Series comparison

-[Qemu-devel] [PULL 00/16] Block patches
+[Qemu-devel] [PULL v2 00/24] Block patches
-The following changes since commit dd25f97c66a75d1508f1d4c6478ed2c95bec428f:
+The following changes since commit 56f9e46b841c7be478ca038d8d4085d776ab4b0d:
-  Merge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20190913' into staging (2019-09-16 10:15:15 +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-2019-09-16
+  git://github.com/stefanha/qemu.git tags/block-pull-request
-for you to fetch changes up to 1825cc0783ccf0ec5d9f0b225a99b340bdd4c68f:
+for you to fetch changes up to a7b91d35bab97a2d3e779d0c64c9b837b52a6cf7:
-  qemu-iotests: Add test for bz #1745922 (2019-09-16 15:37:12 +0200)
+  coroutine-lock: make CoRwlock thread-safe and fair (2017-02-21 11:39:40 +0000)
 ----------------------------------------------------------------
-Block patches:
+Pull request
-- Fix for block jobs when used with I/O threads
-- Fix for a corruption when using qcow2's LUKS encryption mode
+v2:
-- cURL fix
+ * Rebased to resolve scsi conflicts
 - check-block.sh cleanups (for make check)
 - Refactoring
 ----------------------------------------------------------------
-Max Reitz (7):
-  curl: Keep pointer to the CURLState in CURLSocket
-  curl: Keep *socket until the end of curl_sock_cb()
-  curl: Check completion in curl_multi_do()
-  curl: Pass CURLSocket to curl_multi_do()
-  curl: Report only ready sockets
-  curl: Handle success in multi_check_completion
-  curl: Check curl_multi_add_handle()'s return code
-Maxim Levitsky (3):
+Paolo Bonzini (24):
-  block/qcow2: Fix corruption introduced by commit 8ac0f15f335
+  block: move AioContext, QEMUTimer, main-loop to libqemuutil
-  block/qcow2: refactor encryption code
+  aio: introduce aio_co_schedule and aio_co_wake
-  qemu-iotests: Add test for bz #1745922
+  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
-Nir Soffer (2):
+ Makefile.objs                       |   4 -
-  block: Use QEMU_IS_ALIGNED
+ stubs/Makefile.objs                 |   1 +
-  block: Remove unused masks
+ tests/Makefile.include              |  19 +-
+ util/Makefile.objs                  |   6 +-
-Sergio Lopez (1):
+ block/nbd-client.h                  |   2 +-
-  blockjob: update nodes head while removing all bdrv
+ block/qed.h                         |   3 +
+ include/block/aio.h                 |  38 ++-
-Thomas Huth (2):
+ include/block/block_int.h           |  64 +++--
-  tests/qemu-iotests/check: Replace "tests" with "iotests" in final
+ include/io/channel.h                |  72 +++++-
-    status text
+ include/qemu/coroutine.h            |  84 ++++---
-  tests/Makefile: Do not print the name of the check-block.sh shell
+ include/qemu/coroutine_int.h        |  11 +-
-    script
+ include/sysemu/block-backend.h      |  14 +-
+ tests/iothread.h                    |  25 ++
-Vladimir Sementsov-Ogievskiy (1):
+ block/backup.c                      |   2 +-
-  tests/qemu-iotests: Fix qemu-io related output in 026.out.nocache
+ block/blkdebug.c                    |   9 +-
+ block/blkreplay.c                   |   2 +-
- tests/Makefile.include             |   2 +-
+ block/block-backend.c               |  13 +-
- block/qcow2.h                      |   8 +-
+ block/curl.c                        |  44 +++-
- include/block/block.h              |   2 -
+ block/gluster.c                     |   9 +-
- block/bochs.c                      |   4 +-
+ block/io.c                          |  42 +---
- block/cloop.c                      |   4 +-
+ block/iscsi.c                       |  15 +-
- block/curl.c                       | 133 ++++++++++-------------
+ block/linux-aio.c                   |  10 +-
- block/dmg.c                        |   4 +-
+ block/mirror.c                      |  12 +-
- block/io.c                         |   8 +-
+ block/nbd-client.c                  | 119 +++++----
- block/qcow2-cluster.c              |  40 +++----
+ block/nfs.c                         |   9 +-
- block/qcow2-threads.c              |  63 ++++++++---
+ block/qcow2-cluster.c               |   4 +-
- block/qcow2.c                      |   9 +-
+ block/qed-cluster.c                 |   2 +
- block/vvfat.c                      |   8 +-
+ block/qed-table.c                   |  12 +-
- blockjob.c                         |  17 ++-
+ block/qed.c                         |  58 +++--
- migration/block.c                  |   2 +-
+ block/sheepdog.c                    |  31 +--
- qemu-img.c                         |   2 +-
+ block/ssh.c                         |  29 +--
- tests/qemu-iotests/026.out.nocache | 168 ++++++++++++++---------------
+ block/throttle-groups.c             |   4 +-
- tests/qemu-iotests/263             |  91 ++++++++++++++++
+ block/win32-aio.c                   |   9 +-
- tests/qemu-iotests/263.out         |  40 +++++++
+ dma-helpers.c                       |   2 +
- tests/qemu-iotests/check           |   8 +-
+ hw/9pfs/9p.c                        |   2 +-
- tests/qemu-iotests/group           |   1 +
+ hw/block/virtio-blk.c               |  19 +-
-files changed, 380 insertions(+), 234 deletions(-)
+ hw/scsi/scsi-bus.c                  |   2 +
- create mode 100755 tests/qemu-iotests/263
+ hw/scsi/scsi-disk.c                 |  15 ++
- create mode 100644 tests/qemu-iotests/263.out
+ 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

-[Qemu-devel] [PULL 04/16] tests/Makefile: Do not print the name of the check-block.sh shell script
+[Qemu-devel] [PULL v2 02/24] aio: introduce aio_co_schedule and aio_co_wake
-From: Thomas Huth <thuth@redhat.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-The check script is already printing out which iotest is currently
+aio_co_wake provides the infrastructure to start a coroutine on a "home"
-running, so printing out the name of the check-block.sh shell script
+AioContext.  It will be used by CoMutex and CoQueue, so that coroutines
-looks superfluous here.
+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
-Signed-off-by: Thomas Huth <thuth@redhat.com>
+alternative to one-shot bottom halves, and saves the effort of tracking
-Message-id: 20190906113534.10907-1-thuth@redhat.com
+which AioContext a coroutine is running on.
-Acked-by: John Snow <jsnow@redhat.com>
-Reviewed-by: Max Reitz <mreitz@redhat.com>
+aio_co_schedule is the part of aio_co_wake that starts a coroutine
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+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 | 2 +-
+ tests/Makefile.include       |   8 +-
-file changed, 1 insertion(+), 1 deletion(-)
+ 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 @@ QEMU_IOTESTS_HELPERS-$(call land,$(CONFIG_SOFTMMU),$(CONFIG_LINUX)) = tests/qemu
+@@ -XXX,XX +XXX,XX @@ check-unit-y += tests/test-aio$(EXESUF)
- check-tests/check-block.sh: tests/check-block.sh qemu-img$(EXESUF) \
+ gcov-files-test-aio-y = util/async.c util/qemu-timer.o
-         qemu-io$(EXESUF) qemu-nbd$(EXESUF) $(QEMU_IOTESTS_HELPERS-y) \
+ gcov-files-test-aio-$(CONFIG_WIN32) += util/aio-win32.c
-         $(patsubst %,%/all,$(filter %-softmmu,$(TARGET_DIRS)))
+ gcov-files-test-aio-$(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
- .PHONY: $(patsubst %, check-%, $(check-qapi-schema-y))
+ check-unit-y += tests/test-throttle$(EXESUF)
- $(patsubst %, check-%, $(check-qapi-schema-y)): check-%.json: $(SRC_PATH)/%.json
+-gcov-files-test-aio-$(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"
 --
-.21.0
+.9.3

