
++    aio_context_acquire(iscsilun->aio_context);
-index XXXXXXX..XXXXXXX 100644
+     iscsi_service(iscsi, POLLOUT);
---- a/qapi/block-core.json
+     iscsi_set_events(iscsilun);
-+++ b/qapi/block-core.json
++    aio_context_release(iscsilun->aio_context);
-@@ -XXX,XX +XXX,XX @@
+ }
- # Compression type used in qcow2 image file
- #
+ static int64_t sector_lun2qemu(int64_t sector, IscsiLun *iscsilun)
- # @zlib: zlib compression, see <http://zlib.net/>
+diff --git a/block/linux-aio.c b/block/linux-aio.c
-+# @zstd: zstd compression, see <http://github.com/facebook/zstd>
+index XXXXXXX..XXXXXXX 100644
- #
+--- a/block/linux-aio.c
- # Since: 5.1
++++ b/block/linux-aio.c
- ##
+@@ -XXX,XX +XXX,XX @@ static void qemu_laio_completion_cb(EventNotifier *e)
- { 'enum': 'Qcow2CompressionType',
+     LinuxAioState *s = container_of(e, LinuxAioState, e);
--  'data': [ 'zlib' ] }
-+  'data': [ 'zlib', { 'name': 'zstd', 'if': 'defined(CONFIG_ZSTD)' } ] }
+     if (event_notifier_test_and_clear(&s->e)) {
++        aio_context_acquire(s->aio_context);
- ##
+         qemu_laio_process_completions_and_submit(s);
- # @BlockdevCreateOptionsQcow2:
++        aio_context_release(s->aio_context);
-diff --git a/block/qcow2-threads.c b/block/qcow2-threads.c
+     }
-index XXXXXXX..XXXXXXX 100644
+ }
---- a/block/qcow2-threads.c
-+++ b/block/qcow2-threads.c
+@@ -XXX,XX +XXX,XX @@ static bool qemu_laio_poll_cb(void *opaque)
-@@ -XXX,XX +XXX,XX @@
+         return false;
- #define ZLIB_CONST
+     }
- #include <zlib.h>
++    aio_context_acquire(s->aio_context);
-+#ifdef CONFIG_ZSTD
+     qemu_laio_process_completions_and_submit(s);
-+#include <zstd.h>
++    aio_context_release(s->aio_context);
-+#include <zstd_errors.h>
+     return true;
-+#endif
+ }
-+
- #include "qcow2.h"
+diff --git a/block/nfs.c b/block/nfs.c
- #include "block/thread-pool.h"
+index XXXXXXX..XXXXXXX 100644
- #include "crypto.h"
+--- a/block/nfs.c
-@@ -XXX,XX +XXX,XX @@ static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size,
++++ b/block/nfs.c
@@ -XXX,XX +XXX,XX @@ static void nfs_set_events(NFSClient *client)
  static void nfs_process_read(void *arg)
  {
      NFSClient *client = arg;
 +
 +    aio_context_acquire(client->aio_context);
      nfs_service(client->context, POLLIN);
      nfs_set_events(client);
 +    aio_context_release(client->aio_context);
  }
  static void nfs_process_write(void *arg)
  {
      NFSClient *client = arg;
 +
 +    aio_context_acquire(client->aio_context);
      nfs_service(client->context, POLLOUT);
      nfs_set_events(client);
 +    aio_context_release(client->aio_context);
  }
  static void nfs_co_init_task(BlockDriverState *bs, NFSRPC *task)
 diff --git a/block/sheepdog.c b/block/sheepdog.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/sheepdog.c
 +++ b/block/sheepdog.c
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int send_co_req(int sockfd, SheepdogReq *hdr, void *data,
      return ret;
  }
-+#ifdef CONFIG_ZSTD
+-static void restart_co_req(void *opaque)
-+
+-{
-+/*
+-    Coroutine *co = opaque;
-+ * qcow2_zstd_compress()
+-
-+ *
+-    qemu_coroutine_enter(co);
-+ * Compress @src_size bytes of data using zstd compression method
+-}
-+ *
+-
-+ * @dest - destination buffer, @dest_size bytes
+ typedef struct SheepdogReqCo {
-+ * @src - source buffer, @src_size bytes
+     int sockfd;
-+ *
+     BlockDriverState *bs;
-+ * Returns: compressed size on success
+@@ -XXX,XX +XXX,XX @@ typedef struct SheepdogReqCo {
-+ *          -ENOMEM destination buffer is not enough to store compressed data
+     unsigned int *rlen;
-+ *          -EIO    on any other error
+     int ret;
      bool finished;
 +    Coroutine *co;
  } SheepdogReqCo;
 +static void restart_co_req(void *opaque)
 +{
 +    SheepdogReqCo *srco = opaque;
 +
 +    aio_co_wake(srco->co);
 +}
 +
  static coroutine_fn void do_co_req(void *opaque)
  {
      int ret;
 -    Coroutine *co;
      SheepdogReqCo *srco = opaque;
      int sockfd = srco->sockfd;
      SheepdogReq *hdr = srco->hdr;
@@ -XXX,XX +XXX,XX @@ static coroutine_fn void do_co_req(void *opaque)
      unsigned int *wlen = srco->wlen;
      unsigned int *rlen = srco->rlen;
 -    co = qemu_coroutine_self();
 +    srco->co = qemu_coroutine_self();
      aio_set_fd_handler(srco->aio_context, sockfd, false,
 -                       NULL, restart_co_req, NULL, co);
 +                       NULL, restart_co_req, NULL, srco);
      ret = send_co_req(sockfd, hdr, data, wlen);
      if (ret < 0) {
@@ -XXX,XX +XXX,XX @@ static coroutine_fn void do_co_req(void *opaque)
      }
      aio_set_fd_handler(srco->aio_context, sockfd, false,
 -                       restart_co_req, NULL, NULL, co);
 +                       restart_co_req, NULL, NULL, srco);
      ret = qemu_co_recv(sockfd, hdr, sizeof(*hdr));
      if (ret != sizeof(*hdr)) {
@@ -XXX,XX +XXX,XX @@ out:
      aio_set_fd_handler(srco->aio_context, sockfd, false,
                         NULL, NULL, NULL, NULL);
 +    srco->co = NULL;
      srco->ret = ret;
      srco->finished = true;
      if (srco->bs) {
@@ -XXX,XX +XXX,XX @@ static void coroutine_fn aio_read_response(void *opaque)
           * We've finished all requests which belong to the AIOCB, so
           * we can switch back to sd_co_readv/writev now.
           */
 -        qemu_coroutine_enter(acb->coroutine);
 +        aio_co_wake(acb->coroutine);
      }
      return;
@@ -XXX,XX +XXX,XX @@ static void co_read_response(void *opaque)
          s->co_recv = qemu_coroutine_create(aio_read_response, opaque);
      }
 -    qemu_coroutine_enter(s->co_recv);
 +    aio_co_wake(s->co_recv);
  }
  static void co_write_request(void *opaque)
  {
      BDRVSheepdogState *s = opaque;
 -    qemu_coroutine_enter(s->co_send);
 +    aio_co_wake(s->co_send);
  }
  /*
 diff --git a/block/ssh.c b/block/ssh.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/ssh.c
 +++ b/block/ssh.c
@@ -XXX,XX +XXX,XX @@ static void restart_coroutine(void *opaque)
      DPRINTF("co=%p", co);
 -    qemu_coroutine_enter(co);
 +    aio_co_wake(co);
  }
 -static coroutine_fn void set_fd_handler(BDRVSSHState *s, BlockDriverState *bs)
 +/* A non-blocking call returned EAGAIN, so yield, ensuring the
 + * handlers are set up so that we'll be rescheduled when there is an
 + * interesting event on the socket.
 + */
-+static ssize_t qcow2_zstd_compress(void *dest, size_t dest_size,
++static coroutine_fn void co_yield(BDRVSSHState *s, BlockDriverState *bs)
-+                                   const void *src, size_t src_size)
+ {
-+{
+     int r;
-+    ssize_t ret;
+     IOHandler *rd_handler = NULL, *wr_handler = NULL;
-+    size_t zstd_ret;
+@@ -XXX,XX +XXX,XX @@ static coroutine_fn void set_fd_handler(BDRVSSHState *s, BlockDriverState *bs)
-+    ZSTD_outBuffer output = {
-+        .dst = dest,
+     aio_set_fd_handler(bdrv_get_aio_context(bs), s->sock,
-+        .size = dest_size,
+                        false, rd_handler, wr_handler, NULL, co);
-+        .pos = 0
+-}
-+    };
+-
-+    ZSTD_inBuffer input = {
+-static coroutine_fn void clear_fd_handler(BDRVSSHState *s,
-+        .src = src,
+-                                          BlockDriverState *bs)
-+        .size = src_size,
+-{
-+        .pos = 0
+-    DPRINTF("s->sock=%d", s->sock);
-+    };
+-    aio_set_fd_handler(bdrv_get_aio_context(bs), s->sock,
-+    ZSTD_CCtx *cctx = ZSTD_createCCtx();
+-                       false, NULL, NULL, NULL, NULL);
-+
+-}
-+    if (!cctx) {
+-
-+        return -EIO;
+-/* A non-blocking call returned EAGAIN, so yield, ensuring the
-+    }
+- * handlers are set up so that we'll be rescheduled when there is an
-+    /*
+- * interesting event on the socket.
-+     * Use the zstd streamed interface for symmetry with decompression,
+- */
-+     * where streaming is essential since we don't record the exact
+-static coroutine_fn void co_yield(BDRVSSHState *s, BlockDriverState *bs)
-+     * compressed size.
+-{
-+     *
+-    set_fd_handler(s, bs);
-+     * ZSTD_compressStream2() tries to compress everything it could
+     qemu_coroutine_yield();
-+     * with a single call. Although, ZSTD docs says that:
+-    clear_fd_handler(s, bs);
-+     * "You must continue calling ZSTD_compressStream2() with ZSTD_e_end
++    DPRINTF("s->sock=%d - back", s->sock);
-+     * until it returns 0, at which point you are free to start a new frame",
++    aio_set_fd_handler(bdrv_get_aio_context(bs), s->sock, false,
-+     * in out tests we saw the only case when it returned with >0 -
++                       NULL, NULL, NULL, NULL);
-+     * when the output buffer was too small. In that case,
+ }
-+     * ZSTD_compressStream2() expects a bigger buffer on the next call.
-+     * We can't provide a bigger buffer because we are limited with dest_size
+ /* SFTP has a function `libssh2_sftp_seek64' which seeks to a position
-+     * which we pass to the ZSTD_compressStream2() at once.
+diff --git a/block/win32-aio.c b/block/win32-aio.c
-+     * So, we don't need any loops and just abort the compression when we
+index XXXXXXX..XXXXXXX 100644
-+     * don't get 0 result on the first call.
+--- a/block/win32-aio.c
-+     */
++++ b/block/win32-aio.c
-+    zstd_ret = ZSTD_compressStream2(cctx, &output, &input, ZSTD_e_end);
+@@ -XXX,XX +XXX,XX @@ struct QEMUWin32AIOState {
-+
+     HANDLE hIOCP;
-+    if (zstd_ret) {
+     EventNotifier e;
-+        if (zstd_ret > output.size - output.pos) {
+     int count;
-+            ret = -ENOMEM;
+-    bool is_aio_context_attached;
-+        } else {
++    AioContext *aio_ctx;
-+            ret = -EIO;
+ };
-+        }
-+        goto out;
+ typedef struct QEMUWin32AIOCB {
-+    }
+@@ -XXX,XX +XXX,XX @@ static void win32_aio_process_completion(QEMUWin32AIOState *s,
-+
+     }
-+    /* make sure that zstd didn't overflow the dest buffer */
-+    assert(output.pos <= dest_size);
-+    ret = output.pos;
++    aio_context_acquire(s->aio_ctx);
-+out:
+     waiocb->common.cb(waiocb->common.opaque, ret);
-+    ZSTD_freeCCtx(cctx);
++    aio_context_release(s->aio_ctx);
-+    return ret;
+     qemu_aio_unref(waiocb);
-+}
+ }
-+
-+/*
+@@ -XXX,XX +XXX,XX @@ void win32_aio_detach_aio_context(QEMUWin32AIOState *aio,
-+ * qcow2_zstd_decompress()
+                                   AioContext *old_context)
-+ *
+ {
-+ * Decompress some data (not more than @src_size bytes) to produce exactly
+     aio_set_event_notifier(old_context, &aio->e, false, NULL, NULL);
-+ * @dest_size bytes using zstd compression method
+-    aio->is_aio_context_attached = false;
-+ *
++    aio->aio_ctx = NULL;
-+ * @dest - destination buffer, @dest_size bytes
+ }
-+ * @src - source buffer, @src_size bytes
-+ *
+ void win32_aio_attach_aio_context(QEMUWin32AIOState *aio,
-+ * Returns: 0 on success
+                                   AioContext *new_context)
-+ *          -EIO on any error
+ {
-+ */
+-    aio->is_aio_context_attached = true;
-+static ssize_t qcow2_zstd_decompress(void *dest, size_t dest_size,
++    aio->aio_ctx = new_context;
-+                                     const void *src, size_t src_size)
+     aio_set_event_notifier(new_context, &aio->e, false,
-+{
+                            win32_aio_completion_cb, NULL);
-+    size_t zstd_ret = 0;
+ }
-+    ssize_t ret = 0;
+@@ -XXX,XX +XXX,XX @@ out_free_state:
-+    ZSTD_outBuffer output = {
-+        .dst = dest,
+ void win32_aio_cleanup(QEMUWin32AIOState *aio)
-+        .size = dest_size,
+ {
-+        .pos = 0
+-    assert(!aio->is_aio_context_attached);
-+    };
++    assert(!aio->aio_ctx);
-+    ZSTD_inBuffer input = {
+     CloseHandle(aio->hIOCP);
-+        .src = src,
+     event_notifier_cleanup(&aio->e);
-+        .size = src_size,
+     g_free(aio);
-+        .pos = 0
+diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c
-+    };
+index XXXXXXX..XXXXXXX 100644
-+    ZSTD_DCtx *dctx = ZSTD_createDCtx();
+--- a/hw/block/virtio-blk.c
-+
++++ b/hw/block/virtio-blk.c
-+    if (!dctx) {
+@@ -XXX,XX +XXX,XX @@ static void virtio_blk_ioctl_complete(void *opaque, int status)
-+        return -EIO;
+ {
-+    }
+     VirtIOBlockIoctlReq *ioctl_req = opaque;
-+
+     VirtIOBlockReq *req = ioctl_req->req;
-+    /*
+-    VirtIODevice *vdev = VIRTIO_DEVICE(req->dev);
-+     * The compressed stream from the input buffer may consist of more
++    VirtIOBlock *s = req->dev;
-+     * than one zstd frame. So we iterate until we get a fully
++    VirtIODevice *vdev = VIRTIO_DEVICE(s);
-+     * uncompressed cluster.
+     struct virtio_scsi_inhdr *scsi;
-+     * From zstd docs related to ZSTD_decompressStream:
+     struct sg_io_hdr *hdr;
-+     * "return : 0 when a frame is completely decoded and fully flushed"
-+     * We suppose that this means: each time ZSTD_decompressStream reads
+@@ -XXX,XX +XXX,XX @@ bool virtio_blk_handle_vq(VirtIOBlock *s, VirtQueue *vq)
-+     * only ONE full frame and returns 0 if and only if that frame
+     MultiReqBuffer mrb = {};
-+     * is completely decoded and flushed. Only after returning 0,
+     bool progress = false;
-+     * ZSTD_decompressStream reads another ONE full frame.
-+     */
++    aio_context_acquire(blk_get_aio_context(s->blk));
-+    while (output.pos < output.size) {
+     blk_io_plug(s->blk);
-+        size_t last_in_pos = input.pos;
-+        size_t last_out_pos = output.pos;
+     do {
-+        zstd_ret = ZSTD_decompressStream(dctx, &output, &input);
+@@ -XXX,XX +XXX,XX @@ bool virtio_blk_handle_vq(VirtIOBlock *s, VirtQueue *vq)
-+
+     }
-+        if (ZSTD_isError(zstd_ret)) {
-+            ret = -EIO;
+     blk_io_unplug(s->blk);
-+            break;
++    aio_context_release(blk_get_aio_context(s->blk));
-+        }
+     return progress;
-+
+ }
-+        /*
-+         * The ZSTD manual is vague about what to do if it reads
+diff --git a/hw/scsi/virtio-scsi.c b/hw/scsi/virtio-scsi.c
-+         * the buffer partially, and we don't want to get stuck
+index XXXXXXX..XXXXXXX 100644
-+         * in an infinite loop where ZSTD_decompressStream
+--- a/hw/scsi/virtio-scsi.c
-+         * returns > 0 waiting for another input chunk. So, we add
++++ b/hw/scsi/virtio-scsi.c
-+         * a check which ensures that the loop makes some progress
+@@ -XXX,XX +XXX,XX @@ bool virtio_scsi_handle_ctrl_vq(VirtIOSCSI *s, VirtQueue *vq)
-+         * on each step.
+     VirtIOSCSIReq *req;
-+         */
+     bool progress = false;
-+        if (last_in_pos >= input.pos &&
-+            last_out_pos >= output.pos) {
++    virtio_scsi_acquire(s);
-+            ret = -EIO;
+     while ((req = virtio_scsi_pop_req(s, vq))) {
-+            break;
+         progress = true;
-+        }
+         virtio_scsi_handle_ctrl_req(s, req);
-+    }
+     }
-+    /*
++    virtio_scsi_release(s);
-+     * Make sure that we have the frame fully flushed here
+     return progress;
-+     * if not, we somehow managed to get uncompressed cluster
+ }
-+     * greater then the cluster size, possibly because of its
-+     * damage.
+@@ -XXX,XX +XXX,XX @@ bool virtio_scsi_handle_cmd_vq(VirtIOSCSI *s, VirtQueue *vq)
-+     */
-+    if (zstd_ret > 0) {
+     QTAILQ_HEAD(, VirtIOSCSIReq) reqs = QTAILQ_HEAD_INITIALIZER(reqs);
-+        ret = -EIO;
-+    }
++    virtio_scsi_acquire(s);
-+
+     do {
-+    ZSTD_freeDCtx(dctx);
+         virtio_queue_set_notification(vq, 0);
-+    assert(ret == 0 || ret == -EIO);
-+    return ret;
+@@ -XXX,XX +XXX,XX @@ bool virtio_scsi_handle_cmd_vq(VirtIOSCSI *s, VirtQueue *vq)
-+}
+     QTAILQ_FOREACH_SAFE(req, &reqs, next, next) {
-+#endif
+         virtio_scsi_handle_cmd_req_submit(s, req);
-+
+     }
- static int qcow2_compress_pool_func(void *opaque)
++    virtio_scsi_release(s);
- {
+     return progress;
-     Qcow2CompressData *data = opaque;
+ }
-@@ -XXX,XX +XXX,XX @@ qcow2_co_compress(BlockDriverState *bs, void *dest, size_t dest_size,
-         fn = qcow2_zlib_compress;
+@@ -XXX,XX +XXX,XX @@ out:
-         break;
+ bool virtio_scsi_handle_event_vq(VirtIOSCSI *s, VirtQueue *vq)
-+#ifdef CONFIG_ZSTD
+ {
-+    case QCOW2_COMPRESSION_TYPE_ZSTD:
++    virtio_scsi_acquire(s);
-+        fn = qcow2_zstd_compress;
+     if (s->events_dropped) {
-+        break;
+         virtio_scsi_push_event(s, NULL, VIRTIO_SCSI_T_NO_EVENT, 0);
-+#endif
++        virtio_scsi_release(s);
-     default:
+         return true;
-         abort();
+     }
-     }
++    virtio_scsi_release(s);
-@@ -XXX,XX +XXX,XX @@ qcow2_co_decompress(BlockDriverState *bs, void *dest, size_t dest_size,
+     return false;
-         fn = qcow2_zlib_decompress;
+ }
-         break;
+diff --git a/util/aio-posix.c b/util/aio-posix.c
-+#ifdef CONFIG_ZSTD
+index XXXXXXX..XXXXXXX 100644
-+    case QCOW2_COMPRESSION_TYPE_ZSTD:
+--- a/util/aio-posix.c
-+        fn = qcow2_zstd_decompress;
++++ b/util/aio-posix.c
-+        break;
+@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
-+#endif
+             (revents & (G_IO_IN | G_IO_HUP | G_IO_ERR)) &&
-     default:
+             aio_node_check(ctx, node->is_external) &&
-         abort();
+             node->io_read) {
-     }
+-            aio_context_acquire(ctx);
-diff --git a/block/qcow2.c b/block/qcow2.c
+             node->io_read(node->opaque);
-index XXXXXXX..XXXXXXX 100644
+-            aio_context_release(ctx);
---- a/block/qcow2.c
-+++ b/block/qcow2.c
+             /* aio_notify() does not count as progress */
-@@ -XXX,XX +XXX,XX @@ static int validate_compression_type(BDRVQcow2State *s, Error **errp)
+             if (node->opaque != &ctx->notifier) {
- {
+@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
-     switch (s->compression_type) {
+             (revents & (G_IO_OUT | G_IO_ERR)) &&
-     case QCOW2_COMPRESSION_TYPE_ZLIB:
+             aio_node_check(ctx, node->is_external) &&
-+#ifdef CONFIG_ZSTD
+             node->io_write) {
-+    case QCOW2_COMPRESSION_TYPE_ZSTD:
+-            aio_context_acquire(ctx);
-+#endif
+             node->io_write(node->opaque);
-         break;
+-            aio_context_release(ctx);
+             progress = true;
      default:
@@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp)
          }
-         switch (qcow2_opts->compression_type) {
+@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
-+#ifdef CONFIG_ZSTD
+         start = qemu_clock_get_ns(QEMU_CLOCK_REALTIME);
-+        case QCOW2_COMPRESSION_TYPE_ZSTD:
+     }
-+            break;
-+#endif
+-    aio_context_acquire(ctx);
-         default:
+     progress = try_poll_mode(ctx, blocking);
-             error_setg(errp, "Unknown compression type");
+-    aio_context_release(ctx);
-             goto out;
+-
      if (!progress) {
          assert(npfd == 0);
 diff --git a/util/aio-win32.c b/util/aio-win32.c
 index XXXXXXX..XXXXXXX 100644
 --- a/util/aio-win32.c
 +++ b/util/aio-win32.c
@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx, HANDLE event)
              (revents || event_notifier_get_handle(node->e) == event) &&
              node->io_notify) {
              node->pfd.revents = 0;
 -            aio_context_acquire(ctx);
              node->io_notify(node->e);
 -            aio_context_release(ctx);
              /* aio_notify() does not count as progress */
              if (node->e != &ctx->notifier) {
@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx, HANDLE event)
              (node->io_read || node->io_write)) {
              node->pfd.revents = 0;
              if ((revents & G_IO_IN) && node->io_read) {
 -                aio_context_acquire(ctx);
                  node->io_read(node->opaque);
 -                aio_context_release(ctx);
                  progress = true;
              }
              if ((revents & G_IO_OUT) && node->io_write) {
 -                aio_context_acquire(ctx);
                  node->io_write(node->opaque);
 -                aio_context_release(ctx);
                  progress = true;
              }
 --
-.26.2
+.9.3

-New patch
+[Qemu-devel] [PULL v2 14/24] block: explicitly acquire aiocontext in bottom halves that need it
+From: Paolo Bonzini <pbonzini@redhat.com>
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
+Reviewed-by: Fam Zheng <famz@redhat.com>
+Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
+Message-id: 20170213135235.12274-15-pbonzini@redhat.com
+Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
+---
+ block/archipelago.c   |  3 +++
+ block/blkreplay.c     |  2 +-
+ block/block-backend.c |  6 ++++++
+ block/curl.c          | 26 ++++++++++++++++++--------
+ block/gluster.c       |  9 +--------
+ block/io.c            |  6 +++++-
+ block/iscsi.c         |  6 +++++-
+ block/linux-aio.c     | 15 +++++++++------
+ block/nfs.c           |  3 ++-
+ block/null.c          |  4 ++++
+ block/qed.c           |  3 +++
+ block/rbd.c           |  4 ++++
+ dma-helpers.c         |  2 ++
+ hw/block/virtio-blk.c |  2 ++
+ hw/scsi/scsi-bus.c    |  2 ++
+ util/async.c          |  4 ++--
+ util/thread-pool.c    |  2 ++
+files changed, 71 insertions(+), 28 deletions(-)
+diff --git a/block/archipelago.c b/block/archipelago.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/archipelago.c
++++ b/block/archipelago.c
+@@ -XXX,XX +XXX,XX @@ static void qemu_archipelago_complete_aio(void *opaque)
+ {
+     AIORequestData *reqdata = (AIORequestData *) opaque;
+     ArchipelagoAIOCB *aio_cb = (ArchipelagoAIOCB *) reqdata->aio_cb;
++    AioContext *ctx = bdrv_get_aio_context(aio_cb->common.bs);
++    aio_context_acquire(ctx);
+     aio_cb->common.cb(aio_cb->common.opaque, aio_cb->ret);
++    aio_context_release(ctx);
+     aio_cb->status = 0;
+     qemu_aio_unref(aio_cb);
+diff --git a/block/blkreplay.c b/block/blkreplay.c
+index XXXXXXX..XXXXXXX 100755
+--- a/block/blkreplay.c
++++ b/block/blkreplay.c
+@@ -XXX,XX +XXX,XX @@ static int64_t blkreplay_getlength(BlockDriverState *bs)
+ static void blkreplay_bh_cb(void *opaque)
+ {
+     Request *req = opaque;
+-    qemu_coroutine_enter(req->co);
++    aio_co_wake(req->co);
+     qemu_bh_delete(req->bh);
+     g_free(req);
+ }
+diff --git a/block/block-backend.c b/block/block-backend.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/block-backend.c
++++ b/block/block-backend.c
+@@ -XXX,XX +XXX,XX @@ int blk_make_zero(BlockBackend *blk, BdrvRequestFlags flags)
+ static void error_callback_bh(void *opaque)
+ {
+     struct BlockBackendAIOCB *acb = opaque;
++    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
+     bdrv_dec_in_flight(acb->common.bs);
++    aio_context_acquire(ctx);
+     acb->common.cb(acb->common.opaque, acb->ret);
++    aio_context_release(ctx);
+     qemu_aio_unref(acb);
+ }
+@@ -XXX,XX +XXX,XX @@ static void blk_aio_complete(BlkAioEmAIOCB *acb)
+ static void blk_aio_complete_bh(void *opaque)
+ {
+     BlkAioEmAIOCB *acb = opaque;
++    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
+     assert(acb->has_returned);
++    aio_context_acquire(ctx);
+     blk_aio_complete(acb);
++    aio_context_release(ctx);
+ }
+ static BlockAIOCB *blk_aio_prwv(BlockBackend *blk, int64_t offset, int bytes,
+diff --git a/block/curl.c b/block/curl.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/curl.c
++++ b/block/curl.c
+@@ -XXX,XX +XXX,XX @@ static void curl_readv_bh_cb(void *p)
+ {
+     CURLState *state;
+     int running;
++    int ret = -EINPROGRESS;
+     CURLAIOCB *acb = p;
+-    BDRVCURLState *s = acb->common.bs->opaque;
++    BlockDriverState *bs = acb->common.bs;
++    BDRVCURLState *s = bs->opaque;
++    AioContext *ctx = bdrv_get_aio_context(bs);
+     size_t start = acb->sector_num * BDRV_SECTOR_SIZE;
+     size_t end;
++    aio_context_acquire(ctx);
++
+     // In case we have the requested data already (e.g. read-ahead),
+     // we can just call the callback and be done.
+     switch (curl_find_buf(s, start, acb->nb_sectors * BDRV_SECTOR_SIZE, acb)) {
+@@ -XXX,XX +XXX,XX @@ static void curl_readv_bh_cb(void *p)
+             qemu_aio_unref(acb);
+             // fall through
+         case FIND_RET_WAIT:
+-            return;
++            goto out;
+         default:
+             break;
+     }
+@@ -XXX,XX +XXX,XX @@ static void curl_readv_bh_cb(void *p)
+     // No cache found, so let's start a new request
+     state = curl_init_state(acb->common.bs, s);
+     if (!state) {
+-        acb->common.cb(acb->common.opaque, -EIO);
+-        qemu_aio_unref(acb);
+-        return;
++        ret = -EIO;
++        goto out;
+     }
+     acb->start = 0;
+@@ -XXX,XX +XXX,XX @@ static void curl_readv_bh_cb(void *p)
+     state->orig_buf = g_try_malloc(state->buf_len);
+     if (state->buf_len && state->orig_buf == NULL) {
+         curl_clean_state(state);
+-        acb->common.cb(acb->common.opaque, -ENOMEM);
+-        qemu_aio_unref(acb);
+-        return;
++        ret = -ENOMEM;
++        goto out;
+     }
+     state->acb[0] = acb;
+@@ -XXX,XX +XXX,XX @@ static void curl_readv_bh_cb(void *p)
+     /* Tell curl it needs to kick things off */
+     curl_multi_socket_action(s->multi, CURL_SOCKET_TIMEOUT, 0, &running);
++
++out:
++    if (ret != -EINPROGRESS) {
++        acb->common.cb(acb->common.opaque, ret);
++        qemu_aio_unref(acb);
++    }
++    aio_context_release(ctx);
+ }
+ static BlockAIOCB *curl_aio_readv(BlockDriverState *bs,
+diff --git a/block/gluster.c b/block/gluster.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/gluster.c
++++ b/block/gluster.c
+@@ -XXX,XX +XXX,XX @@ static struct glfs *qemu_gluster_init(BlockdevOptionsGluster *gconf,
+     return qemu_gluster_glfs_init(gconf, errp);
+ }
+-static void qemu_gluster_complete_aio(void *opaque)
+-{
+-    GlusterAIOCB *acb = (GlusterAIOCB *)opaque;
+-
+-    qemu_coroutine_enter(acb->coroutine);
+-}
+-
+ /*
+  * AIO callback routine called from GlusterFS thread.
+  */
+@@ -XXX,XX +XXX,XX @@ static void gluster_finish_aiocb(struct glfs_fd *fd, ssize_t ret, void *arg)
+         acb->ret = -EIO; /* Partial read/write - fail it */
+     }
+-    aio_bh_schedule_oneshot(acb->aio_context, qemu_gluster_complete_aio, acb);
++    aio_co_schedule(acb->aio_context, acb->coroutine);
+ }
+ static void qemu_gluster_parse_flags(int bdrv_flags, int *open_flags)
+diff --git a/block/io.c b/block/io.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/io.c
++++ b/block/io.c
+@@ -XXX,XX +XXX,XX @@ static void bdrv_co_drain_bh_cb(void *opaque)
+     bdrv_dec_in_flight(bs);
+     bdrv_drained_begin(bs);
+     data->done = true;
+-    qemu_coroutine_enter(co);
++    aio_co_wake(co);
+ }
+ static void coroutine_fn bdrv_co_yield_to_drain(BlockDriverState *bs)
+@@ -XXX,XX +XXX,XX @@ static void bdrv_co_complete(BlockAIOCBCoroutine *acb)
+ static void bdrv_co_em_bh(void *opaque)
+ {
+     BlockAIOCBCoroutine *acb = opaque;
++    BlockDriverState *bs = acb->common.bs;
++    AioContext *ctx = bdrv_get_aio_context(bs);
+     assert(!acb->need_bh);
++    aio_context_acquire(ctx);
+     bdrv_co_complete(acb);
++    aio_context_release(ctx);
+ }
+ static void bdrv_co_maybe_schedule_bh(BlockAIOCBCoroutine *acb)
+diff --git a/block/iscsi.c b/block/iscsi.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/iscsi.c
++++ b/block/iscsi.c
+@@ -XXX,XX +XXX,XX @@ static void
+ iscsi_bh_cb(void *p)
+ {
+     IscsiAIOCB *acb = p;
++    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
+     qemu_bh_delete(acb->bh);
+     g_free(acb->buf);
+     acb->buf = NULL;
++    aio_context_acquire(ctx);
+     acb->common.cb(acb->common.opaque, acb->status);
++    aio_context_release(ctx);
+     if (acb->task != NULL) {
+         scsi_free_scsi_task(acb->task);
+@@ -XXX,XX +XXX,XX @@ iscsi_schedule_bh(IscsiAIOCB *acb)
+ static void iscsi_co_generic_bh_cb(void *opaque)
+ {
+     struct IscsiTask *iTask = opaque;
++
+     iTask->complete = 1;
+-    qemu_coroutine_enter(iTask->co);
++    aio_co_wake(iTask->co);
+ }
+ static void iscsi_retry_timer_expired(void *opaque)
+diff --git a/block/linux-aio.c b/block/linux-aio.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/linux-aio.c
++++ b/block/linux-aio.c
+@@ -XXX,XX +XXX,XX @@ struct LinuxAioState {
+     io_context_t ctx;
+     EventNotifier e;
+-    /* io queue for submit at batch */
++    /* io queue for submit at batch.  Protected by AioContext lock. */
+     LaioQueue io_q;
+-    /* I/O completion processing */
++    /* I/O completion processing.  Only runs in I/O thread.  */
+     QEMUBH *completion_bh;
+     int event_idx;
+     int event_max;
+@@ -XXX,XX +XXX,XX @@ static inline ssize_t io_event_ret(struct io_event *ev)
+  */
+ static void qemu_laio_process_completion(struct qemu_laiocb *laiocb)
+ {
++    LinuxAioState *s = laiocb->ctx;
+     int ret;
+     ret = laiocb->ret;
+@@ -XXX,XX +XXX,XX @@ static void qemu_laio_process_completion(struct qemu_laiocb *laiocb)
+     }
+     laiocb->ret = ret;
++    aio_context_acquire(s->aio_context);
+     if (laiocb->co) {
+         /* If the coroutine is already entered it must be in ioq_submit() and
+          * will notice laio->ret has been filled in when it eventually runs
+@@ -XXX,XX +XXX,XX @@ static void qemu_laio_process_completion(struct qemu_laiocb *laiocb)
+         laiocb->common.cb(laiocb->common.opaque, ret);
+         qemu_aio_unref(laiocb);
+     }
++    aio_context_release(s->aio_context);
+ }
+ /**
+@@ -XXX,XX +XXX,XX @@ static void qemu_laio_process_completions(LinuxAioState *s)
+ static void qemu_laio_process_completions_and_submit(LinuxAioState *s)
+ {
+     qemu_laio_process_completions(s);
++
++    aio_context_acquire(s->aio_context);
+     if (!s->io_q.plugged && !QSIMPLEQ_EMPTY(&s->io_q.pending)) {
+         ioq_submit(s);
+     }
++    aio_context_release(s->aio_context);
+ }
+ static void qemu_laio_completion_bh(void *opaque)
+@@ -XXX,XX +XXX,XX @@ static void qemu_laio_completion_cb(EventNotifier *e)
+     LinuxAioState *s = container_of(e, LinuxAioState, e);
+     if (event_notifier_test_and_clear(&s->e)) {
+-        aio_context_acquire(s->aio_context);
+         qemu_laio_process_completions_and_submit(s);
+-        aio_context_release(s->aio_context);
+     }
+ }
+@@ -XXX,XX +XXX,XX @@ static bool qemu_laio_poll_cb(void *opaque)
+         return false;
+     }
+-    aio_context_acquire(s->aio_context);
+     qemu_laio_process_completions_and_submit(s);
+-    aio_context_release(s->aio_context);
+     return true;
+ }
+@@ -XXX,XX +XXX,XX @@ void laio_detach_aio_context(LinuxAioState *s, AioContext *old_context)
+ {
+     aio_set_event_notifier(old_context, &s->e, false, NULL, NULL);
+     qemu_bh_delete(s->completion_bh);
++    s->aio_context = NULL;
+ }
+ void laio_attach_aio_context(LinuxAioState *s, AioContext *new_context)
+diff --git a/block/nfs.c b/block/nfs.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/nfs.c
++++ b/block/nfs.c
+@@ -XXX,XX +XXX,XX @@ static void nfs_co_init_task(BlockDriverState *bs, NFSRPC *task)
+ static void nfs_co_generic_bh_cb(void *opaque)
+ {
+     NFSRPC *task = opaque;
++
+     task->complete = 1;
+-    qemu_coroutine_enter(task->co);
++    aio_co_wake(task->co);
+ }
+ static void
+diff --git a/block/null.c b/block/null.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/null.c
++++ b/block/null.c
+@@ -XXX,XX +XXX,XX @@ static const AIOCBInfo null_aiocb_info = {
+ static void null_bh_cb(void *opaque)
+ {
+     NullAIOCB *acb = opaque;
++    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
++
++    aio_context_acquire(ctx);
+     acb->common.cb(acb->common.opaque, 0);
++    aio_context_release(ctx);
+     qemu_aio_unref(acb);
+ }
+diff --git a/block/qed.c b/block/qed.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/qed.c
++++ b/block/qed.c
+@@ -XXX,XX +XXX,XX @@ static void qed_update_l2_table(BDRVQEDState *s, QEDTable *table, int index,
+ static void qed_aio_complete_bh(void *opaque)
+ {
+     QEDAIOCB *acb = opaque;
++    BDRVQEDState *s = acb_to_s(acb);
+     BlockCompletionFunc *cb = acb->common.cb;
+     void *user_opaque = acb->common.opaque;
+     int ret = acb->bh_ret;
+@@ -XXX,XX +XXX,XX @@ static void qed_aio_complete_bh(void *opaque)
+     qemu_aio_unref(acb);
+     /* Invoke callback */
++    qed_acquire(s);
+     cb(user_opaque, ret);
++    qed_release(s);
+ }
+ static void qed_aio_complete(QEDAIOCB *acb, int ret)
+diff --git a/block/rbd.c b/block/rbd.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/rbd.c
++++ b/block/rbd.c
+@@ -XXX,XX +XXX,XX @@ shutdown:
+ static void qemu_rbd_complete_aio(RADOSCB *rcb)
+ {
+     RBDAIOCB *acb = rcb->acb;
++    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
+     int64_t r;
+     r = rcb->ret;
+@@ -XXX,XX +XXX,XX @@ static void qemu_rbd_complete_aio(RADOSCB *rcb)
+         qemu_iovec_from_buf(acb->qiov, 0, acb->bounce, acb->qiov->size);
+     }
+     qemu_vfree(acb->bounce);
++
++    aio_context_acquire(ctx);
+     acb->common.cb(acb->common.opaque, (acb->ret > 0 ? 0 : acb->ret));
++    aio_context_release(ctx);
+     qemu_aio_unref(acb);
+ }
+diff --git a/dma-helpers.c b/dma-helpers.c
+index XXXXXXX..XXXXXXX 100644
+--- a/dma-helpers.c
++++ b/dma-helpers.c
+@@ -XXX,XX +XXX,XX @@ static void dma_blk_cb(void *opaque, int ret)
+                                 QEMU_ALIGN_DOWN(dbs->iov.size, dbs->align));
+     }
++    aio_context_acquire(dbs->ctx);
+     dbs->acb = dbs->io_func(dbs->offset, &dbs->iov,
+                             dma_blk_cb, dbs, dbs->io_func_opaque);
++    aio_context_release(dbs->ctx);
+     assert(dbs->acb);
+ }
+diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c
+index XXXXXXX..XXXXXXX 100644
+--- a/hw/block/virtio-blk.c
++++ b/hw/block/virtio-blk.c
+@@ -XXX,XX +XXX,XX @@ static void virtio_blk_dma_restart_bh(void *opaque)
+     s->rq = NULL;
++    aio_context_acquire(blk_get_aio_context(s->conf.conf.blk));
+     while (req) {
+         VirtIOBlockReq *next = req->next;
+         if (virtio_blk_handle_request(req, &mrb)) {
+@@ -XXX,XX +XXX,XX @@ static void virtio_blk_dma_restart_bh(void *opaque)
+     if (mrb.num_reqs) {
+         virtio_blk_submit_multireq(s->blk, &mrb);
+     }
++    aio_context_release(blk_get_aio_context(s->conf.conf.blk));
+ }
+ static void virtio_blk_dma_restart_cb(void *opaque, int running,
+diff --git a/hw/scsi/scsi-bus.c b/hw/scsi/scsi-bus.c
+index XXXXXXX..XXXXXXX 100644
+--- a/hw/scsi/scsi-bus.c
++++ b/hw/scsi/scsi-bus.c
+@@ -XXX,XX +XXX,XX @@ static void scsi_dma_restart_bh(void *opaque)
+     qemu_bh_delete(s->bh);
+     s->bh = NULL;
++    aio_context_acquire(blk_get_aio_context(s->conf.blk));
+     QTAILQ_FOREACH_SAFE(req, &s->requests, next, next) {
+         scsi_req_ref(req);
+         if (req->retry) {
+@@ -XXX,XX +XXX,XX @@ static void scsi_dma_restart_bh(void *opaque)
+         }
+         scsi_req_unref(req);
+     }
++    aio_context_release(blk_get_aio_context(s->conf.blk));
+ }
+ void scsi_req_retry(SCSIRequest *req)
+diff --git a/util/async.c b/util/async.c
+index XXXXXXX..XXXXXXX 100644
+--- a/util/async.c
++++ b/util/async.c
+@@ -XXX,XX +XXX,XX @@ int aio_bh_poll(AioContext *ctx)
+                 ret = 1;
+             }
+             bh->idle = 0;
+-            aio_context_acquire(ctx);
+             aio_bh_call(bh);
+-            aio_context_release(ctx);
+         }
+         if (bh->deleted) {
+             deleted = true;
+@@ -XXX,XX +XXX,XX @@ static void co_schedule_bh_cb(void *opaque)
+         Coroutine *co = QSLIST_FIRST(&straight);
+         QSLIST_REMOVE_HEAD(&straight, co_scheduled_next);
+         trace_aio_co_schedule_bh_cb(ctx, co);
++        aio_context_acquire(ctx);
+         qemu_coroutine_enter(co);
++        aio_context_release(ctx);
+     }
+ }
+diff --git a/util/thread-pool.c b/util/thread-pool.c
+index XXXXXXX..XXXXXXX 100644
+--- a/util/thread-pool.c
++++ b/util/thread-pool.c
+@@ -XXX,XX +XXX,XX @@ static void thread_pool_completion_bh(void *opaque)
+     ThreadPool *pool = opaque;
+     ThreadPoolElement *elem, *next;
++    aio_context_acquire(pool->ctx);
+ restart:
+     QLIST_FOREACH_SAFE(elem, &pool->head, all, next) {
+         if (elem->state != THREAD_DONE) {
+@@ -XXX,XX +XXX,XX @@ restart:
+             qemu_aio_unref(elem);
+         }
+     }
++    aio_context_release(pool->ctx);
+ }
+ static void thread_pool_cancel(BlockAIOCB *acb)
+--
+.9.3

-New patch
+[Qemu-devel] [PULL v2 15/24] block: explicitly acquire aiocontext in aio callbacks that need it
+From: Paolo Bonzini <pbonzini@redhat.com>
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
+Reviewed-by: Fam Zheng <famz@redhat.com>
+Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
+Message-id: 20170213135235.12274-16-pbonzini@redhat.com
+Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
+---
+ block/archipelago.c    |  3 ---
+ block/block-backend.c  |  7 -------
+ block/curl.c           |  2 +-
+ block/io.c             |  6 +-----
+ block/iscsi.c          |  3 ---
+ block/linux-aio.c      |  5 +----
+ block/mirror.c         | 12 +++++++++---
+ block/null.c           |  8 --------
+ block/qed-cluster.c    |  2 ++
+ block/qed-table.c      | 12 ++++++++++--
+ block/qed.c            |  4 ++--
+ block/rbd.c            |  4 ----
+ block/win32-aio.c      |  3 ---
+ hw/block/virtio-blk.c  | 12 +++++++++++-
+ hw/scsi/scsi-disk.c    | 15 +++++++++++++++
+ hw/scsi/scsi-generic.c | 20 +++++++++++++++++---
+ util/thread-pool.c     |  4 +++-
+files changed, 72 insertions(+), 50 deletions(-)
+diff --git a/block/archipelago.c b/block/archipelago.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/archipelago.c
++++ b/block/archipelago.c
+@@ -XXX,XX +XXX,XX @@ static void qemu_archipelago_complete_aio(void *opaque)
+ {
+     AIORequestData *reqdata = (AIORequestData *) opaque;
+     ArchipelagoAIOCB *aio_cb = (ArchipelagoAIOCB *) reqdata->aio_cb;
+-    AioContext *ctx = bdrv_get_aio_context(aio_cb->common.bs);
+-    aio_context_acquire(ctx);
+     aio_cb->common.cb(aio_cb->common.opaque, aio_cb->ret);
+-    aio_context_release(ctx);
+     aio_cb->status = 0;
+     qemu_aio_unref(aio_cb);
+diff --git a/block/block-backend.c b/block/block-backend.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/block-backend.c
++++ b/block/block-backend.c
+@@ -XXX,XX +XXX,XX @@ int blk_make_zero(BlockBackend *blk, BdrvRequestFlags flags)
+ static void error_callback_bh(void *opaque)
+ {
+     struct BlockBackendAIOCB *acb = opaque;
+-    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
+     bdrv_dec_in_flight(acb->common.bs);
+-    aio_context_acquire(ctx);
+     acb->common.cb(acb->common.opaque, acb->ret);
+-    aio_context_release(ctx);
+     qemu_aio_unref(acb);
+ }
+@@ -XXX,XX +XXX,XX @@ static void blk_aio_complete(BlkAioEmAIOCB *acb)
+ static void blk_aio_complete_bh(void *opaque)
+ {
+     BlkAioEmAIOCB *acb = opaque;
+-    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
+-
+     assert(acb->has_returned);
+-    aio_context_acquire(ctx);
+     blk_aio_complete(acb);
+-    aio_context_release(ctx);
+ }
+ static BlockAIOCB *blk_aio_prwv(BlockBackend *blk, int64_t offset, int bytes,
+diff --git a/block/curl.c b/block/curl.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/curl.c
++++ b/block/curl.c
+@@ -XXX,XX +XXX,XX @@ static void curl_readv_bh_cb(void *p)
+     curl_multi_socket_action(s->multi, CURL_SOCKET_TIMEOUT, 0, &running);
+ out:
++    aio_context_release(ctx);
+     if (ret != -EINPROGRESS) {
+         acb->common.cb(acb->common.opaque, ret);
+         qemu_aio_unref(acb);
+     }
+-    aio_context_release(ctx);
+ }
+ static BlockAIOCB *curl_aio_readv(BlockDriverState *bs,
+diff --git a/block/io.c b/block/io.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/io.c
++++ b/block/io.c
+@@ -XXX,XX +XXX,XX @@ static void bdrv_co_io_em_complete(void *opaque, int ret)
+     CoroutineIOCompletion *co = opaque;
+     co->ret = ret;
+-    qemu_coroutine_enter(co->coroutine);
++    aio_co_wake(co->coroutine);
+ }
+ static int coroutine_fn bdrv_driver_preadv(BlockDriverState *bs,
+@@ -XXX,XX +XXX,XX @@ static void bdrv_co_complete(BlockAIOCBCoroutine *acb)
+ static void bdrv_co_em_bh(void *opaque)
+ {
+     BlockAIOCBCoroutine *acb = opaque;
+-    BlockDriverState *bs = acb->common.bs;
+-    AioContext *ctx = bdrv_get_aio_context(bs);
+     assert(!acb->need_bh);
+-    aio_context_acquire(ctx);
+     bdrv_co_complete(acb);
+-    aio_context_release(ctx);
+ }
+ static void bdrv_co_maybe_schedule_bh(BlockAIOCBCoroutine *acb)
+diff --git a/block/iscsi.c b/block/iscsi.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/iscsi.c
++++ b/block/iscsi.c
+@@ -XXX,XX +XXX,XX @@ static void
+ iscsi_bh_cb(void *p)
+ {
+     IscsiAIOCB *acb = p;
+-    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
+     qemu_bh_delete(acb->bh);
+     g_free(acb->buf);
+     acb->buf = NULL;
+-    aio_context_acquire(ctx);
+     acb->common.cb(acb->common.opaque, acb->status);
+-    aio_context_release(ctx);
+     if (acb->task != NULL) {
+         scsi_free_scsi_task(acb->task);
+diff --git a/block/linux-aio.c b/block/linux-aio.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/linux-aio.c
++++ b/block/linux-aio.c
+@@ -XXX,XX +XXX,XX @@ static inline ssize_t io_event_ret(struct io_event *ev)
+  */
+ static void qemu_laio_process_completion(struct qemu_laiocb *laiocb)
+ {
+-    LinuxAioState *s = laiocb->ctx;
+     int ret;
+     ret = laiocb->ret;
+@@ -XXX,XX +XXX,XX @@ static void qemu_laio_process_completion(struct qemu_laiocb *laiocb)
+     }
+     laiocb->ret = ret;
+-    aio_context_acquire(s->aio_context);
+     if (laiocb->co) {
+         /* If the coroutine is already entered it must be in ioq_submit() and
+          * will notice laio->ret has been filled in when it eventually runs
+@@ -XXX,XX +XXX,XX @@ static void qemu_laio_process_completion(struct qemu_laiocb *laiocb)
+          * that!
+          */
+         if (!qemu_coroutine_entered(laiocb->co)) {
+-            qemu_coroutine_enter(laiocb->co);
++            aio_co_wake(laiocb->co);
+         }
+     } else {
+         laiocb->common.cb(laiocb->common.opaque, ret);
+         qemu_aio_unref(laiocb);
+     }
+-    aio_context_release(s->aio_context);
+ }
+ /**
+diff --git a/block/mirror.c b/block/mirror.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/mirror.c
++++ b/block/mirror.c
+@@ -XXX,XX +XXX,XX @@ static void mirror_write_complete(void *opaque, int ret)
+ {
+     MirrorOp *op = opaque;
+     MirrorBlockJob *s = op->s;
++
++    aio_context_acquire(blk_get_aio_context(s->common.blk));
+     if (ret < 0) {
+         BlockErrorAction action;
+@@ -XXX,XX +XXX,XX @@ static void mirror_write_complete(void *opaque, int ret)
+         }
+     }
+     mirror_iteration_done(op, ret);
++    aio_context_release(blk_get_aio_context(s->common.blk));
+ }
+ static void mirror_read_complete(void *opaque, int ret)
+ {
+     MirrorOp *op = opaque;
+     MirrorBlockJob *s = op->s;
++
++    aio_context_acquire(blk_get_aio_context(s->common.blk));
+     if (ret < 0) {
+         BlockErrorAction action;
+@@ -XXX,XX +XXX,XX @@ static void mirror_read_complete(void *opaque, int ret)
+         }
+         mirror_iteration_done(op, ret);
+-        return;
++    } else {
++        blk_aio_pwritev(s->target, op->sector_num * BDRV_SECTOR_SIZE, &op->qiov,
++                        0, mirror_write_complete, op);
+     }
+-    blk_aio_pwritev(s->target, op->sector_num * BDRV_SECTOR_SIZE, &op->qiov,
+-                    0, mirror_write_complete, op);
++    aio_context_release(blk_get_aio_context(s->common.blk));
+ }
+ static inline void mirror_clip_sectors(MirrorBlockJob *s,
+diff --git a/block/null.c b/block/null.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/null.c
++++ b/block/null.c
+@@ -XXX,XX +XXX,XX @@ static const AIOCBInfo null_aiocb_info = {
+ static void null_bh_cb(void *opaque)
+ {
+     NullAIOCB *acb = opaque;
+-    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
+-
+-    aio_context_acquire(ctx);
+     acb->common.cb(acb->common.opaque, 0);
+-    aio_context_release(ctx);
+     qemu_aio_unref(acb);
+ }
+ static void null_timer_cb(void *opaque)
+ {
+     NullAIOCB *acb = opaque;
+-    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
+-
+-    aio_context_acquire(ctx);
+     acb->common.cb(acb->common.opaque, 0);
+-    aio_context_release(ctx);
+     timer_deinit(&acb->timer);
+     qemu_aio_unref(acb);
+ }
+diff --git a/block/qed-cluster.c b/block/qed-cluster.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/qed-cluster.c
++++ b/block/qed-cluster.c
+@@ -XXX,XX +XXX,XX @@ static void qed_find_cluster_cb(void *opaque, int ret)
+     unsigned int index;
+     unsigned int n;
++    qed_acquire(s);
+     if (ret) {
+         goto out;
+     }
+@@ -XXX,XX +XXX,XX @@ static void qed_find_cluster_cb(void *opaque, int ret)
+ out:
+     find_cluster_cb->cb(find_cluster_cb->opaque, ret, offset, len);
++    qed_release(s);
+     g_free(find_cluster_cb);
+ }
+diff --git a/block/qed-table.c b/block/qed-table.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/qed-table.c
++++ b/block/qed-table.c
+@@ -XXX,XX +XXX,XX @@ static void qed_read_table_cb(void *opaque, int ret)
+ {
+     QEDReadTableCB *read_table_cb = opaque;
+     QEDTable *table = read_table_cb->table;
++    BDRVQEDState *s = read_table_cb->s;
+     int noffsets = read_table_cb->qiov.size / sizeof(uint64_t);
+     int i;
+@@ -XXX,XX +XXX,XX @@ static void qed_read_table_cb(void *opaque, int ret)
+     }
+     /* Byteswap offsets */
++    qed_acquire(s);
+     for (i = 0; i < noffsets; i++) {
+         table->offsets[i] = le64_to_cpu(table->offsets[i]);
+     }
++    qed_release(s);
+ out:
+     /* Completion */
+-    trace_qed_read_table_cb(read_table_cb->s, read_table_cb->table, ret);
++    trace_qed_read_table_cb(s, read_table_cb->table, ret);
+     gencb_complete(&read_table_cb->gencb, ret);
+ }
+@@ -XXX,XX +XXX,XX @@ typedef struct {
+ static void qed_write_table_cb(void *opaque, int ret)
+ {
+     QEDWriteTableCB *write_table_cb = opaque;
++    BDRVQEDState *s = write_table_cb->s;
+-    trace_qed_write_table_cb(write_table_cb->s,
++    trace_qed_write_table_cb(s,
+                              write_table_cb->orig_table,
+                              write_table_cb->flush,
+                              ret);
+@@ -XXX,XX +XXX,XX @@ static void qed_write_table_cb(void *opaque, int ret)
+     if (write_table_cb->flush) {
+         /* We still need to flush first */
+         write_table_cb->flush = false;
++        qed_acquire(s);
+         bdrv_aio_flush(write_table_cb->s->bs, qed_write_table_cb,
+                        write_table_cb);
++        qed_release(s);
+         return;
+     }
+@@ -XXX,XX +XXX,XX @@ static void qed_read_l2_table_cb(void *opaque, int ret)
+     CachedL2Table *l2_table = request->l2_table;
+     uint64_t l2_offset = read_l2_table_cb->l2_offset;
++    qed_acquire(s);
+     if (ret) {
+         /* can't trust loaded L2 table anymore */
+         qed_unref_l2_cache_entry(l2_table);
+@@ -XXX,XX +XXX,XX @@ static void qed_read_l2_table_cb(void *opaque, int ret)
+         request->l2_table = qed_find_l2_cache_entry(&s->l2_cache, l2_offset);
+         assert(request->l2_table != NULL);
+     }
++    qed_release(s);
+     gencb_complete(&read_l2_table_cb->gencb, ret);
+ }
+diff --git a/block/qed.c b/block/qed.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/qed.c
++++ b/block/qed.c
+@@ -XXX,XX +XXX,XX @@ static void qed_is_allocated_cb(void *opaque, int ret, uint64_t offset, size_t l
+     }
+     if (cb->co) {
+-        qemu_coroutine_enter(cb->co);
++        aio_co_wake(cb->co);
+     }
+ }
+@@ -XXX,XX +XXX,XX @@ static void coroutine_fn qed_co_pwrite_zeroes_cb(void *opaque, int ret)
+     cb->done = true;
+     cb->ret = ret;
+     if (cb->co) {
+-        qemu_coroutine_enter(cb->co);
++        aio_co_wake(cb->co);
+     }
+ }
+diff --git a/block/rbd.c b/block/rbd.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/rbd.c
++++ b/block/rbd.c
+@@ -XXX,XX +XXX,XX @@ shutdown:
+ static void qemu_rbd_complete_aio(RADOSCB *rcb)
+ {
+     RBDAIOCB *acb = rcb->acb;
+-    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
+     int64_t r;
+     r = rcb->ret;
+@@ -XXX,XX +XXX,XX @@ static void qemu_rbd_complete_aio(RADOSCB *rcb)
+         qemu_iovec_from_buf(acb->qiov, 0, acb->bounce, acb->qiov->size);
+     }
+     qemu_vfree(acb->bounce);
+-
+-    aio_context_acquire(ctx);
+     acb->common.cb(acb->common.opaque, (acb->ret > 0 ? 0 : acb->ret));
+-    aio_context_release(ctx);
+     qemu_aio_unref(acb);
+ }
+diff --git a/block/win32-aio.c b/block/win32-aio.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/win32-aio.c
++++ b/block/win32-aio.c
+@@ -XXX,XX +XXX,XX @@ static void win32_aio_process_completion(QEMUWin32AIOState *s,
+         qemu_vfree(waiocb->buf);
+     }
+-
+-    aio_context_acquire(s->aio_ctx);
+     waiocb->common.cb(waiocb->common.opaque, ret);
+-    aio_context_release(s->aio_ctx);
+     qemu_aio_unref(waiocb);
+ }
+diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c
+index XXXXXXX..XXXXXXX 100644
+--- a/hw/block/virtio-blk.c
++++ b/hw/block/virtio-blk.c
+@@ -XXX,XX +XXX,XX @@ static int virtio_blk_handle_rw_error(VirtIOBlockReq *req, int error,
+ static void virtio_blk_rw_complete(void *opaque, int ret)
+ {
+     VirtIOBlockReq *next = opaque;
++    VirtIOBlock *s = next->dev;
++    aio_context_acquire(blk_get_aio_context(s->conf.conf.blk));
+     while (next) {
+         VirtIOBlockReq *req = next;
+         next = req->mr_next;
+@@ -XXX,XX +XXX,XX @@ static void virtio_blk_rw_complete(void *opaque, int ret)
+         block_acct_done(blk_get_stats(req->dev->blk), &req->acct);
+         virtio_blk_free_request(req);
+     }
++    aio_context_release(blk_get_aio_context(s->conf.conf.blk));
+ }
+ static void virtio_blk_flush_complete(void *opaque, int ret)
+ {
+     VirtIOBlockReq *req = opaque;
++    VirtIOBlock *s = req->dev;
++    aio_context_acquire(blk_get_aio_context(s->conf.conf.blk));
+     if (ret) {
+         if (virtio_blk_handle_rw_error(req, -ret, 0)) {
+-            return;
++            goto out;
+         }
+     }
+     virtio_blk_req_complete(req, VIRTIO_BLK_S_OK);
+     block_acct_done(blk_get_stats(req->dev->blk), &req->acct);
+     virtio_blk_free_request(req);
++
++out:
++    aio_context_release(blk_get_aio_context(s->conf.conf.blk));
+ }
+ #ifdef __linux__
+@@ -XXX,XX +XXX,XX @@ static void virtio_blk_ioctl_complete(void *opaque, int status)
+     virtio_stl_p(vdev, &scsi->data_len, hdr->dxfer_len);
+ out:
++    aio_context_acquire(blk_get_aio_context(s->conf.conf.blk));
+     virtio_blk_req_complete(req, status);
+     virtio_blk_free_request(req);
++    aio_context_release(blk_get_aio_context(s->conf.conf.blk));
+     g_free(ioctl_req);
+ }
+diff --git a/hw/scsi/scsi-disk.c b/hw/scsi/scsi-disk.c
+index XXXXXXX..XXXXXXX 100644
+--- a/hw/scsi/scsi-disk.c
++++ b/hw/scsi/scsi-disk.c
+@@ -XXX,XX +XXX,XX @@ static void scsi_aio_complete(void *opaque, int ret)
+     assert(r->req.aiocb != NULL);
+     r->req.aiocb = NULL;
++    aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk));
+     if (scsi_disk_req_check_error(r, ret, true)) {
+         goto done;
+     }
+@@ -XXX,XX +XXX,XX @@ static void scsi_aio_complete(void *opaque, int ret)
+     scsi_req_complete(&r->req, GOOD);
+ done:
++    aio_context_release(blk_get_aio_context(s->qdev.conf.blk));
+     scsi_req_unref(&r->req);
+ }
+@@ -XXX,XX +XXX,XX @@ static void scsi_dma_complete(void *opaque, int ret)
+     assert(r->req.aiocb != NULL);
+     r->req.aiocb = NULL;
++    aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk));
+     if (ret < 0) {
+         block_acct_failed(blk_get_stats(s->qdev.conf.blk), &r->acct);
+     } else {
+         block_acct_done(blk_get_stats(s->qdev.conf.blk), &r->acct);
+     }
+     scsi_dma_complete_noio(r, ret);
++    aio_context_release(blk_get_aio_context(s->qdev.conf.blk));
+ }
+ static void scsi_read_complete(void * opaque, int ret)
+@@ -XXX,XX +XXX,XX @@ static void scsi_read_complete(void * opaque, int ret)
+     assert(r->req.aiocb != NULL);
+     r->req.aiocb = NULL;
++    aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk));
+     if (scsi_disk_req_check_error(r, ret, true)) {
+         goto done;
+     }
+@@ -XXX,XX +XXX,XX @@ static void scsi_read_complete(void * opaque, int ret)
+ done:
+     scsi_req_unref(&r->req);
++    aio_context_release(blk_get_aio_context(s->qdev.conf.blk));
+ }
+ /* Actually issue a read to the block device.  */
+@@ -XXX,XX +XXX,XX @@ static void scsi_do_read_cb(void *opaque, int ret)
+     assert (r->req.aiocb != NULL);
+     r->req.aiocb = NULL;
++    aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk));
+     if (ret < 0) {
+         block_acct_failed(blk_get_stats(s->qdev.conf.blk), &r->acct);
+     } else {
+         block_acct_done(blk_get_stats(s->qdev.conf.blk), &r->acct);
+     }
+     scsi_do_read(opaque, ret);
++    aio_context_release(blk_get_aio_context(s->qdev.conf.blk));
+ }
+ /* Read more data from scsi device into buffer.  */
+@@ -XXX,XX +XXX,XX @@ static void scsi_write_complete(void * opaque, int ret)
+     assert (r->req.aiocb != NULL);
+     r->req.aiocb = NULL;
++    aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk));
+     if (ret < 0) {
+         block_acct_failed(blk_get_stats(s->qdev.conf.blk), &r->acct);
+     } else {
+         block_acct_done(blk_get_stats(s->qdev.conf.blk), &r->acct);
+     }
+     scsi_write_complete_noio(r, ret);
++    aio_context_release(blk_get_aio_context(s->qdev.conf.blk));
+ }
+ static void scsi_write_data(SCSIRequest *req)
+@@ -XXX,XX +XXX,XX @@ static void scsi_unmap_complete(void *opaque, int ret)
+ {
+     UnmapCBData *data = opaque;
+     SCSIDiskReq *r = data->r;
++    SCSIDiskState *s = DO_UPCAST(SCSIDiskState, qdev, r->req.dev);
+     assert(r->req.aiocb != NULL);
+     r->req.aiocb = NULL;
++    aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk));
+     scsi_unmap_complete_noio(data, ret);
++    aio_context_release(blk_get_aio_context(s->qdev.conf.blk));
+ }
+ static void scsi_disk_emulate_unmap(SCSIDiskReq *r, uint8_t *inbuf)
+@@ -XXX,XX +XXX,XX @@ static void scsi_write_same_complete(void *opaque, int ret)
+     assert(r->req.aiocb != NULL);
+     r->req.aiocb = NULL;
++    aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk));
+     if (scsi_disk_req_check_error(r, ret, true)) {
+         goto done;
+     }
+@@ -XXX,XX +XXX,XX @@ done:
+     scsi_req_unref(&r->req);
+     qemu_vfree(data->iov.iov_base);
+     g_free(data);
++    aio_context_release(blk_get_aio_context(s->qdev.conf.blk));
+ }
+ static void scsi_disk_emulate_write_same(SCSIDiskReq *r, uint8_t *inbuf)
+diff --git a/hw/scsi/scsi-generic.c b/hw/scsi/scsi-generic.c
+index XXXXXXX..XXXXXXX 100644
+--- a/hw/scsi/scsi-generic.c
++++ b/hw/scsi/scsi-generic.c
+@@ -XXX,XX +XXX,XX @@ done:
+ static void scsi_command_complete(void *opaque, int ret)
+ {
+     SCSIGenericReq *r = (SCSIGenericReq *)opaque;
++    SCSIDevice *s = r->req.dev;
+     assert(r->req.aiocb != NULL);
+     r->req.aiocb = NULL;
++
++    aio_context_acquire(blk_get_aio_context(s->conf.blk));
+     scsi_command_complete_noio(r, ret);
++    aio_context_release(blk_get_aio_context(s->conf.blk));
+ }
+ static int execute_command(BlockBackend *blk,
+@@ -XXX,XX +XXX,XX @@ static void scsi_read_complete(void * opaque, int ret)
+     assert(r->req.aiocb != NULL);
+     r->req.aiocb = NULL;
++    aio_context_acquire(blk_get_aio_context(s->conf.blk));
++
+     if (ret || r->req.io_canceled) {
+         scsi_command_complete_noio(r, ret);
+-        return;
++        goto done;
+     }
+     len = r->io_header.dxfer_len - r->io_header.resid;
+@@ -XXX,XX +XXX,XX @@ static void scsi_read_complete(void * opaque, int ret)
+     r->len = -1;
+     if (len == 0) {
+         scsi_command_complete_noio(r, 0);
+-        return;
++        goto done;
+     }
+     /* Snoop READ CAPACITY output to set the blocksize.  */
+@@ -XXX,XX +XXX,XX @@ static void scsi_read_complete(void * opaque, int ret)
+     }
+     scsi_req_data(&r->req, len);
+     scsi_req_unref(&r->req);
++
++done:
++    aio_context_release(blk_get_aio_context(s->conf.blk));
+ }
+ /* Read more data from scsi device into buffer.  */
+@@ -XXX,XX +XXX,XX @@ static void scsi_write_complete(void * opaque, int ret)
+     assert(r->req.aiocb != NULL);
+     r->req.aiocb = NULL;
++    aio_context_acquire(blk_get_aio_context(s->conf.blk));
++
+     if (ret || r->req.io_canceled) {
+         scsi_command_complete_noio(r, ret);
+-        return;
++        goto done;
+     }
+     if (r->req.cmd.buf[0] == MODE_SELECT && r->req.cmd.buf[4] == 12 &&
+@@ -XXX,XX +XXX,XX @@ static void scsi_write_complete(void * opaque, int ret)
+     }
+     scsi_command_complete_noio(r, ret);
++
++done:
++    aio_context_release(blk_get_aio_context(s->conf.blk));
+ }
+ /* Write data to a scsi device.  Returns nonzero on failure.
+diff --git a/util/thread-pool.c b/util/thread-pool.c
+index XXXXXXX..XXXXXXX 100644
+--- a/util/thread-pool.c
++++ b/util/thread-pool.c
+@@ -XXX,XX +XXX,XX @@ restart:
+              */
+             qemu_bh_schedule(pool->completion_bh);
++            aio_context_release(pool->ctx);
+             elem->common.cb(elem->common.opaque, elem->ret);
++            aio_context_acquire(pool->ctx);
+             qemu_aio_unref(elem);
+             goto restart;
+         } else {
+@@ -XXX,XX +XXX,XX @@ static void thread_pool_co_cb(void *opaque, int ret)
+     ThreadPoolCo *co = opaque;
+     co->ret = ret;
+-    qemu_coroutine_enter(co->co);
++    aio_co_wake(co->co);
+ }
+ int coroutine_fn thread_pool_submit_co(ThreadPool *pool, ThreadPoolFunc *func,
+--
+.9.3

