Series comparison

-[Qemu-devel] [PULL 00/17] Block patches
+[Qemu-devel] [PULL v2 00/24] Block patches
-The following changes since commit 3fbd3405d2b0604ea530fc7a1828f19da1e95ff9:
+The following changes since commit 56f9e46b841c7be478ca038d8d4085d776ab4b0d:
-  Merge remote-tracking branch 'remotes/huth-gitlab/tags/pull-request-2019-08-17' into staging (2019-08-19 14:14:09 +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-2019-08-19
+  git://github.com/stefanha/qemu.git tags/block-pull-request
-for you to fetch changes up to fa27c478102a6b5d1c6b02c005607ad9404b915f:
+for you to fetch changes up to a7b91d35bab97a2d3e779d0c64c9b837b52a6cf7:
-  doc: Preallocation does not require writing zeroes (2019-08-19 17:13:26 +0200)
+  coroutine-lock: make CoRwlock thread-safe and fair (2017-02-21 11:39:40 +0000)
 ----------------------------------------------------------------
-Block patches:
+Pull request
-- preallocation=falloc/full support for LUKS
-- Various minor fixes
+v2:
  * Rebased to resolve scsi conflicts
 ----------------------------------------------------------------
-Max Reitz (16):
-  qemu-img: Fix bdrv_has_zero_init() use in convert
-  mirror: Fix bdrv_has_zero_init() use
-  block: Add bdrv_has_zero_init_truncate()
-  block: Implement .bdrv_has_zero_init_truncate()
-  block: Use bdrv_has_zero_init_truncate()
-  qcow2: Fix .bdrv_has_zero_init()
-  vdi: Fix .bdrv_has_zero_init()
-  vhdx: Fix .bdrv_has_zero_init()
-  iotests: Convert to preallocated encrypted qcow2
-  iotests: Test convert -n to pre-filled image
-  iotests: Full mirror to existing non-zero image
-  vdi: Make block_status recurse for fixed images
-  vmdk: Make block_status recurse for flat extents
-  vpc: Do not return RAW from block_status
-  iotests: Fix 141 when run with qed
-  doc: Preallocation does not require writing zeroes
-Maxim Levitsky (1):
+Paolo Bonzini (24):
-  LUKS: support preallocation
+  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
- qapi/block-core.json             | 15 +++++---
+ Makefile.objs                       |   4 -
- include/block/block.h            |  1 +
+ stubs/Makefile.objs                 |   1 +
- include/block/block_int.h        |  9 +++++
+ tests/Makefile.include              |  19 +-
- block.c                          | 21 +++++++++++
+ util/Makefile.objs                  |   6 +-
- block/crypto.c                   | 30 ++++++++++++++--
+ block/nbd-client.h                  |   2 +-
- block/file-posix.c               |  1 +
+ block/qed.h                         |   3 +
- block/file-win32.c               |  1 +
+ include/block/aio.h                 |  38 ++-
- block/gluster.c                  |  4 +++
+ include/block/block_int.h           |  64 +++--
- block/mirror.c                   | 11 ++++--
+ include/io/channel.h                |  72 +++++-
- block/nfs.c                      |  1 +
+ include/qemu/coroutine.h            |  84 ++++---
- block/parallels.c                |  2 +-
+ include/qemu/coroutine_int.h        |  11 +-
- block/qcow2.c                    | 30 +++++++++++++++-
+ include/sysemu/block-backend.h      |  14 +-
- block/qed.c                      |  1 +
+ tests/iothread.h                    |  25 ++
- block/raw-format.c               |  6 ++++
+ block/backup.c                      |   2 +-
- block/rbd.c                      |  1 +
+ block/blkdebug.c                    |   9 +-
- block/sheepdog.c                 |  1 +
+ block/blkreplay.c                   |   2 +-
- block/ssh.c                      |  1 +
+ block/block-backend.c               |  13 +-
- block/vdi.c                      | 16 +++++++--
+ block/curl.c                        |  44 +++-
- block/vhdx.c                     | 28 +++++++++++++--
+ block/gluster.c                     |   9 +-
- block/vmdk.c                     |  3 ++
+ block/io.c                          |  42 +---
- block/vpc.c                      |  2 +-
+ block/iscsi.c                       |  15 +-
- blockdev.c                       | 16 +++++++--
+ block/linux-aio.c                   |  10 +-
- qemu-img.c                       | 11 ++++--
+ block/mirror.c                      |  12 +-
- tests/test-block-iothread.c      |  2 +-
+ block/nbd-client.c                  | 119 +++++----
- docs/qemu-block-drivers.texi     |  4 +--
+ block/nfs.c                         |   9 +-
- qemu-img.texi                    |  4 +--
+ block/qcow2-cluster.c               |   4 +-
- tests/qemu-iotests/041           | 62 +++++++++++++++++++++++++++++---
+ block/qed-cluster.c                 |   2 +
- tests/qemu-iotests/041.out       |  4 +--
+ block/qed-table.c                   |  12 +-
- tests/qemu-iotests/122           | 17 +++++++++
+ block/qed.c                         |  58 +++--
- tests/qemu-iotests/122.out       |  8 +++++
+ block/sheepdog.c                    |  31 +--
- tests/qemu-iotests/141           |  9 +++--
+ block/ssh.c                         |  29 +--
- tests/qemu-iotests/141.out       |  5 ---
+ block/throttle-groups.c             |   4 +-
- tests/qemu-iotests/188           | 20 ++++++++++-
+ block/win32-aio.c                   |   9 +-
- tests/qemu-iotests/188.out       |  4 +++
+ dma-helpers.c                       |   2 +
- tests/qemu-iotests/common.filter |  5 +++
+ hw/9pfs/9p.c                        |   2 +-
-files changed, 313 insertions(+), 43 deletions(-)
+ 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%)
 --
-.21.0
+.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

-[Qemu-devel] [PULL 17/17] doc: Preallocation does not require writing zeroes
+[Qemu-devel] [PULL v2 03/24] block-backend: allow blk_prw from coroutine context
-When preallocating an encrypted qcow2 image, it just lets the protocol
+From: Paolo Bonzini <pbonzini@redhat.com>
 driver write data and then does not mark the clusters as zero.
 Therefore, reading this image will yield effectively random data.
-As such, we have not fulfilled the promise of always writing zeroes when
+qcow2_create2 calls this.  Do not run a nested event loop, as that
-preallocating an image in a while.  It seems that nobody has really
+breaks when aio_co_wake tries to queue the coroutine on the co_queue_wakeup
-cared, so change the documentation to conform to qemu's actual behavior.
+list of the currently running one.
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Message-id: 20190711132935.13070-1-mreitz@redhat.com
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-Reviewed-by: Eric Blake <eblake@redhat.com>
+Reviewed-by: Fam Zheng <famz@redhat.com>
-Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
+Message-id: 20170213135235.12274-4-pbonzini@redhat.com
-Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
+Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 Signed-off-by: Max Reitz <mreitz@redhat.com>
 ---
- qapi/block-core.json         | 9 +++++----
+ block/block-backend.c | 12 ++++++++----
- docs/qemu-block-drivers.texi | 4 ++--
+file changed, 8 insertions(+), 4 deletions(-)
  qemu-img.texi                | 4 ++--
 files changed, 9 insertions(+), 8 deletions(-)
-diff --git a/qapi/block-core.json b/qapi/block-core.json
+diff --git a/block/block-backend.c b/block/block-backend.c
 index XXXXXXX..XXXXXXX 100644
---- a/qapi/block-core.json
+--- a/block/block-backend.c
-+++ b/qapi/block-core.json
++++ b/block/block-backend.c
-@@ -XXX,XX +XXX,XX @@
+@@ -XXX,XX +XXX,XX @@ static int blk_prw(BlockBackend *blk, int64_t offset, uint8_t *buf,
- # @off: no preallocation
+ {
- # @metadata: preallocate only for metadata
+     QEMUIOVector qiov;
- # @falloc: like @full preallocation but allocate disk space by
+     struct iovec iov;
--#          posix_fallocate() rather than writing zeros.
+-    Coroutine *co;
--# @full: preallocate all data by writing zeros to device to ensure disk
+     BlkRwCo rwco;
--#        space is really available. @full preallocation also sets up
--#        metadata correctly.
+     iov = (struct iovec) {
-+#          posix_fallocate() rather than writing data.
+@@ -XXX,XX +XXX,XX @@ static int blk_prw(BlockBackend *blk, int64_t offset, uint8_t *buf,
-+# @full: preallocate all data by writing it to the device to ensure
+         .ret    = NOT_DONE,
-+#        disk space is really available. This data may or may not be
+     };
-+#        zero, depending on the image format and storage.
-+#        @full preallocation also sets up metadata correctly.
+-    co = qemu_coroutine_create(co_entry, &rwco);
- #
+-    qemu_coroutine_enter(co);
- # Since: 2.2
+-    BDRV_POLL_WHILE(blk_bs(blk), rwco.ret == NOT_DONE);
- ##
++    if (qemu_in_coroutine()) {
-diff --git a/docs/qemu-block-drivers.texi b/docs/qemu-block-drivers.texi
++        /* Fast-path if already in coroutine context */
-index XXXXXXX..XXXXXXX 100644
++        co_entry(&rwco);
---- a/docs/qemu-block-drivers.texi
++    } else {
-+++ b/docs/qemu-block-drivers.texi
++        Coroutine *co = qemu_coroutine_create(co_entry, &rwco);
-@@ -XXX,XX +XXX,XX @@ Supported options:
++        qemu_coroutine_enter(co);
- @item preallocation
++        BDRV_POLL_WHILE(blk_bs(blk), rwco.ret == NOT_DONE);
- Preallocation mode (allowed values: @code{off}, @code{falloc}, @code{full}).
++    }
- @code{falloc} mode preallocates space for image by calling posix_fallocate().
--@code{full} mode preallocates space for image by writing zeros to underlying
+     return rwco.ret;
--storage.
+ }
 +@code{full} mode preallocates space for image by writing data to underlying
 +storage.  This data may or may not be zero, depending on the storage location.
  @end table
  @item qcow2
 diff --git a/qemu-img.texi b/qemu-img.texi
 index XXXXXXX..XXXXXXX 100644
 --- a/qemu-img.texi
 +++ b/qemu-img.texi
@@ -XXX,XX +XXX,XX @@ Supported options:
  @item preallocation
  Preallocation mode (allowed values: @code{off}, @code{falloc}, @code{full}).
  @code{falloc} mode preallocates space for image by calling posix_fallocate().
 -@code{full} mode preallocates space for image by writing zeros to underlying
 -storage.
 +@code{full} mode preallocates space for image by writing data to underlying
 +storage.  This data may or may not be zero, depending on the storage location.
  @end table
  @item qcow2
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 14/17] vmdk: Make block_status recurse for flat extents
+[Qemu-devel] [PULL v2 04/24] test-thread-pool: use generic AioContext infrastructure
-Fixes: 69f47505ee66afaa513305de0c1895a224e52c45
+From: Paolo Bonzini <pbonzini@redhat.com>
-Signed-off-by: Max Reitz <mreitz@redhat.com>
-Message-id: 20190725155512.9827-3-mreitz@redhat.com
+Once the thread pool starts using aio_co_wake, it will also need
-Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
+qemu_get_current_aio_context().  Make test-thread-pool create
-Reviewed-by: John Snow <jsnow@redhat.com>
+an AioContext with qemu_init_main_loop, so that stubs/iothread.c
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+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>
 ---
- block/vmdk.c | 3 +++
+ tests/test-thread-pool.c | 12 +++---------
-file changed, 3 insertions(+)
+file changed, 3 insertions(+), 9 deletions(-)
-diff --git a/block/vmdk.c b/block/vmdk.c
+diff --git a/tests/test-thread-pool.c b/tests/test-thread-pool.c
 index XXXXXXX..XXXXXXX 100644
---- a/block/vmdk.c
+--- a/tests/test-thread-pool.c
-+++ b/block/vmdk.c
++++ b/tests/test-thread-pool.c
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn vmdk_co_block_status(BlockDriverState *bs,
+@@ -XXX,XX +XXX,XX @@
-         if (!extent->compressed) {
+ #include "qapi/error.h"
-             ret |= BDRV_BLOCK_OFFSET_VALID;
+ #include "qemu/timer.h"
-             *map = cluster_offset + index_in_cluster;
+ #include "qemu/error-report.h"
-+            if (extent->flat) {
++#include "qemu/main-loop.h"
-+                ret |= BDRV_BLOCK_RECURSE;
-+            }
+ static AioContext *ctx;
-         }
+ static ThreadPool *pool;
-         *file = extent->file->bs;
+@@ -XXX,XX +XXX,XX @@ static void test_cancel_async(void)
-         break;
+ 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;
  }
 --
-.21.0
+.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