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

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

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

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

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

-[Qemu-devel] [PULL 05/16] tests/qemu-iotests: Fix qemu-io related output in 026.out.nocache
+[Qemu-devel] [PULL v2 08/24] coroutine-lock: reschedule coroutine on the AioContext it was running on
-From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-qemu-io now prefixes its error and warnings with "qemu-io:".
+As a small step towards the introduction of multiqueue, we want
-b9986b08787019e fixed a lot of iotests output but forget about
+coroutines to remain on the same AioContext that started them,
-.out.nocache. Fix it too.
+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.
-Fixes: 99e98d7c9fc1a1639fad ("qemu-io: Use error_[gs]et_progname()")
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-Message-id: 20190816153015.447957-2-vsementsov@virtuozzo.com
+Reviewed-by: Fam Zheng <famz@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: 20170213135235.12274-9-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
- tests/qemu-iotests/026.out.nocache | 168 ++++++++++++++---------------
+ util/qemu-coroutine-lock.c | 5 ++---
-file changed, 84 insertions(+), 84 deletions(-)
+ util/trace-events          | 1 -
 files changed, 2 insertions(+), 4 deletions(-)
-diff --git a/tests/qemu-iotests/026.out.nocache b/tests/qemu-iotests/026.out.nocache
+diff --git a/util/qemu-coroutine-lock.c b/util/qemu-coroutine-lock.c
 index XXXXXXX..XXXXXXX 100644
---- a/tests/qemu-iotests/026.out.nocache
+--- a/util/qemu-coroutine-lock.c
-+++ b/tests/qemu-iotests/026.out.nocache
++++ b/util/qemu-coroutine-lock.c
-@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
+@@ -XXX,XX +XXX,XX @@
- Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
+ #include "qemu/coroutine.h"
+ #include "qemu/coroutine_int.h"
- Event: l1_update; errno: 5; imm: off; once: off; write
+ #include "qemu/queue.h"
--Failed to flush the L2 table cache: Input/output error
++#include "block/aio.h"
--Failed to flush the refcount block cache: Input/output error
+ #include "trace.h"
-+qemu-io: Failed to flush the L2 table cache: Input/output error
-+qemu-io: Failed to flush the refcount block cache: Input/output error
+ void qemu_co_queue_init(CoQueue *queue)
- write failed: Input/output error
+@@ -XXX,XX +XXX,XX @@ void qemu_co_queue_run_restart(Coroutine *co)
-leaked clusters were found on the image.
+ static bool qemu_co_queue_do_restart(CoQueue *queue, bool single)
-@@ -XXX,XX +XXX,XX @@ This means waste of disk space, but no harm to data.
+ {
- Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
+-    Coroutine *self = qemu_coroutine_self();
+     Coroutine *next;
- Event: l1_update; errno: 5; imm: off; once: off; write -b
--Failed to flush the L2 table cache: Input/output error
+     if (QSIMPLEQ_EMPTY(&queue->entries)) {
--Failed to flush the refcount block cache: Input/output error
+@@ -XXX,XX +XXX,XX @@ static bool qemu_co_queue_do_restart(CoQueue *queue, bool single)
-+qemu-io: Failed to flush the L2 table cache: Input/output error
-+qemu-io: Failed to flush the refcount block cache: Input/output error
+     while ((next = QSIMPLEQ_FIRST(&queue->entries)) != NULL) {
- write failed: Input/output error
+         QSIMPLEQ_REMOVE_HEAD(&queue->entries, co_queue_next);
+-        QSIMPLEQ_INSERT_TAIL(&self->co_queue_wakeup, next, co_queue_next);
-leaked clusters were found on the image.
+-        trace_qemu_co_queue_next(next);
-@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
++        aio_co_wake(next);
- Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
+         if (single) {
+             break;
- Event: l1_update; errno: 28; imm: off; once: off; write
+         }
--Failed to flush the L2 table cache: No space left on device
+diff --git a/util/trace-events b/util/trace-events
--Failed to flush the refcount block cache: No space left on device
+index XXXXXXX..XXXXXXX 100644
-+qemu-io: Failed to flush the L2 table cache: No space left on device
+--- a/util/trace-events
-+qemu-io: Failed to flush the refcount block cache: No space left on device
++++ b/util/trace-events
- write failed: No space left on device
+@@ -XXX,XX +XXX,XX @@ qemu_coroutine_terminate(void *co) "self %p"
-leaked clusters were found on the image.
+ # util/qemu-coroutine-lock.c
-@@ -XXX,XX +XXX,XX @@ This means waste of disk space, but no harm to data.
+ qemu_co_queue_run_restart(void *co) "co %p"
- Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
+-qemu_co_queue_next(void *nxt) "next %p"
+ qemu_co_mutex_lock_entry(void *mutex, void *self) "mutex %p self %p"
- Event: l1_update; errno: 28; imm: off; once: off; write -b
+ qemu_co_mutex_lock_return(void *mutex, void *self) "mutex %p self %p"
--Failed to flush the L2 table cache: No space left on device
+ qemu_co_mutex_unlock_entry(void *mutex, void *self) "mutex %p self %p"
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: l2_update; errno: 5; imm: off; once: off; write
 -Failed to flush the L2 table cache: Input/output error
 -Failed to flush the refcount block cache: Input/output error
 +qemu-io: Failed to flush the L2 table cache: Input/output error
 +qemu-io: Failed to flush the refcount block cache: Input/output error
  wrote 131072/131072 bytes at offset 0
 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ This means waste of disk space, but no harm to data.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: l2_update; errno: 5; imm: off; once: off; write -b
 -Failed to flush the L2 table cache: Input/output error
 -Failed to flush the refcount block cache: Input/output error
 +qemu-io: Failed to flush the L2 table cache: Input/output error
 +qemu-io: Failed to flush the refcount block cache: Input/output error
  wrote 131072/131072 bytes at offset 0
 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: l2_update; errno: 28; imm: off; once: off; write
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  wrote 131072/131072 bytes at offset 0
 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ This means waste of disk space, but no harm to data.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: l2_update; errno: 28; imm: off; once: off; write -b
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  wrote 131072/131072 bytes at offset 0
 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: l2_alloc_write; errno: 5; imm: off; once: off; write
 -Failed to flush the L2 table cache: Input/output error
 -Failed to flush the refcount block cache: Input/output error
 +qemu-io: Failed to flush the L2 table cache: Input/output error
 +qemu-io: Failed to flush the refcount block cache: Input/output error
  write failed: Input/output error
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: l2_alloc_write; errno: 5; imm: off; once: off; write -b
 -Failed to flush the L2 table cache: Input/output error
 -Failed to flush the refcount block cache: Input/output error
 +qemu-io: Failed to flush the L2 table cache: Input/output error
 +qemu-io: Failed to flush the refcount block cache: Input/output error
  write failed: Input/output error
 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: l2_alloc_write; errno: 28; imm: off; once: off; write
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: l2_alloc_write; errno: 28; imm: off; once: off; write -b
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: write_aio; errno: 5; imm: off; once: off; write
 -Failed to flush the L2 table cache: Input/output error
 -Failed to flush the refcount block cache: Input/output error
 +qemu-io: Failed to flush the L2 table cache: Input/output error
 +qemu-io: Failed to flush the refcount block cache: Input/output error
  write failed: Input/output error
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: write_aio; errno: 5; imm: off; once: off; write -b
 -Failed to flush the L2 table cache: Input/output error
 -Failed to flush the refcount block cache: Input/output error
 +qemu-io: Failed to flush the L2 table cache: Input/output error
 +qemu-io: Failed to flush the refcount block cache: Input/output error
  write failed: Input/output error
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: write_aio; errno: 28; imm: off; once: off; write
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: write_aio; errno: 28; imm: off; once: off; write -b
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_load; errno: 5; imm: off; once: off; write
 -Failed to flush the L2 table cache: Input/output error
 -Failed to flush the refcount block cache: Input/output error
 +qemu-io: Failed to flush the L2 table cache: Input/output error
 +qemu-io: Failed to flush the refcount block cache: Input/output error
  write failed: Input/output error
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_load; errno: 5; imm: off; once: off; write -b
 -Failed to flush the L2 table cache: Input/output error
 -Failed to flush the refcount block cache: Input/output error
 +qemu-io: Failed to flush the L2 table cache: Input/output error
 +qemu-io: Failed to flush the refcount block cache: Input/output error
  write failed: Input/output error
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_load; errno: 28; imm: off; once: off; write
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_load; errno: 28; imm: off; once: off; write -b
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_update_part; errno: 5; imm: off; once: off; write
 -Failed to flush the L2 table cache: Input/output error
 -Failed to flush the refcount block cache: Input/output error
 +qemu-io: Failed to flush the L2 table cache: Input/output error
 +qemu-io: Failed to flush the refcount block cache: Input/output error
  write failed: Input/output error
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_update_part; errno: 5; imm: off; once: off; write -b
 -Failed to flush the L2 table cache: Input/output error
 -Failed to flush the refcount block cache: Input/output error
 +qemu-io: Failed to flush the L2 table cache: Input/output error
 +qemu-io: Failed to flush the refcount block cache: Input/output error
  write failed: Input/output error
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_update_part; errno: 28; imm: off; once: off; write
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_update_part; errno: 28; imm: off; once: off; write -b
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_alloc; errno: 5; imm: off; once: off; write
 -Failed to flush the L2 table cache: Input/output error
 -Failed to flush the refcount block cache: Input/output error
 +qemu-io: Failed to flush the L2 table cache: Input/output error
 +qemu-io: Failed to flush the refcount block cache: Input/output error
  write failed: Input/output error
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_alloc; errno: 5; imm: off; once: off; write -b
 -Failed to flush the L2 table cache: Input/output error
 -Failed to flush the refcount block cache: Input/output error
 +qemu-io: Failed to flush the L2 table cache: Input/output error
 +qemu-io: Failed to flush the refcount block cache: Input/output error
  write failed: Input/output error
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_alloc; errno: 28; imm: off; once: off; write
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_alloc; errno: 28; imm: off; once: off; write -b
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_alloc_hookup; errno: 28; imm: off; once: off; write
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ This means waste of disk space, but no harm to data.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_alloc_hookup; errno: 28; imm: off; once: off; write -b
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_alloc_write; errno: 28; imm: off; once: off; write
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_alloc_write; errno: 28; imm: off; once: off; write -b
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_alloc_write_blocks; errno: 28; imm: off; once: off; write
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ This means waste of disk space, but no harm to data.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_alloc_write_blocks; errno: 28; imm: off; once: off; write -b
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_alloc_write_table; errno: 28; imm: off; once: off; write
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ This means waste of disk space, but no harm to data.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_alloc_write_table; errno: 28; imm: off; once: off; write -b
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_alloc_switch_table; errno: 28; imm: off; once: off; write
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ This means waste of disk space, but no harm to data.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: refblock_alloc_switch_table; errno: 28; imm: off; once: off; write -b
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: l1_grow_write_table; errno: 5; imm: off; once: off
 -Failed to flush the L2 table cache: Input/output error
 -Failed to flush the refcount block cache: Input/output error
 +qemu-io: Failed to flush the L2 table cache: Input/output error
 +qemu-io: Failed to flush the refcount block cache: Input/output error
  write failed: Input/output error
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: l1_grow_write_table; errno: 28; imm: off; once: off
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
  No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: l1_grow_activate_table; errno: 5; imm: off; once: off
 -Failed to flush the L2 table cache: Input/output error
 -Failed to flush the refcount block cache: Input/output error
 +qemu-io: Failed to flush the L2 table cache: Input/output error
 +qemu-io: Failed to flush the refcount block cache: Input/output error
  write failed: Input/output error
 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
  Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
  Event: l1_grow_activate_table; errno: 28; imm: off; once: off
 -Failed to flush the L2 table cache: No space left on device
 -Failed to flush the refcount block cache: No space left on device
 +qemu-io: Failed to flush the L2 table cache: No space left on device
 +qemu-io: Failed to flush the refcount block cache: No space left on device
  write failed: No space left on device
 leaked clusters were found on the image.
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 03/16] tests/qemu-iotests/check: Replace "tests" with "iotests" in final status text
+[Qemu-devel] [PULL v2 09/24] blkdebug: reschedule coroutine on the AioContext it is running on
-From: Thomas Huth <thuth@redhat.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-When running "make check -j8" or something similar, the iotests are
+Keep the coroutine on the same AioContext.  Without this change,
-running in parallel with the other tests. So when they are printing
+there would be a race between yielding the coroutine and reentering it.
-out "Passed all xx tests" or a similar status message at the end,
+While the race cannot happen now, because the code only runs from a single
-it might not be quite clear that this message belongs to the iotests,
+AioContext, this will change with multiqueue support in the block layer.
 since the output might be mixed with the other tests. Thus change the
 word "tests" here to "iotests" instead to avoid confusion.
-Signed-off-by: Thomas Huth <thuth@redhat.com>
+While doing the change, replace custom bottom half with aio_co_schedule.
-Message-id: 20190906113920.11271-1-thuth@redhat.com
-Reviewed-by: John Snow <jsnow@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-Signed-off-by: Max Reitz <mreitz@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>
 ---
- tests/qemu-iotests/check | 8 ++++----
+ block/blkdebug.c | 9 +--------
-file changed, 4 insertions(+), 4 deletions(-)
+file changed, 1 insertion(+), 8 deletions(-)
-diff --git a/tests/qemu-iotests/check b/tests/qemu-iotests/check
+diff --git a/block/blkdebug.c b/block/blkdebug.c
-index XXXXXXX..XXXXXXX 100755
+index XXXXXXX..XXXXXXX 100644
---- a/tests/qemu-iotests/check
+--- a/block/blkdebug.c
-+++ b/tests/qemu-iotests/check
++++ b/block/blkdebug.c
-@@ -XXX,XX +XXX,XX @@ END        { if (NR > 0) {
+@@ -XXX,XX +XXX,XX @@ out:
-         if [ ! -z "$n_bad" -a $n_bad != 0 ]
+     return ret;
-         then
+ }
-             echo "Failures:$bad"
--            echo "Failed $n_bad of $try tests"
+-static void error_callback_bh(void *opaque)
-+            echo "Failed $n_bad of $try iotests"
+-{
-             echo "Failures:$bad" | fmt >>check.log
+-    Coroutine *co = opaque;
--            echo "Failed $n_bad of $try tests" >>check.log
+-    qemu_coroutine_enter(co);
-+            echo "Failed $n_bad of $try iotests" >>check.log
+-}
-         else
+-
--            echo "Passed all $try tests"
+ static int inject_error(BlockDriverState *bs, BlkdebugRule *rule)
--            echo "Passed all $try tests" >>check.log
+ {
-+            echo "Passed all $try iotests"
+     BDRVBlkdebugState *s = bs->opaque;
-+            echo "Passed all $try iotests" >>check.log
+@@ -XXX,XX +XXX,XX @@ static int inject_error(BlockDriverState *bs, BlkdebugRule *rule)
-         fi
+     }
-         needwrap=false
-     fi
+     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 10/16] curl: Report only ready sockets
+[Qemu-devel] [PULL v2 10/24] qed: introduce qed_aio_start_io and qed_aio_next_io_cb
-Instead of reporting all sockets to cURL, only report the one that has
+From: Paolo Bonzini <pbonzini@redhat.com>
 caused curl_multi_do_locked() to be called.  This lets us get rid of the
 QLIST_FOREACH_SAFE() list, which was actually wrong: SAFE foreaches are
 only safe when the current element is removed in each iteration.  If it
 possible for the list to be concurrently modified, we cannot guarantee
 that only the current element will be removed.  Therefore, we must not
 use QLIST_FOREACH_SAFE() here.
-Fixes: ff5ca1664af85b24a4180d595ea6873fd3deac57
+qed_aio_start_io and qed_aio_next_io will not have to acquire/release
-Cc: qemu-stable@nongnu.org
+the AioContext, while qed_aio_next_io_cb will.  Split the functionality
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+and gain a little type-safety in the process.
-Message-id: 20190910124136.10565-6-mreitz@redhat.com
-Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Reviewed-by: John Snow <jsnow@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+Reviewed-by: Fam Zheng <famz@redhat.com>
 Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
 Message-id: 20170213135235.12274-11-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
- block/curl.c | 17 ++++++-----------
+ block/qed.c | 39 +++++++++++++++++++++++++--------------
-file changed, 6 insertions(+), 11 deletions(-)
+file changed, 25 insertions(+), 14 deletions(-)
-diff --git a/block/curl.c b/block/curl.c
+diff --git a/block/qed.c b/block/qed.c
 index XXXXXXX..XXXXXXX 100644
---- a/block/curl.c
+--- a/block/qed.c
-+++ b/block/curl.c
++++ b/block/qed.c
-@@ -XXX,XX +XXX,XX @@ static void curl_multi_check_completion(BDRVCURLState *s)
+@@ -XXX,XX +XXX,XX @@ static CachedL2Table *qed_new_l2_table(BDRVQEDState *s)
      return l2_table;
  }
- /* Called with s->mutex held.  */
+-static void qed_aio_next_io(void *opaque, int ret);
--static void curl_multi_do_locked(CURLSocket *ready_socket)
++static void qed_aio_next_io(QEDAIOCB *acb, int ret);
-+static void curl_multi_do_locked(CURLSocket *socket)
++
 +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)
  {
--    CURLSocket *socket, *next_socket;
+@@ -XXX,XX +XXX,XX @@ static void qed_unplug_allocating_write_reqs(BDRVQEDState *s)
--    CURLState *s = ready_socket->state;
-+    BDRVCURLState *s = socket->state->s;
+     acb = QSIMPLEQ_FIRST(&s->allocating_write_reqs);
-     int running;
+     if (acb) {
-     int r;
+-        qed_aio_next_io(acb, 0);
++        qed_aio_start_io(acb);
--    if (!s->s->multi) {
+     }
-+    if (!s->multi) {
+ }
@@ -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;
      }
--    /* Need to use _SAFE because curl_multi_socket_action() may trigger
+     BLKDBG_EVENT(bs->file, BLKDBG_READ_AIO);
--     * curl_sock_cb() which might modify this list */
+     bdrv_aio_readv(bs->file, offset / BDRV_SECTOR_SIZE,
--    QLIST_FOREACH_SAFE(socket, &s->sockets, next, next_socket) {
+                    &acb->cur_qiov, acb->cur_qiov.size / BDRV_SECTOR_SIZE,
--        do {
+-                   qed_aio_next_io, acb);
--            r = curl_multi_socket_action(s->s->multi, socket->fd, 0, &running);
++                   qed_aio_next_io_cb, acb);
--        } while (r == CURLM_CALL_MULTI_PERFORM);
+     return;
--    }
-+    do {
+ err:
-+        r = curl_multi_socket_action(s->multi, socket->fd, 0, &running);
+@@ -XXX,XX +XXX,XX @@ err:
-+    } while (r == CURLM_CALL_MULTI_PERFORM);
+ /**
   * 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;
  }
- static void curl_multi_do(void *arg)
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 13/16] blockjob: update nodes head while removing all bdrv
+[Qemu-devel] [PULL v2 11/24] aio: push aio_context_acquire/release down to dispatching
-From: Sergio Lopez <slp@redhat.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-block_job_remove_all_bdrv() iterates through job->nodes, calling
+The AioContext data structures are now protected by list_lock and/or
-bdrv_root_unref_child() for each entry. The call to the latter may
+they are walked with FOREACH_RCU primitives.  There is no need anymore
-reach child_job_[can_]set_aio_ctx(), which will also attempt to
+to acquire the AioContext for the entire duration of aio_dispatch.
-traverse job->nodes, potentially finding entries that where freed
+Instead, just acquire it before and after invoking the callbacks.
-on previous iterations.
+The next step is then to push it further down.
-To avoid this situation, update job->nodes head on each iteration to
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-ensure that already freed entries are no longer linked to the list.
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
 Reviewed-by: Fam Zheng <famz@redhat.com>
 Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
 Message-id: 20170213135235.12274-12-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
  util/aio-posix.c | 25 +++++++++++--------------
  util/aio-win32.c | 15 +++++++--------
  util/async.c     |  2 ++
 files changed, 20 insertions(+), 22 deletions(-)
-RHBZ: https://bugzilla.redhat.com/show_bug.cgi?id=1746631
+diff --git a/util/aio-posix.c b/util/aio-posix.c
 Signed-off-by: Sergio Lopez <slp@redhat.com>
 Cc: qemu-stable@nongnu.org
 Signed-off-by: Max Reitz <mreitz@redhat.com>
 Message-id: 20190911100316.32282-1-mreitz@redhat.com
 Reviewed-by: Sergio Lopez <slp@redhat.com>
 Signed-off-by: Max Reitz <mreitz@redhat.com>
 ---
  blockjob.c | 17 +++++++++++++----
 file changed, 13 insertions(+), 4 deletions(-)
 diff --git a/blockjob.c b/blockjob.c
 index XXXXXXX..XXXXXXX 100644
---- a/blockjob.c
+--- a/util/aio-posix.c
-+++ b/blockjob.c
++++ b/util/aio-posix.c
-@@ -XXX,XX +XXX,XX @@ static const BdrvChildRole child_job = {
+@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
+             (revents & (G_IO_IN | G_IO_HUP | G_IO_ERR)) &&
- void block_job_remove_all_bdrv(BlockJob *job)
+             aio_node_check(ctx, node->is_external) &&
- {
+             node->io_read) {
--    GSList *l;
++            aio_context_acquire(ctx);
--    for (l = job->nodes; l; l = l->next) {
+             node->io_read(node->opaque);
-+    /*
++            aio_context_release(ctx);
-+     * bdrv_root_unref_child() may reach child_job_[can_]set_aio_ctx(),
-+     * which will also traverse job->nodes, so consume the list one by
+             /* aio_notify() does not count as progress */
-+     * one to make sure that such a concurrent access does not attempt
+             if (node->opaque != &ctx->notifier) {
-+     * to process an already freed BdrvChild.
+@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
-+     */
+             (revents & (G_IO_OUT | G_IO_ERR)) &&
-+    while (job->nodes) {
+             aio_node_check(ctx, node->is_external) &&
-+        GSList *l = job->nodes;
+             node->io_write) {
-         BdrvChild *c = l->data;
++            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);
 +
-+        job->nodes = l->next;
++    if (!progress) {
-+
+         assert(npfd == 0);
-         bdrv_op_unblock_all(c->bs, job->blocker);
-         bdrv_root_unref_child(c);
+         /* fill pollfds */
-+
+@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
-+        g_slist_free_1(l);
+         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);
 -        }
      }
--    g_slist_free(job->nodes);
--    job->nodes = NULL;
+     if (blocking) {
@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
          progress = true;
      }
 -    aio_context_release(ctx);
 -
      return progress;
  }
- bool block_job_has_bdrv(BlockJob *job, BlockDriverState *bs)
+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