-[PULL 2/5] qcow2: rework the cluster compression routine
+[Qemu-devel] [PULL v2 16/24] aio-posix: partially inline aio_dispatch into aio_poll
-From: Denis Plotnikov <dplotnikov@virtuozzo.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-The patch enables processing the image compression type defined
+This patch prepares for the removal of unnecessary lockcnt inc/dec pairs.
-for the image and chooses an appropriate method for image clusters
+Extract the dispatching loop for file descriptor handlers into a new
-(de)compression.
+function aio_dispatch_handlers, and then inline aio_dispatch into
 aio_poll.
-Signed-off-by: Denis Plotnikov <dplotnikov@virtuozzo.com>
+aio_dispatch can now become void.
-Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Reviewed-by: Alberto Garcia <berto@igalia.com>
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Reviewed-by: Max Reitz <mreitz@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-Message-Id: <20200507082521.29210-3-dplotnikov@virtuozzo.com>
+Reviewed-by: Fam Zheng <famz@redhat.com>
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
 Message-id: 20170213135235.12274-17-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
- block/qcow2-threads.c | 71 ++++++++++++++++++++++++++++++++++++-------
+ include/block/aio.h |  6 +-----
-file changed, 60 insertions(+), 11 deletions(-)
+ util/aio-posix.c    | 44 ++++++++++++++------------------------------
  util/aio-win32.c    | 13 ++++---------
  util/async.c        |  2 +-
 files changed, 20 insertions(+), 45 deletions(-)
-diff --git a/block/qcow2-threads.c b/block/qcow2-threads.c
+diff --git a/include/block/aio.h b/include/block/aio.h
 index XXXXXXX..XXXXXXX 100644
---- a/block/qcow2-threads.c
+--- a/include/block/aio.h
-+++ b/block/qcow2-threads.c
++++ b/include/block/aio.h
-@@ -XXX,XX +XXX,XX @@ typedef struct Qcow2CompressData {
+@@ -XXX,XX +XXX,XX @@ bool aio_pending(AioContext *ctx);
- } Qcow2CompressData;
+ /* Dispatch any pending callbacks from the GSource attached to the AioContext.
  /*
 - * qcow2_compress()
 + * qcow2_zlib_compress()
 + *
 + * Compress @src_size bytes of data using zlib compression method
   *
-  * @dest - destination buffer, @dest_size bytes
+  * This is used internally in the implementation of the GSource.
-  * @src - source buffer, @src_size bytes
+- *
-@@ -XXX,XX +XXX,XX @@ typedef struct Qcow2CompressData {
+- * @dispatch_fds: true to process fds, false to skip them
-  *          -ENOMEM destination buffer is not enough to store compressed data
+- *                (can be used as an optimization by callers that know there
-  *          -EIO    on any other error
+- *                are no fds ready)
   */
--static ssize_t qcow2_compress(void *dest, size_t dest_size,
+-bool aio_dispatch(AioContext *ctx, bool dispatch_fds);
--                              const void *src, size_t src_size)
++void aio_dispatch(AioContext *ctx);
-+static ssize_t qcow2_zlib_compress(void *dest, size_t dest_size,
-+                                   const void *src, size_t src_size)
+ /* Progress in completing AIO work to occur.  This can issue new pending
   * aio as a result of executing I/O completion or bh callbacks.
 diff --git a/util/aio-posix.c b/util/aio-posix.c
 index XXXXXXX..XXXXXXX 100644
 --- a/util/aio-posix.c
 +++ b/util/aio-posix.c
@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
      AioHandler *node, *tmp;
      bool progress = false;
 -    /*
 -     * We have to walk very carefully in case aio_set_fd_handler is
 -     * called while we're walking.
 -     */
 -    qemu_lockcnt_inc(&ctx->list_lock);
 -
      QLIST_FOREACH_SAFE_RCU(node, &ctx->aio_handlers, node, tmp) {
          int revents;
@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
          }
      }
 -    qemu_lockcnt_dec(&ctx->list_lock);
      return progress;
  }
 -/*
 - * Note that dispatch_fds == false has the side-effect of post-poning the
 - * freeing of deleted handlers.
 - */
 -bool aio_dispatch(AioContext *ctx, bool dispatch_fds)
 +void aio_dispatch(AioContext *ctx)
  {
-     ssize_t ret;
+-    bool progress;
-     z_stream strm;
++    aio_bh_poll(ctx);
-@@ -XXX,XX +XXX,XX @@ static ssize_t qcow2_compress(void *dest, size_t dest_size,
 -    /*
 -     * If there are callbacks left that have been queued, we need to call them.
 -     * Do not call select in this case, because it is possible that the caller
 -     * does not need a complete flush (as is the case for aio_poll loops).
 -     */
 -    progress = aio_bh_poll(ctx);
 +    qemu_lockcnt_inc(&ctx->list_lock);
 +    aio_dispatch_handlers(ctx);
 +    qemu_lockcnt_dec(&ctx->list_lock);
 -    if (dispatch_fds) {
 -        progress |= aio_dispatch_handlers(ctx);
 -    }
 -
 -    /* Run our timers */
 -    progress |= timerlistgroup_run_timers(&ctx->tlg);
 -
 -    return progress;
 +    timerlistgroup_run_timers(&ctx->tlg);
  }
- /*
+ /* These thread-local variables are used only in a small part of aio_poll
-- * qcow2_decompress()
+@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
-+ * qcow2_zlib_decompress()
+     npfd = 0;
-  *
+     qemu_lockcnt_dec(&ctx->list_lock);
-  * Decompress some data (not more than @src_size bytes) to produce exactly
-- * @dest_size bytes.
+-    /* Run dispatch even if there were no readable fds to run timers */
-+ * @dest_size bytes using zlib compression method
+-    if (aio_dispatch(ctx, ret > 0)) {
-  *
+-        progress = true;
-  * @dest - destination buffer, @dest_size bytes
++    progress |= aio_bh_poll(ctx);
-  * @src - source buffer, @src_size bytes
++
-@@ -XXX,XX +XXX,XX @@ static ssize_t qcow2_compress(void *dest, size_t dest_size,
++    if (ret > 0) {
-  * Returns: 0 on success
++        qemu_lockcnt_inc(&ctx->list_lock);
-  *          -EIO on fail
++        progress |= aio_dispatch_handlers(ctx);
-  */
++        qemu_lockcnt_dec(&ctx->list_lock);
--static ssize_t qcow2_decompress(void *dest, size_t dest_size,
+     }
--                                const void *src, size_t src_size)
-+static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size,
++    progress |= timerlistgroup_run_timers(&ctx->tlg);
-+                                     const void *src, size_t src_size)
++
      return progress;
  }
 diff --git a/util/aio-win32.c b/util/aio-win32.c
 index XXXXXXX..XXXXXXX 100644
 --- a/util/aio-win32.c
 +++ b/util/aio-win32.c
@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx, HANDLE event)
      return progress;
  }
 -bool aio_dispatch(AioContext *ctx, bool dispatch_fds)
 +void aio_dispatch(AioContext *ctx)
  {
-     int ret;
+-    bool progress;
-     z_stream strm;
+-
-@@ -XXX,XX +XXX,XX @@ qcow2_co_do_compress(BlockDriverState *bs, void *dest, size_t dest_size,
+-    progress = aio_bh_poll(ctx);
-     return arg.ret;
+-    if (dispatch_fds) {
 -        progress |= aio_dispatch_handlers(ctx, INVALID_HANDLE_VALUE);
 -    }
 -    progress |= timerlistgroup_run_timers(&ctx->tlg);
 -    return progress;
 +    aio_bh_poll(ctx);
 +    aio_dispatch_handlers(ctx, INVALID_HANDLE_VALUE);
 +    timerlistgroup_run_timers(&ctx->tlg);
  }
-+/*
+ bool aio_poll(AioContext *ctx, bool blocking)
-+ * qcow2_co_compress()
+diff --git a/util/async.c b/util/async.c
-+ *
+index XXXXXXX..XXXXXXX 100644
-+ * Compress @src_size bytes of data using the compression
+--- a/util/async.c
-+ * method defined by the image compression type
++++ b/util/async.c
-+ *
+@@ -XXX,XX +XXX,XX @@ aio_ctx_dispatch(GSource     *source,
-+ * @dest - destination buffer, @dest_size bytes
+     AioContext *ctx = (AioContext *) source;
-+ * @src - source buffer, @src_size bytes
-+ *
+     assert(callback == NULL);
-+ * Returns: compressed size on success
+-    aio_dispatch(ctx, true);
-+ *          a negative error code on failure
++    aio_dispatch(ctx);
-+ */
+     return true;
  ssize_t coroutine_fn
  qcow2_co_compress(BlockDriverState *bs, void *dest, size_t dest_size,
                    const void *src, size_t src_size)
  {
 -    return qcow2_co_do_compress(bs, dest, dest_size, src, src_size,
 -                                qcow2_compress);
 +    BDRVQcow2State *s = bs->opaque;
 +    Qcow2CompressFunc fn;
 +
 +    switch (s->compression_type) {
 +    case QCOW2_COMPRESSION_TYPE_ZLIB:
 +        fn = qcow2_zlib_compress;
 +        break;
 +
 +    default:
 +        abort();
 +    }
 +
 +    return qcow2_co_do_compress(bs, dest, dest_size, src, src_size, fn);
  }
-+/*
-+ * qcow2_co_decompress()
-+ *
-+ * Decompress some data (not more than @src_size bytes) to produce exactly
-+ * @dest_size bytes using the compression method defined by the image
-+ * compression type
-+ *
-+ * @dest - destination buffer, @dest_size bytes
-+ * @src - source buffer, @src_size bytes
-+ *
-+ * Returns: 0 on success
-+ *          a negative error code on failure
-+ */
- ssize_t coroutine_fn
- qcow2_co_decompress(BlockDriverState *bs, void *dest, size_t dest_size,
-                     const void *src, size_t src_size)
- {
--    return qcow2_co_do_compress(bs, dest, dest_size, src, src_size,
--                                qcow2_decompress);
-+    BDRVQcow2State *s = bs->opaque;
-+    Qcow2CompressFunc fn;
-+
-+    switch (s->compression_type) {
-+    case QCOW2_COMPRESSION_TYPE_ZLIB:
-+        fn = qcow2_zlib_decompress;
-+        break;
-+
-+    default:
-+        abort();
-+    }
-+
-+    return qcow2_co_do_compress(bs, dest, dest_size, src, src_size, fn);
- }
 --
-.26.2
+.9.3

-New patch
+[Qemu-devel] [PULL v2 17/24] async: remove unnecessary inc/dec pairs
+From: Paolo Bonzini <pbonzini@redhat.com>
+Pull the increment/decrement pair out of aio_bh_poll and into the
+callers.
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
+Reviewed-by: Fam Zheng <famz@redhat.com>
+Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
+Message-id: 20170213135235.12274-18-pbonzini@redhat.com
+Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
+---
+ util/aio-posix.c |  8 +++-----
+ util/aio-win32.c |  8 ++++----
+ util/async.c     | 12 ++++++------
+files changed, 13 insertions(+), 15 deletions(-)
+diff --git a/util/aio-posix.c b/util/aio-posix.c
+index XXXXXXX..XXXXXXX 100644
+--- a/util/aio-posix.c
++++ b/util/aio-posix.c
+@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
+ void aio_dispatch(AioContext *ctx)
+ {
++    qemu_lockcnt_inc(&ctx->list_lock);
+     aio_bh_poll(ctx);
+-
+-    qemu_lockcnt_inc(&ctx->list_lock);
+     aio_dispatch_handlers(ctx);
+     qemu_lockcnt_dec(&ctx->list_lock);
+@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
+     }
+     npfd = 0;
+-    qemu_lockcnt_dec(&ctx->list_lock);
+     progress |= aio_bh_poll(ctx);
+     if (ret > 0) {
+-        qemu_lockcnt_inc(&ctx->list_lock);
+         progress |= aio_dispatch_handlers(ctx);
+-        qemu_lockcnt_dec(&ctx->list_lock);
+     }
++    qemu_lockcnt_dec(&ctx->list_lock);
++
+     progress |= timerlistgroup_run_timers(&ctx->tlg);
+     return progress;
+diff --git a/util/aio-win32.c b/util/aio-win32.c
+index XXXXXXX..XXXXXXX 100644
+--- a/util/aio-win32.c
++++ b/util/aio-win32.c
+@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx, HANDLE event)
+     bool progress = false;
+     AioHandler *tmp;
+-    qemu_lockcnt_inc(&ctx->list_lock);
+-
+     /*
+      * We have to walk very carefully in case aio_set_fd_handler is
+      * called while we're walking.
+@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx, HANDLE event)
+         }
+     }
+-    qemu_lockcnt_dec(&ctx->list_lock);
+     return progress;
+ }
+ void aio_dispatch(AioContext *ctx)
+ {
++    qemu_lockcnt_inc(&ctx->list_lock);
+     aio_bh_poll(ctx);
+     aio_dispatch_handlers(ctx, INVALID_HANDLE_VALUE);
++    qemu_lockcnt_dec(&ctx->list_lock);
+     timerlistgroup_run_timers(&ctx->tlg);
+ }
+@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
+         }
+     }
+-    qemu_lockcnt_dec(&ctx->list_lock);
+     first = true;
+     /* ctx->notifier is always registered.  */
+@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
+         progress |= aio_dispatch_handlers(ctx, event);
+     } while (count > 0);
++    qemu_lockcnt_dec(&ctx->list_lock);
++
+     progress |= timerlistgroup_run_timers(&ctx->tlg);
+     return progress;
+ }
+diff --git a/util/async.c b/util/async.c
+index XXXXXXX..XXXXXXX 100644
+--- a/util/async.c
++++ b/util/async.c
+@@ -XXX,XX +XXX,XX @@ void aio_bh_call(QEMUBH *bh)
+     bh->cb(bh->opaque);
+ }
+-/* Multiple occurrences of aio_bh_poll cannot be called concurrently */
++/* Multiple occurrences of aio_bh_poll cannot be called concurrently.
++ * The count in ctx->list_lock is incremented before the call, and is
++ * not affected by the call.
++ */
+ int aio_bh_poll(AioContext *ctx)
+ {
+     QEMUBH *bh, **bhp, *next;
+     int ret;
+     bool deleted = false;
+-    qemu_lockcnt_inc(&ctx->list_lock);
+-
+     ret = 0;
+     for (bh = atomic_rcu_read(&ctx->first_bh); bh; bh = next) {
+         next = atomic_rcu_read(&bh->next);
+@@ -XXX,XX +XXX,XX @@ int aio_bh_poll(AioContext *ctx)
+     /* remove deleted bhs */
+     if (!deleted) {
+-        qemu_lockcnt_dec(&ctx->list_lock);
+         return ret;
+     }
+-    if (qemu_lockcnt_dec_and_lock(&ctx->list_lock)) {
++    if (qemu_lockcnt_dec_if_lock(&ctx->list_lock)) {
+         bhp = &ctx->first_bh;
+         while (*bhp) {
+             bh = *bhp;
+@@ -XXX,XX +XXX,XX @@ int aio_bh_poll(AioContext *ctx)
+                 bhp = &bh->next;
+             }
+         }
+-        qemu_lockcnt_unlock(&ctx->list_lock);
++        qemu_lockcnt_inc_and_unlock(&ctx->list_lock);
+     }
+     return ret;
+ }
+--
+.9.3

-[PULL 1/5] qcow2: introduce compression type feature
+[Qemu-devel] [PULL v2 18/24] block: document fields protected by AioContext lock
-From: Denis Plotnikov <dplotnikov@virtuozzo.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-The patch adds some preparation parts for incompatible compression type
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-feature to qcow2 allowing the use different compression methods for
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-image clusters (de)compressing.
+Reviewed-by: Fam Zheng <famz@redhat.com>
 Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
 Message-id: 20170213135235.12274-19-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
  include/block/block_int.h      | 64 +++++++++++++++++++++++++-----------------
  include/sysemu/block-backend.h | 14 ++++++---
 files changed, 49 insertions(+), 29 deletions(-)
-It is implied that the compression type is set on the image creation and
-can be changed only later by image conversion, thus compression type
-defines the only compression algorithm used for the image, and thus,
-for all image clusters.
-The goal of the feature is to add support of other compression methods
-to qcow2. For example, ZSTD which is more effective on compression than ZLIB.
-The default compression is ZLIB. Images created with ZLIB compression type
-are backward compatible with older qemu versions.
-Adding of the compression type breaks a number of tests because now the
-compression type is reported on image creation and there are some changes
-in the qcow2 header in size and offsets.
-The tests are fixed in the following ways:
-    * filter out compression_type for many tests
-    * fix header size, feature table size and backing file offset
-      affected tests: 031, 036, 061, 080
-      header_size +=8: 1 byte compression type
-bytes padding
-      feature_table += 48: incompatible feature compression type
-      backing_file_offset += 56 (8 + 48 -> header_change + feature_table_change)
-    * add "compression type" for test output matching when it isn't filtered
-      affected tests: 049, 060, 061, 065, 082, 085, 144, 182, 185, 198, 206,
-, 255, 274, 280
-Signed-off-by: Denis Plotnikov <dplotnikov@virtuozzo.com>
-Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Reviewed-by: Eric Blake <eblake@redhat.com>
-Reviewed-by: Max Reitz <mreitz@redhat.com>
-QAPI part:
-Acked-by: Markus Armbruster <armbru@redhat.com>
-Message-Id: <20200507082521.29210-2-dplotnikov@virtuozzo.com>
-Signed-off-by: Max Reitz <mreitz@redhat.com>
----
- qapi/block-core.json             |  22 +++++-
- block/qcow2.h                    |  20 +++++-
- include/block/block_int.h        |   1 +
- block/qcow2.c                    | 113 +++++++++++++++++++++++++++++++
- tests/qemu-iotests/031.out       |  14 ++--
- tests/qemu-iotests/036.out       |   4 +-
- tests/qemu-iotests/049.out       | 102 ++++++++++++++--------------
- tests/qemu-iotests/060.out       |   1 +
- tests/qemu-iotests/061.out       |  34 ++++++----
- tests/qemu-iotests/065           |  28 +++++---
- tests/qemu-iotests/080           |   2 +-
- tests/qemu-iotests/082.out       |  48 +++++++++++--
- tests/qemu-iotests/085.out       |  38 +++++------
- tests/qemu-iotests/144.out       |   4 +-
- tests/qemu-iotests/182.out       |   2 +-
- tests/qemu-iotests/185.out       |   8 +--
- tests/qemu-iotests/198.out       |   2 +
- tests/qemu-iotests/206.out       |   5 ++
- tests/qemu-iotests/242.out       |   5 ++
- tests/qemu-iotests/255.out       |   8 +--
- tests/qemu-iotests/274.out       |  49 +++++++-------
- tests/qemu-iotests/280.out       |   2 +-
- tests/qemu-iotests/common.filter |   3 +-
-files changed, 365 insertions(+), 150 deletions(-)
-diff --git a/qapi/block-core.json b/qapi/block-core.json
-index XXXXXXX..XXXXXXX 100644
---- a/qapi/block-core.json
-+++ b/qapi/block-core.json
-@@ -XXX,XX +XXX,XX @@
- #
- # @bitmaps: A list of qcow2 bitmap details (since 4.0)
- #
-+# @compression-type: the image cluster compression method (since 5.1)
-+#
- # Since: 1.7
- ##
- { 'struct': 'ImageInfoSpecificQCow2',
-@@ -XXX,XX +XXX,XX @@
-       '*corrupt': 'bool',
-       'refcount-bits': 'int',
-       '*encrypt': 'ImageInfoSpecificQCow2Encryption',
--      '*bitmaps': ['Qcow2BitmapInfo']
-+      '*bitmaps': ['Qcow2BitmapInfo'],
-+      'compression-type': 'Qcow2CompressionType'
-   } }
- ##
-@@ -XXX,XX +XXX,XX @@
-   'data': [ 'v2', 'v3' ] }
-+##
-+# @Qcow2CompressionType:
-+#
-+# Compression type used in qcow2 image file
-+#
-+# @zlib: zlib compression, see <http://zlib.net/>
-+#
-+# Since: 5.1
-+##
-+{ 'enum': 'Qcow2CompressionType',
-+  'data': [ 'zlib' ] }
-+
- ##
- # @BlockdevCreateOptionsQcow2:
- #
-@@ -XXX,XX +XXX,XX @@
- #                 allowed values: off, falloc, full, metadata)
- # @lazy-refcounts: True if refcounts may be updated lazily (default: off)
- # @refcount-bits: Width of reference counts in bits (default: 16)
-+# @compression-type: The image cluster compression method
-+#                    (default: zlib, since 5.1)
- #
- # Since: 2.12
- ##
-@@ -XXX,XX +XXX,XX @@
-             '*cluster-size':    'size',
-             '*preallocation':   'PreallocMode',
-             '*lazy-refcounts':  'bool',
--            '*refcount-bits':   'int' } }
-+            '*refcount-bits':   'int',
-+            '*compression-type':'Qcow2CompressionType' } }
- ##
- # @BlockdevCreateOptionsQed:
-diff --git a/block/qcow2.h b/block/qcow2.h
-index XXXXXXX..XXXXXXX 100644
---- a/block/qcow2.h
-+++ b/block/qcow2.h
-@@ -XXX,XX +XXX,XX @@ typedef struct QCowHeader {
-     uint32_t refcount_order;
-     uint32_t header_length;
-+
-+    /* Additional fields */
-+    uint8_t compression_type;
-+
-+    /* header must be a multiple of 8 */
-+    uint8_t padding[7];
- } QEMU_PACKED QCowHeader;
-+QEMU_BUILD_BUG_ON(!QEMU_IS_ALIGNED(sizeof(QCowHeader), 8));
-+
- typedef struct QEMU_PACKED QCowSnapshotHeader {
-     /* header is 8 byte aligned */
-     uint64_t l1_table_offset;
-@@ -XXX,XX +XXX,XX @@ enum {
-     QCOW2_INCOMPAT_DIRTY_BITNR      = 0,
-     QCOW2_INCOMPAT_CORRUPT_BITNR    = 1,
-     QCOW2_INCOMPAT_DATA_FILE_BITNR  = 2,
-+    QCOW2_INCOMPAT_COMPRESSION_BITNR = 3,
-     QCOW2_INCOMPAT_DIRTY            = 1 << QCOW2_INCOMPAT_DIRTY_BITNR,
-     QCOW2_INCOMPAT_CORRUPT          = 1 << QCOW2_INCOMPAT_CORRUPT_BITNR,
-     QCOW2_INCOMPAT_DATA_FILE        = 1 << QCOW2_INCOMPAT_DATA_FILE_BITNR,
-+    QCOW2_INCOMPAT_COMPRESSION      = 1 << QCOW2_INCOMPAT_COMPRESSION_BITNR,
-     QCOW2_INCOMPAT_MASK             = QCOW2_INCOMPAT_DIRTY
-                                     | QCOW2_INCOMPAT_CORRUPT
--                                    | QCOW2_INCOMPAT_DATA_FILE,
-+                                    | QCOW2_INCOMPAT_DATA_FILE
-+                                    | QCOW2_INCOMPAT_COMPRESSION,
- };
- /* Compatible feature bits */
-@@ -XXX,XX +XXX,XX @@ typedef struct BDRVQcow2State {
-     bool metadata_preallocation_checked;
-     bool metadata_preallocation;
-+    /*
-+     * Compression type used for the image. Default: 0 - ZLIB
-+     * The image compression type is set on image creation.
-+     * For now, the only way to change the compression type
-+     * is to convert the image with the desired compression type set.
-+     */
-+    Qcow2CompressionType compression_type;
- } BDRVQcow2State;
- typedef struct Qcow2COWRegion {
 diff --git a/include/block/block_int.h b/include/block/block_int.h
 index XXXXXXX..XXXXXXX 100644
 --- a/include/block/block_int.h
 +++ b/include/block/block_int.h
-@@ -XXX,XX +XXX,XX @@
+@@ -XXX,XX +XXX,XX @@ struct BdrvChild {
- #define BLOCK_OPT_REFCOUNT_BITS     "refcount_bits"
+  * copied as well.
- #define BLOCK_OPT_DATA_FILE         "data_file"
+  */
- #define BLOCK_OPT_DATA_FILE_RAW     "data_file_raw"
+ struct BlockDriverState {
-+#define BLOCK_OPT_COMPRESSION_TYPE  "compression_type"
+-    int64_t total_sectors; /* if we are reading a disk image, give its
+-                              size in sectors */
- #define BLOCK_PROBE_BUF_SIZE        512
++    /* Protected by big QEMU lock or read-only after opening.  No special
++     * locking needed during I/O...
-diff --git a/block/qcow2.c b/block/qcow2.c
++     */
      int open_flags; /* flags used to open the file, re-used for re-open */
      bool read_only; /* if true, the media is read only */
      bool encrypted; /* if true, the media is encrypted */
@@ -XXX,XX +XXX,XX @@ struct BlockDriverState {
      bool sg;        /* if true, the device is a /dev/sg* */
      bool probed;    /* if true, format was probed rather than specified */
 -    int copy_on_read; /* if nonzero, copy read backing sectors into image.
 -                         note this is a reference count */
 -
 -    CoQueue flush_queue;            /* Serializing flush queue */
 -    bool active_flush_req;          /* Flush request in flight? */
 -    unsigned int write_gen;         /* Current data generation */
 -    unsigned int flushed_gen;       /* Flushed write generation */
 -
      BlockDriver *drv; /* NULL means no media */
      void *opaque;
@@ -XXX,XX +XXX,XX @@ struct BlockDriverState {
      BdrvChild *backing;
      BdrvChild *file;
 -    /* Callback before write request is processed */
 -    NotifierWithReturnList before_write_notifiers;
 -
 -    /* number of in-flight requests; overall and serialising */
 -    unsigned int in_flight;
 -    unsigned int serialising_in_flight;
 -
 -    bool wakeup;
 -
 -    /* Offset after the highest byte written to */
 -    uint64_t wr_highest_offset;
 -
      /* I/O Limits */
      BlockLimits bl;
@@ -XXX,XX +XXX,XX @@ struct BlockDriverState {
      QTAILQ_ENTRY(BlockDriverState) bs_list;
      /* element of the list of monitor-owned BDS */
      QTAILQ_ENTRY(BlockDriverState) monitor_list;
 -    QLIST_HEAD(, BdrvDirtyBitmap) dirty_bitmaps;
      int refcnt;
 -    QLIST_HEAD(, BdrvTrackedRequest) tracked_requests;
 -
      /* operation blockers */
      QLIST_HEAD(, BdrvOpBlocker) op_blockers[BLOCK_OP_TYPE_MAX];
@@ -XXX,XX +XXX,XX @@ struct BlockDriverState {
      /* The error object in use for blocking operations on backing_hd */
      Error *backing_blocker;
 +    /* Protected by AioContext lock */
 +
 +    /* If true, copy read backing sectors into image.  Can be >1 if more
 +     * than one client has requested copy-on-read.
 +     */
 +    int copy_on_read;
 +
 +    /* If we are reading a disk image, give its size in sectors.
 +     * Generally read-only; it is written to by load_vmstate and save_vmstate,
 +     * but the block layer is quiescent during those.
 +     */
 +    int64_t total_sectors;
 +
 +    /* Callback before write request is processed */
 +    NotifierWithReturnList before_write_notifiers;
 +
 +    /* number of in-flight requests; overall and serialising */
 +    unsigned int in_flight;
 +    unsigned int serialising_in_flight;
 +
 +    bool wakeup;
 +
 +    /* Offset after the highest byte written to */
 +    uint64_t wr_highest_offset;
 +
      /* threshold limit for writes, in bytes. "High water mark". */
      uint64_t write_threshold_offset;
      NotifierWithReturn write_threshold_notifier;
@@ -XXX,XX +XXX,XX @@ struct BlockDriverState {
      /* counter for nested bdrv_io_plug */
      unsigned io_plugged;
 +    QLIST_HEAD(, BdrvTrackedRequest) tracked_requests;
 +    CoQueue flush_queue;                  /* Serializing flush queue */
 +    bool active_flush_req;                /* Flush request in flight? */
 +    unsigned int write_gen;               /* Current data generation */
 +    unsigned int flushed_gen;             /* Flushed write generation */
 +
 +    QLIST_HEAD(, BdrvDirtyBitmap) dirty_bitmaps;
 +
 +    /* do we need to tell the quest if we have a volatile write cache? */
 +    int enable_write_cache;
 +
      int quiesce_counter;
  };
 diff --git a/include/sysemu/block-backend.h b/include/sysemu/block-backend.h
 index XXXXXXX..XXXXXXX 100644
---- a/block/qcow2.c
+--- a/include/sysemu/block-backend.h
-+++ b/block/qcow2.c
++++ b/include/sysemu/block-backend.h
-@@ -XXX,XX +XXX,XX @@ static int qcow2_update_options(BlockDriverState *bs, QDict *options,
+@@ -XXX,XX +XXX,XX @@ typedef struct BlockDevOps {
-     return ret;
+  * fields that must be public. This is in particular for QLIST_ENTRY() and
- }
+  * friends so that BlockBackends can be kept in lists outside block-backend.c */
+ typedef struct BlockBackendPublic {
-+static int validate_compression_type(BDRVQcow2State *s, Error **errp)
+-    /* I/O throttling.
-+{
+-     * throttle_state tells us if this BlockBackend has I/O limits configured.
-+    switch (s->compression_type) {
+-     * io_limits_disabled tells us if they are currently being enforced */
-+    case QCOW2_COMPRESSION_TYPE_ZLIB:
++    /* I/O throttling has its own locking, but also some fields are
-+        break;
++     * protected by the AioContext lock.
 +     */
 +
-+    default:
++    /* Protected by AioContext lock.  */
-+        error_setg(errp, "qcow2: unknown compression type: %u",
+     CoQueue      throttled_reqs[2];
 +                   s->compression_type);
 +        return -ENOTSUP;
 +    }
 +
-+    /*
++    /* Nonzero if the I/O limits are currently being ignored; generally
-+     * if the compression type differs from QCOW2_COMPRESSION_TYPE_ZLIB
++     * it is zero.  */
-+     * the incompatible feature flag must be set
+     unsigned int io_limits_disabled;
-+     */
-+    if (s->compression_type == QCOW2_COMPRESSION_TYPE_ZLIB) {
+     /* The following fields are protected by the ThrottleGroup lock.
-+        if (s->incompatible_features & QCOW2_INCOMPAT_COMPRESSION) {
+-     * See the ThrottleGroup documentation for details. */
-+            error_setg(errp, "qcow2: Compression type incompatible feature "
++     * See the ThrottleGroup documentation for details.
-+                             "bit must not be set");
++     * throttle_state tells us if I/O limits are configured. */
-+            return -EINVAL;
+     ThrottleState *throttle_state;
-+        }
+     ThrottleTimers throttle_timers;
-+    } else {
+     unsigned       pending_reqs[2];
 +        if (!(s->incompatible_features & QCOW2_INCOMPAT_COMPRESSION)) {
 +            error_setg(errp, "qcow2: Compression type incompatible feature "
 +                             "bit must be set");
 +            return -EINVAL;
 +        }
 +    }
 +
 +    return 0;
 +}
 +
  /* Called with s->lock held.  */
  static int coroutine_fn qcow2_do_open(BlockDriverState *bs, QDict *options,
                                        int flags, Error **errp)
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_do_open(BlockDriverState *bs, QDict *options,
      s->compatible_features      = header.compatible_features;
      s->autoclear_features       = header.autoclear_features;
 +    /*
 +     * Handle compression type
 +     * Older qcow2 images don't contain the compression type header.
 +     * Distinguish them by the header length and use
 +     * the only valid (default) compression type in that case
 +     */
 +    if (header.header_length > offsetof(QCowHeader, compression_type)) {
 +        s->compression_type = header.compression_type;
 +    } else {
 +        s->compression_type = QCOW2_COMPRESSION_TYPE_ZLIB;
 +    }
 +
 +    ret = validate_compression_type(s, errp);
 +    if (ret) {
 +        goto fail;
 +    }
 +
      if (s->incompatible_features & ~QCOW2_INCOMPAT_MASK) {
          void *feature_table = NULL;
          qcow2_read_extensions(bs, header.header_length, ext_end,
@@ -XXX,XX +XXX,XX @@ int qcow2_update_header(BlockDriverState *bs)
      total_size = bs->total_sectors * BDRV_SECTOR_SIZE;
      refcount_table_clusters = s->refcount_table_size >> (s->cluster_bits - 3);
 +    ret = validate_compression_type(s, NULL);
 +    if (ret) {
 +        goto fail;
 +    }
 +
      *header = (QCowHeader) {
          /* Version 2 fields */
          .magic                  = cpu_to_be32(QCOW_MAGIC),
@@ -XXX,XX +XXX,XX @@ int qcow2_update_header(BlockDriverState *bs)
          .autoclear_features     = cpu_to_be64(s->autoclear_features),
          .refcount_order         = cpu_to_be32(s->refcount_order),
          .header_length          = cpu_to_be32(header_length),
 +        .compression_type       = s->compression_type,
      };
      /* For older versions, write a shorter header */
@@ -XXX,XX +XXX,XX @@ int qcow2_update_header(BlockDriverState *bs)
                  .bit  = QCOW2_INCOMPAT_DATA_FILE_BITNR,
                  .name = "external data file",
              },
 +            {
 +                .type = QCOW2_FEAT_TYPE_INCOMPATIBLE,
 +                .bit  = QCOW2_INCOMPAT_COMPRESSION_BITNR,
 +                .name = "compression type",
 +            },
              {
                  .type = QCOW2_FEAT_TYPE_COMPATIBLE,
                  .bit  = QCOW2_COMPAT_LAZY_REFCOUNTS_BITNR,
@@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp)
      uint64_t* refcount_table;
      Error *local_err = NULL;
      int ret;
 +    uint8_t compression_type = QCOW2_COMPRESSION_TYPE_ZLIB;
      assert(create_options->driver == BLOCKDEV_DRIVER_QCOW2);
      qcow2_opts = &create_options->u.qcow2;
@@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp)
          }
      }
 +    if (qcow2_opts->has_compression_type &&
 +        qcow2_opts->compression_type != QCOW2_COMPRESSION_TYPE_ZLIB) {
 +
 +        ret = -EINVAL;
 +
 +        if (version < 3) {
 +            error_setg(errp, "Non-zlib compression type is only supported with "
 +                       "compatibility level 1.1 and above (use version=v3 or "
 +                       "greater)");
 +            goto out;
 +        }
 +
 +        switch (qcow2_opts->compression_type) {
 +        default:
 +            error_setg(errp, "Unknown compression type");
 +            goto out;
 +        }
 +
 +        compression_type = qcow2_opts->compression_type;
 +    }
 +
      /* Create BlockBackend to write to the image */
      blk = blk_new_with_bs(bs, BLK_PERM_WRITE | BLK_PERM_RESIZE, BLK_PERM_ALL,
                            errp);
@@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp)
          .refcount_table_offset      = cpu_to_be64(cluster_size),
          .refcount_table_clusters    = cpu_to_be32(1),
          .refcount_order             = cpu_to_be32(refcount_order),
 +        /* don't deal with endianness since compression_type is 1 byte long */
 +        .compression_type           = compression_type,
          .header_length              = cpu_to_be32(sizeof(*header)),
      };
@@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp)
          header->autoclear_features |=
              cpu_to_be64(QCOW2_AUTOCLEAR_DATA_FILE_RAW);
      }
 +    if (compression_type != QCOW2_COMPRESSION_TYPE_ZLIB) {
 +        header->incompatible_features |=
 +            cpu_to_be64(QCOW2_INCOMPAT_COMPRESSION);
 +    }
      ret = blk_pwrite(blk, 0, header, cluster_size, 0);
      g_free(header);
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_create_opts(BlockDriver *drv,
          { BLOCK_OPT_ENCRYPT,            BLOCK_OPT_ENCRYPT_FORMAT },
          { BLOCK_OPT_COMPAT_LEVEL,       "version" },
          { BLOCK_OPT_DATA_FILE_RAW,      "data-file-raw" },
 +        { BLOCK_OPT_COMPRESSION_TYPE,   "compression-type" },
          { NULL, NULL },
      };
@@ -XXX,XX +XXX,XX @@ static ImageInfoSpecific *qcow2_get_specific_info(BlockDriverState *bs,
              .data_file          = g_strdup(s->image_data_file),
              .has_data_file_raw  = has_data_file(bs),
              .data_file_raw      = data_file_is_raw(bs),
 +            .compression_type   = s->compression_type,
          };
      } else {
          /* if this assertion fails, this probably means a new version was
@@ -XXX,XX +XXX,XX @@ static int qcow2_amend_options(BlockDriverState *bs, QemuOpts *opts,
                                   "images");
                  return -EINVAL;
              }
 +        } else if (!strcmp(desc->name, BLOCK_OPT_COMPRESSION_TYPE)) {
 +            const char *ct_name =
 +                qemu_opt_get(opts, BLOCK_OPT_COMPRESSION_TYPE);
 +            int compression_type =
 +                qapi_enum_parse(&Qcow2CompressionType_lookup, ct_name, -1,
 +                                NULL);
 +            if (compression_type == -1) {
 +                error_setg(errp, "Unknown compression type: %s", ct_name);
 +                return -ENOTSUP;
 +            }
 +
 +            if (compression_type != s->compression_type) {
 +                error_setg(errp, "Changing the compression type "
 +                                 "is not supported");
 +                return -ENOTSUP;
 +            }
          } else {
              /* if this point is reached, this probably means a new option was
               * added without having it covered here */
@@ -XXX,XX +XXX,XX @@ static QemuOptsList qcow2_create_opts = {
              .help = "Width of a reference count entry in bits",
              .def_value_str = "16"
          },
 +        {
 +            .name = BLOCK_OPT_COMPRESSION_TYPE,
 +            .type = QEMU_OPT_STRING,
 +            .help = "Compression method used for image cluster compression",
 +            .def_value_str = "zlib"
 +        },
          { /* end of list */ }
      }
  };
 diff --git a/tests/qemu-iotests/031.out b/tests/qemu-iotests/031.out
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/031.out
 +++ b/tests/qemu-iotests/031.out
@@ -XXX,XX +XXX,XX @@ incompatible_features     []
  compatible_features       []
  autoclear_features        []
  refcount_order            4
 -header_length             104
 +header_length             112
  Header extension:
  magic                     0x6803f857
 -length                    288
 +length                    336
  data                      <binary>
  Header extension:
@@ -XXX,XX +XXX,XX @@ incompatible_features     []
  compatible_features       []
  autoclear_features        []
  refcount_order            4
 -header_length             104
 +header_length             112
  Header extension:
  magic                     0x6803f857
 -length                    288
 +length                    336
  data                      <binary>
  Header extension:
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  magic                     0x514649fb
  version                   3
 -backing_file_offset       0x1d8
 +backing_file_offset       0x210
  backing_file_size         0x17
  cluster_bits              16
  size                      67108864
@@ -XXX,XX +XXX,XX @@ incompatible_features     []
  compatible_features       []
  autoclear_features        []
  refcount_order            4
 -header_length             104
 +header_length             112
  Header extension:
  magic                     0xe2792aca
@@ -XXX,XX +XXX,XX @@ data                      'host_device'
  Header extension:
  magic                     0x6803f857
 -length                    288
 +length                    336
  data                      <binary>
  Header extension:
 diff --git a/tests/qemu-iotests/036.out b/tests/qemu-iotests/036.out
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/036.out
 +++ b/tests/qemu-iotests/036.out
@@ -XXX,XX +XXX,XX @@ compatible_features       []
  autoclear_features        [63]
  Header extension:
  magic                     0x6803f857
 -length                    288
 +length                    336
  data                      <binary>
@@ -XXX,XX +XXX,XX @@ compatible_features       []
  autoclear_features        []
  Header extension:
  magic                     0x6803f857
 -length                    288
 +length                    336
  data                      <binary>
  *** done
 diff --git a/tests/qemu-iotests/049.out b/tests/qemu-iotests/049.out
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/049.out
 +++ b/tests/qemu-iotests/049.out
@@ -XXX,XX +XXX,XX @@ QA output created by 049
  == 1. Traditional size parameter ==
  qemu-img create -f qcow2 TEST_DIR/t.qcow2 1024
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 TEST_DIR/t.qcow2 1024b
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 TEST_DIR/t.qcow2 1k
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 TEST_DIR/t.qcow2 1K
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 TEST_DIR/t.qcow2 1M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1048576 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1048576 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 TEST_DIR/t.qcow2 1G
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1073741824 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1073741824 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 TEST_DIR/t.qcow2 1T
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1099511627776 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1099511627776 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 TEST_DIR/t.qcow2 1024.0
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 TEST_DIR/t.qcow2 1024.0b
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 TEST_DIR/t.qcow2 1.5k
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1536 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1536 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 TEST_DIR/t.qcow2 1.5K
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1536 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1536 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 TEST_DIR/t.qcow2 1.5M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1572864 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1572864 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 TEST_DIR/t.qcow2 1.5G
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1610612736 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1610612736 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 TEST_DIR/t.qcow2 1.5T
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1649267441664 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1649267441664 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  == 2. Specifying size via -o ==
  qemu-img create -f qcow2 -o size=1024 TEST_DIR/t.qcow2
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o size=1024b TEST_DIR/t.qcow2
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o size=1k TEST_DIR/t.qcow2
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o size=1K TEST_DIR/t.qcow2
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o size=1M TEST_DIR/t.qcow2
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1048576 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1048576 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o size=1G TEST_DIR/t.qcow2
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1073741824 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1073741824 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o size=1T TEST_DIR/t.qcow2
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1099511627776 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1099511627776 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o size=1024.0 TEST_DIR/t.qcow2
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o size=1024.0b TEST_DIR/t.qcow2
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o size=1.5k TEST_DIR/t.qcow2
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1536 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1536 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o size=1.5K TEST_DIR/t.qcow2
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1536 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1536 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o size=1.5M TEST_DIR/t.qcow2
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1572864 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1572864 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o size=1.5G TEST_DIR/t.qcow2
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1610612736 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1610612736 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o size=1.5T TEST_DIR/t.qcow2
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1649267441664 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1649267441664 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  == 3. Invalid sizes ==
@@ -XXX,XX +XXX,XX @@ qemu-img: TEST_DIR/t.qcow2: The image size must be specified only once
  == Check correct interpretation of suffixes for cluster size ==
  qemu-img create -f qcow2 -o cluster_size=1024 TEST_DIR/t.qcow2 64M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o cluster_size=1024b TEST_DIR/t.qcow2 64M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o cluster_size=1k TEST_DIR/t.qcow2 64M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o cluster_size=1K TEST_DIR/t.qcow2 64M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o cluster_size=1M TEST_DIR/t.qcow2 64M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1048576 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1048576 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o cluster_size=1024.0 TEST_DIR/t.qcow2 64M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o cluster_size=1024.0b TEST_DIR/t.qcow2 64M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o cluster_size=0.5k TEST_DIR/t.qcow2 64M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=512 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=512 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o cluster_size=0.5K TEST_DIR/t.qcow2 64M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=512 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=512 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o cluster_size=0.5M TEST_DIR/t.qcow2 64M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=524288 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=524288 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  == Check compat level option ==
  qemu-img create -f qcow2 -o compat=0.10 TEST_DIR/t.qcow2 64M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=0.10 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=0.10 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o compat=1.1 TEST_DIR/t.qcow2 64M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=1.1 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=1.1 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o compat=0.42 TEST_DIR/t.qcow2 64M
  qemu-img: TEST_DIR/t.qcow2: Invalid parameter '0.42'
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=0.42 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=0.42 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o compat=foobar TEST_DIR/t.qcow2 64M
  qemu-img: TEST_DIR/t.qcow2: Invalid parameter 'foobar'
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=foobar cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=foobar cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  == Check preallocation option ==
  qemu-img create -f qcow2 -o preallocation=off TEST_DIR/t.qcow2 64M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=65536 preallocation=off lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=65536 preallocation=off lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o preallocation=metadata TEST_DIR/t.qcow2 64M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=65536 preallocation=metadata lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=65536 preallocation=metadata lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o preallocation=1234 TEST_DIR/t.qcow2 64M
  qemu-img: TEST_DIR/t.qcow2: Invalid parameter '1234'
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=65536 preallocation=1234 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=65536 preallocation=1234 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  == Check encryption option ==
  qemu-img create -f qcow2 -o encryption=off TEST_DIR/t.qcow2 64M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 encryption=off cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 encryption=off cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 --object secret,id=sec0,data=123456 -o encryption=on,encrypt.key-secret=sec0 TEST_DIR/t.qcow2 64M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 encryption=on encrypt.key-secret=sec0 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 encryption=on encrypt.key-secret=sec0 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  == Check lazy_refcounts option (only with v3) ==
  qemu-img create -f qcow2 -o compat=1.1,lazy_refcounts=off TEST_DIR/t.qcow2 64M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=1.1 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=1.1 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o compat=1.1,lazy_refcounts=on TEST_DIR/t.qcow2 64M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=1.1 cluster_size=65536 lazy_refcounts=on refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=1.1 cluster_size=65536 lazy_refcounts=on refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o compat=0.10,lazy_refcounts=off TEST_DIR/t.qcow2 64M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=0.10 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=0.10 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  qemu-img create -f qcow2 -o compat=0.10,lazy_refcounts=on TEST_DIR/t.qcow2 64M
  qemu-img: TEST_DIR/t.qcow2: Lazy refcounts only supported with compatibility level 1.1 and above (use version=v3 or greater)
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=0.10 cluster_size=65536 lazy_refcounts=on refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=0.10 cluster_size=65536 lazy_refcounts=on refcount_bits=16 compression_type=zlib
  *** done
 diff --git a/tests/qemu-iotests/060.out b/tests/qemu-iotests/060.out
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/060.out
 +++ b/tests/qemu-iotests/060.out
@@ -XXX,XX +XXX,XX @@ virtual size: 64 MiB (67108864 bytes)
  cluster_size: 65536
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: false
      refcount bits: 16
      corrupt: true
 diff --git a/tests/qemu-iotests/061.out b/tests/qemu-iotests/061.out
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/061.out
 +++ b/tests/qemu-iotests/061.out
@@ -XXX,XX +XXX,XX @@ incompatible_features     []
  compatible_features       [0]
  autoclear_features        []
  refcount_order            4
 -header_length             104
 +header_length             112
  Header extension:
  magic                     0x6803f857
 -length                    288
 +length                    336
  data                      <binary>
  magic                     0x514649fb
@@ -XXX,XX +XXX,XX @@ incompatible_features     []
  compatible_features       [0]
  autoclear_features        []
  refcount_order            4
 -header_length             104
 +header_length             112
  Header extension:
  magic                     0x6803f857
 -length                    288
 +length                    336
  data                      <binary>
  magic                     0x514649fb
@@ -XXX,XX +XXX,XX @@ incompatible_features     [0]
  compatible_features       [0]
  autoclear_features        []
  refcount_order            4
 -header_length             104
 +header_length             112
  Header extension:
  magic                     0x6803f857
 -length                    288
 +length                    336
  data                      <binary>
  ERROR cluster 5 refcount=0 reference=1
@@ -XXX,XX +XXX,XX @@ incompatible_features     []
  compatible_features       [42]
  autoclear_features        [42]
  refcount_order            4
 -header_length             104
 +header_length             112
  Header extension:
  magic                     0x6803f857
 -length                    288
 +length                    336
  data                      <binary>
  magic                     0x514649fb
@@ -XXX,XX +XXX,XX @@ incompatible_features     []
  compatible_features       [0]
  autoclear_features        []
  refcount_order            4
 -header_length             104
 +header_length             112
  Header extension:
  magic                     0x6803f857
 -length                    288
 +length                    336
  data                      <binary>
  read 65536/65536 bytes at offset 44040192
@@ -XXX,XX +XXX,XX @@ incompatible_features     [0]
  compatible_features       [0]
  autoclear_features        []
  refcount_order            4
 -header_length             104
 +header_length             112
  Header extension:
  magic                     0x6803f857
 -length                    288
 +length                    336
  data                      <binary>
  ERROR cluster 5 refcount=0 reference=1
@@ -XXX,XX +XXX,XX @@ incompatible_features     []
  compatible_features       []
  autoclear_features        []
  refcount_order            4
 -header_length             104
 +header_length             112
  Header extension:
  magic                     0x6803f857
 -length                    288
 +length                    336
  data                      <binary>
  read 131072/131072 bytes at offset 0
@@ -XXX,XX +XXX,XX @@ virtual size: 64 MiB (67108864 bytes)
  cluster_size: 65536
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: false
      refcount bits: 16
      data file: TEST_DIR/t.IMGFMT.data
@@ -XXX,XX +XXX,XX @@ virtual size: 64 MiB (67108864 bytes)
  cluster_size: 65536
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: false
      refcount bits: 16
      data file: foo
@@ -XXX,XX +XXX,XX @@ virtual size: 64 MiB (67108864 bytes)
  cluster_size: 65536
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: false
      refcount bits: 16
      data file raw: false
@@ -XXX,XX +XXX,XX @@ virtual size: 64 MiB (67108864 bytes)
  cluster_size: 65536
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: false
      refcount bits: 16
      data file: TEST_DIR/t.IMGFMT.data
@@ -XXX,XX +XXX,XX @@ virtual size: 64 MiB (67108864 bytes)
  cluster_size: 65536
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: false
      refcount bits: 16
      data file: TEST_DIR/t.IMGFMT.data
@@ -XXX,XX +XXX,XX @@ virtual size: 64 MiB (67108864 bytes)
  cluster_size: 65536
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: false
      refcount bits: 16
      data file: TEST_DIR/t.IMGFMT.data
 diff --git a/tests/qemu-iotests/065 b/tests/qemu-iotests/065
 index XXXXXXX..XXXXXXX 100755
 --- a/tests/qemu-iotests/065
 +++ b/tests/qemu-iotests/065
@@ -XXX,XX +XXX,XX @@ class TestQMP(TestImageInfoSpecific):
  class TestQCow2(TestQemuImgInfo):
      '''Testing a qcow2 version 2 image'''
      img_options = 'compat=0.10'
 -    json_compare = { 'compat': '0.10', 'refcount-bits': 16 }
 -    human_compare = [ 'compat: 0.10', 'refcount bits: 16' ]
 +    json_compare = { 'compat': '0.10', 'refcount-bits': 16,
 +                     'compression-type': 'zlib' }
 +    human_compare = [ 'compat: 0.10', 'compression type: zlib',
 +                      'refcount bits: 16' ]
  class TestQCow3NotLazy(TestQemuImgInfo):
      '''Testing a qcow2 version 3 image with lazy refcounts disabled'''
      img_options = 'compat=1.1,lazy_refcounts=off'
      json_compare = { 'compat': '1.1', 'lazy-refcounts': False,
 -                     'refcount-bits': 16, 'corrupt': False }
 -    human_compare = [ 'compat: 1.1', 'lazy refcounts: false',
 -                      'refcount bits: 16', 'corrupt: false' ]
 +                     'refcount-bits': 16, 'corrupt': False,
 +                     'compression-type': 'zlib' }
 +    human_compare = [ 'compat: 1.1', 'compression type: zlib',
 +                      'lazy refcounts: false', 'refcount bits: 16',
 +                      'corrupt: false' ]
  class TestQCow3Lazy(TestQemuImgInfo):
      '''Testing a qcow2 version 3 image with lazy refcounts enabled'''
      img_options = 'compat=1.1,lazy_refcounts=on'
      json_compare = { 'compat': '1.1', 'lazy-refcounts': True,
 -                     'refcount-bits': 16, 'corrupt': False }
 -    human_compare = [ 'compat: 1.1', 'lazy refcounts: true',
 -                      'refcount bits: 16', 'corrupt: false' ]
 +                     'refcount-bits': 16, 'corrupt': False,
 +                     'compression-type': 'zlib' }
 +    human_compare = [ 'compat: 1.1', 'compression type: zlib',
 +                      'lazy refcounts: true', 'refcount bits: 16',
 +                      'corrupt: false' ]
  class TestQCow3NotLazyQMP(TestQMP):
      '''Testing a qcow2 version 3 image with lazy refcounts disabled, opening
@@ -XXX,XX +XXX,XX @@ class TestQCow3NotLazyQMP(TestQMP):
      img_options = 'compat=1.1,lazy_refcounts=off'
      qemu_options = 'lazy-refcounts=on'
      compare = { 'compat': '1.1', 'lazy-refcounts': False,
 -                'refcount-bits': 16, 'corrupt': False }
 +                'refcount-bits': 16, 'corrupt': False,
 +                'compression-type': 'zlib' }
  class TestQCow3LazyQMP(TestQMP):
@@ -XXX,XX +XXX,XX @@ class TestQCow3LazyQMP(TestQMP):
      img_options = 'compat=1.1,lazy_refcounts=on'
      qemu_options = 'lazy-refcounts=off'
      compare = { 'compat': '1.1', 'lazy-refcounts': True,
 -                'refcount-bits': 16, 'corrupt': False }
 +                'refcount-bits': 16, 'corrupt': False,
 +                'compression-type': 'zlib' }
  TestImageInfoSpecific = None
  TestQemuImgInfo = None
 diff --git a/tests/qemu-iotests/080 b/tests/qemu-iotests/080
 index XXXXXXX..XXXXXXX 100755
 --- a/tests/qemu-iotests/080
 +++ b/tests/qemu-iotests/080
@@ -XXX,XX +XXX,XX @@ _supported_os Linux
  # - This is generally a test for compat=1.1 images
  _unsupported_imgopts 'refcount_bits=1[^0-9]' data_file 'compat=0.10'
 -header_size=104
 +header_size=112
  offset_backing_file_offset=8
  offset_backing_file_size=16
 diff --git a/tests/qemu-iotests/082.out b/tests/qemu-iotests/082.out
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/082.out
 +++ b/tests/qemu-iotests/082.out
@@ -XXX,XX +XXX,XX @@ QA output created by 082
  === create: Options specified more than once ===
  Testing: create -f foo -f qcow2 TEST_DIR/t.qcow2 128M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  image: TEST_DIR/t.IMGFMT
  file format: IMGFMT
  virtual size: 128 MiB (134217728 bytes)
  cluster_size: 65536
  Testing: create -f qcow2 -o cluster_size=4k -o lazy_refcounts=on TEST_DIR/t.qcow2 128M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=4096 lazy_refcounts=on refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=4096 lazy_refcounts=on refcount_bits=16 compression_type=zlib
  image: TEST_DIR/t.IMGFMT
  file format: IMGFMT
  virtual size: 128 MiB (134217728 bytes)
  cluster_size: 4096
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: true
      refcount bits: 16
      corrupt: false
  Testing: create -f qcow2 -o cluster_size=4k -o lazy_refcounts=on -o cluster_size=8k TEST_DIR/t.qcow2 128M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=8192 lazy_refcounts=on refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=8192 lazy_refcounts=on refcount_bits=16 compression_type=zlib
  image: TEST_DIR/t.IMGFMT
  file format: IMGFMT
  virtual size: 128 MiB (134217728 bytes)
  cluster_size: 8192
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: true
      refcount bits: 16
      corrupt: false
  Testing: create -f qcow2 -o cluster_size=4k,cluster_size=8k TEST_DIR/t.qcow2 128M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=8192 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=8192 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  image: TEST_DIR/t.IMGFMT
  file format: IMGFMT
  virtual size: 128 MiB (134217728 bytes)
@@ -XXX,XX +XXX,XX @@ Supported options:
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
    size=<size>            - Virtual disk size
  Testing: create -f qcow2 -u -o backing_file=TEST_DIR/t.qcow2,,help TEST_DIR/t.qcow2 128M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/t.qcow2,,help cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/t.qcow2,,help cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  Testing: create -f qcow2 -u -o backing_file=TEST_DIR/t.qcow2,,? TEST_DIR/t.qcow2 128M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/t.qcow2,,? cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/t.qcow2,,? cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  Testing: create -f qcow2 -o backing_file=TEST_DIR/t.qcow2, -o help TEST_DIR/t.qcow2 128M
  qemu-img: Invalid option list: backing_file=TEST_DIR/t.qcow2,
@@ -XXX,XX +XXX,XX @@ Supported qcow2 options:
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ qemu-img: Format driver 'bochs' does not support image creation
  === convert: Options specified more than once ===
  Testing: create -f qcow2 TEST_DIR/t.qcow2 128M
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  Testing: convert -f foo -f qcow2 TEST_DIR/t.qcow2 TEST_DIR/t.qcow2.base
  image: TEST_DIR/t.IMGFMT.base
@@ -XXX,XX +XXX,XX @@ virtual size: 128 MiB (134217728 bytes)
  cluster_size: 4096
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: true
      refcount bits: 16
      corrupt: false
@@ -XXX,XX +XXX,XX @@ virtual size: 128 MiB (134217728 bytes)
  cluster_size: 8192
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: true
      refcount bits: 16
      corrupt: false
@@ -XXX,XX +XXX,XX @@ Supported options:
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported qcow2 options:
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ virtual size: 128 MiB (134217728 bytes)
  cluster_size: 65536
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: true
      refcount bits: 16
      corrupt: false
@@ -XXX,XX +XXX,XX @@ virtual size: 130 MiB (136314880 bytes)
  cluster_size: 65536
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: false
      refcount bits: 16
      corrupt: false
@@ -XXX,XX +XXX,XX @@ virtual size: 132 MiB (138412032 bytes)
  cluster_size: 65536
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: true
      refcount bits: 16
      corrupt: false
@@ -XXX,XX +XXX,XX @@ Creation options for 'qcow2':
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Creation options for 'qcow2':
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Creation options for 'qcow2':
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Creation options for 'qcow2':
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Creation options for 'qcow2':
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Creation options for 'qcow2':
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Creation options for 'qcow2':
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Creation options for 'qcow2':
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Creation options for 'qcow2':
    backing_fmt=<str>      - Image format of the base image
    cluster_size=<size>    - qcow2 cluster size
    compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
 +  compression_type=<str> - Compression method used for image cluster compression
    data_file=<str>        - File name of an external data file
    data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
    encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
 diff --git a/tests/qemu-iotests/085.out b/tests/qemu-iotests/085.out
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/085.out
 +++ b/tests/qemu-iotests/085.out
@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/t.IMGFMT.2', fmt=IMGFMT size=134217728
  === Create a single snapshot on virtio0 ===
  { 'execute': 'blockdev-snapshot-sync', 'arguments': { 'device': 'virtio0', 'snapshot-file':'TEST_DIR/1-snapshot-v0.IMGFMT', 'format': 'IMGFMT' } }
 -Formatting 'TEST_DIR/1-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/t.qcow2.1 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/1-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/t.qcow2.1 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  {"return": {}}
  === Invalid command - missing device and nodename ===
@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/1-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file
  === Create several transactional group snapshots ===
  { 'execute': 'transaction', 'arguments': {'actions': [ { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio0', 'snapshot-file': 'TEST_DIR/2-snapshot-v0.IMGFMT' } }, { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio1', 'snapshot-file': 'TEST_DIR/2-snapshot-v1.IMGFMT' } } ] } }
 -Formatting 'TEST_DIR/2-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/1-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 -Formatting 'TEST_DIR/2-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/t.qcow2.2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/2-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/1-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 +Formatting 'TEST_DIR/2-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/t.qcow2.2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  {"return": {}}
  { 'execute': 'transaction', 'arguments': {'actions': [ { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio0', 'snapshot-file': 'TEST_DIR/3-snapshot-v0.IMGFMT' } }, { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio1', 'snapshot-file': 'TEST_DIR/3-snapshot-v1.IMGFMT' } } ] } }
 -Formatting 'TEST_DIR/3-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/2-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 -Formatting 'TEST_DIR/3-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/2-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/3-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/2-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 +Formatting 'TEST_DIR/3-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/2-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  {"return": {}}
  { 'execute': 'transaction', 'arguments': {'actions': [ { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio0', 'snapshot-file': 'TEST_DIR/4-snapshot-v0.IMGFMT' } }, { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio1', 'snapshot-file': 'TEST_DIR/4-snapshot-v1.IMGFMT' } } ] } }
 -Formatting 'TEST_DIR/4-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/3-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 -Formatting 'TEST_DIR/4-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/3-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/4-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/3-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 +Formatting 'TEST_DIR/4-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/3-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  {"return": {}}
  { 'execute': 'transaction', 'arguments': {'actions': [ { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio0', 'snapshot-file': 'TEST_DIR/5-snapshot-v0.IMGFMT' } }, { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio1', 'snapshot-file': 'TEST_DIR/5-snapshot-v1.IMGFMT' } } ] } }
 -Formatting 'TEST_DIR/5-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/4-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 -Formatting 'TEST_DIR/5-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/4-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/5-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/4-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 +Formatting 'TEST_DIR/5-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/4-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  {"return": {}}
  { 'execute': 'transaction', 'arguments': {'actions': [ { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio0', 'snapshot-file': 'TEST_DIR/6-snapshot-v0.IMGFMT' } }, { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio1', 'snapshot-file': 'TEST_DIR/6-snapshot-v1.IMGFMT' } } ] } }
 -Formatting 'TEST_DIR/6-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/5-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 -Formatting 'TEST_DIR/6-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/5-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/6-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/5-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 +Formatting 'TEST_DIR/6-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/5-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  {"return": {}}
  { 'execute': 'transaction', 'arguments': {'actions': [ { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio0', 'snapshot-file': 'TEST_DIR/7-snapshot-v0.IMGFMT' } }, { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio1', 'snapshot-file': 'TEST_DIR/7-snapshot-v1.IMGFMT' } } ] } }
 -Formatting 'TEST_DIR/7-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/6-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 -Formatting 'TEST_DIR/7-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/6-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/7-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/6-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 +Formatting 'TEST_DIR/7-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/6-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  {"return": {}}
  { 'execute': 'transaction', 'arguments': {'actions': [ { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio0', 'snapshot-file': 'TEST_DIR/8-snapshot-v0.IMGFMT' } }, { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio1', 'snapshot-file': 'TEST_DIR/8-snapshot-v1.IMGFMT' } } ] } }
 -Formatting 'TEST_DIR/8-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/7-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 -Formatting 'TEST_DIR/8-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/7-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/8-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/7-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 +Formatting 'TEST_DIR/8-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/7-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  {"return": {}}
  { 'execute': 'transaction', 'arguments': {'actions': [ { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio0', 'snapshot-file': 'TEST_DIR/9-snapshot-v0.IMGFMT' } }, { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio1', 'snapshot-file': 'TEST_DIR/9-snapshot-v1.IMGFMT' } } ] } }
 -Formatting 'TEST_DIR/9-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/8-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 -Formatting 'TEST_DIR/9-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/8-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/9-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/8-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 +Formatting 'TEST_DIR/9-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/8-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  {"return": {}}
  { 'execute': 'transaction', 'arguments': {'actions': [ { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio0', 'snapshot-file': 'TEST_DIR/10-snapshot-v0.IMGFMT' } }, { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio1', 'snapshot-file': 'TEST_DIR/10-snapshot-v1.IMGFMT' } } ] } }
 -Formatting 'TEST_DIR/10-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/9-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 -Formatting 'TEST_DIR/10-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/9-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/10-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/9-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 +Formatting 'TEST_DIR/10-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/9-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  {"return": {}}
  === Create a couple of snapshots using blockdev-snapshot ===
 diff --git a/tests/qemu-iotests/144.out b/tests/qemu-iotests/144.out
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/144.out
 +++ b/tests/qemu-iotests/144.out
@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=536870912
  { 'execute': 'qmp_capabilities' }
  {"return": {}}
  { 'execute': 'blockdev-snapshot-sync', 'arguments': { 'device': 'virtio0', 'snapshot-file':'TEST_DIR/tmp.IMGFMT', 'format': 'IMGFMT' } }
 -Formatting 'TEST_DIR/tmp.qcow2', fmt=qcow2 size=536870912 backing_file=TEST_DIR/t.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/tmp.qcow2', fmt=qcow2 size=536870912 backing_file=TEST_DIR/t.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  {"return": {}}
  === Performing block-commit on active layer ===
@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/tmp.qcow2', fmt=qcow2 size=536870912 backing_file=TEST_DIR/
  === Performing Live Snapshot 2 ===
  { 'execute': 'blockdev-snapshot-sync', 'arguments': { 'device': 'virtio0', 'snapshot-file':'TEST_DIR/tmp2.IMGFMT', 'format': 'IMGFMT' } }
 -Formatting 'TEST_DIR/tmp2.qcow2', fmt=qcow2 size=536870912 backing_file=TEST_DIR/t.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/tmp2.qcow2', fmt=qcow2 size=536870912 backing_file=TEST_DIR/t.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  {"return": {}}
  *** done
 diff --git a/tests/qemu-iotests/182.out b/tests/qemu-iotests/182.out
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/182.out
 +++ b/tests/qemu-iotests/182.out
@@ -XXX,XX +XXX,XX @@ Is another process using the image [TEST_DIR/t.qcow2]?
  {'execute': 'blockdev-add', 'arguments': { 'node-name': 'node0', 'driver': 'file', 'filename': 'TEST_DIR/t.IMGFMT', 'locking': 'on' } }
  {"return": {}}
  {'execute': 'blockdev-snapshot-sync', 'arguments': { 'node-name': 'node0', 'snapshot-file': 'TEST_DIR/t.IMGFMT.overlay', 'snapshot-node-name': 'node1' } }
 -Formatting 'TEST_DIR/t.qcow2.overlay', fmt=qcow2 size=197120 backing_file=TEST_DIR/t.qcow2 backing_fmt=file cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2.overlay', fmt=qcow2 size=197120 backing_file=TEST_DIR/t.qcow2 backing_fmt=file cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  {"return": {}}
  {'execute': 'blockdev-add', 'arguments': { 'node-name': 'node1', 'driver': 'file', 'filename': 'TEST_DIR/t.IMGFMT', 'locking': 'on' } }
  {"return": {}}
 diff --git a/tests/qemu-iotests/185.out b/tests/qemu-iotests/185.out
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/185.out
 +++ b/tests/qemu-iotests/185.out
@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/t.IMGFMT.base', fmt=IMGFMT size=67108864
  === Creating backing chain ===
  { 'execute': 'blockdev-snapshot-sync', 'arguments': { 'device': 'disk', 'snapshot-file': 'TEST_DIR/t.IMGFMT.mid', 'format': 'IMGFMT', 'mode': 'absolute-paths' } }
 -Formatting 'TEST_DIR/t.qcow2.mid', fmt=qcow2 size=67108864 backing_file=TEST_DIR/t.qcow2.base backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2.mid', fmt=qcow2 size=67108864 backing_file=TEST_DIR/t.qcow2.base backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  {"return": {}}
  { 'execute': 'human-monitor-command', 'arguments': { 'command-line': 'qemu-io disk "write 0 4M"' } }
  wrote 4194304/4194304 bytes at offset 0
 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
  {"return": ""}
  { 'execute': 'blockdev-snapshot-sync', 'arguments': { 'device': 'disk', 'snapshot-file': 'TEST_DIR/t.IMGFMT', 'format': 'IMGFMT', 'mode': 'absolute-paths' } }
 -Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 backing_file=TEST_DIR/t.qcow2.mid backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 backing_file=TEST_DIR/t.qcow2.mid backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  {"return": {}}
  === Start commit job and exit qemu ===
@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 backing_file=TEST_DIR/t.q
  { 'execute': 'qmp_capabilities' }
  {"return": {}}
  { 'execute': 'drive-mirror', 'arguments': { 'device': 'disk', 'target': 'TEST_DIR/t.IMGFMT.copy', 'format': 'IMGFMT', 'sync': 'full', 'speed': 65536 } }
 -Formatting 'TEST_DIR/t.qcow2.copy', fmt=qcow2 size=67108864 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2.copy', fmt=qcow2 size=67108864 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "disk"}}
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "disk"}}
  {"return": {}}
@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/t.qcow2.copy', fmt=qcow2 size=67108864 cluster_size=65536 l
  { 'execute': 'qmp_capabilities' }
  {"return": {}}
  { 'execute': 'drive-backup', 'arguments': { 'device': 'disk', 'target': 'TEST_DIR/t.IMGFMT.copy', 'format': 'IMGFMT', 'sync': 'full', 'speed': 65536 } }
 -Formatting 'TEST_DIR/t.qcow2.copy', fmt=qcow2 size=67108864 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/t.qcow2.copy', fmt=qcow2 size=67108864 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "disk"}}
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "disk"}}
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "paused", "id": "disk"}}
 diff --git a/tests/qemu-iotests/198.out b/tests/qemu-iotests/198.out
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/198.out
 +++ b/tests/qemu-iotests/198.out
@@ -XXX,XX +XXX,XX @@ image: json:{ /* filtered */ }
  file format: IMGFMT
  virtual size: 16 MiB (16777216 bytes)
  Format specific information:
 +    compression type: zlib
      encrypt:
          ivgen alg: plain64
          hash alg: sha256
@@ -XXX,XX +XXX,XX @@ file format: IMGFMT
  virtual size: 16 MiB (16777216 bytes)
  backing file: TEST_DIR/t.IMGFMT.base
  Format specific information:
 +    compression type: zlib
      encrypt:
          ivgen alg: plain64
          hash alg: sha256
 diff --git a/tests/qemu-iotests/206.out b/tests/qemu-iotests/206.out
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/206.out
 +++ b/tests/qemu-iotests/206.out
@@ -XXX,XX +XXX,XX @@ virtual size: 128 MiB (134217728 bytes)
  cluster_size: 65536
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: false
      refcount bits: 16
      corrupt: false
@@ -XXX,XX +XXX,XX @@ virtual size: 64 MiB (67108864 bytes)
  cluster_size: 65536
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: false
      refcount bits: 16
      corrupt: false
@@ -XXX,XX +XXX,XX @@ virtual size: 32 MiB (33554432 bytes)
  cluster_size: 2097152
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: true
      refcount bits: 1
      corrupt: false
@@ -XXX,XX +XXX,XX @@ backing file: TEST_IMG.base
  backing file format: IMGFMT
  Format specific information:
      compat: 0.10
 +    compression type: zlib
      refcount bits: 16
  === Successful image creation (encrypted) ===
@@ -XXX,XX +XXX,XX @@ encrypted: yes
  cluster_size: 65536
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: false
      refcount bits: 16
      encrypt:
 diff --git a/tests/qemu-iotests/242.out b/tests/qemu-iotests/242.out
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/242.out
 +++ b/tests/qemu-iotests/242.out
@@ -XXX,XX +XXX,XX @@ virtual size: 1 MiB (1048576 bytes)
  cluster_size: 65536
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: false
      refcount bits: 16
      corrupt: false
@@ -XXX,XX +XXX,XX @@ virtual size: 1 MiB (1048576 bytes)
  cluster_size: 65536
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: false
      bitmaps:
          [0]:
@@ -XXX,XX +XXX,XX @@ virtual size: 1 MiB (1048576 bytes)
  cluster_size: 65536
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: false
      bitmaps:
          [0]:
@@ -XXX,XX +XXX,XX @@ virtual size: 1 MiB (1048576 bytes)
  cluster_size: 65536
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: false
      bitmaps:
          [0]:
@@ -XXX,XX +XXX,XX @@ virtual size: 1 MiB (1048576 bytes)
  cluster_size: 65536
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: false
      bitmaps:
          [0]:
 diff --git a/tests/qemu-iotests/255.out b/tests/qemu-iotests/255.out
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/255.out
 +++ b/tests/qemu-iotests/255.out
@@ -XXX,XX +XXX,XX @@ Finishing a commit job with background reads
  === Create backing chain and start VM ===
 -Formatting 'TEST_DIR/PID-t.qcow2.mid', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-t.qcow2.mid', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 -Formatting 'TEST_DIR/PID-t.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-t.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  === Start background read requests ===
@@ -XXX,XX +XXX,XX @@ Closing the VM while a job is being cancelled
  === Create images and start VM ===
 -Formatting 'TEST_DIR/PID-src.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-src.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 -Formatting 'TEST_DIR/PID-dst.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-dst.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  wrote 1048576/1048576 bytes at offset 0
 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 diff --git a/tests/qemu-iotests/274.out b/tests/qemu-iotests/274.out
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/274.out
 +++ b/tests/qemu-iotests/274.out
@@ -XXX,XX +XXX,XX @@
  == Commit tests ==
 -Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=2097152 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=2097152 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 -Formatting 'TEST_DIR/PID-mid', fmt=qcow2 size=1048576 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-mid', fmt=qcow2 size=1048576 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 -Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=2097152 backing_file=TEST_DIR/PID-mid cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=2097152 backing_file=TEST_DIR/PID-mid cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  wrote 2097152/2097152 bytes at offset 0
 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ cluster_size: 65536
  backing file: TEST_DIR/PID-base
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: false
      refcount bits: 16
      corrupt: false
@@ -XXX,XX +XXX,XX @@ read 1048576/1048576 bytes at offset 1048576
 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
  === Testing HMP commit (top -> mid) ===
 -Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=2097152 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=2097152 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 -Formatting 'TEST_DIR/PID-mid', fmt=qcow2 size=1048576 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-mid', fmt=qcow2 size=1048576 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 -Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=2097152 backing_file=TEST_DIR/PID-mid cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=2097152 backing_file=TEST_DIR/PID-mid cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  wrote 2097152/2097152 bytes at offset 0
 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ cluster_size: 65536
  backing file: TEST_DIR/PID-base
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: false
      refcount bits: 16
      corrupt: false
@@ -XXX,XX +XXX,XX @@ read 1048576/1048576 bytes at offset 1048576
 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
  === Testing QMP active commit (top -> mid) ===
 -Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=2097152 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=2097152 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 -Formatting 'TEST_DIR/PID-mid', fmt=qcow2 size=1048576 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-mid', fmt=qcow2 size=1048576 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 -Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=2097152 backing_file=TEST_DIR/PID-mid cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=2097152 backing_file=TEST_DIR/PID-mid cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  wrote 2097152/2097152 bytes at offset 0
 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ cluster_size: 65536
  backing file: TEST_DIR/PID-base
  Format specific information:
      compat: 1.1
 +    compression type: zlib
      lazy refcounts: false
      refcount bits: 16
      corrupt: false
@@ -XXX,XX +XXX,XX @@ read 1048576/1048576 bytes at offset 1048576
  == Resize tests ==
  === preallocation=off ===
 -Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=6442450944 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=6442450944 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 -Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=1073741824 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=1073741824 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  wrote 65536/65536 bytes at offset 5368709120
 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ read 65536/65536 bytes at offset 5368709120
  { "start": 1073741824, "length": 7516192768, "depth": 0, "zero": true, "data": false}]
  === preallocation=metadata ===
 -Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=34359738368 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=34359738368 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 -Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=32212254720 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=32212254720 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  wrote 65536/65536 bytes at offset 33285996544
 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ read 65536/65536 bytes at offset 33285996544
  { "start": 34896609280, "length": 536870912, "depth": 0, "zero": true, "data": false, "offset": 2685075456}]
  === preallocation=falloc ===
 -Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=10485760 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=10485760 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 -Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=5242880 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=5242880 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  wrote 65536/65536 bytes at offset 9437184
 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ read 65536/65536 bytes at offset 9437184
  { "start": 5242880, "length": 10485760, "depth": 0, "zero": false, "data": true, "offset": 327680}]
  === preallocation=full ===
 -Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=16777216 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=16777216 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 -Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=8388608 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=8388608 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  wrote 65536/65536 bytes at offset 11534336
 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ read 65536/65536 bytes at offset 11534336
  { "start": 8388608, "length": 4194304, "depth": 0, "zero": false, "data": true, "offset": 327680}]
  === preallocation=off ===
 -Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=393216 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=393216 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 -Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=259072 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=259072 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  wrote 65536/65536 bytes at offset 259072
 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ read 65536/65536 bytes at offset 259072
  { "start": 262144, "length": 262144, "depth": 0, "zero": true, "data": false}]
  === preallocation=off ===
 -Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=409600 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=409600 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 -Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=262144 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=262144 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  wrote 65536/65536 bytes at offset 344064
 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ read 65536/65536 bytes at offset 344064
  { "start": 262144, "length": 262144, "depth": 0, "zero": true, "data": false}]
  === preallocation=off ===
 -Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=524288 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=524288 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 -Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=262144 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=262144 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  wrote 65536/65536 bytes at offset 446464
 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 diff --git a/tests/qemu-iotests/280.out b/tests/qemu-iotests/280.out
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/280.out
 +++ b/tests/qemu-iotests/280.out
@@ -XXX,XX +XXX,XX @@
 -Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=67108864 cluster_size=65536 lazy_refcounts=off refcount_bits=16
 +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=67108864 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
  === Launch VM ===
  Enabling migration QMP events on VM...
 diff --git a/tests/qemu-iotests/common.filter b/tests/qemu-iotests/common.filter
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/common.filter
 +++ b/tests/qemu-iotests/common.filter
@@ -XXX,XX +XXX,XX @@ _filter_img_create()
          -e "s# refcount_bits=[0-9]\\+##g" \
          -e "s# key-secret=[a-zA-Z0-9]\\+##g" \
          -e "s# iter-time=[0-9]\\+##g" \
 -        -e "s# force_size=\\(on\\|off\\)##g"
 +        -e "s# force_size=\\(on\\|off\\)##g" \
 +        -e "s# compression_type=[a-zA-Z0-9]\\+##g"
  }
  _filter_img_info()
 --
-.26.2
+.9.3