-[Qemu-devel] [PULL 12/17] iotests: Full mirror to existing non-zero image
+[Qemu-devel] [PULL v2 07/24] nbd: convert to use qio_channel_yield
-The result of a sync=full mirror should always be the equal to the
+From: Paolo Bonzini <pbonzini@redhat.com>
 input.  Therefore, existing images should be treated as potentially
 non-zero and thus should be explicitly initialized to be zero
 beforehand.
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+In the client, read the reply headers from a coroutine, switching the
-Message-id: 20190724171239.8764-12-mreitz@redhat.com
+read side between the "read header" coroutine and the I/O coroutine that
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+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>
 ---
- tests/qemu-iotests/041     | 62 +++++++++++++++++++++++++++++++++++---
+ block/nbd-client.h |   2 +-
- tests/qemu-iotests/041.out |  4 +--
+ block/nbd-client.c | 117 ++++++++++++++++++++++++-----------------------------
-files changed, 60 insertions(+), 6 deletions(-)
+ nbd/client.c       |   2 +-
  nbd/common.c       |   9 +----
  nbd/server.c       |  94 +++++++++++++-----------------------------
 files changed, 83 insertions(+), 141 deletions(-)
-diff --git a/tests/qemu-iotests/041 b/tests/qemu-iotests/041
+diff --git a/block/nbd-client.h b/block/nbd-client.h
-index XXXXXXX..XXXXXXX 100755
+index XXXXXXX..XXXXXXX 100644
---- a/tests/qemu-iotests/041
+--- a/block/nbd-client.h
-+++ b/tests/qemu-iotests/041
++++ b/block/nbd-client.h
-@@ -XXX,XX +XXX,XX @@ class TestUnbackedSource(iotests.QMPTestCase):
+@@ -XXX,XX +XXX,XX @@ typedef struct NBDClientSession {
-     def setUp(self):
-         qemu_img('create', '-f', iotests.imgfmt, test_img,
+     CoMutex send_mutex;
-                  str(TestUnbackedSource.image_len))
+     CoQueue free_sema;
--        self.vm = iotests.VM().add_drive(test_img)
+-    Coroutine *send_coroutine;
-+        self.vm = iotests.VM()
++    Coroutine *read_reply_co;
-         self.vm.launch()
+     int in_flight;
-+        result = self.vm.qmp('blockdev-add', node_name='drive0',
-+                             driver=iotests.imgfmt,
+     Coroutine *recv_coroutine[MAX_NBD_REQUESTS];
-+                             file={
+diff --git a/block/nbd-client.c b/block/nbd-client.c
-+                                 'driver': 'file',
+index XXXXXXX..XXXXXXX 100644
-+                                 'filename': test_img,
+--- a/block/nbd-client.c
-+                             })
++++ b/block/nbd-client.c
-+        self.assert_qmp(result, 'return', {})
+@@ -XXX,XX +XXX,XX @@
+ #define HANDLE_TO_INDEX(bs, handle) ((handle) ^ ((uint64_t)(intptr_t)bs))
-     def tearDown(self):
+ #define INDEX_TO_HANDLE(bs, index)  ((index)  ^ ((uint64_t)(intptr_t)bs))
-         self.vm.shutdown()
-@@ -XXX,XX +XXX,XX @@ class TestUnbackedSource(iotests.QMPTestCase):
+-static void nbd_recv_coroutines_enter_all(NBDClientSession *s)
++static void nbd_recv_coroutines_enter_all(BlockDriverState *bs)
-     def test_absolute_paths_full(self):
+ {
-         self.assert_no_active_block_jobs()
++    NBDClientSession *s = nbd_get_client_session(bs);
--        result = self.vm.qmp('drive-mirror', device='drive0',
+     int i;
-+        result = self.vm.qmp('drive-mirror', job_id='drive0', device='drive0',
-                              sync='full', target=target_img,
+     for (i = 0; i < MAX_NBD_REQUESTS; i++) {
-                              mode='absolute-paths')
+@@ -XXX,XX +XXX,XX @@ static void nbd_recv_coroutines_enter_all(NBDClientSession *s)
-         self.assert_qmp(result, 'return', {})
+             qemu_coroutine_enter(s->recv_coroutine[i]);
-@@ -XXX,XX +XXX,XX @@ class TestUnbackedSource(iotests.QMPTestCase):
+         }
+     }
-     def test_absolute_paths_top(self):
++    BDRV_POLL_WHILE(bs, s->read_reply_co);
-         self.assert_no_active_block_jobs()
+ }
--        result = self.vm.qmp('drive-mirror', device='drive0',
-+        result = self.vm.qmp('drive-mirror', job_id='drive0', device='drive0',
+ static void nbd_teardown_connection(BlockDriverState *bs)
-                              sync='top', target=target_img,
+@@ -XXX,XX +XXX,XX @@ static void nbd_teardown_connection(BlockDriverState *bs)
-                              mode='absolute-paths')
+     qio_channel_shutdown(client->ioc,
-         self.assert_qmp(result, 'return', {})
+                          QIO_CHANNEL_SHUTDOWN_BOTH,
-@@ -XXX,XX +XXX,XX @@ class TestUnbackedSource(iotests.QMPTestCase):
+                          NULL);
+-    nbd_recv_coroutines_enter_all(client);
-     def test_absolute_paths_none(self):
++    nbd_recv_coroutines_enter_all(bs);
-         self.assert_no_active_block_jobs()
--        result = self.vm.qmp('drive-mirror', device='drive0',
+     nbd_client_detach_aio_context(bs);
-+        result = self.vm.qmp('drive-mirror', job_id='drive0', device='drive0',
+     object_unref(OBJECT(client->sioc));
-                              sync='none', target=target_img,
+@@ -XXX,XX +XXX,XX @@ static void nbd_teardown_connection(BlockDriverState *bs)
-                              mode='absolute-paths')
+     client->ioc = NULL;
-         self.assert_qmp(result, 'return', {})
+ }
-         self.complete_and_wait()
-         self.assert_no_active_block_jobs()
+-static void nbd_reply_ready(void *opaque)
++static coroutine_fn void nbd_read_reply_entry(void *opaque)
-+    def test_existing_full(self):
+ {
-+        qemu_img('create', '-f', iotests.imgfmt, target_img,
+-    BlockDriverState *bs = opaque;
-+                 str(self.image_len))
+-    NBDClientSession *s = nbd_get_client_session(bs);
-+        qemu_io('-c', 'write -P 42 0 64k', target_img)
++    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);
 +
-+        self.assert_no_active_block_jobs()
+     s->recv_coroutine[i] = NULL;
-+        result = self.vm.qmp('drive-mirror', job_id='drive0', device='drive0',
+-    if (s->in_flight-- == MAX_NBD_REQUESTS) {
-+                             sync='full', target=target_img, mode='existing')
+-        qemu_co_queue_next(&s->free_sema);
-+        self.assert_qmp(result, 'return', {})
++    s->in_flight--;
-+        self.complete_and_wait()
++    qemu_co_queue_next(&s->free_sema);
 +        self.assert_no_active_block_jobs()
 +
-+        result = self.vm.qmp('blockdev-del', node_name='drive0')
++    /* Kick the read_reply_co to get the next reply.  */
-+        self.assert_qmp(result, 'return', {})
++    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);
 +
-+        self.assertTrue(iotests.compare_images(test_img, target_img),
+     nbd_client_put(client);
-+                        'target image does not match source after mirroring')
+ }
@@ -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);
      }
 +
-+    def test_blockdev_full(self):
++    nbd_client_receive_next_request(client);
 +        qemu_img('create', '-f', iotests.imgfmt, target_img,
 +                 str(self.image_len))
 +        qemu_io('-c', 'write -P 42 0 64k', target_img)
 +
-+        result = self.vm.qmp('blockdev-add', node_name='target',
+ out:
-+                             driver=iotests.imgfmt,
+     g_free(data);
-+                             file={
+ }
-+                                 'driver': 'file',
+@@ -XXX,XX +XXX,XX @@ void nbd_client_new(NBDExport *exp,
-+                                 'filename': target_img,
+     object_ref(OBJECT(client->sioc));
-+                             })
+     client->ioc = QIO_CHANNEL(sioc);
-+        self.assert_qmp(result, 'return', {})
+     object_ref(OBJECT(client->ioc));
-+
+-    client->can_read = true;
-+        self.assert_no_active_block_jobs()
+     client->close = close_fn;
-+        result = self.vm.qmp('blockdev-mirror', job_id='drive0', device='drive0',
-+                             sync='full', target='target')
+     data->client = client;
 +        self.assert_qmp(result, 'return', {})
 +        self.complete_and_wait()
 +        self.assert_no_active_block_jobs()
 +
 +        result = self.vm.qmp('blockdev-del', node_name='drive0')
 +        self.assert_qmp(result, 'return', {})
 +
 +        result = self.vm.qmp('blockdev-del', node_name='target')
 +        self.assert_qmp(result, 'return', {})
 +
 +        self.assertTrue(iotests.compare_images(test_img, target_img),
 +                        'target image does not match source after mirroring')
 +
  class TestGranularity(iotests.QMPTestCase):
      image_len = 10 * 1024 * 1024 # MB
 diff --git a/tests/qemu-iotests/041.out b/tests/qemu-iotests/041.out
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/041.out
 +++ b/tests/qemu-iotests/041.out
@@ -XXX,XX +XXX,XX @@
 -........................................................................................
 +..........................................................................................
  ----------------------------------------------------------------------
 -Ran 88 tests
 +Ran 90 tests
  OK
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 10/17] iotests: Convert to preallocated encrypted qcow2
+[Qemu-devel] [PULL v2 08/24] coroutine-lock: reschedule coroutine on the AioContext it was running on
-Add a test case for converting an empty image (which only returns zeroes
+From: Paolo Bonzini <pbonzini@redhat.com>
 when read) to a preallocated encrypted qcow2 image.
 qcow2_has_zero_init() should return 0 then, thus forcing qemu-img
 convert to create zero clusters.
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+As a small step towards the introduction of multiqueue, we want
-Acked-by: Stefano Garzarella <sgarzare@redhat.com>
+coroutines to remain on the same AioContext that started them,
-Tested-by: Stefano Garzarella <sgarzare@redhat.com>
+unless they are moved explicitly with e.g. aio_co_schedule.  This patch
-Message-id: 20190724171239.8764-10-mreitz@redhat.com
+avoids that coroutines switch AioContext when they use a CoMutex.
-Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
+For now it does not make much of a difference, because the CoMutex
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+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>
 ---
- tests/qemu-iotests/188     | 20 +++++++++++++++++++-
+ util/qemu-coroutine-lock.c | 5 ++---
- tests/qemu-iotests/188.out |  4 ++++
+ util/trace-events          | 1 -
-files changed, 23 insertions(+), 1 deletion(-)
+files changed, 2 insertions(+), 4 deletions(-)
-diff --git a/tests/qemu-iotests/188 b/tests/qemu-iotests/188
+diff --git a/util/qemu-coroutine-lock.c b/util/qemu-coroutine-lock.c
 index XXXXXXX..XXXXXXX 100755
 --- a/tests/qemu-iotests/188
 +++ b/tests/qemu-iotests/188
@@ -XXX,XX +XXX,XX @@ SECRETALT="secret,id=sec0,data=platypus"
  _make_test_img --object $SECRET -o "encrypt.format=luks,encrypt.key-secret=sec0,encrypt.iter-time=10" $size
 -IMGSPEC="driver=$IMGFMT,file.filename=$TEST_IMG,encrypt.key-secret=sec0"
 +IMGSPEC="driver=$IMGFMT,encrypt.key-secret=sec0,file.filename=$TEST_IMG"
  QEMU_IO_OPTIONS=$QEMU_IO_OPTIONS_NO_FMT
@@ -XXX,XX +XXX,XX @@ echo
  echo "== verify open failure with wrong password =="
  $QEMU_IO --object $SECRETALT -c "read -P 0xa 0 $size" --image-opts $IMGSPEC | _filter_qemu_io | _filter_testdir
 +_cleanup_test_img
 +
 +echo
 +echo "== verify that has_zero_init returns false when preallocating =="
 +
 +# Empty source file
 +if [ -n "$TEST_IMG_FILE" ]; then
 +    TEST_IMG_FILE="${TEST_IMG_FILE}.orig" _make_test_img $size
 +else
 +    TEST_IMG="${TEST_IMG}.orig" _make_test_img $size
 +fi
 +
 +$QEMU_IMG convert -O "$IMGFMT" --object $SECRET \
 +    -o "encrypt.format=luks,encrypt.key-secret=sec0,encrypt.iter-time=10,preallocation=metadata" \
 +    "${TEST_IMG}.orig" "$TEST_IMG"
 +
 +$QEMU_IMG compare --object $SECRET --image-opts "${IMGSPEC}.orig" "$IMGSPEC"
 +
  # success, all done
  echo "*** done"
 diff --git a/tests/qemu-iotests/188.out b/tests/qemu-iotests/188.out
 index XXXXXXX..XXXXXXX 100644
---- a/tests/qemu-iotests/188.out
+--- a/util/qemu-coroutine-lock.c
-+++ b/tests/qemu-iotests/188.out
++++ b/util/qemu-coroutine-lock.c
-@@ -XXX,XX +XXX,XX @@ read 16777216/16777216 bytes at offset 0
+@@ -XXX,XX +XXX,XX @@
+ #include "qemu/coroutine.h"
- == verify open failure with wrong password ==
+ #include "qemu/coroutine_int.h"
- qemu-io: can't open: Invalid password, cannot unlock any keyslot
+ #include "qemu/queue.h"
-+
++#include "block/aio.h"
-+== verify that has_zero_init returns false when preallocating ==
+ #include "trace.h"
-+Formatting 'TEST_DIR/t.IMGFMT.orig', fmt=IMGFMT size=16777216
-+Images are identical.
+ void qemu_co_queue_init(CoQueue *queue)
- *** done
+@@ -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"
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 16/17] iotests: Fix 141 when run with qed
+[Qemu-devel] [PULL v2 09/24] blkdebug: reschedule coroutine on the AioContext it is running on
-f47505ee has changed qcow2 in such a way that the commit job run in
+From: Paolo Bonzini <pbonzini@redhat.com>
 test 141 (and 144[1]) returns before it emits the READY event.  However,
 also runs with qed, where the order is still the other way around.
 Just filter out the {"return": {}} so the test passes for qed again.
-[1] 144 only runs with qcow2, so it is fine as it is.
+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.
-Suggested-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
+While doing the change, replace custom bottom half with aio_co_schedule.
-Fixes: 69f47505ee66afaa513305de0c1895a224e52c45
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-Message-id: 20190809185253.17535-1-mreitz@redhat.com
+Reviewed-by: Fam Zheng <famz@redhat.com>
-Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Reviewed-by: John Snow <jsnow@redhat.com>
+Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+Message-id: 20170213135235.12274-10-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
- tests/qemu-iotests/141           | 9 +++++++--
+ block/blkdebug.c | 9 +--------
- tests/qemu-iotests/141.out       | 5 -----
+file changed, 1 insertion(+), 8 deletions(-)
  tests/qemu-iotests/common.filter | 5 +++++
 files changed, 12 insertions(+), 7 deletions(-)
-diff --git a/tests/qemu-iotests/141 b/tests/qemu-iotests/141
+diff --git a/block/blkdebug.c b/block/blkdebug.c
 index XXXXXXX..XXXXXXX 100755
 --- a/tests/qemu-iotests/141
 +++ b/tests/qemu-iotests/141
@@ -XXX,XX +XXX,XX @@ test_blockjob()
                }}}" \
          'return'
 +    # If "$2" is an event, we may or may not see it before the
 +    # {"return": {}}.  Therefore, filter the {"return": {}} out both
 +    # here and in the next command.  (Naturally, if we do not see it
 +    # here, we will see it before the next command can be executed,
 +    # so it will appear in the next _send_qemu_cmd's output.)
      _send_qemu_cmd $QEMU_HANDLE \
          "$1" \
          "$2" \
 -        | _filter_img_create
 +        | _filter_img_create | _filter_qmp_empty_return
      # We want this to return an error because the block job is still running
      _send_qemu_cmd $QEMU_HANDLE \
          "{'execute': 'blockdev-del',
            'arguments': {'node-name': 'drv0'}}" \
 -        'error' | _filter_generated_node_ids
 +        'error' | _filter_generated_node_ids | _filter_qmp_empty_return
      _send_qemu_cmd $QEMU_HANDLE \
          "{'execute': 'block-job-cancel',
 diff --git a/tests/qemu-iotests/141.out b/tests/qemu-iotests/141.out
 index XXXXXXX..XXXXXXX 100644
---- a/tests/qemu-iotests/141.out
+--- a/block/blkdebug.c
-+++ b/tests/qemu-iotests/141.out
++++ b/block/blkdebug.c
-@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1048576 backing_file=TEST_DIR/m.
+@@ -XXX,XX +XXX,XX @@ out:
- Formatting 'TEST_DIR/o.IMGFMT', fmt=IMGFMT size=1048576 backing_file=TEST_DIR/t.IMGFMT backing_fmt=IMGFMT
+     return ret;
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "job0"}}
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "job0"}}
 -{"return": {}}
  {"error": {"class": "GenericError", "desc": "Node drv0 is in use"}}
  {"return": {}}
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "job0"}}
@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/o.IMGFMT', fmt=IMGFMT size=1048576 backing_file=TEST_DIR/t.
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "job0"}}
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "job0"}}
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "job0", "len": 0, "offset": 0, "speed": 0, "type": "mirror"}}
 -{"return": {}}
  {"error": {"class": "GenericError", "desc": "Node 'drv0' is busy: block device is in use by block job: mirror"}}
  {"return": {}}
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "job0"}}
@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/o.IMGFMT', fmt=IMGFMT size=1048576 backing_file=TEST_DIR/t.
  {"return": {}}
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "job0"}}
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "job0"}}
 -{"return": {}}
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "job0"}}
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "job0", "len": 0, "offset": 0, "speed": 0, "type": "commit"}}
  {"error": {"class": "GenericError", "desc": "Node 'drv0' is busy: block device is in use by block job: commit"}}
@@ -XXX,XX +XXX,XX @@ wrote 1048576/1048576 bytes at offset 0
  {"return": {}}
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "job0"}}
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "job0"}}
 -{"return": {}}
  {"error": {"class": "GenericError", "desc": "Node drv0 is in use"}}
  {"return": {}}
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "job0"}}
@@ -XXX,XX +XXX,XX @@ wrote 1048576/1048576 bytes at offset 0
  {"return": {}}
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "job0"}}
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "job0"}}
 -{"return": {}}
  {"error": {"class": "GenericError", "desc": "Node drv0 is in use"}}
  {"return": {}}
  {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "job0"}}
 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_nbd()
          -e 's#\(foo\|PORT/\?\|.sock\): Failed to .*$#\1#'
  }
-+_filter_qmp_empty_return()
+-static void error_callback_bh(void *opaque)
-+{
+-{
-+    grep -v '{"return": {}}'
+-    Coroutine *co = opaque;
-+}
+-    qemu_coroutine_enter(co);
-+
+-}
- # make sure this script returns success
+-
- true
+ 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();
      }
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 08/17] vdi: Fix .bdrv_has_zero_init()
+[Qemu-devel] [PULL v2 10/24] qed: introduce qed_aio_start_io and qed_aio_next_io_cb
-Static VDI images cannot guarantee to be zero-initialized.  If the image
+From: Paolo Bonzini <pbonzini@redhat.com>
 has been statically allocated, forward the call to the underlying
 storage node.
-Reported-by: Stefano Garzarella <sgarzare@redhat.com>
+qed_aio_start_io and qed_aio_next_io will not have to acquire/release
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+the AioContext, while qed_aio_next_io_cb will.  Split the functionality
-Reviewed-by: Stefan Weil <sw@weilnetz.de>
+and gain a little type-safety in the process.
-Acked-by: Stefano Garzarella <sgarzare@redhat.com>
-Tested-by: Stefano Garzarella <sgarzare@redhat.com>
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Message-id: 20190724171239.8764-8-mreitz@redhat.com
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-Reviewed-by: Maxim Levitsky <mlevitsk@redhat.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: 20170213135235.12274-11-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
- block/vdi.c | 13 ++++++++++++-
+ block/qed.c | 39 +++++++++++++++++++++++++--------------
-file changed, 12 insertions(+), 1 deletion(-)
+file changed, 25 insertions(+), 14 deletions(-)
-diff --git a/block/vdi.c b/block/vdi.c
+diff --git a/block/qed.c b/block/qed.c
 index XXXXXXX..XXXXXXX 100644
---- a/block/vdi.c
+--- a/block/qed.c
-+++ b/block/vdi.c
++++ b/block/qed.c
-@@ -XXX,XX +XXX,XX @@ static void vdi_close(BlockDriverState *bs)
+@@ -XXX,XX +XXX,XX @@ static CachedL2Table *qed_new_l2_table(BDRVQEDState *s)
-     error_free(s->migration_blocker);
+     return l2_table;
  }
-+static int vdi_has_zero_init(BlockDriverState *bs)
+-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)
 +{
-+    BDRVVdiState *s = bs->opaque;
++    qed_aio_next_io(acb, 0);
 +
 +    if (s->header.image_type == VDI_TYPE_STATIC) {
 +        return bdrv_has_zero_init(bs->file->bs);
 +    } else {
 +        return 1;
 +    }
 +}
 +
- static QemuOptsList vdi_create_opts = {
++static void qed_aio_next_io_cb(void *opaque, int ret)
-     .name = "vdi-create-opts",
++{
-     .head = QTAILQ_HEAD_INITIALIZER(vdi_create_opts.head),
++    QEDAIOCB *acb = opaque;
-@@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_vdi = {
++
-     .bdrv_child_perm          = bdrv_format_default_perms,
++    qed_aio_next_io(acb, ret);
-     .bdrv_co_create      = vdi_co_create,
++}
-     .bdrv_co_create_opts = vdi_co_create_opts,
--    .bdrv_has_zero_init = bdrv_has_zero_init_1,
+ static void qed_plug_allocating_write_reqs(BDRVQEDState *s)
-+    .bdrv_has_zero_init  = vdi_has_zero_init,
+ {
-     .bdrv_co_block_status = vdi_co_block_status,
+@@ -XXX,XX +XXX,XX @@ static void qed_unplug_allocating_write_reqs(BDRVQEDState *s)
-     .bdrv_make_empty = vdi_make_empty,
      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;
  }
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 15/17] vpc: Do not return RAW from block_status
+[Qemu-devel] [PULL v2 11/24] aio: push aio_context_acquire/release down to dispatching
-vpc is not really a passthrough driver, even when using the fixed
+From: Paolo Bonzini <pbonzini@redhat.com>
 subformat (where host and guest offsets are equal).  It should handle
 preallocation like all other drivers do, namely by returning
 DATA | RECURSE instead of RAW.
-There is no tangible difference but the fact that bdrv_is_allocated() no
+The AioContext data structures are now protected by list_lock and/or
-longer falls through to the protocol layer.
+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.
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Message-id: 20190725155512.9827-4-mreitz@redhat.com
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-Reviewed-by: John Snow <jsnow@redhat.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: 20170213135235.12274-12-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
- block/vpc.c | 2 +-
+ util/aio-posix.c | 25 +++++++++++--------------
-file changed, 1 insertion(+), 1 deletion(-)
+ util/aio-win32.c | 15 +++++++--------
  util/async.c     |  2 ++
 files changed, 20 insertions(+), 22 deletions(-)
-diff --git a/block/vpc.c b/block/vpc.c
+diff --git a/util/aio-posix.c b/util/aio-posix.c
 index XXXXXXX..XXXXXXX 100644
---- a/block/vpc.c
+--- a/util/aio-posix.c
-+++ b/block/vpc.c
++++ b/util/aio-posix.c
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn vpc_co_block_status(BlockDriverState *bs,
+@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
-         *pnum = bytes;
+             (revents & (G_IO_IN | G_IO_HUP | G_IO_ERR)) &&
-         *map = offset;
+             aio_node_check(ctx, node->is_external) &&
-         *file = bs->file->bs;
+             node->io_read) {
--        return BDRV_BLOCK_RAW | BDRV_BLOCK_OFFSET_VALID;
++            aio_context_acquire(ctx);
-+        return BDRV_BLOCK_DATA | BDRV_BLOCK_OFFSET_VALID | BDRV_BLOCK_RECURSE;
+             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)
      }
-     qemu_co_mutex_lock(&s->lock);
+     /* 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;
 --
-.21.0
+.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