-[Qemu-devel] [PULL 11/16] curl: Handle success in multi_check_completion
+[Qemu-devel] [PULL v2 12/24] block: explicitly acquire aiocontext in timers that need it
-Background: As of cURL 7.59.0, it verifies that several functions are
+From: Paolo Bonzini <pbonzini@redhat.com>
-not called from within a callback.  Among these functions is
-curl_multi_add_handle().
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-curl_read_cb() is a callback from cURL and not a coroutine.  Waking up
+Reviewed-by: Fam Zheng <famz@redhat.com>
-acb->co will lead to entering it then and there, which means the current
+Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
-request will settle and the caller (if it runs in the same coroutine)
+Message-id: 20170213135235.12274-13-pbonzini@redhat.com
-may then issue the next request.  In such a case, we will enter
+Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 curl_setup_preadv() effectively from within curl_read_cb().
 Calling curl_multi_add_handle() will then fail and the new request will
 not be processed.
 Fix this by not letting curl_read_cb() wake up acb->co.  Instead, leave
 the whole business of settling the AIOCB objects to
 curl_multi_check_completion() (which is called from our timer callback
 and our FD handler, so not from any cURL callbacks).
 Reported-by: Natalie Gavrielov <ngavrilo@redhat.com>
 Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1740193
 Cc: qemu-stable@nongnu.org
 Signed-off-by: Max Reitz <mreitz@redhat.com>
 Message-id: 20190910124136.10565-7-mreitz@redhat.com
 Reviewed-by: John Snow <jsnow@redhat.com>
 Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
 Signed-off-by: Max Reitz <mreitz@redhat.com>
 ---
- block/curl.c | 69 ++++++++++++++++++++++------------------------------
+ block/qed.h                 |  3 +++
-file changed, 29 insertions(+), 40 deletions(-)
+ 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 size_t curl_read_cb(void *ptr, size_t size, size_t nmemb, void *opaque)
+@@ -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)
  {
-     CURLState *s = ((CURLState*)opaque);
+     IscsiLun *iscsilun = opaque;
-     size_t realsize = size * nmemb;
--    int i;
++    aio_context_acquire(iscsilun->aio_context);
+     if (iscsi_get_nops_in_flight(iscsilun->iscsi) >= MAX_NOP_FAILURES) {
-     trace_curl_read_cb(realsize);
+         error_report("iSCSI: NOP timeout. Reconnecting...");
+         iscsilun->request_timed_out = true;
-@@ -XXX,XX +XXX,XX @@ static size_t curl_read_cb(void *ptr, size_t size, size_t nmemb, void *opaque)
+     } else if (iscsi_nop_out_async(iscsilun->iscsi, NULL, NULL, 0, NULL) != 0) {
-     memcpy(s->orig_buf + s->buf_off, ptr, realsize);
+         error_report("iSCSI: failed to sent NOP-Out. Disabling NOP messages.");
-     s->buf_off += realsize;
+-        return;
++        goto out;
--    for(i=0; i<CURL_NUM_ACB; i++) {
+     }
--        CURLAIOCB *acb = s->acb[i];
--
+     timer_mod(iscsilun->nop_timer, qemu_clock_get_ms(QEMU_CLOCK_REALTIME) + NOP_INTERVAL);
--        if (!acb)
+     iscsi_set_events(iscsilun);
--            continue;
++
--
++out:
--        if ((s->buf_off >= acb->end)) {
++    aio_context_release(iscsilun->aio_context);
--            size_t request_length = acb->bytes;
+ }
--
--            qemu_iovec_from_buf(acb->qiov, 0, s->orig_buf + acb->start,
+ static void iscsi_readcapacity_sync(IscsiLun *iscsilun, Error **errp)
--                                acb->end - acb->start);
+diff --git a/block/null.c b/block/null.c
--
+index XXXXXXX..XXXXXXX 100644
--            if (acb->end - acb->start < request_length) {
+--- a/block/null.c
--                size_t offset = acb->end - acb->start;
++++ b/block/null.c
--                qemu_iovec_memset(acb->qiov, offset, 0,
+@@ -XXX,XX +XXX,XX @@ static void null_bh_cb(void *opaque)
--                                  request_length - offset);
+ static void null_timer_cb(void *opaque)
--            }
+ {
--
+     NullAIOCB *acb = opaque;
--            acb->ret = 0;
++    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
--            s->acb[i] = NULL;
++
--            qemu_mutex_unlock(&s->s->mutex);
++    aio_context_acquire(ctx);
--            aio_co_wake(acb->co);
+     acb->common.cb(acb->common.opaque, 0);
--            qemu_mutex_lock(&s->s->mutex);
++    aio_context_release(ctx);
--        }
+     timer_deinit(&acb->timer);
--    }
+     qemu_aio_unref(acb);
--
+ }
- read_end:
+diff --git a/block/qed.c b/block/qed.c
-     /* curl will error out if we do not return this value */
+index XXXXXXX..XXXXXXX 100644
-     return size * nmemb;
+--- a/block/qed.c
-@@ -XXX,XX +XXX,XX @@ static void curl_multi_check_completion(BDRVCURLState *s)
++++ b/block/qed.c
-             break;
+@@ -XXX,XX +XXX,XX @@ static void qed_need_check_timer_cb(void *opaque)
-         if (msg->msg == CURLMSG_DONE) {
+     trace_qed_need_check_timer_cb(s);
-+            int i;
-             CURLState *state = NULL;
++    qed_acquire(s);
-+            bool error = msg->data.result != CURLE_OK;
+     qed_plug_allocating_write_reqs(s);
-+
-             curl_easy_getinfo(msg->easy_handle, CURLINFO_PRIVATE,
+     /* Ensure writes are on disk before clearing flag */
-                               (char **)&state);
+     bdrv_aio_flush(s->bs->file->bs, qed_clear_need_check, s);
++    qed_release(s);
--            /* ACBs for successful messages get completed in curl_read_cb */
++}
--            if (msg->data.result != CURLE_OK) {
++
--                int i;
++void qed_acquire(BDRVQEDState *s)
-+            if (error) {
++{
-                 static int errcount = 100;
++    aio_context_acquire(bdrv_get_aio_context(s->bs));
++}
-                 /* Don't lose the original error message from curl, since
++
-@@ -XXX,XX +XXX,XX @@ static void curl_multi_check_completion(BDRVCURLState *s)
++void qed_release(BDRVQEDState *s)
-                         error_report("curl: further errors suppressed");
++{
-                     }
++    aio_context_release(bdrv_get_aio_context(s->bs));
-                 }
+ }
-+            }
+ static void qed_start_need_check_timer(BDRVQEDState *s)
--                for (i = 0; i < CURL_NUM_ACB; i++) {
+diff --git a/block/throttle-groups.c b/block/throttle-groups.c
--                    CURLAIOCB *acb = state->acb[i];
+index XXXXXXX..XXXXXXX 100644
-+            for (i = 0; i < CURL_NUM_ACB; i++) {
+--- a/block/throttle-groups.c
-+                CURLAIOCB *acb = state->acb[i];
++++ b/block/throttle-groups.c
+@@ -XXX,XX +XXX,XX @@ static void timer_cb(BlockBackend *blk, bool is_write)
--                    if (acb == NULL) {
+     qemu_mutex_unlock(&tg->lock);
--                        continue;
--                    }
+     /* Run the request that was waiting for this timer */
-+                if (acb == NULL) {
++    aio_context_acquire(blk_get_aio_context(blk));
-+                    continue;
+     empty_queue = !qemu_co_enter_next(&blkp->throttled_reqs[is_write]);
-+                }
++    aio_context_release(blk_get_aio_context(blk));
-+
-+                if (!error) {
+     /* If the request queue was empty then we have to take care of
-+                    /* Assert that we have read all data */
+      * scheduling the next one */
-+                    assert(state->buf_off >= acb->end);
+diff --git a/util/aio-posix.c b/util/aio-posix.c
-+
+index XXXXXXX..XXXXXXX 100644
-+                    qemu_iovec_from_buf(acb->qiov, 0,
+--- a/util/aio-posix.c
-+                                        state->orig_buf + acb->start,
++++ b/util/aio-posix.c
-+                                        acb->end - acb->start);
+@@ -XXX,XX +XXX,XX @@ bool aio_dispatch(AioContext *ctx, bool dispatch_fds)
+     }
--                    acb->ret = -EIO;
--                    state->acb[i] = NULL;
+     /* Run our timers */
--                    qemu_mutex_unlock(&s->mutex);
+-    aio_context_acquire(ctx);
--                    aio_co_wake(acb->co);
+     progress |= timerlistgroup_run_timers(&ctx->tlg);
--                    qemu_mutex_lock(&s->mutex);
+-    aio_context_release(ctx);
-+                    if (acb->end - acb->start < acb->bytes) {
-+                        size_t offset = acb->end - acb->start;
+     return progress;
-+                        qemu_iovec_memset(acb->qiov, offset, 0,
+ }
-+                                          acb->bytes - offset);
+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
-+                acb->ret = error ? -EIO : 0;
+@@ -XXX,XX +XXX,XX @@ bool aio_poll(AioContext *ctx, bool blocking)
-+                state->acb[i] = NULL;
+         progress |= aio_dispatch_handlers(ctx, event);
-+                qemu_mutex_unlock(&s->mutex);
+     } while (count > 0);
-+                aio_co_wake(acb->co);
-+                qemu_mutex_lock(&s->mutex);
+-    aio_context_acquire(ctx);
-             }
+     progress |= timerlistgroup_run_timers(&ctx->tlg);
+-    aio_context_release(ctx);
-             curl_clean_state(state);
+     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,
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 08/16] curl: Check completion in curl_multi_do()
+[Qemu-devel] [PULL v2 13/24] block: explicitly acquire aiocontext in callbacks that need it
-While it is more likely that transfers complete after some file
+From: Paolo Bonzini <pbonzini@redhat.com>
 descriptor has data ready to read, we probably should not rely on it.
 Better be safe than sorry and call curl_multi_check_completion() in
 curl_multi_do(), too, just like it is done in curl_multi_read().
-With this change, curl_multi_do() and curl_multi_read() are actually the
+This covers both file descriptor callbacks and polling callbacks,
-same, so drop curl_multi_read() and use curl_multi_do() as the sole FD
+since they execute related code.
 handler.
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Message-id: 20190910124136.10565-4-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>
-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: 20170213135235.12274-14-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
- block/curl.c | 14 ++------------
+ block/curl.c          | 16 +++++++++++++---
-file changed, 2 insertions(+), 12 deletions(-)
+ 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(-)
 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 @@ typedef struct BDRVCURLState {
+@@ -XXX,XX +XXX,XX @@ static void curl_multi_check_completion(BDRVCURLState *s)
+     }
- static void curl_clean_state(CURLState *s);
+ }
- static void curl_multi_do(void *arg);
--static void curl_multi_read(void *arg);
+-static void curl_multi_do(void *arg)
++static void curl_multi_do_locked(CURLState *s)
- #ifdef NEED_CURL_TIMER_CALLBACK
+ {
- /* Called from curl_multi_do_locked, with s->mutex held.  */
+-    CURLState *s = (CURLState *)arg;
-@@ -XXX,XX +XXX,XX @@ static int curl_sock_cb(CURL *curl, curl_socket_t fd, int action,
+     CURLSocket *socket, *next_socket;
-     switch (action) {
+     int running;
-         case CURL_POLL_IN:
+     int r;
              aio_set_fd_handler(s->aio_context, fd, false,
 -                               curl_multi_read, NULL, NULL, state);
 +                               curl_multi_do, NULL, NULL, state);
              break;
          case CURL_POLL_OUT:
              aio_set_fd_handler(s->aio_context, fd, false,
@@ -XXX,XX +XXX,XX @@ static int curl_sock_cb(CURL *curl, curl_socket_t fd, int action,
              break;
          case CURL_POLL_INOUT:
              aio_set_fd_handler(s->aio_context, fd, false,
 -                               curl_multi_read, curl_multi_do, NULL, state);
 +                               curl_multi_do, curl_multi_do, NULL, state);
              break;
          case CURL_POLL_REMOVE:
              aio_set_fd_handler(s->aio_context, fd, false,
 @@ -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;
--    qemu_mutex_lock(&s->s->mutex);
+-    curl_multi_do(arg);
--    curl_multi_do_locked(s);
++    aio_context_acquire(s->s->aio_context);
--    qemu_mutex_unlock(&s->s->mutex);
++    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);
 -}
 -
--static void curl_multi_read(void *arg)
+ 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)
 -{
--    CURLState *s = (CURLState *)arg;
+-    DPRINTF("s->sock=%d", s->sock);
 -    aio_set_fd_handler(bdrv_get_aio_context(bs), s->sock,
 -                       false, NULL, NULL, NULL, NULL);
 -}
 -
-     qemu_mutex_lock(&s->s->mutex);
+-/* A non-blocking call returned EAGAIN, so yield, ensuring the
-     curl_multi_do_locked(s);
+- * handlers are set up so that we'll be rescheduled when there is an
-     curl_multi_check_completion(s->s);
+- * 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 12/16] curl: Check curl_multi_add_handle()'s return code
+[Qemu-devel] [PULL v2 14/24] block: explicitly acquire aiocontext in bottom halves that need it
-If we had done that all along, debugging would have been much simpler.
+From: Paolo Bonzini <pbonzini@redhat.com>
 (Also, I/O errors are better than hangs.)
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Message-id: 20190910124136.10565-8-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>
-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: 20170213135235.12274-15-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
- block/curl.c | 8 +++++++-
+ block/archipelago.c   |  3 +++
-file changed, 7 insertions(+), 1 deletion(-)
+ block/blkreplay.c     |  2 +-
  block/block-backend.c |  6 ++++++
  block/curl.c          | 26 ++++++++++++++++++--------
  block/gluster.c       |  9 +--------
  block/io.c            |  6 +++++-
  block/iscsi.c         |  6 +++++-
  block/linux-aio.c     | 15 +++++++++------
  block/nfs.c           |  3 ++-
  block/null.c          |  4 ++++
  block/qed.c           |  3 +++
  block/rbd.c           |  4 ++++
  dma-helpers.c         |  2 ++
  hw/block/virtio-blk.c |  2 ++
  hw/scsi/scsi-bus.c    |  2 ++
  util/async.c          |  4 ++--
  util/thread-pool.c    |  2 ++
 files changed, 71 insertions(+), 28 deletions(-)
+diff --git a/block/archipelago.c b/block/archipelago.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/archipelago.c
++++ b/block/archipelago.c
+@@ -XXX,XX +XXX,XX @@ static void qemu_archipelago_complete_aio(void *opaque)
+ {
+     AIORequestData *reqdata = (AIORequestData *) opaque;
+     ArchipelagoAIOCB *aio_cb = (ArchipelagoAIOCB *) reqdata->aio_cb;
++    AioContext *ctx = bdrv_get_aio_context(aio_cb->common.bs);
++    aio_context_acquire(ctx);
+     aio_cb->common.cb(aio_cb->common.opaque, aio_cb->ret);
++    aio_context_release(ctx);
+     aio_cb->status = 0;
+     qemu_aio_unref(aio_cb);
+diff --git a/block/blkreplay.c b/block/blkreplay.c
+index XXXXXXX..XXXXXXX 100755
+--- a/block/blkreplay.c
++++ b/block/blkreplay.c
+@@ -XXX,XX +XXX,XX @@ static int64_t blkreplay_getlength(BlockDriverState *bs)
+ static void blkreplay_bh_cb(void *opaque)
+ {
+     Request *req = opaque;
+-    qemu_coroutine_enter(req->co);
++    aio_co_wake(req->co);
+     qemu_bh_delete(req->bh);
+     g_free(req);
+ }
+diff --git a/block/block-backend.c b/block/block-backend.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/block-backend.c
++++ b/block/block-backend.c
+@@ -XXX,XX +XXX,XX @@ int blk_make_zero(BlockBackend *blk, BdrvRequestFlags flags)
+ static void error_callback_bh(void *opaque)
+ {
+     struct BlockBackendAIOCB *acb = opaque;
++    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
+     bdrv_dec_in_flight(acb->common.bs);
++    aio_context_acquire(ctx);
+     acb->common.cb(acb->common.opaque, acb->ret);
++    aio_context_release(ctx);
+     qemu_aio_unref(acb);
+ }
+@@ -XXX,XX +XXX,XX @@ static void blk_aio_complete(BlkAioEmAIOCB *acb)
+ static void blk_aio_complete_bh(void *opaque)
+ {
+     BlkAioEmAIOCB *acb = opaque;
++    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
+     assert(acb->has_returned);
++    aio_context_acquire(ctx);
+     blk_aio_complete(acb);
++    aio_context_release(ctx);
+ }
+ static BlockAIOCB *blk_aio_prwv(BlockBackend *blk, int64_t offset, int bytes,
 diff --git a/block/curl.c b/block/curl.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/curl.c
 +++ b/block/curl.c
-@@ -XXX,XX +XXX,XX @@ static void curl_setup_preadv(BlockDriverState *bs, CURLAIOCB *acb)
+@@ -XXX,XX +XXX,XX @@ static void curl_readv_bh_cb(void *p)
-     trace_curl_setup_preadv(acb->bytes, start, state->range);
+ {
-     curl_easy_setopt(state->curl, CURLOPT_RANGE, state->range);
+     CURLState *state;
+     int running;
--    curl_multi_add_handle(s->multi, state->curl);
++    int ret = -EINPROGRESS;
-+    if (curl_multi_add_handle(s->multi, state->curl) != CURLM_OK) {
-+        state->acb[0] = NULL;
+     CURLAIOCB *acb = p;
-+        acb->ret = -EIO;
+-    BDRVCURLState *s = acb->common.bs->opaque;
-+
++    BlockDriverState *bs = acb->common.bs;
-+        curl_clean_state(state);
++    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)
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 06/16] curl: Keep pointer to the CURLState in CURLSocket
+[Qemu-devel] [PULL v2 15/24] block: explicitly acquire aiocontext in aio callbacks that need it
-A follow-up patch will make curl_multi_do() and curl_multi_read() take a
+From: Paolo Bonzini <pbonzini@redhat.com>
 CURLSocket instead of the CURLState.  They still need the latter,
 though, so add a pointer to it to the former.
-Cc: qemu-stable@nongnu.org
+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: John Snow <jsnow@redhat.com>
+Reviewed-by: Fam Zheng <famz@redhat.com>
-Message-id: 20190910124136.10565-2-mreitz@redhat.com
+Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
-Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
+Message-id: 20170213135235.12274-16-pbonzini@redhat.com
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
- block/curl.c | 3 +++
+ block/archipelago.c    |  3 ---
-file changed, 3 insertions(+)
+ block/block-backend.c  |  7 -------
  block/curl.c           |  2 +-
  block/io.c             |  6 +-----
  block/iscsi.c          |  3 ---
  block/linux-aio.c      |  5 +----
  block/mirror.c         | 12 +++++++++---
  block/null.c           |  8 --------
  block/qed-cluster.c    |  2 ++
  block/qed-table.c      | 12 ++++++++++--
  block/qed.c            |  4 ++--
  block/rbd.c            |  4 ----
  block/win32-aio.c      |  3 ---
  hw/block/virtio-blk.c  | 12 +++++++++++-
  hw/scsi/scsi-disk.c    | 15 +++++++++++++++
  hw/scsi/scsi-generic.c | 20 +++++++++++++++++---
  util/thread-pool.c     |  4 +++-
 files changed, 72 insertions(+), 50 deletions(-)
+diff --git a/block/archipelago.c b/block/archipelago.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/archipelago.c
++++ b/block/archipelago.c
+@@ -XXX,XX +XXX,XX @@ static void qemu_archipelago_complete_aio(void *opaque)
+ {
+     AIORequestData *reqdata = (AIORequestData *) opaque;
+     ArchipelagoAIOCB *aio_cb = (ArchipelagoAIOCB *) reqdata->aio_cb;
+-    AioContext *ctx = bdrv_get_aio_context(aio_cb->common.bs);
+-    aio_context_acquire(ctx);
+     aio_cb->common.cb(aio_cb->common.opaque, aio_cb->ret);
+-    aio_context_release(ctx);
+     aio_cb->status = 0;
+     qemu_aio_unref(aio_cb);
+diff --git a/block/block-backend.c b/block/block-backend.c
+index XXXXXXX..XXXXXXX 100644
+--- a/block/block-backend.c
++++ b/block/block-backend.c
+@@ -XXX,XX +XXX,XX @@ int blk_make_zero(BlockBackend *blk, BdrvRequestFlags flags)
+ static void error_callback_bh(void *opaque)
+ {
+     struct BlockBackendAIOCB *acb = opaque;
+-    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
+     bdrv_dec_in_flight(acb->common.bs);
+-    aio_context_acquire(ctx);
+     acb->common.cb(acb->common.opaque, acb->ret);
+-    aio_context_release(ctx);
+     qemu_aio_unref(acb);
+ }
+@@ -XXX,XX +XXX,XX @@ static void blk_aio_complete(BlkAioEmAIOCB *acb)
+ static void blk_aio_complete_bh(void *opaque)
+ {
+     BlkAioEmAIOCB *acb = opaque;
+-    AioContext *ctx = bdrv_get_aio_context(acb->common.bs);
+-
+     assert(acb->has_returned);
+-    aio_context_acquire(ctx);
+     blk_aio_complete(acb);
+-    aio_context_release(ctx);
+ }
+ static BlockAIOCB *blk_aio_prwv(BlockBackend *blk, int64_t offset, int bytes,
 diff --git a/block/curl.c b/block/curl.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/curl.c
 +++ b/block/curl.c
-@@ -XXX,XX +XXX,XX @@ static CURLMcode __curl_multi_socket_action(CURLM *multi_handle,
+@@ -XXX,XX +XXX,XX @@ static void curl_readv_bh_cb(void *p)
- #define CURL_BLOCK_OPT_TIMEOUT_DEFAULT 5
+     curl_multi_socket_action(s->multi, CURL_SOCKET_TIMEOUT, 0, &running);
- struct BDRVCURLState;
+ out:
-+struct CURLState;
++    aio_context_release(ctx);
+     if (ret != -EINPROGRESS) {
- static bool libcurl_initialized;
+         acb->common.cb(acb->common.opaque, ret);
+         qemu_aio_unref(acb);
-@@ -XXX,XX +XXX,XX @@ typedef struct CURLAIOCB {
+     }
+-    aio_context_release(ctx);
- typedef struct CURLSocket {
+ }
-     int fd;
-+    struct CURLState *state;
+ static BlockAIOCB *curl_aio_readv(BlockDriverState *bs,
-     QLIST_ENTRY(CURLSocket) next;
+diff --git a/block/io.c b/block/io.c
- } CURLSocket;
+index XXXXXXX..XXXXXXX 100644
+--- a/block/io.c
-@@ -XXX,XX +XXX,XX @@ static int curl_sock_cb(CURL *curl, curl_socket_t fd, int action,
++++ b/block/io.c
-     if (!socket) {
+@@ -XXX,XX +XXX,XX @@ static void bdrv_co_io_em_complete(void *opaque, int ret)
-         socket = g_new0(CURLSocket, 1);
+     CoroutineIOCompletion *co = opaque;
-         socket->fd = fd;
-+        socket->state = state;
+     co->ret = ret;
-         QLIST_INSERT_HEAD(&state->sockets, socket, next);
+-    qemu_coroutine_enter(co->coroutine);
-     }
++    aio_co_wake(co->coroutine);
-     socket = NULL;
+ }
  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,
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 09/16] curl: Pass CURLSocket to curl_multi_do()
+[Qemu-devel] [PULL v2 16/24] aio-posix: partially inline aio_dispatch into aio_poll
-curl_multi_do_locked() currently marks all sockets as ready.  That is
+From: Paolo Bonzini <pbonzini@redhat.com>
 not only inefficient, but in fact unsafe (the loop is).  A follow-up
 patch will change that, but to do so, curl_multi_do_locked() needs to
 know exactly which socket is ready; and that is accomplished by this
 patch here.
-Cc: qemu-stable@nongnu.org
+This patch prepares for the removal of unnecessary lockcnt inc/dec pairs.
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+Extract the dispatching loop for file descriptor handlers into a new
-Message-id: 20190910124136.10565-5-mreitz@redhat.com
+function aio_dispatch_handlers, and then inline aio_dispatch into
-Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
+aio_poll.
-Reviewed-by: John Snow <jsnow@redhat.com>
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+aio_dispatch can now become void.
 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: 20170213135235.12274-17-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
- block/curl.c | 20 +++++++++++---------
+ include/block/aio.h |  6 +-----
-file changed, 11 insertions(+), 9 deletions(-)
+ util/aio-posix.c    | 44 ++++++++++++++------------------------------
  util/aio-win32.c    | 13 ++++---------
  util/async.c        |  2 +-
 files changed, 20 insertions(+), 45 deletions(-)
-diff --git a/block/curl.c b/block/curl.c
+diff --git a/include/block/aio.h b/include/block/aio.h
 index XXXXXXX..XXXXXXX 100644
---- a/block/curl.c
+--- a/include/block/aio.h
-+++ b/block/curl.c
++++ b/include/block/aio.h
-@@ -XXX,XX +XXX,XX @@ static int curl_sock_cb(CURL *curl, curl_socket_t fd, int action,
+@@ -XXX,XX +XXX,XX @@ bool aio_pending(AioContext *ctx);
-     switch (action) {
+ /* Dispatch any pending callbacks from the GSource attached to the AioContext.
-         case CURL_POLL_IN:
+  *
-             aio_set_fd_handler(s->aio_context, fd, false,
+  * This is used internally in the implementation of the GSource.
--                               curl_multi_do, NULL, NULL, state);
+- *
-+                               curl_multi_do, NULL, NULL, socket);
+- * @dispatch_fds: true to process fds, false to skip them
-             break;
+- *                (can be used as an optimization by callers that know there
-         case CURL_POLL_OUT:
+- *                are no fds ready)
-             aio_set_fd_handler(s->aio_context, fd, false,
+  */
--                               NULL, curl_multi_do, NULL, state);
+-bool aio_dispatch(AioContext *ctx, bool dispatch_fds);
-+                               NULL, curl_multi_do, NULL, socket);
++void aio_dispatch(AioContext *ctx);
-             break;
-         case CURL_POLL_INOUT:
+ /* Progress in completing AIO work to occur.  This can issue new pending
-             aio_set_fd_handler(s->aio_context, fd, false,
+  * aio as a result of executing I/O completion or bh callbacks.
--                               curl_multi_do, curl_multi_do, NULL, state);
+diff --git a/util/aio-posix.c b/util/aio-posix.c
-+                               curl_multi_do, curl_multi_do, NULL, socket);
+index XXXXXXX..XXXXXXX 100644
-             break;
+--- a/util/aio-posix.c
-         case CURL_POLL_REMOVE:
++++ b/util/aio-posix.c
-             aio_set_fd_handler(s->aio_context, fd, false,
+@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
-@@ -XXX,XX +XXX,XX @@ static void curl_multi_check_completion(BDRVCURLState *s)
+     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;
  }
- /* Called with s->mutex held.  */
+-/*
--static void curl_multi_do_locked(CURLState *s)
+- * Note that dispatch_fds == false has the side-effect of post-poning the
-+static void curl_multi_do_locked(CURLSocket *ready_socket)
+- * freeing of deleted handlers.
 - */
 -bool aio_dispatch(AioContext *ctx, bool dispatch_fds)
 +void aio_dispatch(AioContext *ctx)
  {
-     CURLSocket *socket, *next_socket;
+-    bool progress;
-+    CURLState *s = ready_socket->state;
++    aio_bh_poll(ctx);
-     int running;
-     int r;
+-    /*
+-     * If there are callbacks left that have been queued, we need to call them.
-@@ -XXX,XX +XXX,XX @@ static void curl_multi_do_locked(CURLState *s)
+-     * 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).
- static void curl_multi_do(void *arg)
+-     */
 -    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)
  {
--    CURLState *s = (CURLState *)arg;
+-    bool progress;
-+    CURLSocket *socket = arg;
+-
-+    BDRVCURLState *s = socket->state->s;
+-    progress = aio_bh_poll(ctx);
+-    if (dispatch_fds) {
--    qemu_mutex_lock(&s->s->mutex);
+-        progress |= aio_dispatch_handlers(ctx, INVALID_HANDLE_VALUE);
--    curl_multi_do_locked(s);
+-    }
--    curl_multi_check_completion(s->s);
+-    progress |= timerlistgroup_run_timers(&ctx->tlg);
--    qemu_mutex_unlock(&s->s->mutex);
+-    return progress;
-+    qemu_mutex_lock(&s->mutex);
++    aio_bh_poll(ctx);
-+    curl_multi_do_locked(socket);
++    aio_dispatch_handlers(ctx, INVALID_HANDLE_VALUE);
-+    curl_multi_check_completion(s);
++    timerlistgroup_run_timers(&ctx->tlg);
 +    qemu_mutex_unlock(&s->mutex);
  }
- static void curl_multi_timeout_do(void *arg)
+ 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 14/16] block/qcow2: Fix corruption introduced by commit 8ac0f15f335
+[Qemu-devel] [PULL v2 17/24] async: remove unnecessary inc/dec pairs
-From: Maxim Levitsky <mlevitsk@redhat.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-This fixes subtle corruption introduced by luks threaded encryption
+Pull the increment/decrement pair out of aio_bh_poll and into the
-in commit 8ac0f15f335
+callers.
-Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1745922
+Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
 Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
 Reviewed-by: Fam Zheng <famz@redhat.com>
 Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
 Message-id: 20170213135235.12274-18-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
  util/aio-posix.c |  8 +++-----
  util/aio-win32.c |  8 ++++----
  util/async.c     | 12 ++++++------
 files changed, 13 insertions(+), 15 deletions(-)
-The corruption happens when we do a write that
+diff --git a/util/aio-posix.c b/util/aio-posix.c
    * writes to two or more unallocated clusters at once
    * doesn't fully cover the first sector
    * doesn't fully cover the last sector
    * uses luks encryption
 In this case, when allocating the new clusters we COW both areas
 prior to the write and after the write, and we encrypt them.
 The above mentioned commit accidentally made it so we encrypt the
 second COW area using the physical cluster offset of the first area.