-New patch
+[Qemu-devel] [PULL v2 19/24] coroutine-lock: make CoMutex thread-safe
+From: Paolo Bonzini <pbonzini@redhat.com>
 This uses the lock-free mutex described in the paper '"Blocking without
 Locking", or LFTHREADS: A lock-free thread library' by Gidenstam and
 Papatriantafilou.  The same technique is used in OSv, and in fact
 the code is essentially a conversion to C of OSv's code.
 [Added missing coroutine_fn in tests/test-aio-multithread.c.
 --Stefan]
 Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
 Reviewed-by: Fam Zheng <famz@redhat.com>
 Message-id: 20170213181244.16297-2-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
  include/qemu/coroutine.h     |  17 ++++-
  tests/test-aio-multithread.c |  86 ++++++++++++++++++++++++
  util/qemu-coroutine-lock.c   | 155 ++++++++++++++++++++++++++++++++++++++++---
  util/trace-events            |   1 +
 files changed, 246 insertions(+), 13 deletions(-)
 diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
 index XXXXXXX..XXXXXXX 100644
 --- a/include/qemu/coroutine.h
 +++ b/include/qemu/coroutine.h
@@ -XXX,XX +XXX,XX @@ bool qemu_co_queue_empty(CoQueue *queue);
  /**
   * Provides a mutex that can be used to synchronise coroutines
   */
 +struct CoWaitRecord;
  typedef struct CoMutex {
 -    bool locked;
 +    /* Count of pending lockers; 0 for a free mutex, 1 for an
 +     * uncontended mutex.
 +     */
 +    unsigned locked;
 +
 +    /* A queue of waiters.  Elements are added atomically in front of
 +     * from_push.  to_pop is only populated, and popped from, by whoever
 +     * is in charge of the next wakeup.  This can be an unlocker or,
 +     * through the handoff protocol, a locker that is about to go to sleep.
 +     */
 +    QSLIST_HEAD(, CoWaitRecord) from_push, to_pop;
 +
 +    unsigned handoff, sequence;
 +
      Coroutine *holder;
 -    CoQueue queue;
  } CoMutex;
  /**
 diff --git a/tests/test-aio-multithread.c b/tests/test-aio-multithread.c
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/test-aio-multithread.c
 +++ b/tests/test-aio-multithread.c
@@ -XXX,XX +XXX,XX @@ static void test_multi_co_schedule_10(void)
      test_multi_co_schedule(10);
  }
 +/* CoMutex thread-safety.  */
 +
 +static uint32_t atomic_counter;
 +static uint32_t running;
 +static uint32_t counter;
 +static CoMutex comutex;
 +
 +static void coroutine_fn test_multi_co_mutex_entry(void *opaque)
 +{
 +    while (!atomic_mb_read(&now_stopping)) {
 +        qemu_co_mutex_lock(&comutex);
 +        counter++;
 +        qemu_co_mutex_unlock(&comutex);
 +
 +        /* Increase atomic_counter *after* releasing the mutex.  Otherwise
 +         * there is a chance (it happens about 1 in 3 runs) that the iothread
 +         * exits before the coroutine is woken up, causing a spurious
 +         * assertion failure.
 +         */
 +        atomic_inc(&atomic_counter);
 +    }
 +    atomic_dec(&running);
 +}
 +
 +static void test_multi_co_mutex(int threads, int seconds)
 +{
 +    int i;
 +
 +    qemu_co_mutex_init(&comutex);
 +    counter = 0;
 +    atomic_counter = 0;
 +    now_stopping = false;
 +
 +    create_aio_contexts();
 +    assert(threads <= NUM_CONTEXTS);
 +    running = threads;
 +    for (i = 0; i < threads; i++) {
 +        Coroutine *co1 = qemu_coroutine_create(test_multi_co_mutex_entry, NULL);
 +        aio_co_schedule(ctx[i], co1);
 +    }
 +
 +    g_usleep(seconds * 1000000);
 +
 +    atomic_mb_set(&now_stopping, true);
 +    while (running > 0) {
 +        g_usleep(100000);
 +    }
 +
 +    join_aio_contexts();
 +    g_test_message("%d iterations/second\n", counter / seconds);
 +    g_assert_cmpint(counter, ==, atomic_counter);
 +}
 +
 +/* Testing with NUM_CONTEXTS threads focuses on the queue.  The mutex however
 + * is too contended (and the threads spend too much time in aio_poll)
 + * to actually stress the handoff protocol.
 + */
 +static void test_multi_co_mutex_1(void)
 +{
 +    test_multi_co_mutex(NUM_CONTEXTS, 1);
 +}
 +
 +static void test_multi_co_mutex_10(void)
 +{
 +    test_multi_co_mutex(NUM_CONTEXTS, 10);
 +}
 +
 +/* Testing with fewer threads stresses the handoff protocol too.  Still, the
 + * case where the locker _can_ pick up a handoff is very rare, happening
 + * about 10 times in 1 million, so increase the runtime a bit compared to
 + * other "quick" testcases that only run for 1 second.
 + */
 +static void test_multi_co_mutex_2_3(void)
 +{
 +    test_multi_co_mutex(2, 3);
 +}
 +
 +static void test_multi_co_mutex_2_30(void)
 +{
 +    test_multi_co_mutex(2, 30);
 +}
 +
  /* End of tests.  */
  int main(int argc, char **argv)
@@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv)
      g_test_add_func("/aio/multi/lifecycle", test_lifecycle);
      if (g_test_quick()) {
          g_test_add_func("/aio/multi/schedule", test_multi_co_schedule_1);
 +        g_test_add_func("/aio/multi/mutex/contended", test_multi_co_mutex_1);
 +        g_test_add_func("/aio/multi/mutex/handoff", test_multi_co_mutex_2_3);
      } else {
          g_test_add_func("/aio/multi/schedule", test_multi_co_schedule_10);
 +        g_test_add_func("/aio/multi/mutex/contended", test_multi_co_mutex_10);
 +        g_test_add_func("/aio/multi/mutex/handoff", test_multi_co_mutex_2_30);
      }
      return g_test_run();
  }
 diff --git a/util/qemu-coroutine-lock.c b/util/qemu-coroutine-lock.c
 index XXXXXXX..XXXXXXX 100644
 --- a/util/qemu-coroutine-lock.c
 +++ b/util/qemu-coroutine-lock.c
@@ -XXX,XX +XXX,XX @@
   * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
   * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
   * THE SOFTWARE.
 + *
 + * The lock-free mutex implementation is based on OSv
 + * (core/lfmutex.cc, include/lockfree/mutex.hh).
 + * Copyright (C) 2013 Cloudius Systems, Ltd.
   */
  #include "qemu/osdep.h"
@@ -XXX,XX +XXX,XX @@ bool qemu_co_queue_empty(CoQueue *queue)
      return QSIMPLEQ_FIRST(&queue->entries) == NULL;
  }
 +/* The wait records are handled with a multiple-producer, single-consumer
 + * lock-free queue.  There cannot be two concurrent pop_waiter() calls
 + * because pop_waiter() can only be called while mutex->handoff is zero.
 + * This can happen in three cases:
 + * - in qemu_co_mutex_unlock, before the hand-off protocol has started.
 + *   In this case, qemu_co_mutex_lock will see mutex->handoff == 0 and
 + *   not take part in the handoff.
 + * - in qemu_co_mutex_lock, if it steals the hand-off responsibility from
 + *   qemu_co_mutex_unlock.  In this case, qemu_co_mutex_unlock will fail
 + *   the cmpxchg (it will see either 0 or the next sequence value) and
 + *   exit.  The next hand-off cannot begin until qemu_co_mutex_lock has
 + *   woken up someone.
 + * - in qemu_co_mutex_unlock, if it takes the hand-off token itself.
 + *   In this case another iteration starts with mutex->handoff == 0;
 + *   a concurrent qemu_co_mutex_lock will fail the cmpxchg, and
 + *   qemu_co_mutex_unlock will go back to case (1).
 + *
 + * The following functions manage this queue.
 + */
 +typedef struct CoWaitRecord {
 +    Coroutine *co;
 +    QSLIST_ENTRY(CoWaitRecord) next;
 +} CoWaitRecord;
 +
 +static void push_waiter(CoMutex *mutex, CoWaitRecord *w)
 +{
 +    w->co = qemu_coroutine_self();
 +    QSLIST_INSERT_HEAD_ATOMIC(&mutex->from_push, w, next);
 +}
 +
 +static void move_waiters(CoMutex *mutex)
 +{
 +    QSLIST_HEAD(, CoWaitRecord) reversed;
 +    QSLIST_MOVE_ATOMIC(&reversed, &mutex->from_push);
 +    while (!QSLIST_EMPTY(&reversed)) {
 +        CoWaitRecord *w = QSLIST_FIRST(&reversed);
 +        QSLIST_REMOVE_HEAD(&reversed, next);
 +        QSLIST_INSERT_HEAD(&mutex->to_pop, w, next);
 +    }
 +}
 +
 +static CoWaitRecord *pop_waiter(CoMutex *mutex)
 +{
 +    CoWaitRecord *w;
 +
 +    if (QSLIST_EMPTY(&mutex->to_pop)) {
 +        move_waiters(mutex);
 +        if (QSLIST_EMPTY(&mutex->to_pop)) {
 +            return NULL;
 +        }
 +    }
 +    w = QSLIST_FIRST(&mutex->to_pop);
 +    QSLIST_REMOVE_HEAD(&mutex->to_pop, next);
 +    return w;
 +}
 +
 +static bool has_waiters(CoMutex *mutex)
 +{
 +    return QSLIST_EMPTY(&mutex->to_pop) || QSLIST_EMPTY(&mutex->from_push);
 +}
 +
  void qemu_co_mutex_init(CoMutex *mutex)
  {
      memset(mutex, 0, sizeof(*mutex));
 -    qemu_co_queue_init(&mutex->queue);
  }
 -void coroutine_fn qemu_co_mutex_lock(CoMutex *mutex)
 +static void coroutine_fn qemu_co_mutex_lock_slowpath(CoMutex *mutex)
  {
      Coroutine *self = qemu_coroutine_self();
 +    CoWaitRecord w;
 +    unsigned old_handoff;
      trace_qemu_co_mutex_lock_entry(mutex, self);
 +    w.co = self;
 +    push_waiter(mutex, &w);
 -    while (mutex->locked) {
 -        qemu_co_queue_wait(&mutex->queue);
 +    /* This is the "Responsibility Hand-Off" protocol; a lock() picks from
 +     * a concurrent unlock() the responsibility of waking somebody up.
 +     */
 +    old_handoff = atomic_mb_read(&mutex->handoff);
 +    if (old_handoff &&
 +        has_waiters(mutex) &&
 +        atomic_cmpxchg(&mutex->handoff, old_handoff, 0) == old_handoff) {
 +        /* There can be no concurrent pops, because there can be only
 +         * one active handoff at a time.
 +         */
 +        CoWaitRecord *to_wake = pop_waiter(mutex);
 +        Coroutine *co = to_wake->co;
 +        if (co == self) {
 +            /* We got the lock ourselves!  */
 +            assert(to_wake == &w);
 +            return;
 +        }
 +
 +        aio_co_wake(co);
      }
 -    mutex->locked = true;
 -    mutex->holder = self;
 -    self->locks_held++;
 -
 +    qemu_coroutine_yield();
      trace_qemu_co_mutex_lock_return(mutex, self);
  }
 +void coroutine_fn qemu_co_mutex_lock(CoMutex *mutex)
 +{
 +    Coroutine *self = qemu_coroutine_self();
 +
 +    if (atomic_fetch_inc(&mutex->locked) == 0) {
 +        /* Uncontended.  */
 +        trace_qemu_co_mutex_lock_uncontended(mutex, self);
 +    } else {
 +        qemu_co_mutex_lock_slowpath(mutex);
 +    }
 +    mutex->holder = self;
 +    self->locks_held++;
 +}
 +
  void coroutine_fn qemu_co_mutex_unlock(CoMutex *mutex)
  {
      Coroutine *self = qemu_coroutine_self();
      trace_qemu_co_mutex_unlock_entry(mutex, self);
 -    assert(mutex->locked == true);
 +    assert(mutex->locked);
      assert(mutex->holder == self);
      assert(qemu_in_coroutine());
 -    mutex->locked = false;
      mutex->holder = NULL;
      self->locks_held--;
 -    qemu_co_queue_next(&mutex->queue);
 +    if (atomic_fetch_dec(&mutex->locked) == 1) {
 +        /* No waiting qemu_co_mutex_lock().  Pfew, that was easy!  */
 +        return;
 +    }
 +
 +    for (;;) {
 +        CoWaitRecord *to_wake = pop_waiter(mutex);
 +        unsigned our_handoff;
 +
 +        if (to_wake) {
 +            Coroutine *co = to_wake->co;
 +            aio_co_wake(co);
 +            break;
 +        }
 +
 +        /* Some concurrent lock() is in progress (we know this because
 +         * mutex->locked was >1) but it hasn't yet put itself on the wait
 +         * queue.  Pick a sequence number for the handoff protocol (not 0).
 +         */
 +        if (++mutex->sequence == 0) {
 +            mutex->sequence = 1;
 +        }
 +
 +        our_handoff = mutex->sequence;
 +        atomic_mb_set(&mutex->handoff, our_handoff);
 +        if (!has_waiters(mutex)) {
 +            /* The concurrent lock has not added itself yet, so it
 +             * will be able to pick our handoff.
 +             */
 +            break;
 +        }
 +
 +        /* Try to do the handoff protocol ourselves; if somebody else has
 +         * already taken it, however, we're done and they're responsible.
 +         */
 +        if (atomic_cmpxchg(&mutex->handoff, our_handoff, 0) != our_handoff) {
 +            break;
 +        }
 +    }
      trace_qemu_co_mutex_unlock_return(mutex, self);
  }
 diff --git a/util/trace-events b/util/trace-events
 index XXXXXXX..XXXXXXX 100644
 --- a/util/trace-events
 +++ b/util/trace-events
@@ -XXX,XX +XXX,XX @@ qemu_coroutine_terminate(void *co) "self %p"
  # util/qemu-coroutine-lock.c
  qemu_co_queue_run_restart(void *co) "co %p"
 +qemu_co_mutex_lock_uncontended(void *mutex, void *self) "mutex %p self %p"
  qemu_co_mutex_lock_entry(void *mutex, void *self) "mutex %p self %p"
  qemu_co_mutex_lock_return(void *mutex, void *self) "mutex %p self %p"
  qemu_co_mutex_unlock_entry(void *mutex, void *self) "mutex %p self %p"
 --
 .9.3

-New patch
+[Qemu-devel] [PULL v2 20/24] coroutine-lock: add limited spinning to CoMutex
+From: Paolo Bonzini <pbonzini@redhat.com>
+Running a very small critical section on pthread_mutex_t and CoMutex
+shows that pthread_mutex_t is much faster because it doesn't actually
+go to sleep.  What happens is that the critical section is shorter
+than the latency of entering the kernel and thus FUTEX_WAIT always
+fails.  With CoMutex there is no such latency but you still want to
+avoid wait and wakeup.  So introduce it artificially.
+This only works with one waiters; because CoMutex is fair, it will
+always have more waits and wakeups than a pthread_mutex_t.
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
+Reviewed-by: Fam Zheng <famz@redhat.com>
+Message-id: 20170213181244.16297-3-pbonzini@redhat.com
+Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
+---
+ include/qemu/coroutine.h   |  5 +++++
+ util/qemu-coroutine-lock.c | 51 ++++++++++++++++++++++++++++++++++++++++------
+ util/qemu-coroutine.c      |  2 +-
+files changed, 51 insertions(+), 7 deletions(-)
+diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
+index XXXXXXX..XXXXXXX 100644
+--- a/include/qemu/coroutine.h
++++ b/include/qemu/coroutine.h
+@@ -XXX,XX +XXX,XX @@ typedef struct CoMutex {
+      */
+     unsigned locked;
++    /* Context that is holding the lock.  Useful to avoid spinning
++     * when two coroutines on the same AioContext try to get the lock. :)
++     */
++    AioContext *ctx;
++
+     /* A queue of waiters.  Elements are added atomically in front of
+      * from_push.  to_pop is only populated, and popped from, by whoever
+      * is in charge of the next wakeup.  This can be an unlocker or,
+diff --git a/util/qemu-coroutine-lock.c b/util/qemu-coroutine-lock.c
+index XXXXXXX..XXXXXXX 100644
+--- a/util/qemu-coroutine-lock.c
++++ b/util/qemu-coroutine-lock.c
+@@ -XXX,XX +XXX,XX @@
+ #include "qemu-common.h"
+ #include "qemu/coroutine.h"
+ #include "qemu/coroutine_int.h"
++#include "qemu/processor.h"
+ #include "qemu/queue.h"
+ #include "block/aio.h"
+ #include "trace.h"
+@@ -XXX,XX +XXX,XX @@ void qemu_co_mutex_init(CoMutex *mutex)
+     memset(mutex, 0, sizeof(*mutex));
+ }
+-static void coroutine_fn qemu_co_mutex_lock_slowpath(CoMutex *mutex)
++static void coroutine_fn qemu_co_mutex_wake(CoMutex *mutex, Coroutine *co)
++{
++    /* Read co before co->ctx; pairs with smp_wmb() in
++     * qemu_coroutine_enter().
++     */
++    smp_read_barrier_depends();
++    mutex->ctx = co->ctx;
++    aio_co_wake(co);
++}
++
++static void coroutine_fn qemu_co_mutex_lock_slowpath(AioContext *ctx,
++                                                     CoMutex *mutex)
+ {
+     Coroutine *self = qemu_coroutine_self();
+     CoWaitRecord w;
+@@ -XXX,XX +XXX,XX @@ static void coroutine_fn qemu_co_mutex_lock_slowpath(CoMutex *mutex)
+         if (co == self) {
+             /* We got the lock ourselves!  */
+             assert(to_wake == &w);
++            mutex->ctx = ctx;
+             return;
+         }
+-        aio_co_wake(co);
++        qemu_co_mutex_wake(mutex, co);
+     }
+     qemu_coroutine_yield();
+@@ -XXX,XX +XXX,XX @@ static void coroutine_fn qemu_co_mutex_lock_slowpath(CoMutex *mutex)
+ void coroutine_fn qemu_co_mutex_lock(CoMutex *mutex)
+ {
++    AioContext *ctx = qemu_get_current_aio_context();
+     Coroutine *self = qemu_coroutine_self();
++    int waiters, i;
+-    if (atomic_fetch_inc(&mutex->locked) == 0) {
++    /* Running a very small critical section on pthread_mutex_t and CoMutex
++     * shows that pthread_mutex_t is much faster because it doesn't actually
++     * go to sleep.  What happens is that the critical section is shorter
++     * than the latency of entering the kernel and thus FUTEX_WAIT always
++     * fails.  With CoMutex there is no such latency but you still want to
++     * avoid wait and wakeup.  So introduce it artificially.
++     */
++    i = 0;
++retry_fast_path:
++    waiters = atomic_cmpxchg(&mutex->locked, 0, 1);
++    if (waiters != 0) {
++        while (waiters == 1 && ++i < 1000) {
++            if (atomic_read(&mutex->ctx) == ctx) {
++                break;
++            }
++            if (atomic_read(&mutex->locked) == 0) {
++                goto retry_fast_path;
++            }
++            cpu_relax();
++        }
++        waiters = atomic_fetch_inc(&mutex->locked);
++    }
++
++    if (waiters == 0) {
+         /* Uncontended.  */
+         trace_qemu_co_mutex_lock_uncontended(mutex, self);
++        mutex->ctx = ctx;
+     } else {
+-        qemu_co_mutex_lock_slowpath(mutex);
++        qemu_co_mutex_lock_slowpath(ctx, mutex);
+     }
+     mutex->holder = self;
+     self->locks_held++;
+@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_mutex_unlock(CoMutex *mutex)
+     assert(mutex->holder == self);
+     assert(qemu_in_coroutine());
++    mutex->ctx = NULL;
+     mutex->holder = NULL;
+     self->locks_held--;
+     if (atomic_fetch_dec(&mutex->locked) == 1) {
+@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_mutex_unlock(CoMutex *mutex)
+         unsigned our_handoff;
+         if (to_wake) {
+-            Coroutine *co = to_wake->co;
+-            aio_co_wake(co);
++            qemu_co_mutex_wake(mutex, to_wake->co);
+             break;
+         }
+diff --git a/util/qemu-coroutine.c b/util/qemu-coroutine.c
+index XXXXXXX..XXXXXXX 100644
+--- a/util/qemu-coroutine.c
++++ b/util/qemu-coroutine.c
+@@ -XXX,XX +XXX,XX @@ void qemu_coroutine_enter(Coroutine *co)
+     co->ctx = qemu_get_current_aio_context();
+     /* Store co->ctx before anything that stores co.  Matches
+-     * barrier in aio_co_wake.
++     * barrier in aio_co_wake and qemu_co_mutex_wake.
+      */
+     smp_wmb();
+--
+.9.3

-New patch
+[Qemu-devel] [PULL v2 21/24] test-aio-multithread: add performance comparison with thread-based mutexes
+From: Paolo Bonzini <pbonzini@redhat.com>
 Add two implementations of the same benchmark as the previous patch,
 but using pthreads.  One uses a normal QemuMutex, the other is Linux
 only and implements a fair mutex based on MCS locks and futexes.
 This shows that the slower performance of the 5-thread case is due to
 the fairness of CoMutex, rather than to coroutines.  If fairness does
 not matter, as is the case with two threads, CoMutex can actually be
 faster than pthreads.
 Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
 Reviewed-by: Fam Zheng <famz@redhat.com>
 Message-id: 20170213181244.16297-4-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
  tests/test-aio-multithread.c | 164 +++++++++++++++++++++++++++++++++++++++++++
 file changed, 164 insertions(+)
 diff --git a/tests/test-aio-multithread.c b/tests/test-aio-multithread.c
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/test-aio-multithread.c
 +++ b/tests/test-aio-multithread.c
@@ -XXX,XX +XXX,XX @@ static void test_multi_co_mutex_2_30(void)
      test_multi_co_mutex(2, 30);
  }
 +/* Same test with fair mutexes, for performance comparison.  */
 +
 +#ifdef CONFIG_LINUX
 +#include "qemu/futex.h"
 +
 +/* The nodes for the mutex reside in this structure (on which we try to avoid
 + * false sharing).  The head of the mutex is in the "mutex_head" variable.
 + */
 +static struct {
 +    int next, locked;
 +    int padding[14];
 +} nodes[NUM_CONTEXTS] __attribute__((__aligned__(64)));
 +
 +static int mutex_head = -1;
 +
 +static void mcs_mutex_lock(void)
 +{
 +    int prev;
 +
 +    nodes[id].next = -1;
 +    nodes[id].locked = 1;
 +    prev = atomic_xchg(&mutex_head, id);
 +    if (prev != -1) {
 +        atomic_set(&nodes[prev].next, id);
 +        qemu_futex_wait(&nodes[id].locked, 1);
 +    }
 +}
 +
 +static void mcs_mutex_unlock(void)
 +{
 +    int next;
 +    if (nodes[id].next == -1) {
 +        if (atomic_read(&mutex_head) == id &&
 +            atomic_cmpxchg(&mutex_head, id, -1) == id) {
 +            /* Last item in the list, exit.  */
 +            return;
 +        }
 +        while (atomic_read(&nodes[id].next) == -1) {
 +            /* mcs_mutex_lock did the xchg, but has not updated
 +             * nodes[prev].next yet.
 +             */
 +        }
 +    }
 +
 +    /* Wake up the next in line.  */
 +    next = nodes[id].next;
 +    nodes[next].locked = 0;
 +    qemu_futex_wake(&nodes[next].locked, 1);
 +}
 +
 +static void test_multi_fair_mutex_entry(void *opaque)
 +{
 +    while (!atomic_mb_read(&now_stopping)) {
 +        mcs_mutex_lock();
 +        counter++;
 +        mcs_mutex_unlock();
 +        atomic_inc(&atomic_counter);
 +    }
 +    atomic_dec(&running);
 +}
 +
 +static void test_multi_fair_mutex(int threads, int seconds)
 +{
 +    int i;
 +
 +    assert(mutex_head == -1);
 +    counter = 0;
 +    atomic_counter = 0;
 +    now_stopping = false;
 +
 +    create_aio_contexts();
 +    assert(threads <= NUM_CONTEXTS);
 +    running = threads;
 +    for (i = 0; i < threads; i++) {
 +        Coroutine *co1 = qemu_coroutine_create(test_multi_fair_mutex_entry, NULL);
 +        aio_co_schedule(ctx[i], co1);
 +    }
 +
 +    g_usleep(seconds * 1000000);
 +
 +    atomic_mb_set(&now_stopping, true);
 +    while (running > 0) {
 +        g_usleep(100000);
 +    }
 +
 +    join_aio_contexts();
 +    g_test_message("%d iterations/second\n", counter / seconds);
 +    g_assert_cmpint(counter, ==, atomic_counter);
 +}
 +
 +static void test_multi_fair_mutex_1(void)
 +{
 +    test_multi_fair_mutex(NUM_CONTEXTS, 1);
 +}
 +
 +static void test_multi_fair_mutex_10(void)
 +{
 +    test_multi_fair_mutex(NUM_CONTEXTS, 10);
 +}
 +#endif
 +
 +/* Same test with pthread mutexes, for performance comparison and
 + * portability.  */
 +
 +static QemuMutex mutex;
 +
 +static void test_multi_mutex_entry(void *opaque)
 +{
 +    while (!atomic_mb_read(&now_stopping)) {
 +        qemu_mutex_lock(&mutex);
 +        counter++;
 +        qemu_mutex_unlock(&mutex);
 +        atomic_inc(&atomic_counter);
 +    }
 +    atomic_dec(&running);
 +}
 +
 +static void test_multi_mutex(int threads, int seconds)
 +{
 +    int i;
 +
 +    qemu_mutex_init(&mutex);
 +    counter = 0;
 +    atomic_counter = 0;
 +    now_stopping = false;
 +
 +    create_aio_contexts();
 +    assert(threads <= NUM_CONTEXTS);
 +    running = threads;
 +    for (i = 0; i < threads; i++) {
 +        Coroutine *co1 = qemu_coroutine_create(test_multi_mutex_entry, NULL);
 +        aio_co_schedule(ctx[i], co1);
 +    }
 +
 +    g_usleep(seconds * 1000000);
 +
 +    atomic_mb_set(&now_stopping, true);
 +    while (running > 0) {
 +        g_usleep(100000);
 +    }
 +
 +    join_aio_contexts();
 +    g_test_message("%d iterations/second\n", counter / seconds);
 +    g_assert_cmpint(counter, ==, atomic_counter);
 +}
 +
 +static void test_multi_mutex_1(void)
 +{
 +    test_multi_mutex(NUM_CONTEXTS, 1);
 +}
 +
 +static void test_multi_mutex_10(void)
 +{
 +    test_multi_mutex(NUM_CONTEXTS, 10);
 +}
 +
  /* End of tests.  */
  int main(int argc, char **argv)
@@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv)
          g_test_add_func("/aio/multi/schedule", test_multi_co_schedule_1);
          g_test_add_func("/aio/multi/mutex/contended", test_multi_co_mutex_1);
          g_test_add_func("/aio/multi/mutex/handoff", test_multi_co_mutex_2_3);
 +#ifdef CONFIG_LINUX
 +        g_test_add_func("/aio/multi/mutex/mcs", test_multi_fair_mutex_1);
 +#endif
 +        g_test_add_func("/aio/multi/mutex/pthread", test_multi_mutex_1);
      } else {
          g_test_add_func("/aio/multi/schedule", test_multi_co_schedule_10);
          g_test_add_func("/aio/multi/mutex/contended", test_multi_co_mutex_10);
          g_test_add_func("/aio/multi/mutex/handoff", test_multi_co_mutex_2_30);
 +#ifdef CONFIG_LINUX
 +        g_test_add_func("/aio/multi/mutex/mcs", test_multi_fair_mutex_10);
 +#endif
 +        g_test_add_func("/aio/multi/mutex/pthread", test_multi_mutex_10);
      }
      return g_test_run();
  }
 --
 .9.3

-[PULL 5/5] block/block-copy: fix use-after-free of task pointer
+[Qemu-devel] [PULL v2 22/24] coroutine-lock: place CoMutex before CoQueue in header
-From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-Obviously, we should g_free the task after trace point and offset
+This will avoid forward references in the next patch.  It is also
-update.
+more logical because CoQueue is not anymore the basic primitive.
-Reported-by: Coverity (CID 1428756)
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-Fixes: 4ce5dd3e9b5ee0fac18625860eb3727399ee965e
+Reviewed-by: Fam Zheng <famz@redhat.com>
-Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
+Message-id: 20170213181244.16297-5-pbonzini@redhat.com
-Message-Id: <20200507183800.22626-1-vsementsov@virtuozzo.com>
+Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 Reviewed-by: Eric Blake <eblake@redhat.com>
 Signed-off-by: Max Reitz <mreitz@redhat.com>
 ---
- block/block-copy.c | 2 +-
+ include/qemu/coroutine.h | 89 ++++++++++++++++++++++++------------------------
-file changed, 1 insertion(+), 1 deletion(-)
+file changed, 44 insertions(+), 45 deletions(-)
-diff --git a/block/block-copy.c b/block/block-copy.c
+diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
 index XXXXXXX..XXXXXXX 100644
---- a/block/block-copy.c
+--- a/include/qemu/coroutine.h
-+++ b/block/block-copy.c
++++ b/include/qemu/coroutine.h
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn block_copy_dirty_clusters(BlockCopyState *s,
+@@ -XXX,XX +XXX,XX @@ bool qemu_in_coroutine(void);
-         }
+  */
-         if (s->skip_unallocated && !(ret & BDRV_BLOCK_ALLOCATED)) {
+ bool qemu_coroutine_entered(Coroutine *co);
-             block_copy_task_end(task, 0);
--            g_free(task);
+-
-             progress_set_remaining(s->progress,
+-/**
-                                    bdrv_get_dirty_count(s->copy_bitmap) +
+- * CoQueues are a mechanism to queue coroutines in order to continue executing
-                                    s->in_flight_bytes);
+- * them later. They provide the fundamental primitives on which coroutine locks
-             trace_block_copy_skip_range(s, task->offset, task->bytes);
+- * are built.
-             offset = task_end(task);
+- */
-             bytes = end - offset;
+-typedef struct CoQueue {
-+            g_free(task);
+-    QSIMPLEQ_HEAD(, Coroutine) entries;
-             continue;
+-} CoQueue;
-         }
+-
-         task->zeroes = ret & BDRV_BLOCK_ZERO;
+-/**
 - * Initialise a CoQueue. This must be called before any other operation is used
 - * on the CoQueue.
 - */
 -void qemu_co_queue_init(CoQueue *queue);
 -
 -/**
 - * Adds the current coroutine to the CoQueue and transfers control to the
 - * caller of the coroutine.
 - */
 -void coroutine_fn qemu_co_queue_wait(CoQueue *queue);
 -
 -/**
 - * Restarts the next coroutine in the CoQueue and removes it from the queue.
 - *
 - * Returns true if a coroutine was restarted, false if the queue is empty.
 - */
 -bool coroutine_fn qemu_co_queue_next(CoQueue *queue);
 -
 -/**
 - * Restarts all coroutines in the CoQueue and leaves the queue empty.
 - */
 -void coroutine_fn qemu_co_queue_restart_all(CoQueue *queue);
 -
 -/**
 - * Enter the next coroutine in the queue
 - */
 -bool qemu_co_enter_next(CoQueue *queue);
 -
 -/**
 - * Checks if the CoQueue is empty.
 - */
 -bool qemu_co_queue_empty(CoQueue *queue);
 -
 -
  /**
   * Provides a mutex that can be used to synchronise coroutines
   */
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_mutex_lock(CoMutex *mutex);
   */
  void coroutine_fn qemu_co_mutex_unlock(CoMutex *mutex);
 +
 +/**
 + * CoQueues are a mechanism to queue coroutines in order to continue executing
 + * them later.
 + */
 +typedef struct CoQueue {
 +    QSIMPLEQ_HEAD(, Coroutine) entries;
 +} CoQueue;
 +
 +/**
 + * Initialise a CoQueue. This must be called before any other operation is used
 + * on the CoQueue.
 + */
 +void qemu_co_queue_init(CoQueue *queue);
 +
 +/**
 + * Adds the current coroutine to the CoQueue and transfers control to the
 + * caller of the coroutine.
 + */
 +void coroutine_fn qemu_co_queue_wait(CoQueue *queue);
 +
 +/**
 + * Restarts the next coroutine in the CoQueue and removes it from the queue.
 + *
 + * Returns true if a coroutine was restarted, false if the queue is empty.
 + */
 +bool coroutine_fn qemu_co_queue_next(CoQueue *queue);
 +
 +/**
 + * Restarts all coroutines in the CoQueue and leaves the queue empty.
 + */
 +void coroutine_fn qemu_co_queue_restart_all(CoQueue *queue);
 +
 +/**
 + * Enter the next coroutine in the queue
 + */
 +bool qemu_co_enter_next(CoQueue *queue);
 +
 +/**
 + * Checks if the CoQueue is empty.
 + */
 +bool qemu_co_queue_empty(CoQueue *queue);
 +
 +
  typedef struct CoRwlock {
      bool writer;
      int reader;
 --
-.26.2
+.9.3

-New patch
+[Qemu-devel] [PULL v2 23/24] coroutine-lock: add mutex argument to CoQueue APIs
+From: Paolo Bonzini <pbonzini@redhat.com>
 All that CoQueue needs in order to become thread-safe is help
 from an external mutex.  Add this to the API.
 Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
 Reviewed-by: Fam Zheng <famz@redhat.com>
 Message-id: 20170213181244.16297-6-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
  include/qemu/coroutine.h   |  8 +++++---
  block/backup.c             |  2 +-
  block/io.c                 |  4 ++--
  block/nbd-client.c         |  2 +-
  block/qcow2-cluster.c      |  4 +---
  block/sheepdog.c           |  2 +-
  block/throttle-groups.c    |  2 +-
  hw/9pfs/9p.c               |  2 +-
  util/qemu-coroutine-lock.c | 24 +++++++++++++++++++++---
 files changed, 34 insertions(+), 16 deletions(-)
 diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
 index XXXXXXX..XXXXXXX 100644
 --- a/include/qemu/coroutine.h
 +++ b/include/qemu/coroutine.h
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_mutex_unlock(CoMutex *mutex);
  /**
   * CoQueues are a mechanism to queue coroutines in order to continue executing
 - * them later.
 + * them later.  They are similar to condition variables, but they need help
 + * from an external mutex in order to maintain thread-safety.
   */
  typedef struct CoQueue {
      QSIMPLEQ_HEAD(, Coroutine) entries;
@@ -XXX,XX +XXX,XX @@ void qemu_co_queue_init(CoQueue *queue);
  /**
   * Adds the current coroutine to the CoQueue and transfers control to the
 - * caller of the coroutine.
 + * caller of the coroutine.  The mutex is unlocked during the wait and
 + * locked again afterwards.
   */
 -void coroutine_fn qemu_co_queue_wait(CoQueue *queue);
 +void coroutine_fn qemu_co_queue_wait(CoQueue *queue, CoMutex *mutex);
  /**
   * Restarts the next coroutine in the CoQueue and removes it from the queue.
 diff --git a/block/backup.c b/block/backup.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/backup.c
 +++ b/block/backup.c
@@ -XXX,XX +XXX,XX @@ static void coroutine_fn wait_for_overlapping_requests(BackupBlockJob *job,
          retry = false;
          QLIST_FOREACH(req, &job->inflight_reqs, list) {
              if (end > req->start && start < req->end) {
 -                qemu_co_queue_wait(&req->wait_queue);
 +                qemu_co_queue_wait(&req->wait_queue, NULL);
                  retry = true;
                  break;
              }
 diff --git a/block/io.c b/block/io.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/io.c
 +++ b/block/io.c
@@ -XXX,XX +XXX,XX @@ static bool coroutine_fn wait_serialising_requests(BdrvTrackedRequest *self)
                   * (instead of producing a deadlock in the former case). */
                  if (!req->waiting_for) {
                      self->waiting_for = req;
 -                    qemu_co_queue_wait(&req->wait_queue);
 +                    qemu_co_queue_wait(&req->wait_queue, NULL);
                      self->waiting_for = NULL;
                      retry = true;
                      waited = true;
@@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_flush(BlockDriverState *bs)
      /* Wait until any previous flushes are completed */
      while (bs->active_flush_req) {
 -        qemu_co_queue_wait(&bs->flush_queue);
 +        qemu_co_queue_wait(&bs->flush_queue, NULL);
      }
      bs->active_flush_req = true;
 diff --git a/block/nbd-client.c b/block/nbd-client.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/nbd-client.c
 +++ b/block/nbd-client.c
@@ -XXX,XX +XXX,XX @@ static void nbd_coroutine_start(NBDClientSession *s,
      /* Poor man semaphore.  The free_sema is locked when no other request
       * can be accepted, and unlocked after receiving one reply.  */
      if (s->in_flight == MAX_NBD_REQUESTS) {
 -        qemu_co_queue_wait(&s->free_sema);
 +        qemu_co_queue_wait(&s->free_sema, NULL);
          assert(s->in_flight < MAX_NBD_REQUESTS);
      }
      s->in_flight++;
 diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/qcow2-cluster.c
 +++ b/block/qcow2-cluster.c
@@ -XXX,XX +XXX,XX @@ static int handle_dependencies(BlockDriverState *bs, uint64_t guest_offset,
              if (bytes == 0) {
                  /* Wait for the dependency to complete. We need to recheck
                   * the free/allocated clusters when we continue. */
 -                qemu_co_mutex_unlock(&s->lock);
 -                qemu_co_queue_wait(&old_alloc->dependent_requests);
 -                qemu_co_mutex_lock(&s->lock);
 +                qemu_co_queue_wait(&old_alloc->dependent_requests, &s->lock);
                  return -EAGAIN;
              }
          }
 diff --git a/block/sheepdog.c b/block/sheepdog.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/sheepdog.c
 +++ b/block/sheepdog.c
@@ -XXX,XX +XXX,XX @@ static void wait_for_overlapping_aiocb(BDRVSheepdogState *s, SheepdogAIOCB *acb)
  retry:
      QLIST_FOREACH(cb, &s->inflight_aiocb_head, aiocb_siblings) {
          if (AIOCBOverlapping(acb, cb)) {
 -            qemu_co_queue_wait(&s->overlapping_queue);
 +            qemu_co_queue_wait(&s->overlapping_queue, NULL);
              goto retry;
          }
      }
 diff --git a/block/throttle-groups.c b/block/throttle-groups.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/throttle-groups.c
 +++ b/block/throttle-groups.c
@@ -XXX,XX +XXX,XX @@ void coroutine_fn throttle_group_co_io_limits_intercept(BlockBackend *blk,
      if (must_wait || blkp->pending_reqs[is_write]) {
          blkp->pending_reqs[is_write]++;
          qemu_mutex_unlock(&tg->lock);
 -        qemu_co_queue_wait(&blkp->throttled_reqs[is_write]);
 +        qemu_co_queue_wait(&blkp->throttled_reqs[is_write], NULL);
          qemu_mutex_lock(&tg->lock);
          blkp->pending_reqs[is_write]--;
      }
 diff --git a/hw/9pfs/9p.c b/hw/9pfs/9p.c
 index XXXXXXX..XXXXXXX 100644
 --- a/hw/9pfs/9p.c
 +++ b/hw/9pfs/9p.c
@@ -XXX,XX +XXX,XX @@ static void coroutine_fn v9fs_flush(void *opaque)
          /*
           * Wait for pdu to complete.
           */
 -        qemu_co_queue_wait(&cancel_pdu->complete);
 +        qemu_co_queue_wait(&cancel_pdu->complete, NULL);
          cancel_pdu->cancelled = 0;
          pdu_free(cancel_pdu);
      }
 diff --git a/util/qemu-coroutine-lock.c b/util/qemu-coroutine-lock.c
 index XXXXXXX..XXXXXXX 100644
 --- a/util/qemu-coroutine-lock.c
 +++ b/util/qemu-coroutine-lock.c
@@ -XXX,XX +XXX,XX @@ void qemu_co_queue_init(CoQueue *queue)
      QSIMPLEQ_INIT(&queue->entries);
  }
 -void coroutine_fn qemu_co_queue_wait(CoQueue *queue)
 +void coroutine_fn qemu_co_queue_wait(CoQueue *queue, CoMutex *mutex)
  {
      Coroutine *self = qemu_coroutine_self();
      QSIMPLEQ_INSERT_TAIL(&queue->entries, self, co_queue_next);
 +
 +    if (mutex) {
 +        qemu_co_mutex_unlock(mutex);
 +    }
 +
 +    /* There is no race condition here.  Other threads will call
 +     * aio_co_schedule on our AioContext, which can reenter this
 +     * coroutine but only after this yield and after the main loop
 +     * has gone through the next iteration.
 +     */
      qemu_coroutine_yield();
      assert(qemu_in_coroutine());
 +
 +    /* TODO: OSv implements wait morphing here, where the wakeup
 +     * primitive automatically places the woken coroutine on the
 +     * mutex's queue.  This avoids the thundering herd effect.
 +     */
 +    if (mutex) {
 +        qemu_co_mutex_lock(mutex);
 +    }
  }
  /**
@@ -XXX,XX +XXX,XX @@ void qemu_co_rwlock_rdlock(CoRwlock *lock)
      Coroutine *self = qemu_coroutine_self();
      while (lock->writer) {
 -        qemu_co_queue_wait(&lock->queue);
 +        qemu_co_queue_wait(&lock->queue, NULL);
      }
      lock->reader++;
      self->locks_held++;
@@ -XXX,XX +XXX,XX @@ void qemu_co_rwlock_wrlock(CoRwlock *lock)
      Coroutine *self = qemu_coroutine_self();
      while (lock->writer || lock->reader) {
 -        qemu_co_queue_wait(&lock->queue);
 +        qemu_co_queue_wait(&lock->queue, NULL);
      }
      lock->writer = true;
      self->locks_held++;
 --
 .9.3