-[Qemu-devel] [PULL 07/17] qcow2: Fix .bdrv_has_zero_init()
+[Qemu-devel] [PULL v2 13/24] block: explicitly acquire aiocontext in callbacks that need it
-If a qcow2 file is preallocated, it can no longer guarantee that it
+From: Paolo Bonzini <pbonzini@redhat.com>
 initially appears as filled with zeroes.
-So implement .bdrv_has_zero_init() by checking whether the file is
+This covers both file descriptor callbacks and polling callbacks,
-preallocated; if so, forward the call to the underlying storage node,
+since they execute related code.
 except for when it is encrypted: Encrypted preallocated images always
 return effectively random data, so .bdrv_has_zero_init() must always
 return 0 for them.
-.bdrv_has_zero_init_truncate() can remain bdrv_has_zero_init_1(),
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-because it presupposes PREALLOC_MODE_OFF.
+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 ------
 files changed, 68 insertions(+), 55 deletions(-)
-Reported-by: Stefano Garzarella <sgarzare@redhat.com>
+diff --git a/block/curl.c b/block/curl.c
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+index XXXXXXX..XXXXXXX 100644
-Message-id: 20190724171239.8764-7-mreitz@redhat.com
+--- a/block/curl.c
-Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
++++ b/block/curl.c
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+@@ -XXX,XX +XXX,XX @@ static void curl_multi_check_completion(BDRVCURLState *s)
----
+     }
- block/qcow2.c | 29 ++++++++++++++++++++++++++++-
+ }
-file changed, 28 insertions(+), 1 deletion(-)
+-static void curl_multi_do(void *arg)
-diff --git a/block/qcow2.c b/block/qcow2.c
++static void curl_multi_do_locked(CURLState *s)
-index XXXXXXX..XXXXXXX 100644
+ {
---- a/block/qcow2.c
+-    CURLState *s = (CURLState *)arg;
-+++ b/block/qcow2.c
+     CURLSocket *socket, *next_socket;
-@@ -XXX,XX +XXX,XX @@ static ImageInfoSpecific *qcow2_get_specific_info(BlockDriverState *bs,
+     int running;
-     return spec_info;
+     int r;
- }
+@@ -XXX,XX +XXX,XX @@ static void curl_multi_do(void *arg)
+     }
-+static int qcow2_has_zero_init(BlockDriverState *bs)
+ }
 +static void curl_multi_do(void *arg)
 +{
-+    BDRVQcow2State *s = bs->opaque;
++    CURLState *s = (CURLState *)arg;
-+    bool preallocated;
++
-+
++    aio_context_acquire(s->s->aio_context);
-+    if (qemu_in_coroutine()) {
++    curl_multi_do_locked(s);
-+        qemu_co_mutex_lock(&s->lock);
++    aio_context_release(s->s->aio_context);
 +    }
 +    /*
 +     * Check preallocation status: Preallocated images have all L2
 +     * tables allocated, nonpreallocated images have none.  It is
 +     * therefore enough to check the first one.
 +     */
 +    preallocated = s->l1_size > 0 && s->l1_table[0] != 0;
 +    if (qemu_in_coroutine()) {
 +        qemu_co_mutex_unlock(&s->lock);
 +    }
 +
 +    if (!preallocated) {
 +        return 1;
 +    } else if (bs->encrypted) {
 +        return 0;
 +    } else {
 +        return bdrv_has_zero_init(s->data_file->bs);
 +    }
 +}
 +
- static int qcow2_save_vmstate(BlockDriverState *bs, QEMUIOVector *qiov,
+ static void curl_multi_read(void *arg)
-                               int64_t pos)
+ {
- {
+     CURLState *s = (CURLState *)arg;
-@@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_qcow2 = {
-     .bdrv_child_perm      = bdrv_format_default_perms,
+-    curl_multi_do(arg);
-     .bdrv_co_create_opts  = qcow2_co_create_opts,
++    aio_context_acquire(s->s->aio_context);
-     .bdrv_co_create       = qcow2_co_create,
++    curl_multi_do_locked(s);
--    .bdrv_has_zero_init = bdrv_has_zero_init_1,
+     curl_multi_check_completion(s->s);
-+    .bdrv_has_zero_init   = qcow2_has_zero_init,
++    aio_context_release(s->s->aio_context);
-     .bdrv_has_zero_init_truncate = bdrv_has_zero_init_1,
+ }
-     .bdrv_co_block_status = qcow2_co_block_status,
  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;
              }
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 05/17] block: Implement .bdrv_has_zero_init_truncate()
+[Qemu-devel] [PULL v2 14/24] block: explicitly acquire aiocontext in bottom halves that need it
-We need to implement .bdrv_has_zero_init_truncate() for every block
+From: Paolo Bonzini <pbonzini@redhat.com>
 driver that supports truncation and has a .bdrv_has_zero_init()
 implementation.
-Implement it the same way each driver implements .bdrv_has_zero_init().
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-This is at least not any more unsafe than what we had before.
+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(-)
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+diff --git a/block/archipelago.c b/block/archipelago.c
-Message-id: 20190724171239.8764-5-mreitz@redhat.com
+index XXXXXXX..XXXXXXX 100644
-Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
+--- a/block/archipelago.c
-Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
++++ b/block/archipelago.c
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+@@ -XXX,XX +XXX,XX @@ static void qemu_archipelago_complete_aio(void *opaque)
----
+ {
- block/file-posix.c | 1 +
+     AIORequestData *reqdata = (AIORequestData *) opaque;
- block/file-win32.c | 1 +
+     ArchipelagoAIOCB *aio_cb = (ArchipelagoAIOCB *) reqdata->aio_cb;
- block/gluster.c    | 4 ++++
++    AioContext *ctx = bdrv_get_aio_context(aio_cb->common.bs);
- block/nfs.c        | 1 +
- block/qcow2.c      | 1 +
++    aio_context_acquire(ctx);
- block/qed.c        | 1 +
+     aio_cb->common.cb(aio_cb->common.opaque, aio_cb->ret);
- block/raw-format.c | 6 ++++++
++    aio_context_release(ctx);
- block/rbd.c        | 1 +
+     aio_cb->status = 0;
- block/sheepdog.c   | 1 +
- block/ssh.c        | 1 +
+     qemu_aio_unref(aio_cb);
-files changed, 18 insertions(+)
+diff --git a/block/blkreplay.c b/block/blkreplay.c
+index XXXXXXX..XXXXXXX 100755
-diff --git a/block/file-posix.c b/block/file-posix.c
+--- a/block/blkreplay.c
-index XXXXXXX..XXXXXXX 100644
++++ b/block/blkreplay.c
---- a/block/file-posix.c
+@@ -XXX,XX +XXX,XX @@ static int64_t blkreplay_getlength(BlockDriverState *bs)
-+++ b/block/file-posix.c
+ static void blkreplay_bh_cb(void *opaque)
-@@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_file = {
+ {
-     .bdrv_co_create = raw_co_create,
+     Request *req = opaque;
-     .bdrv_co_create_opts = raw_co_create_opts,
+-    qemu_coroutine_enter(req->co);
-     .bdrv_has_zero_init = bdrv_has_zero_init_1,
++    aio_co_wake(req->co);
-+    .bdrv_has_zero_init_truncate = bdrv_has_zero_init_1,
+     qemu_bh_delete(req->bh);
-     .bdrv_co_block_status = raw_co_block_status,
+     g_free(req);
-     .bdrv_co_invalidate_cache = raw_co_invalidate_cache,
+ }
-     .bdrv_co_pwrite_zeroes = raw_co_pwrite_zeroes,
+diff --git a/block/block-backend.c b/block/block-backend.c
-diff --git a/block/file-win32.c b/block/file-win32.c
+index XXXXXXX..XXXXXXX 100644
-index XXXXXXX..XXXXXXX 100644
+--- a/block/block-backend.c
---- a/block/file-win32.c
++++ b/block/block-backend.c
-+++ b/block/file-win32.c
+@@ -XXX,XX +XXX,XX @@ int blk_make_zero(BlockBackend *blk, BdrvRequestFlags flags)
-@@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_file = {
+ static void error_callback_bh(void *opaque)
-     .bdrv_close         = raw_close,
+ {
-     .bdrv_co_create_opts = raw_co_create_opts,
+     struct BlockBackendAIOCB *acb = opaque;
-     .bdrv_has_zero_init = bdrv_has_zero_init_1,
++    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
-+    .bdrv_has_zero_init_truncate = bdrv_has_zero_init_1,
+     bdrv_dec_in_flight(acb->common.bs);
-     .bdrv_aio_preadv    = raw_aio_preadv,
++    aio_context_acquire(ctx);
-     .bdrv_aio_pwritev   = raw_aio_pwritev,
+     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 BlockDriver bdrv_gluster = {
+@@ -XXX,XX +XXX,XX @@ static struct glfs *qemu_gluster_init(BlockdevOptionsGluster *gconf,
-     .bdrv_co_writev               = qemu_gluster_co_writev,
+     return qemu_gluster_glfs_init(gconf, errp);
-     .bdrv_co_flush_to_disk        = qemu_gluster_co_flush_to_disk,
+ }
-     .bdrv_has_zero_init           = qemu_gluster_has_zero_init,
-+    .bdrv_has_zero_init_truncate  = qemu_gluster_has_zero_init,
+-static void qemu_gluster_complete_aio(void *opaque)
- #ifdef CONFIG_GLUSTERFS_DISCARD
+-{
-     .bdrv_co_pdiscard             = qemu_gluster_co_pdiscard,
+-    GlusterAIOCB *acb = (GlusterAIOCB *)opaque;
- #endif
+-
-@@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_gluster_tcp = {
+-    qemu_coroutine_enter(acb->coroutine);
-     .bdrv_co_writev               = qemu_gluster_co_writev,
+-}
-     .bdrv_co_flush_to_disk        = qemu_gluster_co_flush_to_disk,
+-
-     .bdrv_has_zero_init           = qemu_gluster_has_zero_init,
+ /*
-+    .bdrv_has_zero_init_truncate  = qemu_gluster_has_zero_init,
+  * AIO callback routine called from GlusterFS thread.
- #ifdef CONFIG_GLUSTERFS_DISCARD
+  */
-     .bdrv_co_pdiscard             = qemu_gluster_co_pdiscard,
+@@ -XXX,XX +XXX,XX @@ static void gluster_finish_aiocb(struct glfs_fd *fd, ssize_t ret, void *arg)
- #endif
+         acb->ret = -EIO; /* Partial read/write - fail it */
-@@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_gluster_unix = {
+     }
-     .bdrv_co_writev               = qemu_gluster_co_writev,
-     .bdrv_co_flush_to_disk        = qemu_gluster_co_flush_to_disk,
+-    aio_bh_schedule_oneshot(acb->aio_context, qemu_gluster_complete_aio, acb);
-     .bdrv_has_zero_init           = qemu_gluster_has_zero_init,
++    aio_co_schedule(acb->aio_context, acb->coroutine);
-+    .bdrv_has_zero_init_truncate  = qemu_gluster_has_zero_init,
+ }
- #ifdef CONFIG_GLUSTERFS_DISCARD
-     .bdrv_co_pdiscard             = qemu_gluster_co_pdiscard,
+ static void qemu_gluster_parse_flags(int bdrv_flags, int *open_flags)
- #endif
+diff --git a/block/io.c b/block/io.c
-@@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_gluster_rdma = {
+index XXXXXXX..XXXXXXX 100644
-     .bdrv_co_writev               = qemu_gluster_co_writev,
+--- a/block/io.c
-     .bdrv_co_flush_to_disk        = qemu_gluster_co_flush_to_disk,
++++ b/block/io.c
-     .bdrv_has_zero_init           = qemu_gluster_has_zero_init,
+@@ -XXX,XX +XXX,XX @@ static void bdrv_co_drain_bh_cb(void *opaque)
-+    .bdrv_has_zero_init_truncate  = qemu_gluster_has_zero_init,
+     bdrv_dec_in_flight(bs);
- #ifdef CONFIG_GLUSTERFS_DISCARD
+     bdrv_drained_begin(bs);
-     .bdrv_co_pdiscard             = qemu_gluster_co_pdiscard,
+     data->done = true;
- #endif
+-    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 BlockDriver bdrv_nfs = {
+@@ -XXX,XX +XXX,XX @@ static void nfs_co_init_task(BlockDriverState *bs, NFSRPC *task)
-     .create_opts                    = &nfs_create_opts,
+ static void nfs_co_generic_bh_cb(void *opaque)
+ {
-     .bdrv_has_zero_init             = nfs_has_zero_init,
+     NFSRPC *task = opaque;
-+    .bdrv_has_zero_init_truncate    = nfs_has_zero_init,
++
-     .bdrv_get_allocated_file_size   = nfs_get_allocated_file_size,
+     task->complete = 1;
-     .bdrv_co_truncate               = nfs_file_co_truncate,
+-    qemu_coroutine_enter(task->co);
++    aio_co_wake(task->co);
-diff --git a/block/qcow2.c b/block/qcow2.c
+ }
-index XXXXXXX..XXXXXXX 100644
---- a/block/qcow2.c
+ static void
-+++ b/block/qcow2.c
+diff --git a/block/null.c b/block/null.c
-@@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_qcow2 = {
+index XXXXXXX..XXXXXXX 100644
-     .bdrv_co_create_opts  = qcow2_co_create_opts,
+--- a/block/null.c
-     .bdrv_co_create       = qcow2_co_create,
++++ b/block/null.c
-     .bdrv_has_zero_init = bdrv_has_zero_init_1,
+@@ -XXX,XX +XXX,XX @@ static const AIOCBInfo null_aiocb_info = {
-+    .bdrv_has_zero_init_truncate = bdrv_has_zero_init_1,
+ static void null_bh_cb(void *opaque)
-     .bdrv_co_block_status = qcow2_co_block_status,
+ {
+     NullAIOCB *acb = opaque;
-     .bdrv_co_preadv         = qcow2_co_preadv,
++    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 BlockDriver bdrv_qed = {
+@@ -XXX,XX +XXX,XX @@ static void qed_update_l2_table(BDRVQEDState *s, QEDTable *table, int index,
-     .bdrv_co_create           = bdrv_qed_co_create,
+ static void qed_aio_complete_bh(void *opaque)
-     .bdrv_co_create_opts      = bdrv_qed_co_create_opts,
+ {
-     .bdrv_has_zero_init       = bdrv_has_zero_init_1,
+     QEDAIOCB *acb = opaque;
-+    .bdrv_has_zero_init_truncate = bdrv_has_zero_init_1,
++    BDRVQEDState *s = acb_to_s(acb);
-     .bdrv_co_block_status     = bdrv_qed_co_block_status,
+     BlockCompletionFunc *cb = acb->common.cb;
-     .bdrv_co_readv            = bdrv_qed_co_readv,
+     void *user_opaque = acb->common.opaque;
-     .bdrv_co_writev           = bdrv_qed_co_writev,
+     int ret = acb->bh_ret;
-diff --git a/block/raw-format.c b/block/raw-format.c
+@@ -XXX,XX +XXX,XX @@ static void qed_aio_complete_bh(void *opaque)
-index XXXXXXX..XXXXXXX 100644
+     qemu_aio_unref(acb);
---- a/block/raw-format.c
-+++ b/block/raw-format.c
+     /* Invoke callback */
-@@ -XXX,XX +XXX,XX @@ static int raw_has_zero_init(BlockDriverState *bs)
++    qed_acquire(s);
-     return bdrv_has_zero_init(bs->file->bs);
+     cb(user_opaque, ret);
- }
++    qed_release(s);
+ }
-+static int raw_has_zero_init_truncate(BlockDriverState *bs)
-+{
+ static void qed_aio_complete(QEDAIOCB *acb, int ret)
 +    return bdrv_has_zero_init_truncate(bs->file->bs);
 +}
 +
  static int coroutine_fn raw_co_create_opts(const char *filename, QemuOpts *opts,
                                             Error **errp)
  {
@@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_raw = {
      .bdrv_co_ioctl        = &raw_co_ioctl,
      .create_opts          = &raw_create_opts,
      .bdrv_has_zero_init   = &raw_has_zero_init,
 +    .bdrv_has_zero_init_truncate = &raw_has_zero_init_truncate,
      .strong_runtime_opts  = raw_strong_runtime_opts,
      .mutable_opts         = mutable_opts,
  };
 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 @@ static BlockDriver bdrv_rbd = {
+@@ -XXX,XX +XXX,XX @@ shutdown:
-     .bdrv_co_create         = qemu_rbd_co_create,
+ static void qemu_rbd_complete_aio(RADOSCB *rcb)
-     .bdrv_co_create_opts    = qemu_rbd_co_create_opts,
+ {
-     .bdrv_has_zero_init     = bdrv_has_zero_init_1,
+     RBDAIOCB *acb = rcb->acb;
-+    .bdrv_has_zero_init_truncate = bdrv_has_zero_init_1,
++    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
-     .bdrv_get_info          = qemu_rbd_getinfo,
+     int64_t r;
-     .create_opts            = &qemu_rbd_create_opts,
-     .bdrv_getlength         = qemu_rbd_getlength,
+     r = rcb->ret;
-diff --git a/block/sheepdog.c b/block/sheepdog.c
+@@ -XXX,XX +XXX,XX @@ static void qemu_rbd_complete_aio(RADOSCB *rcb)
-index XXXXXXX..XXXXXXX 100644
+         qemu_iovec_from_buf(acb->qiov, 0, acb->bounce, acb->qiov->size);
---- a/block/sheepdog.c
+     }
-+++ b/block/sheepdog.c
+     qemu_vfree(acb->bounce);
-@@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_sheepdog = {
++
-     .bdrv_co_create               = sd_co_create,
++    aio_context_acquire(ctx);
-     .bdrv_co_create_opts          = sd_co_create_opts,
+     acb->common.cb(acb->common.opaque, (acb->ret > 0 ? 0 : acb->ret));
-     .bdrv_has_zero_init           = bdrv_has_zero_init_1,
++    aio_context_release(ctx);
-+    .bdrv_has_zero_init_truncate  = bdrv_has_zero_init_1,
-     .bdrv_getlength               = sd_getlength,
+     qemu_aio_unref(acb);
-     .bdrv_get_allocated_file_size = sd_get_allocated_file_size,
+ }
-     .bdrv_co_truncate             = sd_co_truncate,
+diff --git a/dma-helpers.c b/dma-helpers.c
-diff --git a/block/ssh.c b/block/ssh.c
+index XXXXXXX..XXXXXXX 100644
-index XXXXXXX..XXXXXXX 100644
+--- a/dma-helpers.c
---- a/block/ssh.c
++++ b/dma-helpers.c
-+++ b/block/ssh.c
+@@ -XXX,XX +XXX,XX @@ static void dma_blk_cb(void *opaque, int ret)
-@@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_ssh = {
+                                 QEMU_ALIGN_DOWN(dbs->iov.size, dbs->align));
-     .bdrv_co_create_opts          = ssh_co_create_opts,
+     }
-     .bdrv_close                   = ssh_close,
-     .bdrv_has_zero_init           = ssh_has_zero_init,
++    aio_context_acquire(dbs->ctx);
-+    .bdrv_has_zero_init_truncate  = ssh_has_zero_init,
+     dbs->acb = dbs->io_func(dbs->offset, &dbs->iov,
-     .bdrv_co_readv                = ssh_co_readv,
+                             dma_blk_cb, dbs, dbs->io_func_opaque);
-     .bdrv_co_writev               = ssh_co_writev,
++    aio_context_release(dbs->ctx);
-     .bdrv_getlength               = ssh_getlength,
+     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)
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 03/17] mirror: Fix bdrv_has_zero_init() use
+[Qemu-devel] [PULL v2 15/24] block: explicitly acquire aiocontext in aio callbacks that need it
-bdrv_has_zero_init() only has meaning for newly created images or image
+From: Paolo Bonzini <pbonzini@redhat.com>
 areas.  If the mirror job itself did not create the image, it cannot
 rely on bdrv_has_zero_init()'s result to carry any meaning.
-This is the case for drive-mirror with mode=existing and always for
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-blockdev-mirror.
+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(-)
-Note that we only have to zero-initialize the target with sync=full,
+diff --git a/block/archipelago.c b/block/archipelago.c
-because other modes actually do not promise that the target will contain
+index XXXXXXX..XXXXXXX 100644
-the same data as the source after the job -- sync=top only promises to
+--- a/block/archipelago.c
-copy anything allocated in the top layer, and sync=none will only copy
++++ b/block/archipelago.c
-new I/O.  (Which is how mirror has always handled it.)
+@@ -XXX,XX +XXX,XX @@ static void qemu_archipelago_complete_aio(void *opaque)
+ {
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+     AIORequestData *reqdata = (AIORequestData *) opaque;
-Message-id: 20190724171239.8764-3-mreitz@redhat.com
+     ArchipelagoAIOCB *aio_cb = (ArchipelagoAIOCB *) reqdata->aio_cb;
-Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
+-    AioContext *ctx = bdrv_get_aio_context(aio_cb->common.bs);
-Signed-off-by: Max Reitz <mreitz@redhat.com>
----
+-    aio_context_acquire(ctx);
- include/block/block_int.h   |  2 ++
+     aio_cb->common.cb(aio_cb->common.opaque, aio_cb->ret);
- block/mirror.c              | 11 ++++++++---
+-    aio_context_release(ctx);
- blockdev.c                  | 16 +++++++++++++---
+     aio_cb->status = 0;
- tests/test-block-iothread.c |  2 +-
-files changed, 24 insertions(+), 7 deletions(-)
+     qemu_aio_unref(aio_cb);
+diff --git a/block/block-backend.c b/block/block-backend.c
-diff --git a/include/block/block_int.h b/include/block/block_int.h
+index XXXXXXX..XXXXXXX 100644
-index XXXXXXX..XXXXXXX 100644
+--- a/block/block-backend.c
---- a/include/block/block_int.h
++++ b/block/block-backend.c
-+++ b/include/block/block_int.h
+@@ -XXX,XX +XXX,XX @@ int blk_make_zero(BlockBackend *blk, BdrvRequestFlags flags)
-@@ -XXX,XX +XXX,XX @@ BlockJob *commit_active_start(const char *job_id, BlockDriverState *bs,
+ static void error_callback_bh(void *opaque)
-  * @buf_size: The amount of data that can be in flight at one time.
+ {
-  * @mode: Whether to collapse all images in the chain to the target.
+     struct BlockBackendAIOCB *acb = opaque;
-  * @backing_mode: How to establish the target's backing chain after completion.
+-    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
-+ * @zero_target: Whether the target should be explicitly zero-initialized
-  * @on_source_error: The action to take upon error reading from the source.
+     bdrv_dec_in_flight(acb->common.bs);
-  * @on_target_error: The action to take upon error writing to the target.
+-    aio_context_acquire(ctx);
-  * @unmap: Whether to unmap target where source sectors only contain zeroes.
+     acb->common.cb(acb->common.opaque, acb->ret);
-@@ -XXX,XX +XXX,XX @@ void mirror_start(const char *job_id, BlockDriverState *bs,
+-    aio_context_release(ctx);
-                   int creation_flags, int64_t speed,
+     qemu_aio_unref(acb);
-                   uint32_t granularity, int64_t buf_size,
+ }
-                   MirrorSyncMode mode, BlockMirrorBackingMode backing_mode,
-+                  bool zero_target,
+@@ -XXX,XX +XXX,XX @@ static void blk_aio_complete(BlkAioEmAIOCB *acb)
-                   BlockdevOnError on_source_error,
+ static void blk_aio_complete_bh(void *opaque)
-                   BlockdevOnError on_target_error,
+ {
-                   bool unmap, const char *filter_node_name,
+     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 @@ typedef struct MirrorBlockJob {
+@@ -XXX,XX +XXX,XX @@ static void mirror_write_complete(void *opaque, int ret)
-     Error *replace_blocker;
+ {
-     bool is_none_mode;
+     MirrorOp *op = opaque;
-     BlockMirrorBackingMode backing_mode;
+     MirrorBlockJob *s = op->s;
-+    /* Whether the target image requires explicit zero-initialization */
++
-+    bool zero_target;
++    aio_context_acquire(blk_get_aio_context(s->common.blk));
-     MirrorCopyMode copy_mode;
+     if (ret < 0) {
-     BlockdevOnError on_source_error, on_target_error;
+         BlockErrorAction action;
-     bool synced;
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn mirror_dirty_init(MirrorBlockJob *s)
+@@ -XXX,XX +XXX,XX @@ static void mirror_write_complete(void *opaque, int ret)
-     int ret;
+         }
-     int64_t count;
+     }
+     mirror_iteration_done(op, ret);
--    if (base == NULL && !bdrv_has_zero_init(target_bs)) {
++    aio_context_release(blk_get_aio_context(s->common.blk));
-+    if (s->zero_target) {
+ }
-         if (!bdrv_can_write_zeroes_with_unmap(target_bs)) {
-             bdrv_set_dirty_bitmap(s->dirty_bitmap, 0, s->bdev_length);
+ static void mirror_read_complete(void *opaque, int ret)
-             return 0;
+ {
-@@ -XXX,XX +XXX,XX @@ static BlockJob *mirror_start_job(
+     MirrorOp *op = opaque;
-                              const char *replaces, int64_t speed,
+     MirrorBlockJob *s = op->s;
-                              uint32_t granularity, int64_t buf_size,
++
-                              BlockMirrorBackingMode backing_mode,
++    aio_context_acquire(blk_get_aio_context(s->common.blk));
-+                             bool zero_target,
+     if (ret < 0) {
-                              BlockdevOnError on_source_error,
+         BlockErrorAction action;
-                              BlockdevOnError on_target_error,
-                              bool unmap,
+@@ -XXX,XX +XXX,XX @@ static void mirror_read_complete(void *opaque, int ret)
-@@ -XXX,XX +XXX,XX @@ static BlockJob *mirror_start_job(
+         }
-     s->on_target_error = on_target_error;
-     s->is_none_mode = is_none_mode;
+         mirror_iteration_done(op, ret);
-     s->backing_mode = backing_mode;
+-        return;
-+    s->zero_target = zero_target;
++    } else {
-     s->copy_mode = copy_mode;
++        blk_aio_pwritev(s->target, op->sector_num * BDRV_SECTOR_SIZE, &op->qiov,
-     s->base = base;
++                        0, mirror_write_complete, op);
-     s->granularity = granularity;
+     }
-@@ -XXX,XX +XXX,XX @@ void mirror_start(const char *job_id, BlockDriverState *bs,
+-    blk_aio_pwritev(s->target, op->sector_num * BDRV_SECTOR_SIZE, &op->qiov,
-                   int creation_flags, int64_t speed,
+-                    0, mirror_write_complete, op);
-                   uint32_t granularity, int64_t buf_size,
++    aio_context_release(blk_get_aio_context(s->common.blk));
-                   MirrorSyncMode mode, BlockMirrorBackingMode backing_mode,
+ }
-+                  bool zero_target,
-                   BlockdevOnError on_source_error,
+ static inline void mirror_clip_sectors(MirrorBlockJob *s,
-                   BlockdevOnError on_target_error,
+diff --git a/block/null.c b/block/null.c
-                   bool unmap, const char *filter_node_name,
+index XXXXXXX..XXXXXXX 100644
-@@ -XXX,XX +XXX,XX @@ void mirror_start(const char *job_id, BlockDriverState *bs,
+--- a/block/null.c
-     is_none_mode = mode == MIRROR_SYNC_MODE_NONE;
++++ b/block/null.c
-     base = mode == MIRROR_SYNC_MODE_TOP ? backing_bs(bs) : NULL;
+@@ -XXX,XX +XXX,XX @@ static const AIOCBInfo null_aiocb_info = {
-     mirror_start_job(job_id, bs, creation_flags, target, replaces,
+ static void null_bh_cb(void *opaque)
--                     speed, granularity, buf_size, backing_mode,
+ {
-+                     speed, granularity, buf_size, backing_mode, zero_target,
+     NullAIOCB *acb = opaque;
-                      on_source_error, on_target_error, unmap, NULL, NULL,
+-    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
-                      &mirror_job_driver, is_none_mode, base, false,
+-
-                      filter_node_name, true, copy_mode, errp);
+-    aio_context_acquire(ctx);
-@@ -XXX,XX +XXX,XX @@ BlockJob *commit_active_start(const char *job_id, BlockDriverState *bs,
+     acb->common.cb(acb->common.opaque, 0);
+-    aio_context_release(ctx);
-     ret = mirror_start_job(
+     qemu_aio_unref(acb);
-                      job_id, bs, creation_flags, base, NULL, speed, 0, 0,
+ }
--                     MIRROR_LEAVE_BACKING_CHAIN,
-+                     MIRROR_LEAVE_BACKING_CHAIN, false,
+ static void null_timer_cb(void *opaque)
-                      on_error, on_error, true, cb, opaque,
+ {
-                      &commit_active_job_driver, false, base, auto_complete,
+     NullAIOCB *acb = opaque;
-                      filter_node_name, false, MIRROR_COPY_MODE_BACKGROUND,
+-    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
-diff --git a/blockdev.c b/blockdev.c
+-
-index XXXXXXX..XXXXXXX 100644
+-    aio_context_acquire(ctx);
---- a/blockdev.c
+     acb->common.cb(acb->common.opaque, 0);
-+++ b/blockdev.c
+-    aio_context_release(ctx);
-@@ -XXX,XX +XXX,XX @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
+     timer_deinit(&acb->timer);
-                                    bool has_replaces, const char *replaces,
+     qemu_aio_unref(acb);
-                                    enum MirrorSyncMode sync,
+ }
-                                    BlockMirrorBackingMode backing_mode,
+diff --git a/block/qed-cluster.c b/block/qed-cluster.c
-+                                   bool zero_target,
+index XXXXXXX..XXXXXXX 100644
-                                    bool has_speed, int64_t speed,
+--- a/block/qed-cluster.c
-                                    bool has_granularity, uint32_t granularity,
++++ b/block/qed-cluster.c
-                                    bool has_buf_size, int64_t buf_size,
+@@ -XXX,XX +XXX,XX @@ static void qed_find_cluster_cb(void *opaque, int ret)
-@@ -XXX,XX +XXX,XX @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
+     unsigned int index;
-      */
+     unsigned int n;
-     mirror_start(job_id, bs, target,
-                  has_replaces ? replaces : NULL, job_flags,
++    qed_acquire(s);
--                 speed, granularity, buf_size, sync, backing_mode,
+     if (ret) {
 +                 speed, granularity, buf_size, sync, backing_mode, zero_target,
                   on_source_error, on_target_error, unmap, filter_node_name,
                   copy_mode, errp);
  }
@@ -XXX,XX +XXX,XX @@ void qmp_drive_mirror(DriveMirror *arg, Error **errp)
      int flags;
      int64_t size;
      const char *format = arg->format;
 +    bool zero_target;
      int ret;
      bs = qmp_get_root_bs(arg->device, errp);
@@ -XXX,XX +XXX,XX @@ void qmp_drive_mirror(DriveMirror *arg, Error **errp)
          goto out;
      }
+@@ -XXX,XX +XXX,XX @@ static void qed_find_cluster_cb(void *opaque, int ret)
-+    zero_target = (arg->sync == MIRROR_SYNC_MODE_FULL &&
-+                   (arg->mode == NEW_IMAGE_MODE_EXISTING ||
+ out:
-+                    !bdrv_has_zero_init(target_bs)));
+     find_cluster_cb->cb(find_cluster_cb->opaque, ret, offset, len);
-+
++    qed_release(s);
-     ret = bdrv_try_set_aio_context(target_bs, aio_context, errp);
+     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) {
-         bdrv_unref(target_bs);
+         block_acct_failed(blk_get_stats(s->qdev.conf.blk), &r->acct);
-@@ -XXX,XX +XXX,XX @@ void qmp_drive_mirror(DriveMirror *arg, Error **errp)
+     } else {
+         block_acct_done(blk_get_stats(s->qdev.conf.blk), &r->acct);
-     blockdev_mirror_common(arg->has_job_id ? arg->job_id : NULL, bs, target_bs,
+     }
-                            arg->has_replaces, arg->replaces, arg->sync,
+     scsi_dma_complete_noio(r, ret);
--                           backing_mode, arg->has_speed, arg->speed,
++    aio_context_release(blk_get_aio_context(s->qdev.conf.blk));
-+                           backing_mode, zero_target,
+ }
-+                           arg->has_speed, arg->speed,
-                            arg->has_granularity, arg->granularity,
+ static void scsi_read_complete(void * opaque, int ret)
-                            arg->has_buf_size, arg->buf_size,
+@@ -XXX,XX +XXX,XX @@ static void scsi_read_complete(void * opaque, int ret)
-                            arg->has_on_source_error, arg->on_source_error,
-@@ -XXX,XX +XXX,XX @@ void qmp_blockdev_mirror(bool has_job_id, const char *job_id,
+     assert(r->req.aiocb != NULL);
-     AioContext *aio_context;
+     r->req.aiocb = NULL;
-     BlockMirrorBackingMode backing_mode = MIRROR_LEAVE_BACKING_CHAIN;
++    aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk));
-     Error *local_err = NULL;
+     if (scsi_disk_req_check_error(r, ret, true)) {
-+    bool zero_target;
+         goto done;
-     int ret;
+     }
+@@ -XXX,XX +XXX,XX @@ static void scsi_read_complete(void * opaque, int ret)
-     bs = qmp_get_root_bs(device, errp);
-@@ -XXX,XX +XXX,XX @@ void qmp_blockdev_mirror(bool has_job_id, const char *job_id,
+ done:
-         return;
+     scsi_req_unref(&r->req);
-     }
++    aio_context_release(blk_get_aio_context(s->qdev.conf.blk));
+ }
-+    zero_target = (sync == MIRROR_SYNC_MODE_FULL);
-+
+ /* Actually issue a read to the block device.  */
-     aio_context = bdrv_get_aio_context(bs);
+@@ -XXX,XX +XXX,XX @@ static void scsi_do_read_cb(void *opaque, int ret)
-     aio_context_acquire(aio_context);
+     assert (r->req.aiocb != NULL);
+     r->req.aiocb = NULL;
-@@ -XXX,XX +XXX,XX @@ void qmp_blockdev_mirror(bool has_job_id, const char *job_id,
++    aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk));
-     blockdev_mirror_common(has_job_id ? job_id : NULL, bs, target_bs,
+     if (ret < 0) {
-                            has_replaces, replaces, sync, backing_mode,
+         block_acct_failed(blk_get_stats(s->qdev.conf.blk), &r->acct);
--                           has_speed, speed,
+     } else {
-+                           zero_target, has_speed, speed,
+         block_acct_done(blk_get_stats(s->qdev.conf.blk), &r->acct);
-                            has_granularity, granularity,
+     }
-                            has_buf_size, buf_size,
+     scsi_do_read(opaque, ret);
-                            has_on_source_error, on_source_error,
++    aio_context_release(blk_get_aio_context(s->qdev.conf.blk));
-diff --git a/tests/test-block-iothread.c b/tests/test-block-iothread.c
+ }
-index XXXXXXX..XXXXXXX 100644
---- a/tests/test-block-iothread.c
+ /* Read more data from scsi device into buffer.  */
-+++ b/tests/test-block-iothread.c
+@@ -XXX,XX +XXX,XX @@ static void scsi_write_complete(void * opaque, int ret)
-@@ -XXX,XX +XXX,XX @@ static void test_propagate_mirror(void)
+     assert (r->req.aiocb != NULL);
+     r->req.aiocb = NULL;
-     /* Start a mirror job */
-     mirror_start("job0", src, target, NULL, JOB_DEFAULT, 0, 0, 0,
++    aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk));
--                 MIRROR_SYNC_MODE_NONE, MIRROR_OPEN_BACKING_CHAIN,
+     if (ret < 0) {
-+                 MIRROR_SYNC_MODE_NONE, MIRROR_OPEN_BACKING_CHAIN, false,
+         block_acct_failed(blk_get_stats(s->qdev.conf.blk), &r->acct);
-                  BLOCKDEV_ON_ERROR_REPORT, BLOCKDEV_ON_ERROR_REPORT,
+     } else {
-                  false, "filter_node", MIRROR_COPY_MODE_BACKGROUND,
+         block_acct_done(blk_get_stats(s->qdev.conf.blk), &r->acct);
-                  &error_abort);
+     }
      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,
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 06/17] block: Use bdrv_has_zero_init_truncate()
+[Qemu-devel] [PULL v2 16/24] aio-posix: partially inline aio_dispatch into aio_poll
-vhdx and parallels call bdrv_has_zero_init() when they do not really
+From: Paolo Bonzini <pbonzini@redhat.com>
 care about an image's post-create state but only about what happens when
 you grow an image.  That is a bit ugly, and also overly safe when
 growing preallocated images without preallocating the new areas.
-Let them use bdrv_has_zero_init_truncate() instead.
+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.
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+aio_dispatch can now become void.
-Message-id: 20190724171239.8764-6-mreitz@redhat.com
-Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-[mreitz: Added commit message]
+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/parallels.c | 2 +-
+ include/block/aio.h |  6 +-----
- block/vhdx.c      | 2 +-
+ util/aio-posix.c    | 44 ++++++++++++++------------------------------
-files changed, 2 insertions(+), 2 deletions(-)
+ util/aio-win32.c    | 13 ++++---------
  util/async.c        |  2 +-
 files changed, 20 insertions(+), 45 deletions(-)
-diff --git a/block/parallels.c b/block/parallels.c
+diff --git a/include/block/aio.h b/include/block/aio.h
 index XXXXXXX..XXXXXXX 100644
---- a/block/parallels.c
+--- a/include/block/aio.h
-+++ b/block/parallels.c
++++ b/include/block/aio.h
-@@ -XXX,XX +XXX,XX @@ static int parallels_open(BlockDriverState *bs, QDict *options, int flags,
+@@ -XXX,XX +XXX,XX @@ bool aio_pending(AioContext *ctx);
-         goto fail_options;
+ /* 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)
          }
      }
--    if (!bdrv_has_zero_init(bs->file->bs)) {
+-    qemu_lockcnt_dec(&ctx->list_lock);
-+    if (!bdrv_has_zero_init_truncate(bs->file->bs)) {
+     return progress;
-         s->prealloc_mode = PRL_PREALLOC_MODE_FALLOCATE;
+ }
 -/*
 - * 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);
      }
-diff --git a/block/vhdx.c b/block/vhdx.c
++    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/block/vhdx.c
+--- a/util/aio-win32.c
-+++ b/block/vhdx.c
++++ b/util/aio-win32.c
-@@ -XXX,XX +XXX,XX @@ static coroutine_fn int vhdx_co_writev(BlockDriverState *bs, int64_t sector_num,
+@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx, HANDLE event)
-                 /* Queue another write of zero buffers if the underlying file
+     return progress;
-                  * does not zero-fill on file extension */
+ }
--                if (bdrv_has_zero_init(bs->file->bs) == 0) {
+-bool aio_dispatch(AioContext *ctx, bool dispatch_fds)
-+                if (bdrv_has_zero_init_truncate(bs->file->bs) == 0) {
++void aio_dispatch(AioContext *ctx)
-                     use_zero_buffers = true;
+ {
+-    bool progress;
-                     /* zero fill the front, if any */
+-
 -    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;
  }
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 02/17] qemu-img: Fix bdrv_has_zero_init() use in convert
+[Qemu-devel] [PULL v2 17/24] async: remove unnecessary inc/dec pairs
-bdrv_has_zero_init() only has meaning for newly created images or image
+From: Paolo Bonzini <pbonzini@redhat.com>
 areas.  If qemu-img convert did not create the image itself, it cannot
 rely on bdrv_has_zero_init()'s result to carry any meaning.
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+Pull the increment/decrement pair out of aio_bh_poll and into the
-Message-id: 20190724171239.8764-2-mreitz@redhat.com
+callers.
-Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
-Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Signed-off-by: Max Reitz <mreitz@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>
 ---
- qemu-img.c | 11 ++++++++---
+ util/aio-posix.c |  8 +++-----
-file changed, 8 insertions(+), 3 deletions(-)
+ util/aio-win32.c |  8 ++++----
  util/async.c     | 12 ++++++------
 files changed, 13 insertions(+), 15 deletions(-)
-diff --git a/qemu-img.c b/qemu-img.c
+diff --git a/util/aio-posix.c b/util/aio-posix.c
 index XXXXXXX..XXXXXXX 100644
---- a/qemu-img.c
+--- a/util/aio-posix.c
-+++ b/qemu-img.c
++++ b/util/aio-posix.c
-@@ -XXX,XX +XXX,XX @@ typedef struct ImgConvertState {
+@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
-     bool has_zero_init;
-     bool compressed;
+ void aio_dispatch(AioContext *ctx)
-     bool unallocated_blocks_are_zero;
+ {
-+    bool target_is_new;
++    qemu_lockcnt_inc(&ctx->list_lock);
-     bool target_has_backing;
+     aio_bh_poll(ctx);
-     int64_t target_backing_sectors; /* negative if unknown */
+-
-     bool wr_in_order;
+-    qemu_lockcnt_inc(&ctx->list_lock);
-@@ -XXX,XX +XXX,XX @@ static int convert_do_copy(ImgConvertState *s)
+     aio_dispatch_handlers(ctx);
-     int64_t sector_num = 0;
+     qemu_lockcnt_dec(&ctx->list_lock);
-     /* Check whether we have zero initialisation or can get it efficiently */
+@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
--    s->has_zero_init = s->min_sparse && !s->target_has_backing
+     }
--                     ? bdrv_has_zero_init(blk_bs(s->target))
--                     : false;
+     npfd = 0;
-+    if (s->target_is_new && s->min_sparse && !s->target_has_backing) {
+-    qemu_lockcnt_dec(&ctx->list_lock);
-+        s->has_zero_init = bdrv_has_zero_init(blk_bs(s->target));
-+    } else {
+     progress |= aio_bh_poll(ctx);
-+        s->has_zero_init = false;
-+    }
+     if (ret > 0) {
+-        qemu_lockcnt_inc(&ctx->list_lock);
-     if (!s->has_zero_init && !s->target_has_backing &&
+         progress |= aio_dispatch_handlers(ctx);
-         bdrv_can_write_zeroes_with_unmap(blk_bs(s->target)))
+-        qemu_lockcnt_dec(&ctx->list_lock);
-@@ -XXX,XX +XXX,XX @@ static int img_convert(int argc, char **argv)
+     }
 +    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)
          }
      }
-+    s.target_is_new = !skip_create;
+-    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);
 +
-     flags = s.min_sparse ? (BDRV_O_RDWR | BDRV_O_UNMAP) : BDRV_O_RDWR;
+     progress |= timerlistgroup_run_timers(&ctx->tlg);
-     ret = bdrv_parse_cache_mode(cache, &flags, &writethrough);
+     return progress;
-     if (ret < 0) {
+ }
 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;
  }
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 04/17] block: Add bdrv_has_zero_init_truncate()
+[Qemu-devel] [PULL v2 18/24] block: document fields protected by AioContext lock
-No .bdrv_has_zero_init() implementation returns 1 if growing the file
+From: Paolo Bonzini <pbonzini@redhat.com>
 would add non-zero areas (at least with PREALLOC_MODE_OFF), so using it
 in lieu of this new function was always safe.
-But on the other hand, it is possible that growing an image that is not
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-zero-initialized would still add a zero-initialized area, like when
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-using nonpreallocating truncation on a preallocated image.  For callers
+Reviewed-by: Fam Zheng <famz@redhat.com>
-that care only about truncation, not about creation with potential
+Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
-preallocation, this new function is useful.
+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(-)
-Alternatively, we could have added a PreallocMode parameter to
-bdrv_has_zero_init().  But the only user would have been qemu-img
-convert, which does not have a plain PreallocMode value right now -- it
-would have to parse the creation option to obtain it.  Therefore, the
-simpler solution is to let bdrv_has_zero_init() inquire the
-preallocation status and add the new bdrv_has_zero_init_truncate() that
-presupposes PREALLOC_MODE_OFF.
-Signed-off-by: Max Reitz <mreitz@redhat.com>
-Message-id: 20190724171239.8764-4-mreitz@redhat.com
-Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
-Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
-Signed-off-by: Max Reitz <mreitz@redhat.com>
----
- include/block/block.h     |  1 +
- include/block/block_int.h |  7 +++++++
- block.c                   | 21 +++++++++++++++++++++
-files changed, 29 insertions(+)
-diff --git a/include/block/block.h b/include/block/block.h
-index XXXXXXX..XXXXXXX 100644
---- a/include/block/block.h
-+++ b/include/block/block.h
-@@ -XXX,XX +XXX,XX @@ int bdrv_pdiscard(BdrvChild *child, int64_t offset, int64_t bytes);
- int bdrv_co_pdiscard(BdrvChild *child, int64_t offset, int64_t bytes);
- int bdrv_has_zero_init_1(BlockDriverState *bs);
- int bdrv_has_zero_init(BlockDriverState *bs);
-+int bdrv_has_zero_init_truncate(BlockDriverState *bs);
- bool bdrv_unallocated_blocks_are_zero(BlockDriverState *bs);
- bool bdrv_can_write_zeroes_with_unmap(BlockDriverState *bs);
- int bdrv_block_status(BlockDriverState *bs, int64_t offset,
 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 BlockDriver {
+@@ -XXX,XX +XXX,XX @@ struct BdrvChild {
-     /*
+  * copied as well.
-      * Returns 1 if newly created images are guaranteed to contain only
+  */
-      * zeros, 0 otherwise.
+ struct BlockDriverState {
-+     * Must return 0 if .bdrv_has_zero_init_truncate() returns 0.
+-    int64_t total_sectors; /* if we are reading a disk image, give its
-      */
+-                              size in sectors */
-     int (*bdrv_has_zero_init)(BlockDriverState *bs);
++    /* Protected by big QEMU lock or read-only after opening.  No special
++     * locking needed during I/O...
 +    /*
 +     * Returns 1 if new areas added by growing the image with
 +     * PREALLOC_MODE_OFF contain only zeros, 0 otherwise.
 +     */
-+    int (*bdrv_has_zero_init_truncate)(BlockDriverState *bs);
+     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 */
 +
-     /* Remove fd handlers, timers, and other event loop callbacks so the event
++    /* If true, copy read backing sectors into image.  Can be >1 if more
-      * loop is no longer in use.  Called with no in-flight requests and in
++     * than one client has requested copy-on-read.
-      * depth-first traversal order with parents before child nodes.
++     */
-diff --git a/block.c b/block.c
++    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.c
+--- a/include/sysemu/block-backend.h
-+++ b/block.c
++++ b/include/sysemu/block-backend.h
-@@ -XXX,XX +XXX,XX @@ int bdrv_has_zero_init(BlockDriverState *bs)
+@@ -XXX,XX +XXX,XX @@ typedef struct BlockDevOps {
-     return 0;
+  * 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 {
-+int bdrv_has_zero_init_truncate(BlockDriverState *bs)
+-    /* I/O throttling.
-+{
+-     * throttle_state tells us if this BlockBackend has I/O limits configured.
-+    if (!bs->drv) {
+-     * io_limits_disabled tells us if they are currently being enforced */
-+        return 0;
++    /* I/O throttling has its own locking, but also some fields are
-+    }
++     * protected by the AioContext lock.
 +     */
 +
-+    if (bs->backing) {
++    /* Protected by AioContext lock.  */
-+        /* Depends on the backing image length, but better safe than sorry */
+     CoQueue      throttled_reqs[2];
 +        return 0;
 +    }
 +    if (bs->drv->bdrv_has_zero_init_truncate) {
 +        return bs->drv->bdrv_has_zero_init_truncate(bs);
 +    }
 +    if (bs->file && bs->drv->is_filter) {
 +        return bdrv_has_zero_init_truncate(bs->file->bs);
 +    }
 +
-+    /* safe default */
++    /* Nonzero if the I/O limits are currently being ignored; generally
-+    return 0;
++     * it is zero.  */
-+}
+     unsigned int io_limits_disabled;
-+
- bool bdrv_unallocated_blocks_are_zero(BlockDriverState *bs)
+     /* The following fields are protected by the ThrottleGroup lock.
- {
+-     * See the ThrottleGroup documentation for details. */
-     BlockDriverInfo bdi;
++     * 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];
 --
-.21.0
+.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