 The problem is that offset_in_cluster in do_perform_cow_encrypt
 can be larger that the cluster size, thus cluster_offset
 will no longer point to the start of the cluster at which encrypted
 area starts.
 Next patch in this series will refactor the code to avoid all these
 assumptions.
 In the bugreport that was triggered by rebasing a luks image to new,
 zero filled base, which lot of such writes, and causes some files
 with zero areas to contain garbage there instead.
 But as described above it can happen elsewhere as well
 Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
 Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
 Message-id: 20190915203655.21638-2-mlevitsk@redhat.com
 Reviewed-by: Max Reitz <mreitz@redhat.com>
 Signed-off-by: Max Reitz <mreitz@redhat.com>
 ---
  block/qcow2-cluster.c | 7 ++++---
 file changed, 4 insertions(+), 3 deletions(-)
 diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c
 index XXXXXXX..XXXXXXX 100644
---- a/block/qcow2-cluster.c
+--- a/util/aio-posix.c
-+++ b/block/qcow2-cluster.c
++++ b/util/aio-posix.c
-@@ -XXX,XX +XXX,XX @@ static bool coroutine_fn do_perform_cow_encrypt(BlockDriverState *bs,
+@@ -XXX,XX +XXX,XX @@ static bool aio_dispatch_handlers(AioContext *ctx)
-         assert(QEMU_IS_ALIGNED(offset_in_cluster, BDRV_SECTOR_SIZE));
-         assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
+ void aio_dispatch(AioContext *ctx)
-         assert(s->crypto);
+ {
--        if (qcow2_co_encrypt(bs, cluster_offset,
++    qemu_lockcnt_inc(&ctx->list_lock);
--                             src_cluster_offset + offset_in_cluster,
+     aio_bh_poll(ctx);
--                             buffer, bytes) < 0) {
+-
-+        if (qcow2_co_encrypt(bs,
+-    qemu_lockcnt_inc(&ctx->list_lock);
-+                start_of_cluster(s, cluster_offset + offset_in_cluster),
+     aio_dispatch_handlers(ctx);
-+                src_cluster_offset + offset_in_cluster,
+     qemu_lockcnt_dec(&ctx->list_lock);
-+                buffer, bytes) < 0) {
-             return false;
+@@ -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;
+ }
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 02/16] block: Remove unused masks
+[Qemu-devel] [PULL v2 18/24] block: document fields protected by AioContext lock
-From: Nir Soffer <nirsof@gmail.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-Replace confusing usage:
+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: 20170213135235.12274-19-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(-)
-    ~BDRV_SECTOR_MASK
+diff --git a/include/block/block_int.h b/include/block/block_int.h
 With more clear:
     (BDRV_SECTOR_SIZE - 1)
 Remove BDRV_SECTOR_MASK and the unused BDRV_BLOCK_OFFSET_MASK which was
 it's last user.
 Signed-off-by: Nir Soffer <nsoffer@redhat.com>
 Message-id: 20190827185913.27427-3-nsoffer@redhat.com
 Reviewed-by: Juan Quintela <quintela@redhat.com>
 Reviewed-by: John Snow <jsnow@redhat.com>
 Signed-off-by: Max Reitz <mreitz@redhat.com>
 ---
  include/block/block.h | 2 --
  migration/block.c     | 2 +-
 files changed, 1 insertion(+), 3 deletions(-)
 diff --git a/include/block/block.h b/include/block/block.h
 index XXXXXXX..XXXXXXX 100644
---- a/include/block/block.h
+--- a/include/block/block_int.h
-+++ b/include/block/block.h
++++ b/include/block/block_int.h
-@@ -XXX,XX +XXX,XX @@ typedef struct HDGeometry {
+@@ -XXX,XX +XXX,XX @@ struct BdrvChild {
+  * copied as well.
- #define BDRV_SECTOR_BITS   9
+  */
- #define BDRV_SECTOR_SIZE   (1ULL << BDRV_SECTOR_BITS)
+ struct BlockDriverState {
--#define BDRV_SECTOR_MASK   ~(BDRV_SECTOR_SIZE - 1)
+-    int64_t total_sectors; /* if we are reading a disk image, give its
+-                              size in sectors */
- #define BDRV_REQUEST_MAX_SECTORS MIN(SIZE_MAX >> BDRV_SECTOR_BITS, \
++    /* Protected by big QEMU lock or read-only after opening.  No special
-                                      INT_MAX >> BDRV_SECTOR_BITS)
++     * locking needed during I/O...
-@@ -XXX,XX +XXX,XX @@ typedef struct HDGeometry {
++     */
- #define BDRV_BLOCK_ALLOCATED    0x10
+     int open_flags; /* flags used to open the file, re-used for re-open */
- #define BDRV_BLOCK_EOF          0x20
+     bool read_only; /* if true, the media is read only */
- #define BDRV_BLOCK_RECURSE      0x40
+     bool encrypted; /* if true, the media is encrypted */
--#define BDRV_BLOCK_OFFSET_MASK  BDRV_SECTOR_MASK
+@@ -XXX,XX +XXX,XX @@ struct BlockDriverState {
+     bool sg;        /* if true, the device is a /dev/sg* */
- typedef QSIMPLEQ_HEAD(BlockReopenQueue, BlockReopenQueueEntry) BlockReopenQueue;
+     bool probed;    /* if true, format was probed rather than specified */
-diff --git a/migration/block.c b/migration/block.c
+-    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/migration/block.c
+--- a/include/sysemu/block-backend.h
-+++ b/migration/block.c
++++ b/include/sysemu/block-backend.h
-@@ -XXX,XX +XXX,XX @@ static int block_load(QEMUFile *f, void *opaque, int version_id)
+@@ -XXX,XX +XXX,XX @@ typedef struct BlockDevOps {
-     do {
+  * fields that must be public. This is in particular for QLIST_ENTRY() and
-         addr = qemu_get_be64(f);
+  * friends so that BlockBackends can be kept in lists outside block-backend.c */
+ typedef struct BlockBackendPublic {
--        flags = addr & ~BDRV_SECTOR_MASK;
+-    /* I/O throttling.
-+        flags = addr & (BDRV_SECTOR_SIZE - 1);
+-     * throttle_state tells us if this BlockBackend has I/O limits configured.
-         addr >>= BDRV_SECTOR_BITS;
+-     * io_limits_disabled tells us if they are currently being enforced */
++    /* I/O throttling has its own locking, but also some fields are
-         if (flags & BLK_MIG_FLAG_DEVICE_BLOCK) {
++     * 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];
 --
-.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 07/16] curl: Keep *socket until the end of curl_sock_cb()
+[Qemu-devel] [PULL v2 20/24] coroutine-lock: add limited spinning to CoMutex
-This does not really change anything, but it makes the code a bit easier
+From: Paolo Bonzini <pbonzini@redhat.com>
 to follow once we use @socket as the opaque pointer for
 aio_set_fd_handler().
-Cc: qemu-stable@nongnu.org
+Running a very small critical section on pthread_mutex_t and CoMutex
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+shows that pthread_mutex_t is much faster because it doesn't actually
-Message-id: 20190910124136.10565-3-mreitz@redhat.com
+go to sleep.  What happens is that the critical section is shorter
-Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
+than the latency of entering the kernel and thus FUTEX_WAIT always
-Reviewed-by: John Snow <jsnow@redhat.com>
+fails.  With CoMutex there is no such latency but you still want to
-Signed-off-by: Max Reitz <mreitz@redhat.com>
+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>
 ---
- block/curl.c | 10 +++++-----
+ include/qemu/coroutine.h   |  5 +++++
-file changed, 5 insertions(+), 5 deletions(-)
+ util/qemu-coroutine-lock.c | 51 ++++++++++++++++++++++++++++++++++++++++------
  util/qemu-coroutine.c      |  2 +-
 files changed, 51 insertions(+), 7 deletions(-)
-diff --git a/block/curl.c b/block/curl.c
+diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
 index XXXXXXX..XXXXXXX 100644
---- a/block/curl.c
+--- a/include/qemu/coroutine.h
-+++ b/block/curl.c
++++ b/include/qemu/coroutine.h
-@@ -XXX,XX +XXX,XX @@ static int curl_sock_cb(CURL *curl, curl_socket_t fd, int action,
+@@ -XXX,XX +XXX,XX @@ typedef struct CoMutex {
+      */
-     QLIST_FOREACH(socket, &state->sockets, next) {
+     unsigned locked;
-         if (socket->fd == fd) {
--            if (action == CURL_POLL_REMOVE) {
++    /* Context that is holding the lock.  Useful to avoid spinning
--                QLIST_REMOVE(socket, next);
++     * when two coroutines on the same AioContext try to get the lock. :)
--                g_free(socket);
++     */
--            }
++    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;
          }
-     }
-@@ -XXX,XX +XXX,XX @@ static int curl_sock_cb(CURL *curl, curl_socket_t fd, int action,
+diff --git a/util/qemu-coroutine.c b/util/qemu-coroutine.c
-         socket->state = state;
+index XXXXXXX..XXXXXXX 100644
-         QLIST_INSERT_HEAD(&state->sockets, socket, next);
+--- a/util/qemu-coroutine.c
-     }
++++ b/util/qemu-coroutine.c
--    socket = NULL;
+@@ -XXX,XX +XXX,XX @@ void qemu_coroutine_enter(Coroutine *co)
+     co->ctx = qemu_get_current_aio_context();
-     trace_curl_sock_cb(action, (int)fd);
-     switch (action) {
+     /* Store co->ctx before anything that stores co.  Matches
-@@ -XXX,XX +XXX,XX @@ static int curl_sock_cb(CURL *curl, curl_socket_t fd, int action,
+-     * barrier in aio_co_wake.
-             break;
++     * barrier in aio_co_wake and qemu_co_mutex_wake.
-     }
+      */
+     smp_wmb();
 +    if (action == CURL_POLL_REMOVE) {
 +        QLIST_REMOVE(socket, next);
 +        g_free(socket);
 +    }
 +
      return 0;
  }
 --
-.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 16/16] qemu-iotests: Add test for bz #1745922
+[Qemu-devel] [PULL v2 22/24] coroutine-lock: place CoMutex before CoQueue in header
-From: Maxim Levitsky <mlevitsk@redhat.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
+This will avoid forward references in the next patch.  It is also
-Tested-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
+more logical because CoQueue is not anymore the basic primitive.
-Message-id: 20190915203655.21638-4-mlevitsk@redhat.com
-Reviewed-by: Max Reitz <mreitz@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-Signed-off-by: Max Reitz <mreitz@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/263     | 91 ++++++++++++++++++++++++++++++++++++++
+ include/qemu/coroutine.h | 89 ++++++++++++++++++++++++------------------------
- tests/qemu-iotests/263.out | 40 +++++++++++++++++
+file changed, 44 insertions(+), 45 deletions(-)
  tests/qemu-iotests/group   |  1 +
 files changed, 132 insertions(+)
  create mode 100755 tests/qemu-iotests/263
  create mode 100644 tests/qemu-iotests/263.out
-diff --git a/tests/qemu-iotests/263 b/tests/qemu-iotests/263
+diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
-new file mode 100755
+index XXXXXXX..XXXXXXX 100644
-index XXXXXXX..XXXXXXX
+--- a/include/qemu/coroutine.h
---- /dev/null
++++ b/include/qemu/coroutine.h
-+++ b/tests/qemu-iotests/263
+@@ -XXX,XX +XXX,XX @@ bool qemu_in_coroutine(void);
-@@ -XXX,XX +XXX,XX @@
+  */
-+#!/usr/bin/env bash
+ bool qemu_coroutine_entered(Coroutine *co);
-+#
-+# Test encrypted write that crosses cluster boundary of two unallocated clusters
+-
-+# Based on 188
+-/**
-+#
+- * CoQueues are a mechanism to queue coroutines in order to continue executing
-+# Copyright (C) 2019 Red Hat, Inc.
+- * them later. They provide the fundamental primitives on which coroutine locks
-+#
+- * are built.
-+# This program is free software; you can redistribute it and/or modify
+- */
-+# it under the terms of the GNU General Public License as published by
+-typedef struct CoQueue {
-+# the Free Software Foundation; either version 2 of the License, or
+-    QSIMPLEQ_HEAD(, Coroutine) entries;
-+# (at your option) any later version.
+-} CoQueue;
-+#
+-
-+# This program is distributed in the hope that it will be useful,
+-/**
-+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+- * Initialise a CoQueue. This must be called before any other operation is used
-+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+- * on the CoQueue.
-+# GNU General Public License for more details.
+- */
-+#
+-void qemu_co_queue_init(CoQueue *queue);
-+# You should have received a copy of the GNU General Public License
+-
-+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+-/**
-+#
+- * 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);
 +
-+# creator
++/**
-+owner=mlevitsk@redhat.com
++ * CoQueues are a mechanism to queue coroutines in order to continue executing
 + * them later.
 + */
 +typedef struct CoQueue {
 +    QSIMPLEQ_HEAD(, Coroutine) entries;
 +} CoQueue;
 +
-+seq=`basename $0`
++/**
-+echo "QA output created by $seq"
++ * Initialise a CoQueue. This must be called before any other operation is used
 + * on the CoQueue.
 + */
 +void qemu_co_queue_init(CoQueue *queue);
 +
-+status=1    # failure is the default!
++/**
 + * Adds the current coroutine to the CoQueue and transfers control to the
 + * caller of the coroutine.
 + */
 +void coroutine_fn qemu_co_queue_wait(CoQueue *queue);
 +
-+_cleanup()
++/**
-+{
++ * Restarts the next coroutine in the CoQueue and removes it from the queue.
-+    _cleanup_test_img
++ *
-+}
++ * Returns true if a coroutine was restarted, false if the queue is empty.
-+trap "_cleanup; exit \$status" 0 1 2 3 15
++ */
 +bool coroutine_fn qemu_co_queue_next(CoQueue *queue);
 +
-+# get standard environment, filters and checks
++/**
-+. ./common.rc
++ * Restarts all coroutines in the CoQueue and leaves the queue empty.
-+. ./common.filter
++ */
 +void coroutine_fn qemu_co_queue_restart_all(CoQueue *queue);
 +
-+_supported_fmt qcow2
++/**
-+_supported_proto generic
++ * Enter the next coroutine in the queue
-+_supported_os Linux
++ */
 +bool qemu_co_enter_next(CoQueue *queue);
 +
 +/**
 + * Checks if the CoQueue is empty.
 + */
 +bool qemu_co_queue_empty(CoQueue *queue);
 +
 +
-+size=1M
+ typedef struct CoRwlock {
-+
+     bool writer;
-+SECRET="secret,id=sec0,data=astrochicken"
+     int reader;
 +QEMU_IO_OPTIONS=$QEMU_IO_OPTIONS_NO_FMT
 +
 +
 +_run_test()
 +{
 +    echo "== reading the whole image =="
 +    $QEMU_IO --object $SECRET -c "read -P 0 0 $size" --image-opts "$1" | _filter_qemu_io | _filter_testdir
 +
 +    echo
 +    echo "== write two 512 byte sectors on a cluster boundary =="
 +    $QEMU_IO --object $SECRET -c "write -P 0xAA 0xFE00 0x400" --image-opts "$1" | _filter_qemu_io | _filter_testdir
 +
 +    echo
 +    echo "== verify that the rest of the image is not changed =="
 +    $QEMU_IO --object $SECRET -c "read -P 0x00 0x00000 0xFE00" --image-opts "$1" | _filter_qemu_io | _filter_testdir
 +    $QEMU_IO --object $SECRET -c "read -P 0xAA 0x0FE00 0x400" --image-opts "$1" | _filter_qemu_io | _filter_testdir
 +    $QEMU_IO --object $SECRET -c "read -P 0x00 0x10200 0xEFE00" --image-opts "$1" | _filter_qemu_io | _filter_testdir
 +
 +}
 +
 +
 +echo
 +echo "testing LUKS qcow2 encryption"
 +echo
 +
 +_make_test_img --object $SECRET -o "encrypt.format=luks,encrypt.key-secret=sec0,encrypt.iter-time=10,cluster_size=64K" $size
 +_run_test "driver=$IMGFMT,encrypt.key-secret=sec0,file.filename=$TEST_IMG"
 +_cleanup_test_img
 +
 +echo
 +echo "testing legacy AES qcow2 encryption"
 +echo
 +
 +
 +_make_test_img --object $SECRET -o "encrypt.format=aes,encrypt.key-secret=sec0,cluster_size=64K" $size
 +_run_test "driver=$IMGFMT,encrypt.key-secret=sec0,file.filename=$TEST_IMG"
 +_cleanup_test_img
 +
 +
 +
 +# success, all done
 +echo "*** done"
 +rm -f $seq.full
 +status=0
 diff --git a/tests/qemu-iotests/263.out b/tests/qemu-iotests/263.out
 new file mode 100644
 index XXXXXXX..XXXXXXX
 --- /dev/null
 +++ b/tests/qemu-iotests/263.out
@@ -XXX,XX +XXX,XX @@
 +QA output created by 263
 +
 +testing LUKS qcow2 encryption
 +
 +Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1048576 encrypt.format=luks encrypt.key-secret=sec0 encrypt.iter-time=10
 +== reading the whole image ==
 +read 1048576/1048576 bytes at offset 0
 +1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +
 +== write two 512 byte sectors on a cluster boundary ==
 +wrote 1024/1024 bytes at offset 65024
 +1 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +
 +== verify that the rest of the image is not changed ==
 +read 65024/65024 bytes at offset 0
 +63.500 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +read 1024/1024 bytes at offset 65024
 +1 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +read 982528/982528 bytes at offset 66048
 +959.500 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +
 +testing legacy AES qcow2 encryption
 +
 +Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1048576 encrypt.format=aes encrypt.key-secret=sec0
 +== reading the whole image ==
 +read 1048576/1048576 bytes at offset 0
 +1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +
 +== write two 512 byte sectors on a cluster boundary ==
 +wrote 1024/1024 bytes at offset 65024
 +1 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +
 +== verify that the rest of the image is not changed ==
 +read 65024/65024 bytes at offset 0
 +63.500 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +read 1024/1024 bytes at offset 65024
 +1 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +read 982528/982528 bytes at offset 66048
 +959.500 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 +*** done
 diff --git a/tests/qemu-iotests/group b/tests/qemu-iotests/group
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qemu-iotests/group
 +++ b/tests/qemu-iotests/group
@@ -XXX,XX +XXX,XX @@
 rw
 rw quick
 rw quick migration
 +263 rw quick
 rw auto quick
 rw quick
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 01/16] block: Use QEMU_IS_ALIGNED
+[Qemu-devel] [PULL v2 23/24] coroutine-lock: add mutex argument to CoQueue APIs
-From: Nir Soffer <nirsof@gmail.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-Replace instances of:
+All that CoQueue needs in order to become thread-safe is help
+from an external mutex.  Add this to the API.
-    (n & (BDRV_SECTOR_SIZE - 1)) == 0
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-And:
+Reviewed-by: Fam Zheng <famz@redhat.com>
+Message-id: 20170213181244.16297-6-pbonzini@redhat.com
-   (n & ~BDRV_SECTOR_MASK) == 0
+Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 With:
     QEMU_IS_ALIGNED(n, BDRV_SECTOR_SIZE)
 Which reveals the intent of the code better, and makes it easier to
 locate the code checking alignment.
 Signed-off-by: Nir Soffer <nsoffer@redhat.com>
 Message-id: 20190827185913.27427-2-nsoffer@redhat.com
 Reviewed-by: John Snow <jsnow@redhat.com>
 Signed-off-by: Max Reitz <mreitz@redhat.com>
 ---
- block/bochs.c         | 4 ++--
+ include/qemu/coroutine.h   |  8 +++++---
- block/cloop.c         | 4 ++--
+ block/backup.c             |  2 +-
- block/dmg.c           | 4 ++--
+ block/io.c                 |  4 ++--
- block/io.c            | 8 ++++----
+ block/nbd-client.c         |  2 +-
- block/qcow2-cluster.c | 4 ++--
+ block/qcow2-cluster.c      |  4 +---
- block/qcow2.c         | 4 ++--
+ block/sheepdog.c           |  2 +-
- block/vvfat.c         | 8 ++++----
+ block/throttle-groups.c    |  2 +-
- qemu-img.c            | 2 +-
+ hw/9pfs/9p.c               |  2 +-
-files changed, 19 insertions(+), 19 deletions(-)
+ util/qemu-coroutine-lock.c | 24 +++++++++++++++++++++---
+files changed, 34 insertions(+), 16 deletions(-)
-diff --git a/block/bochs.c b/block/bochs.c
-index XXXXXXX..XXXXXXX 100644
+diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
---- a/block/bochs.c
+index XXXXXXX..XXXXXXX 100644
-+++ b/block/bochs.c
+--- a/include/qemu/coroutine.h
-@@ -XXX,XX +XXX,XX @@ bochs_co_preadv(BlockDriverState *bs, uint64_t offset, uint64_t bytes,
++++ b/include/qemu/coroutine.h
-     QEMUIOVector local_qiov;
+@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_mutex_unlock(CoMutex *mutex);
-     int ret;
+ /**
--    assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
+  * CoQueues are a mechanism to queue coroutines in order to continue executing
--    assert((bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
+- * them later.
-+    assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
++ * them later.  They are similar to condition variables, but they need help
-+    assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
++ * from an external mutex in order to maintain thread-safety.
+  */
-     qemu_iovec_init(&local_qiov, qiov->niov);
+ typedef struct CoQueue {
-     qemu_co_mutex_lock(&s->lock);
+     QSIMPLEQ_HEAD(, Coroutine) entries;
-diff --git a/block/cloop.c b/block/cloop.c
+@@ -XXX,XX +XXX,XX @@ void qemu_co_queue_init(CoQueue *queue);
-index XXXXXXX..XXXXXXX 100644
---- a/block/cloop.c
+ /**
-+++ b/block/cloop.c
+  * Adds the current coroutine to the CoQueue and transfers control to the
-@@ -XXX,XX +XXX,XX @@ cloop_co_preadv(BlockDriverState *bs, uint64_t offset, uint64_t bytes,
+- * caller of the coroutine.
-     int nb_sectors = bytes >> BDRV_SECTOR_BITS;
++ * caller of the coroutine.  The mutex is unlocked during the wait and
-     int ret, i;
++ * locked again afterwards.
+  */
--    assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
+-void coroutine_fn qemu_co_queue_wait(CoQueue *queue);
--    assert((bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
++void coroutine_fn qemu_co_queue_wait(CoQueue *queue, CoMutex *mutex);
-+    assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
-+    assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
+ /**
+  * Restarts the next coroutine in the CoQueue and removes it from the queue.
-     qemu_co_mutex_lock(&s->lock);
+diff --git a/block/backup.c b/block/backup.c
+index XXXXXXX..XXXXXXX 100644
-diff --git a/block/dmg.c b/block/dmg.c
+--- a/block/backup.c
-index XXXXXXX..XXXXXXX 100644
++++ b/block/backup.c
---- a/block/dmg.c
+@@ -XXX,XX +XXX,XX @@ static void coroutine_fn wait_for_overlapping_requests(BackupBlockJob *job,
-+++ b/block/dmg.c
+         retry = false;
-@@ -XXX,XX +XXX,XX @@ dmg_co_preadv(BlockDriverState *bs, uint64_t offset, uint64_t bytes,
+         QLIST_FOREACH(req, &job->inflight_reqs, list) {
-     int nb_sectors = bytes >> BDRV_SECTOR_BITS;
+             if (end > req->start && start < req->end) {
-     int ret, i;
+-                qemu_co_queue_wait(&req->wait_queue);
++                qemu_co_queue_wait(&req->wait_queue, NULL);
--    assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
+                 retry = true;
--    assert((bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
+                 break;
-+    assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
+             }
 +    assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
      qemu_co_mutex_lock(&s->lock);
 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 int coroutine_fn bdrv_driver_preadv(BlockDriverState *bs,
+@@ -XXX,XX +XXX,XX @@ static bool coroutine_fn wait_serialising_requests(BdrvTrackedRequest *self)
-     sector_num = offset >> BDRV_SECTOR_BITS;
+                  * (instead of producing a deadlock in the former case). */
-     nb_sectors = bytes >> BDRV_SECTOR_BITS;
+                 if (!req->waiting_for) {
+                     self->waiting_for = req;
--    assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
+-                    qemu_co_queue_wait(&req->wait_queue);
--    assert((bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
++                    qemu_co_queue_wait(&req->wait_queue, NULL);
-+    assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
+                     self->waiting_for = NULL;
-+    assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
+                     retry = true;
-     assert(bytes <= BDRV_REQUEST_MAX_BYTES);
+                     waited = true;
-     assert(drv->bdrv_co_readv);
+@@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_flush(BlockDriverState *bs)
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_driver_pwritev(BlockDriverState *bs,
+     /* Wait until any previous flushes are completed */
-     sector_num = offset >> BDRV_SECTOR_BITS;
+     while (bs->active_flush_req) {
-     nb_sectors = bytes >> BDRV_SECTOR_BITS;
+-        qemu_co_queue_wait(&bs->flush_queue);
++        qemu_co_queue_wait(&bs->flush_queue, NULL);
--    assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
+     }
--    assert((bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
-+    assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
+     bs->active_flush_req = true;
-+    assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
+diff --git a/block/nbd-client.c b/block/nbd-client.c
-     assert(bytes <= BDRV_REQUEST_MAX_BYTES);
+index XXXXXXX..XXXXXXX 100644
+--- a/block/nbd-client.c
-     assert(drv->bdrv_co_writev);
++++ 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 bool coroutine_fn do_perform_cow_encrypt(BlockDriverState *bs,
+@@ -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)
  {
-     if (bytes && bs->encrypted) {
+     Coroutine *self = qemu_coroutine_self();
-         BDRVQcow2State *s = bs->opaque;
+     QSIMPLEQ_INSERT_TAIL(&queue->entries, self, co_queue_next);
--        assert((offset_in_cluster & ~BDRV_SECTOR_MASK) == 0);
++
--        assert((bytes & ~BDRV_SECTOR_MASK) == 0);
++    if (mutex) {
-+        assert(QEMU_IS_ALIGNED(offset_in_cluster, BDRV_SECTOR_SIZE));
++        qemu_co_mutex_unlock(mutex);
-+        assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
++    }
-         assert(s->crypto);
++
-         if (qcow2_co_encrypt(bs, cluster_offset,
++    /* There is no race condition here.  Other threads will call
-                              src_cluster_offset + offset_in_cluster,
++     * aio_co_schedule on our AioContext, which can reenter this
-diff --git a/block/qcow2.c b/block/qcow2.c
++     * coroutine but only after this yield and after the main loop
-index XXXXXXX..XXXXXXX 100644
++     * has gone through the next iteration.
---- a/block/qcow2.c
++     */
-+++ b/block/qcow2.c
+     qemu_coroutine_yield();
-@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_preadv_part(BlockDriverState *bs,
+     assert(qemu_in_coroutine());
-                     goto fail;
++
-                 }
++    /* TODO: OSv implements wait morphing here, where the wakeup
++     * primitive automatically places the woken coroutine on the
--                assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
++     * mutex's queue.  This avoids the thundering herd effect.
--                assert((cur_bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
++     */
-+                assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
++    if (mutex) {
-+                assert(QEMU_IS_ALIGNED(cur_bytes, BDRV_SECTOR_SIZE));
++        qemu_co_mutex_lock(mutex);
-                 if (qcow2_co_decrypt(bs, cluster_offset, offset,
++    }
-                                      cluster_data, cur_bytes) < 0) {
+ }
-                     ret = -EIO;
-diff --git a/block/vvfat.c b/block/vvfat.c
+ /**
-index XXXXXXX..XXXXXXX 100644
+@@ -XXX,XX +XXX,XX @@ void qemu_co_rwlock_rdlock(CoRwlock *lock)
---- a/block/vvfat.c
+     Coroutine *self = qemu_coroutine_self();
-+++ b/block/vvfat.c
-@@ -XXX,XX +XXX,XX @@ vvfat_co_preadv(BlockDriverState *bs, uint64_t offset, uint64_t bytes,
+     while (lock->writer) {
-     int nb_sectors = bytes >> BDRV_SECTOR_BITS;
+-        qemu_co_queue_wait(&lock->queue);
-     void *buf;
++        qemu_co_queue_wait(&lock->queue, NULL);
+     }
--    assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
+     lock->reader++;
--    assert((bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
+     self->locks_held++;
-+    assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
+@@ -XXX,XX +XXX,XX @@ void qemu_co_rwlock_wrlock(CoRwlock *lock)
-+    assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
+     Coroutine *self = qemu_coroutine_self();
-     buf = g_try_malloc(bytes);
+     while (lock->writer || lock->reader) {
-     if (bytes && buf == NULL) {
+-        qemu_co_queue_wait(&lock->queue);
-@@ -XXX,XX +XXX,XX @@ vvfat_co_pwritev(BlockDriverState *bs, uint64_t offset, uint64_t bytes,
++        qemu_co_queue_wait(&lock->queue, NULL);
-     int nb_sectors = bytes >> BDRV_SECTOR_BITS;
+     }
-     void *buf;
+     lock->writer = true;
+     self->locks_held++;
 -    assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
 -    assert((bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
 +    assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
 +    assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
      buf = g_try_malloc(bytes);
      if (bytes && buf == NULL) {
 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 @@ static int img_convert(int argc, char **argv)
              int64_t sval;
              sval = cvtnum(optarg);
 -            if (sval < 0 || sval & (BDRV_SECTOR_SIZE - 1) ||
 +            if (sval < 0 || !QEMU_IS_ALIGNED(sval, BDRV_SECTOR_SIZE) ||
                  sval / BDRV_SECTOR_SIZE > MAX_BUF_SECTORS) {
                  error_report("Invalid buffer size for sparse output specified. "
                      "Valid sizes are multiples of %llu up to %llu. Select "
 --
-.21.0
+.9.3