-[PULL 4/5] iotests: 287: add qcow2 compression type test
+[Qemu-devel] [PULL v2 24/24] coroutine-lock: make CoRwlock thread-safe and fair
-From: Denis Plotnikov <dplotnikov@virtuozzo.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-The test checks fulfilling qcow2 requirements for the compression
+This adds a CoMutex around the existing CoQueue.  Because the write-side
-type feature and zstd compression type operability.
+can just take CoMutex, the old "writer" field is not necessary anymore.
 Instead of removing it altogether, count the number of pending writers
 during a read-side critical section and forbid further readers from
 entering.
-Signed-off-by: Denis Plotnikov <dplotnikov@virtuozzo.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
+Reviewed-by: Fam Zheng <famz@redhat.com>
-Tested-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
+Message-id: 20170213181244.16297-7-pbonzini@redhat.com
-Reviewed-by: Eric Blake <eblake@redhat.com>
+Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 Message-Id: <20200507082521.29210-5-dplotnikov@virtuozzo.com>
 Signed-off-by: Max Reitz <mreitz@redhat.com>
 ---
- tests/qemu-iotests/287     | 152 +++++++++++++++++++++++++++++++++++++
+ include/qemu/coroutine.h   |  3 ++-
- tests/qemu-iotests/287.out |  67 ++++++++++++++++
+ util/qemu-coroutine-lock.c | 35 ++++++++++++++++++++++++-----------
- tests/qemu-iotests/group   |   1 +
+files changed, 26 insertions(+), 12 deletions(-)
 files changed, 220 insertions(+)
  create mode 100755 tests/qemu-iotests/287
  create mode 100644 tests/qemu-iotests/287.out
-diff --git a/tests/qemu-iotests/287 b/tests/qemu-iotests/287
+diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
-new file mode 100755
+index XXXXXXX..XXXXXXX 100644
-index XXXXXXX..XXXXXXX
+--- a/include/qemu/coroutine.h
---- /dev/null
++++ b/include/qemu/coroutine.h
-+++ b/tests/qemu-iotests/287
+@@ -XXX,XX +XXX,XX @@ bool qemu_co_queue_empty(CoQueue *queue);
-@@ -XXX,XX +XXX,XX @@
-+#!/usr/bin/env bash
-+#
+ typedef struct CoRwlock {
-+# Test case for an image using zstd compression
+-    bool writer;
-+#
++    int pending_writer;
-+# Copyright (c) 2020 Virtuozzo International GmbH
+     int reader;
-+#
++    CoMutex mutex;
-+# This program is free software; you can redistribute it and/or modify
+     CoQueue queue;
-+# it under the terms of the GNU General Public License as published by
+ } CoRwlock;
-+# the Free Software Foundation; either version 2 of the License, or
-+# (at your option) any later version.
+diff --git a/util/qemu-coroutine-lock.c b/util/qemu-coroutine-lock.c
-+#
+index XXXXXXX..XXXXXXX 100644
-+# This program is distributed in the hope that it will be useful,
+--- a/util/qemu-coroutine-lock.c
-+# but WITHOUT ANY WARRANTY; without even the implied warranty of
++++ b/util/qemu-coroutine-lock.c
-+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+@@ -XXX,XX +XXX,XX @@ void qemu_co_rwlock_init(CoRwlock *lock)
-+# GNU General Public License for more details.
+ {
-+#
+     memset(lock, 0, sizeof(*lock));
-+# You should have received a copy of the GNU General Public License
+     qemu_co_queue_init(&lock->queue);
-+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
++    qemu_co_mutex_init(&lock->mutex);
-+#
+ }
  void qemu_co_rwlock_rdlock(CoRwlock *lock)
  {
      Coroutine *self = qemu_coroutine_self();
 -    while (lock->writer) {
 -        qemu_co_queue_wait(&lock->queue, NULL);
 +    qemu_co_mutex_lock(&lock->mutex);
 +    /* For fairness, wait if a writer is in line.  */
 +    while (lock->pending_writer) {
 +        qemu_co_queue_wait(&lock->queue, &lock->mutex);
      }
      lock->reader++;
 +    qemu_co_mutex_unlock(&lock->mutex);
 +
-+# creator
++    /* The rest of the read-side critical section is run without the mutex.  */
-+owner=dplotnikov@virtuozzo.com
+     self->locks_held++;
  }
@@ -XXX,XX +XXX,XX @@ void qemu_co_rwlock_unlock(CoRwlock *lock)
      Coroutine *self = qemu_coroutine_self();
      assert(qemu_in_coroutine());
 -    if (lock->writer) {
 -        lock->writer = false;
 +    if (!lock->reader) {
 +        /* The critical section started in qemu_co_rwlock_wrlock.  */
          qemu_co_queue_restart_all(&lock->queue);
      } else {
 +        self->locks_held--;
 +
-+seq="$(basename $0)"
++        qemu_co_mutex_lock(&lock->mutex);
-+echo "QA output created by $seq"
+         lock->reader--;
          assert(lock->reader >= 0);
          /* Wakeup only one waiting writer */
@@ -XXX,XX +XXX,XX @@ void qemu_co_rwlock_unlock(CoRwlock *lock)
              qemu_co_queue_next(&lock->queue);
          }
      }
 -    self->locks_held--;
 +    qemu_co_mutex_unlock(&lock->mutex);
  }
  void qemu_co_rwlock_wrlock(CoRwlock *lock)
  {
 -    Coroutine *self = qemu_coroutine_self();
 -
 -    while (lock->writer || lock->reader) {
 -        qemu_co_queue_wait(&lock->queue, NULL);
 +    qemu_co_mutex_lock(&lock->mutex);
 +    lock->pending_writer++;
 +    while (lock->reader) {
 +        qemu_co_queue_wait(&lock->queue, &lock->mutex);
      }
 -    lock->writer = true;
 -    self->locks_held++;
 +    lock->pending_writer--;
 +
-+status=1    # failure is the default!
++    /* The rest of the write-side critical section is run with
-+
++     * the mutex taken, so that lock->reader remains zero.
-+# standard environment
++     * There is no need to update self->locks_held.
-+. ./common.rc
++     */
-+. ./common.filter
+ }
 +
 +# This tests qocw2-specific low-level functionality
 +_supported_fmt qcow2
 +_supported_proto file
 +_supported_os Linux
 +_unsupported_imgopts 'compat=0.10' data_file
 +
 +COMPR_IMG="$TEST_IMG.compressed"
 +RAND_FILE="$TEST_DIR/rand_data"
 +
 +_cleanup()
 +{
 +    _cleanup_test_img
 +    _rm_test_img "$COMPR_IMG"
 +    rm -f "$RAND_FILE"
 +}
 +trap "_cleanup; exit \$status" 0 1 2 3 15
 +
 +# for all the cases
 +CLUSTER_SIZE=65536
 +
 +# Check if we can run this test.
 +if IMGOPTS='compression_type=zstd' _make_test_img 64M |
 +    grep "Invalid parameter 'zstd'"; then
 +    _notrun "ZSTD is disabled"
 +fi
 +
 +echo
 +echo "=== Testing compression type incompatible bit setting for zlib ==="
 +echo
 +_make_test_img -o compression_type=zlib 64M
 +$PYTHON qcow2.py "$TEST_IMG" dump-header | grep incompatible_features
 +
 +echo
 +echo "=== Testing compression type incompatible bit setting for zstd ==="
 +echo
 +_make_test_img -o compression_type=zstd 64M
 +$PYTHON qcow2.py "$TEST_IMG" dump-header | grep incompatible_features
 +
 +echo
 +echo "=== Testing zlib with incompatible bit set ==="
 +echo
 +_make_test_img -o compression_type=zlib 64M
 +$PYTHON qcow2.py "$TEST_IMG" set-feature-bit incompatible 3
 +# to make sure the bit was actually set
 +$PYTHON qcow2.py "$TEST_IMG" dump-header | grep incompatible_features
 +
 +if $QEMU_IMG info "$TEST_IMG" >/dev/null 2>&1 ; then
 +    echo "Error: The image opened successfully. The image must not be opened."
 +fi
 +
 +echo
 +echo "=== Testing zstd with incompatible bit unset ==="
 +echo
 +_make_test_img -o compression_type=zstd 64M
 +$PYTHON qcow2.py "$TEST_IMG" set-header incompatible_features 0
 +# to make sure the bit was actually unset
 +$PYTHON qcow2.py "$TEST_IMG" dump-header | grep incompatible_features
 +
 +if $QEMU_IMG info "$TEST_IMG" >/dev/null 2>&1 ; then
 +    echo "Error: The image opened successfully. The image must not be opened."
 +fi
 +
 +echo
 +echo "=== Testing compression type values ==="
 +echo
 +# zlib=0
 +_make_test_img -o compression_type=zlib 64M
 +peek_file_be "$TEST_IMG" 104 1
 +echo
 +
 +# zstd=1
 +_make_test_img -o compression_type=zstd 64M
 +peek_file_be "$TEST_IMG" 104 1
 +echo
 +
 +echo
 +echo "=== Testing simple reading and writing with zstd ==="
 +echo
 +_make_test_img -o compression_type=zstd 64M
 +$QEMU_IO -c "write -c -P 0xAC 64K 64K " "$TEST_IMG" | _filter_qemu_io
 +$QEMU_IO -c "read -P 0xAC 64K 64K " "$TEST_IMG" | _filter_qemu_io
 +# read on the cluster boundaries
 +$QEMU_IO -c "read -v 131070 8 " "$TEST_IMG" | _filter_qemu_io
 +$QEMU_IO -c "read -v 65534 8" "$TEST_IMG" | _filter_qemu_io
 +
 +echo
 +echo "=== Testing adjacent clusters reading and writing with zstd ==="
 +echo
 +_make_test_img -o compression_type=zstd 64M
 +$QEMU_IO -c "write -c -P 0xAB 0 64K " "$TEST_IMG" | _filter_qemu_io
 +$QEMU_IO -c "write -c -P 0xAC 64K 64K " "$TEST_IMG" | _filter_qemu_io
 +$QEMU_IO -c "write -c -P 0xAD 128K 64K " "$TEST_IMG" | _filter_qemu_io
 +
 +$QEMU_IO -c "read -P 0xAB 0 64k " "$TEST_IMG" | _filter_qemu_io
 +$QEMU_IO -c "read -P 0xAC 64K 64k " "$TEST_IMG" | _filter_qemu_io
 +$QEMU_IO -c "read -P 0xAD 128K 64k " "$TEST_IMG" | _filter_qemu_io
 +
 +echo
 +echo "=== Testing incompressible cluster processing with zstd ==="
 +echo
 +# create a 2M image and fill it with 1M likely incompressible data
 +# and 1M compressible data
 +dd if=/dev/urandom of="$RAND_FILE" bs=1M count=1 seek=1
 +QEMU_IO_OPTIONS="$QEMU_IO_OPTIONS_NO_FMT" \
 +$QEMU_IO -f raw -c "write -P 0xFA 0 1M" "$RAND_FILE" | _filter_qemu_io
 +
 +$QEMU_IMG convert -f raw -O $IMGFMT -c \
 +-o "$(_optstr_add "$IMGOPTS" "compression_type=zlib")" "$RAND_FILE" \
 +"$TEST_IMG" | _filter_qemu_io
 +
 +$QEMU_IMG convert -O $IMGFMT -c \
 +-o "$(_optstr_add "$IMGOPTS" "compression_type=zstd")" "$TEST_IMG" \
 +"$COMPR_IMG" | _filter_qemu_io
 +
 +$QEMU_IMG compare "$TEST_IMG" "$COMPR_IMG"
 +
 +# success, all done
 +echo "*** done"
 +rm -f $seq.full
 +status=0
 diff --git a/tests/qemu-iotests/287.out b/tests/qemu-iotests/287.out
 new file mode 100644
 index XXXXXXX..XXXXXXX
 --- /dev/null
 +++ b/tests/qemu-iotests/287.out
@@ -XXX,XX +XXX,XX @@
 +QA output created by 287
 +
 +=== Testing compression type incompatible bit setting for zlib ===
 +
 +Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
 +incompatible_features     []
 +
 +=== Testing compression type incompatible bit setting for zstd ===
 +
 +Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
 +incompatible_features     [3]
 +
 +=== Testing zlib with incompatible bit set  ===
 +
 +Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
 +incompatible_features     [3]
 +
 +=== Testing zstd with incompatible bit unset  ===
 +
 +Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
 +incompatible_features     []
 +
 +=== Testing compression type values  ===
 +
 +Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
 +   0
 +Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
 +   1
 +
 +=== Testing simple reading and writing with zstd ===
 +
 +Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
 +wrote 65536/65536 bytes at offset 65536
 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +read 65536/65536 bytes at offset 65536
 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +0001fffe:  ac ac 00 00 00 00 00 00  ........
 +read 8/8 bytes at offset 131070
 +8 bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +0000fffe:  00 00 ac ac ac ac ac ac  ........
 +read 8/8 bytes at offset 65534
 +8 bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +
 +=== Testing adjacent clusters reading and writing with zstd ===
 +
 +Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
 +wrote 65536/65536 bytes at offset 0
 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +wrote 65536/65536 bytes at offset 65536
 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +wrote 65536/65536 bytes at offset 131072
 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +read 65536/65536 bytes at offset 0
 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +read 65536/65536 bytes at offset 65536
 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +read 65536/65536 bytes at offset 131072
 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +
 +=== Testing incompressible cluster processing with zstd ===
 +
 +1+0 records in
 +1+0 records out
 +wrote 1048576/1048576 bytes at offset 0
 +1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +Images are identical.
 +*** done
 diff --git a/tests/qemu-iotests/group b/tests/qemu-iotests/group
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/group
 +++ b/tests/qemu-iotests/group
@@ -XXX,XX +XXX,XX @@
 auto quick
 rw
 rw quick
 +287 auto quick
 quick
 rw quick
 rw auto quick
 --
-.26.2
+.9.3

The following changes since commit d5c75ec500d96f1d93447f990cd5a4ef5ba27fae:

Merge remote-tracking branch 'remotes/stefanberger/tags/pull-tpm-2020-05-08-1' into staging (2020-05-12 17:00:10 +0100)

are available in the Git repository at:

https://github.com/XanClic/qemu.git tags/pull-block-2020-05-13

for you to fetch changes up to fc9aefc8c0d3c6392656ea661ce72c1583b70bbd:

block/block-copy: fix use-after-free of task pointer (2020-05-13 14:20:31 +0200)

----------------------------------------------------------------
Block patches:
- zstd compression for qcow2
- Fix use-after-free

----------------------------------------------------------------
Denis Plotnikov (4):
  qcow2: introduce compression type feature
  qcow2: rework the cluster compression routine
  qcow2: add zstd cluster compression
  iotests: 287: add qcow2 compression type test

Vladimir Sementsov-Ogievskiy (1):
  block/block-copy: fix use-after-free of task pointer

-- 
2.26.2

From: Denis Plotnikov <dplotnikov@virtuozzo.com>

The patch adds some preparation parts for incompatible compression type
feature to qcow2 allowing the use different compression methods for
image clusters (de)compressing.

It is implied that the compression type is set on the image creation and
can be changed only later by image conversion, thus compression type
defines the only compression algorithm used for the image, and thus,
for all image clusters.

The goal of the feature is to add support of other compression methods
to qcow2. For example, ZSTD which is more effective on compression than ZLIB.

The default compression is ZLIB. Images created with ZLIB compression type
are backward compatible with older qemu versions.

Adding of the compression type breaks a number of tests because now the
compression type is reported on image creation and there are some changes
in the qcow2 header in size and offsets.

The tests are fixed in the following ways:
    * filter out compression_type for many tests
    * fix header size, feature table size and backing file offset
      affected tests: 031, 036, 061, 080
      header_size +=8: 1 byte compression type
                       7 bytes padding
      feature_table += 48: incompatible feature compression type
      backing_file_offset += 56 (8 + 48 -> header_change + feature_table_change)
    * add "compression type" for test output matching when it isn't filtered
      affected tests: 049, 060, 061, 065, 082, 085, 144, 182, 185, 198, 206,
                      242, 255, 274, 280

Signed-off-by: Denis Plotnikov <dplotnikov@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
QAPI part:
Acked-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20200507082521.29210-2-dplotnikov@virtuozzo.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 qapi/block-core.json             |  22 +++++-
 block/qcow2.h                    |  20 +++++-
 include/block/block_int.h        |   1 +
 block/qcow2.c                    | 113 +++++++++++++++++++++++++++++++
 tests/qemu-iotests/031.out       |  14 ++--
 tests/qemu-iotests/036.out       |   4 +-
 tests/qemu-iotests/049.out       | 102 ++++++++++++++--------------
 tests/qemu-iotests/060.out       |   1 +
 tests/qemu-iotests/061.out       |  34 ++++++----
 tests/qemu-iotests/065           |  28 +++++---
 tests/qemu-iotests/080           |   2 +-
 tests/qemu-iotests/082.out       |  48 +++++++++++--
 tests/qemu-iotests/085.out       |  38 +++++------
 tests/qemu-iotests/144.out       |   4 +-
 tests/qemu-iotests/182.out       |   2 +-
 tests/qemu-iotests/185.out       |   8 +--
 tests/qemu-iotests/198.out       |   2 +
 tests/qemu-iotests/206.out       |   5 ++
 tests/qemu-iotests/242.out       |   5 ++
 tests/qemu-iotests/255.out       |   8 +--
 tests/qemu-iotests/274.out       |  49 +++++++-------
 tests/qemu-iotests/280.out       |   2 +-
 tests/qemu-iotests/common.filter |   3 +-
 23 files changed, 365 insertions(+), 150 deletions(-)

diff --git a/qapi/block-core.json b/qapi/block-core.json
index XXXXXXX..XXXXXXX 100644
--- a/qapi/block-core.json
+++ b/qapi/block-core.json
@@ -XXX,XX +XXX,XX @@
 #
 # @bitmaps: A list of qcow2 bitmap details (since 4.0)
 #
+# @compression-type: the image cluster compression method (since 5.1)
+#
 # Since: 1.7
 ##
 { 'struct': 'ImageInfoSpecificQCow2',
@@ -XXX,XX +XXX,XX @@
       '*corrupt': 'bool',
       'refcount-bits': 'int',
       '*encrypt': 'ImageInfoSpecificQCow2Encryption',
-      '*bitmaps': ['Qcow2BitmapInfo']
+      '*bitmaps': ['Qcow2BitmapInfo'],
+      'compression-type': 'Qcow2CompressionType'
   } }
 
 ##
@@ -XXX,XX +XXX,XX @@
   'data': [ 'v2', 'v3' ] }
 
 
+##
+# @Qcow2CompressionType:
+#
+# Compression type used in qcow2 image file
+#
+# @zlib: zlib compression, see <http://zlib.net/>
+#
+# Since: 5.1
+##
+{ 'enum': 'Qcow2CompressionType',
+  'data': [ 'zlib' ] }
+
 ##
 # @BlockdevCreateOptionsQcow2:
 #
@@ -XXX,XX +XXX,XX @@
 #                 allowed values: off, falloc, full, metadata)
 # @lazy-refcounts: True if refcounts may be updated lazily (default: off)
 # @refcount-bits: Width of reference counts in bits (default: 16)
+# @compression-type: The image cluster compression method
+#                    (default: zlib, since 5.1)
 #
 # Since: 2.12
 ##
@@ -XXX,XX +XXX,XX @@
             '*cluster-size':    'size',
             '*preallocation':   'PreallocMode',
             '*lazy-refcounts':  'bool',
-            '*refcount-bits':   'int' } }
+            '*refcount-bits':   'int',
+            '*compression-type':'Qcow2CompressionType' } }
 
 ##
 # @BlockdevCreateOptionsQed:
diff --git a/block/qcow2.h b/block/qcow2.h
index XXXXXXX..XXXXXXX 100644
--- a/block/qcow2.h
+++ b/block/qcow2.h
@@ -XXX,XX +XXX,XX @@ typedef struct QCowHeader {
 
     uint32_t refcount_order;
     uint32_t header_length;
+
+    /* Additional fields */
+    uint8_t compression_type;
+
+    /* header must be a multiple of 8 */
+    uint8_t padding[7];
 } QEMU_PACKED QCowHeader;
 
+QEMU_BUILD_BUG_ON(!QEMU_IS_ALIGNED(sizeof(QCowHeader), 8));
+
 typedef struct QEMU_PACKED QCowSnapshotHeader {
     /* header is 8 byte aligned */
     uint64_t l1_table_offset;
@@ -XXX,XX +XXX,XX @@ enum {
     QCOW2_INCOMPAT_DIRTY_BITNR      = 0,
     QCOW2_INCOMPAT_CORRUPT_BITNR    = 1,
     QCOW2_INCOMPAT_DATA_FILE_BITNR  = 2,
+    QCOW2_INCOMPAT_COMPRESSION_BITNR = 3,
     QCOW2_INCOMPAT_DIRTY            = 1 << QCOW2_INCOMPAT_DIRTY_BITNR,
     QCOW2_INCOMPAT_CORRUPT          = 1 << QCOW2_INCOMPAT_CORRUPT_BITNR,
     QCOW2_INCOMPAT_DATA_FILE        = 1 << QCOW2_INCOMPAT_DATA_FILE_BITNR,
+    QCOW2_INCOMPAT_COMPRESSION      = 1 << QCOW2_INCOMPAT_COMPRESSION_BITNR,
 
     QCOW2_INCOMPAT_MASK             = QCOW2_INCOMPAT_DIRTY
                                     | QCOW2_INCOMPAT_CORRUPT
-                                    | QCOW2_INCOMPAT_DATA_FILE,
+                                    | QCOW2_INCOMPAT_DATA_FILE
+                                    | QCOW2_INCOMPAT_COMPRESSION,
 };
 
 /* Compatible feature bits */
@@ -XXX,XX +XXX,XX @@ typedef struct BDRVQcow2State {
 
     bool metadata_preallocation_checked;
     bool metadata_preallocation;
+    /*
+     * Compression type used for the image. Default: 0 - ZLIB
+     * The image compression type is set on image creation.
+     * For now, the only way to change the compression type
+     * is to convert the image with the desired compression type set.
+     */
+    Qcow2CompressionType compression_type;
 } BDRVQcow2State;
 
 typedef struct Qcow2COWRegion {
diff --git a/include/block/block_int.h b/include/block/block_int.h
index XXXXXXX..XXXXXXX 100644
--- a/include/block/block_int.h
+++ b/include/block/block_int.h
@@ -XXX,XX +XXX,XX @@
 #define BLOCK_OPT_REFCOUNT_BITS     "refcount_bits"
 #define BLOCK_OPT_DATA_FILE         "data_file"
 #define BLOCK_OPT_DATA_FILE_RAW     "data_file_raw"
+#define BLOCK_OPT_COMPRESSION_TYPE  "compression_type"
 
 #define BLOCK_PROBE_BUF_SIZE        512
 
diff --git a/block/qcow2.c b/block/qcow2.c
index XXXXXXX..XXXXXXX 100644
--- a/block/qcow2.c
+++ b/block/qcow2.c
@@ -XXX,XX +XXX,XX @@ static int qcow2_update_options(BlockDriverState *bs, QDict *options,
     return ret;
 }
 
+static int validate_compression_type(BDRVQcow2State *s, Error **errp)
+{
+    switch (s->compression_type) {
+    case QCOW2_COMPRESSION_TYPE_ZLIB:
+        break;
+
+    default:
+        error_setg(errp, "qcow2: unknown compression type: %u",
+                   s->compression_type);
+        return -ENOTSUP;
+    }
+
+    /*
+     * if the compression type differs from QCOW2_COMPRESSION_TYPE_ZLIB
+     * the incompatible feature flag must be set
+     */
+    if (s->compression_type == QCOW2_COMPRESSION_TYPE_ZLIB) {
+        if (s->incompatible_features & QCOW2_INCOMPAT_COMPRESSION) {
+            error_setg(errp, "qcow2: Compression type incompatible feature "
+                             "bit must not be set");
+            return -EINVAL;
+        }
+    } else {
+        if (!(s->incompatible_features & QCOW2_INCOMPAT_COMPRESSION)) {
+            error_setg(errp, "qcow2: Compression type incompatible feature "
+                             "bit must be set");
+            return -EINVAL;
+        }
+    }
+
+    return 0;
+}
+
 /* Called with s->lock held.  */
 static int coroutine_fn qcow2_do_open(BlockDriverState *bs, QDict *options,
                                       int flags, Error **errp)
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_do_open(BlockDriverState *bs, QDict *options,
     s->compatible_features      = header.compatible_features;
     s->autoclear_features       = header.autoclear_features;
 
+    /*
+     * Handle compression type
+     * Older qcow2 images don't contain the compression type header.
+     * Distinguish them by the header length and use
+     * the only valid (default) compression type in that case
+     */
+    if (header.header_length > offsetof(QCowHeader, compression_type)) {
+        s->compression_type = header.compression_type;
+    } else {
+        s->compression_type = QCOW2_COMPRESSION_TYPE_ZLIB;
+    }
+
+    ret = validate_compression_type(s, errp);
+    if (ret) {
+        goto fail;
+    }
+
     if (s->incompatible_features & ~QCOW2_INCOMPAT_MASK) {
         void *feature_table = NULL;
         qcow2_read_extensions(bs, header.header_length, ext_end,
@@ -XXX,XX +XXX,XX @@ int qcow2_update_header(BlockDriverState *bs)
     total_size = bs->total_sectors * BDRV_SECTOR_SIZE;
     refcount_table_clusters = s->refcount_table_size >> (s->cluster_bits - 3);
 
+    ret = validate_compression_type(s, NULL);
+    if (ret) {
+        goto fail;
+    }
+
     *header = (QCowHeader) {
         /* Version 2 fields */
         .magic                  = cpu_to_be32(QCOW_MAGIC),
@@ -XXX,XX +XXX,XX @@ int qcow2_update_header(BlockDriverState *bs)
         .autoclear_features     = cpu_to_be64(s->autoclear_features),
         .refcount_order         = cpu_to_be32(s->refcount_order),
         .header_length          = cpu_to_be32(header_length),
+        .compression_type       = s->compression_type,
     };
 
     /* For older versions, write a shorter header */
@@ -XXX,XX +XXX,XX @@ int qcow2_update_header(BlockDriverState *bs)
                 .bit  = QCOW2_INCOMPAT_DATA_FILE_BITNR,
                 .name = "external data file",
             },
+            {
+                .type = QCOW2_FEAT_TYPE_INCOMPATIBLE,
+                .bit  = QCOW2_INCOMPAT_COMPRESSION_BITNR,
+                .name = "compression type",
+            },
             {
                 .type = QCOW2_FEAT_TYPE_COMPATIBLE,
                 .bit  = QCOW2_COMPAT_LAZY_REFCOUNTS_BITNR,
@@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp)
     uint64_t* refcount_table;
     Error *local_err = NULL;
     int ret;
+    uint8_t compression_type = QCOW2_COMPRESSION_TYPE_ZLIB;
 
     assert(create_options->driver == BLOCKDEV_DRIVER_QCOW2);
     qcow2_opts = &create_options->u.qcow2;
@@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp)
         }
     }
 
+    if (qcow2_opts->has_compression_type &&
+        qcow2_opts->compression_type != QCOW2_COMPRESSION_TYPE_ZLIB) {
+
+        ret = -EINVAL;
+
+        if (version < 3) {
+            error_setg(errp, "Non-zlib compression type is only supported with "
+                       "compatibility level 1.1 and above (use version=v3 or "
+                       "greater)");
+            goto out;
+        }
+
+        switch (qcow2_opts->compression_type) {
+        default:
+            error_setg(errp, "Unknown compression type");
+            goto out;
+        }
+
+        compression_type = qcow2_opts->compression_type;
+    }
+
     /* Create BlockBackend to write to the image */
     blk = blk_new_with_bs(bs, BLK_PERM_WRITE | BLK_PERM_RESIZE, BLK_PERM_ALL,
                           errp);
@@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp)
         .refcount_table_offset      = cpu_to_be64(cluster_size),
         .refcount_table_clusters    = cpu_to_be32(1),
         .refcount_order             = cpu_to_be32(refcount_order),
+        /* don't deal with endianness since compression_type is 1 byte long */
+        .compression_type           = compression_type,
         .header_length              = cpu_to_be32(sizeof(*header)),
     };
 
@@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp)
         header->autoclear_features |=
             cpu_to_be64(QCOW2_AUTOCLEAR_DATA_FILE_RAW);
     }
+    if (compression_type != QCOW2_COMPRESSION_TYPE_ZLIB) {
+        header->incompatible_features |=
+            cpu_to_be64(QCOW2_INCOMPAT_COMPRESSION);
+    }
 
     ret = blk_pwrite(blk, 0, header, cluster_size, 0);
     g_free(header);
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_create_opts(BlockDriver *drv,
         { BLOCK_OPT_ENCRYPT,            BLOCK_OPT_ENCRYPT_FORMAT },
         { BLOCK_OPT_COMPAT_LEVEL,       "version" },
         { BLOCK_OPT_DATA_FILE_RAW,      "data-file-raw" },
+        { BLOCK_OPT_COMPRESSION_TYPE,   "compression-type" },
         { NULL, NULL },
     };
 
@@ -XXX,XX +XXX,XX @@ static ImageInfoSpecific *qcow2_get_specific_info(BlockDriverState *bs,
             .data_file          = g_strdup(s->image_data_file),
             .has_data_file_raw  = has_data_file(bs),
             .data_file_raw      = data_file_is_raw(bs),
+            .compression_type   = s->compression_type,
         };
     } else {
         /* if this assertion fails, this probably means a new version was
@@ -XXX,XX +XXX,XX @@ static int qcow2_amend_options(BlockDriverState *bs, QemuOpts *opts,
                                  "images");
                 return -EINVAL;
             }
+        } else if (!strcmp(desc->name, BLOCK_OPT_COMPRESSION_TYPE)) {
+            const char *ct_name =
+                qemu_opt_get(opts, BLOCK_OPT_COMPRESSION_TYPE);
+            int compression_type =
+                qapi_enum_parse(&Qcow2CompressionType_lookup, ct_name, -1,
+                                NULL);
+            if (compression_type == -1) {
+                error_setg(errp, "Unknown compression type: %s", ct_name);
+                return -ENOTSUP;
+            }
+
+            if (compression_type != s->compression_type) {
+                error_setg(errp, "Changing the compression type "
+                                 "is not supported");
+                return -ENOTSUP;
+            }
         } else {
             /* if this point is reached, this probably means a new option was
              * added without having it covered here */
@@ -XXX,XX +XXX,XX @@ static QemuOptsList qcow2_create_opts = {
             .help = "Width of a reference count entry in bits",
             .def_value_str = "16"
         },
+        {
+            .name = BLOCK_OPT_COMPRESSION_TYPE,
+            .type = QEMU_OPT_STRING,
+            .help = "Compression method used for image cluster compression",
+            .def_value_str = "zlib"
+        },
         { /* end of list */ }
     }
 };
diff --git a/tests/qemu-iotests/031.out b/tests/qemu-iotests/031.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/031.out
+++ b/tests/qemu-iotests/031.out
@@ -XXX,XX +XXX,XX @@ incompatible_features     []
 compatible_features       []
 autoclear_features        []
 refcount_order            4
-header_length             104
+header_length             112
 
 Header extension:
 magic                     0x6803f857
-length                    288
+length                    336
 data                      <binary>
 
 Header extension:
@@ -XXX,XX +XXX,XX @@ incompatible_features     []
 compatible_features       []
 autoclear_features        []
 refcount_order            4
-header_length             104
+header_length             112
 
 Header extension:
 magic                     0x6803f857
-length                    288
+length                    336
 data                      <binary>
 
 Header extension:
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 
 magic                     0x514649fb
 version                   3
-backing_file_offset       0x1d8
+backing_file_offset       0x210
 backing_file_size         0x17
 cluster_bits              16
 size                      67108864
@@ -XXX,XX +XXX,XX @@ incompatible_features     []
 compatible_features       []
 autoclear_features        []
 refcount_order            4
-header_length             104
+header_length             112
 
 Header extension:
 magic                     0xe2792aca
@@ -XXX,XX +XXX,XX @@ data                      'host_device'
 
 Header extension:
 magic                     0x6803f857
-length                    288
+length                    336
 data                      <binary>
 
 Header extension:
diff --git a/tests/qemu-iotests/036.out b/tests/qemu-iotests/036.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/036.out
+++ b/tests/qemu-iotests/036.out
@@ -XXX,XX +XXX,XX @@ compatible_features       []
 autoclear_features        [63]
 Header extension:
 magic                     0x6803f857
-length                    288
+length                    336
 data                      <binary>
 
 
@@ -XXX,XX +XXX,XX @@ compatible_features       []
 autoclear_features        []
 Header extension:
 magic                     0x6803f857
-length                    288
+length                    336
 data                      <binary>
 
 *** done
diff --git a/tests/qemu-iotests/049.out b/tests/qemu-iotests/049.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/049.out
+++ b/tests/qemu-iotests/049.out
@@ -XXX,XX +XXX,XX @@ QA output created by 049
 == 1. Traditional size parameter ==
 
 qemu-img create -f qcow2 TEST_DIR/t.qcow2 1024
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 TEST_DIR/t.qcow2 1024b
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 TEST_DIR/t.qcow2 1k
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 TEST_DIR/t.qcow2 1K
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 TEST_DIR/t.qcow2 1M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1048576 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1048576 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 TEST_DIR/t.qcow2 1G
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1073741824 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1073741824 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 TEST_DIR/t.qcow2 1T
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1099511627776 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1099511627776 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 TEST_DIR/t.qcow2 1024.0
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 TEST_DIR/t.qcow2 1024.0b
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 TEST_DIR/t.qcow2 1.5k
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1536 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1536 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 TEST_DIR/t.qcow2 1.5K
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1536 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1536 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 TEST_DIR/t.qcow2 1.5M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1572864 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1572864 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 TEST_DIR/t.qcow2 1.5G
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1610612736 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1610612736 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 TEST_DIR/t.qcow2 1.5T
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1649267441664 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1649267441664 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 == 2. Specifying size via -o ==
 
 qemu-img create -f qcow2 -o size=1024 TEST_DIR/t.qcow2
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o size=1024b TEST_DIR/t.qcow2
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o size=1k TEST_DIR/t.qcow2
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o size=1K TEST_DIR/t.qcow2
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o size=1M TEST_DIR/t.qcow2
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1048576 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1048576 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o size=1G TEST_DIR/t.qcow2
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1073741824 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1073741824 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o size=1T TEST_DIR/t.qcow2
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1099511627776 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1099511627776 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o size=1024.0 TEST_DIR/t.qcow2
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o size=1024.0b TEST_DIR/t.qcow2
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o size=1.5k TEST_DIR/t.qcow2
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1536 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1536 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o size=1.5K TEST_DIR/t.qcow2
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1536 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1536 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o size=1.5M TEST_DIR/t.qcow2
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1572864 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1572864 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o size=1.5G TEST_DIR/t.qcow2
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1610612736 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1610612736 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o size=1.5T TEST_DIR/t.qcow2
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1649267441664 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1649267441664 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 == 3. Invalid sizes ==
 
@@ -XXX,XX +XXX,XX @@ qemu-img: TEST_DIR/t.qcow2: The image size must be specified only once
 == Check correct interpretation of suffixes for cluster size ==
 
 qemu-img create -f qcow2 -o cluster_size=1024 TEST_DIR/t.qcow2 64M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o cluster_size=1024b TEST_DIR/t.qcow2 64M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o cluster_size=1k TEST_DIR/t.qcow2 64M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o cluster_size=1K TEST_DIR/t.qcow2 64M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o cluster_size=1M TEST_DIR/t.qcow2 64M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1048576 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1048576 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o cluster_size=1024.0 TEST_DIR/t.qcow2 64M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o cluster_size=1024.0b TEST_DIR/t.qcow2 64M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=1024 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o cluster_size=0.5k TEST_DIR/t.qcow2 64M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=512 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=512 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o cluster_size=0.5K TEST_DIR/t.qcow2 64M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=512 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=512 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o cluster_size=0.5M TEST_DIR/t.qcow2 64M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=524288 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=524288 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 == Check compat level option ==
 
 qemu-img create -f qcow2 -o compat=0.10 TEST_DIR/t.qcow2 64M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=0.10 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=0.10 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o compat=1.1 TEST_DIR/t.qcow2 64M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=1.1 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=1.1 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o compat=0.42 TEST_DIR/t.qcow2 64M
 qemu-img: TEST_DIR/t.qcow2: Invalid parameter '0.42'
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=0.42 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=0.42 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o compat=foobar TEST_DIR/t.qcow2 64M
 qemu-img: TEST_DIR/t.qcow2: Invalid parameter 'foobar'
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=foobar cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=foobar cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 == Check preallocation option ==
 
 qemu-img create -f qcow2 -o preallocation=off TEST_DIR/t.qcow2 64M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=65536 preallocation=off lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=65536 preallocation=off lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o preallocation=metadata TEST_DIR/t.qcow2 64M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=65536 preallocation=metadata lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=65536 preallocation=metadata lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o preallocation=1234 TEST_DIR/t.qcow2 64M
 qemu-img: TEST_DIR/t.qcow2: Invalid parameter '1234'
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=65536 preallocation=1234 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=65536 preallocation=1234 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 == Check encryption option ==
 
 qemu-img create -f qcow2 -o encryption=off TEST_DIR/t.qcow2 64M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 encryption=off cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 encryption=off cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 --object secret,id=sec0,data=123456 -o encryption=on,encrypt.key-secret=sec0 TEST_DIR/t.qcow2 64M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 encryption=on encrypt.key-secret=sec0 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 encryption=on encrypt.key-secret=sec0 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 == Check lazy_refcounts option (only with v3) ==
 
 qemu-img create -f qcow2 -o compat=1.1,lazy_refcounts=off TEST_DIR/t.qcow2 64M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=1.1 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=1.1 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o compat=1.1,lazy_refcounts=on TEST_DIR/t.qcow2 64M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=1.1 cluster_size=65536 lazy_refcounts=on refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=1.1 cluster_size=65536 lazy_refcounts=on refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o compat=0.10,lazy_refcounts=off TEST_DIR/t.qcow2 64M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=0.10 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=0.10 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 qemu-img create -f qcow2 -o compat=0.10,lazy_refcounts=on TEST_DIR/t.qcow2 64M
 qemu-img: TEST_DIR/t.qcow2: Lazy refcounts only supported with compatibility level 1.1 and above (use version=v3 or greater)
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=0.10 cluster_size=65536 lazy_refcounts=on refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=0.10 cluster_size=65536 lazy_refcounts=on refcount_bits=16 compression_type=zlib
 
 *** done
diff --git a/tests/qemu-iotests/060.out b/tests/qemu-iotests/060.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/060.out
+++ b/tests/qemu-iotests/060.out
@@ -XXX,XX +XXX,XX @@ virtual size: 64 MiB (67108864 bytes)
 cluster_size: 65536
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: false
     refcount bits: 16
     corrupt: true
diff --git a/tests/qemu-iotests/061.out b/tests/qemu-iotests/061.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/061.out
+++ b/tests/qemu-iotests/061.out
@@ -XXX,XX +XXX,XX @@ incompatible_features     []
 compatible_features       [0]
 autoclear_features        []
 refcount_order            4
-header_length             104
+header_length             112
 
 Header extension:
 magic                     0x6803f857
-length                    288
+length                    336
 data                      <binary>
 
 magic                     0x514649fb
@@ -XXX,XX +XXX,XX @@ incompatible_features     []
 compatible_features       [0]
 autoclear_features        []
 refcount_order            4
-header_length             104
+header_length             112
 
 Header extension:
 magic                     0x6803f857
-length                    288
+length                    336
 data                      <binary>
 
 magic                     0x514649fb
@@ -XXX,XX +XXX,XX @@ incompatible_features     [0]
 compatible_features       [0]
 autoclear_features        []
 refcount_order            4
-header_length             104
+header_length             112
 
 Header extension:
 magic                     0x6803f857
-length                    288
+length                    336
 data                      <binary>
 
 ERROR cluster 5 refcount=0 reference=1
@@ -XXX,XX +XXX,XX @@ incompatible_features     []
 compatible_features       [42]
 autoclear_features        [42]
 refcount_order            4
-header_length             104
+header_length             112
 
 Header extension:
 magic                     0x6803f857
-length                    288
+length                    336
 data                      <binary>
 
 magic                     0x514649fb
@@ -XXX,XX +XXX,XX @@ incompatible_features     []
 compatible_features       [0]
 autoclear_features        []
 refcount_order            4
-header_length             104
+header_length             112
 
 Header extension:
 magic                     0x6803f857
-length                    288
+length                    336
 data                      <binary>
 
 read 65536/65536 bytes at offset 44040192
@@ -XXX,XX +XXX,XX @@ incompatible_features     [0]
 compatible_features       [0]
 autoclear_features        []
 refcount_order            4
-header_length             104
+header_length             112
 
 Header extension:
 magic                     0x6803f857
-length                    288
+length                    336
 data                      <binary>
 
 ERROR cluster 5 refcount=0 reference=1
@@ -XXX,XX +XXX,XX @@ incompatible_features     []
 compatible_features       []
 autoclear_features        []
 refcount_order            4
-header_length             104
+header_length             112
 
 Header extension:
 magic                     0x6803f857
-length                    288
+length                    336
 data                      <binary>
 
 read 131072/131072 bytes at offset 0
@@ -XXX,XX +XXX,XX @@ virtual size: 64 MiB (67108864 bytes)
 cluster_size: 65536
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: false
     refcount bits: 16
     data file: TEST_DIR/t.IMGFMT.data
@@ -XXX,XX +XXX,XX @@ virtual size: 64 MiB (67108864 bytes)
 cluster_size: 65536
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: false
     refcount bits: 16
     data file: foo
@@ -XXX,XX +XXX,XX @@ virtual size: 64 MiB (67108864 bytes)
 cluster_size: 65536
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: false
     refcount bits: 16
     data file raw: false
@@ -XXX,XX +XXX,XX @@ virtual size: 64 MiB (67108864 bytes)
 cluster_size: 65536
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: false
     refcount bits: 16
     data file: TEST_DIR/t.IMGFMT.data
@@ -XXX,XX +XXX,XX @@ virtual size: 64 MiB (67108864 bytes)
 cluster_size: 65536
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: false
     refcount bits: 16
     data file: TEST_DIR/t.IMGFMT.data
@@ -XXX,XX +XXX,XX @@ virtual size: 64 MiB (67108864 bytes)
 cluster_size: 65536
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: false
     refcount bits: 16
     data file: TEST_DIR/t.IMGFMT.data
diff --git a/tests/qemu-iotests/065 b/tests/qemu-iotests/065
index XXXXXXX..XXXXXXX 100755
--- a/tests/qemu-iotests/065
+++ b/tests/qemu-iotests/065
@@ -XXX,XX +XXX,XX @@ class TestQMP(TestImageInfoSpecific):
 class TestQCow2(TestQemuImgInfo):
     '''Testing a qcow2 version 2 image'''
     img_options = 'compat=0.10'
-    json_compare = { 'compat': '0.10', 'refcount-bits': 16 }
-    human_compare = [ 'compat: 0.10', 'refcount bits: 16' ]
+    json_compare = { 'compat': '0.10', 'refcount-bits': 16,
+                     'compression-type': 'zlib' }
+    human_compare = [ 'compat: 0.10', 'compression type: zlib',
+                      'refcount bits: 16' ]
 
 class TestQCow3NotLazy(TestQemuImgInfo):
     '''Testing a qcow2 version 3 image with lazy refcounts disabled'''
     img_options = 'compat=1.1,lazy_refcounts=off'
     json_compare = { 'compat': '1.1', 'lazy-refcounts': False,
-                     'refcount-bits': 16, 'corrupt': False }
-    human_compare = [ 'compat: 1.1', 'lazy refcounts: false',
-                      'refcount bits: 16', 'corrupt: false' ]
+                     'refcount-bits': 16, 'corrupt': False,
+                     'compression-type': 'zlib' }
+    human_compare = [ 'compat: 1.1', 'compression type: zlib',
+                      'lazy refcounts: false', 'refcount bits: 16',
+                      'corrupt: false' ]
 
 class TestQCow3Lazy(TestQemuImgInfo):
     '''Testing a qcow2 version 3 image with lazy refcounts enabled'''
     img_options = 'compat=1.1,lazy_refcounts=on'
     json_compare = { 'compat': '1.1', 'lazy-refcounts': True,
-                     'refcount-bits': 16, 'corrupt': False }
-    human_compare = [ 'compat: 1.1', 'lazy refcounts: true',
-                      'refcount bits: 16', 'corrupt: false' ]
+                     'refcount-bits': 16, 'corrupt': False,
+                     'compression-type': 'zlib' }
+    human_compare = [ 'compat: 1.1', 'compression type: zlib',
+                      'lazy refcounts: true', 'refcount bits: 16',
+                      'corrupt: false' ]
 
 class TestQCow3NotLazyQMP(TestQMP):
     '''Testing a qcow2 version 3 image with lazy refcounts disabled, opening
@@ -XXX,XX +XXX,XX @@ class TestQCow3NotLazyQMP(TestQMP):
     img_options = 'compat=1.1,lazy_refcounts=off'
     qemu_options = 'lazy-refcounts=on'
     compare = { 'compat': '1.1', 'lazy-refcounts': False,
-                'refcount-bits': 16, 'corrupt': False }
+                'refcount-bits': 16, 'corrupt': False,
+                'compression-type': 'zlib' }
 
 
 class TestQCow3LazyQMP(TestQMP):
@@ -XXX,XX +XXX,XX @@ class TestQCow3LazyQMP(TestQMP):
     img_options = 'compat=1.1,lazy_refcounts=on'
     qemu_options = 'lazy-refcounts=off'
     compare = { 'compat': '1.1', 'lazy-refcounts': True,
-                'refcount-bits': 16, 'corrupt': False }
+                'refcount-bits': 16, 'corrupt': False,
+                'compression-type': 'zlib' }
 
 TestImageInfoSpecific = None
 TestQemuImgInfo = None
diff --git a/tests/qemu-iotests/080 b/tests/qemu-iotests/080
index XXXXXXX..XXXXXXX 100755
--- a/tests/qemu-iotests/080
+++ b/tests/qemu-iotests/080
@@ -XXX,XX +XXX,XX @@ _supported_os Linux
 # - This is generally a test for compat=1.1 images
 _unsupported_imgopts 'refcount_bits=1[^0-9]' data_file 'compat=0.10'
 
-header_size=104
+header_size=112
 
 offset_backing_file_offset=8
 offset_backing_file_size=16
diff --git a/tests/qemu-iotests/082.out b/tests/qemu-iotests/082.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/082.out
+++ b/tests/qemu-iotests/082.out
@@ -XXX,XX +XXX,XX @@ QA output created by 082
 === create: Options specified more than once ===
 
 Testing: create -f foo -f qcow2 TEST_DIR/t.qcow2 128M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 image: TEST_DIR/t.IMGFMT
 file format: IMGFMT
 virtual size: 128 MiB (134217728 bytes)
 cluster_size: 65536
 
 Testing: create -f qcow2 -o cluster_size=4k -o lazy_refcounts=on TEST_DIR/t.qcow2 128M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=4096 lazy_refcounts=on refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=4096 lazy_refcounts=on refcount_bits=16 compression_type=zlib
 image: TEST_DIR/t.IMGFMT
 file format: IMGFMT
 virtual size: 128 MiB (134217728 bytes)
 cluster_size: 4096
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: true
     refcount bits: 16
     corrupt: false
 
 Testing: create -f qcow2 -o cluster_size=4k -o lazy_refcounts=on -o cluster_size=8k TEST_DIR/t.qcow2 128M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=8192 lazy_refcounts=on refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=8192 lazy_refcounts=on refcount_bits=16 compression_type=zlib
 image: TEST_DIR/t.IMGFMT
 file format: IMGFMT
 virtual size: 128 MiB (134217728 bytes)
 cluster_size: 8192
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: true
     refcount bits: 16
     corrupt: false
 
 Testing: create -f qcow2 -o cluster_size=4k,cluster_size=8k TEST_DIR/t.qcow2 128M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=8192 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=8192 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 image: TEST_DIR/t.IMGFMT
 file format: IMGFMT
 virtual size: 128 MiB (134217728 bytes)
@@ -XXX,XX +XXX,XX @@ Supported options:
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
   size=<size>            - Virtual disk size
 
 Testing: create -f qcow2 -u -o backing_file=TEST_DIR/t.qcow2,,help TEST_DIR/t.qcow2 128M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/t.qcow2,,help cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/t.qcow2,,help cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 Testing: create -f qcow2 -u -o backing_file=TEST_DIR/t.qcow2,,? TEST_DIR/t.qcow2 128M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/t.qcow2,,? cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/t.qcow2,,? cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 Testing: create -f qcow2 -o backing_file=TEST_DIR/t.qcow2, -o help TEST_DIR/t.qcow2 128M
 qemu-img: Invalid option list: backing_file=TEST_DIR/t.qcow2,
@@ -XXX,XX +XXX,XX @@ Supported qcow2 options:
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ qemu-img: Format driver 'bochs' does not support image creation
 === convert: Options specified more than once ===
 
 Testing: create -f qcow2 TEST_DIR/t.qcow2 128M
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 Testing: convert -f foo -f qcow2 TEST_DIR/t.qcow2 TEST_DIR/t.qcow2.base
 image: TEST_DIR/t.IMGFMT.base
@@ -XXX,XX +XXX,XX @@ virtual size: 128 MiB (134217728 bytes)
 cluster_size: 4096
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: true
     refcount bits: 16
     corrupt: false
@@ -XXX,XX +XXX,XX @@ virtual size: 128 MiB (134217728 bytes)
 cluster_size: 8192
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: true
     refcount bits: 16
     corrupt: false
@@ -XXX,XX +XXX,XX @@ Supported options:
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported options:
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Supported qcow2 options:
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ virtual size: 128 MiB (134217728 bytes)
 cluster_size: 65536
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: true
     refcount bits: 16
     corrupt: false
@@ -XXX,XX +XXX,XX @@ virtual size: 130 MiB (136314880 bytes)
 cluster_size: 65536
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: false
     refcount bits: 16
     corrupt: false
@@ -XXX,XX +XXX,XX @@ virtual size: 132 MiB (138412032 bytes)
 cluster_size: 65536
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: true
     refcount bits: 16
     corrupt: false
@@ -XXX,XX +XXX,XX @@ Creation options for 'qcow2':
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Creation options for 'qcow2':
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Creation options for 'qcow2':
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Creation options for 'qcow2':
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Creation options for 'qcow2':
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Creation options for 'qcow2':
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Creation options for 'qcow2':
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Creation options for 'qcow2':
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
@@ -XXX,XX +XXX,XX @@ Creation options for 'qcow2':
   backing_fmt=<str>      - Image format of the base image
   cluster_size=<size>    - qcow2 cluster size
   compat=<str>           - Compatibility level (v2 [0.10] or v3 [1.1])
+  compression_type=<str> - Compression method used for image cluster compression
   data_file=<str>        - File name of an external data file
   data_file_raw=<bool (on/off)> - The external data file must stay valid as a raw image
   encrypt.cipher-alg=<str> - Name of encryption cipher algorithm
diff --git a/tests/qemu-iotests/085.out b/tests/qemu-iotests/085.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/085.out
+++ b/tests/qemu-iotests/085.out
@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/t.IMGFMT.2', fmt=IMGFMT size=134217728
 === Create a single snapshot on virtio0 ===
 
 { 'execute': 'blockdev-snapshot-sync', 'arguments': { 'device': 'virtio0', 'snapshot-file':'TEST_DIR/1-snapshot-v0.IMGFMT', 'format': 'IMGFMT' } }
-Formatting 'TEST_DIR/1-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/t.qcow2.1 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/1-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/t.qcow2.1 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 {"return": {}}
 
 === Invalid command - missing device and nodename ===
@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/1-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file
 === Create several transactional group snapshots ===
 
 { 'execute': 'transaction', 'arguments': {'actions': [ { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio0', 'snapshot-file': 'TEST_DIR/2-snapshot-v0.IMGFMT' } }, { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio1', 'snapshot-file': 'TEST_DIR/2-snapshot-v1.IMGFMT' } } ] } }
-Formatting 'TEST_DIR/2-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/1-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
-Formatting 'TEST_DIR/2-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/t.qcow2.2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/2-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/1-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
+Formatting 'TEST_DIR/2-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/t.qcow2.2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 {"return": {}}
 { 'execute': 'transaction', 'arguments': {'actions': [ { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio0', 'snapshot-file': 'TEST_DIR/3-snapshot-v0.IMGFMT' } }, { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio1', 'snapshot-file': 'TEST_DIR/3-snapshot-v1.IMGFMT' } } ] } }
-Formatting 'TEST_DIR/3-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/2-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
-Formatting 'TEST_DIR/3-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/2-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/3-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/2-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
+Formatting 'TEST_DIR/3-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/2-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 {"return": {}}
 { 'execute': 'transaction', 'arguments': {'actions': [ { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio0', 'snapshot-file': 'TEST_DIR/4-snapshot-v0.IMGFMT' } }, { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio1', 'snapshot-file': 'TEST_DIR/4-snapshot-v1.IMGFMT' } } ] } }
-Formatting 'TEST_DIR/4-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/3-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
-Formatting 'TEST_DIR/4-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/3-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/4-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/3-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
+Formatting 'TEST_DIR/4-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/3-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 {"return": {}}
 { 'execute': 'transaction', 'arguments': {'actions': [ { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio0', 'snapshot-file': 'TEST_DIR/5-snapshot-v0.IMGFMT' } }, { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio1', 'snapshot-file': 'TEST_DIR/5-snapshot-v1.IMGFMT' } } ] } }
-Formatting 'TEST_DIR/5-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/4-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
-Formatting 'TEST_DIR/5-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/4-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/5-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/4-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
+Formatting 'TEST_DIR/5-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/4-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 {"return": {}}
 { 'execute': 'transaction', 'arguments': {'actions': [ { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio0', 'snapshot-file': 'TEST_DIR/6-snapshot-v0.IMGFMT' } }, { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio1', 'snapshot-file': 'TEST_DIR/6-snapshot-v1.IMGFMT' } } ] } }
-Formatting 'TEST_DIR/6-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/5-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
-Formatting 'TEST_DIR/6-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/5-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/6-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/5-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
+Formatting 'TEST_DIR/6-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/5-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 {"return": {}}
 { 'execute': 'transaction', 'arguments': {'actions': [ { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio0', 'snapshot-file': 'TEST_DIR/7-snapshot-v0.IMGFMT' } }, { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio1', 'snapshot-file': 'TEST_DIR/7-snapshot-v1.IMGFMT' } } ] } }
-Formatting 'TEST_DIR/7-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/6-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
-Formatting 'TEST_DIR/7-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/6-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/7-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/6-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
+Formatting 'TEST_DIR/7-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/6-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 {"return": {}}
 { 'execute': 'transaction', 'arguments': {'actions': [ { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio0', 'snapshot-file': 'TEST_DIR/8-snapshot-v0.IMGFMT' } }, { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio1', 'snapshot-file': 'TEST_DIR/8-snapshot-v1.IMGFMT' } } ] } }
-Formatting 'TEST_DIR/8-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/7-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
-Formatting 'TEST_DIR/8-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/7-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/8-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/7-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
+Formatting 'TEST_DIR/8-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/7-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 {"return": {}}
 { 'execute': 'transaction', 'arguments': {'actions': [ { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio0', 'snapshot-file': 'TEST_DIR/9-snapshot-v0.IMGFMT' } }, { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio1', 'snapshot-file': 'TEST_DIR/9-snapshot-v1.IMGFMT' } } ] } }
-Formatting 'TEST_DIR/9-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/8-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
-Formatting 'TEST_DIR/9-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/8-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/9-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/8-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
+Formatting 'TEST_DIR/9-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/8-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 {"return": {}}
 { 'execute': 'transaction', 'arguments': {'actions': [ { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio0', 'snapshot-file': 'TEST_DIR/10-snapshot-v0.IMGFMT' } }, { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'virtio1', 'snapshot-file': 'TEST_DIR/10-snapshot-v1.IMGFMT' } } ] } }
-Formatting 'TEST_DIR/10-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/9-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
-Formatting 'TEST_DIR/10-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/9-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/10-snapshot-v0.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/9-snapshot-v0.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
+Formatting 'TEST_DIR/10-snapshot-v1.qcow2', fmt=qcow2 size=134217728 backing_file=TEST_DIR/9-snapshot-v1.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 {"return": {}}
 
 === Create a couple of snapshots using blockdev-snapshot ===
diff --git a/tests/qemu-iotests/144.out b/tests/qemu-iotests/144.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/144.out
+++ b/tests/qemu-iotests/144.out
@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=536870912
 { 'execute': 'qmp_capabilities' }
 {"return": {}}
 { 'execute': 'blockdev-snapshot-sync', 'arguments': { 'device': 'virtio0', 'snapshot-file':'TEST_DIR/tmp.IMGFMT', 'format': 'IMGFMT' } }
-Formatting 'TEST_DIR/tmp.qcow2', fmt=qcow2 size=536870912 backing_file=TEST_DIR/t.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/tmp.qcow2', fmt=qcow2 size=536870912 backing_file=TEST_DIR/t.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 {"return": {}}
 
 === Performing block-commit on active layer ===
@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/tmp.qcow2', fmt=qcow2 size=536870912 backing_file=TEST_DIR/
 === Performing Live Snapshot 2 ===
 
 { 'execute': 'blockdev-snapshot-sync', 'arguments': { 'device': 'virtio0', 'snapshot-file':'TEST_DIR/tmp2.IMGFMT', 'format': 'IMGFMT' } }
-Formatting 'TEST_DIR/tmp2.qcow2', fmt=qcow2 size=536870912 backing_file=TEST_DIR/t.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/tmp2.qcow2', fmt=qcow2 size=536870912 backing_file=TEST_DIR/t.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 {"return": {}}
 *** done
diff --git a/tests/qemu-iotests/182.out b/tests/qemu-iotests/182.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/182.out
+++ b/tests/qemu-iotests/182.out
@@ -XXX,XX +XXX,XX @@ Is another process using the image [TEST_DIR/t.qcow2]?
 {'execute': 'blockdev-add', 'arguments': { 'node-name': 'node0', 'driver': 'file', 'filename': 'TEST_DIR/t.IMGFMT', 'locking': 'on' } }
 {"return": {}}
 {'execute': 'blockdev-snapshot-sync', 'arguments': { 'node-name': 'node0', 'snapshot-file': 'TEST_DIR/t.IMGFMT.overlay', 'snapshot-node-name': 'node1' } }
-Formatting 'TEST_DIR/t.qcow2.overlay', fmt=qcow2 size=197120 backing_file=TEST_DIR/t.qcow2 backing_fmt=file cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2.overlay', fmt=qcow2 size=197120 backing_file=TEST_DIR/t.qcow2 backing_fmt=file cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 {"return": {}}
 {'execute': 'blockdev-add', 'arguments': { 'node-name': 'node1', 'driver': 'file', 'filename': 'TEST_DIR/t.IMGFMT', 'locking': 'on' } }
 {"return": {}}
diff --git a/tests/qemu-iotests/185.out b/tests/qemu-iotests/185.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/185.out
+++ b/tests/qemu-iotests/185.out
@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/t.IMGFMT.base', fmt=IMGFMT size=67108864
 === Creating backing chain ===
 
 { 'execute': 'blockdev-snapshot-sync', 'arguments': { 'device': 'disk', 'snapshot-file': 'TEST_DIR/t.IMGFMT.mid', 'format': 'IMGFMT', 'mode': 'absolute-paths' } }
-Formatting 'TEST_DIR/t.qcow2.mid', fmt=qcow2 size=67108864 backing_file=TEST_DIR/t.qcow2.base backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2.mid', fmt=qcow2 size=67108864 backing_file=TEST_DIR/t.qcow2.base backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 {"return": {}}
 { 'execute': 'human-monitor-command', 'arguments': { 'command-line': 'qemu-io disk "write 0 4M"' } }
 wrote 4194304/4194304 bytes at offset 0
 4 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 {"return": ""}
 { 'execute': 'blockdev-snapshot-sync', 'arguments': { 'device': 'disk', 'snapshot-file': 'TEST_DIR/t.IMGFMT', 'format': 'IMGFMT', 'mode': 'absolute-paths' } }
-Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 backing_file=TEST_DIR/t.qcow2.mid backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 backing_file=TEST_DIR/t.qcow2.mid backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 {"return": {}}
 
 === Start commit job and exit qemu ===
@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 backing_file=TEST_DIR/t.q
 { 'execute': 'qmp_capabilities' }
 {"return": {}}
 { 'execute': 'drive-mirror', 'arguments': { 'device': 'disk', 'target': 'TEST_DIR/t.IMGFMT.copy', 'format': 'IMGFMT', 'sync': 'full', 'speed': 65536 } }
-Formatting 'TEST_DIR/t.qcow2.copy', fmt=qcow2 size=67108864 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2.copy', fmt=qcow2 size=67108864 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "disk"}}
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "disk"}}
 {"return": {}}
@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/t.qcow2.copy', fmt=qcow2 size=67108864 cluster_size=65536 l
 { 'execute': 'qmp_capabilities' }
 {"return": {}}
 { 'execute': 'drive-backup', 'arguments': { 'device': 'disk', 'target': 'TEST_DIR/t.IMGFMT.copy', 'format': 'IMGFMT', 'sync': 'full', 'speed': 65536 } }
-Formatting 'TEST_DIR/t.qcow2.copy', fmt=qcow2 size=67108864 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/t.qcow2.copy', fmt=qcow2 size=67108864 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "disk"}}
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "disk"}}
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "paused", "id": "disk"}}
diff --git a/tests/qemu-iotests/198.out b/tests/qemu-iotests/198.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/198.out
+++ b/tests/qemu-iotests/198.out
@@ -XXX,XX +XXX,XX @@ image: json:{ /* filtered */ }
 file format: IMGFMT
 virtual size: 16 MiB (16777216 bytes)
 Format specific information:
+    compression type: zlib
     encrypt:
         ivgen alg: plain64
         hash alg: sha256
@@ -XXX,XX +XXX,XX @@ file format: IMGFMT
 virtual size: 16 MiB (16777216 bytes)
 backing file: TEST_DIR/t.IMGFMT.base
 Format specific information:
+    compression type: zlib
     encrypt:
         ivgen alg: plain64
         hash alg: sha256
diff --git a/tests/qemu-iotests/206.out b/tests/qemu-iotests/206.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/206.out
+++ b/tests/qemu-iotests/206.out
@@ -XXX,XX +XXX,XX @@ virtual size: 128 MiB (134217728 bytes)
 cluster_size: 65536
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: false
     refcount bits: 16
     corrupt: false
@@ -XXX,XX +XXX,XX @@ virtual size: 64 MiB (67108864 bytes)
 cluster_size: 65536
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: false
     refcount bits: 16
     corrupt: false
@@ -XXX,XX +XXX,XX @@ virtual size: 32 MiB (33554432 bytes)
 cluster_size: 2097152
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: true
     refcount bits: 1
     corrupt: false
@@ -XXX,XX +XXX,XX @@ backing file: TEST_IMG.base
 backing file format: IMGFMT
 Format specific information:
     compat: 0.10
+    compression type: zlib
     refcount bits: 16
 
 === Successful image creation (encrypted) ===
@@ -XXX,XX +XXX,XX @@ encrypted: yes
 cluster_size: 65536
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: false
     refcount bits: 16
     encrypt:
diff --git a/tests/qemu-iotests/242.out b/tests/qemu-iotests/242.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/242.out
+++ b/tests/qemu-iotests/242.out
@@ -XXX,XX +XXX,XX @@ virtual size: 1 MiB (1048576 bytes)
 cluster_size: 65536
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: false
     refcount bits: 16
     corrupt: false
@@ -XXX,XX +XXX,XX @@ virtual size: 1 MiB (1048576 bytes)
 cluster_size: 65536
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: false
     bitmaps:
         [0]:
@@ -XXX,XX +XXX,XX @@ virtual size: 1 MiB (1048576 bytes)
 cluster_size: 65536
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: false
     bitmaps:
         [0]:
@@ -XXX,XX +XXX,XX @@ virtual size: 1 MiB (1048576 bytes)
 cluster_size: 65536
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: false
     bitmaps:
         [0]:
@@ -XXX,XX +XXX,XX @@ virtual size: 1 MiB (1048576 bytes)
 cluster_size: 65536
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: false
     bitmaps:
         [0]:
diff --git a/tests/qemu-iotests/255.out b/tests/qemu-iotests/255.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/255.out
+++ b/tests/qemu-iotests/255.out
@@ -XXX,XX +XXX,XX @@ Finishing a commit job with background reads
 
 === Create backing chain and start VM ===
 
-Formatting 'TEST_DIR/PID-t.qcow2.mid', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-t.qcow2.mid', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
-Formatting 'TEST_DIR/PID-t.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-t.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 === Start background read requests ===
 
@@ -XXX,XX +XXX,XX @@ Closing the VM while a job is being cancelled
 
 === Create images and start VM ===
 
-Formatting 'TEST_DIR/PID-src.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-src.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
-Formatting 'TEST_DIR/PID-dst.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-dst.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 wrote 1048576/1048576 bytes at offset 0
 1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
diff --git a/tests/qemu-iotests/274.out b/tests/qemu-iotests/274.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/274.out
+++ b/tests/qemu-iotests/274.out
@@ -XXX,XX +XXX,XX @@
 == Commit tests ==
-Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=2097152 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=2097152 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
-Formatting 'TEST_DIR/PID-mid', fmt=qcow2 size=1048576 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-mid', fmt=qcow2 size=1048576 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
-Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=2097152 backing_file=TEST_DIR/PID-mid cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=2097152 backing_file=TEST_DIR/PID-mid cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 wrote 2097152/2097152 bytes at offset 0
 2 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ cluster_size: 65536
 backing file: TEST_DIR/PID-base
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: false
     refcount bits: 16
     corrupt: false
@@ -XXX,XX +XXX,XX @@ read 1048576/1048576 bytes at offset 1048576
 1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 
 === Testing HMP commit (top -> mid) ===
-Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=2097152 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=2097152 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
-Formatting 'TEST_DIR/PID-mid', fmt=qcow2 size=1048576 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-mid', fmt=qcow2 size=1048576 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
-Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=2097152 backing_file=TEST_DIR/PID-mid cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=2097152 backing_file=TEST_DIR/PID-mid cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 wrote 2097152/2097152 bytes at offset 0
 2 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ cluster_size: 65536
 backing file: TEST_DIR/PID-base
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: false
     refcount bits: 16
     corrupt: false
@@ -XXX,XX +XXX,XX @@ read 1048576/1048576 bytes at offset 1048576
 1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 
 === Testing QMP active commit (top -> mid) ===
-Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=2097152 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=2097152 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
-Formatting 'TEST_DIR/PID-mid', fmt=qcow2 size=1048576 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-mid', fmt=qcow2 size=1048576 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
-Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=2097152 backing_file=TEST_DIR/PID-mid cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=2097152 backing_file=TEST_DIR/PID-mid cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 wrote 2097152/2097152 bytes at offset 0
 2 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ cluster_size: 65536
 backing file: TEST_DIR/PID-base
 Format specific information:
     compat: 1.1
+    compression type: zlib
     lazy refcounts: false
     refcount bits: 16
     corrupt: false
@@ -XXX,XX +XXX,XX @@ read 1048576/1048576 bytes at offset 1048576
 
 == Resize tests ==
 === preallocation=off ===
-Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=6442450944 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=6442450944 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
-Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=1073741824 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=1073741824 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 wrote 65536/65536 bytes at offset 5368709120
 64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ read 65536/65536 bytes at offset 5368709120
 { "start": 1073741824, "length": 7516192768, "depth": 0, "zero": true, "data": false}]
 
 === preallocation=metadata ===
-Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=34359738368 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=34359738368 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
-Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=32212254720 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=32212254720 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 wrote 65536/65536 bytes at offset 33285996544
 64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ read 65536/65536 bytes at offset 33285996544
 { "start": 34896609280, "length": 536870912, "depth": 0, "zero": true, "data": false, "offset": 2685075456}]
 
 === preallocation=falloc ===
-Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=10485760 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=10485760 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
-Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=5242880 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=5242880 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 wrote 65536/65536 bytes at offset 9437184
 64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ read 65536/65536 bytes at offset 9437184
 { "start": 5242880, "length": 10485760, "depth": 0, "zero": false, "data": true, "offset": 327680}]
 
 === preallocation=full ===
-Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=16777216 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=16777216 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
-Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=8388608 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=8388608 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 wrote 65536/65536 bytes at offset 11534336
 64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ read 65536/65536 bytes at offset 11534336
 { "start": 8388608, "length": 4194304, "depth": 0, "zero": false, "data": true, "offset": 327680}]
 
 === preallocation=off ===
-Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=393216 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=393216 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
-Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=259072 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=259072 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 wrote 65536/65536 bytes at offset 259072
 64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ read 65536/65536 bytes at offset 259072
 { "start": 262144, "length": 262144, "depth": 0, "zero": true, "data": false}]
 
 === preallocation=off ===
-Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=409600 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=409600 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
-Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=262144 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=262144 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 wrote 65536/65536 bytes at offset 344064
 64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ read 65536/65536 bytes at offset 344064
 { "start": 262144, "length": 262144, "depth": 0, "zero": true, "data": false}]
 
 === preallocation=off ===
-Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=524288 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=524288 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
-Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=262144 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=262144 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 wrote 65536/65536 bytes at offset 446464
 64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
diff --git a/tests/qemu-iotests/280.out b/tests/qemu-iotests/280.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/280.out
+++ b/tests/qemu-iotests/280.out
@@ -XXX,XX +XXX,XX @@
-Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=67108864 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=67108864 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib
 
 === Launch VM ===
 Enabling migration QMP events on VM...
diff --git a/tests/qemu-iotests/common.filter b/tests/qemu-iotests/common.filter
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/common.filter
+++ b/tests/qemu-iotests/common.filter
@@ -XXX,XX +XXX,XX @@ _filter_img_create()
         -e "s# refcount_bits=[0-9]\\+##g" \
         -e "s# key-secret=[a-zA-Z0-9]\\+##g" \
         -e "s# iter-time=[0-9]\\+##g" \
-        -e "s# force_size=\$on\\|off\$##g"
+        -e "s# force_size=\$on\\|off\$##g" \
+        -e "s# compression_type=[a-zA-Z0-9]\\+##g"
 }
 
 _filter_img_info()
-- 
2.26.2

From: Denis Plotnikov <dplotnikov@virtuozzo.com>

The patch enables processing the image compression type defined
for the image and chooses an appropriate method for image clusters
(de)compression.