-[Qemu-devel] [PULL 01/17] LUKS: support preallocation
+[Qemu-devel] [PULL v2 20/24] coroutine-lock: add limited spinning to CoMutex
-From: Maxim Levitsky <mlevitsk@redhat.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-preallocation=off and preallocation=metadata
+Running a very small critical section on pthread_mutex_t and CoMutex
-both allocate luks header only, and preallocation=falloc/full
+shows that pthread_mutex_t is much faster because it doesn't actually
-is passed to underlying file.
+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.
-Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1534951
+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: Maxim Levitsky <mlevitsk@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-Message-id: 20190716161901.1430-1-mlevitsk@redhat.com
+Reviewed-by: Fam Zheng <famz@redhat.com>
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+Message-id: 20170213181244.16297-3-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
- qapi/block-core.json |  6 +++++-
+ include/qemu/coroutine.h   |  5 +++++
- block/crypto.c       | 30 +++++++++++++++++++++++++++---
+ util/qemu-coroutine-lock.c | 51 ++++++++++++++++++++++++++++++++++++++++------
-files changed, 32 insertions(+), 4 deletions(-)
+ util/qemu-coroutine.c      |  2 +-
 files changed, 51 insertions(+), 7 deletions(-)
-diff --git a/qapi/block-core.json b/qapi/block-core.json
+diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
 index XXXXXXX..XXXXXXX 100644
---- a/qapi/block-core.json
+--- a/include/qemu/coroutine.h
-+++ b/qapi/block-core.json
++++ 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"
- # @file             Node to create the image format on
+ #include "qemu/coroutine.h"
- # @size             Size of the virtual disk in bytes
+ #include "qemu/coroutine_int.h"
-+# @preallocation    Preallocation mode for the new image
++#include "qemu/processor.h"
-+#                   (since: 4.2)
+ #include "qemu/queue.h"
-+#                   (default: off; allowed values: off, metadata, falloc, full)
+ #include "block/aio.h"
- #
+ #include "trace.h"
- # Since: 2.12
+@@ -XXX,XX +XXX,XX @@ void qemu_co_mutex_init(CoMutex *mutex)
- ##
+     memset(mutex, 0, sizeof(*mutex));
  { 'struct': 'BlockdevCreateOptionsLUKS',
    'base': 'QCryptoBlockCreateOptionsLUKS',
    'data': { 'file':             'BlockdevRef',
 -            'size':             'size' } }
 +            'size':             'size',
 +            '*preallocation':   'PreallocMode' } }
  ##
  # @BlockdevCreateOptionsNfs:
 diff --git a/block/crypto.c b/block/crypto.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/crypto.c
 +++ b/block/crypto.c
@@ -XXX,XX +XXX,XX @@ static ssize_t block_crypto_read_func(QCryptoBlock *block,
  struct BlockCryptoCreateData {
      BlockBackend *blk;
      uint64_t size;
 +    PreallocMode prealloc;
  };
@@ -XXX,XX +XXX,XX @@ static ssize_t block_crypto_init_func(QCryptoBlock *block,
       * available to the guest, so we must take account of that
       * which will be used by the crypto header
       */
 -    return blk_truncate(data->blk, data->size + headerlen, PREALLOC_MODE_OFF,
 +    return blk_truncate(data->blk, data->size + headerlen, data->prealloc,
                          errp);
  }
-@@ -XXX,XX +XXX,XX @@ static int block_crypto_open_generic(QCryptoBlockFormat format,
+-static void coroutine_fn qemu_co_mutex_lock_slowpath(CoMutex *mutex)
- static int block_crypto_co_create_generic(BlockDriverState *bs,
++static void coroutine_fn qemu_co_mutex_wake(CoMutex *mutex, Coroutine *co)
-                                           int64_t size,
++{
-                                           QCryptoBlockCreateOptions *opts,
++    /* Read co before co->ctx; pairs with smp_wmb() in
-+                                          PreallocMode prealloc,
++     * qemu_coroutine_enter().
-                                           Error **errp)
++     */
 +    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)
  {
-     int ret;
+     Coroutine *self = qemu_coroutine_self();
-@@ -XXX,XX +XXX,XX @@ static int block_crypto_co_create_generic(BlockDriverState *bs,
+     CoWaitRecord w;
-         goto cleanup;
+@@ -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);
      }
-+    if (prealloc == PREALLOC_MODE_METADATA) {
+     qemu_coroutine_yield();
-+        prealloc = PREALLOC_MODE_OFF;
+@@ -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);
 +    }
 +
-     data = (struct BlockCryptoCreateData) {
++    if (waiters == 0) {
-         .blk = blk,
+         /* Uncontended.  */
-         .size = size,
+         trace_qemu_co_mutex_lock_uncontended(mutex, self);
-+        .prealloc = prealloc,
++        mutex->ctx = ctx;
-     };
+     } else {
+-        qemu_co_mutex_lock_slowpath(mutex);
-     crypto = qcrypto_block_create(opts, NULL,
++        qemu_co_mutex_lock_slowpath(ctx, mutex);
@@ -XXX,XX +XXX,XX @@ block_crypto_co_create_luks(BlockdevCreateOptions *create_options, Error **errp)
      BlockdevCreateOptionsLUKS *luks_opts;
      BlockDriverState *bs = NULL;
      QCryptoBlockCreateOptions create_opts;
 +    PreallocMode preallocation = PREALLOC_MODE_OFF;
      int ret;
      assert(create_options->driver == BLOCKDEV_DRIVER_LUKS);
@@ -XXX,XX +XXX,XX @@ block_crypto_co_create_luks(BlockdevCreateOptions *create_options, Error **errp)
          .u.luks = *qapi_BlockdevCreateOptionsLUKS_base(luks_opts),
      };
 +    if (luks_opts->has_preallocation) {
 +        preallocation = luks_opts->preallocation;
 +    }
 +
      ret = block_crypto_co_create_generic(bs, luks_opts->size, &create_opts,
 -                                         errp);
 +                                         preallocation, errp);
      if (ret < 0) {
          goto fail;
      }
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn block_crypto_co_create_opts_luks(const char *filename,
+     mutex->holder = self;
-     QCryptoBlockCreateOptions *create_opts = NULL;
+     self->locks_held++;
-     BlockDriverState *bs = NULL;
+@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_mutex_unlock(CoMutex *mutex)
-     QDict *cryptoopts;
+     assert(mutex->holder == self);
-+    PreallocMode prealloc;
+     assert(qemu_in_coroutine());
-+    char *buf = NULL;
-     int64_t size;
++    mutex->ctx = NULL;
-     int ret;
+     mutex->holder = NULL;
-+    Error *local_err = NULL;
+     self->locks_held--;
+     if (atomic_fetch_dec(&mutex->locked) == 1) {
-     /* Parse options */
+@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_mutex_unlock(CoMutex *mutex)
-     size = qemu_opt_get_size_del(opts, BLOCK_OPT_SIZE, 0);
+         unsigned our_handoff;
-+    buf = qemu_opt_get_del(opts, BLOCK_OPT_PREALLOC);
+         if (to_wake) {
-+    prealloc = qapi_enum_parse(&PreallocMode_lookup, buf,
+-            Coroutine *co = to_wake->co;
-+                               PREALLOC_MODE_OFF, &local_err);
+-            aio_co_wake(co);
-+    g_free(buf);
++            qemu_co_mutex_wake(mutex, to_wake->co);
-+    if (local_err) {
+             break;
-+        error_propagate(errp, local_err);
+         }
-+        return -EINVAL;
-+    }
+diff --git a/util/qemu-coroutine.c b/util/qemu-coroutine.c
-+
+index XXXXXXX..XXXXXXX 100644
-     cryptoopts = qemu_opts_to_qdict_filtered(opts, NULL,
+--- a/util/qemu-coroutine.c
-                                              &block_crypto_create_opts_luks,
++++ b/util/qemu-coroutine.c
-                                              true);
+@@ -XXX,XX +XXX,XX @@ void qemu_coroutine_enter(Coroutine *co)
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn block_crypto_co_create_opts_luks(const char *filename,
+     co->ctx = qemu_get_current_aio_context();
-     }
+     /* Store co->ctx before anything that stores co.  Matches
-     /* Create format layer */
+-     * barrier in aio_co_wake.
--    ret = block_crypto_co_create_generic(bs, size, create_opts, errp);
++     * barrier in aio_co_wake and qemu_co_mutex_wake.
-+    ret = block_crypto_co_create_generic(bs, size, create_opts, prealloc, errp);
+      */
-     if (ret < 0) {
+     smp_wmb();
-         goto fail;
      }
 --
-.21.0
+.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