-[Qemu-devel] [PULL 15/16] block/qcow2: refactor encryption code
+[Qemu-devel] [PULL v2 24/24] coroutine-lock: make CoRwlock thread-safe and fair
-From: Maxim Levitsky <mlevitsk@redhat.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-* Change the qcow2_co_{encrypt|decrypt} to just receive full host and
+This adds a CoMutex around the existing CoQueue.  Because the write-side
-  guest offsets and use this function directly instead of calling
+can just take CoMutex, the old "writer" field is not necessary anymore.
-  do_perform_cow_encrypt (which is removed by that patch).
+Instead of removing it altogether, count the number of pending writers
 during a read-side critical section and forbid further readers from
 entering.
-* Adjust qcow2_co_encdec to take full host and guest offsets as well.
+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 ++++++++++++++++++++++++-----------
 files changed, 26 insertions(+), 12 deletions(-)
-* Document the qcow2_co_{encrypt|decrypt} arguments
+diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
   to prevent the bug fixed in former commit from hopefully
   happening again.
 Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
 Message-id: 20190915203655.21638-3-mlevitsk@redhat.com
 Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
 [mreitz: Let perform_cow() return the error value returned by
          qcow2_co_encrypt(), as proposed by Vladimir]
 Signed-off-by: Max Reitz <mreitz@redhat.com>
 ---
  block/qcow2.h         |  8 +++---
  block/qcow2-cluster.c | 41 +++++++++-------------------
  block/qcow2-threads.c | 63 +++++++++++++++++++++++++++++++++----------
  block/qcow2.c         |  5 ++--
 files changed, 69 insertions(+), 48 deletions(-)
 diff --git a/block/qcow2.h b/block/qcow2.h
 index XXXXXXX..XXXXXXX 100644
---- a/block/qcow2.h
+--- a/include/qemu/coroutine.h
-+++ b/block/qcow2.h
++++ b/include/qemu/coroutine.h
-@@ -XXX,XX +XXX,XX @@ ssize_t coroutine_fn
+@@ -XXX,XX +XXX,XX @@ bool qemu_co_queue_empty(CoQueue *queue);
- qcow2_co_decompress(BlockDriverState *bs, void *dest, size_t dest_size,
-                     const void *src, size_t src_size);
- int coroutine_fn
+ typedef struct CoRwlock {
--qcow2_co_encrypt(BlockDriverState *bs, uint64_t file_cluster_offset,
+-    bool writer;
--                 uint64_t offset, void *buf, size_t len);
++    int pending_writer;
-+qcow2_co_encrypt(BlockDriverState *bs, uint64_t host_offset,
+     int reader;
-+                 uint64_t guest_offset, void *buf, size_t len);
++    CoMutex mutex;
- int coroutine_fn
+     CoQueue queue;
--qcow2_co_decrypt(BlockDriverState *bs, uint64_t file_cluster_offset,
+ } CoRwlock;
--                 uint64_t offset, void *buf, size_t len);
-+qcow2_co_decrypt(BlockDriverState *bs, uint64_t host_offset,
+diff --git a/util/qemu-coroutine-lock.c b/util/qemu-coroutine-lock.c
 +                 uint64_t guest_offset, void *buf, size_t len);
  #endif
 diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c
 index XXXXXXX..XXXXXXX 100644
---- a/block/qcow2-cluster.c
+--- a/util/qemu-coroutine-lock.c
-+++ b/block/qcow2-cluster.c
++++ b/util/qemu-coroutine-lock.c
-@@ -XXX,XX +XXX,XX @@ static int coroutine_fn do_perform_cow_read(BlockDriverState *bs,
+@@ -XXX,XX +XXX,XX @@ void qemu_co_rwlock_init(CoRwlock *lock)
-     return 0;
+ {
      memset(lock, 0, sizeof(*lock));
      qemu_co_queue_init(&lock->queue);
 +    qemu_co_mutex_init(&lock->mutex);
  }
--static bool coroutine_fn do_perform_cow_encrypt(BlockDriverState *bs,
+ void qemu_co_rwlock_rdlock(CoRwlock *lock)
--                                                uint64_t src_cluster_offset,
+ {
--                                                uint64_t cluster_offset,
+     Coroutine *self = qemu_coroutine_self();
--                                                unsigned offset_in_cluster,
--                                                uint8_t *buffer,
+-    while (lock->writer) {
--                                                unsigned bytes)
+-        qemu_co_queue_wait(&lock->queue, NULL);
--{
++    qemu_co_mutex_lock(&lock->mutex);
--    if (bytes && bs->encrypted) {
++    /* For fairness, wait if a writer is in line.  */
--        BDRVQcow2State *s = bs->opaque;
++    while (lock->pending_writer) {
--        assert(QEMU_IS_ALIGNED(offset_in_cluster, BDRV_SECTOR_SIZE));
++        qemu_co_queue_wait(&lock->queue, &lock->mutex);
--        assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
+     }
--        assert(s->crypto);
+     lock->reader++;
--        if (qcow2_co_encrypt(bs,
++    qemu_co_mutex_unlock(&lock->mutex);
 -                start_of_cluster(s, cluster_offset + offset_in_cluster),
 -                src_cluster_offset + offset_in_cluster,
 -                buffer, bytes) < 0) {
 -            return false;
 -        }
 -    }
 -    return true;
 -}
 -
  static int coroutine_fn do_perform_cow_write(BlockDriverState *bs,
                                               uint64_t cluster_offset,
                                               unsigned offset_in_cluster,
@@ -XXX,XX +XXX,XX @@ static int perform_cow(BlockDriverState *bs, QCowL2Meta *m)
      /* Encrypt the data if necessary before writing it */
      if (bs->encrypted) {
 -        if (!do_perform_cow_encrypt(bs, m->offset, m->alloc_offset,
 -                                    start->offset, start_buffer,
 -                                    start->nb_bytes) ||
 -            !do_perform_cow_encrypt(bs, m->offset, m->alloc_offset,
 -                                    end->offset, end_buffer, end->nb_bytes)) {
 -            ret = -EIO;
 +        ret = qcow2_co_encrypt(bs,
 +                               m->alloc_offset + start->offset,
 +                               m->offset + start->offset,
 +                               start_buffer, start->nb_bytes);
 +        if (ret < 0) {
 +            goto fail;
 +        }
 +
-+        ret = qcow2_co_encrypt(bs,
++    /* The rest of the read-side critical section is run without the mutex.  */
-+                               m->alloc_offset + end->offset,
+     self->locks_held++;
-+                               m->offset + end->offset,
+ }
-+                               end_buffer, end->nb_bytes);
-+        if (ret < 0) {
+@@ -XXX,XX +XXX,XX @@ void qemu_co_rwlock_unlock(CoRwlock *lock)
-             goto fail;
+     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);
          }
      }
-diff --git a/block/qcow2-threads.c b/block/qcow2-threads.c
+-    self->locks_held--;
-index XXXXXXX..XXXXXXX 100644
++    qemu_co_mutex_unlock(&lock->mutex);
 --- a/block/qcow2-threads.c
 +++ b/block/qcow2-threads.c
@@ -XXX,XX +XXX,XX @@ static int qcow2_encdec_pool_func(void *opaque)
  }
- static int coroutine_fn
+ void qemu_co_rwlock_wrlock(CoRwlock *lock)
 -qcow2_co_encdec(BlockDriverState *bs, uint64_t file_cluster_offset,
 -                  uint64_t offset, void *buf, size_t len, Qcow2EncDecFunc func)
 +qcow2_co_encdec(BlockDriverState *bs, uint64_t host_offset,
 +                uint64_t guest_offset, void *buf, size_t len,
 +                Qcow2EncDecFunc func)
  {
-     BDRVQcow2State *s = bs->opaque;
+-    Coroutine *self = qemu_coroutine_self();
-     Qcow2EncDecData arg = {
+-
-         .block = s->crypto,
+-    while (lock->writer || lock->reader) {
--        .offset = s->crypt_physical_offset ?
+-        qemu_co_queue_wait(&lock->queue, NULL);
--                      file_cluster_offset + offset_into_cluster(s, offset) :
++    qemu_co_mutex_lock(&lock->mutex);
--                      offset,
++    lock->pending_writer++;
-+        .offset = s->crypt_physical_offset ? host_offset : guest_offset,
++    while (lock->reader) {
-         .buf = buf,
++        qemu_co_queue_wait(&lock->queue, &lock->mutex);
-         .len = len,
+     }
-         .func = func,
+-    lock->writer = true;
-     };
+-    self->locks_held++;
++    lock->pending_writer--;
 -    return qcow2_co_process(bs, qcow2_encdec_pool_func, &arg);
 +    assert(QEMU_IS_ALIGNED(guest_offset, BDRV_SECTOR_SIZE));
 +    assert(QEMU_IS_ALIGNED(host_offset, BDRV_SECTOR_SIZE));
 +    assert(QEMU_IS_ALIGNED(len, BDRV_SECTOR_SIZE));
 +    assert(s->crypto);
 +
-+    return len == 0 ? 0 : qcow2_co_process(bs, qcow2_encdec_pool_func, &arg);
++    /* 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.
 +     */
  }
-+/*
-+ * qcow2_co_encrypt()
-+ *
-+ * Encrypts one or more contiguous aligned sectors
-+ *
-+ * @host_offset - underlying storage offset of the first sector of the
-+ * data to be encrypted
-+ *
-+ * @guest_offset - guest (virtual) offset of the first sector of the
-+ * data to be encrypted
-+ *
-+ * @buf - buffer with the data to encrypt, that after encryption
-+ *        will be written to the underlying storage device at
-+ *        @host_offset
-+ *
-+ * @len - length of the buffer (must be a BDRV_SECTOR_SIZE multiple)
-+ *
-+ * Depending on the encryption method, @host_offset and/or @guest_offset
-+ * may be used for generating the initialization vector for
-+ * encryption.
-+ *
-+ * Note that while the whole range must be aligned on sectors, it
-+ * does not have to be aligned on clusters and can also cross cluster
-+ * boundaries
-+ */
- int coroutine_fn
--qcow2_co_encrypt(BlockDriverState *bs, uint64_t file_cluster_offset,
--                 uint64_t offset, void *buf, size_t len)
-+qcow2_co_encrypt(BlockDriverState *bs, uint64_t host_offset,
-+                 uint64_t guest_offset, void *buf, size_t len)
- {
--    return qcow2_co_encdec(bs, file_cluster_offset, offset, buf, len,
--                             qcrypto_block_encrypt);
-+    return qcow2_co_encdec(bs, host_offset, guest_offset, buf, len,
-+                           qcrypto_block_encrypt);
- }
-+/*
-+ * qcow2_co_decrypt()
-+ *
-+ * Decrypts one or more contiguous aligned sectors
-+ * Similar to qcow2_co_encrypt
-+ */
- int coroutine_fn
--qcow2_co_decrypt(BlockDriverState *bs, uint64_t file_cluster_offset,
--                 uint64_t offset, void *buf, size_t len)
-+qcow2_co_decrypt(BlockDriverState *bs, uint64_t host_offset,
-+                 uint64_t guest_offset, void *buf, size_t len)
- {
--    return qcow2_co_encdec(bs, file_cluster_offset, offset, buf, len,
--                             qcrypto_block_decrypt);
-+    return qcow2_co_encdec(bs, host_offset, guest_offset, buf, len,
-+                           qcrypto_block_decrypt);
- }
-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 coroutine_fn int qcow2_co_preadv_part(BlockDriverState *bs,
-                 assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
-                 assert(QEMU_IS_ALIGNED(cur_bytes, BDRV_SECTOR_SIZE));
--                if (qcow2_co_decrypt(bs, cluster_offset, offset,
-+                if (qcow2_co_decrypt(bs, cluster_offset + offset_in_cluster,
-+                                     offset,
-                                      cluster_data, cur_bytes) < 0) {
-                     ret = -EIO;
-                     goto fail;
-@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_pwritev_part(
-             qemu_iovec_to_buf(qiov, qiov_offset + bytes_done,
-                               cluster_data, cur_bytes);
--            if (qcow2_co_encrypt(bs, cluster_offset, offset,
-+            if (qcow2_co_encrypt(bs, cluster_offset + offset_in_cluster, offset,
-                                  cluster_data, cur_bytes) < 0) {
-                 ret = -EIO;
-                 goto out_unlocked;
 --
-.21.0
+.9.3