Signed-off-by: Denis Plotnikov <dplotnikov@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200507082521.29210-3-dplotnikov@virtuozzo.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block/qcow2-threads.c | 71 ++++++++++++++++++++++++++++++++++++-------
 1 file changed, 60 insertions(+), 11 deletions(-)

diff --git a/block/qcow2-threads.c b/block/qcow2-threads.c
index XXXXXXX..XXXXXXX 100644
--- a/block/qcow2-threads.c
+++ b/block/qcow2-threads.c
@@ -XXX,XX +XXX,XX @@ typedef struct Qcow2CompressData {
 } Qcow2CompressData;
 
 /*
- * qcow2_compress()
+ * qcow2_zlib_compress()
+ *
+ * Compress @src_size bytes of data using zlib compression method
  *
  * @dest - destination buffer, @dest_size bytes
  * @src - source buffer, @src_size bytes
@@ -XXX,XX +XXX,XX @@ typedef struct Qcow2CompressData {
  *          -ENOMEM destination buffer is not enough to store compressed data
  *          -EIO    on any other error
  */
-static ssize_t qcow2_compress(void *dest, size_t dest_size,
-                              const void *src, size_t src_size)
+static ssize_t qcow2_zlib_compress(void *dest, size_t dest_size,
+                                   const void *src, size_t src_size)
 {
     ssize_t ret;
     z_stream strm;
@@ -XXX,XX +XXX,XX @@ static ssize_t qcow2_compress(void *dest, size_t dest_size,
 }
 
 /*
- * qcow2_decompress()
+ * qcow2_zlib_decompress()
  *
  * Decompress some data (not more than @src_size bytes) to produce exactly
- * @dest_size bytes.
+ * @dest_size bytes using zlib compression method
  *
  * @dest - destination buffer, @dest_size bytes
  * @src - source buffer, @src_size bytes
@@ -XXX,XX +XXX,XX @@ static ssize_t qcow2_compress(void *dest, size_t dest_size,
  * Returns: 0 on success
  *          -EIO on fail
  */
-static ssize_t qcow2_decompress(void *dest, size_t dest_size,
-                                const void *src, size_t src_size)
+static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size,
+                                     const void *src, size_t src_size)
 {
     int ret;
     z_stream strm;
@@ -XXX,XX +XXX,XX @@ qcow2_co_do_compress(BlockDriverState *bs, void *dest, size_t dest_size,
     return arg.ret;
 }
 
+/*
+ * qcow2_co_compress()
+ *
+ * Compress @src_size bytes of data using the compression
+ * method defined by the image compression type
+ *
+ * @dest - destination buffer, @dest_size bytes
+ * @src - source buffer, @src_size bytes
+ *
+ * Returns: compressed size on success
+ *          a negative error code on failure
+ */
 ssize_t coroutine_fn
 qcow2_co_compress(BlockDriverState *bs, void *dest, size_t dest_size,
                   const void *src, size_t src_size)
 {
-    return qcow2_co_do_compress(bs, dest, dest_size, src, src_size,
-                                qcow2_compress);
+    BDRVQcow2State *s = bs->opaque;
+    Qcow2CompressFunc fn;
+
+    switch (s->compression_type) {
+    case QCOW2_COMPRESSION_TYPE_ZLIB:
+        fn = qcow2_zlib_compress;
+        break;
+
+    default:
+        abort();
+    }
+
+    return qcow2_co_do_compress(bs, dest, dest_size, src, src_size, fn);
 }
 
+/*
+ * qcow2_co_decompress()
+ *
+ * Decompress some data (not more than @src_size bytes) to produce exactly
+ * @dest_size bytes using the compression method defined by the image
+ * compression type
+ *
+ * @dest - destination buffer, @dest_size bytes
+ * @src - source buffer, @src_size bytes
+ *
+ * Returns: 0 on success
+ *          a negative error code on failure
+ */
 ssize_t coroutine_fn
 qcow2_co_decompress(BlockDriverState *bs, void *dest, size_t dest_size,
                     const void *src, size_t src_size)
 {
-    return qcow2_co_do_compress(bs, dest, dest_size, src, src_size,
-                                qcow2_decompress);
+    BDRVQcow2State *s = bs->opaque;
+    Qcow2CompressFunc fn;
+
+    switch (s->compression_type) {
+    case QCOW2_COMPRESSION_TYPE_ZLIB:
+        fn = qcow2_zlib_decompress;
+        break;
+
+    default:
+        abort();
+    }
+
+    return qcow2_co_do_compress(bs, dest, dest_size, src, src_size, fn);
 }
 
 
-- 
2.26.2

From: Denis Plotnikov <dplotnikov@virtuozzo.com>

zstd significantly reduces cluster compression time.
It provides better compression performance maintaining
the same level of the compression ratio in comparison with
zlib, which, at the moment, is the only compression
method available.

The performance test results:
Test compresses and decompresses qemu qcow2 image with just
installed rhel-7.6 guest.
Image cluster size: 64K. Image on disk size: 2.2G

The test was conducted with brd disk to reduce the influence
of disk subsystem to the test results.
The results is given in seconds.

compress cmd:
  time ./qemu-img convert -O qcow2 -c -o compression_type=[zlib|zstd]
                  src.img [zlib|zstd]_compressed.img
decompress cmd
  time ./qemu-img convert -O qcow2
                  [zlib|zstd]_compressed.img uncompressed.img

compression               decompression
         zlib       zstd           zlib         zstd
------------------------------------------------------------
real     65.5       16.3 (-75 %)    1.9          1.6 (-16 %)
user     65.0       15.8            5.3          2.5
sys       3.3        0.2            2.0          2.0

Both ZLIB and ZSTD gave the same compression ratio: 1.57
compressed image size in both cases: 1.4G

Signed-off-by: Denis Plotnikov <dplotnikov@virtuozzo.com>
QAPI part:
Acked-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20200507082521.29210-4-dplotnikov@virtuozzo.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 docs/interop/qcow2.txt |   1 +
 configure              |   2 +-
 qapi/block-core.json   |   3 +-
 block/qcow2-threads.c  | 169 +++++++++++++++++++++++++++++++++++++++++
 block/qcow2.c          |   7 ++
 5 files changed, 180 insertions(+), 2 deletions(-)

diff --git a/docs/interop/qcow2.txt b/docs/interop/qcow2.txt
index XXXXXXX..XXXXXXX 100644
--- a/docs/interop/qcow2.txt
+++ b/docs/interop/qcow2.txt
@@ -XXX,XX +XXX,XX @@ version 2.
 
                     Available compression type values:
                         0: zlib <https://www.zlib.net/>
+                        1: zstd <http://github.com/facebook/zstd>
 
 
 === Header padding ===
diff --git a/configure b/configure
index XXXXXXX..XXXXXXX 100755
--- a/configure
+++ b/configure
@@ -XXX,XX +XXX,XX @@ disabled with --disable-FEATURE, default is enabled if available:
   lzfse           support of lzfse compression library
                   (for reading lzfse-compressed dmg images)
   zstd            support for zstd compression library
-                  (for migration compression)
+                  (for migration compression and qcow2 cluster compression)
   seccomp         seccomp support
   coroutine-pool  coroutine freelist (better performance)
   glusterfs       GlusterFS backend
diff --git a/qapi/block-core.json b/qapi/block-core.json
index XXXXXXX..XXXXXXX 100644
--- a/qapi/block-core.json
+++ b/qapi/block-core.json
@@ -XXX,XX +XXX,XX @@
 # Compression type used in qcow2 image file
 #
 # @zlib: zlib compression, see <http://zlib.net/>
+# @zstd: zstd compression, see <http://github.com/facebook/zstd>
 #
 # Since: 5.1
 ##
 { 'enum': 'Qcow2CompressionType',
-  'data': [ 'zlib' ] }
+  'data': [ 'zlib', { 'name': 'zstd', 'if': 'defined(CONFIG_ZSTD)' } ] }
 
 ##
 # @BlockdevCreateOptionsQcow2:
diff --git a/block/qcow2-threads.c b/block/qcow2-threads.c
index XXXXXXX..XXXXXXX 100644
--- a/block/qcow2-threads.c
+++ b/block/qcow2-threads.c
@@ -XXX,XX +XXX,XX @@
 #define ZLIB_CONST
 #include <zlib.h>
 
+#ifdef CONFIG_ZSTD
+#include <zstd.h>
+#include <zstd_errors.h>
+#endif
+
 #include "qcow2.h"
 #include "block/thread-pool.h"
 #include "crypto.h"
@@ -XXX,XX +XXX,XX @@ static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size,
     return ret;
 }
 
+#ifdef CONFIG_ZSTD
+
+/*
+ * qcow2_zstd_compress()
+ *
+ * Compress @src_size bytes of data using zstd compression method
+ *
+ * @dest - destination buffer, @dest_size bytes
+ * @src - source buffer, @src_size bytes
+ *
+ * Returns: compressed size on success
+ *          -ENOMEM destination buffer is not enough to store compressed data
+ *          -EIO    on any other error
+ */
+static ssize_t qcow2_zstd_compress(void *dest, size_t dest_size,
+                                   const void *src, size_t src_size)
+{
+    ssize_t ret;
+    size_t zstd_ret;
+    ZSTD_outBuffer output = {
+        .dst = dest,
+        .size = dest_size,
+        .pos = 0
+    };
+    ZSTD_inBuffer input = {
+        .src = src,
+        .size = src_size,
+        .pos = 0
+    };
+    ZSTD_CCtx *cctx = ZSTD_createCCtx();
+
+    if (!cctx) {
+        return -EIO;
+    }
+    /*
+     * Use the zstd streamed interface for symmetry with decompression,
+     * where streaming is essential since we don't record the exact
+     * compressed size.
+     *
+     * ZSTD_compressStream2() tries to compress everything it could
+     * with a single call. Although, ZSTD docs says that:
+     * "You must continue calling ZSTD_compressStream2() with ZSTD_e_end
+     * until it returns 0, at which point you are free to start a new frame",
+     * in out tests we saw the only case when it returned with >0 -
+     * when the output buffer was too small. In that case,
+     * ZSTD_compressStream2() expects a bigger buffer on the next call.
+     * We can't provide a bigger buffer because we are limited with dest_size
+     * which we pass to the ZSTD_compressStream2() at once.
+     * So, we don't need any loops and just abort the compression when we
+     * don't get 0 result on the first call.
+     */
+    zstd_ret = ZSTD_compressStream2(cctx, &output, &input, ZSTD_e_end);
+
+    if (zstd_ret) {
+        if (zstd_ret > output.size - output.pos) {
+            ret = -ENOMEM;
+        } else {
+            ret = -EIO;
+        }
+        goto out;
+    }
+
+    /* make sure that zstd didn't overflow the dest buffer */
+    assert(output.pos <= dest_size);
+    ret = output.pos;
+out:
+    ZSTD_freeCCtx(cctx);
+    return ret;
+}
+
+/*
+ * qcow2_zstd_decompress()
+ *
+ * Decompress some data (not more than @src_size bytes) to produce exactly
+ * @dest_size bytes using zstd compression method
+ *
+ * @dest - destination buffer, @dest_size bytes
+ * @src - source buffer, @src_size bytes
+ *
+ * Returns: 0 on success
+ *          -EIO on any error
+ */
+static ssize_t qcow2_zstd_decompress(void *dest, size_t dest_size,
+                                     const void *src, size_t src_size)
+{
+    size_t zstd_ret = 0;
+    ssize_t ret = 0;
+    ZSTD_outBuffer output = {
+        .dst = dest,
+        .size = dest_size,
+        .pos = 0
+    };
+    ZSTD_inBuffer input = {
+        .src = src,
+        .size = src_size,
+        .pos = 0
+    };
+    ZSTD_DCtx *dctx = ZSTD_createDCtx();
+
+    if (!dctx) {
+        return -EIO;
+    }
+
+    /*
+     * The compressed stream from the input buffer may consist of more
+     * than one zstd frame. So we iterate until we get a fully
+     * uncompressed cluster.
+     * From zstd docs related to ZSTD_decompressStream:
+     * "return : 0 when a frame is completely decoded and fully flushed"
+     * We suppose that this means: each time ZSTD_decompressStream reads
+     * only ONE full frame and returns 0 if and only if that frame
+     * is completely decoded and flushed. Only after returning 0,
+     * ZSTD_decompressStream reads another ONE full frame.
+     */
+    while (output.pos < output.size) {
+        size_t last_in_pos = input.pos;
+        size_t last_out_pos = output.pos;
+        zstd_ret = ZSTD_decompressStream(dctx, &output, &input);
+
+        if (ZSTD_isError(zstd_ret)) {
+            ret = -EIO;
+            break;
+        }
+
+        /*
+         * The ZSTD manual is vague about what to do if it reads
+         * the buffer partially, and we don't want to get stuck
+         * in an infinite loop where ZSTD_decompressStream
+         * returns > 0 waiting for another input chunk. So, we add
+         * a check which ensures that the loop makes some progress
+         * on each step.
+         */
+        if (last_in_pos >= input.pos &&
+            last_out_pos >= output.pos) {
+            ret = -EIO;
+            break;
+        }
+    }
+    /*
+     * Make sure that we have the frame fully flushed here
+     * if not, we somehow managed to get uncompressed cluster
+     * greater then the cluster size, possibly because of its
+     * damage.
+     */
+    if (zstd_ret > 0) {
+        ret = -EIO;
+    }
+
+    ZSTD_freeDCtx(dctx);
+    assert(ret == 0 || ret == -EIO);
+    return ret;
+}
+#endif
+
 static int qcow2_compress_pool_func(void *opaque)
 {
     Qcow2CompressData *data = opaque;
@@ -XXX,XX +XXX,XX @@ qcow2_co_compress(BlockDriverState *bs, void *dest, size_t dest_size,
         fn = qcow2_zlib_compress;
         break;
 
+#ifdef CONFIG_ZSTD
+    case QCOW2_COMPRESSION_TYPE_ZSTD:
+        fn = qcow2_zstd_compress;
+        break;
+#endif
     default:
         abort();
     }
@@ -XXX,XX +XXX,XX @@ qcow2_co_decompress(BlockDriverState *bs, void *dest, size_t dest_size,
         fn = qcow2_zlib_decompress;
         break;
 
+#ifdef CONFIG_ZSTD
+    case QCOW2_COMPRESSION_TYPE_ZSTD:
+        fn = qcow2_zstd_decompress;
+        break;
+#endif
     default:
         abort();
     }
diff --git a/block/qcow2.c b/block/qcow2.c
index XXXXXXX..XXXXXXX 100644
--- a/block/qcow2.c
+++ b/block/qcow2.c
@@ -XXX,XX +XXX,XX @@ static int validate_compression_type(BDRVQcow2State *s, Error **errp)
 {
     switch (s->compression_type) {
     case QCOW2_COMPRESSION_TYPE_ZLIB:
+#ifdef CONFIG_ZSTD
+    case QCOW2_COMPRESSION_TYPE_ZSTD:
+#endif
         break;
 
     default:
@@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp)
         }
 
         switch (qcow2_opts->compression_type) {
+#ifdef CONFIG_ZSTD
+        case QCOW2_COMPRESSION_TYPE_ZSTD:
+            break;
+#endif
         default:
             error_setg(errp, "Unknown compression type");
             goto out;
-- 
2.26.2

From: Denis Plotnikov <dplotnikov@virtuozzo.com>

The test checks fulfilling qcow2 requirements for the compression
type feature and zstd compression type operability.

Signed-off-by: Denis Plotnikov <dplotnikov@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Tested-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200507082521.29210-5-dplotnikov@virtuozzo.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 tests/qemu-iotests/287     | 152 +++++++++++++++++++++++++++++++++++++
 tests/qemu-iotests/287.out |  67 ++++++++++++++++
 tests/qemu-iotests/group   |   1 +
 3 files changed, 220 insertions(+)
 create mode 100755 tests/qemu-iotests/287
 create mode 100644 tests/qemu-iotests/287.out

diff --git a/tests/qemu-iotests/287 b/tests/qemu-iotests/287
new file mode 100755
index XXXXXXX..XXXXXXX
--- /dev/null
+++ b/tests/qemu-iotests/287
@@ -XXX,XX +XXX,XX @@
+#!/usr/bin/env bash
+#
+# Test case for an image using zstd compression
+#
+# Copyright (c) 2020 Virtuozzo International GmbH
+#
+# 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/>.
+#
+
+# creator
+owner=dplotnikov@virtuozzo.com
+
+seq="$(basename $0)"
+echo "QA output created by $seq"
+
+status=1	# failure is the default!
+
+# standard environment
+. ./common.rc
+. ./common.filter
+
+# This tests qocw2-specific low-level functionality
+_supported_fmt qcow2
+_supported_proto file
+_supported_os Linux
+_unsupported_imgopts 'compat=0.10' data_file
+
+COMPR_IMG="$TEST_IMG.compressed"
+RAND_FILE="$TEST_DIR/rand_data"
+
+_cleanup()
+{
+    _cleanup_test_img
+    _rm_test_img "$COMPR_IMG"
+    rm -f "$RAND_FILE"
+}
+trap "_cleanup; exit \$status" 0 1 2 3 15
+
+# for all the cases
+CLUSTER_SIZE=65536
+
+# Check if we can run this test.
+if IMGOPTS='compression_type=zstd' _make_test_img 64M |
+    grep "Invalid parameter 'zstd'"; then
+    _notrun "ZSTD is disabled"
+fi
+
+echo
+echo "=== Testing compression type incompatible bit setting for zlib ==="
+echo
+_make_test_img -o compression_type=zlib 64M
+$PYTHON qcow2.py "$TEST_IMG" dump-header | grep incompatible_features
+
+echo
+echo "=== Testing compression type incompatible bit setting for zstd ==="
+echo
+_make_test_img -o compression_type=zstd 64M
+$PYTHON qcow2.py "$TEST_IMG" dump-header | grep incompatible_features
+
+echo
+echo "=== Testing zlib with incompatible bit set ==="
+echo
+_make_test_img -o compression_type=zlib 64M
+$PYTHON qcow2.py "$TEST_IMG" set-feature-bit incompatible 3
+# to make sure the bit was actually set
+$PYTHON qcow2.py "$TEST_IMG" dump-header | grep incompatible_features
+
+if $QEMU_IMG info "$TEST_IMG" >/dev/null 2>&1 ; then
+    echo "Error: The image opened successfully. The image must not be opened."
+fi
+
+echo
+echo "=== Testing zstd with incompatible bit unset ==="
+echo
+_make_test_img -o compression_type=zstd 64M
+$PYTHON qcow2.py "$TEST_IMG" set-header incompatible_features 0
+# to make sure the bit was actually unset
+$PYTHON qcow2.py "$TEST_IMG" dump-header | grep incompatible_features
+
+if $QEMU_IMG info "$TEST_IMG" >/dev/null 2>&1 ; then
+    echo "Error: The image opened successfully. The image must not be opened."
+fi
+
+echo
+echo "=== Testing compression type values ==="
+echo
+# zlib=0
+_make_test_img -o compression_type=zlib 64M
+peek_file_be "$TEST_IMG" 104 1
+echo
+
+# zstd=1
+_make_test_img -o compression_type=zstd 64M
+peek_file_be "$TEST_IMG" 104 1
+echo
+
+echo
+echo "=== Testing simple reading and writing with zstd ==="
+echo
+_make_test_img -o compression_type=zstd 64M
+$QEMU_IO -c "write -c -P 0xAC 64K 64K " "$TEST_IMG" | _filter_qemu_io
+$QEMU_IO -c "read -P 0xAC 64K 64K " "$TEST_IMG" | _filter_qemu_io
+# read on the cluster boundaries
+$QEMU_IO -c "read -v 131070 8 " "$TEST_IMG" | _filter_qemu_io
+$QEMU_IO -c "read -v 65534 8" "$TEST_IMG" | _filter_qemu_io
+
+echo
+echo "=== Testing adjacent clusters reading and writing with zstd ==="
+echo
+_make_test_img -o compression_type=zstd 64M
+$QEMU_IO -c "write -c -P 0xAB 0 64K " "$TEST_IMG" | _filter_qemu_io
+$QEMU_IO -c "write -c -P 0xAC 64K 64K " "$TEST_IMG" | _filter_qemu_io
+$QEMU_IO -c "write -c -P 0xAD 128K 64K " "$TEST_IMG" | _filter_qemu_io
+
+$QEMU_IO -c "read -P 0xAB 0 64k " "$TEST_IMG" | _filter_qemu_io
+$QEMU_IO -c "read -P 0xAC 64K 64k " "$TEST_IMG" | _filter_qemu_io
+$QEMU_IO -c "read -P 0xAD 128K 64k " "$TEST_IMG" | _filter_qemu_io
+
+echo
+echo "=== Testing incompressible cluster processing with zstd ==="
+echo
+# create a 2M image and fill it with 1M likely incompressible data
+# and 1M compressible data
+dd if=/dev/urandom of="$RAND_FILE" bs=1M count=1 seek=1
+QEMU_IO_OPTIONS="$QEMU_IO_OPTIONS_NO_FMT" \
+$QEMU_IO -f raw -c "write -P 0xFA 0 1M" "$RAND_FILE" | _filter_qemu_io
+
+$QEMU_IMG convert -f raw -O $IMGFMT -c \
+-o "$(_optstr_add "$IMGOPTS" "compression_type=zlib")" "$RAND_FILE" \
+"$TEST_IMG" | _filter_qemu_io
+
+$QEMU_IMG convert -O $IMGFMT -c \
+-o "$(_optstr_add "$IMGOPTS" "compression_type=zstd")" "$TEST_IMG" \
+"$COMPR_IMG" | _filter_qemu_io
+
+$QEMU_IMG compare "$TEST_IMG" "$COMPR_IMG"
+
+# success, all done
+echo "*** done"
+rm -f $seq.full
+status=0
diff --git a/tests/qemu-iotests/287.out b/tests/qemu-iotests/287.out
new file mode 100644
index XXXXXXX..XXXXXXX
--- /dev/null
+++ b/tests/qemu-iotests/287.out
@@ -XXX,XX +XXX,XX @@
+QA output created by 287
+
+=== Testing compression type incompatible bit setting for zlib ===
+
+Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
+incompatible_features     []
+
+=== Testing compression type incompatible bit setting for zstd ===
+
+Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
+incompatible_features     [3]
+
+=== Testing zlib with incompatible bit set  ===
+
+Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
+incompatible_features     [3]
+
+=== Testing zstd with incompatible bit unset  ===
+
+Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
+incompatible_features     []
+
+=== Testing compression type values  ===
+
+Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
+   0
+Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
+   1
+
+=== Testing simple reading and writing with zstd ===
+
+Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
+wrote 65536/65536 bytes at offset 65536
+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+read 65536/65536 bytes at offset 65536
+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+0001fffe:  ac ac 00 00 00 00 00 00  ........
+read 8/8 bytes at offset 131070
+8 bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+0000fffe:  00 00 ac ac ac ac ac ac  ........
+read 8/8 bytes at offset 65534
+8 bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+=== Testing adjacent clusters reading and writing with zstd ===
+
+Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
+wrote 65536/65536 bytes at offset 0
+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+wrote 65536/65536 bytes at offset 65536
+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+wrote 65536/65536 bytes at offset 131072
+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+read 65536/65536 bytes at offset 0
+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+read 65536/65536 bytes at offset 65536
+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+read 65536/65536 bytes at offset 131072
+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+=== Testing incompressible cluster processing with zstd ===
+
+1+0 records in
+1+0 records out
+wrote 1048576/1048576 bytes at offset 0
+1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+Images are identical.
+*** done
diff --git a/tests/qemu-iotests/group b/tests/qemu-iotests/group
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/group
+++ b/tests/qemu-iotests/group
@@ -XXX,XX +XXX,XX @@
 283 auto quick
 284 rw
 286 rw quick
+287 auto quick
 288 quick
 289 rw quick
 290 rw auto quick
-- 
2.26.2

From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

Obviously, we should g_free the task after trace point and offset
update.

Reported-by: Coverity (CID 1428756)
Fixes: 4ce5dd3e9b5ee0fac18625860eb3727399ee965e
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20200507183800.22626-1-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block/block-copy.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/block/block-copy.c b/block/block-copy.c
index XXXXXXX..XXXXXXX 100644
--- a/block/block-copy.c
+++ b/block/block-copy.c
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn block_copy_dirty_clusters(BlockCopyState *s,
         }
         if (s->skip_unallocated && !(ret & BDRV_BLOCK_ALLOCATED)) {
             block_copy_task_end(task, 0);
-            g_free(task);
             progress_set_remaining(s->progress,
                                    bdrv_get_dirty_count(s->copy_bitmap) +
                                    s->in_flight_bytes);
             trace_block_copy_skip_range(s, task->offset, task->bytes);
             offset = task_end(task);
             bytes = end - offset;
+            g_free(task);
             continue;
         }
         task->zeroes = ret & BDRV_BLOCK_ZERO;
-- 
2.26.2

The following changes since commit 56f9e46b841c7be478ca038d8d4085d776ab4b0d:

Merge remote-tracking branch 'remotes/armbru/tags/pull-qapi-2017-02-20' into staging (2017-02-20 17:42:47 +0000)

are available in the git repository at:

git://github.com/stefanha/qemu.git tags/block-pull-request

for you to fetch changes up to a7b91d35bab97a2d3e779d0c64c9b837b52a6cf7:

coroutine-lock: make CoRwlock thread-safe and fair (2017-02-21 11:39:40 +0000)

----------------------------------------------------------------
Pull request

v2:
 * Rebased to resolve scsi conflicts

----------------------------------------------------------------

Paolo Bonzini (24):
  block: move AioContext, QEMUTimer, main-loop to libqemuutil
  aio: introduce aio_co_schedule and aio_co_wake
  block-backend: allow blk_prw from coroutine context
  test-thread-pool: use generic AioContext infrastructure
  io: add methods to set I/O handlers on AioContext
  io: make qio_channel_yield aware of AioContexts
  nbd: convert to use qio_channel_yield
  coroutine-lock: reschedule coroutine on the AioContext it was running
    on
  blkdebug: reschedule coroutine on the AioContext it is running on
  qed: introduce qed_aio_start_io and qed_aio_next_io_cb
  aio: push aio_context_acquire/release down to dispatching
  block: explicitly acquire aiocontext in timers that need it
  block: explicitly acquire aiocontext in callbacks that need it
  block: explicitly acquire aiocontext in bottom halves that need it
  block: explicitly acquire aiocontext in aio callbacks that need it
  aio-posix: partially inline aio_dispatch into aio_poll
  async: remove unnecessary inc/dec pairs
  block: document fields protected by AioContext lock
  coroutine-lock: make CoMutex thread-safe
  coroutine-lock: add limited spinning to CoMutex
  test-aio-multithread: add performance comparison with thread-based
    mutexes
  coroutine-lock: place CoMutex before CoQueue in header
  coroutine-lock: add mutex argument to CoQueue APIs
  coroutine-lock: make CoRwlock thread-safe and fair

-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

AioContext is fairly self contained, the only dependency is QEMUTimer but
that in turn doesn't need anything else.  So move them out of block-obj-y
to avoid introducing a dependency from io/ to block-obj-y.

main-loop and its dependency iohandler also need to be moved, because
later in this series io/ will call iohandler_get_aio_context.

[Changed copyright "the QEMU team" to "other QEMU contributors" as
suggested by Daniel Berrange and agreed by Paolo.
--Stefan]

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 20170213135235.12274-2-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 Makefile.objs                       |  4 ---
 stubs/Makefile.objs                 |  1 +
 tests/Makefile.include              | 11 ++++----
 util/Makefile.objs                  |  6 +++-
 block/io.c                          | 29 -------------------
 stubs/linux-aio.c                   | 32 +++++++++++++++++++++
 stubs/set-fd-handler.c              | 11 --------
 aio-posix.c => util/aio-posix.c     |  2 +-
 aio-win32.c => util/aio-win32.c     |  0
 util/aiocb.c                        | 55 +++++++++++++++++++++++++++++++++++++
 async.c => util/async.c             |  3 +-
 iohandler.c => util/iohandler.c     |  0
 main-loop.c => util/main-loop.c     |  0
 qemu-timer.c => util/qemu-timer.c   |  0
 thread-pool.c => util/thread-pool.c |  2 +-
 trace-events                        | 11 --------
 util/trace-events                   | 11 ++++++++
 17 files changed, 114 insertions(+), 64 deletions(-)
 create mode 100644 stubs/linux-aio.c
 rename aio-posix.c => util/aio-posix.c (99%)
 rename aio-win32.c => util/aio-win32.c (100%)
 create mode 100644 util/aiocb.c
 rename async.c => util/async.c (99%)
 rename iohandler.c => util/iohandler.c (100%)
 rename main-loop.c => util/main-loop.c (100%)
 rename qemu-timer.c => util/qemu-timer.c (100%)
 rename thread-pool.c => util/thread-pool.c (99%)

diff --git a/Makefile.objs b/Makefile.objs
index XXXXXXX..XXXXXXX 100644
--- a/Makefile.objs
+++ b/Makefile.objs
@@ -XXX,XX +XXX,XX @@ chardev-obj-y = chardev/
 #######################################################################
 # block-obj-y is code used by both qemu system emulation and qemu-img
 
-block-obj-y = async.o thread-pool.o
 block-obj-y += nbd/
 block-obj-y += block.o blockjob.o
-block-obj-y += main-loop.o iohandler.o qemu-timer.o
-block-obj-$(CONFIG_POSIX) += aio-posix.o
-block-obj-$(CONFIG_WIN32) += aio-win32.o
 block-obj-y += block/
 block-obj-y += qemu-io-cmds.o
 block-obj-$(CONFIG_REPLICATION) += replication.o
diff --git a/stubs/Makefile.objs b/stubs/Makefile.objs
index XXXXXXX..XXXXXXX 100644
--- a/stubs/Makefile.objs
+++ b/stubs/Makefile.objs
@@ -XXX,XX +XXX,XX @@ stub-obj-y += get-vm-name.o
 stub-obj-y += iothread.o
 stub-obj-y += iothread-lock.o
 stub-obj-y += is-daemonized.o
+stub-obj-$(CONFIG_LINUX_AIO) += linux-aio.o
 stub-obj-y += machine-init-done.o
 stub-obj-y += migr-blocker.o
 stub-obj-y += monitor.o
diff --git a/tests/Makefile.include b/tests/Makefile.include
index XXXXXXX..XXXXXXX 100644
--- a/tests/Makefile.include
+++ b/tests/Makefile.include
@@ -XXX,XX +XXX,XX @@ check-unit-y += tests/test-visitor-serialization$(EXESUF)
 check-unit-y += tests/test-iov$(EXESUF)
 gcov-files-test-iov-y = util/iov.c
 check-unit-y += tests/test-aio$(EXESUF)
+gcov-files-test-aio-y = util/async.c util/qemu-timer.o
+gcov-files-test-aio-$(CONFIG_WIN32) += util/aio-win32.c
+gcov-files-test-aio-$(CONFIG_POSIX) += util/aio-posix.c
 check-unit-y += tests/test-throttle$(EXESUF)
 gcov-files-test-aio-$(CONFIG_WIN32) = aio-win32.c
 gcov-files-test-aio-$(CONFIG_POSIX) = aio-posix.c
@@ -XXX,XX +XXX,XX @@ tests/check-qjson$(EXESUF): tests/check-qjson.o $(test-util-obj-y)
 tests/check-qom-interface$(EXESUF): tests/check-qom-interface.o $(test-qom-obj-y)
 tests/check-qom-proplist$(EXESUF): tests/check-qom-proplist.o $(test-qom-obj-y)
 
-tests/test-char$(EXESUF): tests/test-char.o qemu-timer.o \
-	$(test-util-obj-y) $(qtest-obj-y) $(test-block-obj-y) $(chardev-obj-y)
+tests/test-char$(EXESUF): tests/test-char.o $(test-util-obj-y) $(qtest-obj-y) $(test-io-obj-y) $(chardev-obj-y)
 tests/test-coroutine$(EXESUF): tests/test-coroutine.o $(test-block-obj-y)
 tests/test-aio$(EXESUF): tests/test-aio.o $(test-block-obj-y)
 tests/test-throttle$(EXESUF): tests/test-throttle.o $(test-block-obj-y)
@@ -XXX,XX +XXX,XX @@ tests/test-vmstate$(EXESUF): tests/test-vmstate.o \
 	migration/vmstate.o migration/qemu-file.o \
         migration/qemu-file-channel.o migration/qjson.o \
 	$(test-io-obj-y)
-tests/test-timed-average$(EXESUF): tests/test-timed-average.o qemu-timer.o \
-	$(test-util-obj-y)
+tests/test-timed-average$(EXESUF): tests/test-timed-average.o $(test-util-obj-y)
 tests/test-base64$(EXESUF): tests/test-base64.o \
 	libqemuutil.a libqemustub.a
 tests/ptimer-test$(EXESUF): tests/ptimer-test.o tests/ptimer-test-stubs.o hw/core/ptimer.o libqemustub.a
@@ -XXX,XX +XXX,XX @@ tests/usb-hcd-ehci-test$(EXESUF): tests/usb-hcd-ehci-test.o $(libqos-usb-obj-y)
 tests/usb-hcd-xhci-test$(EXESUF): tests/usb-hcd-xhci-test.o $(libqos-usb-obj-y)
 tests/pc-cpu-test$(EXESUF): tests/pc-cpu-test.o
 tests/postcopy-test$(EXESUF): tests/postcopy-test.o
-tests/vhost-user-test$(EXESUF): tests/vhost-user-test.o qemu-timer.o \
+tests/vhost-user-test$(EXESUF): tests/vhost-user-test.o $(test-util-obj-y) \
 	$(qtest-obj-y) $(test-io-obj-y) $(libqos-virtio-obj-y) $(libqos-pc-obj-y) \
 	$(chardev-obj-y)
 tests/qemu-iotests/socket_scm_helper$(EXESUF): tests/qemu-iotests/socket_scm_helper.o
diff --git a/util/Makefile.objs b/util/Makefile.objs
index XXXXXXX..XXXXXXX 100644
--- a/util/Makefile.objs
+++ b/util/Makefile.objs
@@ -XXX,XX +XXX,XX @@
 util-obj-y = osdep.o cutils.o unicode.o qemu-timer-common.o
 util-obj-y += bufferiszero.o
 util-obj-y += lockcnt.o
+util-obj-y += aiocb.o async.o thread-pool.o qemu-timer.o
+util-obj-y += main-loop.o iohandler.o
+util-obj-$(CONFIG_POSIX) += aio-posix.o
 util-obj-$(CONFIG_POSIX) += compatfd.o
 util-obj-$(CONFIG_POSIX) += event_notifier-posix.o
 util-obj-$(CONFIG_POSIX) += mmap-alloc.o
 util-obj-$(CONFIG_POSIX) += oslib-posix.o
 util-obj-$(CONFIG_POSIX) += qemu-openpty.o
 util-obj-$(CONFIG_POSIX) += qemu-thread-posix.o
-util-obj-$(CONFIG_WIN32) += event_notifier-win32.o
 util-obj-$(CONFIG_POSIX) += memfd.o
+util-obj-$(CONFIG_WIN32) += aio-win32.o
+util-obj-$(CONFIG_WIN32) += event_notifier-win32.o
 util-obj-$(CONFIG_WIN32) += oslib-win32.o
 util-obj-$(CONFIG_WIN32) += qemu-thread-win32.o
 util-obj-y += envlist.o path.o module.o
diff --git a/block/io.c b/block/io.c
index XXXXXXX..XXXXXXX 100644
--- a/block/io.c
+++ b/block/io.c
@@ -XXX,XX +XXX,XX @@ BlockAIOCB *bdrv_aio_flush(BlockDriverState *bs,
     return &acb->common;
 }
 
-void *qemu_aio_get(const AIOCBInfo *aiocb_info, BlockDriverState *bs,
-                   BlockCompletionFunc *cb, void *opaque)
-{
-    BlockAIOCB *acb;
-
-    acb = g_malloc(aiocb_info->aiocb_size);
-    acb->aiocb_info = aiocb_info;
-    acb->bs = bs;
-    acb->cb = cb;
-    acb->opaque = opaque;
-    acb->refcnt = 1;
-    return acb;
-}
-
-void qemu_aio_ref(void *p)
-{
-    BlockAIOCB *acb = p;
-    acb->refcnt++;
-}
-
-void qemu_aio_unref(void *p)
-{
-    BlockAIOCB *acb = p;
-    assert(acb->refcnt > 0);
-    if (--acb->refcnt == 0) {
-        g_free(acb);
-    }
-}
-
 /**************************************************************/
 /* Coroutine block device emulation */
 
diff --git a/stubs/linux-aio.c b/stubs/linux-aio.c
new file mode 100644
index XXXXXXX..XXXXXXX
--- /dev/null
+++ b/stubs/linux-aio.c
@@ -XXX,XX +XXX,XX @@
+/*
+ * Linux native AIO support.
+ *
+ * Copyright (C) 2009 IBM, Corp.
+ * Copyright (C) 2009 Red Hat, Inc.
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 or later.
+ * See the COPYING file in the top-level directory.
+ */
+#include "qemu/osdep.h"
+#include "block/aio.h"
+#include "block/raw-aio.h"
+
+void laio_detach_aio_context(LinuxAioState *s, AioContext *old_context)
+{
+    abort();
+}
+
+void laio_attach_aio_context(LinuxAioState *s, AioContext *new_context)
+{
+    abort();
+}
+
+LinuxAioState *laio_init(void)
+{
+    abort();
+}
+
+void laio_cleanup(LinuxAioState *s)
+{
+    abort();
+}
diff --git a/stubs/set-fd-handler.c b/stubs/set-fd-handler.c
index XXXXXXX..XXXXXXX 100644
--- a/stubs/set-fd-handler.c
+++ b/stubs/set-fd-handler.c
@@ -XXX,XX +XXX,XX @@ void qemu_set_fd_handler(int fd,
 {
     abort();
 }
-
-void aio_set_fd_handler(AioContext *ctx,
-                        int fd,
-                        bool is_external,
-                        IOHandler *io_read,
-                        IOHandler *io_write,
-                        AioPollFn *io_poll,
-                        void *opaque)
-{
-    abort();
-}
diff --git a/aio-posix.c b/util/aio-posix.c
similarity index 99%
rename from aio-posix.c
rename to util/aio-posix.c
index XXXXXXX..XXXXXXX 100644
--- a/aio-posix.c
+++ b/util/aio-posix.c
@@ -XXX,XX +XXX,XX @@
 #include "qemu/rcu_queue.h"
 #include "qemu/sockets.h"
 #include "qemu/cutils.h"
-#include "trace-root.h"
+#include "trace.h"
 #ifdef CONFIG_EPOLL_CREATE1
 #include <sys/epoll.h>
 #endif
diff --git a/aio-win32.c b/util/aio-win32.c
similarity index 100%
rename from aio-win32.c
rename to util/aio-win32.c
diff --git a/util/aiocb.c b/util/aiocb.c
new file mode 100644
index XXXXXXX..XXXXXXX
--- /dev/null
+++ b/util/aiocb.c
@@ -XXX,XX +XXX,XX @@
+/*
+ * BlockAIOCB allocation
+ *
+ * Copyright (c) 2003-2017 Fabrice Bellard and other QEMU contributors
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#include "qemu/osdep.h"
+#include "block/aio.h"
+
+void *qemu_aio_get(const AIOCBInfo *aiocb_info, BlockDriverState *bs,
+                   BlockCompletionFunc *cb, void *opaque)
+{
+    BlockAIOCB *acb;
+
+    acb = g_malloc(aiocb_info->aiocb_size);
+    acb->aiocb_info = aiocb_info;
+    acb->bs = bs;
+    acb->cb = cb;
+    acb->opaque = opaque;
+    acb->refcnt = 1;
+    return acb;
+}
+
+void qemu_aio_ref(void *p)
+{
+    BlockAIOCB *acb = p;
+    acb->refcnt++;
+}
+
+void qemu_aio_unref(void *p)
+{
+    BlockAIOCB *acb = p;
+    assert(acb->refcnt > 0);
+    if (--acb->refcnt == 0) {
+        g_free(acb);
+    }
+}
diff --git a/async.c b/util/async.c
similarity index 99%
rename from async.c
rename to util/async.c
index XXXXXXX..XXXXXXX 100644
--- a/async.c
+++ b/util/async.c
@@ -XXX,XX +XXX,XX @@
 /*
- * QEMU System Emulator
+ * Data plane event loop
  *
  * Copyright (c) 2003-2008 Fabrice Bellard
+ * Copyright (c) 2009-2017 QEMU contributors
  *
  * Permission is hereby granted, free of charge, to any person obtaining a copy
  * of this software and associated documentation files (the "Software"), to deal
diff --git a/iohandler.c b/util/iohandler.c
similarity index 100%
rename from iohandler.c
rename to util/iohandler.c
diff --git a/main-loop.c b/util/main-loop.c
similarity index 100%
rename from main-loop.c
rename to util/main-loop.c
diff --git a/qemu-timer.c b/util/qemu-timer.c
similarity index 100%
rename from qemu-timer.c
rename to util/qemu-timer.c
diff --git a/thread-pool.c b/util/thread-pool.c
similarity index 99%
rename from thread-pool.c
rename to util/thread-pool.c
index XXXXXXX..XXXXXXX 100644
--- a/thread-pool.c
+++ b/util/thread-pool.c
@@ -XXX,XX +XXX,XX @@
 #include "qemu/queue.h"
 #include "qemu/thread.h"
 #include "qemu/coroutine.h"
-#include "trace-root.h"
+#include "trace.h"
 #include "block/thread-pool.h"
 #include "qemu/main-loop.h"
 
diff --git a/trace-events b/trace-events
index XXXXXXX..XXXXXXX 100644
--- a/trace-events
+++ b/trace-events
@@ -XXX,XX +XXX,XX @@
 #
 # The <format-string> should be a sprintf()-compatible format string.
 
-# aio-posix.c
-run_poll_handlers_begin(void *ctx, int64_t max_ns) "ctx %p max_ns %"PRId64
-run_poll_handlers_end(void *ctx, bool progress) "ctx %p progress %d"
-poll_shrink(void *ctx, int64_t old, int64_t new) "ctx %p old %"PRId64" new %"PRId64
-poll_grow(void *ctx, int64_t old, int64_t new) "ctx %p old %"PRId64" new %"PRId64
-
-# thread-pool.c
-thread_pool_submit(void *pool, void *req, void *opaque) "pool %p req %p opaque %p"
-thread_pool_complete(void *pool, void *req, void *opaque, int ret) "pool %p req %p opaque %p ret %d"
-thread_pool_cancel(void *req, void *opaque) "req %p opaque %p"
-
 # ioport.c
 cpu_in(unsigned int addr, char size, unsigned int val) "addr %#x(%c) value %u"
 cpu_out(unsigned int addr, char size, unsigned int val) "addr %#x(%c) value %u"
diff --git a/util/trace-events b/util/trace-events
index XXXXXXX..XXXXXXX 100644
--- a/util/trace-events
+++ b/util/trace-events
@@ -XXX,XX +XXX,XX @@
 # See docs/tracing.txt for syntax documentation.
 
+# util/aio-posix.c
+run_poll_handlers_begin(void *ctx, int64_t max_ns) "ctx %p max_ns %"PRId64
+run_poll_handlers_end(void *ctx, bool progress) "ctx %p progress %d"
+poll_shrink(void *ctx, int64_t old, int64_t new) "ctx %p old %"PRId64" new %"PRId64
+poll_grow(void *ctx, int64_t old, int64_t new) "ctx %p old %"PRId64" new %"PRId64
+
+# util/thread-pool.c
+thread_pool_submit(void *pool, void *req, void *opaque) "pool %p req %p opaque %p"
+thread_pool_complete(void *pool, void *req, void *opaque, int ret) "pool %p req %p opaque %p ret %d"
+thread_pool_cancel(void *req, void *opaque) "req %p opaque %p"
+
 # util/buffer.c
 buffer_resize(const char *buf, size_t olen, size_t len) "%s: old %zd, new %zd"
 buffer_move_empty(const char *buf, size_t len, const char *from) "%s: %zd bytes from %s"
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

aio_co_wake provides the infrastructure to start a coroutine on a "home"
AioContext.  It will be used by CoMutex and CoQueue, so that coroutines
don't jump from one context to another when they go to sleep on a
mutex or waitqueue.  However, it can also be used as a more efficient
alternative to one-shot bottom halves, and saves the effort of tracking
which AioContext a coroutine is running on.

aio_co_schedule is the part of aio_co_wake that starts a coroutine
on a remove AioContext, but it is also useful to implement e.g.
bdrv_set_aio_context callbacks.

The implementation of aio_co_schedule is based on a lock-free
multiple-producer, single-consumer queue.  The multiple producers use
cmpxchg to add to a LIFO stack.  The consumer (a per-AioContext bottom
half) grabs all items added so far, inverts the list to make it FIFO,
and goes through it one item at a time until it's empty.  The data
structure was inspired by OSv, which uses it in the very code we'll
"port" to QEMU for the thread-safe CoMutex.

Most of the new code is really tests.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 20170213135235.12274-3-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 tests/Makefile.include       |   8 +-
 include/block/aio.h          |  32 +++++++
 include/qemu/coroutine_int.h |  11 ++-
 tests/iothread.h             |  25 +++++
 tests/iothread.c             |  91 ++++++++++++++++++
 tests/test-aio-multithread.c | 213 +++++++++++++++++++++++++++++++++++++++++++
 util/async.c                 |  65 +++++++++++++
 util/qemu-coroutine.c        |   8 ++
 util/trace-events            |   4 +
 9 files changed, 453 insertions(+), 4 deletions(-)
 create mode 100644 tests/iothread.h
 create mode 100644 tests/iothread.c
 create mode 100644 tests/test-aio-multithread.c

diff --git a/tests/Makefile.include b/tests/Makefile.include
index XXXXXXX..XXXXXXX 100644
--- a/tests/Makefile.include
+++ b/tests/Makefile.include
@@ -XXX,XX +XXX,XX @@ check-unit-y += tests/test-aio$(EXESUF)
 gcov-files-test-aio-y = util/async.c util/qemu-timer.o
 gcov-files-test-aio-$(CONFIG_WIN32) += util/aio-win32.c
 gcov-files-test-aio-$(CONFIG_POSIX) += util/aio-posix.c
+check-unit-y += tests/test-aio-multithread$(EXESUF)
+gcov-files-test-aio-multithread-y = $(gcov-files-test-aio-y)
+gcov-files-test-aio-multithread-y += util/qemu-coroutine.c tests/iothread.c
 check-unit-y += tests/test-throttle$(EXESUF)
-gcov-files-test-aio-$(CONFIG_WIN32) = aio-win32.c
-gcov-files-test-aio-$(CONFIG_POSIX) = aio-posix.c
 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
@@ -XXX,XX +XXX,XX @@ test-qapi-obj-y = tests/test-qapi-visit.o tests/test-qapi-types.o \
 	$(test-qom-obj-y)
 test-crypto-obj-y = $(crypto-obj-y) $(test-qom-obj-y)
 test-io-obj-y = $(io-obj-y) $(test-crypto-obj-y)
-test-block-obj-y = $(block-obj-y) $(test-io-obj-y)
+test-block-obj-y = $(block-obj-y) $(test-io-obj-y) tests/iothread.o
 
 tests/check-qint$(EXESUF): tests/check-qint.o $(test-util-obj-y)
 tests/check-qstring$(EXESUF): tests/check-qstring.o $(test-util-obj-y)
@@ -XXX,XX +XXX,XX @@ tests/check-qom-proplist$(EXESUF): tests/check-qom-proplist.o $(test-qom-obj-y)
 tests/test-char$(EXESUF): tests/test-char.o $(test-util-obj-y) $(qtest-obj-y) $(test-io-obj-y) $(chardev-obj-y)
 tests/test-coroutine$(EXESUF): tests/test-coroutine.o $(test-block-obj-y)
 tests/test-aio$(EXESUF): tests/test-aio.o $(test-block-obj-y)
+tests/test-aio-multithread$(EXESUF): tests/test-aio-multithread.o $(test-block-obj-y)
 tests/test-throttle$(EXESUF): tests/test-throttle.o $(test-block-obj-y)
 tests/test-blockjob$(EXESUF): tests/test-blockjob.o $(test-block-obj-y) $(test-util-obj-y)
 tests/test-blockjob-txn$(EXESUF): tests/test-blockjob-txn.o $(test-block-obj-y) $(test-util-obj-y)
diff --git a/include/block/aio.h b/include/block/aio.h
index XXXXXXX..XXXXXXX 100644
--- a/include/block/aio.h
+++ b/include/block/aio.h
@@ -XXX,XX +XXX,XX @@ typedef void QEMUBHFunc(void *opaque);
 typedef bool AioPollFn(void *opaque);
 typedef void IOHandler(void *opaque);
 
+struct Coroutine;
 struct ThreadPool;
 struct LinuxAioState;
 
@@ -XXX,XX +XXX,XX @@ struct AioContext {
     bool notified;
     EventNotifier notifier;
 
+    QSLIST_HEAD(, Coroutine) scheduled_coroutines;
+    QEMUBH *co_schedule_bh;
+
     /* Thread pool for performing work and receiving completion callbacks.
      * Has its own locking.
      */
@@ -XXX,XX +XXX,XX @@ static inline bool aio_node_check(AioContext *ctx, bool is_external)
 }
 
 /**
+ * aio_co_schedule:
+ * @ctx: the aio context
+ * @co: the coroutine
+ *
+ * Start a coroutine on a remote AioContext.
+ *
+ * The coroutine must not be entered by anyone else while aio_co_schedule()
+ * is active.  In addition the coroutine must have yielded unless ctx
+ * is the context in which the coroutine is running (i.e. the value of
+ * qemu_get_current_aio_context() from the coroutine itself).
+ */
+void aio_co_schedule(AioContext *ctx, struct Coroutine *co);
+
+/**
+ * aio_co_wake:
+ * @co: the coroutine
+ *
+ * Restart a coroutine on the AioContext where it was running last, thus
+ * preventing coroutines from jumping from one context to another when they
+ * go to sleep.
+ *
+ * aio_co_wake may be executed either in coroutine or non-coroutine
+ * context.  The coroutine must not be entered by anyone else while
+ * aio_co_wake() is active.
+ */
+void aio_co_wake(struct Coroutine *co);
+
+/**
  * Return the AioContext whose event loop runs in the current thread.
  *
  * If called from an IOThread this will be the IOThread's AioContext.  If
diff --git a/include/qemu/coroutine_int.h b/include/qemu/coroutine_int.h
index XXXXXXX..XXXXXXX 100644
--- a/include/qemu/coroutine_int.h
+++ b/include/qemu/coroutine_int.h
@@ -XXX,XX +XXX,XX @@ struct Coroutine {
     CoroutineEntry *entry;
     void *entry_arg;
     Coroutine *caller;
+
+    /* Only used when the coroutine has terminated.  */
     QSLIST_ENTRY(Coroutine) pool_next;
+
     size_t locks_held;
 
-    /* Coroutines that should be woken up when we yield or terminate */
+    /* Coroutines that should be woken up when we yield or terminate.
+     * Only used when the coroutine is running.
+     */
     QSIMPLEQ_HEAD(, Coroutine) co_queue_wakeup;
+
+    /* Only used when the coroutine has yielded.  */
+    AioContext *ctx;
     QSIMPLEQ_ENTRY(Coroutine) co_queue_next;
+    QSLIST_ENTRY(Coroutine) co_scheduled_next;
 };
 
 Coroutine *qemu_coroutine_new(void);
diff --git a/tests/iothread.h b/tests/iothread.h
new file mode 100644
index XXXXXXX..XXXXXXX
--- /dev/null
+++ b/tests/iothread.h
@@ -XXX,XX +XXX,XX @@
+/*
+ * Event loop thread implementation for unit tests
+ *
+ * Copyright Red Hat Inc., 2013, 2016
+ *
+ * Authors:
+ *  Stefan Hajnoczi   <stefanha@redhat.com>
+ *  Paolo Bonzini     <pbonzini@redhat.com>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 or later.
+ * See the COPYING file in the top-level directory.
+ */
+#ifndef TEST_IOTHREAD_H
+#define TEST_IOTHREAD_H
+
+#include "block/aio.h"
+#include "qemu/thread.h"
+
+typedef struct IOThread IOThread;
+
+IOThread *iothread_new(void);
+void iothread_join(IOThread *iothread);
+AioContext *iothread_get_aio_context(IOThread *iothread);
+
+#endif
diff --git a/tests/iothread.c b/tests/iothread.c
new file mode 100644
index XXXXXXX..XXXXXXX
--- /dev/null
+++ b/tests/iothread.c
@@ -XXX,XX +XXX,XX @@
+/*
+ * Event loop thread implementation for unit tests
+ *
+ * Copyright Red Hat Inc., 2013, 2016
+ *
+ * Authors:
+ *  Stefan Hajnoczi   <stefanha@redhat.com>
+ *  Paolo Bonzini     <pbonzini@redhat.com>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 or later.
+ * See the COPYING file in the top-level directory.
+ *
+ */
+
+#include "qemu/osdep.h"
+#include "qapi/error.h"
+#include "block/aio.h"
+#include "qemu/main-loop.h"
+#include "qemu/rcu.h"
+#include "iothread.h"
+
+struct IOThread {
+    AioContext *ctx;
+
+    QemuThread thread;
+    QemuMutex init_done_lock;
+    QemuCond init_done_cond;    /* is thread initialization done? */
+    bool stopping;
+};
+
+static __thread IOThread *my_iothread;
+
+AioContext *qemu_get_current_aio_context(void)
+{
+    return my_iothread ? my_iothread->ctx : qemu_get_aio_context();
+}
+
+static void *iothread_run(void *opaque)
+{
+    IOThread *iothread = opaque;
+
+    rcu_register_thread();
+
+    my_iothread = iothread;
+    qemu_mutex_lock(&iothread->init_done_lock);
+    iothread->ctx = aio_context_new(&error_abort);
+    qemu_cond_signal(&iothread->init_done_cond);
+    qemu_mutex_unlock(&iothread->init_done_lock);
+
+    while (!atomic_read(&iothread->stopping)) {
+        aio_poll(iothread->ctx, true);
+    }
+
+    rcu_unregister_thread();
+    return NULL;
+}
+
+void iothread_join(IOThread *iothread)
+{
+    iothread->stopping = true;
+    aio_notify(iothread->ctx);
+    qemu_thread_join(&iothread->thread);
+    qemu_cond_destroy(&iothread->init_done_cond);
+    qemu_mutex_destroy(&iothread->init_done_lock);
+    aio_context_unref(iothread->ctx);
+    g_free(iothread);
+}
+
+IOThread *iothread_new(void)
+{
+    IOThread *iothread = g_new0(IOThread, 1);
+
+    qemu_mutex_init(&iothread->init_done_lock);
+    qemu_cond_init(&iothread->init_done_cond);
+    qemu_thread_create(&iothread->thread, NULL, iothread_run,
+                       iothread, QEMU_THREAD_JOINABLE);
+
+    /* Wait for initialization to complete */
+    qemu_mutex_lock(&iothread->init_done_lock);
+    while (iothread->ctx == NULL) {
+        qemu_cond_wait(&iothread->init_done_cond,
+                       &iothread->init_done_lock);
+    }
+    qemu_mutex_unlock(&iothread->init_done_lock);
+    return iothread;
+}
+
+AioContext *iothread_get_aio_context(IOThread *iothread)
+{
+    return iothread->ctx;
+}
diff --git a/tests/test-aio-multithread.c b/tests/test-aio-multithread.c
new file mode 100644
index XXXXXXX..XXXXXXX
--- /dev/null
+++ b/tests/test-aio-multithread.c
@@ -XXX,XX +XXX,XX @@
+/*
+ * AioContext multithreading tests
+ *
+ * Copyright Red Hat, Inc. 2016
+ *
+ * Authors:
+ *  Paolo Bonzini    <pbonzini@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 "qemu/osdep.h"
+#include <glib.h>
+#include "block/aio.h"
+#include "qapi/error.h"
+#include "qemu/coroutine.h"
+#include "qemu/thread.h"
+#include "qemu/error-report.h"
+#include "iothread.h"
+
+/* AioContext management */
+
+#define NUM_CONTEXTS 5
+
+static IOThread *threads[NUM_CONTEXTS];
+static AioContext *ctx[NUM_CONTEXTS];
+static __thread int id = -1;
+
+static QemuEvent done_event;
+
+/* Run a function synchronously on a remote iothread. */
+
+typedef struct CtxRunData {
+    QEMUBHFunc *cb;
+    void *arg;
+} CtxRunData;
+
+static void ctx_run_bh_cb(void *opaque)
+{
+    CtxRunData *data = opaque;
+
+    data->cb(data->arg);
+    qemu_event_set(&done_event);
+}
+
+static void ctx_run(int i, QEMUBHFunc *cb, void *opaque)
+{
+    CtxRunData data = {
+        .cb = cb,
+        .arg = opaque
+    };
+
+    qemu_event_reset(&done_event);
+    aio_bh_schedule_oneshot(ctx[i], ctx_run_bh_cb, &data);
+    qemu_event_wait(&done_event);
+}
+
+/* Starting the iothreads. */
+
+static void set_id_cb(void *opaque)
+{
+    int *i = opaque;
+
+    id = *i;
+}
+
+static void create_aio_contexts(void)
+{
+    int i;
+
+    for (i = 0; i < NUM_CONTEXTS; i++) {
+        threads[i] = iothread_new();
+        ctx[i] = iothread_get_aio_context(threads[i]);
+    }
+
+    qemu_event_init(&done_event, false);
+    for (i = 0; i < NUM_CONTEXTS; i++) {
+        ctx_run(i, set_id_cb, &i);
+    }
+}
+
+/* Stopping the iothreads. */
+
+static void join_aio_contexts(void)
+{
+    int i;
+
+    for (i = 0; i < NUM_CONTEXTS; i++) {
+        aio_context_ref(ctx[i]);
+    }
+    for (i = 0; i < NUM_CONTEXTS; i++) {
+        iothread_join(threads[i]);
+    }
+    for (i = 0; i < NUM_CONTEXTS; i++) {
+        aio_context_unref(ctx[i]);
+    }
+    qemu_event_destroy(&done_event);
+}
+
+/* Basic test for the stuff above. */
+
+static void test_lifecycle(void)
+{
+    create_aio_contexts();
+    join_aio_contexts();
+}
+
+/* aio_co_schedule test.  */
+
+static Coroutine *to_schedule[NUM_CONTEXTS];
+
+static bool now_stopping;
+
+static int count_retry;
+static int count_here;
+static int count_other;
+
+static bool schedule_next(int n)
+{
+    Coroutine *co;
+
+    co = atomic_xchg(&to_schedule[n], NULL);
+    if (!co) {
+        atomic_inc(&count_retry);
+        return false;
+    }
+
+    if (n == id) {
+        atomic_inc(&count_here);
+    } else {
+        atomic_inc(&count_other);
+    }
+
+    aio_co_schedule(ctx[n], co);
+    return true;
+}
+
+static void finish_cb(void *opaque)
+{
+    schedule_next(id);
+}
+
+static coroutine_fn void test_multi_co_schedule_entry(void *opaque)
+{
+    g_assert(to_schedule[id] == NULL);
+    atomic_mb_set(&to_schedule[id], qemu_coroutine_self());
+
+    while (!atomic_mb_read(&now_stopping)) {
+        int n;
+
+        n = g_test_rand_int_range(0, NUM_CONTEXTS);
+        schedule_next(n);
+        qemu_coroutine_yield();
+
+        g_assert(to_schedule[id] == NULL);
+        atomic_mb_set(&to_schedule[id], qemu_coroutine_self());
+    }
+}
+
+
+static void test_multi_co_schedule(int seconds)
+{
+    int i;
+
+    count_here = count_other = count_retry = 0;
+    now_stopping = false;
+
+    create_aio_contexts();
+    for (i = 0; i < NUM_CONTEXTS; i++) {
+        Coroutine *co1 = qemu_coroutine_create(test_multi_co_schedule_entry, NULL);
+        aio_co_schedule(ctx[i], co1);
+    }
+
+    g_usleep(seconds * 1000000);
+
+    atomic_mb_set(&now_stopping, true);
+    for (i = 0; i < NUM_CONTEXTS; i++) {
+        ctx_run(i, finish_cb, NULL);
+        to_schedule[i] = NULL;
+    }
+
+    join_aio_contexts();
+    g_test_message("scheduled %d, queued %d, retry %d, total %d\n",
+                  count_other, count_here, count_retry,
+                  count_here + count_other + count_retry);
+}
+
+static void test_multi_co_schedule_1(void)
+{
+    test_multi_co_schedule(1);
+}
+
+static void test_multi_co_schedule_10(void)
+{
+    test_multi_co_schedule(10);
+}
+
+/* End of tests.  */
+
+int main(int argc, char **argv)
+{
+    init_clocks();
+
+    g_test_init(&argc, &argv, NULL);
+    g_test_add_func("/aio/multi/lifecycle", test_lifecycle);
+    if (g_test_quick()) {
+        g_test_add_func("/aio/multi/schedule", test_multi_co_schedule_1);
+    } else {
+        g_test_add_func("/aio/multi/schedule", test_multi_co_schedule_10);
+    }
+    return g_test_run();
+}
diff --git a/util/async.c b/util/async.c
index XXXXXXX..XXXXXXX 100644
--- a/util/async.c
+++ b/util/async.c
@@ -XXX,XX +XXX,XX @@
 #include "qemu/main-loop.h"
 #include "qemu/atomic.h"
 #include "block/raw-aio.h"
+#include "qemu/coroutine_int.h"
+#include "trace.h"
 