-[Qemu-devel] [PULL 11/17] iotests: Test convert -n to pre-filled image
+[Qemu-devel] [PULL v2 22/24] coroutine-lock: place CoMutex before CoQueue in header
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-Message-id: 20190724171239.8764-11-mreitz@redhat.com
-Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
+This will avoid forward references in the next patch.  It is also
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+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>
 ---
- tests/qemu-iotests/122     | 17 +++++++++++++++++
+ include/qemu/coroutine.h | 89 ++++++++++++++++++++++++------------------------
- tests/qemu-iotests/122.out |  8 ++++++++
+file changed, 44 insertions(+), 45 deletions(-)
 files changed, 25 insertions(+)
-diff --git a/tests/qemu-iotests/122 b/tests/qemu-iotests/122
+diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
-index XXXXXXX..XXXXXXX 100755
+index XXXXXXX..XXXXXXX 100644
---- a/tests/qemu-iotests/122
+--- a/include/qemu/coroutine.h
-+++ b/tests/qemu-iotests/122
++++ b/include/qemu/coroutine.h
-@@ -XXX,XX +XXX,XX @@ for min_sparse in 4k 8k; do
+@@ -XXX,XX +XXX,XX @@ bool qemu_in_coroutine(void);
-     $QEMU_IMG map --output=json "$TEST_IMG".orig | _filter_qemu_img_map
+  */
- done
+ 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);
 +
-+echo
++/**
-+echo '=== -n to a non-zero image ==='
++ * CoQueues are a mechanism to queue coroutines in order to continue executing
-+echo
++ * them later.
 + */
 +typedef struct CoQueue {
 +    QSIMPLEQ_HEAD(, Coroutine) entries;
 +} CoQueue;
 +
-+# Keep source zero
++/**
-+_make_test_img 64M
++ * Initialise a CoQueue. This must be called before any other operation is used
 + * on the CoQueue.
 + */
 +void qemu_co_queue_init(CoQueue *queue);
 +
-+# Output is not zero, but has bdrv_has_zero_init() == 1
++/**
-+TEST_IMG="$TEST_IMG".orig _make_test_img 64M
++ * Adds the current coroutine to the CoQueue and transfers control to the
-+$QEMU_IO -c "write -P 42 0 64k" "$TEST_IMG".orig | _filter_qemu_io
++ * caller of the coroutine.
 + */
 +void coroutine_fn qemu_co_queue_wait(CoQueue *queue);
 +
-+# Convert with -n, which should not assume that the target is zeroed
++/**
-+$QEMU_IMG convert -O $IMGFMT -n "$TEST_IMG" "$TEST_IMG".orig
++ * Restarts the next coroutine in the CoQueue and removes it from the queue.
 + *
 + * Returns true if a coroutine was restarted, false if the queue is empty.
 + */
 +bool coroutine_fn qemu_co_queue_next(CoQueue *queue);
 +
-+$QEMU_IMG compare "$TEST_IMG" "$TEST_IMG".orig
++/**
 + * Restarts all coroutines in the CoQueue and leaves the queue empty.
 + */
 +void coroutine_fn qemu_co_queue_restart_all(CoQueue *queue);
 +
- # success, all done
++/**
- echo '*** done'
++ * Enter the next coroutine in the queue
- rm -f $seq.full
++ */
-diff --git a/tests/qemu-iotests/122.out b/tests/qemu-iotests/122.out
++bool qemu_co_enter_next(CoQueue *queue);
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/122.out
 +++ b/tests/qemu-iotests/122.out
@@ -XXX,XX +XXX,XX @@ convert -c -S 8k
  { "start": 9216, "length": 8192, "depth": 0, "zero": true, "data": false},
  { "start": 17408, "length": 1024, "depth": 0, "zero": false, "data": true},
  { "start": 18432, "length": 67090432, "depth": 0, "zero": true, "data": false}]
 +
-+=== -n to a non-zero image ===
++/**
 + * Checks if the CoQueue is empty.
 + */
 +bool qemu_co_queue_empty(CoQueue *queue);
 +
-+Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
++
-+Formatting 'TEST_DIR/t.IMGFMT.orig', fmt=IMGFMT size=67108864
+ typedef struct CoRwlock {
-+wrote 65536/65536 bytes at offset 0
+     bool writer;
-+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+     int reader;
 +Images are identical.
  *** done
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 09/17] vhdx: Fix .bdrv_has_zero_init()
+[Qemu-devel] [PULL v2 23/24] coroutine-lock: add mutex argument to CoQueue APIs
-Fixed VHDX images cannot guarantee to be zero-initialized.  If the image
+From: Paolo Bonzini <pbonzini@redhat.com>
-has the "fixed" subformat, forward the call to the underlying storage
-node.
+All that CoQueue needs in order to become thread-safe is help
+from an external mutex.  Add this to the API.
-Reported-by: Stefano Garzarella <sgarzare@redhat.com>
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-Message-id: 20190724171239.8764-9-mreitz@redhat.com
+Reviewed-by: Fam Zheng <famz@redhat.com>
-Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
+Message-id: 20170213181244.16297-6-pbonzini@redhat.com
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
- block/vhdx.c | 26 +++++++++++++++++++++++++-
+ include/qemu/coroutine.h   |  8 +++++---
-file changed, 25 insertions(+), 1 deletion(-)
+ block/backup.c             |  2 +-
+ block/io.c                 |  4 ++--
-diff --git a/block/vhdx.c b/block/vhdx.c
+ block/nbd-client.c         |  2 +-
-index XXXXXXX..XXXXXXX 100644
+ block/qcow2-cluster.c      |  4 +---
---- a/block/vhdx.c
+ block/sheepdog.c           |  2 +-
-+++ b/block/vhdx.c
+ block/throttle-groups.c    |  2 +-
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn vhdx_co_check(BlockDriverState *bs,
+ hw/9pfs/9p.c               |  2 +-
-     return 0;
+ 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);
  }
-+static int vhdx_has_zero_init(BlockDriverState *bs)
+-void coroutine_fn qemu_co_queue_wait(CoQueue *queue)
-+{
++void coroutine_fn qemu_co_queue_wait(CoQueue *queue, CoMutex *mutex)
-+    BDRVVHDXState *s = bs->opaque;
+ {
-+    int state;
+     Coroutine *self = qemu_coroutine_self();
      QSIMPLEQ_INSERT_TAIL(&queue->entries, self, co_queue_next);
 +
-+    /*
++    if (mutex) {
-+     * Check the subformat: Fixed images have all BAT entries present,
++        qemu_co_mutex_unlock(mutex);
 +     * dynamic images have none (right after creation).  It is
 +     * therefore enough to check the first BAT entry.
 +     */
 +    if (!s->bat_entries) {
 +        return 1;
 +    }
 +
-+    state = s->bat[0] & VHDX_BAT_STATE_BIT_MASK;
++    /* There is no race condition here.  Other threads will call
-+    if (state == PAYLOAD_BLOCK_FULLY_PRESENT) {
++     * aio_co_schedule on our AioContext, which can reenter this
-+        /* Fixed subformat */
++     * coroutine but only after this yield and after the main loop
-+        return bdrv_has_zero_init(bs->file->bs);
++     * 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);
 +    }
-+
+ }
-+    /* Dynamic subformat */
-+    return 1;
+ /**
-+}
+@@ -XXX,XX +XXX,XX @@ void qemu_co_rwlock_rdlock(CoRwlock *lock)
-+
+     Coroutine *self = qemu_coroutine_self();
- static QemuOptsList vhdx_create_opts = {
-     .name = "vhdx-create-opts",
+     while (lock->writer) {
-     .head = QTAILQ_HEAD_INITIALIZER(vhdx_create_opts.head),
+-        qemu_co_queue_wait(&lock->queue);
-@@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_vhdx = {
++        qemu_co_queue_wait(&lock->queue, NULL);
-     .bdrv_co_create_opts    = vhdx_co_create_opts,
+     }
-     .bdrv_get_info          = vhdx_get_info,
+     lock->reader++;
-     .bdrv_co_check          = vhdx_co_check,
+     self->locks_held++;
--    .bdrv_has_zero_init     = bdrv_has_zero_init_1,
+@@ -XXX,XX +XXX,XX @@ void qemu_co_rwlock_wrlock(CoRwlock *lock)
-+    .bdrv_has_zero_init     = vhdx_has_zero_init,
+     Coroutine *self = qemu_coroutine_self();
-     .create_opts            = &vhdx_create_opts,
+     while (lock->writer || lock->reader) {
- };
+-        qemu_co_queue_wait(&lock->queue);
 +        qemu_co_queue_wait(&lock->queue, NULL);
      }
      lock->writer = true;
      self->locks_held++;
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 13/17] vdi: Make block_status recurse for fixed images
+[Qemu-devel] [PULL v2 24/24] coroutine-lock: make CoRwlock thread-safe and fair
-Suggested-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-Fixes: 69f47505ee66afaa513305de0c1895a224e52c45
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+This adds a CoMutex around the existing CoQueue.  Because the write-side
-Message-id: 20190725155512.9827-2-mreitz@redhat.com
+can just take CoMutex, the old "writer" field is not necessary anymore.
-Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
+Instead of removing it altogether, count the number of pending writers
-Reviewed-by: John Snow <jsnow@redhat.com>
+during a read-side critical section and forbid further readers from
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+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>
 ---
- block/vdi.c | 3 ++-
+ include/qemu/coroutine.h   |  3 ++-
-file changed, 2 insertions(+), 1 deletion(-)
+ util/qemu-coroutine-lock.c | 35 ++++++++++++++++++++++++-----------
 files changed, 26 insertions(+), 12 deletions(-)
-diff --git a/block/vdi.c b/block/vdi.c
+diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
 index XXXXXXX..XXXXXXX 100644
---- a/block/vdi.c
+--- a/include/qemu/coroutine.h
-+++ b/block/vdi.c
++++ b/include/qemu/coroutine.h
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn vdi_co_block_status(BlockDriverState *bs,
+@@ -XXX,XX +XXX,XX @@ bool qemu_co_queue_empty(CoQueue *queue);
-     *map = s->header.offset_data + (uint64_t)bmap_entry * s->block_size +
-         index_in_block;
-     *file = bs->file->bs;
+ typedef struct CoRwlock {
--    return BDRV_BLOCK_DATA | BDRV_BLOCK_OFFSET_VALID;
+-    bool writer;
-+    return BDRV_BLOCK_DATA | BDRV_BLOCK_OFFSET_VALID |
++    int pending_writer;
-+        (s->header.image_type == VDI_TYPE_STATIC ? BDRV_BLOCK_RECURSE : 0);
+     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);
  }
- static int coroutine_fn
+ 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.
 +     */
  }
 --
-.21.0
+.9.3