The following changes since commit dd25f97c66a75d1508f1d4c6478ed2c95bec428f:

Merge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20190913' into staging (2019-09-16 10:15:15 +0100)

are available in the Git repository at:

https://github.com/XanClic/qemu.git tags/pull-block-2019-09-16

for you to fetch changes up to 1825cc0783ccf0ec5d9f0b225a99b340bdd4c68f:

qemu-iotests: Add test for bz #1745922 (2019-09-16 15:37:12 +0200)

----------------------------------------------------------------
Block patches:
- Fix for block jobs when used with I/O threads
- Fix for a corruption when using qcow2's LUKS encryption mode
- cURL fix
- check-block.sh cleanups (for make check)
- Refactoring

----------------------------------------------------------------
Max Reitz (7):
  curl: Keep pointer to the CURLState in CURLSocket
  curl: Keep *socket until the end of curl_sock_cb()
  curl: Check completion in curl_multi_do()
  curl: Pass CURLSocket to curl_multi_do()
  curl: Report only ready sockets
  curl: Handle success in multi_check_completion
  curl: Check curl_multi_add_handle()'s return code

Maxim Levitsky (3):
  block/qcow2: Fix corruption introduced by commit 8ac0f15f335
  block/qcow2: refactor encryption code
  qemu-iotests: Add test for bz #1745922

Nir Soffer (2):
  block: Use QEMU_IS_ALIGNED
  block: Remove unused masks

Sergio Lopez (1):
  blockjob: update nodes head while removing all bdrv

Thomas Huth (2):
  tests/qemu-iotests/check: Replace "tests" with "iotests" in final
    status text
  tests/Makefile: Do not print the name of the check-block.sh shell
    script

Vladimir Sementsov-Ogievskiy (1):
  tests/qemu-iotests: Fix qemu-io related output in 026.out.nocache

-- 
2.21.0

From: Nir Soffer <nirsof@gmail.com>

Replace instances of:

(n & (BDRV_SECTOR_SIZE - 1)) == 0

And:

(n & ~BDRV_SECTOR_MASK) == 0

With:

QEMU_IS_ALIGNED(n, BDRV_SECTOR_SIZE)

Which reveals the intent of the code better, and makes it easier to
locate the code checking alignment.