 /***********************************************************/
 /* bottom halves (can be seen as timers which expire ASAP) */
@@ -XXX,XX +XXX,XX @@ aio_ctx_finalize(GSource     *source)
     }
 #endif
 
+    assert(QSLIST_EMPTY(&ctx->scheduled_coroutines));
+    qemu_bh_delete(ctx->co_schedule_bh);
+
     qemu_lockcnt_lock(&ctx->list_lock);
     assert(!qemu_lockcnt_count(&ctx->list_lock));
     while (ctx->first_bh) {
@@ -XXX,XX +XXX,XX @@ static bool event_notifier_poll(void *opaque)
     return atomic_read(&ctx->notified);
 }
 
+static void co_schedule_bh_cb(void *opaque)
+{
+    AioContext *ctx = opaque;
+    QSLIST_HEAD(, Coroutine) straight, reversed;
+
+    QSLIST_MOVE_ATOMIC(&reversed, &ctx->scheduled_coroutines);
+    QSLIST_INIT(&straight);
+
+    while (!QSLIST_EMPTY(&reversed)) {
+        Coroutine *co = QSLIST_FIRST(&reversed);
+        QSLIST_REMOVE_HEAD(&reversed, co_scheduled_next);
+        QSLIST_INSERT_HEAD(&straight, co, co_scheduled_next);
+    }
+
+    while (!QSLIST_EMPTY(&straight)) {
+        Coroutine *co = QSLIST_FIRST(&straight);
+        QSLIST_REMOVE_HEAD(&straight, co_scheduled_next);
+        trace_aio_co_schedule_bh_cb(ctx, co);
+        qemu_coroutine_enter(co);
+    }
+}
+
 AioContext *aio_context_new(Error **errp)
 {
     int ret;
@@ -XXX,XX +XXX,XX @@ AioContext *aio_context_new(Error **errp)
     }
     g_source_set_can_recurse(&ctx->source, true);
     qemu_lockcnt_init(&ctx->list_lock);
+
+    ctx->co_schedule_bh = aio_bh_new(ctx, co_schedule_bh_cb, ctx);
+    QSLIST_INIT(&ctx->scheduled_coroutines);
+
     aio_set_event_notifier(ctx, &ctx->notifier,
                            false,
                            (EventNotifierHandler *)
@@ -XXX,XX +XXX,XX @@ fail:
     return NULL;
 }
 
+void aio_co_schedule(AioContext *ctx, Coroutine *co)
+{
+    trace_aio_co_schedule(ctx, co);
+    QSLIST_INSERT_HEAD_ATOMIC(&ctx->scheduled_coroutines,
+                              co, co_scheduled_next);
+    qemu_bh_schedule(ctx->co_schedule_bh);
+}
+
+void aio_co_wake(struct Coroutine *co)
+{
+    AioContext *ctx;
+
+    /* Read coroutine before co->ctx.  Matches smp_wmb in
+     * qemu_coroutine_enter.
+     */
+    smp_read_barrier_depends();
+    ctx = atomic_read(&co->ctx);
+
+    if (ctx != qemu_get_current_aio_context()) {
+        aio_co_schedule(ctx, co);
+        return;
+    }
+
+    if (qemu_in_coroutine()) {
+        Coroutine *self = qemu_coroutine_self();
+        assert(self != co);
+        QSIMPLEQ_INSERT_TAIL(&self->co_queue_wakeup, co, co_queue_next);
+    } else {
+        aio_context_acquire(ctx);
+        qemu_coroutine_enter(co);
+        aio_context_release(ctx);
+    }
+}
+
 void aio_context_ref(AioContext *ctx)
 {
     g_source_ref(&ctx->source);
diff --git a/util/qemu-coroutine.c b/util/qemu-coroutine.c
index XXXXXXX..XXXXXXX 100644
--- a/util/qemu-coroutine.c
+++ b/util/qemu-coroutine.c
@@ -XXX,XX +XXX,XX @@
 #include "qemu/atomic.h"
 #include "qemu/coroutine.h"
 #include "qemu/coroutine_int.h"
+#include "block/aio.h"
 
 enum {
     POOL_BATCH_SIZE = 64,
@@ -XXX,XX +XXX,XX @@ void qemu_coroutine_enter(Coroutine *co)
     }
 
     co->caller = self;
+    co->ctx = qemu_get_current_aio_context();
+
+    /* Store co->ctx before anything that stores co.  Matches
+     * barrier in aio_co_wake.
+     */
+    smp_wmb();
+
     ret = qemu_coroutine_switch(self, co, COROUTINE_ENTER);
 
     qemu_co_queue_run_restart(co);
diff --git a/util/trace-events b/util/trace-events
index XXXXXXX..XXXXXXX 100644
--- a/util/trace-events
+++ b/util/trace-events
@@ -XXX,XX +XXX,XX @@ run_poll_handlers_end(void *ctx, bool progress) "ctx %p progress %d"
 poll_shrink(void *ctx, int64_t old, int64_t new) "ctx %p old %"PRId64" new %"PRId64
 poll_grow(void *ctx, int64_t old, int64_t new) "ctx %p old %"PRId64" new %"PRId64
 
+# util/async.c
+aio_co_schedule(void *ctx, void *co) "ctx %p co %p"
+aio_co_schedule_bh_cb(void *ctx, void *co) "ctx %p co %p"
+
 # util/thread-pool.c
 thread_pool_submit(void *pool, void *req, void *opaque) "pool %p req %p opaque %p"
 thread_pool_complete(void *pool, void *req, void *opaque, int ret) "pool %p req %p opaque %p ret %d"
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

qcow2_create2 calls this.  Do not run a nested event loop, as that
breaks when aio_co_wake tries to queue the coroutine on the co_queue_wakeup
list of the currently running one.

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 20170213135235.12274-4-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 block/block-backend.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/block/block-backend.c b/block/block-backend.c
index XXXXXXX..XXXXXXX 100644
--- a/block/block-backend.c
+++ b/block/block-backend.c
@@ -XXX,XX +XXX,XX @@ static int blk_prw(BlockBackend *blk, int64_t offset, uint8_t *buf,
 {
     QEMUIOVector qiov;
     struct iovec iov;
-    Coroutine *co;
     BlkRwCo rwco;
 
     iov = (struct iovec) {
@@ -XXX,XX +XXX,XX @@ static int blk_prw(BlockBackend *blk, int64_t offset, uint8_t *buf,
         .ret    = NOT_DONE,
     };
 
-    co = qemu_coroutine_create(co_entry, &rwco);
-    qemu_coroutine_enter(co);
-    BDRV_POLL_WHILE(blk_bs(blk), rwco.ret == NOT_DONE);
+    if (qemu_in_coroutine()) {
+        /* Fast-path if already in coroutine context */
+        co_entry(&rwco);
+    } else {
+        Coroutine *co = qemu_coroutine_create(co_entry, &rwco);
+        qemu_coroutine_enter(co);
+        BDRV_POLL_WHILE(blk_bs(blk), rwco.ret == NOT_DONE);
+    }
 
     return rwco.ret;
 }
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

Once the thread pool starts using aio_co_wake, it will also need
qemu_get_current_aio_context().  Make test-thread-pool create
an AioContext with qemu_init_main_loop, so that stubs/iothread.c
and tests/iothread.c can provide the rest.

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 20170213135235.12274-5-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 tests/test-thread-pool.c | 12 +++---------
 1 file changed, 3 insertions(+), 9 deletions(-)

diff --git a/tests/test-thread-pool.c b/tests/test-thread-pool.c
index XXXXXXX..XXXXXXX 100644
--- a/tests/test-thread-pool.c
+++ b/tests/test-thread-pool.c
@@ -XXX,XX +XXX,XX @@
 #include "qapi/error.h"
 #include "qemu/timer.h"
 #include "qemu/error-report.h"
+#include "qemu/main-loop.h"
 
 static AioContext *ctx;
 static ThreadPool *pool;
@@ -XXX,XX +XXX,XX @@ static void test_cancel_async(void)
 int main(int argc, char **argv)
 {
     int ret;
-    Error *local_error = NULL;
 
-    init_clocks();
-
-    ctx = aio_context_new(&local_error);
-    if (!ctx) {
-        error_reportf_err(local_error, "Failed to create AIO Context: ");
-        exit(1);
-    }
+    qemu_init_main_loop(&error_abort);
+    ctx = qemu_get_current_aio_context();
     pool = aio_get_thread_pool(ctx);
 
     g_test_init(&argc, &argv, NULL);
@@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv)
 
     ret = g_test_run();
 
-    aio_context_unref(ctx);
     return ret;
 }
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

This is in preparation for making qio_channel_yield work on
AioContexts other than the main one.

Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 20170213135235.12274-6-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 include/io/channel.h | 25 +++++++++++++++++++++++++
 io/channel-command.c | 13 +++++++++++++
 io/channel-file.c    | 11 +++++++++++
 io/channel-socket.c  | 16 +++++++++++-----
 io/channel-tls.c     | 12 ++++++++++++
 io/channel-watch.c   |  6 ++++++
 io/channel.c         | 11 +++++++++++
 7 files changed, 89 insertions(+), 5 deletions(-)

diff --git a/include/io/channel.h b/include/io/channel.h
index XXXXXXX..XXXXXXX 100644
--- a/include/io/channel.h
+++ b/include/io/channel.h
@@ -XXX,XX +XXX,XX @@
 
 #include "qemu-common.h"
 #include "qom/object.h"
+#include "block/aio.h"
 
 #define TYPE_QIO_CHANNEL "qio-channel"
 #define QIO_CHANNEL(obj)                                    \
@@ -XXX,XX +XXX,XX @@ struct QIOChannelClass {
                      off_t offset,
                      int whence,
                      Error **errp);
+    void (*io_set_aio_fd_handler)(QIOChannel *ioc,
+                                  AioContext *ctx,
+                                  IOHandler *io_read,
+                                  IOHandler *io_write,
+                                  void *opaque);
 };
 
 /* General I/O handling functions */
@@ -XXX,XX +XXX,XX @@ void qio_channel_yield(QIOChannel *ioc,
 void qio_channel_wait(QIOChannel *ioc,
                       GIOCondition condition);
 
+/**
+ * qio_channel_set_aio_fd_handler:
+ * @ioc: the channel object
+ * @ctx: the AioContext to set the handlers on
+ * @io_read: the read handler
+ * @io_write: the write handler
+ * @opaque: the opaque value passed to the handler
+ *
+ * This is used internally by qio_channel_yield().  It can
+ * be used by channel implementations to forward the handlers
+ * to another channel (e.g. from #QIOChannelTLS to the
+ * underlying socket).
+ */
+void qio_channel_set_aio_fd_handler(QIOChannel *ioc,
+                                    AioContext *ctx,
+                                    IOHandler *io_read,
+                                    IOHandler *io_write,
+                                    void *opaque);
+
 #endif /* QIO_CHANNEL_H */
diff --git a/io/channel-command.c b/io/channel-command.c
index XXXXXXX..XXXXXXX 100644
--- a/io/channel-command.c
+++ b/io/channel-command.c
@@ -XXX,XX +XXX,XX @@ static int qio_channel_command_close(QIOChannel *ioc,
 }
 
 
+static void qio_channel_command_set_aio_fd_handler(QIOChannel *ioc,
+                                                   AioContext *ctx,
+                                                   IOHandler *io_read,
+                                                   IOHandler *io_write,
+                                                   void *opaque)
+{
+    QIOChannelCommand *cioc = QIO_CHANNEL_COMMAND(ioc);
+    aio_set_fd_handler(ctx, cioc->readfd, false, io_read, NULL, NULL, opaque);
+    aio_set_fd_handler(ctx, cioc->writefd, false, NULL, io_write, NULL, opaque);
+}
+
+
 static GSource *qio_channel_command_create_watch(QIOChannel *ioc,
                                                  GIOCondition condition)
 {
@@ -XXX,XX +XXX,XX @@ static void qio_channel_command_class_init(ObjectClass *klass,
     ioc_klass->io_set_blocking = qio_channel_command_set_blocking;
     ioc_klass->io_close = qio_channel_command_close;
     ioc_klass->io_create_watch = qio_channel_command_create_watch;
+    ioc_klass->io_set_aio_fd_handler = qio_channel_command_set_aio_fd_handler;
 }
 
 static const TypeInfo qio_channel_command_info = {
diff --git a/io/channel-file.c b/io/channel-file.c
index XXXXXXX..XXXXXXX 100644
--- a/io/channel-file.c
+++ b/io/channel-file.c
@@ -XXX,XX +XXX,XX @@ static int qio_channel_file_close(QIOChannel *ioc,
 }
 
 
+static void qio_channel_file_set_aio_fd_handler(QIOChannel *ioc,
+                                                AioContext *ctx,
+                                                IOHandler *io_read,
+                                                IOHandler *io_write,
+                                                void *opaque)
+{
+    QIOChannelFile *fioc = QIO_CHANNEL_FILE(ioc);
+    aio_set_fd_handler(ctx, fioc->fd, false, io_read, io_write, NULL, opaque);
+}
+
 static GSource *qio_channel_file_create_watch(QIOChannel *ioc,
                                               GIOCondition condition)
 {
@@ -XXX,XX +XXX,XX @@ static void qio_channel_file_class_init(ObjectClass *klass,
     ioc_klass->io_seek = qio_channel_file_seek;
     ioc_klass->io_close = qio_channel_file_close;
     ioc_klass->io_create_watch = qio_channel_file_create_watch;
+    ioc_klass->io_set_aio_fd_handler = qio_channel_file_set_aio_fd_handler;
 }
 
 static const TypeInfo qio_channel_file_info = {
diff --git a/io/channel-socket.c b/io/channel-socket.c
index XXXXXXX..XXXXXXX 100644
--- a/io/channel-socket.c
+++ b/io/channel-socket.c
@@ -XXX,XX +XXX,XX @@ qio_channel_socket_set_blocking(QIOChannel *ioc,
         qemu_set_block(sioc->fd);
     } else {
         qemu_set_nonblock(sioc->fd);
-#ifdef WIN32
-        WSAEventSelect(sioc->fd, ioc->event,
-                       FD_READ | FD_ACCEPT | FD_CLOSE |
-                       FD_CONNECT | FD_WRITE | FD_OOB);
-#endif
     }
     return 0;
 }
@@ -XXX,XX +XXX,XX @@ qio_channel_socket_shutdown(QIOChannel *ioc,
     return 0;
 }
 
+static void qio_channel_socket_set_aio_fd_handler(QIOChannel *ioc,
+                                                  AioContext *ctx,
+                                                  IOHandler *io_read,
+                                                  IOHandler *io_write,
+                                                  void *opaque)
+{
+    QIOChannelSocket *sioc = QIO_CHANNEL_SOCKET(ioc);
+    aio_set_fd_handler(ctx, sioc->fd, false, io_read, io_write, NULL, opaque);
+}
+
 static GSource *qio_channel_socket_create_watch(QIOChannel *ioc,
                                                 GIOCondition condition)
 {
@@ -XXX,XX +XXX,XX @@ static void qio_channel_socket_class_init(ObjectClass *klass,
     ioc_klass->io_set_cork = qio_channel_socket_set_cork;
     ioc_klass->io_set_delay = qio_channel_socket_set_delay;
     ioc_klass->io_create_watch = qio_channel_socket_create_watch;
+    ioc_klass->io_set_aio_fd_handler = qio_channel_socket_set_aio_fd_handler;
 }
 
 static const TypeInfo qio_channel_socket_info = {
diff --git a/io/channel-tls.c b/io/channel-tls.c
index XXXXXXX..XXXXXXX 100644
--- a/io/channel-tls.c
+++ b/io/channel-tls.c
@@ -XXX,XX +XXX,XX @@ static int qio_channel_tls_close(QIOChannel *ioc,
     return qio_channel_close(tioc->master, errp);
 }
 
+static void qio_channel_tls_set_aio_fd_handler(QIOChannel *ioc,
+                                               AioContext *ctx,
+                                               IOHandler *io_read,
+                                               IOHandler *io_write,
+                                               void *opaque)
+{
+    QIOChannelTLS *tioc = QIO_CHANNEL_TLS(ioc);
+
+    qio_channel_set_aio_fd_handler(tioc->master, ctx, io_read, io_write, opaque);
+}
+
 static GSource *qio_channel_tls_create_watch(QIOChannel *ioc,
                                              GIOCondition condition)
 {
@@ -XXX,XX +XXX,XX @@ static void qio_channel_tls_class_init(ObjectClass *klass,
     ioc_klass->io_close = qio_channel_tls_close;
     ioc_klass->io_shutdown = qio_channel_tls_shutdown;
     ioc_klass->io_create_watch = qio_channel_tls_create_watch;
+    ioc_klass->io_set_aio_fd_handler = qio_channel_tls_set_aio_fd_handler;
 }
 
 static const TypeInfo qio_channel_tls_info = {
diff --git a/io/channel-watch.c b/io/channel-watch.c
index XXXXXXX..XXXXXXX 100644
--- a/io/channel-watch.c
+++ b/io/channel-watch.c
@@ -XXX,XX +XXX,XX @@ GSource *qio_channel_create_socket_watch(QIOChannel *ioc,
     GSource *source;
     QIOChannelSocketSource *ssource;
 
+#ifdef WIN32
+    WSAEventSelect(socket, ioc->event,
+                   FD_READ | FD_ACCEPT | FD_CLOSE |
+                   FD_CONNECT | FD_WRITE | FD_OOB);
+#endif
+
     source = g_source_new(&qio_channel_socket_source_funcs,
                           sizeof(QIOChannelSocketSource));
     ssource = (QIOChannelSocketSource *)source;
diff --git a/io/channel.c b/io/channel.c
index XXXXXXX..XXXXXXX 100644
--- a/io/channel.c
+++ b/io/channel.c
@@ -XXX,XX +XXX,XX @@ GSource *qio_channel_create_watch(QIOChannel *ioc,
 }
 
 
+void qio_channel_set_aio_fd_handler(QIOChannel *ioc,
+                                    AioContext *ctx,
+                                    IOHandler *io_read,
+                                    IOHandler *io_write,
+                                    void *opaque)
+{
+    QIOChannelClass *klass = QIO_CHANNEL_GET_CLASS(ioc);
+
+    klass->io_set_aio_fd_handler(ioc, ctx, io_read, io_write, opaque);
+}
+
 guint qio_channel_add_watch(QIOChannel *ioc,
                             GIOCondition condition,
                             QIOChannelFunc func,
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

Support separate coroutines for reading and writing, and place the
read/write handlers on the AioContext that the QIOChannel is registered
with.

Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 20170213135235.12274-7-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 include/io/channel.h | 47 ++++++++++++++++++++++++++--
 io/channel.c         | 86 +++++++++++++++++++++++++++++++++++++++-------------
 2 files changed, 109 insertions(+), 24 deletions(-)

diff --git a/include/io/channel.h b/include/io/channel.h
index XXXXXXX..XXXXXXX 100644
--- a/include/io/channel.h
+++ b/include/io/channel.h
@@ -XXX,XX +XXX,XX @@
 
 #include "qemu-common.h"
 #include "qom/object.h"
+#include "qemu/coroutine.h"
 #include "block/aio.h"
 
 #define TYPE_QIO_CHANNEL "qio-channel"
@@ -XXX,XX +XXX,XX @@ struct QIOChannel {
     Object parent;
     unsigned int features; /* bitmask of QIOChannelFeatures */
     char *name;
+    AioContext *ctx;
+    Coroutine *read_coroutine;
+    Coroutine *write_coroutine;
 #ifdef _WIN32
     HANDLE event; /* For use with GSource on Win32 */
 #endif
@@ -XXX,XX +XXX,XX @@ guint qio_channel_add_watch(QIOChannel *ioc,
 
 
 /**
+ * qio_channel_attach_aio_context:
+ * @ioc: the channel object
+ * @ctx: the #AioContext to set the handlers on
+ *
+ * Request that qio_channel_yield() sets I/O handlers on
+ * the given #AioContext.  If @ctx is %NULL, qio_channel_yield()
+ * uses QEMU's main thread event loop.
+ *
+ * You can move a #QIOChannel from one #AioContext to another even if
+ * I/O handlers are set for a coroutine.  However, #QIOChannel provides
+ * no synchronization between the calls to qio_channel_yield() and
+ * qio_channel_attach_aio_context().
+ *
+ * Therefore you should first call qio_channel_detach_aio_context()
+ * to ensure that the coroutine is not entered concurrently.  Then,
+ * while the coroutine has yielded, call qio_channel_attach_aio_context(),
+ * and then aio_co_schedule() to place the coroutine on the new
+ * #AioContext.  The calls to qio_channel_detach_aio_context()
+ * and qio_channel_attach_aio_context() should be protected with
+ * aio_context_acquire() and aio_context_release().
+ */
+void qio_channel_attach_aio_context(QIOChannel *ioc,
+                                    AioContext *ctx);
+
+/**
+ * qio_channel_detach_aio_context:
+ * @ioc: the channel object
+ *
+ * Disable any I/O handlers set by qio_channel_yield().  With the
+ * help of aio_co_schedule(), this allows moving a coroutine that was
+ * paused by qio_channel_yield() to another context.
+ */
+void qio_channel_detach_aio_context(QIOChannel *ioc);
+
+/**
  * qio_channel_yield:
  * @ioc: the channel object
  * @condition: the I/O condition to wait for
  *
- * Yields execution from the current coroutine until
- * the condition indicated by @condition becomes
- * available.
+ * Yields execution from the current coroutine until the condition
+ * indicated by @condition becomes available.  @condition must
+ * be either %G_IO_IN or %G_IO_OUT; it cannot contain both.  In
+ * addition, no two coroutine can be waiting on the same condition
+ * and channel at the same time.
  *
  * This must only be called from coroutine context
  */
diff --git a/io/channel.c b/io/channel.c
index XXXXXXX..XXXXXXX 100644
--- a/io/channel.c
+++ b/io/channel.c
@@ -XXX,XX +XXX,XX @@
 #include "qemu/osdep.h"
 #include "io/channel.h"
 #include "qapi/error.h"
-#include "qemu/coroutine.h"
+#include "qemu/main-loop.h"
 
 bool qio_channel_has_feature(QIOChannel *ioc,
                              QIOChannelFeature feature)
@@ -XXX,XX +XXX,XX @@ off_t qio_channel_io_seek(QIOChannel *ioc,
 }
 
 
-typedef struct QIOChannelYieldData QIOChannelYieldData;
-struct QIOChannelYieldData {
-    QIOChannel *ioc;
-    Coroutine *co;
-};
+static void qio_channel_set_aio_fd_handlers(QIOChannel *ioc);
 
+static void qio_channel_restart_read(void *opaque)
+{
+    QIOChannel *ioc = opaque;
+    Coroutine *co = ioc->read_coroutine;
+
+    ioc->read_coroutine = NULL;
+    qio_channel_set_aio_fd_handlers(ioc);
+    aio_co_wake(co);
+}
 
-static gboolean qio_channel_yield_enter(QIOChannel *ioc,
-                                        GIOCondition condition,
-                                        gpointer opaque)
+static void qio_channel_restart_write(void *opaque)
 {
-    QIOChannelYieldData *data = opaque;
-    qemu_coroutine_enter(data->co);
-    return FALSE;
+    QIOChannel *ioc = opaque;
+    Coroutine *co = ioc->write_coroutine;
+
+    ioc->write_coroutine = NULL;
+    qio_channel_set_aio_fd_handlers(ioc);
+    aio_co_wake(co);
 }
 
+static void qio_channel_set_aio_fd_handlers(QIOChannel *ioc)
+{
+    IOHandler *rd_handler = NULL, *wr_handler = NULL;
+    AioContext *ctx;
+
+    if (ioc->read_coroutine) {
+        rd_handler = qio_channel_restart_read;
+    }
+    if (ioc->write_coroutine) {
+        wr_handler = qio_channel_restart_write;
+    }
+
+    ctx = ioc->ctx ? ioc->ctx : iohandler_get_aio_context();
+    qio_channel_set_aio_fd_handler(ioc, ctx, rd_handler, wr_handler, ioc);
+}
+
+void qio_channel_attach_aio_context(QIOChannel *ioc,
+                                    AioContext *ctx)
+{
+    AioContext *old_ctx;
+    if (ioc->ctx == ctx) {
+        return;
+    }
+
+    old_ctx = ioc->ctx ? ioc->ctx : iohandler_get_aio_context();
+    qio_channel_set_aio_fd_handler(ioc, old_ctx, NULL, NULL, NULL);
+    ioc->ctx = ctx;
+    qio_channel_set_aio_fd_handlers(ioc);
+}
+
+void qio_channel_detach_aio_context(QIOChannel *ioc)
+{
+    ioc->read_coroutine = NULL;
+    ioc->write_coroutine = NULL;
+    qio_channel_set_aio_fd_handlers(ioc);
+    ioc->ctx = NULL;
+}
 
 void coroutine_fn qio_channel_yield(QIOChannel *ioc,
                                     GIOCondition condition)
 {
-    QIOChannelYieldData data;
-
     assert(qemu_in_coroutine());
-    data.ioc = ioc;
-    data.co = qemu_coroutine_self();
-    qio_channel_add_watch(ioc,
-                          condition,
-                          qio_channel_yield_enter,
-                          &data,
-                          NULL);
+    if (condition == G_IO_IN) {
+        assert(!ioc->read_coroutine);
+        ioc->read_coroutine = qemu_coroutine_self();
+    } else if (condition == G_IO_OUT) {
+        assert(!ioc->write_coroutine);
+        ioc->write_coroutine = qemu_coroutine_self();
+    } else {
+        abort();
+    }
+    qio_channel_set_aio_fd_handlers(ioc);
     qemu_coroutine_yield();
 }
 
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

In the client, read the reply headers from a coroutine, switching the
read side between the "read header" coroutine and the I/O coroutine that
reads the body of the reply.

In the server, if the server can read more requests it will create a new
"read request" coroutine as soon as a request has been read.  Otherwise,
the new coroutine is created in nbd_request_put.

diff --git a/block/nbd-client.h b/block/nbd-client.h
index XXXXXXX..XXXXXXX 100644
--- a/block/nbd-client.h
+++ b/block/nbd-client.h
@@ -XXX,XX +XXX,XX @@ typedef struct NBDClientSession {
 
     CoMutex send_mutex;
     CoQueue free_sema;
-    Coroutine *send_coroutine;
+    Coroutine *read_reply_co;
     int in_flight;
 
     Coroutine *recv_coroutine[MAX_NBD_REQUESTS];
diff --git a/block/nbd-client.c b/block/nbd-client.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nbd-client.c
+++ b/block/nbd-client.c
@@ -XXX,XX +XXX,XX @@
 #define HANDLE_TO_INDEX(bs, handle) ((handle) ^ ((uint64_t)(intptr_t)bs))
 #define INDEX_TO_HANDLE(bs, index)  ((index)  ^ ((uint64_t)(intptr_t)bs))
 
-static void nbd_recv_coroutines_enter_all(NBDClientSession *s)
+static void nbd_recv_coroutines_enter_all(BlockDriverState *bs)
 {
+    NBDClientSession *s = nbd_get_client_session(bs);
     int i;
 
     for (i = 0; i < MAX_NBD_REQUESTS; i++) {
@@ -XXX,XX +XXX,XX @@ static void nbd_recv_coroutines_enter_all(NBDClientSession *s)
             qemu_coroutine_enter(s->recv_coroutine[i]);
         }
     }
+    BDRV_POLL_WHILE(bs, s->read_reply_co);
 }
 
 static void nbd_teardown_connection(BlockDriverState *bs)
@@ -XXX,XX +XXX,XX @@ static void nbd_teardown_connection(BlockDriverState *bs)
     qio_channel_shutdown(client->ioc,
                          QIO_CHANNEL_SHUTDOWN_BOTH,
                          NULL);
-    nbd_recv_coroutines_enter_all(client);
+    nbd_recv_coroutines_enter_all(bs);
 
     nbd_client_detach_aio_context(bs);
     object_unref(OBJECT(client->sioc));
@@ -XXX,XX +XXX,XX @@ static void nbd_teardown_connection(BlockDriverState *bs)
     client->ioc = NULL;
 }
 
-static void nbd_reply_ready(void *opaque)
+static coroutine_fn void nbd_read_reply_entry(void *opaque)
 {
-    BlockDriverState *bs = opaque;
-    NBDClientSession *s = nbd_get_client_session(bs);
+    NBDClientSession *s = opaque;
     uint64_t i;
     int ret;
 
-    if (!s->ioc) { /* Already closed */
-        return;
-    }
-
-    if (s->reply.handle == 0) {
-        /* No reply already in flight.  Fetch a header.  It is possible
-         * that another thread has done the same thing in parallel, so
-         * the socket is not readable anymore.
-         */
+    for (;;) {
+        assert(s->reply.handle == 0);
         ret = nbd_receive_reply(s->ioc, &s->reply);
-        if (ret == -EAGAIN) {
-            return;
-        }
         if (ret < 0) {
-            s->reply.handle = 0;
-            goto fail;
+            break;
         }
-    }
 
-    /* There's no need for a mutex on the receive side, because the
-     * handler acts as a synchronization point and ensures that only
-     * one coroutine is called until the reply finishes.  */
-    i = HANDLE_TO_INDEX(s, s->reply.handle);
-    if (i >= MAX_NBD_REQUESTS) {
-        goto fail;
-    }
+        /* There's no need for a mutex on the receive side, because the
+         * handler acts as a synchronization point and ensures that only
+         * one coroutine is called until the reply finishes.
+         */
+        i = HANDLE_TO_INDEX(s, s->reply.handle);
+        if (i >= MAX_NBD_REQUESTS || !s->recv_coroutine[i]) {
+            break;
+        }
 
-    if (s->recv_coroutine[i]) {
-        qemu_coroutine_enter(s->recv_coroutine[i]);
-        return;
+        /* We're woken up by the recv_coroutine itself.  Note that there
+         * is no race between yielding and reentering read_reply_co.  This
+         * is because:
+         *
+         * - if recv_coroutine[i] runs on the same AioContext, it is only
+         *   entered after we yield
+         *
+         * - if recv_coroutine[i] runs on a different AioContext, reentering
+         *   read_reply_co happens through a bottom half, which can only
+         *   run after we yield.
+         */
+        aio_co_wake(s->recv_coroutine[i]);
+        qemu_coroutine_yield();
     }
-
-fail:
-    nbd_teardown_connection(bs);
-}
-
-static void nbd_restart_write(void *opaque)
-{
-    BlockDriverState *bs = opaque;
-
-    qemu_coroutine_enter(nbd_get_client_session(bs)->send_coroutine);
+    s->read_reply_co = NULL;
 }
 
 static int nbd_co_send_request(BlockDriverState *bs,
@@ -XXX,XX +XXX,XX @@ static int nbd_co_send_request(BlockDriverState *bs,
                                QEMUIOVector *qiov)
 {
     NBDClientSession *s = nbd_get_client_session(bs);
-    AioContext *aio_context;
     int rc, ret, i;
 
     qemu_co_mutex_lock(&s->send_mutex);
@@ -XXX,XX +XXX,XX @@ static int nbd_co_send_request(BlockDriverState *bs,
         return -EPIPE;
     }
 
-    s->send_coroutine = qemu_coroutine_self();
-    aio_context = bdrv_get_aio_context(bs);
-
-    aio_set_fd_handler(aio_context, s->sioc->fd, false,
-                       nbd_reply_ready, nbd_restart_write, NULL, bs);
     if (qiov) {
         qio_channel_set_cork(s->ioc, true);
         rc = nbd_send_request(s->ioc, request);
@@ -XXX,XX +XXX,XX @@ static int nbd_co_send_request(BlockDriverState *bs,
     } else {
         rc = nbd_send_request(s->ioc, request);
     }
-    aio_set_fd_handler(aio_context, s->sioc->fd, false,
-                       nbd_reply_ready, NULL, NULL, bs);
-    s->send_coroutine = NULL;
     qemu_co_mutex_unlock(&s->send_mutex);
     return rc;
 }
@@ -XXX,XX +XXX,XX @@ static void nbd_co_receive_reply(NBDClientSession *s,
 {
     int ret;
 
-    /* Wait until we're woken up by the read handler.  TODO: perhaps
-     * peek at the next reply and avoid yielding if it's ours?  */
+    /* Wait until we're woken up by nbd_read_reply_entry.  */
     qemu_coroutine_yield();
     *reply = s->reply;
     if (reply->handle != request->handle ||
@@ -XXX,XX +XXX,XX @@ static void nbd_coroutine_start(NBDClientSession *s,
     /* s->recv_coroutine[i] is set as soon as we get the send_lock.  */
 }
 
-static void nbd_coroutine_end(NBDClientSession *s,
+static void nbd_coroutine_end(BlockDriverState *bs,
                               NBDRequest *request)
 {
+    NBDClientSession *s = nbd_get_client_session(bs);
     int i = HANDLE_TO_INDEX(s, request->handle);
+
     s->recv_coroutine[i] = NULL;
-    if (s->in_flight-- == MAX_NBD_REQUESTS) {
-        qemu_co_queue_next(&s->free_sema);
+    s->in_flight--;
+    qemu_co_queue_next(&s->free_sema);
+
+    /* Kick the read_reply_co to get the next reply.  */
+    if (s->read_reply_co) {
+        aio_co_wake(s->read_reply_co);
     }
 }
 
@@ -XXX,XX +XXX,XX @@ int nbd_client_co_preadv(BlockDriverState *bs, uint64_t offset,
     } else {
         nbd_co_receive_reply(client, &request, &reply, qiov);
     }
-    nbd_coroutine_end(client, &request);
+    nbd_coroutine_end(bs, &request);
     return -reply.error;
 }
 
@@ -XXX,XX +XXX,XX @@ int nbd_client_co_pwritev(BlockDriverState *bs, uint64_t offset,
     } else {
         nbd_co_receive_reply(client, &request, &reply, NULL);
     }
-    nbd_coroutine_end(client, &request);
+    nbd_coroutine_end(bs, &request);
     return -reply.error;
 }
 
@@ -XXX,XX +XXX,XX @@ int nbd_client_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset,
     } else {
         nbd_co_receive_reply(client, &request, &reply, NULL);
     }
-    nbd_coroutine_end(client, &request);
+    nbd_coroutine_end(bs, &request);
     return -reply.error;
 }
 
@@ -XXX,XX +XXX,XX @@ int nbd_client_co_flush(BlockDriverState *bs)
     } else {
         nbd_co_receive_reply(client, &request, &reply, NULL);
     }
-    nbd_coroutine_end(client, &request);
+    nbd_coroutine_end(bs, &request);
     return -reply.error;
 }
 
@@ -XXX,XX +XXX,XX @@ int nbd_client_co_pdiscard(BlockDriverState *bs, int64_t offset, int count)
     } else {
         nbd_co_receive_reply(client, &request, &reply, NULL);
     }
-    nbd_coroutine_end(client, &request);
+    nbd_coroutine_end(bs, &request);
     return -reply.error;
 
 }
 
 void nbd_client_detach_aio_context(BlockDriverState *bs)
 {
-    aio_set_fd_handler(bdrv_get_aio_context(bs),
-                       nbd_get_client_session(bs)->sioc->fd,
-                       false, NULL, NULL, NULL, NULL);
+    NBDClientSession *client = nbd_get_client_session(bs);
+    qio_channel_detach_aio_context(QIO_CHANNEL(client->sioc));
 }
 
 void nbd_client_attach_aio_context(BlockDriverState *bs,
                                    AioContext *new_context)
 {
-    aio_set_fd_handler(new_context, nbd_get_client_session(bs)->sioc->fd,
-                       false, nbd_reply_ready, NULL, NULL, bs);
+    NBDClientSession *client = nbd_get_client_session(bs);
+    qio_channel_attach_aio_context(QIO_CHANNEL(client->sioc), new_context);
+    aio_co_schedule(new_context, client->read_reply_co);
 }
 
 void nbd_client_close(BlockDriverState *bs)
@@ -XXX,XX +XXX,XX @@ int nbd_client_init(BlockDriverState *bs,
     /* Now that we're connected, set the socket to be non-blocking and
      * kick the reply mechanism.  */
     qio_channel_set_blocking(QIO_CHANNEL(sioc), false, NULL);
-
+    client->read_reply_co = qemu_coroutine_create(nbd_read_reply_entry, client);
     nbd_client_attach_aio_context(bs, bdrv_get_aio_context(bs));
 
     logout("Established connection with NBD server\n");
diff --git a/nbd/client.c b/nbd/client.c
index XXXXXXX..XXXXXXX 100644
--- a/nbd/client.c
+++ b/nbd/client.c
@@ -XXX,XX +XXX,XX @@ ssize_t nbd_receive_reply(QIOChannel *ioc, NBDReply *reply)
     ssize_t ret;
 
     ret = read_sync(ioc, buf, sizeof(buf));
-    if (ret < 0) {
+    if (ret <= 0) {
         return ret;
     }
 
diff --git a/nbd/common.c b/nbd/common.c
index XXXXXXX..XXXXXXX 100644
--- a/nbd/common.c
+++ b/nbd/common.c
@@ -XXX,XX +XXX,XX @@ ssize_t nbd_wr_syncv(QIOChannel *ioc,
         }
         if (len == QIO_CHANNEL_ERR_BLOCK) {
             if (qemu_in_coroutine()) {
-                /* XXX figure out if we can create a variant on
-                 * qio_channel_yield() that works with AIO contexts
-                 * and consider using that in this branch */
-                qemu_coroutine_yield();
-            } else if (done) {
-                /* XXX this is needed by nbd_reply_ready.  */
-                qio_channel_wait(ioc,
-                                 do_read ? G_IO_IN : G_IO_OUT);
+                qio_channel_yield(ioc, do_read ? G_IO_IN : G_IO_OUT);
             } else {
                 return -EAGAIN;
             }
diff --git a/nbd/server.c b/nbd/server.c
index XXXXXXX..XXXXXXX 100644
--- a/nbd/server.c
+++ b/nbd/server.c
@@ -XXX,XX +XXX,XX @@ struct NBDClient {
     CoMutex send_lock;
     Coroutine *send_coroutine;
 
-    bool can_read;
-
     QTAILQ_ENTRY(NBDClient) next;
     int nb_requests;
     bool closing;
@@ -XXX,XX +XXX,XX @@ struct NBDClient {
 
 /* That's all folks */
 
-static void nbd_set_handlers(NBDClient *client);
-static void nbd_unset_handlers(NBDClient *client);
-static void nbd_update_can_read(NBDClient *client);
+static void nbd_client_receive_next_request(NBDClient *client);
 
 static gboolean nbd_negotiate_continue(QIOChannel *ioc,
                                        GIOCondition condition,
@@ -XXX,XX +XXX,XX @@ void nbd_client_put(NBDClient *client)
          */
         assert(client->closing);
 
-        nbd_unset_handlers(client);
+        qio_channel_detach_aio_context(client->ioc);
         object_unref(OBJECT(client->sioc));
         object_unref(OBJECT(client->ioc));
         if (client->tlscreds) {
@@ -XXX,XX +XXX,XX @@ static NBDRequestData *nbd_request_get(NBDClient *client)
 
     assert(client->nb_requests <= MAX_NBD_REQUESTS - 1);
     client->nb_requests++;
-    nbd_update_can_read(client);
 
     req = g_new0(NBDRequestData, 1);
     nbd_client_get(client);
@@ -XXX,XX +XXX,XX @@ static void nbd_request_put(NBDRequestData *req)
     g_free(req);
 
     client->nb_requests--;
-    nbd_update_can_read(client);
+    nbd_client_receive_next_request(client);
+
     nbd_client_put(client);
 }
 
@@ -XXX,XX +XXX,XX @@ static void blk_aio_attached(AioContext *ctx, void *opaque)
     exp->ctx = ctx;
 
     QTAILQ_FOREACH(client, &exp->clients, next) {
-        nbd_set_handlers(client);
+        qio_channel_attach_aio_context(client->ioc, ctx);
+        if (client->recv_coroutine) {
+            aio_co_schedule(ctx, client->recv_coroutine);
+        }
+        if (client->send_coroutine) {
+            aio_co_schedule(ctx, client->send_coroutine);
+        }
     }
 }
 
@@ -XXX,XX +XXX,XX @@ static void blk_aio_detach(void *opaque)
     TRACE("Export %s: Detaching clients from AIO context %p\n", exp->name, exp->ctx);
 
     QTAILQ_FOREACH(client, &exp->clients, next) {
-        nbd_unset_handlers(client);
+        qio_channel_detach_aio_context(client->ioc);
     }
 
     exp->ctx = NULL;
@@ -XXX,XX +XXX,XX @@ static ssize_t nbd_co_send_reply(NBDRequestData *req, NBDReply *reply,
     g_assert(qemu_in_coroutine());
     qemu_co_mutex_lock(&client->send_lock);
     client->send_coroutine = qemu_coroutine_self();
-    nbd_set_handlers(client);
 
     if (!len) {
         rc = nbd_send_reply(client->ioc, reply);
@@ -XXX,XX +XXX,XX @@ static ssize_t nbd_co_send_reply(NBDRequestData *req, NBDReply *reply,
     }
 
     client->send_coroutine = NULL;
-    nbd_set_handlers(client);
     qemu_co_mutex_unlock(&client->send_lock);
     return rc;
 }
@@ -XXX,XX +XXX,XX @@ static ssize_t nbd_co_receive_request(NBDRequestData *req,
     ssize_t rc;
 
     g_assert(qemu_in_coroutine());
-    client->recv_coroutine = qemu_coroutine_self();
-    nbd_update_can_read(client);
-
+    assert(client->recv_coroutine == qemu_coroutine_self());
     rc = nbd_receive_request(client->ioc, request);
     if (rc < 0) {
         if (rc != -EAGAIN) {
@@ -XXX,XX +XXX,XX @@ static ssize_t nbd_co_receive_request(NBDRequestData *req,
 
 out:
     client->recv_coroutine = NULL;
-    nbd_update_can_read(client);
+    nbd_client_receive_next_request(client);
 
     return rc;
 }
 
-static void nbd_trip(void *opaque)
+/* Owns a reference to the NBDClient passed as opaque.  */
+static coroutine_fn void nbd_trip(void *opaque)
 {
     NBDClient *client = opaque;
     NBDExport *exp = client->exp;
     NBDRequestData *req;
-    NBDRequest request;
+    NBDRequest request = { 0 };    /* GCC thinks it can be used uninitialized */
     NBDReply reply;
     ssize_t ret;
     int flags;
 
     TRACE("Reading request.");
     if (client->closing) {
+        nbd_client_put(client);
         return;
     }
 
@@ -XXX,XX +XXX,XX @@ static void nbd_trip(void *opaque)
 
 done:
     nbd_request_put(req);
+    nbd_client_put(client);
     return;
 
 out:
     nbd_request_put(req);
     client_close(client);
+    nbd_client_put(client);
 }
 
-static void nbd_read(void *opaque)
+static void nbd_client_receive_next_request(NBDClient *client)
 {
-    NBDClient *client = opaque;
-
-    if (client->recv_coroutine) {
-        qemu_coroutine_enter(client->recv_coroutine);
-    } else {
-        qemu_coroutine_enter(qemu_coroutine_create(nbd_trip, client));
-    }
-}
-
-static void nbd_restart_write(void *opaque)
-{
-    NBDClient *client = opaque;
-
-    qemu_coroutine_enter(client->send_coroutine);
-}
-
-static void nbd_set_handlers(NBDClient *client)
-{
-    if (client->exp && client->exp->ctx) {
-        aio_set_fd_handler(client->exp->ctx, client->sioc->fd, true,
-                           client->can_read ? nbd_read : NULL,
-                           client->send_coroutine ? nbd_restart_write : NULL,
-                           NULL, client);
-    }
-}
-
-static void nbd_unset_handlers(NBDClient *client)
-{
-    if (client->exp && client->exp->ctx) {
-        aio_set_fd_handler(client->exp->ctx, client->sioc->fd, true, NULL,
-                           NULL, NULL, NULL);
-    }
-}
-
-static void nbd_update_can_read(NBDClient *client)
-{
-    bool can_read = client->recv_coroutine ||
-                    client->nb_requests < MAX_NBD_REQUESTS;
-
-    if (can_read != client->can_read) {
-        client->can_read = can_read;
-        nbd_set_handlers(client);
-
-        /* There is no need to invoke aio_notify(), since aio_set_fd_handler()
-         * in nbd_set_handlers() will have taken care of that */
+    if (!client->recv_coroutine && client->nb_requests < MAX_NBD_REQUESTS) {
+        nbd_client_get(client);
+        client->recv_coroutine = qemu_coroutine_create(nbd_trip, client);
+        aio_co_schedule(client->exp->ctx, client->recv_coroutine);
     }
 }
 
@@ -XXX,XX +XXX,XX @@ static coroutine_fn void nbd_co_client_start(void *opaque)
         goto out;
     }
     qemu_co_mutex_init(&client->send_lock);
-    nbd_set_handlers(client);
 
     if (exp) {
         QTAILQ_INSERT_TAIL(&exp->clients, client, next);
     }
+
+    nbd_client_receive_next_request(client);
+
 out:
     g_free(data);
 }
@@ -XXX,XX +XXX,XX @@ void nbd_client_new(NBDExport *exp,
     object_ref(OBJECT(client->sioc));
     client->ioc = QIO_CHANNEL(sioc);
     object_ref(OBJECT(client->ioc));
-    client->can_read = true;
     client->close = close_fn;
 
     data->client = client;
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

As a small step towards the introduction of multiqueue, we want
coroutines to remain on the same AioContext that started them,
unless they are moved explicitly with e.g. aio_co_schedule.  This patch
avoids that coroutines switch AioContext when they use a CoMutex.
For now it does not make much of a difference, because the CoMutex
is not thread-safe and the AioContext itself is used to protect the
CoMutex from concurrent access.  However, this is going to change.

diff --git a/util/qemu-coroutine-lock.c b/util/qemu-coroutine-lock.c
index XXXXXXX..XXXXXXX 100644
--- a/util/qemu-coroutine-lock.c
+++ b/util/qemu-coroutine-lock.c
@@ -XXX,XX +XXX,XX @@
 #include "qemu/coroutine.h"
 #include "qemu/coroutine_int.h"
 #include "qemu/queue.h"
+#include "block/aio.h"
 #include "trace.h"
 
 void qemu_co_queue_init(CoQueue *queue)
@@ -XXX,XX +XXX,XX @@ void qemu_co_queue_run_restart(Coroutine *co)
 
 static bool qemu_co_queue_do_restart(CoQueue *queue, bool single)
 {
-    Coroutine *self = qemu_coroutine_self();
     Coroutine *next;
 
     if (QSIMPLEQ_EMPTY(&queue->entries)) {
@@ -XXX,XX +XXX,XX @@ static bool qemu_co_queue_do_restart(CoQueue *queue, bool single)
 
     while ((next = QSIMPLEQ_FIRST(&queue->entries)) != NULL) {
         QSIMPLEQ_REMOVE_HEAD(&queue->entries, co_queue_next);
-        QSIMPLEQ_INSERT_TAIL(&self->co_queue_wakeup, next, co_queue_next);
-        trace_qemu_co_queue_next(next);
+        aio_co_wake(next);
         if (single) {
             break;
         }
diff --git a/util/trace-events b/util/trace-events
index XXXXXXX..XXXXXXX 100644
--- a/util/trace-events
+++ b/util/trace-events
@@ -XXX,XX +XXX,XX @@ qemu_coroutine_terminate(void *co) "self %p"
 
 # util/qemu-coroutine-lock.c
 qemu_co_queue_run_restart(void *co) "co %p"
-qemu_co_queue_next(void *nxt) "next %p"
 qemu_co_mutex_lock_entry(void *mutex, void *self) "mutex %p self %p"
 qemu_co_mutex_lock_return(void *mutex, void *self) "mutex %p self %p"
 qemu_co_mutex_unlock_entry(void *mutex, void *self) "mutex %p self %p"
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

Keep the coroutine on the same AioContext.  Without this change,
there would be a race between yielding the coroutine and reentering it.
While the race cannot happen now, because the code only runs from a single
AioContext, this will change with multiqueue support in the block layer.

While doing the change, replace custom bottom half with aio_co_schedule.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170213135235.12274-10-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 block/blkdebug.c | 9 +--------
 1 file changed, 1 insertion(+), 8 deletions(-)

diff --git a/block/blkdebug.c b/block/blkdebug.c
index XXXXXXX..XXXXXXX 100644
--- a/block/blkdebug.c
+++ b/block/blkdebug.c
@@ -XXX,XX +XXX,XX @@ out:
     return ret;
 }
 
-static void error_callback_bh(void *opaque)
-{
-    Coroutine *co = opaque;
-    qemu_coroutine_enter(co);
-}
-
 static int inject_error(BlockDriverState *bs, BlkdebugRule *rule)
 {
     BDRVBlkdebugState *s = bs->opaque;
@@ -XXX,XX +XXX,XX @@ static int inject_error(BlockDriverState *bs, BlkdebugRule *rule)
     }
 
     if (!immediately) {
-        aio_bh_schedule_oneshot(bdrv_get_aio_context(bs), error_callback_bh,
-                                qemu_coroutine_self());
+        aio_co_schedule(qemu_get_current_aio_context(), qemu_coroutine_self());
         qemu_coroutine_yield();
     }
 
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

qed_aio_start_io and qed_aio_next_io will not have to acquire/release
the AioContext, while qed_aio_next_io_cb will.  Split the functionality
and gain a little type-safety in the process.

diff --git a/block/qed.c b/block/qed.c
index XXXXXXX..XXXXXXX 100644
--- a/block/qed.c
+++ b/block/qed.c
@@ -XXX,XX +XXX,XX @@ static CachedL2Table *qed_new_l2_table(BDRVQEDState *s)
     return l2_table;
 }
 
-static void qed_aio_next_io(void *opaque, int ret);
+static void qed_aio_next_io(QEDAIOCB *acb, int ret);
+
+static void qed_aio_start_io(QEDAIOCB *acb)
+{
+    qed_aio_next_io(acb, 0);
+}
+
+static void qed_aio_next_io_cb(void *opaque, int ret)
+{
+    QEDAIOCB *acb = opaque;
+
+    qed_aio_next_io(acb, ret);
+}
 
 static void qed_plug_allocating_write_reqs(BDRVQEDState *s)
 {
@@ -XXX,XX +XXX,XX @@ static void qed_unplug_allocating_write_reqs(BDRVQEDState *s)
 
     acb = QSIMPLEQ_FIRST(&s->allocating_write_reqs);
     if (acb) {
-        qed_aio_next_io(acb, 0);
+        qed_aio_start_io(acb);
     }
 }
 
@@ -XXX,XX +XXX,XX @@ static void qed_aio_complete(QEDAIOCB *acb, int ret)
         QSIMPLEQ_REMOVE_HEAD(&s->allocating_write_reqs, next);
         acb = QSIMPLEQ_FIRST(&s->allocating_write_reqs);
         if (acb) {
-            qed_aio_next_io(acb, 0);
+            qed_aio_start_io(acb);
         } else if (s->header.features & QED_F_NEED_CHECK) {
             qed_start_need_check_timer(s);
         }
@@ -XXX,XX +XXX,XX @@ static void qed_commit_l2_update(void *opaque, int ret)
     acb->request.l2_table = qed_find_l2_cache_entry(&s->l2_cache, l2_offset);
     assert(acb->request.l2_table != NULL);
 
-    qed_aio_next_io(opaque, ret);
+    qed_aio_next_io(acb, ret);
 }
 
 /**
@@ -XXX,XX +XXX,XX @@ static void qed_aio_write_l2_update(QEDAIOCB *acb, int ret, uint64_t offset)
     if (need_alloc) {
         /* Write out the whole new L2 table */
         qed_write_l2_table(s, &acb->request, 0, s->table_nelems, true,
-                            qed_aio_write_l1_update, acb);
+                           qed_aio_write_l1_update, acb);
     } else {
         /* Write out only the updated part of the L2 table */
         qed_write_l2_table(s, &acb->request, index, acb->cur_nclusters, false,
-                            qed_aio_next_io, acb);
+                           qed_aio_next_io_cb, acb);
     }
     return;
 
@@ -XXX,XX +XXX,XX @@ static void qed_aio_write_main(void *opaque, int ret)
     }
 
     if (acb->find_cluster_ret == QED_CLUSTER_FOUND) {
-        next_fn = qed_aio_next_io;
+        next_fn = qed_aio_next_io_cb;
     } else {
         if (s->bs->backing) {
             next_fn = qed_aio_write_flush_before_l2_update;
@@ -XXX,XX +XXX,XX @@ static void qed_aio_write_alloc(QEDAIOCB *acb, size_t len)
     if (acb->flags & QED_AIOCB_ZERO) {
         /* Skip ahead if the clusters are already zero */
         if (acb->find_cluster_ret == QED_CLUSTER_ZERO) {
-            qed_aio_next_io(acb, 0);
+            qed_aio_start_io(acb);
             return;
         }
 
@@ -XXX,XX +XXX,XX @@ static void qed_aio_read_data(void *opaque, int ret,
     /* Handle zero cluster and backing file reads */
     if (ret == QED_CLUSTER_ZERO) {
         qemu_iovec_memset(&acb->cur_qiov, 0, 0, acb->cur_qiov.size);
-        qed_aio_next_io(acb, 0);
+        qed_aio_start_io(acb);
         return;
     } else if (ret != QED_CLUSTER_FOUND) {
         qed_read_backing_file(s, acb->cur_pos, &acb->cur_qiov,
-                              &acb->backing_qiov, qed_aio_next_io, acb);
+                              &acb->backing_qiov, qed_aio_next_io_cb, acb);
         return;
     }
 
     BLKDBG_EVENT(bs->file, BLKDBG_READ_AIO);
     bdrv_aio_readv(bs->file, offset / BDRV_SECTOR_SIZE,
                    &acb->cur_qiov, acb->cur_qiov.size / BDRV_SECTOR_SIZE,
-                   qed_aio_next_io, acb);
+                   qed_aio_next_io_cb, acb);
     return;
 
 err:
@@ -XXX,XX +XXX,XX @@ err:
 /**
  * Begin next I/O or complete the request
  */
-static void qed_aio_next_io(void *opaque, int ret)
+static void qed_aio_next_io(QEDAIOCB *acb, int ret)
 {
-    QEDAIOCB *acb = opaque;
     BDRVQEDState *s = acb_to_s(acb);
     QEDFindClusterFunc *io_fn = (acb->flags & QED_AIOCB_WRITE) ?
                                 qed_aio_write_data : qed_aio_read_data;
@@ -XXX,XX +XXX,XX @@ static BlockAIOCB *qed_aio_setup(BlockDriverState *bs,
     qemu_iovec_init(&acb->cur_qiov, qiov->niov);
 
     /* Start request */
-    qed_aio_next_io(acb, 0);
+    qed_aio_start_io(acb);
     return &acb->common;
 }
 
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

The AioContext data structures are now protected by list_lock and/or
they are walked with FOREACH_RCU primitives.  There is no need anymore
to acquire the AioContext for the entire duration of aio_dispatch.
Instead, just acquire it before and after invoking the callbacks.
The next step is then to push it further down.

diff --git a/util/aio-posix.c b/util/aio-posix.c
index XXXXXXX..XXXXXXX 100644
--- a/util/aio-posix.c
+++ b/util/aio-posix.c
@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
             (revents & (G_IO_IN | G_IO_HUP | G_IO_ERR)) &&
             aio_node_check(ctx, node->is_external) &&
             node->io_read) {
+            aio_context_acquire(ctx);
             node->io_read(node->opaque);
+            aio_context_release(ctx);
 
             /* aio_notify() does not count as progress */
             if (node->opaque != &ctx->notifier) {
@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
             (revents & (G_IO_OUT | G_IO_ERR)) &&
             aio_node_check(ctx, node->is_external) &&
             node->io_write) {
+            aio_context_acquire(ctx);
             node->io_write(node->opaque);
+            aio_context_release(ctx);
             progress = true;
         }
 
@@ -XXX,XX +XXX,XX @@ bool aio_dispatch(AioContext *ctx, bool dispatch_fds)
     }
 
     /* Run our timers */
+    aio_context_acquire(ctx);
     progress |= timerlistgroup_run_timers(&ctx->tlg);
+    aio_context_release(ctx);
 
     return progress;
 }
@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
     int64_t timeout;
     int64_t start = 0;
 
-    aio_context_acquire(ctx);
-    progress = false;
-
     /* aio_notify can avoid the expensive event_notifier_set if
      * everything (file descriptors, bottom halves, timers) will
      * be re-evaluated before the next blocking poll().  This is
@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
         start = qemu_clock_get_ns(QEMU_CLOCK_REALTIME);
     }
 
-    if (try_poll_mode(ctx, blocking)) {
-        progress = true;
-    } else {
+    aio_context_acquire(ctx);
+    progress = try_poll_mode(ctx, blocking);
+    aio_context_release(ctx);
+
+    if (!progress) {
         assert(npfd == 0);
 
         /* fill pollfds */
@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
         timeout = blocking ? aio_compute_timeout(ctx) : 0;
 
         /* wait until next event */
-        if (timeout) {
-            aio_context_release(ctx);
-        }
         if (aio_epoll_check_poll(ctx, pollfds, npfd, timeout)) {
             AioHandler epoll_handler;
 
@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
         } else  {
             ret = qemu_poll_ns(pollfds, npfd, timeout);
         }
-        if (timeout) {
-            aio_context_acquire(ctx);
-        }
     }
 
     if (blocking) {
@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
         progress = true;
     }
 
-    aio_context_release(ctx);
-
     return progress;
 }
 
diff --git a/util/aio-win32.c b/util/aio-win32.c
index XXXXXXX..XXXXXXX 100644
--- a/util/aio-win32.c
+++ b/util/aio-win32.c
@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx, HANDLE event)
             (revents || event_notifier_get_handle(node->e) == event) &&
             node->io_notify) {
             node->pfd.revents = 0;
+            aio_context_acquire(ctx);
             node->io_notify(node->e);
+            aio_context_release(ctx);
 
             /* aio_notify() does not count as progress */
             if (node->e != &ctx->notifier) {
@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx, HANDLE event)
             (node->io_read || node->io_write)) {
             node->pfd.revents = 0;
             if ((revents & G_IO_IN) && node->io_read) {
+                aio_context_acquire(ctx);
                 node->io_read(node->opaque);
+                aio_context_release(ctx);
                 progress = true;
             }
             if ((revents & G_IO_OUT) && node->io_write) {
+                aio_context_acquire(ctx);
                 node->io_write(node->opaque);
+                aio_context_release(ctx);
                 progress = true;
             }
 
@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
     int count;
     int timeout;
 
-    aio_context_acquire(ctx);
     progress = false;
 
     /* aio_notify can avoid the expensive event_notifier_set if
@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
 
         timeout = blocking && !have_select_revents
             ? qemu_timeout_ns_to_ms(aio_compute_timeout(ctx)) : 0;
-        if (timeout) {
-            aio_context_release(ctx);
-        }
         ret = WaitForMultipleObjects(count, events, FALSE, timeout);
         if (blocking) {
             assert(first);
             atomic_sub(&ctx->notify_me, 2);
         }
-        if (timeout) {
-            aio_context_acquire(ctx);
-        }
 
         if (first) {
             aio_notify_accept(ctx);
@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
         progress |= aio_dispatch_handlers(ctx, event);
     } while (count > 0);
 
+    aio_context_acquire(ctx);
     progress |= timerlistgroup_run_timers(&ctx->tlg);
-
     aio_context_release(ctx);
     return progress;
 }
diff --git a/util/async.c b/util/async.c
index XXXXXXX..XXXXXXX 100644
--- a/util/async.c
+++ b/util/async.c
@@ -XXX,XX +XXX,XX @@ int aio_bh_poll(AioContext *ctx)
                 ret = 1;
             }
             bh->idle = 0;
+            aio_context_acquire(ctx);
             aio_bh_call(bh);
+            aio_context_release(ctx);
         }
         if (bh->deleted) {
             deleted = true;
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170213135235.12274-13-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 block/qed.h                 |  3 +++
 block/curl.c                |  2 ++
 block/io.c                  |  5 +++++
 block/iscsi.c               |  8 ++++++--
 block/null.c                |  4 ++++
 block/qed.c                 | 12 ++++++++++++
 block/throttle-groups.c     |  2 ++
 util/aio-posix.c            |  2 --
 util/aio-win32.c            |  2 --
 util/qemu-coroutine-sleep.c |  2 +-
 10 files changed, 35 insertions(+), 7 deletions(-)

diff --git a/block/qed.h b/block/qed.h
index XXXXXXX..XXXXXXX 100644
--- a/block/qed.h
+++ b/block/qed.h
@@ -XXX,XX +XXX,XX @@ enum {
  */
 typedef void QEDFindClusterFunc(void *opaque, int ret, uint64_t offset, size_t len);
 
+void qed_acquire(BDRVQEDState *s);
+void qed_release(BDRVQEDState *s);
+
 /**
  * Generic callback for chaining async callbacks
  */
diff --git a/block/curl.c b/block/curl.c
index XXXXXXX..XXXXXXX 100644
--- a/block/curl.c
+++ b/block/curl.c
@@ -XXX,XX +XXX,XX @@ static void curl_multi_timeout_do(void *arg)
         return;
     }
 
+    aio_context_acquire(s->aio_context);
     curl_multi_socket_action(s->multi, CURL_SOCKET_TIMEOUT, 0, &running);
 
     curl_multi_check_completion(s);
+    aio_context_release(s->aio_context);
 #else
     abort();
 #endif
diff --git a/block/io.c b/block/io.c
index XXXXXXX..XXXXXXX 100644
--- a/block/io.c
+++ b/block/io.c
@@ -XXX,XX +XXX,XX @@ void bdrv_aio_cancel(BlockAIOCB *acb)
         if (acb->aiocb_info->get_aio_context) {
             aio_poll(acb->aiocb_info->get_aio_context(acb), true);
         } else if (acb->bs) {
+            /* qemu_aio_ref and qemu_aio_unref are not thread-safe, so
+             * assert that we're not using an I/O thread.  Thread-safe
+             * code should use bdrv_aio_cancel_async exclusively.
+             */
+            assert(bdrv_get_aio_context(acb->bs) == qemu_get_aio_context());
             aio_poll(bdrv_get_aio_context(acb->bs), true);
         } else {
             abort();
diff --git a/block/iscsi.c b/block/iscsi.c
index XXXXXXX..XXXXXXX 100644
--- a/block/iscsi.c
+++ b/block/iscsi.c
@@ -XXX,XX +XXX,XX @@ static void iscsi_retry_timer_expired(void *opaque)
     struct IscsiTask *iTask = opaque;
     iTask->complete = 1;
     if (iTask->co) {
-        qemu_coroutine_enter(iTask->co);
+        aio_co_wake(iTask->co);
     }
 }
 
@@ -XXX,XX +XXX,XX @@ static void iscsi_nop_timed_event(void *opaque)
 {
     IscsiLun *iscsilun = opaque;
 
+    aio_context_acquire(iscsilun->aio_context);
     if (iscsi_get_nops_in_flight(iscsilun->iscsi) >= MAX_NOP_FAILURES) {
         error_report("iSCSI: NOP timeout. Reconnecting...");
         iscsilun->request_timed_out = true;
     } else if (iscsi_nop_out_async(iscsilun->iscsi, NULL, NULL, 0, NULL) != 0) {
         error_report("iSCSI: failed to sent NOP-Out. Disabling NOP messages.");
-        return;
+        goto out;
     }
 
     timer_mod(iscsilun->nop_timer, qemu_clock_get_ms(QEMU_CLOCK_REALTIME) + NOP_INTERVAL);
     iscsi_set_events(iscsilun);
+
+out:
+    aio_context_release(iscsilun->aio_context);
 }
 
 static void iscsi_readcapacity_sync(IscsiLun *iscsilun, Error **errp)
diff --git a/block/null.c b/block/null.c
index XXXXXXX..XXXXXXX 100644
--- a/block/null.c
+++ b/block/null.c
@@ -XXX,XX +XXX,XX @@ static void null_bh_cb(void *opaque)
 static void null_timer_cb(void *opaque)
 {
     NullAIOCB *acb = opaque;
+    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
+
+    aio_context_acquire(ctx);
     acb->common.cb(acb->common.opaque, 0);
+    aio_context_release(ctx);
     timer_deinit(&acb->timer);
     qemu_aio_unref(acb);
 }
diff --git a/block/qed.c b/block/qed.c
index XXXXXXX..XXXXXXX 100644
--- a/block/qed.c
+++ b/block/qed.c
@@ -XXX,XX +XXX,XX @@ static void qed_need_check_timer_cb(void *opaque)
 
     trace_qed_need_check_timer_cb(s);
 
+    qed_acquire(s);
     qed_plug_allocating_write_reqs(s);
 
     /* Ensure writes are on disk before clearing flag */
     bdrv_aio_flush(s->bs->file->bs, qed_clear_need_check, s);
+    qed_release(s);
+}
+
+void qed_acquire(BDRVQEDState *s)
+{
+    aio_context_acquire(bdrv_get_aio_context(s->bs));
+}
+
+void qed_release(BDRVQEDState *s)
+{
+    aio_context_release(bdrv_get_aio_context(s->bs));
 }
 
 static void qed_start_need_check_timer(BDRVQEDState *s)
diff --git a/block/throttle-groups.c b/block/throttle-groups.c
index XXXXXXX..XXXXXXX 100644
--- a/block/throttle-groups.c
+++ b/block/throttle-groups.c
@@ -XXX,XX +XXX,XX @@ static void timer_cb(BlockBackend *blk, bool is_write)
     qemu_mutex_unlock(&tg->lock);
 
     /* Run the request that was waiting for this timer */
+    aio_context_acquire(blk_get_aio_context(blk));
     empty_queue = !qemu_co_enter_next(&blkp->throttled_reqs[is_write]);
+    aio_context_release(blk_get_aio_context(blk));
 
     /* If the request queue was empty then we have to take care of
      * scheduling the next one */
diff --git a/util/aio-posix.c b/util/aio-posix.c
index XXXXXXX..XXXXXXX 100644
--- a/util/aio-posix.c
+++ b/util/aio-posix.c
@@ -XXX,XX +XXX,XX @@ bool aio_dispatch(AioContext *ctx, bool dispatch_fds)
     }
 
     /* Run our timers */
-    aio_context_acquire(ctx);
     progress |= timerlistgroup_run_timers(&ctx->tlg);
-    aio_context_release(ctx);
 
     return progress;
 }
diff --git a/util/aio-win32.c b/util/aio-win32.c
index XXXXXXX..XXXXXXX 100644
--- a/util/aio-win32.c
+++ b/util/aio-win32.c
@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
         progress |= aio_dispatch_handlers(ctx, event);
     } while (count > 0);
 
-    aio_context_acquire(ctx);
     progress |= timerlistgroup_run_timers(&ctx->tlg);
-    aio_context_release(ctx);
     return progress;
 }
 
diff --git a/util/qemu-coroutine-sleep.c b/util/qemu-coroutine-sleep.c
index XXXXXXX..XXXXXXX 100644
--- a/util/qemu-coroutine-sleep.c
+++ b/util/qemu-coroutine-sleep.c
@@ -XXX,XX +XXX,XX @@ static void co_sleep_cb(void *opaque)
 {
     CoSleepCB *sleep_cb = opaque;
 
-    qemu_coroutine_enter(sleep_cb->co);
+    aio_co_wake(sleep_cb->co);
 }
 
 void coroutine_fn co_aio_sleep_ns(AioContext *ctx, QEMUClockType type,
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

This covers both file descriptor callbacks and polling callbacks,
since they execute related code.