The following changes since commit 3fbd3405d2b0604ea530fc7a1828f19da1e95ff9:

Merge remote-tracking branch 'remotes/huth-gitlab/tags/pull-request-2019-08-17' into staging (2019-08-19 14:14:09 +0100)

are available in the Git repository at:

https://github.com/XanClic/qemu.git tags/pull-block-2019-08-19

for you to fetch changes up to fa27c478102a6b5d1c6b02c005607ad9404b915f:

doc: Preallocation does not require writing zeroes (2019-08-19 17:13:26 +0200)

----------------------------------------------------------------
Block patches:
- preallocation=falloc/full support for LUKS
- Various minor fixes

----------------------------------------------------------------
Max Reitz (16):
  qemu-img: Fix bdrv_has_zero_init() use in convert
  mirror: Fix bdrv_has_zero_init() use
  block: Add bdrv_has_zero_init_truncate()
  block: Implement .bdrv_has_zero_init_truncate()
  block: Use bdrv_has_zero_init_truncate()
  qcow2: Fix .bdrv_has_zero_init()
  vdi: Fix .bdrv_has_zero_init()
  vhdx: Fix .bdrv_has_zero_init()
  iotests: Convert to preallocated encrypted qcow2
  iotests: Test convert -n to pre-filled image
  iotests: Full mirror to existing non-zero image
  vdi: Make block_status recurse for fixed images
  vmdk: Make block_status recurse for flat extents
  vpc: Do not return RAW from block_status
  iotests: Fix 141 when run with qed
  doc: Preallocation does not require writing zeroes

Maxim Levitsky (1):
  LUKS: support preallocation

-- 
2.21.0

From: Maxim Levitsky <mlevitsk@redhat.com>

preallocation=off and preallocation=metadata
both allocate luks header only, and preallocation=falloc/full
is passed to underlying file.

Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1534951

Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Message-id: 20190716161901.1430-1-mlevitsk@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 qapi/block-core.json |  6 +++++-
 block/crypto.c       | 30 +++++++++++++++++++++++++++---
 2 files changed, 32 insertions(+), 4 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 @@
 #
 # @file             Node to create the image format on
 # @size             Size of the virtual disk in bytes
+# @preallocation    Preallocation mode for the new image
+#                   (since: 4.2)
+#                   (default: off; allowed values: off, metadata, falloc, full)
 #
 # Since: 2.12
 ##
 { 'struct': 'BlockdevCreateOptionsLUKS',
   'base': 'QCryptoBlockCreateOptionsLUKS',
   'data': { 'file':             'BlockdevRef',
-            'size':             'size' } }
+            'size':             'size',
+            '*preallocation':   'PreallocMode' } }
 
 ##
 # @BlockdevCreateOptionsNfs:
diff --git a/block/crypto.c b/block/crypto.c
index XXXXXXX..XXXXXXX 100644
--- a/block/crypto.c
+++ b/block/crypto.c
@@ -XXX,XX +XXX,XX @@ static ssize_t block_crypto_read_func(QCryptoBlock *block,
 struct BlockCryptoCreateData {
     BlockBackend *blk;
     uint64_t size;
+    PreallocMode prealloc;
 };
 
 
@@ -XXX,XX +XXX,XX @@ static ssize_t block_crypto_init_func(QCryptoBlock *block,
      * available to the guest, so we must take account of that
      * which will be used by the crypto header
      */
-    return blk_truncate(data->blk, data->size + headerlen, PREALLOC_MODE_OFF,
+    return blk_truncate(data->blk, data->size + headerlen, data->prealloc,
                         errp);
 }
 
@@ -XXX,XX +XXX,XX @@ static int block_crypto_open_generic(QCryptoBlockFormat format,
 static int block_crypto_co_create_generic(BlockDriverState *bs,
                                           int64_t size,
                                           QCryptoBlockCreateOptions *opts,
+                                          PreallocMode prealloc,
                                           Error **errp)
 {
     int ret;
@@ -XXX,XX +XXX,XX @@ static int block_crypto_co_create_generic(BlockDriverState *bs,
         goto cleanup;
     }
 
+    if (prealloc == PREALLOC_MODE_METADATA) {
+        prealloc = PREALLOC_MODE_OFF;
+    }
+
     data = (struct BlockCryptoCreateData) {
         .blk = blk,
         .size = size,
+        .prealloc = prealloc,
     };
 
     crypto = qcrypto_block_create(opts, NULL,
@@ -XXX,XX +XXX,XX @@ block_crypto_co_create_luks(BlockdevCreateOptions *create_options, Error **errp)
     BlockdevCreateOptionsLUKS *luks_opts;
     BlockDriverState *bs = NULL;
     QCryptoBlockCreateOptions create_opts;
+    PreallocMode preallocation = PREALLOC_MODE_OFF;
     int ret;
 
     assert(create_options->driver == BLOCKDEV_DRIVER_LUKS);
@@ -XXX,XX +XXX,XX @@ block_crypto_co_create_luks(BlockdevCreateOptions *create_options, Error **errp)
         .u.luks = *qapi_BlockdevCreateOptionsLUKS_base(luks_opts),
     };
 
+    if (luks_opts->has_preallocation) {
+        preallocation = luks_opts->preallocation;
+    }
+
     ret = block_crypto_co_create_generic(bs, luks_opts->size, &create_opts,
-                                         errp);
+                                         preallocation, errp);
     if (ret < 0) {
         goto fail;
     }
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn block_crypto_co_create_opts_luks(const char *filename,
     QCryptoBlockCreateOptions *create_opts = NULL;
     BlockDriverState *bs = NULL;
     QDict *cryptoopts;
+    PreallocMode prealloc;
+    char *buf = NULL;
     int64_t size;
     int ret;
+    Error *local_err = NULL;
 
     /* Parse options */
     size = qemu_opt_get_size_del(opts, BLOCK_OPT_SIZE, 0);
 
+    buf = qemu_opt_get_del(opts, BLOCK_OPT_PREALLOC);
+    prealloc = qapi_enum_parse(&PreallocMode_lookup, buf,
+                               PREALLOC_MODE_OFF, &local_err);
+    g_free(buf);
+    if (local_err) {
+        error_propagate(errp, local_err);
+        return -EINVAL;
+    }
+
     cryptoopts = qemu_opts_to_qdict_filtered(opts, NULL,
                                              &block_crypto_create_opts_luks,
                                              true);
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn block_crypto_co_create_opts_luks(const char *filename,
     }
 
     /* Create format layer */
-    ret = block_crypto_co_create_generic(bs, size, create_opts, errp);
+    ret = block_crypto_co_create_generic(bs, size, create_opts, prealloc, errp);
     if (ret < 0) {
         goto fail;
     }
-- 
2.21.0

bdrv_has_zero_init() only has meaning for newly created images or image
areas.  If qemu-img convert did not create the image itself, it cannot
rely on bdrv_has_zero_init()'s result to carry any meaning.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20190724171239.8764-2-mreitz@redhat.com
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 qemu-img.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)

diff --git a/qemu-img.c b/qemu-img.c
index XXXXXXX..XXXXXXX 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@ -XXX,XX +XXX,XX @@ typedef struct ImgConvertState {
     bool has_zero_init;
     bool compressed;
     bool unallocated_blocks_are_zero;
+    bool target_is_new;
     bool target_has_backing;
     int64_t target_backing_sectors; /* negative if unknown */
     bool wr_in_order;
@@ -XXX,XX +XXX,XX @@ static int convert_do_copy(ImgConvertState *s)
     int64_t sector_num = 0;
 
     /* Check whether we have zero initialisation or can get it efficiently */
-    s->has_zero_init = s->min_sparse && !s->target_has_backing
-                     ? bdrv_has_zero_init(blk_bs(s->target))
-                     : false;
+    if (s->target_is_new && s->min_sparse && !s->target_has_backing) {
+        s->has_zero_init = bdrv_has_zero_init(blk_bs(s->target));
+    } else {
+        s->has_zero_init = false;
+    }
 
     if (!s->has_zero_init && !s->target_has_backing &&
         bdrv_can_write_zeroes_with_unmap(blk_bs(s->target)))
@@ -XXX,XX +XXX,XX @@ static int img_convert(int argc, char **argv)
         }
     }
 
+    s.target_is_new = !skip_create;
+
     flags = s.min_sparse ? (BDRV_O_RDWR | BDRV_O_UNMAP) : BDRV_O_RDWR;
     ret = bdrv_parse_cache_mode(cache, &flags, &writethrough);
     if (ret < 0) {
-- 
2.21.0

bdrv_has_zero_init() only has meaning for newly created images or image
areas.  If the mirror job itself did not create the image, it cannot
rely on bdrv_has_zero_init()'s result to carry any meaning.

This is the case for drive-mirror with mode=existing and always for
blockdev-mirror.

Note that we only have to zero-initialize the target with sync=full,
because other modes actually do not promise that the target will contain
the same data as the source after the job -- sync=top only promises to
copy anything allocated in the top layer, and sync=none will only copy
new I/O.  (Which is how mirror has always handled it.)

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20190724171239.8764-3-mreitz@redhat.com
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 include/block/block_int.h   |  2 ++
 block/mirror.c              | 11 ++++++++---
 blockdev.c                  | 16 +++++++++++++---
 tests/test-block-iothread.c |  2 +-
 4 files changed, 24 insertions(+), 7 deletions(-)

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 @@ BlockJob *commit_active_start(const char *job_id, BlockDriverState *bs,
  * @buf_size: The amount of data that can be in flight at one time.
  * @mode: Whether to collapse all images in the chain to the target.
  * @backing_mode: How to establish the target's backing chain after completion.
+ * @zero_target: Whether the target should be explicitly zero-initialized
  * @on_source_error: The action to take upon error reading from the source.
  * @on_target_error: The action to take upon error writing to the target.
  * @unmap: Whether to unmap target where source sectors only contain zeroes.
@@ -XXX,XX +XXX,XX @@ void mirror_start(const char *job_id, BlockDriverState *bs,
                   int creation_flags, int64_t speed,
                   uint32_t granularity, int64_t buf_size,
                   MirrorSyncMode mode, BlockMirrorBackingMode backing_mode,
+                  bool zero_target,
                   BlockdevOnError on_source_error,
                   BlockdevOnError on_target_error,
                   bool unmap, const char *filter_node_name,
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 @@ typedef struct MirrorBlockJob {
     Error *replace_blocker;
     bool is_none_mode;
     BlockMirrorBackingMode backing_mode;
+    /* Whether the target image requires explicit zero-initialization */
+    bool zero_target;
     MirrorCopyMode copy_mode;
     BlockdevOnError on_source_error, on_target_error;
     bool synced;
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn mirror_dirty_init(MirrorBlockJob *s)
     int ret;
     int64_t count;
 
-    if (base == NULL && !bdrv_has_zero_init(target_bs)) {
+    if (s->zero_target) {
         if (!bdrv_can_write_zeroes_with_unmap(target_bs)) {
             bdrv_set_dirty_bitmap(s->dirty_bitmap, 0, s->bdev_length);
             return 0;
@@ -XXX,XX +XXX,XX @@ static BlockJob *mirror_start_job(
                              const char *replaces, int64_t speed,
                              uint32_t granularity, int64_t buf_size,
                              BlockMirrorBackingMode backing_mode,
+                             bool zero_target,
                              BlockdevOnError on_source_error,
                              BlockdevOnError on_target_error,
                              bool unmap,
@@ -XXX,XX +XXX,XX @@ static BlockJob *mirror_start_job(
     s->on_target_error = on_target_error;
     s->is_none_mode = is_none_mode;
     s->backing_mode = backing_mode;
+    s->zero_target = zero_target;
     s->copy_mode = copy_mode;
     s->base = base;
     s->granularity = granularity;
@@ -XXX,XX +XXX,XX @@ void mirror_start(const char *job_id, BlockDriverState *bs,
                   int creation_flags, int64_t speed,
                   uint32_t granularity, int64_t buf_size,
                   MirrorSyncMode mode, BlockMirrorBackingMode backing_mode,
+                  bool zero_target,
                   BlockdevOnError on_source_error,
                   BlockdevOnError on_target_error,
                   bool unmap, const char *filter_node_name,
@@ -XXX,XX +XXX,XX @@ void mirror_start(const char *job_id, BlockDriverState *bs,
     is_none_mode = mode == MIRROR_SYNC_MODE_NONE;
     base = mode == MIRROR_SYNC_MODE_TOP ? backing_bs(bs) : NULL;
     mirror_start_job(job_id, bs, creation_flags, target, replaces,
-                     speed, granularity, buf_size, backing_mode,
+                     speed, granularity, buf_size, backing_mode, zero_target,
                      on_source_error, on_target_error, unmap, NULL, NULL,
                      &mirror_job_driver, is_none_mode, base, false,
                      filter_node_name, true, copy_mode, errp);
@@ -XXX,XX +XXX,XX @@ BlockJob *commit_active_start(const char *job_id, BlockDriverState *bs,
 
     ret = mirror_start_job(
                      job_id, bs, creation_flags, base, NULL, speed, 0, 0,
-                     MIRROR_LEAVE_BACKING_CHAIN,
+                     MIRROR_LEAVE_BACKING_CHAIN, false,
                      on_error, on_error, true, cb, opaque,
                      &commit_active_job_driver, false, base, auto_complete,
                      filter_node_name, false, MIRROR_COPY_MODE_BACKGROUND,
diff --git a/blockdev.c b/blockdev.c
index XXXXXXX..XXXXXXX 100644
--- a/blockdev.c
+++ b/blockdev.c
@@ -XXX,XX +XXX,XX @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
                                    bool has_replaces, const char *replaces,
                                    enum MirrorSyncMode sync,
                                    BlockMirrorBackingMode backing_mode,
+                                   bool zero_target,
                                    bool has_speed, int64_t speed,
                                    bool has_granularity, uint32_t granularity,
                                    bool has_buf_size, int64_t buf_size,
@@ -XXX,XX +XXX,XX @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
      */
     mirror_start(job_id, bs, target,
                  has_replaces ? replaces : NULL, job_flags,
-                 speed, granularity, buf_size, sync, backing_mode,
+                 speed, granularity, buf_size, sync, backing_mode, zero_target,
                  on_source_error, on_target_error, unmap, filter_node_name,
                  copy_mode, errp);
 }
@@ -XXX,XX +XXX,XX @@ void qmp_drive_mirror(DriveMirror *arg, Error **errp)
     int flags;
     int64_t size;
     const char *format = arg->format;
+    bool zero_target;
     int ret;
 
     bs = qmp_get_root_bs(arg->device, errp);
@@ -XXX,XX +XXX,XX @@ void qmp_drive_mirror(DriveMirror *arg, Error **errp)
         goto out;
     }
 
+    zero_target = (arg->sync == MIRROR_SYNC_MODE_FULL &&
+                   (arg->mode == NEW_IMAGE_MODE_EXISTING ||
+                    !bdrv_has_zero_init(target_bs)));
+
     ret = bdrv_try_set_aio_context(target_bs, aio_context, errp);
     if (ret < 0) {
         bdrv_unref(target_bs);
@@ -XXX,XX +XXX,XX @@ void qmp_drive_mirror(DriveMirror *arg, Error **errp)
 
     blockdev_mirror_common(arg->has_job_id ? arg->job_id : NULL, bs, target_bs,
                            arg->has_replaces, arg->replaces, arg->sync,
-                           backing_mode, arg->has_speed, arg->speed,
+                           backing_mode, zero_target,
+                           arg->has_speed, arg->speed,
                            arg->has_granularity, arg->granularity,
                            arg->has_buf_size, arg->buf_size,
                            arg->has_on_source_error, arg->on_source_error,
@@ -XXX,XX +XXX,XX @@ void qmp_blockdev_mirror(bool has_job_id, const char *job_id,
     AioContext *aio_context;
     BlockMirrorBackingMode backing_mode = MIRROR_LEAVE_BACKING_CHAIN;
     Error *local_err = NULL;
+    bool zero_target;
     int ret;
 
     bs = qmp_get_root_bs(device, errp);
@@ -XXX,XX +XXX,XX @@ void qmp_blockdev_mirror(bool has_job_id, const char *job_id,
         return;
     }
 
+    zero_target = (sync == MIRROR_SYNC_MODE_FULL);
+
     aio_context = bdrv_get_aio_context(bs);
     aio_context_acquire(aio_context);
 
@@ -XXX,XX +XXX,XX @@ void qmp_blockdev_mirror(bool has_job_id, const char *job_id,
 
     blockdev_mirror_common(has_job_id ? job_id : NULL, bs, target_bs,
                            has_replaces, replaces, sync, backing_mode,
-                           has_speed, speed,
+                           zero_target, has_speed, speed,
                            has_granularity, granularity,
                            has_buf_size, buf_size,
                            has_on_source_error, on_source_error,
diff --git a/tests/test-block-iothread.c b/tests/test-block-iothread.c
index XXXXXXX..XXXXXXX 100644
--- a/tests/test-block-iothread.c
+++ b/tests/test-block-iothread.c
@@ -XXX,XX +XXX,XX @@ static void test_propagate_mirror(void)
 
     /* Start a mirror job */
     mirror_start("job0", src, target, NULL, JOB_DEFAULT, 0, 0, 0,
-                 MIRROR_SYNC_MODE_NONE, MIRROR_OPEN_BACKING_CHAIN,
+                 MIRROR_SYNC_MODE_NONE, MIRROR_OPEN_BACKING_CHAIN, false,
                  BLOCKDEV_ON_ERROR_REPORT, BLOCKDEV_ON_ERROR_REPORT,
                  false, "filter_node", MIRROR_COPY_MODE_BACKGROUND,
                  &error_abort);
-- 
2.21.0

No .bdrv_has_zero_init() implementation returns 1 if growing the file
would add non-zero areas (at least with PREALLOC_MODE_OFF), so using it
in lieu of this new function was always safe.