Signed-off-by: Nir Soffer <nsoffer@redhat.com>
Message-id: 20190827185913.27427-2-nsoffer@redhat.com
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block/bochs.c         | 4 ++--
 block/cloop.c         | 4 ++--
 block/dmg.c           | 4 ++--
 block/io.c            | 8 ++++----
 block/qcow2-cluster.c | 4 ++--
 block/qcow2.c         | 4 ++--
 block/vvfat.c         | 8 ++++----
 qemu-img.c            | 2 +-
 8 files changed, 19 insertions(+), 19 deletions(-)

diff --git a/block/bochs.c b/block/bochs.c
index XXXXXXX..XXXXXXX 100644
--- a/block/bochs.c
+++ b/block/bochs.c
@@ -XXX,XX +XXX,XX @@ bochs_co_preadv(BlockDriverState *bs, uint64_t offset, uint64_t bytes,
     QEMUIOVector local_qiov;
     int ret;
 
-    assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
-    assert((bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
+    assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
+    assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
 
     qemu_iovec_init(&local_qiov, qiov->niov);
     qemu_co_mutex_lock(&s->lock);
diff --git a/block/cloop.c b/block/cloop.c
index XXXXXXX..XXXXXXX 100644
--- a/block/cloop.c
+++ b/block/cloop.c
@@ -XXX,XX +XXX,XX @@ cloop_co_preadv(BlockDriverState *bs, uint64_t offset, uint64_t bytes,
     int nb_sectors = bytes >> BDRV_SECTOR_BITS;
     int ret, i;
 
-    assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
-    assert((bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
+    assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
+    assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
 
     qemu_co_mutex_lock(&s->lock);
 
diff --git a/block/dmg.c b/block/dmg.c
index XXXXXXX..XXXXXXX 100644
--- a/block/dmg.c
+++ b/block/dmg.c
@@ -XXX,XX +XXX,XX @@ dmg_co_preadv(BlockDriverState *bs, uint64_t offset, uint64_t bytes,
     int nb_sectors = bytes >> BDRV_SECTOR_BITS;
     int ret, i;
 
-    assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
-    assert((bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
+    assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
+    assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
 
     qemu_co_mutex_lock(&s->lock);
 
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 int coroutine_fn bdrv_driver_preadv(BlockDriverState *bs,
     sector_num = offset >> BDRV_SECTOR_BITS;
     nb_sectors = bytes >> BDRV_SECTOR_BITS;
 
-    assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
-    assert((bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
+    assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
+    assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
     assert(bytes <= BDRV_REQUEST_MAX_BYTES);
     assert(drv->bdrv_co_readv);
 
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_driver_pwritev(BlockDriverState *bs,
     sector_num = offset >> BDRV_SECTOR_BITS;
     nb_sectors = bytes >> BDRV_SECTOR_BITS;
 
-    assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
-    assert((bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
+    assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
+    assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
     assert(bytes <= BDRV_REQUEST_MAX_BYTES);
 
     assert(drv->bdrv_co_writev);
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 bool coroutine_fn do_perform_cow_encrypt(BlockDriverState *bs,
 {
     if (bytes && bs->encrypted) {
         BDRVQcow2State *s = bs->opaque;
-        assert((offset_in_cluster & ~BDRV_SECTOR_MASK) == 0);
-        assert((bytes & ~BDRV_SECTOR_MASK) == 0);
+        assert(QEMU_IS_ALIGNED(offset_in_cluster, BDRV_SECTOR_SIZE));
+        assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
         assert(s->crypto);
         if (qcow2_co_encrypt(bs, cluster_offset,
                              src_cluster_offset + offset_in_cluster,
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 coroutine_fn int qcow2_co_preadv_part(BlockDriverState *bs,
                     goto fail;
                 }
 
-                assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
-                assert((cur_bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
+                assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
+                assert(QEMU_IS_ALIGNED(cur_bytes, BDRV_SECTOR_SIZE));
                 if (qcow2_co_decrypt(bs, cluster_offset, offset,
                                      cluster_data, cur_bytes) < 0) {
                     ret = -EIO;
diff --git a/block/vvfat.c b/block/vvfat.c
index XXXXXXX..XXXXXXX 100644
--- a/block/vvfat.c
+++ b/block/vvfat.c
@@ -XXX,XX +XXX,XX @@ vvfat_co_preadv(BlockDriverState *bs, uint64_t offset, uint64_t bytes,
     int nb_sectors = bytes >> BDRV_SECTOR_BITS;
     void *buf;
 
-    assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
-    assert((bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
+    assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
+    assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
 
     buf = g_try_malloc(bytes);
     if (bytes && buf == NULL) {
@@ -XXX,XX +XXX,XX @@ vvfat_co_pwritev(BlockDriverState *bs, uint64_t offset, uint64_t bytes,
     int nb_sectors = bytes >> BDRV_SECTOR_BITS;
     void *buf;
 
-    assert((offset & (BDRV_SECTOR_SIZE - 1)) == 0);
-    assert((bytes & (BDRV_SECTOR_SIZE - 1)) == 0);
+    assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
+    assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
 
     buf = g_try_malloc(bytes);
     if (bytes && buf == NULL) {
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 @@ static int img_convert(int argc, char **argv)
             int64_t sval;
 
             sval = cvtnum(optarg);
-            if (sval < 0 || sval & (BDRV_SECTOR_SIZE - 1) ||
+            if (sval < 0 || !QEMU_IS_ALIGNED(sval, BDRV_SECTOR_SIZE) ||
                 sval / BDRV_SECTOR_SIZE > MAX_BUF_SECTORS) {
                 error_report("Invalid buffer size for sparse output specified. "
                     "Valid sizes are multiples of %llu up to %llu. Select "
-- 
2.21.0

From: Nir Soffer <nirsof@gmail.com>

Replace confusing usage:

~BDRV_SECTOR_MASK

With more clear:

(BDRV_SECTOR_SIZE - 1)

Remove BDRV_SECTOR_MASK and the unused BDRV_BLOCK_OFFSET_MASK which was
it's last user.

Signed-off-by: Nir Soffer <nsoffer@redhat.com>
Message-id: 20190827185913.27427-3-nsoffer@redhat.com
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 include/block/block.h | 2 --
 migration/block.c     | 2 +-
 2 files changed, 1 insertion(+), 3 deletions(-)

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 @@ typedef struct HDGeometry {
 
 #define BDRV_SECTOR_BITS   9
 #define BDRV_SECTOR_SIZE   (1ULL << BDRV_SECTOR_BITS)
-#define BDRV_SECTOR_MASK   ~(BDRV_SECTOR_SIZE - 1)
 
 #define BDRV_REQUEST_MAX_SECTORS MIN(SIZE_MAX >> BDRV_SECTOR_BITS, \
                                      INT_MAX >> BDRV_SECTOR_BITS)
@@ -XXX,XX +XXX,XX @@ typedef struct HDGeometry {
 #define BDRV_BLOCK_ALLOCATED    0x10
 #define BDRV_BLOCK_EOF          0x20
 #define BDRV_BLOCK_RECURSE      0x40
-#define BDRV_BLOCK_OFFSET_MASK  BDRV_SECTOR_MASK
 
 typedef QSIMPLEQ_HEAD(BlockReopenQueue, BlockReopenQueueEntry) BlockReopenQueue;
 
diff --git a/migration/block.c b/migration/block.c
index XXXXXXX..XXXXXXX 100644
--- a/migration/block.c
+++ b/migration/block.c
@@ -XXX,XX +XXX,XX @@ static int block_load(QEMUFile *f, void *opaque, int version_id)
     do {
         addr = qemu_get_be64(f);
 
-        flags = addr & ~BDRV_SECTOR_MASK;
+        flags = addr & (BDRV_SECTOR_SIZE - 1);
         addr >>= BDRV_SECTOR_BITS;
 
         if (flags & BLK_MIG_FLAG_DEVICE_BLOCK) {
-- 
2.21.0

From: Thomas Huth <thuth@redhat.com>

When running "make check -j8" or something similar, the iotests are
running in parallel with the other tests. So when they are printing
out "Passed all xx tests" or a similar status message at the end,
it might not be quite clear that this message belongs to the iotests,
since the output might be mixed with the other tests. Thus change the
word "tests" here to "iotests" instead to avoid confusion.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-id: 20190906113920.11271-1-thuth@redhat.com
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 tests/qemu-iotests/check | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/tests/qemu-iotests/check b/tests/qemu-iotests/check
index XXXXXXX..XXXXXXX 100755
--- a/tests/qemu-iotests/check
+++ b/tests/qemu-iotests/check
@@ -XXX,XX +XXX,XX @@ END        { if (NR > 0) {
         if [ ! -z "$n_bad" -a $n_bad != 0 ]
         then
             echo "Failures:$bad"
-            echo "Failed $n_bad of $try tests"
+            echo "Failed $n_bad of $try iotests"
             echo "Failures:$bad" | fmt >>check.log
-            echo "Failed $n_bad of $try tests" >>check.log
+            echo "Failed $n_bad of $try iotests" >>check.log
         else
-            echo "Passed all $try tests"
-            echo "Passed all $try tests" >>check.log
+            echo "Passed all $try iotests"
+            echo "Passed all $try iotests" >>check.log
         fi
         needwrap=false
     fi
-- 
2.21.0

From: Thomas Huth <thuth@redhat.com>

The check script is already printing out which iotest is currently
running, so printing out the name of the check-block.sh shell script
looks superfluous here.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-id: 20190906113534.10907-1-thuth@redhat.com
Acked-by: John Snow <jsnow@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 tests/Makefile.include | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

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 @@ QEMU_IOTESTS_HELPERS-$(call land,$(CONFIG_SOFTMMU),$(CONFIG_LINUX)) = tests/qemu
 check-tests/check-block.sh: tests/check-block.sh qemu-img$(EXESUF) \
 		qemu-io$(EXESUF) qemu-nbd$(EXESUF) $(QEMU_IOTESTS_HELPERS-y) \
 		$(patsubst %,%/all,$(filter %-softmmu,$(TARGET_DIRS)))
-	$<
+	@$<
 
 .PHONY: $(patsubst %, check-%, $(check-qapi-schema-y))
 $(patsubst %, check-%, $(check-qapi-schema-y)): check-%.json: $(SRC_PATH)/%.json
-- 
2.21.0

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

qemu-io now prefixes its error and warnings with "qemu-io:".
36b9986b08787019e fixed a lot of iotests output but forget about
026.out.nocache. Fix it too.

Fixes: 99e98d7c9fc1a1639fad ("qemu-io: Use error_[gs]et_progname()")
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-id: 20190816153015.447957-2-vsementsov@virtuozzo.com
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 tests/qemu-iotests/026.out.nocache | 168 ++++++++++++++---------------
 1 file changed, 84 insertions(+), 84 deletions(-)

diff --git a/tests/qemu-iotests/026.out.nocache b/tests/qemu-iotests/026.out.nocache
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/026.out.nocache
+++ b/tests/qemu-iotests/026.out.nocache
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: l1_update; errno: 5; imm: off; once: off; write 
-Failed to flush the L2 table cache: Input/output error
-Failed to flush the refcount block cache: Input/output error
+qemu-io: Failed to flush the L2 table cache: Input/output error
+qemu-io: Failed to flush the refcount block cache: Input/output error
 write failed: Input/output error
 
 1 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ This means waste of disk space, but no harm to data.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: l1_update; errno: 5; imm: off; once: off; write -b
-Failed to flush the L2 table cache: Input/output error
-Failed to flush the refcount block cache: Input/output error
+qemu-io: Failed to flush the L2 table cache: Input/output error
+qemu-io: Failed to flush the refcount block cache: Input/output error
 write failed: Input/output error
 
 1 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: l1_update; errno: 28; imm: off; once: off; write 
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 
 1 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ This means waste of disk space, but no harm to data.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: l1_update; errno: 28; imm: off; once: off; write -b
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 
 1 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
 
 Event: l2_update; errno: 5; imm: off; once: off; write 
-Failed to flush the L2 table cache: Input/output error
-Failed to flush the refcount block cache: Input/output error
+qemu-io: Failed to flush the L2 table cache: Input/output error
+qemu-io: Failed to flush the refcount block cache: Input/output error
 wrote 131072/131072 bytes at offset 0
 128 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 
@@ -XXX,XX +XXX,XX @@ This means waste of disk space, but no harm to data.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: l2_update; errno: 5; imm: off; once: off; write -b
-Failed to flush the L2 table cache: Input/output error
-Failed to flush the refcount block cache: Input/output error
+qemu-io: Failed to flush the L2 table cache: Input/output error
+qemu-io: Failed to flush the refcount block cache: Input/output error
 wrote 131072/131072 bytes at offset 0
 128 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824
 
 Event: l2_update; errno: 28; imm: off; once: off; write 
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 wrote 131072/131072 bytes at offset 0
 128 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 
@@ -XXX,XX +XXX,XX @@ This means waste of disk space, but no harm to data.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: l2_update; errno: 28; imm: off; once: off; write -b
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 wrote 131072/131072 bytes at offset 0
 128 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
 
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: l2_alloc_write; errno: 5; imm: off; once: off; write 
-Failed to flush the L2 table cache: Input/output error
-Failed to flush the refcount block cache: Input/output error
+qemu-io: Failed to flush the L2 table cache: Input/output error
+qemu-io: Failed to flush the refcount block cache: Input/output error
 write failed: Input/output error
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: l2_alloc_write; errno: 5; imm: off; once: off; write -b
-Failed to flush the L2 table cache: Input/output error
-Failed to flush the refcount block cache: Input/output error
+qemu-io: Failed to flush the L2 table cache: Input/output error
+qemu-io: Failed to flush the refcount block cache: Input/output error
 write failed: Input/output error
 
 1 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: l2_alloc_write; errno: 28; imm: off; once: off; write 
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: l2_alloc_write; errno: 28; imm: off; once: off; write -b
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 
 1 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: write_aio; errno: 5; imm: off; once: off; write 
-Failed to flush the L2 table cache: Input/output error
-Failed to flush the refcount block cache: Input/output error
+qemu-io: Failed to flush the L2 table cache: Input/output error
+qemu-io: Failed to flush the refcount block cache: Input/output error
 write failed: Input/output error
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: write_aio; errno: 5; imm: off; once: off; write -b
-Failed to flush the L2 table cache: Input/output error
-Failed to flush the refcount block cache: Input/output error
+qemu-io: Failed to flush the L2 table cache: Input/output error
+qemu-io: Failed to flush the refcount block cache: Input/output error
 write failed: Input/output error
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: write_aio; errno: 28; imm: off; once: off; write 
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: write_aio; errno: 28; imm: off; once: off; write -b
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_load; errno: 5; imm: off; once: off; write 
-Failed to flush the L2 table cache: Input/output error
-Failed to flush the refcount block cache: Input/output error
+qemu-io: Failed to flush the L2 table cache: Input/output error
+qemu-io: Failed to flush the refcount block cache: Input/output error
 write failed: Input/output error
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_load; errno: 5; imm: off; once: off; write -b
-Failed to flush the L2 table cache: Input/output error
-Failed to flush the refcount block cache: Input/output error
+qemu-io: Failed to flush the L2 table cache: Input/output error
+qemu-io: Failed to flush the refcount block cache: Input/output error
 write failed: Input/output error
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_load; errno: 28; imm: off; once: off; write 
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_load; errno: 28; imm: off; once: off; write -b
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_update_part; errno: 5; imm: off; once: off; write 
-Failed to flush the L2 table cache: Input/output error
-Failed to flush the refcount block cache: Input/output error
+qemu-io: Failed to flush the L2 table cache: Input/output error
+qemu-io: Failed to flush the refcount block cache: Input/output error
 write failed: Input/output error
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_update_part; errno: 5; imm: off; once: off; write -b
-Failed to flush the L2 table cache: Input/output error
-Failed to flush the refcount block cache: Input/output error
+qemu-io: Failed to flush the L2 table cache: Input/output error
+qemu-io: Failed to flush the refcount block cache: Input/output error
 write failed: Input/output error
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_update_part; errno: 28; imm: off; once: off; write 
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_update_part; errno: 28; imm: off; once: off; write -b
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_alloc; errno: 5; imm: off; once: off; write 
-Failed to flush the L2 table cache: Input/output error
-Failed to flush the refcount block cache: Input/output error
+qemu-io: Failed to flush the L2 table cache: Input/output error
+qemu-io: Failed to flush the refcount block cache: Input/output error
 write failed: Input/output error
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_alloc; errno: 5; imm: off; once: off; write -b
-Failed to flush the L2 table cache: Input/output error
-Failed to flush the refcount block cache: Input/output error
+qemu-io: Failed to flush the L2 table cache: Input/output error
+qemu-io: Failed to flush the refcount block cache: Input/output error
 write failed: Input/output error
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_alloc; errno: 28; imm: off; once: off; write 
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_alloc; errno: 28; imm: off; once: off; write -b
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_alloc_hookup; errno: 28; imm: off; once: off; write 
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 
 55 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ This means waste of disk space, but no harm to data.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_alloc_hookup; errno: 28; imm: off; once: off; write -b
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 
 251 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_alloc_write; errno: 28; imm: off; once: off; write 
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_alloc_write; errno: 28; imm: off; once: off; write -b
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_alloc_write_blocks; errno: 28; imm: off; once: off; write 
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 
 10 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ This means waste of disk space, but no harm to data.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_alloc_write_blocks; errno: 28; imm: off; once: off; write -b
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 
 23 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_alloc_write_table; errno: 28; imm: off; once: off; write 
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 
 10 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ This means waste of disk space, but no harm to data.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_alloc_write_table; errno: 28; imm: off; once: off; write -b
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 
 23 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_alloc_switch_table; errno: 28; imm: off; once: off; write 
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 
 10 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ This means waste of disk space, but no harm to data.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: refblock_alloc_switch_table; errno: 28; imm: off; once: off; write -b
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 
 23 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: l1_grow_write_table; errno: 5; imm: off; once: off
-Failed to flush the L2 table cache: Input/output error
-Failed to flush the refcount block cache: Input/output error
+qemu-io: Failed to flush the L2 table cache: Input/output error
+qemu-io: Failed to flush the refcount block cache: Input/output error
 write failed: Input/output error
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: l1_grow_write_table; errno: 28; imm: off; once: off
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: l1_grow_activate_table; errno: 5; imm: off; once: off
-Failed to flush the L2 table cache: Input/output error
-Failed to flush the refcount block cache: Input/output error
+qemu-io: Failed to flush the L2 table cache: Input/output error
+qemu-io: Failed to flush the refcount block cache: Input/output error
 write failed: Input/output error
 
 96 leaked clusters were found on the image.
@@ -XXX,XX +XXX,XX @@ No errors were found on the image.
 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 
 
 Event: l1_grow_activate_table; errno: 28; imm: off; once: off
-Failed to flush the L2 table cache: No space left on device
-Failed to flush the refcount block cache: No space left on device
+qemu-io: Failed to flush the L2 table cache: No space left on device
+qemu-io: Failed to flush the refcount block cache: No space left on device
 write failed: No space left on device
 
 96 leaked clusters were found on the image.
-- 
2.21.0

A follow-up patch will make curl_multi_do() and curl_multi_read() take a
CURLSocket instead of the CURLState.  They still need the latter,
though, so add a pointer to it to the former.