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170213135235.12274-14-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 block/curl.c          | 16 +++++++++++++---
 block/iscsi.c         |  4 ++++
 block/linux-aio.c     |  4 ++++
 block/nfs.c           |  6 ++++++
 block/sheepdog.c      | 29 +++++++++++++++--------------
 block/ssh.c           | 29 +++++++++--------------------
 block/win32-aio.c     | 10 ++++++----
 hw/block/virtio-blk.c |  5 ++++-
 hw/scsi/virtio-scsi.c |  7 +++++++
 util/aio-posix.c      |  7 -------
 util/aio-win32.c      |  6 ------
 11 files changed, 68 insertions(+), 55 deletions(-)

diff --git a/block/curl.c b/block/curl.c
index XXXXXXX..XXXXXXX 100644
--- a/block/curl.c
+++ b/block/curl.c
@@ -XXX,XX +XXX,XX @@ static void curl_multi_check_completion(BDRVCURLState *s)
     }
 }
 
-static void curl_multi_do(void *arg)
+static void curl_multi_do_locked(CURLState *s)
 {
-    CURLState *s = (CURLState *)arg;
     CURLSocket *socket, *next_socket;
     int running;
     int r;
@@ -XXX,XX +XXX,XX @@ static void curl_multi_do(void *arg)
     }
 }
 
+static void curl_multi_do(void *arg)
+{
+    CURLState *s = (CURLState *)arg;
+
+    aio_context_acquire(s->s->aio_context);
+    curl_multi_do_locked(s);
+    aio_context_release(s->s->aio_context);
+}
+
 static void curl_multi_read(void *arg)
 {
     CURLState *s = (CURLState *)arg;
 
-    curl_multi_do(arg);
+    aio_context_acquire(s->s->aio_context);
+    curl_multi_do_locked(s);
     curl_multi_check_completion(s->s);
+    aio_context_release(s->s->aio_context);
 }
 
 static void curl_multi_timeout_do(void *arg)
diff --git a/block/iscsi.c b/block/iscsi.c
index XXXXXXX..XXXXXXX 100644
--- a/block/iscsi.c
+++ b/block/iscsi.c
@@ -XXX,XX +XXX,XX @@ iscsi_process_read(void *arg)
     IscsiLun *iscsilun = arg;
     struct iscsi_context *iscsi = iscsilun->iscsi;
 
+    aio_context_acquire(iscsilun->aio_context);
     iscsi_service(iscsi, POLLIN);
     iscsi_set_events(iscsilun);
+    aio_context_release(iscsilun->aio_context);
 }
 
 static void
@@ -XXX,XX +XXX,XX @@ iscsi_process_write(void *arg)
     IscsiLun *iscsilun = arg;
     struct iscsi_context *iscsi = iscsilun->iscsi;
 
+    aio_context_acquire(iscsilun->aio_context);
     iscsi_service(iscsi, POLLOUT);
     iscsi_set_events(iscsilun);
+    aio_context_release(iscsilun->aio_context);
 }
 
 static int64_t sector_lun2qemu(int64_t sector, IscsiLun *iscsilun)
diff --git a/block/linux-aio.c b/block/linux-aio.c
index XXXXXXX..XXXXXXX 100644
--- a/block/linux-aio.c
+++ b/block/linux-aio.c
@@ -XXX,XX +XXX,XX @@ static void qemu_laio_completion_cb(EventNotifier *e)
     LinuxAioState *s = container_of(e, LinuxAioState, e);
 
     if (event_notifier_test_and_clear(&s->e)) {
+        aio_context_acquire(s->aio_context);
         qemu_laio_process_completions_and_submit(s);
+        aio_context_release(s->aio_context);
     }
 }
 
@@ -XXX,XX +XXX,XX @@ static bool qemu_laio_poll_cb(void *opaque)
         return false;
     }
 
+    aio_context_acquire(s->aio_context);
     qemu_laio_process_completions_and_submit(s);
+    aio_context_release(s->aio_context);
     return true;
 }
 
diff --git a/block/nfs.c b/block/nfs.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nfs.c
+++ b/block/nfs.c
@@ -XXX,XX +XXX,XX @@ static void nfs_set_events(NFSClient *client)
 static void nfs_process_read(void *arg)
 {
     NFSClient *client = arg;
+
+    aio_context_acquire(client->aio_context);
     nfs_service(client->context, POLLIN);
     nfs_set_events(client);
+    aio_context_release(client->aio_context);
 }
 
 static void nfs_process_write(void *arg)
 {
     NFSClient *client = arg;
+
+    aio_context_acquire(client->aio_context);
     nfs_service(client->context, POLLOUT);
     nfs_set_events(client);
+    aio_context_release(client->aio_context);
 }
 
 static void nfs_co_init_task(BlockDriverState *bs, NFSRPC *task)
diff --git a/block/sheepdog.c b/block/sheepdog.c
index XXXXXXX..XXXXXXX 100644
--- a/block/sheepdog.c
+++ b/block/sheepdog.c
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int send_co_req(int sockfd, SheepdogReq *hdr, void *data,
     return ret;
 }
 
-static void restart_co_req(void *opaque)
-{
-    Coroutine *co = opaque;
-
-    qemu_coroutine_enter(co);
-}
-
 typedef struct SheepdogReqCo {
     int sockfd;
     BlockDriverState *bs;
@@ -XXX,XX +XXX,XX @@ typedef struct SheepdogReqCo {
     unsigned int *rlen;
     int ret;
     bool finished;
+    Coroutine *co;
 } SheepdogReqCo;
 
+static void restart_co_req(void *opaque)
+{
+    SheepdogReqCo *srco = opaque;
+
+    aio_co_wake(srco->co);
+}
+
 static coroutine_fn void do_co_req(void *opaque)
 {
     int ret;
-    Coroutine *co;
     SheepdogReqCo *srco = opaque;
     int sockfd = srco->sockfd;
     SheepdogReq *hdr = srco->hdr;
@@ -XXX,XX +XXX,XX @@ static coroutine_fn void do_co_req(void *opaque)
     unsigned int *wlen = srco->wlen;
     unsigned int *rlen = srco->rlen;
 
-    co = qemu_coroutine_self();
+    srco->co = qemu_coroutine_self();
     aio_set_fd_handler(srco->aio_context, sockfd, false,
-                       NULL, restart_co_req, NULL, co);
+                       NULL, restart_co_req, NULL, srco);
 
     ret = send_co_req(sockfd, hdr, data, wlen);
     if (ret < 0) {
@@ -XXX,XX +XXX,XX @@ static coroutine_fn void do_co_req(void *opaque)
     }
 
     aio_set_fd_handler(srco->aio_context, sockfd, false,
-                       restart_co_req, NULL, NULL, co);
+                       restart_co_req, NULL, NULL, srco);
 
     ret = qemu_co_recv(sockfd, hdr, sizeof(*hdr));
     if (ret != sizeof(*hdr)) {
@@ -XXX,XX +XXX,XX @@ out:
     aio_set_fd_handler(srco->aio_context, sockfd, false,
                        NULL, NULL, NULL, NULL);
 
+    srco->co = NULL;
     srco->ret = ret;
     srco->finished = true;
     if (srco->bs) {
@@ -XXX,XX +XXX,XX @@ static void coroutine_fn aio_read_response(void *opaque)
          * We've finished all requests which belong to the AIOCB, so
          * we can switch back to sd_co_readv/writev now.
          */
-        qemu_coroutine_enter(acb->coroutine);
+        aio_co_wake(acb->coroutine);
     }
 
     return;
@@ -XXX,XX +XXX,XX @@ static void co_read_response(void *opaque)
         s->co_recv = qemu_coroutine_create(aio_read_response, opaque);
     }
 
-    qemu_coroutine_enter(s->co_recv);
+    aio_co_wake(s->co_recv);
 }
 
 static void co_write_request(void *opaque)
 {
     BDRVSheepdogState *s = opaque;
 
-    qemu_coroutine_enter(s->co_send);
+    aio_co_wake(s->co_send);
 }
 
 /*
diff --git a/block/ssh.c b/block/ssh.c
index XXXXXXX..XXXXXXX 100644
--- a/block/ssh.c
+++ b/block/ssh.c
@@ -XXX,XX +XXX,XX @@ static void restart_coroutine(void *opaque)
 
     DPRINTF("co=%p", co);
 
-    qemu_coroutine_enter(co);
+    aio_co_wake(co);
 }
 
-static coroutine_fn void set_fd_handler(BDRVSSHState *s, BlockDriverState *bs)
+/* A non-blocking call returned EAGAIN, so yield, ensuring the
+ * handlers are set up so that we'll be rescheduled when there is an
+ * interesting event on the socket.
+ */
+static coroutine_fn void co_yield(BDRVSSHState *s, BlockDriverState *bs)
 {
     int r;
     IOHandler *rd_handler = NULL, *wr_handler = NULL;
@@ -XXX,XX +XXX,XX @@ static coroutine_fn void set_fd_handler(BDRVSSHState *s, BlockDriverState *bs)
 
     aio_set_fd_handler(bdrv_get_aio_context(bs), s->sock,
                        false, rd_handler, wr_handler, NULL, co);
-}
-
-static coroutine_fn void clear_fd_handler(BDRVSSHState *s,
-                                          BlockDriverState *bs)
-{
-    DPRINTF("s->sock=%d", s->sock);
-    aio_set_fd_handler(bdrv_get_aio_context(bs), s->sock,
-                       false, NULL, NULL, NULL, NULL);
-}
-
-/* A non-blocking call returned EAGAIN, so yield, ensuring the
- * handlers are set up so that we'll be rescheduled when there is an
- * interesting event on the socket.
- */
-static coroutine_fn void co_yield(BDRVSSHState *s, BlockDriverState *bs)
-{
-    set_fd_handler(s, bs);
     qemu_coroutine_yield();
-    clear_fd_handler(s, bs);
+    DPRINTF("s->sock=%d - back", s->sock);
+    aio_set_fd_handler(bdrv_get_aio_context(bs), s->sock, false,
+                       NULL, NULL, NULL, NULL);
 }
 
 /* SFTP has a function `libssh2_sftp_seek64' which seeks to a position
diff --git a/block/win32-aio.c b/block/win32-aio.c
index XXXXXXX..XXXXXXX 100644
--- a/block/win32-aio.c
+++ b/block/win32-aio.c
@@ -XXX,XX +XXX,XX @@ struct QEMUWin32AIOState {
     HANDLE hIOCP;
     EventNotifier e;
     int count;
-    bool is_aio_context_attached;
+    AioContext *aio_ctx;
 };
 
 typedef struct QEMUWin32AIOCB {
@@ -XXX,XX +XXX,XX @@ static void win32_aio_process_completion(QEMUWin32AIOState *s,
     }
 
 
+    aio_context_acquire(s->aio_ctx);
     waiocb->common.cb(waiocb->common.opaque, ret);
+    aio_context_release(s->aio_ctx);
     qemu_aio_unref(waiocb);
 }
 
@@ -XXX,XX +XXX,XX @@ void win32_aio_detach_aio_context(QEMUWin32AIOState *aio,
                                   AioContext *old_context)
 {
     aio_set_event_notifier(old_context, &aio->e, false, NULL, NULL);
-    aio->is_aio_context_attached = false;
+    aio->aio_ctx = NULL;
 }
 
 void win32_aio_attach_aio_context(QEMUWin32AIOState *aio,
                                   AioContext *new_context)
 {
-    aio->is_aio_context_attached = true;
+    aio->aio_ctx = new_context;
     aio_set_event_notifier(new_context, &aio->e, false,
                            win32_aio_completion_cb, NULL);
 }
@@ -XXX,XX +XXX,XX @@ out_free_state:
 
 void win32_aio_cleanup(QEMUWin32AIOState *aio)
 {
-    assert(!aio->is_aio_context_attached);
+    assert(!aio->aio_ctx);
     CloseHandle(aio->hIOCP);
     event_notifier_cleanup(&aio->e);
     g_free(aio);
diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/block/virtio-blk.c
+++ b/hw/block/virtio-blk.c
@@ -XXX,XX +XXX,XX @@ static void virtio_blk_ioctl_complete(void *opaque, int status)
 {
     VirtIOBlockIoctlReq *ioctl_req = opaque;
     VirtIOBlockReq *req = ioctl_req->req;
-    VirtIODevice *vdev = VIRTIO_DEVICE(req->dev);
+    VirtIOBlock *s = req->dev;
+    VirtIODevice *vdev = VIRTIO_DEVICE(s);
     struct virtio_scsi_inhdr *scsi;
     struct sg_io_hdr *hdr;
 
@@ -XXX,XX +XXX,XX @@ bool virtio_blk_handle_vq(VirtIOBlock *s, VirtQueue *vq)
     MultiReqBuffer mrb = {};
     bool progress = false;
 
+    aio_context_acquire(blk_get_aio_context(s->blk));
     blk_io_plug(s->blk);
 
     do {
@@ -XXX,XX +XXX,XX @@ bool virtio_blk_handle_vq(VirtIOBlock *s, VirtQueue *vq)
     }
 
     blk_io_unplug(s->blk);
+    aio_context_release(blk_get_aio_context(s->blk));
     return progress;
 }
 
diff --git a/hw/scsi/virtio-scsi.c b/hw/scsi/virtio-scsi.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/scsi/virtio-scsi.c
+++ b/hw/scsi/virtio-scsi.c
@@ -XXX,XX +XXX,XX @@ bool virtio_scsi_handle_ctrl_vq(VirtIOSCSI *s, VirtQueue *vq)
     VirtIOSCSIReq *req;
     bool progress = false;
 
+    virtio_scsi_acquire(s);
     while ((req = virtio_scsi_pop_req(s, vq))) {
         progress = true;
         virtio_scsi_handle_ctrl_req(s, req);
     }
+    virtio_scsi_release(s);
     return progress;
 }
 
@@ -XXX,XX +XXX,XX @@ bool virtio_scsi_handle_cmd_vq(VirtIOSCSI *s, VirtQueue *vq)
 
     QTAILQ_HEAD(, VirtIOSCSIReq) reqs = QTAILQ_HEAD_INITIALIZER(reqs);
 
+    virtio_scsi_acquire(s);
     do {
         virtio_queue_set_notification(vq, 0);
 
@@ -XXX,XX +XXX,XX @@ bool virtio_scsi_handle_cmd_vq(VirtIOSCSI *s, VirtQueue *vq)
     QTAILQ_FOREACH_SAFE(req, &reqs, next, next) {
         virtio_scsi_handle_cmd_req_submit(s, req);
     }
+    virtio_scsi_release(s);
     return progress;
 }
 
@@ -XXX,XX +XXX,XX @@ out:
 
 bool virtio_scsi_handle_event_vq(VirtIOSCSI *s, VirtQueue *vq)
 {
+    virtio_scsi_acquire(s);
     if (s->events_dropped) {
         virtio_scsi_push_event(s, NULL, VIRTIO_SCSI_T_NO_EVENT, 0);
+        virtio_scsi_release(s);
         return true;
     }
+    virtio_scsi_release(s);
     return false;
 }
 
diff --git a/util/aio-posix.c b/util/aio-posix.c
index XXXXXXX..XXXXXXX 100644
--- a/util/aio-posix.c
+++ b/util/aio-posix.c
@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
             (revents & (G_IO_IN | G_IO_HUP | G_IO_ERR)) &&
             aio_node_check(ctx, node->is_external) &&
             node->io_read) {
-            aio_context_acquire(ctx);
             node->io_read(node->opaque);
-            aio_context_release(ctx);
 
             /* aio_notify() does not count as progress */
             if (node->opaque != &ctx->notifier) {
@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
             (revents & (G_IO_OUT | G_IO_ERR)) &&
             aio_node_check(ctx, node->is_external) &&
             node->io_write) {
-            aio_context_acquire(ctx);
             node->io_write(node->opaque);
-            aio_context_release(ctx);
             progress = true;
         }
 
@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
         start = qemu_clock_get_ns(QEMU_CLOCK_REALTIME);
     }
 
-    aio_context_acquire(ctx);
     progress = try_poll_mode(ctx, blocking);
-    aio_context_release(ctx);
-
     if (!progress) {
         assert(npfd == 0);
 
diff --git a/util/aio-win32.c b/util/aio-win32.c
index XXXXXXX..XXXXXXX 100644
--- a/util/aio-win32.c
+++ b/util/aio-win32.c
@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx, HANDLE event)
             (revents || event_notifier_get_handle(node->e) == event) &&
             node->io_notify) {
             node->pfd.revents = 0;
-            aio_context_acquire(ctx);
             node->io_notify(node->e);
-            aio_context_release(ctx);
 
             /* aio_notify() does not count as progress */
             if (node->e != &ctx->notifier) {
@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx, HANDLE event)
             (node->io_read || node->io_write)) {
             node->pfd.revents = 0;
             if ((revents & G_IO_IN) && node->io_read) {
-                aio_context_acquire(ctx);
                 node->io_read(node->opaque);
-                aio_context_release(ctx);
                 progress = true;
             }
             if ((revents & G_IO_OUT) && node->io_write) {
-                aio_context_acquire(ctx);
                 node->io_write(node->opaque);
-                aio_context_release(ctx);
                 progress = true;
             }
 
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170213135235.12274-15-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 block/archipelago.c   |  3 +++
 block/blkreplay.c     |  2 +-
 block/block-backend.c |  6 ++++++
 block/curl.c          | 26 ++++++++++++++++++--------
 block/gluster.c       |  9 +--------
 block/io.c            |  6 +++++-
 block/iscsi.c         |  6 +++++-
 block/linux-aio.c     | 15 +++++++++------
 block/nfs.c           |  3 ++-
 block/null.c          |  4 ++++
 block/qed.c           |  3 +++
 block/rbd.c           |  4 ++++
 dma-helpers.c         |  2 ++
 hw/block/virtio-blk.c |  2 ++
 hw/scsi/scsi-bus.c    |  2 ++
 util/async.c          |  4 ++--
 util/thread-pool.c    |  2 ++
 17 files changed, 71 insertions(+), 28 deletions(-)

diff --git a/block/archipelago.c b/block/archipelago.c
index XXXXXXX..XXXXXXX 100644
--- a/block/archipelago.c
+++ b/block/archipelago.c
@@ -XXX,XX +XXX,XX @@ static void qemu_archipelago_complete_aio(void *opaque)
 {
     AIORequestData *reqdata = (AIORequestData *) opaque;
     ArchipelagoAIOCB *aio_cb = (ArchipelagoAIOCB *) reqdata->aio_cb;
+    AioContext *ctx = bdrv_get_aio_context(aio_cb->common.bs);
 
+    aio_context_acquire(ctx);
     aio_cb->common.cb(aio_cb->common.opaque, aio_cb->ret);
+    aio_context_release(ctx);
     aio_cb->status = 0;
 
     qemu_aio_unref(aio_cb);
diff --git a/block/blkreplay.c b/block/blkreplay.c
index XXXXXXX..XXXXXXX 100755
--- a/block/blkreplay.c
+++ b/block/blkreplay.c
@@ -XXX,XX +XXX,XX @@ static int64_t blkreplay_getlength(BlockDriverState *bs)
 static void blkreplay_bh_cb(void *opaque)
 {
     Request *req = opaque;
-    qemu_coroutine_enter(req->co);
+    aio_co_wake(req->co);
     qemu_bh_delete(req->bh);
     g_free(req);
 }
diff --git a/block/block-backend.c b/block/block-backend.c
index XXXXXXX..XXXXXXX 100644
--- a/block/block-backend.c
+++ b/block/block-backend.c
@@ -XXX,XX +XXX,XX @@ int blk_make_zero(BlockBackend *blk, BdrvRequestFlags flags)
 static void error_callback_bh(void *opaque)
 {
     struct BlockBackendAIOCB *acb = opaque;
+    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
 
     bdrv_dec_in_flight(acb->common.bs);
+    aio_context_acquire(ctx);
     acb->common.cb(acb->common.opaque, acb->ret);
+    aio_context_release(ctx);
     qemu_aio_unref(acb);
 }
 
@@ -XXX,XX +XXX,XX @@ static void blk_aio_complete(BlkAioEmAIOCB *acb)
 static void blk_aio_complete_bh(void *opaque)
 {
     BlkAioEmAIOCB *acb = opaque;
+    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
 
     assert(acb->has_returned);
+    aio_context_acquire(ctx);
     blk_aio_complete(acb);
+    aio_context_release(ctx);
 }
 
 static BlockAIOCB *blk_aio_prwv(BlockBackend *blk, int64_t offset, int bytes,
diff --git a/block/curl.c b/block/curl.c
index XXXXXXX..XXXXXXX 100644
--- a/block/curl.c
+++ b/block/curl.c
@@ -XXX,XX +XXX,XX @@ static void curl_readv_bh_cb(void *p)
 {
     CURLState *state;
     int running;
+    int ret = -EINPROGRESS;
 
     CURLAIOCB *acb = p;
-    BDRVCURLState *s = acb->common.bs->opaque;
+    BlockDriverState *bs = acb->common.bs;
+    BDRVCURLState *s = bs->opaque;
+    AioContext *ctx = bdrv_get_aio_context(bs);
 
     size_t start = acb->sector_num * BDRV_SECTOR_SIZE;
     size_t end;
 
+    aio_context_acquire(ctx);
+
     // In case we have the requested data already (e.g. read-ahead),
     // we can just call the callback and be done.
     switch (curl_find_buf(s, start, acb->nb_sectors * BDRV_SECTOR_SIZE, acb)) {
@@ -XXX,XX +XXX,XX @@ static void curl_readv_bh_cb(void *p)
             qemu_aio_unref(acb);
             // fall through
         case FIND_RET_WAIT:
-            return;
+            goto out;
         default:
             break;
     }
@@ -XXX,XX +XXX,XX @@ static void curl_readv_bh_cb(void *p)
     // No cache found, so let's start a new request
     state = curl_init_state(acb->common.bs, s);
     if (!state) {
-        acb->common.cb(acb->common.opaque, -EIO);
-        qemu_aio_unref(acb);
-        return;
+        ret = -EIO;
+        goto out;
     }
 
     acb->start = 0;
@@ -XXX,XX +XXX,XX @@ static void curl_readv_bh_cb(void *p)
     state->orig_buf = g_try_malloc(state->buf_len);
     if (state->buf_len && state->orig_buf == NULL) {
         curl_clean_state(state);
-        acb->common.cb(acb->common.opaque, -ENOMEM);
-        qemu_aio_unref(acb);
-        return;
+        ret = -ENOMEM;
+        goto out;
     }
     state->acb[0] = acb;
 
@@ -XXX,XX +XXX,XX @@ static void curl_readv_bh_cb(void *p)
 
     /* Tell curl it needs to kick things off */
     curl_multi_socket_action(s->multi, CURL_SOCKET_TIMEOUT, 0, &running);
+
+out:
+    if (ret != -EINPROGRESS) {
+        acb->common.cb(acb->common.opaque, ret);
+        qemu_aio_unref(acb);
+    }
+    aio_context_release(ctx);
 }
 
 static BlockAIOCB *curl_aio_readv(BlockDriverState *bs,
diff --git a/block/gluster.c b/block/gluster.c
index XXXXXXX..XXXXXXX 100644
--- a/block/gluster.c
+++ b/block/gluster.c
@@ -XXX,XX +XXX,XX @@ static struct glfs *qemu_gluster_init(BlockdevOptionsGluster *gconf,
     return qemu_gluster_glfs_init(gconf, errp);
 }
 
-static void qemu_gluster_complete_aio(void *opaque)
-{
-    GlusterAIOCB *acb = (GlusterAIOCB *)opaque;
-
-    qemu_coroutine_enter(acb->coroutine);
-}
-
 /*
  * AIO callback routine called from GlusterFS thread.
  */
@@ -XXX,XX +XXX,XX @@ static void gluster_finish_aiocb(struct glfs_fd *fd, ssize_t ret, void *arg)
         acb->ret = -EIO; /* Partial read/write - fail it */
     }
 
-    aio_bh_schedule_oneshot(acb->aio_context, qemu_gluster_complete_aio, acb);
+    aio_co_schedule(acb->aio_context, acb->coroutine);
 }
 
 static void qemu_gluster_parse_flags(int bdrv_flags, int *open_flags)
diff --git a/block/io.c b/block/io.c
index XXXXXXX..XXXXXXX 100644
--- a/block/io.c
+++ b/block/io.c
@@ -XXX,XX +XXX,XX @@ static void bdrv_co_drain_bh_cb(void *opaque)
     bdrv_dec_in_flight(bs);
     bdrv_drained_begin(bs);
     data->done = true;
-    qemu_coroutine_enter(co);
+    aio_co_wake(co);
 }
 
 static void coroutine_fn bdrv_co_yield_to_drain(BlockDriverState *bs)
@@ -XXX,XX +XXX,XX @@ static void bdrv_co_complete(BlockAIOCBCoroutine *acb)
 static void bdrv_co_em_bh(void *opaque)
 {
     BlockAIOCBCoroutine *acb = opaque;
+    BlockDriverState *bs = acb->common.bs;
+    AioContext *ctx = bdrv_get_aio_context(bs);
 
     assert(!acb->need_bh);
+    aio_context_acquire(ctx);
     bdrv_co_complete(acb);
+    aio_context_release(ctx);
 }
 
 static void bdrv_co_maybe_schedule_bh(BlockAIOCBCoroutine *acb)
diff --git a/block/iscsi.c b/block/iscsi.c
index XXXXXXX..XXXXXXX 100644
--- a/block/iscsi.c
+++ b/block/iscsi.c
@@ -XXX,XX +XXX,XX @@ static void
 iscsi_bh_cb(void *p)
 {
     IscsiAIOCB *acb = p;
+    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
 
     qemu_bh_delete(acb->bh);
 
     g_free(acb->buf);
     acb->buf = NULL;
 
+    aio_context_acquire(ctx);
     acb->common.cb(acb->common.opaque, acb->status);
+    aio_context_release(ctx);
 
     if (acb->task != NULL) {
         scsi_free_scsi_task(acb->task);
@@ -XXX,XX +XXX,XX @@ iscsi_schedule_bh(IscsiAIOCB *acb)
 static void iscsi_co_generic_bh_cb(void *opaque)
 {
     struct IscsiTask *iTask = opaque;
+
     iTask->complete = 1;
-    qemu_coroutine_enter(iTask->co);
+    aio_co_wake(iTask->co);
 }
 
 static void iscsi_retry_timer_expired(void *opaque)
diff --git a/block/linux-aio.c b/block/linux-aio.c
index XXXXXXX..XXXXXXX 100644
--- a/block/linux-aio.c
+++ b/block/linux-aio.c
@@ -XXX,XX +XXX,XX @@ struct LinuxAioState {
     io_context_t ctx;
     EventNotifier e;
 
-    /* io queue for submit at batch */
+    /* io queue for submit at batch.  Protected by AioContext lock. */
     LaioQueue io_q;
 
-    /* I/O completion processing */
+    /* I/O completion processing.  Only runs in I/O thread.  */
     QEMUBH *completion_bh;
     int event_idx;
     int event_max;
@@ -XXX,XX +XXX,XX @@ static inline ssize_t io_event_ret(struct io_event *ev)
  */
 static void qemu_laio_process_completion(struct qemu_laiocb *laiocb)
 {
+    LinuxAioState *s = laiocb->ctx;
     int ret;
 
     ret = laiocb->ret;
@@ -XXX,XX +XXX,XX @@ static void qemu_laio_process_completion(struct qemu_laiocb *laiocb)
     }
 
     laiocb->ret = ret;
+    aio_context_acquire(s->aio_context);
     if (laiocb->co) {
         /* If the coroutine is already entered it must be in ioq_submit() and
          * will notice laio->ret has been filled in when it eventually runs
@@ -XXX,XX +XXX,XX @@ static void qemu_laio_process_completion(struct qemu_laiocb *laiocb)
         laiocb->common.cb(laiocb->common.opaque, ret);
         qemu_aio_unref(laiocb);
     }
+    aio_context_release(s->aio_context);
 }
 
 /**
@@ -XXX,XX +XXX,XX @@ static void qemu_laio_process_completions(LinuxAioState *s)
 static void qemu_laio_process_completions_and_submit(LinuxAioState *s)
 {
     qemu_laio_process_completions(s);
+
+    aio_context_acquire(s->aio_context);
     if (!s->io_q.plugged && !QSIMPLEQ_EMPTY(&s->io_q.pending)) {
         ioq_submit(s);
     }
+    aio_context_release(s->aio_context);
 }
 
 static void qemu_laio_completion_bh(void *opaque)
@@ -XXX,XX +XXX,XX @@ static void qemu_laio_completion_cb(EventNotifier *e)
     LinuxAioState *s = container_of(e, LinuxAioState, e);
 
     if (event_notifier_test_and_clear(&s->e)) {
-        aio_context_acquire(s->aio_context);
         qemu_laio_process_completions_and_submit(s);
-        aio_context_release(s->aio_context);
     }
 }
 
@@ -XXX,XX +XXX,XX @@ static bool qemu_laio_poll_cb(void *opaque)
         return false;
     }
 
-    aio_context_acquire(s->aio_context);
     qemu_laio_process_completions_and_submit(s);
-    aio_context_release(s->aio_context);
     return true;
 }
 
@@ -XXX,XX +XXX,XX @@ void laio_detach_aio_context(LinuxAioState *s, AioContext *old_context)
 {
     aio_set_event_notifier(old_context, &s->e, false, NULL, NULL);
     qemu_bh_delete(s->completion_bh);
+    s->aio_context = NULL;
 }
 
 void laio_attach_aio_context(LinuxAioState *s, AioContext *new_context)
diff --git a/block/nfs.c b/block/nfs.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nfs.c
+++ b/block/nfs.c
@@ -XXX,XX +XXX,XX @@ static void nfs_co_init_task(BlockDriverState *bs, NFSRPC *task)
 static void nfs_co_generic_bh_cb(void *opaque)
 {
     NFSRPC *task = opaque;
+
     task->complete = 1;
-    qemu_coroutine_enter(task->co);
+    aio_co_wake(task->co);
 }
 
 static void
diff --git a/block/null.c b/block/null.c
index XXXXXXX..XXXXXXX 100644
--- a/block/null.c
+++ b/block/null.c
@@ -XXX,XX +XXX,XX @@ static const AIOCBInfo null_aiocb_info = {
 static void null_bh_cb(void *opaque)
 {
     NullAIOCB *acb = opaque;
+    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
+
+    aio_context_acquire(ctx);
     acb->common.cb(acb->common.opaque, 0);
+    aio_context_release(ctx);
     qemu_aio_unref(acb);
 }
 
diff --git a/block/qed.c b/block/qed.c
index XXXXXXX..XXXXXXX 100644
--- a/block/qed.c
+++ b/block/qed.c
@@ -XXX,XX +XXX,XX @@ static void qed_update_l2_table(BDRVQEDState *s, QEDTable *table, int index,
 static void qed_aio_complete_bh(void *opaque)
 {
     QEDAIOCB *acb = opaque;
+    BDRVQEDState *s = acb_to_s(acb);
     BlockCompletionFunc *cb = acb->common.cb;
     void *user_opaque = acb->common.opaque;
     int ret = acb->bh_ret;
@@ -XXX,XX +XXX,XX @@ static void qed_aio_complete_bh(void *opaque)
     qemu_aio_unref(acb);
 
     /* Invoke callback */
+    qed_acquire(s);
     cb(user_opaque, ret);
+    qed_release(s);
 }
 
 static void qed_aio_complete(QEDAIOCB *acb, int ret)
diff --git a/block/rbd.c b/block/rbd.c
index XXXXXXX..XXXXXXX 100644
--- a/block/rbd.c
+++ b/block/rbd.c
@@ -XXX,XX +XXX,XX @@ shutdown:
 static void qemu_rbd_complete_aio(RADOSCB *rcb)
 {
     RBDAIOCB *acb = rcb->acb;
+    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
     int64_t r;
 
     r = rcb->ret;
@@ -XXX,XX +XXX,XX @@ static void qemu_rbd_complete_aio(RADOSCB *rcb)
         qemu_iovec_from_buf(acb->qiov, 0, acb->bounce, acb->qiov->size);
     }
     qemu_vfree(acb->bounce);
+
+    aio_context_acquire(ctx);
     acb->common.cb(acb->common.opaque, (acb->ret > 0 ? 0 : acb->ret));
+    aio_context_release(ctx);
 
     qemu_aio_unref(acb);
 }
diff --git a/dma-helpers.c b/dma-helpers.c
index XXXXXXX..XXXXXXX 100644
--- a/dma-helpers.c
+++ b/dma-helpers.c
@@ -XXX,XX +XXX,XX @@ static void dma_blk_cb(void *opaque, int ret)
                                 QEMU_ALIGN_DOWN(dbs->iov.size, dbs->align));
     }
 
+    aio_context_acquire(dbs->ctx);
     dbs->acb = dbs->io_func(dbs->offset, &dbs->iov,
                             dma_blk_cb, dbs, dbs->io_func_opaque);
+    aio_context_release(dbs->ctx);
     assert(dbs->acb);
 }
 
diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/block/virtio-blk.c
+++ b/hw/block/virtio-blk.c
@@ -XXX,XX +XXX,XX @@ static void virtio_blk_dma_restart_bh(void *opaque)
 
     s->rq = NULL;
 
+    aio_context_acquire(blk_get_aio_context(s->conf.conf.blk));
     while (req) {
         VirtIOBlockReq *next = req->next;
         if (virtio_blk_handle_request(req, &mrb)) {
@@ -XXX,XX +XXX,XX @@ static void virtio_blk_dma_restart_bh(void *opaque)
     if (mrb.num_reqs) {
         virtio_blk_submit_multireq(s->blk, &mrb);
     }
+    aio_context_release(blk_get_aio_context(s->conf.conf.blk));
 }
 
 static void virtio_blk_dma_restart_cb(void *opaque, int running,
diff --git a/hw/scsi/scsi-bus.c b/hw/scsi/scsi-bus.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/scsi/scsi-bus.c
+++ b/hw/scsi/scsi-bus.c
@@ -XXX,XX +XXX,XX @@ static void scsi_dma_restart_bh(void *opaque)
     qemu_bh_delete(s->bh);
     s->bh = NULL;
 
+    aio_context_acquire(blk_get_aio_context(s->conf.blk));
     QTAILQ_FOREACH_SAFE(req, &s->requests, next, next) {
         scsi_req_ref(req);
         if (req->retry) {
@@ -XXX,XX +XXX,XX @@ static void scsi_dma_restart_bh(void *opaque)
         }
         scsi_req_unref(req);
     }
+    aio_context_release(blk_get_aio_context(s->conf.blk));
 }
 
 void scsi_req_retry(SCSIRequest *req)
diff --git a/util/async.c b/util/async.c
index XXXXXXX..XXXXXXX 100644
--- a/util/async.c
+++ b/util/async.c
@@ -XXX,XX +XXX,XX @@ int aio_bh_poll(AioContext *ctx)
                 ret = 1;
             }
             bh->idle = 0;
-            aio_context_acquire(ctx);
             aio_bh_call(bh);
-            aio_context_release(ctx);
         }
         if (bh->deleted) {
             deleted = true;
@@ -XXX,XX +XXX,XX @@ static void co_schedule_bh_cb(void *opaque)
         Coroutine *co = QSLIST_FIRST(&straight);
         QSLIST_REMOVE_HEAD(&straight, co_scheduled_next);
         trace_aio_co_schedule_bh_cb(ctx, co);
+        aio_context_acquire(ctx);
         qemu_coroutine_enter(co);
+        aio_context_release(ctx);
     }
 }
 
diff --git a/util/thread-pool.c b/util/thread-pool.c
index XXXXXXX..XXXXXXX 100644
--- a/util/thread-pool.c
+++ b/util/thread-pool.c
@@ -XXX,XX +XXX,XX @@ static void thread_pool_completion_bh(void *opaque)
     ThreadPool *pool = opaque;
     ThreadPoolElement *elem, *next;
 
+    aio_context_acquire(pool->ctx);
 restart:
     QLIST_FOREACH_SAFE(elem, &pool->head, all, next) {
         if (elem->state != THREAD_DONE) {
@@ -XXX,XX +XXX,XX @@ restart:
             qemu_aio_unref(elem);
         }
     }
+    aio_context_release(pool->ctx);
 }
 
 static void thread_pool_cancel(BlockAIOCB *acb)
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170213135235.12274-16-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 block/archipelago.c    |  3 ---
 block/block-backend.c  |  7 -------
 block/curl.c           |  2 +-
 block/io.c             |  6 +-----
 block/iscsi.c          |  3 ---
 block/linux-aio.c      |  5 +----
 block/mirror.c         | 12 +++++++++---
 block/null.c           |  8 --------
 block/qed-cluster.c    |  2 ++
 block/qed-table.c      | 12 ++++++++++--
 block/qed.c            |  4 ++--
 block/rbd.c            |  4 ----
 block/win32-aio.c      |  3 ---
 hw/block/virtio-blk.c  | 12 +++++++++++-
 hw/scsi/scsi-disk.c    | 15 +++++++++++++++
 hw/scsi/scsi-generic.c | 20 +++++++++++++++++---
 util/thread-pool.c     |  4 +++-
 17 files changed, 72 insertions(+), 50 deletions(-)

diff --git a/block/archipelago.c b/block/archipelago.c
index XXXXXXX..XXXXXXX 100644
--- a/block/archipelago.c
+++ b/block/archipelago.c
@@ -XXX,XX +XXX,XX @@ static void qemu_archipelago_complete_aio(void *opaque)
 {
     AIORequestData *reqdata = (AIORequestData *) opaque;
     ArchipelagoAIOCB *aio_cb = (ArchipelagoAIOCB *) reqdata->aio_cb;
-    AioContext *ctx = bdrv_get_aio_context(aio_cb->common.bs);
 
-    aio_context_acquire(ctx);
     aio_cb->common.cb(aio_cb->common.opaque, aio_cb->ret);
-    aio_context_release(ctx);
     aio_cb->status = 0;
 
     qemu_aio_unref(aio_cb);
diff --git a/block/block-backend.c b/block/block-backend.c
index XXXXXXX..XXXXXXX 100644
--- a/block/block-backend.c
+++ b/block/block-backend.c
@@ -XXX,XX +XXX,XX @@ int blk_make_zero(BlockBackend *blk, BdrvRequestFlags flags)
 static void error_callback_bh(void *opaque)
 {
     struct BlockBackendAIOCB *acb = opaque;
-    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
 
     bdrv_dec_in_flight(acb->common.bs);
-    aio_context_acquire(ctx);
     acb->common.cb(acb->common.opaque, acb->ret);
-    aio_context_release(ctx);
     qemu_aio_unref(acb);
 }
 
@@ -XXX,XX +XXX,XX @@ static void blk_aio_complete(BlkAioEmAIOCB *acb)
 static void blk_aio_complete_bh(void *opaque)
 {
     BlkAioEmAIOCB *acb = opaque;
-    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
-
     assert(acb->has_returned);
-    aio_context_acquire(ctx);
     blk_aio_complete(acb);
-    aio_context_release(ctx);
 }
 
 static BlockAIOCB *blk_aio_prwv(BlockBackend *blk, int64_t offset, int bytes,
diff --git a/block/curl.c b/block/curl.c
index XXXXXXX..XXXXXXX 100644
--- a/block/curl.c
+++ b/block/curl.c
@@ -XXX,XX +XXX,XX @@ static void curl_readv_bh_cb(void *p)
     curl_multi_socket_action(s->multi, CURL_SOCKET_TIMEOUT, 0, &running);
 
 out:
+    aio_context_release(ctx);
     if (ret != -EINPROGRESS) {
         acb->common.cb(acb->common.opaque, ret);
         qemu_aio_unref(acb);
     }
-    aio_context_release(ctx);
 }
 
 static BlockAIOCB *curl_aio_readv(BlockDriverState *bs,
diff --git a/block/io.c b/block/io.c
index XXXXXXX..XXXXXXX 100644
--- a/block/io.c
+++ b/block/io.c
@@ -XXX,XX +XXX,XX @@ static void bdrv_co_io_em_complete(void *opaque, int ret)
     CoroutineIOCompletion *co = opaque;
 
     co->ret = ret;
-    qemu_coroutine_enter(co->coroutine);
+    aio_co_wake(co->coroutine);
 }
 
 static int coroutine_fn bdrv_driver_preadv(BlockDriverState *bs,
@@ -XXX,XX +XXX,XX @@ static void bdrv_co_complete(BlockAIOCBCoroutine *acb)
 static void bdrv_co_em_bh(void *opaque)
 {
     BlockAIOCBCoroutine *acb = opaque;
-    BlockDriverState *bs = acb->common.bs;
-    AioContext *ctx = bdrv_get_aio_context(bs);
 
     assert(!acb->need_bh);
-    aio_context_acquire(ctx);
     bdrv_co_complete(acb);
-    aio_context_release(ctx);
 }
 
 static void bdrv_co_maybe_schedule_bh(BlockAIOCBCoroutine *acb)
diff --git a/block/iscsi.c b/block/iscsi.c
index XXXXXXX..XXXXXXX 100644
--- a/block/iscsi.c
+++ b/block/iscsi.c
@@ -XXX,XX +XXX,XX @@ static void
 iscsi_bh_cb(void *p)
 {
     IscsiAIOCB *acb = p;
-    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
 
     qemu_bh_delete(acb->bh);
 
     g_free(acb->buf);
     acb->buf = NULL;
 
-    aio_context_acquire(ctx);
     acb->common.cb(acb->common.opaque, acb->status);
-    aio_context_release(ctx);
 
     if (acb->task != NULL) {
         scsi_free_scsi_task(acb->task);
diff --git a/block/linux-aio.c b/block/linux-aio.c
index XXXXXXX..XXXXXXX 100644
--- a/block/linux-aio.c
+++ b/block/linux-aio.c
@@ -XXX,XX +XXX,XX @@ static inline ssize_t io_event_ret(struct io_event *ev)
  */
 static void qemu_laio_process_completion(struct qemu_laiocb *laiocb)
 {
-    LinuxAioState *s = laiocb->ctx;
     int ret;
 
     ret = laiocb->ret;
@@ -XXX,XX +XXX,XX @@ static void qemu_laio_process_completion(struct qemu_laiocb *laiocb)
     }
 
     laiocb->ret = ret;
-    aio_context_acquire(s->aio_context);
     if (laiocb->co) {
         /* If the coroutine is already entered it must be in ioq_submit() and
          * will notice laio->ret has been filled in when it eventually runs
@@ -XXX,XX +XXX,XX @@ static void qemu_laio_process_completion(struct qemu_laiocb *laiocb)
          * that!
          */
         if (!qemu_coroutine_entered(laiocb->co)) {
-            qemu_coroutine_enter(laiocb->co);
+            aio_co_wake(laiocb->co);
         }
     } else {
         laiocb->common.cb(laiocb->common.opaque, ret);
         qemu_aio_unref(laiocb);
     }
-    aio_context_release(s->aio_context);
 }
 
 /**
diff --git a/block/mirror.c b/block/mirror.c
index XXXXXXX..XXXXXXX 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -XXX,XX +XXX,XX @@ static void mirror_write_complete(void *opaque, int ret)
 {
     MirrorOp *op = opaque;
     MirrorBlockJob *s = op->s;
+
+    aio_context_acquire(blk_get_aio_context(s->common.blk));
     if (ret < 0) {
         BlockErrorAction action;
 
@@ -XXX,XX +XXX,XX @@ static void mirror_write_complete(void *opaque, int ret)
         }
     }
     mirror_iteration_done(op, ret);
+    aio_context_release(blk_get_aio_context(s->common.blk));
 }
 
 static void mirror_read_complete(void *opaque, int ret)
 {
     MirrorOp *op = opaque;
     MirrorBlockJob *s = op->s;
+
+    aio_context_acquire(blk_get_aio_context(s->common.blk));
     if (ret < 0) {
         BlockErrorAction action;
 
@@ -XXX,XX +XXX,XX @@ static void mirror_read_complete(void *opaque, int ret)
         }
 
         mirror_iteration_done(op, ret);
-        return;
+    } else {
+        blk_aio_pwritev(s->target, op->sector_num * BDRV_SECTOR_SIZE, &op->qiov,
+                        0, mirror_write_complete, op);
     }
-    blk_aio_pwritev(s->target, op->sector_num * BDRV_SECTOR_SIZE, &op->qiov,
-                    0, mirror_write_complete, op);
+    aio_context_release(blk_get_aio_context(s->common.blk));
 }
 
 static inline void mirror_clip_sectors(MirrorBlockJob *s,
diff --git a/block/null.c b/block/null.c
index XXXXXXX..XXXXXXX 100644
--- a/block/null.c
+++ b/block/null.c
@@ -XXX,XX +XXX,XX @@ static const AIOCBInfo null_aiocb_info = {
 static void null_bh_cb(void *opaque)
 {
     NullAIOCB *acb = opaque;
-    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
-
-    aio_context_acquire(ctx);
     acb->common.cb(acb->common.opaque, 0);
-    aio_context_release(ctx);
     qemu_aio_unref(acb);
 }
 
 static void null_timer_cb(void *opaque)
 {
     NullAIOCB *acb = opaque;
-    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
-
-    aio_context_acquire(ctx);
     acb->common.cb(acb->common.opaque, 0);
-    aio_context_release(ctx);
     timer_deinit(&acb->timer);
     qemu_aio_unref(acb);
 }
diff --git a/block/qed-cluster.c b/block/qed-cluster.c
index XXXXXXX..XXXXXXX 100644
--- a/block/qed-cluster.c
+++ b/block/qed-cluster.c
@@ -XXX,XX +XXX,XX @@ static void qed_find_cluster_cb(void *opaque, int ret)
     unsigned int index;
     unsigned int n;
 
+    qed_acquire(s);
     if (ret) {
         goto out;
     }
@@ -XXX,XX +XXX,XX @@ static void qed_find_cluster_cb(void *opaque, int ret)
 
 out:
     find_cluster_cb->cb(find_cluster_cb->opaque, ret, offset, len);
+    qed_release(s);
     g_free(find_cluster_cb);
 }
 
diff --git a/block/qed-table.c b/block/qed-table.c
index XXXXXXX..XXXXXXX 100644
--- a/block/qed-table.c
+++ b/block/qed-table.c
@@ -XXX,XX +XXX,XX @@ static void qed_read_table_cb(void *opaque, int ret)
 {
     QEDReadTableCB *read_table_cb = opaque;
     QEDTable *table = read_table_cb->table;
+    BDRVQEDState *s = read_table_cb->s;
     int noffsets = read_table_cb->qiov.size / sizeof(uint64_t);
     int i;
 
@@ -XXX,XX +XXX,XX @@ static void qed_read_table_cb(void *opaque, int ret)
     }
 
     /* Byteswap offsets */
+    qed_acquire(s);
     for (i = 0; i < noffsets; i++) {
         table->offsets[i] = le64_to_cpu(table->offsets[i]);
     }
+    qed_release(s);
 
 out:
     /* Completion */
-    trace_qed_read_table_cb(read_table_cb->s, read_table_cb->table, ret);
+    trace_qed_read_table_cb(s, read_table_cb->table, ret);
     gencb_complete(&read_table_cb->gencb, ret);
 }
 
@@ -XXX,XX +XXX,XX @@ typedef struct {
 static void qed_write_table_cb(void *opaque, int ret)
 {
     QEDWriteTableCB *write_table_cb = opaque;
+    BDRVQEDState *s = write_table_cb->s;
 
-    trace_qed_write_table_cb(write_table_cb->s,
+    trace_qed_write_table_cb(s,
                              write_table_cb->orig_table,
                              write_table_cb->flush,
                              ret);
@@ -XXX,XX +XXX,XX @@ static void qed_write_table_cb(void *opaque, int ret)
     if (write_table_cb->flush) {
         /* We still need to flush first */
         write_table_cb->flush = false;
+        qed_acquire(s);
         bdrv_aio_flush(write_table_cb->s->bs, qed_write_table_cb,
                        write_table_cb);
+        qed_release(s);
         return;
     }
 
@@ -XXX,XX +XXX,XX @@ static void qed_read_l2_table_cb(void *opaque, int ret)
     CachedL2Table *l2_table = request->l2_table;
     uint64_t l2_offset = read_l2_table_cb->l2_offset;
 
+    qed_acquire(s);
     if (ret) {
         /* can't trust loaded L2 table anymore */
         qed_unref_l2_cache_entry(l2_table);
@@ -XXX,XX +XXX,XX @@ static void qed_read_l2_table_cb(void *opaque, int ret)
         request->l2_table = qed_find_l2_cache_entry(&s->l2_cache, l2_offset);
         assert(request->l2_table != NULL);
     }
+    qed_release(s);
 
     gencb_complete(&read_l2_table_cb->gencb, ret);
 }
diff --git a/block/qed.c b/block/qed.c
index XXXXXXX..XXXXXXX 100644
--- a/block/qed.c
+++ b/block/qed.c
@@ -XXX,XX +XXX,XX @@ static void qed_is_allocated_cb(void *opaque, int ret, uint64_t offset, size_t l
     }
 
     if (cb->co) {
-        qemu_coroutine_enter(cb->co);
+        aio_co_wake(cb->co);
     }
 }
 
@@ -XXX,XX +XXX,XX @@ static void coroutine_fn qed_co_pwrite_zeroes_cb(void *opaque, int ret)
     cb->done = true;
     cb->ret = ret;
     if (cb->co) {
-        qemu_coroutine_enter(cb->co);
+        aio_co_wake(cb->co);
     }
 }
 
diff --git a/block/rbd.c b/block/rbd.c
index XXXXXXX..XXXXXXX 100644
--- a/block/rbd.c
+++ b/block/rbd.c
@@ -XXX,XX +XXX,XX @@ shutdown:
 static void qemu_rbd_complete_aio(RADOSCB *rcb)
 {
     RBDAIOCB *acb = rcb->acb;
-    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
     int64_t r;
 
     r = rcb->ret;
@@ -XXX,XX +XXX,XX @@ static void qemu_rbd_complete_aio(RADOSCB *rcb)
         qemu_iovec_from_buf(acb->qiov, 0, acb->bounce, acb->qiov->size);
     }
     qemu_vfree(acb->bounce);
-
-    aio_context_acquire(ctx);
     acb->common.cb(acb->common.opaque, (acb->ret > 0 ? 0 : acb->ret));
-    aio_context_release(ctx);
 
     qemu_aio_unref(acb);
 }
diff --git a/block/win32-aio.c b/block/win32-aio.c
index XXXXXXX..XXXXXXX 100644
--- a/block/win32-aio.c
+++ b/block/win32-aio.c
@@ -XXX,XX +XXX,XX @@ static void win32_aio_process_completion(QEMUWin32AIOState *s,
         qemu_vfree(waiocb->buf);
     }
 
-
-    aio_context_acquire(s->aio_ctx);
     waiocb->common.cb(waiocb->common.opaque, ret);
-    aio_context_release(s->aio_ctx);
     qemu_aio_unref(waiocb);
 }
 
diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/block/virtio-blk.c
+++ b/hw/block/virtio-blk.c
@@ -XXX,XX +XXX,XX @@ static int virtio_blk_handle_rw_error(VirtIOBlockReq *req, int error,
 static void virtio_blk_rw_complete(void *opaque, int ret)
 {
     VirtIOBlockReq *next = opaque;
+    VirtIOBlock *s = next->dev;
 
+    aio_context_acquire(blk_get_aio_context(s->conf.conf.blk));
     while (next) {
         VirtIOBlockReq *req = next;
         next = req->mr_next;
@@ -XXX,XX +XXX,XX @@ static void virtio_blk_rw_complete(void *opaque, int ret)
         block_acct_done(blk_get_stats(req->dev->blk), &req->acct);
         virtio_blk_free_request(req);
     }
+    aio_context_release(blk_get_aio_context(s->conf.conf.blk));
 }
 
 static void virtio_blk_flush_complete(void *opaque, int ret)
 {
     VirtIOBlockReq *req = opaque;
+    VirtIOBlock *s = req->dev;
 
+    aio_context_acquire(blk_get_aio_context(s->conf.conf.blk));
     if (ret) {
         if (virtio_blk_handle_rw_error(req, -ret, 0)) {
-            return;
+            goto out;
         }
     }
 
     virtio_blk_req_complete(req, VIRTIO_BLK_S_OK);
     block_acct_done(blk_get_stats(req->dev->blk), &req->acct);
     virtio_blk_free_request(req);
+
+out:
+    aio_context_release(blk_get_aio_context(s->conf.conf.blk));
 }
 
 #ifdef __linux__
@@ -XXX,XX +XXX,XX @@ static void virtio_blk_ioctl_complete(void *opaque, int status)
     virtio_stl_p(vdev, &scsi->data_len, hdr->dxfer_len);
 
 out:
+    aio_context_acquire(blk_get_aio_context(s->conf.conf.blk));
     virtio_blk_req_complete(req, status);
     virtio_blk_free_request(req);
+    aio_context_release(blk_get_aio_context(s->conf.conf.blk));
     g_free(ioctl_req);
 }
 
diff --git a/hw/scsi/scsi-disk.c b/hw/scsi/scsi-disk.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/scsi/scsi-disk.c
+++ b/hw/scsi/scsi-disk.c
@@ -XXX,XX +XXX,XX @@ static void scsi_aio_complete(void *opaque, int ret)
 
     assert(r->req.aiocb != NULL);
     r->req.aiocb = NULL;
+    aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk));
     if (scsi_disk_req_check_error(r, ret, true)) {
         goto done;
     }
@@ -XXX,XX +XXX,XX @@ static void scsi_aio_complete(void *opaque, int ret)
     scsi_req_complete(&r->req, GOOD);
 
 done:
+    aio_context_release(blk_get_aio_context(s->qdev.conf.blk));
     scsi_req_unref(&r->req);
 }
 
@@ -XXX,XX +XXX,XX @@ static void scsi_dma_complete(void *opaque, int ret)
     assert(r->req.aiocb != NULL);
     r->req.aiocb = NULL;
 