But on the other hand, it is possible that growing an image that is not
zero-initialized would still add a zero-initialized area, like when
using nonpreallocating truncation on a preallocated image.  For callers
that care only about truncation, not about creation with potential
preallocation, this new function is useful.

Alternatively, we could have added a PreallocMode parameter to
bdrv_has_zero_init().  But the only user would have been qemu-img
convert, which does not have a plain PreallocMode value right now -- it
would have to parse the creation option to obtain it.  Therefore, the
simpler solution is to let bdrv_has_zero_init() inquire the
preallocation status and add the new bdrv_has_zero_init_truncate() that
presupposes PREALLOC_MODE_OFF.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20190724171239.8764-4-mreitz@redhat.com
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 include/block/block.h     |  1 +
 include/block/block_int.h |  7 +++++++
 block.c                   | 21 +++++++++++++++++++++
 3 files changed, 29 insertions(+)

diff --git a/include/block/block.h b/include/block/block.h
index XXXXXXX..XXXXXXX 100644
--- a/include/block/block.h
+++ b/include/block/block.h
@@ -XXX,XX +XXX,XX @@ int bdrv_pdiscard(BdrvChild *child, int64_t offset, int64_t bytes);
 int bdrv_co_pdiscard(BdrvChild *child, int64_t offset, int64_t bytes);
 int bdrv_has_zero_init_1(BlockDriverState *bs);
 int bdrv_has_zero_init(BlockDriverState *bs);
+int bdrv_has_zero_init_truncate(BlockDriverState *bs);
 bool bdrv_unallocated_blocks_are_zero(BlockDriverState *bs);
 bool bdrv_can_write_zeroes_with_unmap(BlockDriverState *bs);
 int bdrv_block_status(BlockDriverState *bs, int64_t offset,
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 BlockDriver {
     /*
      * Returns 1 if newly created images are guaranteed to contain only
      * zeros, 0 otherwise.
+     * Must return 0 if .bdrv_has_zero_init_truncate() returns 0.
      */
     int (*bdrv_has_zero_init)(BlockDriverState *bs);
 
+    /*
+     * Returns 1 if new areas added by growing the image with
+     * PREALLOC_MODE_OFF contain only zeros, 0 otherwise.
+     */
+    int (*bdrv_has_zero_init_truncate)(BlockDriverState *bs);
+
     /* Remove fd handlers, timers, and other event loop callbacks so the event
      * loop is no longer in use.  Called with no in-flight requests and in
      * depth-first traversal order with parents before child nodes.
diff --git a/block.c b/block.c
index XXXXXXX..XXXXXXX 100644
--- a/block.c
+++ b/block.c
@@ -XXX,XX +XXX,XX @@ int bdrv_has_zero_init(BlockDriverState *bs)
     return 0;
 }
 
+int bdrv_has_zero_init_truncate(BlockDriverState *bs)
+{
+    if (!bs->drv) {
+        return 0;
+    }
+
+    if (bs->backing) {
+        /* Depends on the backing image length, but better safe than sorry */
+        return 0;
+    }
+    if (bs->drv->bdrv_has_zero_init_truncate) {
+        return bs->drv->bdrv_has_zero_init_truncate(bs);
+    }
+    if (bs->file && bs->drv->is_filter) {
+        return bdrv_has_zero_init_truncate(bs->file->bs);
+    }
+
+    /* safe default */
+    return 0;
+}
+
 bool bdrv_unallocated_blocks_are_zero(BlockDriverState *bs)
 {
     BlockDriverInfo bdi;
-- 
2.21.0

We need to implement .bdrv_has_zero_init_truncate() for every block
driver that supports truncation and has a .bdrv_has_zero_init()
implementation.

Implement it the same way each driver implements .bdrv_has_zero_init().
This is at least not any more unsafe than what we had before.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20190724171239.8764-5-mreitz@redhat.com
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block/file-posix.c | 1 +
 block/file-win32.c | 1 +
 block/gluster.c    | 4 ++++
 block/nfs.c        | 1 +
 block/qcow2.c      | 1 +
 block/qed.c        | 1 +
 block/raw-format.c | 6 ++++++
 block/rbd.c        | 1 +
 block/sheepdog.c   | 1 +
 block/ssh.c        | 1 +
 10 files changed, 18 insertions(+)

diff --git a/block/file-posix.c b/block/file-posix.c
index XXXXXXX..XXXXXXX 100644
--- a/block/file-posix.c
+++ b/block/file-posix.c
@@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_file = {
     .bdrv_co_create = raw_co_create,
     .bdrv_co_create_opts = raw_co_create_opts,
     .bdrv_has_zero_init = bdrv_has_zero_init_1,
+    .bdrv_has_zero_init_truncate = bdrv_has_zero_init_1,
     .bdrv_co_block_status = raw_co_block_status,
     .bdrv_co_invalidate_cache = raw_co_invalidate_cache,
     .bdrv_co_pwrite_zeroes = raw_co_pwrite_zeroes,
diff --git a/block/file-win32.c b/block/file-win32.c
index XXXXXXX..XXXXXXX 100644
--- a/block/file-win32.c
+++ b/block/file-win32.c
@@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_file = {
     .bdrv_close         = raw_close,
     .bdrv_co_create_opts = raw_co_create_opts,
     .bdrv_has_zero_init = bdrv_has_zero_init_1,
+    .bdrv_has_zero_init_truncate = bdrv_has_zero_init_1,
 
     .bdrv_aio_preadv    = raw_aio_preadv,
     .bdrv_aio_pwritev   = raw_aio_pwritev,
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 BlockDriver bdrv_gluster = {
     .bdrv_co_writev               = qemu_gluster_co_writev,
     .bdrv_co_flush_to_disk        = qemu_gluster_co_flush_to_disk,
     .bdrv_has_zero_init           = qemu_gluster_has_zero_init,
+    .bdrv_has_zero_init_truncate  = qemu_gluster_has_zero_init,
 #ifdef CONFIG_GLUSTERFS_DISCARD
     .bdrv_co_pdiscard             = qemu_gluster_co_pdiscard,
 #endif
@@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_gluster_tcp = {
     .bdrv_co_writev               = qemu_gluster_co_writev,
     .bdrv_co_flush_to_disk        = qemu_gluster_co_flush_to_disk,
     .bdrv_has_zero_init           = qemu_gluster_has_zero_init,
+    .bdrv_has_zero_init_truncate  = qemu_gluster_has_zero_init,
 #ifdef CONFIG_GLUSTERFS_DISCARD
     .bdrv_co_pdiscard             = qemu_gluster_co_pdiscard,
 #endif
@@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_gluster_unix = {
     .bdrv_co_writev               = qemu_gluster_co_writev,
     .bdrv_co_flush_to_disk        = qemu_gluster_co_flush_to_disk,
     .bdrv_has_zero_init           = qemu_gluster_has_zero_init,
+    .bdrv_has_zero_init_truncate  = qemu_gluster_has_zero_init,
 #ifdef CONFIG_GLUSTERFS_DISCARD
     .bdrv_co_pdiscard             = qemu_gluster_co_pdiscard,
 #endif
@@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_gluster_rdma = {
     .bdrv_co_writev               = qemu_gluster_co_writev,
     .bdrv_co_flush_to_disk        = qemu_gluster_co_flush_to_disk,
     .bdrv_has_zero_init           = qemu_gluster_has_zero_init,
+    .bdrv_has_zero_init_truncate  = qemu_gluster_has_zero_init,
 #ifdef CONFIG_GLUSTERFS_DISCARD
     .bdrv_co_pdiscard             = qemu_gluster_co_pdiscard,
 #endif
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 BlockDriver bdrv_nfs = {
     .create_opts                    = &nfs_create_opts,
 
     .bdrv_has_zero_init             = nfs_has_zero_init,
+    .bdrv_has_zero_init_truncate    = nfs_has_zero_init,
     .bdrv_get_allocated_file_size   = nfs_get_allocated_file_size,
     .bdrv_co_truncate               = nfs_file_co_truncate,
 
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 @@ BlockDriver bdrv_qcow2 = {
     .bdrv_co_create_opts  = qcow2_co_create_opts,
     .bdrv_co_create       = qcow2_co_create,
     .bdrv_has_zero_init = bdrv_has_zero_init_1,
+    .bdrv_has_zero_init_truncate = bdrv_has_zero_init_1,
     .bdrv_co_block_status = qcow2_co_block_status,
 
     .bdrv_co_preadv         = qcow2_co_preadv,
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 BlockDriver bdrv_qed = {
     .bdrv_co_create           = bdrv_qed_co_create,
     .bdrv_co_create_opts      = bdrv_qed_co_create_opts,
     .bdrv_has_zero_init       = bdrv_has_zero_init_1,
+    .bdrv_has_zero_init_truncate = bdrv_has_zero_init_1,
     .bdrv_co_block_status     = bdrv_qed_co_block_status,
     .bdrv_co_readv            = bdrv_qed_co_readv,
     .bdrv_co_writev           = bdrv_qed_co_writev,
diff --git a/block/raw-format.c b/block/raw-format.c
index XXXXXXX..XXXXXXX 100644
--- a/block/raw-format.c
+++ b/block/raw-format.c
@@ -XXX,XX +XXX,XX @@ static int raw_has_zero_init(BlockDriverState *bs)
     return bdrv_has_zero_init(bs->file->bs);
 }
 
+static int raw_has_zero_init_truncate(BlockDriverState *bs)
+{
+    return bdrv_has_zero_init_truncate(bs->file->bs);
+}
+
 static int coroutine_fn raw_co_create_opts(const char *filename, QemuOpts *opts,
                                            Error **errp)
 {
@@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_raw = {
     .bdrv_co_ioctl        = &raw_co_ioctl,
     .create_opts          = &raw_create_opts,
     .bdrv_has_zero_init   = &raw_has_zero_init,
+    .bdrv_has_zero_init_truncate = &raw_has_zero_init_truncate,
     .strong_runtime_opts  = raw_strong_runtime_opts,
     .mutable_opts         = mutable_opts,
 };
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 @@ static BlockDriver bdrv_rbd = {
     .bdrv_co_create         = qemu_rbd_co_create,
     .bdrv_co_create_opts    = qemu_rbd_co_create_opts,
     .bdrv_has_zero_init     = bdrv_has_zero_init_1,
+    .bdrv_has_zero_init_truncate = bdrv_has_zero_init_1,
     .bdrv_get_info          = qemu_rbd_getinfo,
     .create_opts            = &qemu_rbd_create_opts,
     .bdrv_getlength         = qemu_rbd_getlength,
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 BlockDriver bdrv_sheepdog = {
     .bdrv_co_create               = sd_co_create,
     .bdrv_co_create_opts          = sd_co_create_opts,
     .bdrv_has_zero_init           = bdrv_has_zero_init_1,
+    .bdrv_has_zero_init_truncate  = bdrv_has_zero_init_1,
     .bdrv_getlength               = sd_getlength,
     .bdrv_get_allocated_file_size = sd_get_allocated_file_size,
     .bdrv_co_truncate             = sd_co_truncate,
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 BlockDriver bdrv_ssh = {
     .bdrv_co_create_opts          = ssh_co_create_opts,
     .bdrv_close                   = ssh_close,
     .bdrv_has_zero_init           = ssh_has_zero_init,
+    .bdrv_has_zero_init_truncate  = ssh_has_zero_init,
     .bdrv_co_readv                = ssh_co_readv,
     .bdrv_co_writev               = ssh_co_writev,
     .bdrv_getlength               = ssh_getlength,
-- 
2.21.0

vhdx and parallels call bdrv_has_zero_init() when they do not really
care about an image's post-create state but only about what happens when
you grow an image.  That is a bit ugly, and also overly safe when
growing preallocated images without preallocating the new areas.

Let them use bdrv_has_zero_init_truncate() instead.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20190724171239.8764-6-mreitz@redhat.com
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
[mreitz: Added commit message]
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block/parallels.c | 2 +-
 block/vhdx.c      | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/block/parallels.c b/block/parallels.c
index XXXXXXX..XXXXXXX 100644
--- a/block/parallels.c
+++ b/block/parallels.c
@@ -XXX,XX +XXX,XX @@ static int parallels_open(BlockDriverState *bs, QDict *options, int flags,
         goto fail_options;
     }
 
-    if (!bdrv_has_zero_init(bs->file->bs)) {
+    if (!bdrv_has_zero_init_truncate(bs->file->bs)) {
         s->prealloc_mode = PRL_PREALLOC_MODE_FALLOCATE;
     }
 
diff --git a/block/vhdx.c b/block/vhdx.c
index XXXXXXX..XXXXXXX 100644
--- a/block/vhdx.c
+++ b/block/vhdx.c
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int vhdx_co_writev(BlockDriverState *bs, int64_t sector_num,
                 /* Queue another write of zero buffers if the underlying file
                  * does not zero-fill on file extension */
 
-                if (bdrv_has_zero_init(bs->file->bs) == 0) {
+                if (bdrv_has_zero_init_truncate(bs->file->bs) == 0) {
                     use_zero_buffers = true;
 
                     /* zero fill the front, if any */
-- 
2.21.0

If a qcow2 file is preallocated, it can no longer guarantee that it
initially appears as filled with zeroes.

So implement .bdrv_has_zero_init() by checking whether the file is
preallocated; if so, forward the call to the underlying storage node,
except for when it is encrypted: Encrypted preallocated images always
return effectively random data, so .bdrv_has_zero_init() must always
return 0 for them.

.bdrv_has_zero_init_truncate() can remain bdrv_has_zero_init_1(),
because it presupposes PREALLOC_MODE_OFF.

Reported-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20190724171239.8764-7-mreitz@redhat.com
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block/qcow2.c | 29 ++++++++++++++++++++++++++++-
 1 file changed, 28 insertions(+), 1 deletion(-)

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 ImageInfoSpecific *qcow2_get_specific_info(BlockDriverState *bs,
     return spec_info;
 }
 
+static int qcow2_has_zero_init(BlockDriverState *bs)
+{
+    BDRVQcow2State *s = bs->opaque;
+    bool preallocated;
+
+    if (qemu_in_coroutine()) {
+        qemu_co_mutex_lock(&s->lock);
+    }
+    /*
+     * Check preallocation status: Preallocated images have all L2
+     * tables allocated, nonpreallocated images have none.  It is
+     * therefore enough to check the first one.
+     */
+    preallocated = s->l1_size > 0 && s->l1_table[0] != 0;
+    if (qemu_in_coroutine()) {
+        qemu_co_mutex_unlock(&s->lock);
+    }
+
+    if (!preallocated) {
+        return 1;
+    } else if (bs->encrypted) {
+        return 0;
+    } else {
+        return bdrv_has_zero_init(s->data_file->bs);
+    }
+}
+
 static int qcow2_save_vmstate(BlockDriverState *bs, QEMUIOVector *qiov,
                               int64_t pos)
 {
@@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_qcow2 = {
     .bdrv_child_perm      = bdrv_format_default_perms,
     .bdrv_co_create_opts  = qcow2_co_create_opts,
     .bdrv_co_create       = qcow2_co_create,
-    .bdrv_has_zero_init = bdrv_has_zero_init_1,
+    .bdrv_has_zero_init   = qcow2_has_zero_init,
     .bdrv_has_zero_init_truncate = bdrv_has_zero_init_1,
     .bdrv_co_block_status = qcow2_co_block_status,
 
-- 
2.21.0

Static VDI images cannot guarantee to be zero-initialized.  If the image
has been statically allocated, forward the call to the underlying
storage node.

Reported-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Stefan Weil <sw@weilnetz.de>
Acked-by: Stefano Garzarella <sgarzare@redhat.com>
Tested-by: Stefano Garzarella <sgarzare@redhat.com>
Message-id: 20190724171239.8764-8-mreitz@redhat.com
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block/vdi.c | 13 ++++++++++++-
 1 file changed, 12 insertions(+), 1 deletion(-)

diff --git a/block/vdi.c b/block/vdi.c
index XXXXXXX..XXXXXXX 100644
--- a/block/vdi.c
+++ b/block/vdi.c
@@ -XXX,XX +XXX,XX @@ static void vdi_close(BlockDriverState *bs)
     error_free(s->migration_blocker);
 }
 
+static int vdi_has_zero_init(BlockDriverState *bs)
+{
+    BDRVVdiState *s = bs->opaque;
+
+    if (s->header.image_type == VDI_TYPE_STATIC) {
+        return bdrv_has_zero_init(bs->file->bs);
+    } else {
+        return 1;
+    }
+}
+
 static QemuOptsList vdi_create_opts = {
     .name = "vdi-create-opts",
     .head = QTAILQ_HEAD_INITIALIZER(vdi_create_opts.head),
@@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_vdi = {
     .bdrv_child_perm          = bdrv_format_default_perms,
     .bdrv_co_create      = vdi_co_create,
     .bdrv_co_create_opts = vdi_co_create_opts,
-    .bdrv_has_zero_init = bdrv_has_zero_init_1,
+    .bdrv_has_zero_init  = vdi_has_zero_init,
     .bdrv_co_block_status = vdi_co_block_status,
     .bdrv_make_empty = vdi_make_empty,
 
-- 
2.21.0

Fixed VHDX images cannot guarantee to be zero-initialized.  If the image
has the "fixed" subformat, forward the call to the underlying storage
node.

Reported-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20190724171239.8764-9-mreitz@redhat.com
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block/vhdx.c | 26 +++++++++++++++++++++++++-
 1 file changed, 25 insertions(+), 1 deletion(-)

diff --git a/block/vhdx.c b/block/vhdx.c
index XXXXXXX..XXXXXXX 100644
--- a/block/vhdx.c
+++ b/block/vhdx.c
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn vhdx_co_check(BlockDriverState *bs,
     return 0;
 }
 
+static int vhdx_has_zero_init(BlockDriverState *bs)
+{
+    BDRVVHDXState *s = bs->opaque;
+    int state;
+
+    /*
+     * Check the subformat: Fixed images have all BAT entries present,
+     * dynamic images have none (right after creation).  It is
+     * therefore enough to check the first BAT entry.
+     */
+    if (!s->bat_entries) {
+        return 1;
+    }
+
+    state = s->bat[0] & VHDX_BAT_STATE_BIT_MASK;
+    if (state == PAYLOAD_BLOCK_FULLY_PRESENT) {
+        /* Fixed subformat */
+        return bdrv_has_zero_init(bs->file->bs);
+    }
+
+    /* Dynamic subformat */
+    return 1;
+}
+
 static QemuOptsList vhdx_create_opts = {
     .name = "vhdx-create-opts",
     .head = QTAILQ_HEAD_INITIALIZER(vhdx_create_opts.head),
@@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_vhdx = {
     .bdrv_co_create_opts    = vhdx_co_create_opts,
     .bdrv_get_info          = vhdx_get_info,
     .bdrv_co_check          = vhdx_co_check,
-    .bdrv_has_zero_init     = bdrv_has_zero_init_1,
+    .bdrv_has_zero_init     = vhdx_has_zero_init,
 
     .create_opts            = &vhdx_create_opts,
 };
-- 
2.21.0

Add a test case for converting an empty image (which only returns zeroes
when read) to a preallocated encrypted qcow2 image.
qcow2_has_zero_init() should return 0 then, thus forcing qemu-img
convert to create zero clusters.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Acked-by: Stefano Garzarella <sgarzare@redhat.com>
Tested-by: Stefano Garzarella <sgarzare@redhat.com>
Message-id: 20190724171239.8764-10-mreitz@redhat.com
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 tests/qemu-iotests/188     | 20 +++++++++++++++++++-
 tests/qemu-iotests/188.out |  4 ++++
 2 files changed, 23 insertions(+), 1 deletion(-)

diff --git a/tests/qemu-iotests/188 b/tests/qemu-iotests/188
index XXXXXXX..XXXXXXX 100755
--- a/tests/qemu-iotests/188
+++ b/tests/qemu-iotests/188
@@ -XXX,XX +XXX,XX @@ SECRETALT="secret,id=sec0,data=platypus"
 
 _make_test_img --object $SECRET -o "encrypt.format=luks,encrypt.key-secret=sec0,encrypt.iter-time=10" $size
 
-IMGSPEC="driver=$IMGFMT,file.filename=$TEST_IMG,encrypt.key-secret=sec0"
+IMGSPEC="driver=$IMGFMT,encrypt.key-secret=sec0,file.filename=$TEST_IMG"
 
 QEMU_IO_OPTIONS=$QEMU_IO_OPTIONS_NO_FMT
 
@@ -XXX,XX +XXX,XX @@ echo
 echo "== verify open failure with wrong password =="
 $QEMU_IO --object $SECRETALT -c "read -P 0xa 0 $size" --image-opts $IMGSPEC | _filter_qemu_io | _filter_testdir
 
+_cleanup_test_img
+
+echo
+echo "== verify that has_zero_init returns false when preallocating =="
+
+# Empty source file
+if [ -n "$TEST_IMG_FILE" ]; then
+    TEST_IMG_FILE="${TEST_IMG_FILE}.orig" _make_test_img $size
+else
+    TEST_IMG="${TEST_IMG}.orig" _make_test_img $size
+fi
+
+$QEMU_IMG convert -O "$IMGFMT" --object $SECRET \
+    -o "encrypt.format=luks,encrypt.key-secret=sec0,encrypt.iter-time=10,preallocation=metadata" \
+    "${TEST_IMG}.orig" "$TEST_IMG"
+
+$QEMU_IMG compare --object $SECRET --image-opts "${IMGSPEC}.orig" "$IMGSPEC"
+
 
 # success, all done
 echo "*** done"
diff --git a/tests/qemu-iotests/188.out b/tests/qemu-iotests/188.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/188.out
+++ b/tests/qemu-iotests/188.out
@@ -XXX,XX +XXX,XX @@ read 16777216/16777216 bytes at offset 0
 
 == verify open failure with wrong password ==
 qemu-io: can't open: Invalid password, cannot unlock any keyslot
+
+== verify that has_zero_init returns false when preallocating ==
+Formatting 'TEST_DIR/t.IMGFMT.orig', fmt=IMGFMT size=16777216
+Images are identical.
 *** done
-- 
2.21.0

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20190724171239.8764-11-mreitz@redhat.com
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 tests/qemu-iotests/122     | 17 +++++++++++++++++
 tests/qemu-iotests/122.out |  8 ++++++++
 2 files changed, 25 insertions(+)

diff --git a/tests/qemu-iotests/122 b/tests/qemu-iotests/122
index XXXXXXX..XXXXXXX 100755
--- a/tests/qemu-iotests/122
+++ b/tests/qemu-iotests/122
@@ -XXX,XX +XXX,XX @@ for min_sparse in 4k 8k; do
     $QEMU_IMG map --output=json "$TEST_IMG".orig | _filter_qemu_img_map
 done
 
+
+echo
+echo '=== -n to a non-zero image ==='
+echo
+
+# Keep source zero
+_make_test_img 64M
+
+# Output is not zero, but has bdrv_has_zero_init() == 1
+TEST_IMG="$TEST_IMG".orig _make_test_img 64M
+$QEMU_IO -c "write -P 42 0 64k" "$TEST_IMG".orig | _filter_qemu_io
+
+# Convert with -n, which should not assume that the target is zeroed
+$QEMU_IMG convert -O $IMGFMT -n "$TEST_IMG" "$TEST_IMG".orig
+
+$QEMU_IMG compare "$TEST_IMG" "$TEST_IMG".orig
+
 # success, all done
 echo '*** done'
 rm -f $seq.full
diff --git a/tests/qemu-iotests/122.out b/tests/qemu-iotests/122.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/122.out
+++ b/tests/qemu-iotests/122.out
@@ -XXX,XX +XXX,XX @@ convert -c -S 8k
 { "start": 9216, "length": 8192, "depth": 0, "zero": true, "data": false},
 { "start": 17408, "length": 1024, "depth": 0, "zero": false, "data": true},
 { "start": 18432, "length": 67090432, "depth": 0, "zero": true, "data": false}]
+
+=== -n to a non-zero image ===
+
+Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
+Formatting 'TEST_DIR/t.IMGFMT.orig', 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)
+Images are identical.
 *** done
-- 
2.21.0

The result of a sync=full mirror should always be the equal to the
input.  Therefore, existing images should be treated as potentially
non-zero and thus should be explicitly initialized to be zero
beforehand.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20190724171239.8764-12-mreitz@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 tests/qemu-iotests/041     | 62 +++++++++++++++++++++++++++++++++++---
 tests/qemu-iotests/041.out |  4 +--
 2 files changed, 60 insertions(+), 6 deletions(-)

diff --git a/tests/qemu-iotests/041 b/tests/qemu-iotests/041
index XXXXXXX..XXXXXXX 100755
--- a/tests/qemu-iotests/041
+++ b/tests/qemu-iotests/041
@@ -XXX,XX +XXX,XX @@ class TestUnbackedSource(iotests.QMPTestCase):
     def setUp(self):
         qemu_img('create', '-f', iotests.imgfmt, test_img,
                  str(TestUnbackedSource.image_len))
-        self.vm = iotests.VM().add_drive(test_img)
+        self.vm = iotests.VM()
         self.vm.launch()
+        result = self.vm.qmp('blockdev-add', node_name='drive0',
+                             driver=iotests.imgfmt,
+                             file={
+                                 'driver': 'file',
+                                 'filename': test_img,
+                             })
+        self.assert_qmp(result, 'return', {})
 
     def tearDown(self):
         self.vm.shutdown()
@@ -XXX,XX +XXX,XX @@ class TestUnbackedSource(iotests.QMPTestCase):
 
     def test_absolute_paths_full(self):
         self.assert_no_active_block_jobs()
-        result = self.vm.qmp('drive-mirror', device='drive0',
+        result = self.vm.qmp('drive-mirror', job_id='drive0', device='drive0',
                              sync='full', target=target_img,
                              mode='absolute-paths')
         self.assert_qmp(result, 'return', {})
@@ -XXX,XX +XXX,XX @@ class TestUnbackedSource(iotests.QMPTestCase):
 
     def test_absolute_paths_top(self):
         self.assert_no_active_block_jobs()
-        result = self.vm.qmp('drive-mirror', device='drive0',
+        result = self.vm.qmp('drive-mirror', job_id='drive0', device='drive0',
                              sync='top', target=target_img,
                              mode='absolute-paths')
         self.assert_qmp(result, 'return', {})
@@ -XXX,XX +XXX,XX @@ class TestUnbackedSource(iotests.QMPTestCase):
 
     def test_absolute_paths_none(self):
         self.assert_no_active_block_jobs()
-        result = self.vm.qmp('drive-mirror', device='drive0',
+        result = self.vm.qmp('drive-mirror', job_id='drive0', device='drive0',
                              sync='none', target=target_img,
                              mode='absolute-paths')
         self.assert_qmp(result, 'return', {})
         self.complete_and_wait()
         self.assert_no_active_block_jobs()
 
+    def test_existing_full(self):
+        qemu_img('create', '-f', iotests.imgfmt, target_img,
+                 str(self.image_len))
+        qemu_io('-c', 'write -P 42 0 64k', target_img)
+
+        self.assert_no_active_block_jobs()
+        result = self.vm.qmp('drive-mirror', job_id='drive0', device='drive0',
+                             sync='full', target=target_img, mode='existing')
+        self.assert_qmp(result, 'return', {})
+        self.complete_and_wait()
+        self.assert_no_active_block_jobs()
+
+        result = self.vm.qmp('blockdev-del', node_name='drive0')
+        self.assert_qmp(result, 'return', {})
+
+        self.assertTrue(iotests.compare_images(test_img, target_img),
+                        'target image does not match source after mirroring')
+
+    def test_blockdev_full(self):
+        qemu_img('create', '-f', iotests.imgfmt, target_img,
+                 str(self.image_len))
+        qemu_io('-c', 'write -P 42 0 64k', target_img)
+
+        result = self.vm.qmp('blockdev-add', node_name='target',
+                             driver=iotests.imgfmt,
+                             file={
+                                 'driver': 'file',
+                                 'filename': target_img,
+                             })
+        self.assert_qmp(result, 'return', {})
+
+        self.assert_no_active_block_jobs()
+        result = self.vm.qmp('blockdev-mirror', job_id='drive0', device='drive0',
+                             sync='full', target='target')
+        self.assert_qmp(result, 'return', {})
+        self.complete_and_wait()
+        self.assert_no_active_block_jobs()
+
+        result = self.vm.qmp('blockdev-del', node_name='drive0')
+        self.assert_qmp(result, 'return', {})
+
+        result = self.vm.qmp('blockdev-del', node_name='target')
+        self.assert_qmp(result, 'return', {})
+
+        self.assertTrue(iotests.compare_images(test_img, target_img),
+                        'target image does not match source after mirroring')
+
 class TestGranularity(iotests.QMPTestCase):
     image_len = 10 * 1024 * 1024 # MB
 
diff --git a/tests/qemu-iotests/041.out b/tests/qemu-iotests/041.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/041.out
+++ b/tests/qemu-iotests/041.out
@@ -XXX,XX +XXX,XX @@
-........................................................................................
+..........................................................................................
 ----------------------------------------------------------------------
-Ran 88 tests
+Ran 90 tests
 
 OK
-- 
2.21.0

vpc is not really a passthrough driver, even when using the fixed
subformat (where host and guest offsets are equal).  It should handle
preallocation like all other drivers do, namely by returning
DATA | RECURSE instead of RAW.

There is no tangible difference but the fact that bdrv_is_allocated() no
longer falls through to the protocol layer.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20190725155512.9827-4-mreitz@redhat.com
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block/vpc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/block/vpc.c b/block/vpc.c
index XXXXXXX..XXXXXXX 100644
--- a/block/vpc.c
+++ b/block/vpc.c
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn vpc_co_block_status(BlockDriverState *bs,
         *pnum = bytes;
         *map = offset;
         *file = bs->file->bs;
-        return BDRV_BLOCK_RAW | BDRV_BLOCK_OFFSET_VALID;
+        return BDRV_BLOCK_DATA | BDRV_BLOCK_OFFSET_VALID | BDRV_BLOCK_RECURSE;
     }
 
     qemu_co_mutex_lock(&s->lock);
-- 
2.21.0

69f47505ee has changed qcow2 in such a way that the commit job run in
test 141 (and 144[1]) returns before it emits the READY event.  However,
141 also runs with qed, where the order is still the other way around.
Just filter out the {"return": {}} so the test passes for qed again.

[1] 144 only runs with qcow2, so it is fine as it is.

Suggested-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Fixes: 69f47505ee66afaa513305de0c1895a224e52c45
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20190809185253.17535-1-mreitz@redhat.com
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 tests/qemu-iotests/141           | 9 +++++++--
 tests/qemu-iotests/141.out       | 5 -----
 tests/qemu-iotests/common.filter | 5 +++++
 3 files changed, 12 insertions(+), 7 deletions(-)

diff --git a/tests/qemu-iotests/141 b/tests/qemu-iotests/141
index XXXXXXX..XXXXXXX 100755
--- a/tests/qemu-iotests/141
+++ b/tests/qemu-iotests/141
@@ -XXX,XX +XXX,XX @@ test_blockjob()
               }}}" \
         'return'
 
+    # If "$2" is an event, we may or may not see it before the
+    # {"return": {}}.  Therefore, filter the {"return": {}} out both
+    # here and in the next command.  (Naturally, if we do not see it
+    # here, we will see it before the next command can be executed,
+    # so it will appear in the next _send_qemu_cmd's output.)
     _send_qemu_cmd $QEMU_HANDLE \
         "$1" \
         "$2" \
-        | _filter_img_create
+        | _filter_img_create | _filter_qmp_empty_return
 
     # We want this to return an error because the block job is still running
     _send_qemu_cmd $QEMU_HANDLE \
         "{'execute': 'blockdev-del',
           'arguments': {'node-name': 'drv0'}}" \
-        'error' | _filter_generated_node_ids
+        'error' | _filter_generated_node_ids | _filter_qmp_empty_return
 
     _send_qemu_cmd $QEMU_HANDLE \
         "{'execute': 'block-job-cancel',
diff --git a/tests/qemu-iotests/141.out b/tests/qemu-iotests/141.out
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/141.out
+++ b/tests/qemu-iotests/141.out
@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1048576 backing_file=TEST_DIR/m.
 Formatting 'TEST_DIR/o.IMGFMT', fmt=IMGFMT size=1048576 backing_file=TEST_DIR/t.IMGFMT backing_fmt=IMGFMT
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "job0"}}
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "job0"}}
-{"return": {}}
 {"error": {"class": "GenericError", "desc": "Node drv0 is in use"}}
 {"return": {}}
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "job0"}}
@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/o.IMGFMT', fmt=IMGFMT size=1048576 backing_file=TEST_DIR/t.
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "job0"}}
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "job0"}}
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "job0", "len": 0, "offset": 0, "speed": 0, "type": "mirror"}}
-{"return": {}}
 {"error": {"class": "GenericError", "desc": "Node 'drv0' is busy: block device is in use by block job: mirror"}}
 {"return": {}}
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "job0"}}
@@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/o.IMGFMT', fmt=IMGFMT size=1048576 backing_file=TEST_DIR/t.
 {"return": {}}
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "job0"}}
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "job0"}}
-{"return": {}}
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "job0"}}
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "job0", "len": 0, "offset": 0, "speed": 0, "type": "commit"}}
 {"error": {"class": "GenericError", "desc": "Node 'drv0' is busy: block device is in use by block job: commit"}}
@@ -XXX,XX +XXX,XX @@ wrote 1048576/1048576 bytes at offset 0
 {"return": {}}
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "job0"}}
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "job0"}}
-{"return": {}}
 {"error": {"class": "GenericError", "desc": "Node drv0 is in use"}}
 {"return": {}}
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "job0"}}
@@ -XXX,XX +XXX,XX @@ wrote 1048576/1048576 bytes at offset 0
 {"return": {}}
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "job0"}}
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "job0"}}
-{"return": {}}
 {"error": {"class": "GenericError", "desc": "Node drv0 is in use"}}
 {"return": {}}
 {"timestamp": {"seconds":  TIMESTAMP, "microseconds":  TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "job0"}}
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_nbd()
         -e 's#$foo\|PORT/\?\|.sock$: Failed to .*$#\1#'
 }
 
+_filter_qmp_empty_return()
+{
+    grep -v '{"return": {}}'
+}
+
 # make sure this script returns success
 true
-- 
2.21.0

When preallocating an encrypted qcow2 image, it just lets the protocol
driver write data and then does not mark the clusters as zero.
Therefore, reading this image will yield effectively random data.

As such, we have not fulfilled the promise of always writing zeroes when
preallocating an image in a while.  It seems that nobody has really
cared, so change the documentation to conform to qemu's actual behavior.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20190711132935.13070-1-mreitz@redhat.com
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 qapi/block-core.json         | 9 +++++----
 docs/qemu-block-drivers.texi | 4 ++--
 qemu-img.texi                | 4 ++--
 3 files changed, 9 insertions(+), 8 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 @@
 # @off: no preallocation
 # @metadata: preallocate only for metadata
 # @falloc: like @full preallocation but allocate disk space by
-#          posix_fallocate() rather than writing zeros.
-# @full: preallocate all data by writing zeros to device to ensure disk
-#        space is really available. @full preallocation also sets up
-#        metadata correctly.
+#          posix_fallocate() rather than writing data.
+# @full: preallocate all data by writing it to the device to ensure
+#        disk space is really available. This data may or may not be
+#        zero, depending on the image format and storage.
+#        @full preallocation also sets up metadata correctly.
 #
 # Since: 2.2
 ##
diff --git a/docs/qemu-block-drivers.texi b/docs/qemu-block-drivers.texi
index XXXXXXX..XXXXXXX 100644
--- a/docs/qemu-block-drivers.texi
+++ b/docs/qemu-block-drivers.texi
@@ -XXX,XX +XXX,XX @@ Supported options:
 @item preallocation
 Preallocation mode (allowed values: @code{off}, @code{falloc}, @code{full}).
 @code{falloc} mode preallocates space for image by calling posix_fallocate().
-@code{full} mode preallocates space for image by writing zeros to underlying
-storage.
+@code{full} mode preallocates space for image by writing data to underlying
+storage.  This data may or may not be zero, depending on the storage location.
 @end table
 
 @item qcow2
diff --git a/qemu-img.texi b/qemu-img.texi
index XXXXXXX..XXXXXXX 100644
--- a/qemu-img.texi
+++ b/qemu-img.texi
@@ -XXX,XX +XXX,XX @@ Supported options:
 @item preallocation
 Preallocation mode (allowed values: @code{off}, @code{falloc}, @code{full}).
 @code{falloc} mode preallocates space for image by calling posix_fallocate().
-@code{full} mode preallocates space for image by writing zeros to underlying
-storage.
+@code{full} mode preallocates space for image by writing data to underlying
+storage.  This data may or may not be zero, depending on the storage location.
 @end table
 
 @item qcow2
-- 
2.21.0

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