+    aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk));
     if (ret < 0) {
         block_acct_failed(blk_get_stats(s->qdev.conf.blk), &r->acct);
     } else {
         block_acct_done(blk_get_stats(s->qdev.conf.blk), &r->acct);
     }
     scsi_dma_complete_noio(r, ret);
+    aio_context_release(blk_get_aio_context(s->qdev.conf.blk));
 }
 
 static void scsi_read_complete(void * opaque, int ret)
@@ -XXX,XX +XXX,XX @@ static void scsi_read_complete(void * opaque, int ret)
 
     assert(r->req.aiocb != NULL);
     r->req.aiocb = NULL;
+    aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk));
     if (scsi_disk_req_check_error(r, ret, true)) {
         goto done;
     }
@@ -XXX,XX +XXX,XX @@ static void scsi_read_complete(void * opaque, int ret)
 
 done:
     scsi_req_unref(&r->req);
+    aio_context_release(blk_get_aio_context(s->qdev.conf.blk));
 }
 
 /* Actually issue a read to the block device.  */
@@ -XXX,XX +XXX,XX @@ static void scsi_do_read_cb(void *opaque, int ret)
     assert (r->req.aiocb != NULL);
     r->req.aiocb = NULL;
 
+    aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk));
     if (ret < 0) {
         block_acct_failed(blk_get_stats(s->qdev.conf.blk), &r->acct);
     } else {
         block_acct_done(blk_get_stats(s->qdev.conf.blk), &r->acct);
     }
     scsi_do_read(opaque, ret);
+    aio_context_release(blk_get_aio_context(s->qdev.conf.blk));
 }
 
 /* Read more data from scsi device into buffer.  */
@@ -XXX,XX +XXX,XX @@ static void scsi_write_complete(void * opaque, int ret)
     assert (r->req.aiocb != NULL);
     r->req.aiocb = NULL;
 
+    aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk));
     if (ret < 0) {
         block_acct_failed(blk_get_stats(s->qdev.conf.blk), &r->acct);
     } else {
         block_acct_done(blk_get_stats(s->qdev.conf.blk), &r->acct);
     }
     scsi_write_complete_noio(r, ret);
+    aio_context_release(blk_get_aio_context(s->qdev.conf.blk));
 }
 
 static void scsi_write_data(SCSIRequest *req)
@@ -XXX,XX +XXX,XX @@ static void scsi_unmap_complete(void *opaque, int ret)
 {
     UnmapCBData *data = opaque;
     SCSIDiskReq *r = data->r;
+    SCSIDiskState *s = DO_UPCAST(SCSIDiskState, qdev, r->req.dev);
 
     assert(r->req.aiocb != NULL);
     r->req.aiocb = NULL;
 
+    aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk));
     scsi_unmap_complete_noio(data, ret);
+    aio_context_release(blk_get_aio_context(s->qdev.conf.blk));
 }
 
 static void scsi_disk_emulate_unmap(SCSIDiskReq *r, uint8_t *inbuf)
@@ -XXX,XX +XXX,XX @@ static void scsi_write_same_complete(void *opaque, int ret)
 
     assert(r->req.aiocb != NULL);
     r->req.aiocb = NULL;
+    aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk));
     if (scsi_disk_req_check_error(r, ret, true)) {
         goto done;
     }
@@ -XXX,XX +XXX,XX @@ done:
     scsi_req_unref(&r->req);
     qemu_vfree(data->iov.iov_base);
     g_free(data);
+    aio_context_release(blk_get_aio_context(s->qdev.conf.blk));
 }
 
 static void scsi_disk_emulate_write_same(SCSIDiskReq *r, uint8_t *inbuf)
diff --git a/hw/scsi/scsi-generic.c b/hw/scsi/scsi-generic.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/scsi/scsi-generic.c
+++ b/hw/scsi/scsi-generic.c
@@ -XXX,XX +XXX,XX @@ done:
 static void scsi_command_complete(void *opaque, int ret)
 {
     SCSIGenericReq *r = (SCSIGenericReq *)opaque;
+    SCSIDevice *s = r->req.dev;
 
     assert(r->req.aiocb != NULL);
     r->req.aiocb = NULL;
+
+    aio_context_acquire(blk_get_aio_context(s->conf.blk));
     scsi_command_complete_noio(r, ret);
+    aio_context_release(blk_get_aio_context(s->conf.blk));
 }
 
 static int execute_command(BlockBackend *blk,
@@ -XXX,XX +XXX,XX @@ static void scsi_read_complete(void * opaque, int ret)
     assert(r->req.aiocb != NULL);
     r->req.aiocb = NULL;
 
+    aio_context_acquire(blk_get_aio_context(s->conf.blk));
+
     if (ret || r->req.io_canceled) {
         scsi_command_complete_noio(r, ret);
-        return;
+        goto done;
     }
 
     len = r->io_header.dxfer_len - r->io_header.resid;
@@ -XXX,XX +XXX,XX @@ static void scsi_read_complete(void * opaque, int ret)
     r->len = -1;
     if (len == 0) {
         scsi_command_complete_noio(r, 0);
-        return;
+        goto done;
     }
 
     /* Snoop READ CAPACITY output to set the blocksize.  */
@@ -XXX,XX +XXX,XX @@ static void scsi_read_complete(void * opaque, int ret)
     }
     scsi_req_data(&r->req, len);
     scsi_req_unref(&r->req);
+
+done:
+    aio_context_release(blk_get_aio_context(s->conf.blk));
 }
 
 /* Read more data from scsi device into buffer.  */
@@ -XXX,XX +XXX,XX @@ static void scsi_write_complete(void * opaque, int ret)
     assert(r->req.aiocb != NULL);
     r->req.aiocb = NULL;
 
+    aio_context_acquire(blk_get_aio_context(s->conf.blk));
+
     if (ret || r->req.io_canceled) {
         scsi_command_complete_noio(r, ret);
-        return;
+        goto done;
     }
 
     if (r->req.cmd.buf[0] == MODE_SELECT && r->req.cmd.buf[4] == 12 &&
@@ -XXX,XX +XXX,XX @@ static void scsi_write_complete(void * opaque, int ret)
     }
 
     scsi_command_complete_noio(r, ret);
+
+done:
+    aio_context_release(blk_get_aio_context(s->conf.blk));
 }
 
 /* Write data to a scsi device.  Returns nonzero on failure.
diff --git a/util/thread-pool.c b/util/thread-pool.c
index XXXXXXX..XXXXXXX 100644
--- a/util/thread-pool.c
+++ b/util/thread-pool.c
@@ -XXX,XX +XXX,XX @@ restart:
              */
             qemu_bh_schedule(pool->completion_bh);
 
+            aio_context_release(pool->ctx);
             elem->common.cb(elem->common.opaque, elem->ret);
+            aio_context_acquire(pool->ctx);
             qemu_aio_unref(elem);
             goto restart;
         } else {
@@ -XXX,XX +XXX,XX @@ static void thread_pool_co_cb(void *opaque, int ret)
     ThreadPoolCo *co = opaque;
 
     co->ret = ret;
-    qemu_coroutine_enter(co->co);
+    aio_co_wake(co->co);
 }
 
 int coroutine_fn thread_pool_submit_co(ThreadPool *pool, ThreadPoolFunc *func,
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

This patch prepares for the removal of unnecessary lockcnt inc/dec pairs.
Extract the dispatching loop for file descriptor handlers into a new
function aio_dispatch_handlers, and then inline aio_dispatch into
aio_poll.

aio_dispatch can now become void.

diff --git a/include/block/aio.h b/include/block/aio.h
index XXXXXXX..XXXXXXX 100644
--- a/include/block/aio.h
+++ b/include/block/aio.h
@@ -XXX,XX +XXX,XX @@ bool aio_pending(AioContext *ctx);
 /* Dispatch any pending callbacks from the GSource attached to the AioContext.
  *
  * This is used internally in the implementation of the GSource.
- *
- * @dispatch_fds: true to process fds, false to skip them
- *                (can be used as an optimization by callers that know there
- *                are no fds ready)
  */
-bool aio_dispatch(AioContext *ctx, bool dispatch_fds);
+void aio_dispatch(AioContext *ctx);
 
 /* Progress in completing AIO work to occur.  This can issue new pending
  * aio as a result of executing I/O completion or bh callbacks.
diff --git a/util/aio-posix.c b/util/aio-posix.c
index XXXXXXX..XXXXXXX 100644
--- a/util/aio-posix.c
+++ b/util/aio-posix.c
@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
     AioHandler *node, *tmp;
     bool progress = false;
 
-    /*
-     * We have to walk very carefully in case aio_set_fd_handler is
-     * called while we're walking.
-     */
-    qemu_lockcnt_inc(&ctx->list_lock);
-
     QLIST_FOREACH_SAFE_RCU(node, &ctx->aio_handlers, node, tmp) {
         int revents;
 
@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
         }
     }
 
-    qemu_lockcnt_dec(&ctx->list_lock);
     return progress;
 }
 
-/*
- * Note that dispatch_fds == false has the side-effect of post-poning the
- * freeing of deleted handlers.
- */
-bool aio_dispatch(AioContext *ctx, bool dispatch_fds)
+void aio_dispatch(AioContext *ctx)
 {
-    bool progress;
+    aio_bh_poll(ctx);
 
-    /*
-     * If there are callbacks left that have been queued, we need to call them.
-     * Do not call select in this case, because it is possible that the caller
-     * does not need a complete flush (as is the case for aio_poll loops).
-     */
-    progress = aio_bh_poll(ctx);
+    qemu_lockcnt_inc(&ctx->list_lock);
+    aio_dispatch_handlers(ctx);
+    qemu_lockcnt_dec(&ctx->list_lock);
 
-    if (dispatch_fds) {
-        progress |= aio_dispatch_handlers(ctx);
-    }
-
-    /* Run our timers */
-    progress |= timerlistgroup_run_timers(&ctx->tlg);
-
-    return progress;
+    timerlistgroup_run_timers(&ctx->tlg);
 }
 
 /* These thread-local variables are used only in a small part of aio_poll
@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
     npfd = 0;
     qemu_lockcnt_dec(&ctx->list_lock);
 
-    /* Run dispatch even if there were no readable fds to run timers */
-    if (aio_dispatch(ctx, ret > 0)) {
-        progress = true;
+    progress |= aio_bh_poll(ctx);
+
+    if (ret > 0) {
+        qemu_lockcnt_inc(&ctx->list_lock);
+        progress |= aio_dispatch_handlers(ctx);
+        qemu_lockcnt_dec(&ctx->list_lock);
     }
 
+    progress |= timerlistgroup_run_timers(&ctx->tlg);
+
     return progress;
 }
 
diff --git a/util/aio-win32.c b/util/aio-win32.c
index XXXXXXX..XXXXXXX 100644
--- a/util/aio-win32.c
+++ b/util/aio-win32.c
@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx, HANDLE event)
     return progress;
 }
 
-bool aio_dispatch(AioContext *ctx, bool dispatch_fds)
+void aio_dispatch(AioContext *ctx)
 {
-    bool progress;
-
-    progress = aio_bh_poll(ctx);
-    if (dispatch_fds) {
-        progress |= aio_dispatch_handlers(ctx, INVALID_HANDLE_VALUE);
-    }
-    progress |= timerlistgroup_run_timers(&ctx->tlg);
-    return progress;
+    aio_bh_poll(ctx);
+    aio_dispatch_handlers(ctx, INVALID_HANDLE_VALUE);
+    timerlistgroup_run_timers(&ctx->tlg);
 }
 
 bool aio_poll(AioContext *ctx, bool blocking)
diff --git a/util/async.c b/util/async.c
index XXXXXXX..XXXXXXX 100644
--- a/util/async.c
+++ b/util/async.c
@@ -XXX,XX +XXX,XX @@ aio_ctx_dispatch(GSource     *source,
     AioContext *ctx = (AioContext *) source;
 
     assert(callback == NULL);
-    aio_dispatch(ctx, true);
+    aio_dispatch(ctx);
     return true;
 }
 
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

Pull the increment/decrement pair out of aio_bh_poll and into the
callers.

diff --git a/util/aio-posix.c b/util/aio-posix.c
index XXXXXXX..XXXXXXX 100644
--- a/util/aio-posix.c
+++ b/util/aio-posix.c
@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
 
 void aio_dispatch(AioContext *ctx)
 {
+    qemu_lockcnt_inc(&ctx->list_lock);
     aio_bh_poll(ctx);
-
-    qemu_lockcnt_inc(&ctx->list_lock);
     aio_dispatch_handlers(ctx);
     qemu_lockcnt_dec(&ctx->list_lock);
 
@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
     }
 
     npfd = 0;
-    qemu_lockcnt_dec(&ctx->list_lock);
 
     progress |= aio_bh_poll(ctx);
 
     if (ret > 0) {
-        qemu_lockcnt_inc(&ctx->list_lock);
         progress |= aio_dispatch_handlers(ctx);
-        qemu_lockcnt_dec(&ctx->list_lock);
     }
 
+    qemu_lockcnt_dec(&ctx->list_lock);
+
     progress |= timerlistgroup_run_timers(&ctx->tlg);
 
     return progress;
diff --git a/util/aio-win32.c b/util/aio-win32.c
index XXXXXXX..XXXXXXX 100644
--- a/util/aio-win32.c
+++ b/util/aio-win32.c
@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx, HANDLE event)
     bool progress = false;
     AioHandler *tmp;
 
-    qemu_lockcnt_inc(&ctx->list_lock);
-
     /*
      * We have to walk very carefully in case aio_set_fd_handler is
      * called while we're walking.
@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx, HANDLE event)
         }
     }
 
-    qemu_lockcnt_dec(&ctx->list_lock);
     return progress;
 }
 
 void aio_dispatch(AioContext *ctx)
 {
+    qemu_lockcnt_inc(&ctx->list_lock);
     aio_bh_poll(ctx);
     aio_dispatch_handlers(ctx, INVALID_HANDLE_VALUE);
+    qemu_lockcnt_dec(&ctx->list_lock);
     timerlistgroup_run_timers(&ctx->tlg);
 }
 
@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
         }
     }
 
-    qemu_lockcnt_dec(&ctx->list_lock);
     first = true;
 
     /* ctx->notifier is always registered.  */
@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
         progress |= aio_dispatch_handlers(ctx, event);
     } while (count > 0);
 
+    qemu_lockcnt_dec(&ctx->list_lock);
+
     progress |= timerlistgroup_run_timers(&ctx->tlg);
     return progress;
 }
diff --git a/util/async.c b/util/async.c
index XXXXXXX..XXXXXXX 100644
--- a/util/async.c
+++ b/util/async.c
@@ -XXX,XX +XXX,XX @@ void aio_bh_call(QEMUBH *bh)
     bh->cb(bh->opaque);
 }
 
-/* Multiple occurrences of aio_bh_poll cannot be called concurrently */
+/* Multiple occurrences of aio_bh_poll cannot be called concurrently.
+ * The count in ctx->list_lock is incremented before the call, and is
+ * not affected by the call.
+ */
 int aio_bh_poll(AioContext *ctx)
 {
     QEMUBH *bh, **bhp, *next;
     int ret;
     bool deleted = false;
 
-    qemu_lockcnt_inc(&ctx->list_lock);
-
     ret = 0;
     for (bh = atomic_rcu_read(&ctx->first_bh); bh; bh = next) {
         next = atomic_rcu_read(&bh->next);
@@ -XXX,XX +XXX,XX @@ int aio_bh_poll(AioContext *ctx)
 
     /* remove deleted bhs */
     if (!deleted) {
-        qemu_lockcnt_dec(&ctx->list_lock);
         return ret;
     }
 
-    if (qemu_lockcnt_dec_and_lock(&ctx->list_lock)) {
+    if (qemu_lockcnt_dec_if_lock(&ctx->list_lock)) {
         bhp = &ctx->first_bh;
         while (*bhp) {
             bh = *bhp;
@@ -XXX,XX +XXX,XX @@ int aio_bh_poll(AioContext *ctx)
                 bhp = &bh->next;
             }
         }
-        qemu_lockcnt_unlock(&ctx->list_lock);
+        qemu_lockcnt_inc_and_unlock(&ctx->list_lock);
     }
     return ret;
 }
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

diff --git a/include/block/block_int.h b/include/block/block_int.h
index XXXXXXX..XXXXXXX 100644
--- a/include/block/block_int.h
+++ b/include/block/block_int.h
@@ -XXX,XX +XXX,XX @@ struct BdrvChild {
  * copied as well.
  */
 struct BlockDriverState {
-    int64_t total_sectors; /* if we are reading a disk image, give its
-                              size in sectors */
+    /* Protected by big QEMU lock or read-only after opening.  No special
+     * locking needed during I/O...
+     */
     int open_flags; /* flags used to open the file, re-used for re-open */
     bool read_only; /* if true, the media is read only */
     bool encrypted; /* if true, the media is encrypted */
@@ -XXX,XX +XXX,XX @@ struct BlockDriverState {
     bool sg;        /* if true, the device is a /dev/sg* */
     bool probed;    /* if true, format was probed rather than specified */
 
-    int copy_on_read; /* if nonzero, copy read backing sectors into image.
-                         note this is a reference count */
-
-    CoQueue flush_queue;            /* Serializing flush queue */
-    bool active_flush_req;          /* Flush request in flight? */
-    unsigned int write_gen;         /* Current data generation */
-    unsigned int flushed_gen;       /* Flushed write generation */
-
     BlockDriver *drv; /* NULL means no media */
     void *opaque;
 
@@ -XXX,XX +XXX,XX @@ struct BlockDriverState {
     BdrvChild *backing;
     BdrvChild *file;
 
-    /* Callback before write request is processed */
-    NotifierWithReturnList before_write_notifiers;
-
-    /* number of in-flight requests; overall and serialising */
-    unsigned int in_flight;
-    unsigned int serialising_in_flight;
-
-    bool wakeup;
-
-    /* Offset after the highest byte written to */
-    uint64_t wr_highest_offset;
-
     /* I/O Limits */
     BlockLimits bl;
 
@@ -XXX,XX +XXX,XX @@ struct BlockDriverState {
     QTAILQ_ENTRY(BlockDriverState) bs_list;
     /* element of the list of monitor-owned BDS */
     QTAILQ_ENTRY(BlockDriverState) monitor_list;
-    QLIST_HEAD(, BdrvDirtyBitmap) dirty_bitmaps;
     int refcnt;
 
-    QLIST_HEAD(, BdrvTrackedRequest) tracked_requests;
-
     /* operation blockers */
     QLIST_HEAD(, BdrvOpBlocker) op_blockers[BLOCK_OP_TYPE_MAX];
 
@@ -XXX,XX +XXX,XX @@ struct BlockDriverState {
     /* The error object in use for blocking operations on backing_hd */
     Error *backing_blocker;
 
+    /* Protected by AioContext lock */
+
+    /* If true, copy read backing sectors into image.  Can be >1 if more
+     * than one client has requested copy-on-read.
+     */
+    int copy_on_read;
+
+    /* If we are reading a disk image, give its size in sectors.
+     * Generally read-only; it is written to by load_vmstate and save_vmstate,
+     * but the block layer is quiescent during those.
+     */
+    int64_t total_sectors;
+
+    /* Callback before write request is processed */
+    NotifierWithReturnList before_write_notifiers;
+
+    /* number of in-flight requests; overall and serialising */
+    unsigned int in_flight;
+    unsigned int serialising_in_flight;
+
+    bool wakeup;
+
+    /* Offset after the highest byte written to */
+    uint64_t wr_highest_offset;
+
     /* threshold limit for writes, in bytes. "High water mark". */
     uint64_t write_threshold_offset;
     NotifierWithReturn write_threshold_notifier;
@@ -XXX,XX +XXX,XX @@ struct BlockDriverState {
     /* counter for nested bdrv_io_plug */
     unsigned io_plugged;
 
+    QLIST_HEAD(, BdrvTrackedRequest) tracked_requests;
+    CoQueue flush_queue;                  /* Serializing flush queue */
+    bool active_flush_req;                /* Flush request in flight? */
+    unsigned int write_gen;               /* Current data generation */
+    unsigned int flushed_gen;             /* Flushed write generation */
+
+    QLIST_HEAD(, BdrvDirtyBitmap) dirty_bitmaps;
+
+    /* do we need to tell the quest if we have a volatile write cache? */
+    int enable_write_cache;
+
     int quiesce_counter;
 };
 
diff --git a/include/sysemu/block-backend.h b/include/sysemu/block-backend.h
index XXXXXXX..XXXXXXX 100644
--- a/include/sysemu/block-backend.h
+++ b/include/sysemu/block-backend.h
@@ -XXX,XX +XXX,XX @@ typedef struct BlockDevOps {
  * fields that must be public. This is in particular for QLIST_ENTRY() and
  * friends so that BlockBackends can be kept in lists outside block-backend.c */
 typedef struct BlockBackendPublic {
-    /* I/O throttling.
-     * throttle_state tells us if this BlockBackend has I/O limits configured.
-     * io_limits_disabled tells us if they are currently being enforced */
+    /* I/O throttling has its own locking, but also some fields are
+     * protected by the AioContext lock.
+     */
+
+    /* Protected by AioContext lock.  */
     CoQueue      throttled_reqs[2];
+
+    /* Nonzero if the I/O limits are currently being ignored; generally
+     * it is zero.  */
     unsigned int io_limits_disabled;
 
     /* The following fields are protected by the ThrottleGroup lock.
-     * See the ThrottleGroup documentation for details. */
+     * See the ThrottleGroup documentation for details.
+     * throttle_state tells us if I/O limits are configured. */
     ThrottleState *throttle_state;
     ThrottleTimers throttle_timers;
     unsigned       pending_reqs[2];
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

This uses the lock-free mutex described in the paper '"Blocking without
Locking", or LFTHREADS: A lock-free thread library' by Gidenstam and
Papatriantafilou.  The same technique is used in OSv, and in fact
the code is essentially a conversion to C of OSv's code.

[Added missing coroutine_fn in tests/test-aio-multithread.c.
--Stefan]

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 20170213181244.16297-2-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 include/qemu/coroutine.h     |  17 ++++-
 tests/test-aio-multithread.c |  86 ++++++++++++++++++++++++
 util/qemu-coroutine-lock.c   | 155 ++++++++++++++++++++++++++++++++++++++++---
 util/trace-events            |   1 +
 4 files changed, 246 insertions(+), 13 deletions(-)

diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
index XXXXXXX..XXXXXXX 100644
--- a/include/qemu/coroutine.h
+++ b/include/qemu/coroutine.h
@@ -XXX,XX +XXX,XX @@ bool qemu_co_queue_empty(CoQueue *queue);
 /**
  * Provides a mutex that can be used to synchronise coroutines
  */
+struct CoWaitRecord;
 typedef struct CoMutex {
-    bool locked;
+    /* Count of pending lockers; 0 for a free mutex, 1 for an
+     * uncontended mutex.
+     */
+    unsigned locked;
+
+    /* A queue of waiters.  Elements are added atomically in front of
+     * from_push.  to_pop is only populated, and popped from, by whoever
+     * is in charge of the next wakeup.  This can be an unlocker or,
+     * through the handoff protocol, a locker that is about to go to sleep.
+     */
+    QSLIST_HEAD(, CoWaitRecord) from_push, to_pop;
+
+    unsigned handoff, sequence;
+
     Coroutine *holder;
-    CoQueue queue;
 } CoMutex;
 
 /**
diff --git a/tests/test-aio-multithread.c b/tests/test-aio-multithread.c
index XXXXXXX..XXXXXXX 100644
--- a/tests/test-aio-multithread.c
+++ b/tests/test-aio-multithread.c
@@ -XXX,XX +XXX,XX @@ static void test_multi_co_schedule_10(void)
     test_multi_co_schedule(10);
 }
 
+/* CoMutex thread-safety.  */
+
+static uint32_t atomic_counter;
+static uint32_t running;
+static uint32_t counter;
+static CoMutex comutex;
+
+static void coroutine_fn test_multi_co_mutex_entry(void *opaque)
+{
+    while (!atomic_mb_read(&now_stopping)) {
+        qemu_co_mutex_lock(&comutex);
+        counter++;
+        qemu_co_mutex_unlock(&comutex);
+
+        /* Increase atomic_counter *after* releasing the mutex.  Otherwise
+         * there is a chance (it happens about 1 in 3 runs) that the iothread
+         * exits before the coroutine is woken up, causing a spurious
+         * assertion failure.
+         */
+        atomic_inc(&atomic_counter);
+    }
+    atomic_dec(&running);
+}
+
+static void test_multi_co_mutex(int threads, int seconds)
+{
+    int i;
+
+    qemu_co_mutex_init(&comutex);
+    counter = 0;
+    atomic_counter = 0;
+    now_stopping = false;
+
+    create_aio_contexts();
+    assert(threads <= NUM_CONTEXTS);
+    running = threads;
+    for (i = 0; i < threads; i++) {
+        Coroutine *co1 = qemu_coroutine_create(test_multi_co_mutex_entry, NULL);
+        aio_co_schedule(ctx[i], co1);
+    }
+
+    g_usleep(seconds * 1000000);
+
+    atomic_mb_set(&now_stopping, true);
+    while (running > 0) {
+        g_usleep(100000);
+    }
+
+    join_aio_contexts();
+    g_test_message("%d iterations/second\n", counter / seconds);
+    g_assert_cmpint(counter, ==, atomic_counter);
+}
+
+/* Testing with NUM_CONTEXTS threads focuses on the queue.  The mutex however
+ * is too contended (and the threads spend too much time in aio_poll)
+ * to actually stress the handoff protocol.
+ */
+static void test_multi_co_mutex_1(void)
+{
+    test_multi_co_mutex(NUM_CONTEXTS, 1);
+}
+
+static void test_multi_co_mutex_10(void)
+{
+    test_multi_co_mutex(NUM_CONTEXTS, 10);
+}
+
+/* Testing with fewer threads stresses the handoff protocol too.  Still, the
+ * case where the locker _can_ pick up a handoff is very rare, happening
+ * about 10 times in 1 million, so increase the runtime a bit compared to
+ * other "quick" testcases that only run for 1 second.
+ */
+static void test_multi_co_mutex_2_3(void)
+{
+    test_multi_co_mutex(2, 3);
+}
+
+static void test_multi_co_mutex_2_30(void)
+{
+    test_multi_co_mutex(2, 30);
+}
+
 /* End of tests.  */
 
 int main(int argc, char **argv)
@@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv)
     g_test_add_func("/aio/multi/lifecycle", test_lifecycle);
     if (g_test_quick()) {
         g_test_add_func("/aio/multi/schedule", test_multi_co_schedule_1);
+        g_test_add_func("/aio/multi/mutex/contended", test_multi_co_mutex_1);
+        g_test_add_func("/aio/multi/mutex/handoff", test_multi_co_mutex_2_3);
     } else {
         g_test_add_func("/aio/multi/schedule", test_multi_co_schedule_10);
+        g_test_add_func("/aio/multi/mutex/contended", test_multi_co_mutex_10);
+        g_test_add_func("/aio/multi/mutex/handoff", test_multi_co_mutex_2_30);
     }
     return g_test_run();
 }
diff --git a/util/qemu-coroutine-lock.c b/util/qemu-coroutine-lock.c
index XXXXXXX..XXXXXXX 100644
--- a/util/qemu-coroutine-lock.c
+++ b/util/qemu-coroutine-lock.c
@@ -XXX,XX +XXX,XX @@
  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
  * THE SOFTWARE.
+ *
+ * The lock-free mutex implementation is based on OSv
+ * (core/lfmutex.cc, include/lockfree/mutex.hh).
+ * Copyright (C) 2013 Cloudius Systems, Ltd.
  */
 
 #include "qemu/osdep.h"
@@ -XXX,XX +XXX,XX @@ bool qemu_co_queue_empty(CoQueue *queue)
     return QSIMPLEQ_FIRST(&queue->entries) == NULL;
 }
 
+/* The wait records are handled with a multiple-producer, single-consumer
+ * lock-free queue.  There cannot be two concurrent pop_waiter() calls
+ * because pop_waiter() can only be called while mutex->handoff is zero.
+ * This can happen in three cases:
+ * - in qemu_co_mutex_unlock, before the hand-off protocol has started.
+ *   In this case, qemu_co_mutex_lock will see mutex->handoff == 0 and
+ *   not take part in the handoff.
+ * - in qemu_co_mutex_lock, if it steals the hand-off responsibility from
+ *   qemu_co_mutex_unlock.  In this case, qemu_co_mutex_unlock will fail
+ *   the cmpxchg (it will see either 0 or the next sequence value) and
+ *   exit.  The next hand-off cannot begin until qemu_co_mutex_lock has
+ *   woken up someone.
+ * - in qemu_co_mutex_unlock, if it takes the hand-off token itself.
+ *   In this case another iteration starts with mutex->handoff == 0;
+ *   a concurrent qemu_co_mutex_lock will fail the cmpxchg, and
+ *   qemu_co_mutex_unlock will go back to case (1).
+ *
+ * The following functions manage this queue.
+ */
+typedef struct CoWaitRecord {
+    Coroutine *co;
+    QSLIST_ENTRY(CoWaitRecord) next;
+} CoWaitRecord;
+
+static void push_waiter(CoMutex *mutex, CoWaitRecord *w)
+{
+    w->co = qemu_coroutine_self();
+    QSLIST_INSERT_HEAD_ATOMIC(&mutex->from_push, w, next);
+}
+
+static void move_waiters(CoMutex *mutex)
+{
+    QSLIST_HEAD(, CoWaitRecord) reversed;
+    QSLIST_MOVE_ATOMIC(&reversed, &mutex->from_push);
+    while (!QSLIST_EMPTY(&reversed)) {
+        CoWaitRecord *w = QSLIST_FIRST(&reversed);
+        QSLIST_REMOVE_HEAD(&reversed, next);
+        QSLIST_INSERT_HEAD(&mutex->to_pop, w, next);
+    }
+}
+
+static CoWaitRecord *pop_waiter(CoMutex *mutex)
+{
+    CoWaitRecord *w;
+
+    if (QSLIST_EMPTY(&mutex->to_pop)) {
+        move_waiters(mutex);
+        if (QSLIST_EMPTY(&mutex->to_pop)) {
+            return NULL;
+        }
+    }
+    w = QSLIST_FIRST(&mutex->to_pop);
+    QSLIST_REMOVE_HEAD(&mutex->to_pop, next);
+    return w;
+}
+
+static bool has_waiters(CoMutex *mutex)
+{
+    return QSLIST_EMPTY(&mutex->to_pop) || QSLIST_EMPTY(&mutex->from_push);
+}
+
 void qemu_co_mutex_init(CoMutex *mutex)
 {
     memset(mutex, 0, sizeof(*mutex));
-    qemu_co_queue_init(&mutex->queue);
 }
 
-void coroutine_fn qemu_co_mutex_lock(CoMutex *mutex)
+static void coroutine_fn qemu_co_mutex_lock_slowpath(CoMutex *mutex)
 {
     Coroutine *self = qemu_coroutine_self();
+    CoWaitRecord w;
+    unsigned old_handoff;
 
     trace_qemu_co_mutex_lock_entry(mutex, self);
+    w.co = self;
+    push_waiter(mutex, &w);
 
-    while (mutex->locked) {
-        qemu_co_queue_wait(&mutex->queue);
+    /* This is the "Responsibility Hand-Off" protocol; a lock() picks from
+     * a concurrent unlock() the responsibility of waking somebody up.
+     */
+    old_handoff = atomic_mb_read(&mutex->handoff);
+    if (old_handoff &&
+        has_waiters(mutex) &&
+        atomic_cmpxchg(&mutex->handoff, old_handoff, 0) == old_handoff) {
+        /* There can be no concurrent pops, because there can be only
+         * one active handoff at a time.
+         */
+        CoWaitRecord *to_wake = pop_waiter(mutex);
+        Coroutine *co = to_wake->co;
+        if (co == self) {
+            /* We got the lock ourselves!  */
+            assert(to_wake == &w);
+            return;
+        }
+
+        aio_co_wake(co);
     }
 
-    mutex->locked = true;
-    mutex->holder = self;
-    self->locks_held++;
-
+    qemu_coroutine_yield();
     trace_qemu_co_mutex_lock_return(mutex, self);
 }
 
+void coroutine_fn qemu_co_mutex_lock(CoMutex *mutex)
+{
+    Coroutine *self = qemu_coroutine_self();
+
+    if (atomic_fetch_inc(&mutex->locked) == 0) {
+        /* Uncontended.  */
+        trace_qemu_co_mutex_lock_uncontended(mutex, self);
+    } else {
+        qemu_co_mutex_lock_slowpath(mutex);
+    }
+    mutex->holder = self;
+    self->locks_held++;
+}
+
 void coroutine_fn qemu_co_mutex_unlock(CoMutex *mutex)
 {
     Coroutine *self = qemu_coroutine_self();
 
     trace_qemu_co_mutex_unlock_entry(mutex, self);
 
-    assert(mutex->locked == true);
+    assert(mutex->locked);
     assert(mutex->holder == self);
     assert(qemu_in_coroutine());
 
-    mutex->locked = false;
     mutex->holder = NULL;
     self->locks_held--;
-    qemu_co_queue_next(&mutex->queue);
+    if (atomic_fetch_dec(&mutex->locked) == 1) {
+        /* No waiting qemu_co_mutex_lock().  Pfew, that was easy!  */
+        return;
+    }
+
+    for (;;) {
+        CoWaitRecord *to_wake = pop_waiter(mutex);
+        unsigned our_handoff;
+
+        if (to_wake) {
+            Coroutine *co = to_wake->co;
+            aio_co_wake(co);
+            break;
+        }
+
+        /* Some concurrent lock() is in progress (we know this because
+         * mutex->locked was >1) but it hasn't yet put itself on the wait
+         * queue.  Pick a sequence number for the handoff protocol (not 0).
+         */
+        if (++mutex->sequence == 0) {
+            mutex->sequence = 1;
+        }
+
+        our_handoff = mutex->sequence;
+        atomic_mb_set(&mutex->handoff, our_handoff);
+        if (!has_waiters(mutex)) {
+            /* The concurrent lock has not added itself yet, so it
+             * will be able to pick our handoff.
+             */
+            break;
+        }
+
+        /* Try to do the handoff protocol ourselves; if somebody else has
+         * already taken it, however, we're done and they're responsible.
+         */
+        if (atomic_cmpxchg(&mutex->handoff, our_handoff, 0) != our_handoff) {
+            break;
+        }
+    }
 
     trace_qemu_co_mutex_unlock_return(mutex, self);
 }
diff --git a/util/trace-events b/util/trace-events
index XXXXXXX..XXXXXXX 100644
--- a/util/trace-events
+++ b/util/trace-events
@@ -XXX,XX +XXX,XX @@ qemu_coroutine_terminate(void *co) "self %p"
 
 # util/qemu-coroutine-lock.c
 qemu_co_queue_run_restart(void *co) "co %p"
+qemu_co_mutex_lock_uncontended(void *mutex, void *self) "mutex %p self %p"
 qemu_co_mutex_lock_entry(void *mutex, void *self) "mutex %p self %p"
 qemu_co_mutex_lock_return(void *mutex, void *self) "mutex %p self %p"
 qemu_co_mutex_unlock_entry(void *mutex, void *self) "mutex %p self %p"
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

Running a very small critical section on pthread_mutex_t and CoMutex
shows that pthread_mutex_t is much faster because it doesn't actually
go to sleep.  What happens is that the critical section is shorter
than the latency of entering the kernel and thus FUTEX_WAIT always
fails.  With CoMutex there is no such latency but you still want to
avoid wait and wakeup.  So introduce it artificially.

This only works with one waiters; because CoMutex is fair, it will
always have more waits and wakeups than a pthread_mutex_t.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 20170213181244.16297-3-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 include/qemu/coroutine.h   |  5 +++++
 util/qemu-coroutine-lock.c | 51 ++++++++++++++++++++++++++++++++++++++++------
 util/qemu-coroutine.c      |  2 +-
 3 files changed, 51 insertions(+), 7 deletions(-)

diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
index XXXXXXX..XXXXXXX 100644
--- a/include/qemu/coroutine.h
+++ b/include/qemu/coroutine.h
@@ -XXX,XX +XXX,XX @@ typedef struct CoMutex {
      */
     unsigned locked;
 
+    /* Context that is holding the lock.  Useful to avoid spinning
+     * when two coroutines on the same AioContext try to get the lock. :)
+     */
+    AioContext *ctx;
+
     /* A queue of waiters.  Elements are added atomically in front of
      * from_push.  to_pop is only populated, and popped from, by whoever
      * is in charge of the next wakeup.  This can be an unlocker or,
diff --git a/util/qemu-coroutine-lock.c b/util/qemu-coroutine-lock.c
index XXXXXXX..XXXXXXX 100644
--- a/util/qemu-coroutine-lock.c
+++ b/util/qemu-coroutine-lock.c
@@ -XXX,XX +XXX,XX @@
 #include "qemu-common.h"
 #include "qemu/coroutine.h"
 #include "qemu/coroutine_int.h"
+#include "qemu/processor.h"
 #include "qemu/queue.h"
 #include "block/aio.h"
 #include "trace.h"
@@ -XXX,XX +XXX,XX @@ void qemu_co_mutex_init(CoMutex *mutex)
     memset(mutex, 0, sizeof(*mutex));
 }
 
-static void coroutine_fn qemu_co_mutex_lock_slowpath(CoMutex *mutex)
+static void coroutine_fn qemu_co_mutex_wake(CoMutex *mutex, Coroutine *co)
+{
+    /* Read co before co->ctx; pairs with smp_wmb() in
+     * qemu_coroutine_enter().
+     */
+    smp_read_barrier_depends();
+    mutex->ctx = co->ctx;
+    aio_co_wake(co);
+}
+
+static void coroutine_fn qemu_co_mutex_lock_slowpath(AioContext *ctx,
+                                                     CoMutex *mutex)
 {
     Coroutine *self = qemu_coroutine_self();
     CoWaitRecord w;
@@ -XXX,XX +XXX,XX @@ static void coroutine_fn qemu_co_mutex_lock_slowpath(CoMutex *mutex)
         if (co == self) {
             /* We got the lock ourselves!  */
             assert(to_wake == &w);
+            mutex->ctx = ctx;
             return;
         }
 
-        aio_co_wake(co);
+        qemu_co_mutex_wake(mutex, co);
     }
 
     qemu_coroutine_yield();
@@ -XXX,XX +XXX,XX @@ static void coroutine_fn qemu_co_mutex_lock_slowpath(CoMutex *mutex)
 
 void coroutine_fn qemu_co_mutex_lock(CoMutex *mutex)
 {
+    AioContext *ctx = qemu_get_current_aio_context();
     Coroutine *self = qemu_coroutine_self();
+    int waiters, i;
 
-    if (atomic_fetch_inc(&mutex->locked) == 0) {
+    /* Running a very small critical section on pthread_mutex_t and CoMutex
+     * shows that pthread_mutex_t is much faster because it doesn't actually
+     * go to sleep.  What happens is that the critical section is shorter
+     * than the latency of entering the kernel and thus FUTEX_WAIT always
+     * fails.  With CoMutex there is no such latency but you still want to
+     * avoid wait and wakeup.  So introduce it artificially.
+     */
+    i = 0;
+retry_fast_path:
+    waiters = atomic_cmpxchg(&mutex->locked, 0, 1);
+    if (waiters != 0) {
+        while (waiters == 1 && ++i < 1000) {
+            if (atomic_read(&mutex->ctx) == ctx) {
+                break;
+            }
+            if (atomic_read(&mutex->locked) == 0) {
+                goto retry_fast_path;
+            }
+            cpu_relax();
+        }
+        waiters = atomic_fetch_inc(&mutex->locked);
+    }
+
+    if (waiters == 0) {
         /* Uncontended.  */
         trace_qemu_co_mutex_lock_uncontended(mutex, self);
+        mutex->ctx = ctx;
     } else {
-        qemu_co_mutex_lock_slowpath(mutex);
+        qemu_co_mutex_lock_slowpath(ctx, mutex);
     }
     mutex->holder = self;
     self->locks_held++;
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_mutex_unlock(CoMutex *mutex)
     assert(mutex->holder == self);
     assert(qemu_in_coroutine());
 
+    mutex->ctx = NULL;
     mutex->holder = NULL;
     self->locks_held--;
     if (atomic_fetch_dec(&mutex->locked) == 1) {
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_mutex_unlock(CoMutex *mutex)
         unsigned our_handoff;
 
         if (to_wake) {
-            Coroutine *co = to_wake->co;
-            aio_co_wake(co);
+            qemu_co_mutex_wake(mutex, to_wake->co);
             break;
         }
 
diff --git a/util/qemu-coroutine.c b/util/qemu-coroutine.c
index XXXXXXX..XXXXXXX 100644
--- a/util/qemu-coroutine.c
+++ b/util/qemu-coroutine.c
@@ -XXX,XX +XXX,XX @@ void qemu_coroutine_enter(Coroutine *co)
     co->ctx = qemu_get_current_aio_context();
 
     /* Store co->ctx before anything that stores co.  Matches
-     * barrier in aio_co_wake.
+     * barrier in aio_co_wake and qemu_co_mutex_wake.
      */
     smp_wmb();
 
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

Add two implementations of the same benchmark as the previous patch,
but using pthreads.  One uses a normal QemuMutex, the other is Linux
only and implements a fair mutex based on MCS locks and futexes.
This shows that the slower performance of the 5-thread case is due to
the fairness of CoMutex, rather than to coroutines.  If fairness does
not matter, as is the case with two threads, CoMutex can actually be
faster than pthreads.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 20170213181244.16297-4-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 tests/test-aio-multithread.c | 164 +++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 164 insertions(+)

diff --git a/tests/test-aio-multithread.c b/tests/test-aio-multithread.c
index XXXXXXX..XXXXXXX 100644
--- a/tests/test-aio-multithread.c
+++ b/tests/test-aio-multithread.c
@@ -XXX,XX +XXX,XX @@ static void test_multi_co_mutex_2_30(void)
     test_multi_co_mutex(2, 30);
 }
 
+/* Same test with fair mutexes, for performance comparison.  */
+
+#ifdef CONFIG_LINUX
+#include "qemu/futex.h"
+
+/* The nodes for the mutex reside in this structure (on which we try to avoid
+ * false sharing).  The head of the mutex is in the "mutex_head" variable.
+ */
+static struct {
+    int next, locked;
+    int padding[14];
+} nodes[NUM_CONTEXTS] __attribute__((__aligned__(64)));
+
+static int mutex_head = -1;
+
+static void mcs_mutex_lock(void)
+{
+    int prev;
+
+    nodes[id].next = -1;
+    nodes[id].locked = 1;
+    prev = atomic_xchg(&mutex_head, id);
+    if (prev != -1) {
+        atomic_set(&nodes[prev].next, id);
+        qemu_futex_wait(&nodes[id].locked, 1);
+    }
+}
+
+static void mcs_mutex_unlock(void)
+{
+    int next;
+    if (nodes[id].next == -1) {
+        if (atomic_read(&mutex_head) == id &&
+            atomic_cmpxchg(&mutex_head, id, -1) == id) {
+            /* Last item in the list, exit.  */
+            return;
+        }
+        while (atomic_read(&nodes[id].next) == -1) {
+            /* mcs_mutex_lock did the xchg, but has not updated
+             * nodes[prev].next yet.
+             */
+        }
+    }
+
+    /* Wake up the next in line.  */
+    next = nodes[id].next;
+    nodes[next].locked = 0;
+    qemu_futex_wake(&nodes[next].locked, 1);
+}
+
+static void test_multi_fair_mutex_entry(void *opaque)
+{
+    while (!atomic_mb_read(&now_stopping)) {
+        mcs_mutex_lock();
+        counter++;
+        mcs_mutex_unlock();
+        atomic_inc(&atomic_counter);
+    }
+    atomic_dec(&running);
+}
+
+static void test_multi_fair_mutex(int threads, int seconds)
+{
+    int i;
+
+    assert(mutex_head == -1);
+    counter = 0;
+    atomic_counter = 0;
+    now_stopping = false;
+
+    create_aio_contexts();
+    assert(threads <= NUM_CONTEXTS);
+    running = threads;
+    for (i = 0; i < threads; i++) {
+        Coroutine *co1 = qemu_coroutine_create(test_multi_fair_mutex_entry, NULL);
+        aio_co_schedule(ctx[i], co1);
+    }
+
+    g_usleep(seconds * 1000000);
+
+    atomic_mb_set(&now_stopping, true);
+    while (running > 0) {
+        g_usleep(100000);
+    }
+
+    join_aio_contexts();
+    g_test_message("%d iterations/second\n", counter / seconds);
+    g_assert_cmpint(counter, ==, atomic_counter);
+}
+
+static void test_multi_fair_mutex_1(void)
+{
+    test_multi_fair_mutex(NUM_CONTEXTS, 1);
+}
+
+static void test_multi_fair_mutex_10(void)
+{
+    test_multi_fair_mutex(NUM_CONTEXTS, 10);
+}
+#endif
+
+/* Same test with pthread mutexes, for performance comparison and
+ * portability.  */
+
+static QemuMutex mutex;
+
+static void test_multi_mutex_entry(void *opaque)
+{
+    while (!atomic_mb_read(&now_stopping)) {
+        qemu_mutex_lock(&mutex);
+        counter++;
+        qemu_mutex_unlock(&mutex);
+        atomic_inc(&atomic_counter);
+    }
+    atomic_dec(&running);
+}
+
+static void test_multi_mutex(int threads, int seconds)
+{
+    int i;
+
+    qemu_mutex_init(&mutex);
+    counter = 0;
+    atomic_counter = 0;
+    now_stopping = false;
+
+    create_aio_contexts();
+    assert(threads <= NUM_CONTEXTS);
+    running = threads;
+    for (i = 0; i < threads; i++) {
+        Coroutine *co1 = qemu_coroutine_create(test_multi_mutex_entry, NULL);
+        aio_co_schedule(ctx[i], co1);
+    }
+
+    g_usleep(seconds * 1000000);
+
+    atomic_mb_set(&now_stopping, true);
+    while (running > 0) {
+        g_usleep(100000);
+    }
+
+    join_aio_contexts();
+    g_test_message("%d iterations/second\n", counter / seconds);
+    g_assert_cmpint(counter, ==, atomic_counter);
+}
+
+static void test_multi_mutex_1(void)
+{
+    test_multi_mutex(NUM_CONTEXTS, 1);
+}
+
+static void test_multi_mutex_10(void)
+{
+    test_multi_mutex(NUM_CONTEXTS, 10);
+}
+
 /* End of tests.  */
 
 int main(int argc, char **argv)
@@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv)
         g_test_add_func("/aio/multi/schedule", test_multi_co_schedule_1);
         g_test_add_func("/aio/multi/mutex/contended", test_multi_co_mutex_1);
         g_test_add_func("/aio/multi/mutex/handoff", test_multi_co_mutex_2_3);
+#ifdef CONFIG_LINUX
+        g_test_add_func("/aio/multi/mutex/mcs", test_multi_fair_mutex_1);
+#endif
+        g_test_add_func("/aio/multi/mutex/pthread", test_multi_mutex_1);
     } else {
         g_test_add_func("/aio/multi/schedule", test_multi_co_schedule_10);
         g_test_add_func("/aio/multi/mutex/contended", test_multi_co_mutex_10);
         g_test_add_func("/aio/multi/mutex/handoff", test_multi_co_mutex_2_30);
+#ifdef CONFIG_LINUX
+        g_test_add_func("/aio/multi/mutex/mcs", test_multi_fair_mutex_10);
+#endif
+        g_test_add_func("/aio/multi/mutex/pthread", test_multi_mutex_10);
     }
     return g_test_run();
 }
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

This will avoid forward references in the next patch.  It is also
more logical because CoQueue is not anymore the basic primitive.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 20170213181244.16297-5-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 include/qemu/coroutine.h | 89 ++++++++++++++++++++++++------------------------
 1 file changed, 44 insertions(+), 45 deletions(-)

diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
index XXXXXXX..XXXXXXX 100644
--- a/include/qemu/coroutine.h
+++ b/include/qemu/coroutine.h
@@ -XXX,XX +XXX,XX @@ bool qemu_in_coroutine(void);
  */
 bool qemu_coroutine_entered(Coroutine *co);
 
-
-/**
- * CoQueues are a mechanism to queue coroutines in order to continue executing
- * them later. They provide the fundamental primitives on which coroutine locks
- * are built.
- */
-typedef struct CoQueue {
-    QSIMPLEQ_HEAD(, Coroutine) entries;
-} CoQueue;
-
-/**
- * Initialise a CoQueue. This must be called before any other operation is used
- * on the CoQueue.
- */
-void qemu_co_queue_init(CoQueue *queue);
-
-/**
- * Adds the current coroutine to the CoQueue and transfers control to the
- * caller of the coroutine.
- */
-void coroutine_fn qemu_co_queue_wait(CoQueue *queue);
-
-/**
- * Restarts the next coroutine in the CoQueue and removes it from the queue.
- *
- * Returns true if a coroutine was restarted, false if the queue is empty.
- */
-bool coroutine_fn qemu_co_queue_next(CoQueue *queue);
-
-/**
- * Restarts all coroutines in the CoQueue and leaves the queue empty.
- */
-void coroutine_fn qemu_co_queue_restart_all(CoQueue *queue);
-
-/**
- * Enter the next coroutine in the queue
- */
-bool qemu_co_enter_next(CoQueue *queue);
-
-/**
- * Checks if the CoQueue is empty.
- */
-bool qemu_co_queue_empty(CoQueue *queue);
-
-
 /**
  * Provides a mutex that can be used to synchronise coroutines
  */
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_mutex_lock(CoMutex *mutex);
  */
 void coroutine_fn qemu_co_mutex_unlock(CoMutex *mutex);
 
+
+/**
+ * CoQueues are a mechanism to queue coroutines in order to continue executing
+ * them later.
+ */
+typedef struct CoQueue {
+    QSIMPLEQ_HEAD(, Coroutine) entries;
+} CoQueue;
+
+/**
+ * Initialise a CoQueue. This must be called before any other operation is used
+ * on the CoQueue.
+ */
+void qemu_co_queue_init(CoQueue *queue);
+
+/**
+ * Adds the current coroutine to the CoQueue and transfers control to the
+ * caller of the coroutine.
+ */
+void coroutine_fn qemu_co_queue_wait(CoQueue *queue);
+
+/**
+ * Restarts the next coroutine in the CoQueue and removes it from the queue.
+ *
+ * Returns true if a coroutine was restarted, false if the queue is empty.
+ */
+bool coroutine_fn qemu_co_queue_next(CoQueue *queue);
+
+/**
+ * Restarts all coroutines in the CoQueue and leaves the queue empty.
+ */
+void coroutine_fn qemu_co_queue_restart_all(CoQueue *queue);
+
+/**
+ * Enter the next coroutine in the queue
+ */
+bool qemu_co_enter_next(CoQueue *queue);
+
+/**
+ * Checks if the CoQueue is empty.
+ */
+bool qemu_co_queue_empty(CoQueue *queue);
+
+
 typedef struct CoRwlock {
     bool writer;
     int reader;
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

All that CoQueue needs in order to become thread-safe is help
from an external mutex.  Add this to the API.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 20170213181244.16297-6-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 include/qemu/coroutine.h   |  8 +++++---
 block/backup.c             |  2 +-
 block/io.c                 |  4 ++--
 block/nbd-client.c         |  2 +-
 block/qcow2-cluster.c      |  4 +---
 block/sheepdog.c           |  2 +-
 block/throttle-groups.c    |  2 +-
 hw/9pfs/9p.c               |  2 +-
 util/qemu-coroutine-lock.c | 24 +++++++++++++++++++++---
 9 files changed, 34 insertions(+), 16 deletions(-)

diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
index XXXXXXX..XXXXXXX 100644
--- a/include/qemu/coroutine.h
+++ b/include/qemu/coroutine.h
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_mutex_unlock(CoMutex *mutex);
 
 /**
  * CoQueues are a mechanism to queue coroutines in order to continue executing
- * them later.
+ * them later.  They are similar to condition variables, but they need help
+ * from an external mutex in order to maintain thread-safety.
  */
 typedef struct CoQueue {
     QSIMPLEQ_HEAD(, Coroutine) entries;
@@ -XXX,XX +XXX,XX @@ void qemu_co_queue_init(CoQueue *queue);
 
 /**
  * Adds the current coroutine to the CoQueue and transfers control to the
- * caller of the coroutine.
+ * caller of the coroutine.  The mutex is unlocked during the wait and
+ * locked again afterwards.
  */
-void coroutine_fn qemu_co_queue_wait(CoQueue *queue);
+void coroutine_fn qemu_co_queue_wait(CoQueue *queue, CoMutex *mutex);
 
 /**
  * Restarts the next coroutine in the CoQueue and removes it from the queue.
diff --git a/block/backup.c b/block/backup.c
index XXXXXXX..XXXXXXX 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -XXX,XX +XXX,XX @@ static void coroutine_fn wait_for_overlapping_requests(BackupBlockJob *job,
         retry = false;
         QLIST_FOREACH(req, &job->inflight_reqs, list) {
             if (end > req->start && start < req->end) {
-                qemu_co_queue_wait(&req->wait_queue);
+                qemu_co_queue_wait(&req->wait_queue, NULL);
                 retry = true;
                 break;
             }
diff --git a/block/io.c b/block/io.c
index XXXXXXX..XXXXXXX 100644
--- a/block/io.c
+++ b/block/io.c
@@ -XXX,XX +XXX,XX @@ static bool coroutine_fn wait_serialising_requests(BdrvTrackedRequest *self)
                  * (instead of producing a deadlock in the former case). */
                 if (!req->waiting_for) {
                     self->waiting_for = req;
-                    qemu_co_queue_wait(&req->wait_queue);
+                    qemu_co_queue_wait(&req->wait_queue, NULL);
                     self->waiting_for = NULL;
                     retry = true;
                     waited = true;
@@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_flush(BlockDriverState *bs)
 
     /* Wait until any previous flushes are completed */
     while (bs->active_flush_req) {
-        qemu_co_queue_wait(&bs->flush_queue);
+        qemu_co_queue_wait(&bs->flush_queue, NULL);
     }
 
     bs->active_flush_req = true;
diff --git a/block/nbd-client.c b/block/nbd-client.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nbd-client.c
+++ b/block/nbd-client.c
@@ -XXX,XX +XXX,XX @@ static void nbd_coroutine_start(NBDClientSession *s,
     /* Poor man semaphore.  The free_sema is locked when no other request
      * can be accepted, and unlocked after receiving one reply.  */
     if (s->in_flight == MAX_NBD_REQUESTS) {
-        qemu_co_queue_wait(&s->free_sema);
+        qemu_co_queue_wait(&s->free_sema, NULL);
         assert(s->in_flight < MAX_NBD_REQUESTS);
     }
     s->in_flight++;
diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c
index XXXXXXX..XXXXXXX 100644
--- a/block/qcow2-cluster.c
+++ b/block/qcow2-cluster.c
@@ -XXX,XX +XXX,XX @@ static int handle_dependencies(BlockDriverState *bs, uint64_t guest_offset,
             if (bytes == 0) {
                 /* Wait for the dependency to complete. We need to recheck
                  * the free/allocated clusters when we continue. */
-                qemu_co_mutex_unlock(&s->lock);
-                qemu_co_queue_wait(&old_alloc->dependent_requests);
-                qemu_co_mutex_lock(&s->lock);
+                qemu_co_queue_wait(&old_alloc->dependent_requests, &s->lock);
                 return -EAGAIN;
             }
         }
diff --git a/block/sheepdog.c b/block/sheepdog.c
index XXXXXXX..XXXXXXX 100644
--- a/block/sheepdog.c
+++ b/block/sheepdog.c
@@ -XXX,XX +XXX,XX @@ static void wait_for_overlapping_aiocb(BDRVSheepdogState *s, SheepdogAIOCB *acb)
 retry:
     QLIST_FOREACH(cb, &s->inflight_aiocb_head, aiocb_siblings) {
         if (AIOCBOverlapping(acb, cb)) {
-            qemu_co_queue_wait(&s->overlapping_queue);
+            qemu_co_queue_wait(&s->overlapping_queue, NULL);
             goto retry;
         }
     }
diff --git a/block/throttle-groups.c b/block/throttle-groups.c
index XXXXXXX..XXXXXXX 100644
--- a/block/throttle-groups.c
+++ b/block/throttle-groups.c
@@ -XXX,XX +XXX,XX @@ void coroutine_fn throttle_group_co_io_limits_intercept(BlockBackend *blk,
     if (must_wait || blkp->pending_reqs[is_write]) {
         blkp->pending_reqs[is_write]++;
         qemu_mutex_unlock(&tg->lock);
-        qemu_co_queue_wait(&blkp->throttled_reqs[is_write]);
+        qemu_co_queue_wait(&blkp->throttled_reqs[is_write], NULL);
         qemu_mutex_lock(&tg->lock);
         blkp->pending_reqs[is_write]--;
     }
diff --git a/hw/9pfs/9p.c b/hw/9pfs/9p.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/9pfs/9p.c
+++ b/hw/9pfs/9p.c
@@ -XXX,XX +XXX,XX @@ static void coroutine_fn v9fs_flush(void *opaque)
         /*
          * Wait for pdu to complete.
          */
-        qemu_co_queue_wait(&cancel_pdu->complete);
+        qemu_co_queue_wait(&cancel_pdu->complete, NULL);
         cancel_pdu->cancelled = 0;
         pdu_free(cancel_pdu);
     }
diff --git a/util/qemu-coroutine-lock.c b/util/qemu-coroutine-lock.c
index XXXXXXX..XXXXXXX 100644
--- a/util/qemu-coroutine-lock.c
+++ b/util/qemu-coroutine-lock.c
@@ -XXX,XX +XXX,XX @@ void qemu_co_queue_init(CoQueue *queue)
     QSIMPLEQ_INIT(&queue->entries);
 }
 
-void coroutine_fn qemu_co_queue_wait(CoQueue *queue)
+void coroutine_fn qemu_co_queue_wait(CoQueue *queue, CoMutex *mutex)
 {
     Coroutine *self = qemu_coroutine_self();
     QSIMPLEQ_INSERT_TAIL(&queue->entries, self, co_queue_next);
+
+    if (mutex) {
+        qemu_co_mutex_unlock(mutex);
+    }
+
+    /* There is no race condition here.  Other threads will call
+     * aio_co_schedule on our AioContext, which can reenter this
+     * coroutine but only after this yield and after the main loop
+     * has gone through the next iteration.
+     */
     qemu_coroutine_yield();
     assert(qemu_in_coroutine());
+
+    /* TODO: OSv implements wait morphing here, where the wakeup
+     * primitive automatically places the woken coroutine on the
+     * mutex's queue.  This avoids the thundering herd effect.
+     */
+    if (mutex) {
+        qemu_co_mutex_lock(mutex);
+    }
 }
 
 /**
@@ -XXX,XX +XXX,XX @@ void qemu_co_rwlock_rdlock(CoRwlock *lock)
     Coroutine *self = qemu_coroutine_self();
 
     while (lock->writer) {
-        qemu_co_queue_wait(&lock->queue);
+        qemu_co_queue_wait(&lock->queue, NULL);
     }
     lock->reader++;
     self->locks_held++;
@@ -XXX,XX +XXX,XX @@ void qemu_co_rwlock_wrlock(CoRwlock *lock)
     Coroutine *self = qemu_coroutine_self();
 
     while (lock->writer || lock->reader) {
-        qemu_co_queue_wait(&lock->queue);
+        qemu_co_queue_wait(&lock->queue, NULL);
     }
     lock->writer = true;
     self->locks_held++;
-- 
2.9.3

From: Paolo Bonzini <pbonzini@redhat.com>

This adds a CoMutex around the existing CoQueue.  Because the write-side
can just take CoMutex, the old "writer" field is not necessary anymore.
Instead of removing it altogether, count the number of pending writers
during a read-side critical section and forbid further readers from
entering.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 20170213181244.16297-7-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 include/qemu/coroutine.h   |  3 ++-
 util/qemu-coroutine-lock.c | 35 ++++++++++++++++++++++++-----------
 2 files changed, 26 insertions(+), 12 deletions(-)

diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
index XXXXXXX..XXXXXXX 100644
--- a/include/qemu/coroutine.h
+++ b/include/qemu/coroutine.h
@@ -XXX,XX +XXX,XX @@ bool qemu_co_queue_empty(CoQueue *queue);
 
 
 typedef struct CoRwlock {
-    bool writer;
+    int pending_writer;
     int reader;
+    CoMutex mutex;
     CoQueue queue;
 } CoRwlock;
 
diff --git a/util/qemu-coroutine-lock.c b/util/qemu-coroutine-lock.c
index XXXXXXX..XXXXXXX 100644
--- a/util/qemu-coroutine-lock.c
+++ b/util/qemu-coroutine-lock.c
@@ -XXX,XX +XXX,XX @@ void qemu_co_rwlock_init(CoRwlock *lock)
 {
     memset(lock, 0, sizeof(*lock));
     qemu_co_queue_init(&lock->queue);
+    qemu_co_mutex_init(&lock->mutex);
 }
 
 void qemu_co_rwlock_rdlock(CoRwlock *lock)
 {
     Coroutine *self = qemu_coroutine_self();
 
-    while (lock->writer) {
-        qemu_co_queue_wait(&lock->queue, NULL);
+    qemu_co_mutex_lock(&lock->mutex);
+    /* For fairness, wait if a writer is in line.  */
+    while (lock->pending_writer) {
+        qemu_co_queue_wait(&lock->queue, &lock->mutex);
     }
     lock->reader++;
+    qemu_co_mutex_unlock(&lock->mutex);
+
+    /* The rest of the read-side critical section is run without the mutex.  */
     self->locks_held++;
 }
 
@@ -XXX,XX +XXX,XX @@ void qemu_co_rwlock_unlock(CoRwlock *lock)
     Coroutine *self = qemu_coroutine_self();
 
     assert(qemu_in_coroutine());
-    if (lock->writer) {
-        lock->writer = false;
+    if (!lock->reader) {
+        /* The critical section started in qemu_co_rwlock_wrlock.  */
         qemu_co_queue_restart_all(&lock->queue);
     } else {
+        self->locks_held--;
+
+        qemu_co_mutex_lock(&lock->mutex);
         lock->reader--;
         assert(lock->reader >= 0);
         /* Wakeup only one waiting writer */
@@ -XXX,XX +XXX,XX @@ void qemu_co_rwlock_unlock(CoRwlock *lock)
             qemu_co_queue_next(&lock->queue);
         }
     }
-    self->locks_held--;
+    qemu_co_mutex_unlock(&lock->mutex);
 }
 
 void qemu_co_rwlock_wrlock(CoRwlock *lock)
 {
-    Coroutine *self = qemu_coroutine_self();
-
-    while (lock->writer || lock->reader) {
-        qemu_co_queue_wait(&lock->queue, NULL);
+    qemu_co_mutex_lock(&lock->mutex);
+    lock->pending_writer++;
+    while (lock->reader) {
+        qemu_co_queue_wait(&lock->queue, &lock->mutex);
     }
-    lock->writer = true;
-    self->locks_held++;
+    lock->pending_writer--;
+
+    /* The rest of the write-side critical section is run with
+     * the mutex taken, so that lock->reader remains zero.
+     * There is no need to update self->locks_held.
+     */
 }
-- 
2.9.3