Cc: qemu-stable@nongnu.org
Signed-off-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Message-id: 20190910124136.10565-2-mreitz@redhat.com
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block/curl.c | 3 +++
 1 file changed, 3 insertions(+)

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 CURLMcode __curl_multi_socket_action(CURLM *multi_handle,
 #define CURL_BLOCK_OPT_TIMEOUT_DEFAULT 5
 
 struct BDRVCURLState;
+struct CURLState;
 
 static bool libcurl_initialized;
 
@@ -XXX,XX +XXX,XX @@ typedef struct CURLAIOCB {
 
 typedef struct CURLSocket {
     int fd;
+    struct CURLState *state;
     QLIST_ENTRY(CURLSocket) next;
 } CURLSocket;
 
@@ -XXX,XX +XXX,XX @@ static int curl_sock_cb(CURL *curl, curl_socket_t fd, int action,
     if (!socket) {
         socket = g_new0(CURLSocket, 1);
         socket->fd = fd;
+        socket->state = state;
         QLIST_INSERT_HEAD(&state->sockets, socket, next);
     }
     socket = NULL;
-- 
2.21.0

This does not really change anything, but it makes the code a bit easier
to follow once we use @socket as the opaque pointer for
aio_set_fd_handler().

Cc: qemu-stable@nongnu.org
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20190910124136.10565-3-mreitz@redhat.com
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block/curl.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 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 int curl_sock_cb(CURL *curl, curl_socket_t fd, int action,
 
     QLIST_FOREACH(socket, &state->sockets, next) {
         if (socket->fd == fd) {
-            if (action == CURL_POLL_REMOVE) {
-                QLIST_REMOVE(socket, next);
-                g_free(socket);
-            }
             break;
         }
     }
@@ -XXX,XX +XXX,XX @@ static int curl_sock_cb(CURL *curl, curl_socket_t fd, int action,
         socket->state = state;
         QLIST_INSERT_HEAD(&state->sockets, socket, next);
     }
-    socket = NULL;
 
     trace_curl_sock_cb(action, (int)fd);
     switch (action) {
@@ -XXX,XX +XXX,XX @@ static int curl_sock_cb(CURL *curl, curl_socket_t fd, int action,
             break;
     }
 
+    if (action == CURL_POLL_REMOVE) {
+        QLIST_REMOVE(socket, next);
+        g_free(socket);
+    }
+
     return 0;
 }
 
-- 
2.21.0

While it is more likely that transfers complete after some file
descriptor has data ready to read, we probably should not rely on it.
Better be safe than sorry and call curl_multi_check_completion() in
curl_multi_do(), too, just like it is done in curl_multi_read().

With this change, curl_multi_do() and curl_multi_read() are actually the
same, so drop curl_multi_read() and use curl_multi_do() as the sole FD
handler.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20190910124136.10565-4-mreitz@redhat.com
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block/curl.c | 14 ++------------
 1 file changed, 2 insertions(+), 12 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 @@ typedef struct BDRVCURLState {
 
 static void curl_clean_state(CURLState *s);
 static void curl_multi_do(void *arg);
-static void curl_multi_read(void *arg);
 
 #ifdef NEED_CURL_TIMER_CALLBACK
 /* Called from curl_multi_do_locked, with s->mutex held.  */
@@ -XXX,XX +XXX,XX @@ static int curl_sock_cb(CURL *curl, curl_socket_t fd, int action,
     switch (action) {
         case CURL_POLL_IN:
             aio_set_fd_handler(s->aio_context, fd, false,
-                               curl_multi_read, NULL, NULL, state);
+                               curl_multi_do, NULL, NULL, state);
             break;
         case CURL_POLL_OUT:
             aio_set_fd_handler(s->aio_context, fd, false,
@@ -XXX,XX +XXX,XX @@ static int curl_sock_cb(CURL *curl, curl_socket_t fd, int action,
             break;
         case CURL_POLL_INOUT:
             aio_set_fd_handler(s->aio_context, fd, false,
-                               curl_multi_read, curl_multi_do, NULL, state);
+                               curl_multi_do, curl_multi_do, NULL, state);
             break;
         case CURL_POLL_REMOVE:
             aio_set_fd_handler(s->aio_context, fd, false,
@@ -XXX,XX +XXX,XX @@ static void curl_multi_do(void *arg)
 {
     CURLState *s = (CURLState *)arg;
 
-    qemu_mutex_lock(&s->s->mutex);
-    curl_multi_do_locked(s);
-    qemu_mutex_unlock(&s->s->mutex);
-}
-
-static void curl_multi_read(void *arg)
-{
-    CURLState *s = (CURLState *)arg;
-
     qemu_mutex_lock(&s->s->mutex);
     curl_multi_do_locked(s);
     curl_multi_check_completion(s->s);
-- 
2.21.0

curl_multi_do_locked() currently marks all sockets as ready.  That is
not only inefficient, but in fact unsafe (the loop is).  A follow-up
patch will change that, but to do so, curl_multi_do_locked() needs to
know exactly which socket is ready; and that is accomplished by this
patch here.

Cc: qemu-stable@nongnu.org
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20190910124136.10565-5-mreitz@redhat.com
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block/curl.c | 20 +++++++++++---------
 1 file changed, 11 insertions(+), 9 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 int curl_sock_cb(CURL *curl, curl_socket_t fd, int action,
     switch (action) {
         case CURL_POLL_IN:
             aio_set_fd_handler(s->aio_context, fd, false,
-                               curl_multi_do, NULL, NULL, state);
+                               curl_multi_do, NULL, NULL, socket);
             break;
         case CURL_POLL_OUT:
             aio_set_fd_handler(s->aio_context, fd, false,
-                               NULL, curl_multi_do, NULL, state);
+                               NULL, curl_multi_do, NULL, socket);
             break;
         case CURL_POLL_INOUT:
             aio_set_fd_handler(s->aio_context, fd, false,
-                               curl_multi_do, curl_multi_do, NULL, state);
+                               curl_multi_do, curl_multi_do, NULL, socket);
             break;
         case CURL_POLL_REMOVE:
             aio_set_fd_handler(s->aio_context, fd, false,
@@ -XXX,XX +XXX,XX @@ static void curl_multi_check_completion(BDRVCURLState *s)
 }
 
 /* Called with s->mutex held.  */
-static void curl_multi_do_locked(CURLState *s)
+static void curl_multi_do_locked(CURLSocket *ready_socket)
 {
     CURLSocket *socket, *next_socket;
+    CURLState *s = ready_socket->state;
     int running;
     int r;
 
@@ -XXX,XX +XXX,XX @@ static void curl_multi_do_locked(CURLState *s)
 
 static void curl_multi_do(void *arg)
 {
-    CURLState *s = (CURLState *)arg;
+    CURLSocket *socket = arg;
+    BDRVCURLState *s = socket->state->s;
 
-    qemu_mutex_lock(&s->s->mutex);
-    curl_multi_do_locked(s);
-    curl_multi_check_completion(s->s);
-    qemu_mutex_unlock(&s->s->mutex);
+    qemu_mutex_lock(&s->mutex);
+    curl_multi_do_locked(socket);
+    curl_multi_check_completion(s);
+    qemu_mutex_unlock(&s->mutex);
 }
 
 static void curl_multi_timeout_do(void *arg)
-- 
2.21.0

Instead of reporting all sockets to cURL, only report the one that has
caused curl_multi_do_locked() to be called.  This lets us get rid of the
QLIST_FOREACH_SAFE() list, which was actually wrong: SAFE foreaches are
only safe when the current element is removed in each iteration.  If it
possible for the list to be concurrently modified, we cannot guarantee
that only the current element will be removed.  Therefore, we must not
use QLIST_FOREACH_SAFE() here.

Fixes: ff5ca1664af85b24a4180d595ea6873fd3deac57
Cc: qemu-stable@nongnu.org
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20190910124136.10565-6-mreitz@redhat.com
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block/curl.c | 17 ++++++-----------
 1 file changed, 6 insertions(+), 11 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)
 }
 
 /* Called with s->mutex held.  */
-static void curl_multi_do_locked(CURLSocket *ready_socket)
+static void curl_multi_do_locked(CURLSocket *socket)
 {
-    CURLSocket *socket, *next_socket;
-    CURLState *s = ready_socket->state;
+    BDRVCURLState *s = socket->state->s;
     int running;
     int r;
 
-    if (!s->s->multi) {
+    if (!s->multi) {
         return;
     }
 
-    /* Need to use _SAFE because curl_multi_socket_action() may trigger
-     * curl_sock_cb() which might modify this list */
-    QLIST_FOREACH_SAFE(socket, &s->sockets, next, next_socket) {
-        do {
-            r = curl_multi_socket_action(s->s->multi, socket->fd, 0, &running);
-        } while (r == CURLM_CALL_MULTI_PERFORM);
-    }
+    do {
+        r = curl_multi_socket_action(s->multi, socket->fd, 0, &running);
+    } while (r == CURLM_CALL_MULTI_PERFORM);
 }
 
 static void curl_multi_do(void *arg)
-- 
2.21.0

Background: As of cURL 7.59.0, it verifies that several functions are
not called from within a callback.  Among these functions is
curl_multi_add_handle().

curl_read_cb() is a callback from cURL and not a coroutine.  Waking up
acb->co will lead to entering it then and there, which means the current
request will settle and the caller (if it runs in the same coroutine)
may then issue the next request.  In such a case, we will enter
curl_setup_preadv() effectively from within curl_read_cb().

Calling curl_multi_add_handle() will then fail and the new request will
not be processed.

Fix this by not letting curl_read_cb() wake up acb->co.  Instead, leave
the whole business of settling the AIOCB objects to
curl_multi_check_completion() (which is called from our timer callback
and our FD handler, so not from any cURL callbacks).

Reported-by: Natalie Gavrielov <ngavrilo@redhat.com>
Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1740193
Cc: qemu-stable@nongnu.org
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20190910124136.10565-7-mreitz@redhat.com
Reviewed-by: John Snow <jsnow@redhat.com>
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block/curl.c | 69 ++++++++++++++++++++++------------------------------
 1 file changed, 29 insertions(+), 40 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 size_t curl_read_cb(void *ptr, size_t size, size_t nmemb, void *opaque)
 {
     CURLState *s = ((CURLState*)opaque);
     size_t realsize = size * nmemb;
-    int i;
 
     trace_curl_read_cb(realsize);
 
@@ -XXX,XX +XXX,XX @@ static size_t curl_read_cb(void *ptr, size_t size, size_t nmemb, void *opaque)
     memcpy(s->orig_buf + s->buf_off, ptr, realsize);
     s->buf_off += realsize;
 
-    for(i=0; i<CURL_NUM_ACB; i++) {
-        CURLAIOCB *acb = s->acb[i];
-
-        if (!acb)
-            continue;
-
-        if ((s->buf_off >= acb->end)) {
-            size_t request_length = acb->bytes;
-
-            qemu_iovec_from_buf(acb->qiov, 0, s->orig_buf + acb->start,
-                                acb->end - acb->start);
-
-            if (acb->end - acb->start < request_length) {
-                size_t offset = acb->end - acb->start;
-                qemu_iovec_memset(acb->qiov, offset, 0,
-                                  request_length - offset);
-            }
-
-            acb->ret = 0;
-            s->acb[i] = NULL;
-            qemu_mutex_unlock(&s->s->mutex);
-            aio_co_wake(acb->co);
-            qemu_mutex_lock(&s->s->mutex);
-        }
-    }
-
 read_end:
     /* curl will error out if we do not return this value */
     return size * nmemb;
@@ -XXX,XX +XXX,XX @@ static void curl_multi_check_completion(BDRVCURLState *s)
             break;
 
         if (msg->msg == CURLMSG_DONE) {
+            int i;
             CURLState *state = NULL;
+            bool error = msg->data.result != CURLE_OK;
+
             curl_easy_getinfo(msg->easy_handle, CURLINFO_PRIVATE,
                               (char **)&state);
 
-            /* ACBs for successful messages get completed in curl_read_cb */
-            if (msg->data.result != CURLE_OK) {
-                int i;
+            if (error) {
                 static int errcount = 100;
 
                 /* Don't lose the original error message from curl, since
@@ -XXX,XX +XXX,XX @@ static void curl_multi_check_completion(BDRVCURLState *s)
                         error_report("curl: further errors suppressed");
                     }
                 }
+            }
 
-                for (i = 0; i < CURL_NUM_ACB; i++) {
-                    CURLAIOCB *acb = state->acb[i];
+            for (i = 0; i < CURL_NUM_ACB; i++) {
+                CURLAIOCB *acb = state->acb[i];
 
-                    if (acb == NULL) {
-                        continue;
-                    }
+                if (acb == NULL) {
+                    continue;
+                }
+
+                if (!error) {
+                    /* Assert that we have read all data */
+                    assert(state->buf_off >= acb->end);
+
+                    qemu_iovec_from_buf(acb->qiov, 0,
+                                        state->orig_buf + acb->start,
+                                        acb->end - acb->start);
 
-                    acb->ret = -EIO;
-                    state->acb[i] = NULL;
-                    qemu_mutex_unlock(&s->mutex);
-                    aio_co_wake(acb->co);
-                    qemu_mutex_lock(&s->mutex);
+                    if (acb->end - acb->start < acb->bytes) {
+                        size_t offset = acb->end - acb->start;
+                        qemu_iovec_memset(acb->qiov, offset, 0,
+                                          acb->bytes - offset);
+                    }
                 }
+
+                acb->ret = error ? -EIO : 0;
+                state->acb[i] = NULL;
+                qemu_mutex_unlock(&s->mutex);
+                aio_co_wake(acb->co);
+                qemu_mutex_lock(&s->mutex);
             }
 
             curl_clean_state(state);
-- 
2.21.0

If we had done that all along, debugging would have been much simpler.
(Also, I/O errors are better than hangs.)

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20190910124136.10565-8-mreitz@redhat.com
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block/curl.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

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_setup_preadv(BlockDriverState *bs, CURLAIOCB *acb)
     trace_curl_setup_preadv(acb->bytes, start, state->range);
     curl_easy_setopt(state->curl, CURLOPT_RANGE, state->range);
 
-    curl_multi_add_handle(s->multi, state->curl);
+    if (curl_multi_add_handle(s->multi, state->curl) != CURLM_OK) {
+        state->acb[0] = NULL;
+        acb->ret = -EIO;
+
+        curl_clean_state(state);
+        goto out;
+    }
 
     /* Tell curl it needs to kick things off */
     curl_multi_socket_action(s->multi, CURL_SOCKET_TIMEOUT, 0, &running);
-- 
2.21.0

From: Sergio Lopez <slp@redhat.com>

block_job_remove_all_bdrv() iterates through job->nodes, calling
bdrv_root_unref_child() for each entry. The call to the latter may
reach child_job_[can_]set_aio_ctx(), which will also attempt to
traverse job->nodes, potentially finding entries that where freed
on previous iterations.

To avoid this situation, update job->nodes head on each iteration to
ensure that already freed entries are no longer linked to the list.

RHBZ: https://bugzilla.redhat.com/show_bug.cgi?id=1746631
Signed-off-by: Sergio Lopez <slp@redhat.com>
Cc: qemu-stable@nongnu.org
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20190911100316.32282-1-mreitz@redhat.com
Reviewed-by: Sergio Lopez <slp@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 blockjob.c | 17 +++++++++++++----
 1 file changed, 13 insertions(+), 4 deletions(-)

diff --git a/blockjob.c b/blockjob.c
index XXXXXXX..XXXXXXX 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -XXX,XX +XXX,XX @@ static const BdrvChildRole child_job = {
 
 void block_job_remove_all_bdrv(BlockJob *job)
 {
-    GSList *l;
-    for (l = job->nodes; l; l = l->next) {
+    /*
+     * bdrv_root_unref_child() may reach child_job_[can_]set_aio_ctx(),
+     * which will also traverse job->nodes, so consume the list one by
+     * one to make sure that such a concurrent access does not attempt
+     * to process an already freed BdrvChild.
+     */
+    while (job->nodes) {
+        GSList *l = job->nodes;
         BdrvChild *c = l->data;
+
+        job->nodes = l->next;
+
         bdrv_op_unblock_all(c->bs, job->blocker);
         bdrv_root_unref_child(c);
+
+        g_slist_free_1(l);
     }
-    g_slist_free(job->nodes);
-    job->nodes = NULL;
 }
 
 bool block_job_has_bdrv(BlockJob *job, BlockDriverState *bs)
-- 
2.21.0

From: Maxim Levitsky <mlevitsk@redhat.com>

This fixes subtle corruption introduced by luks threaded encryption
in commit 8ac0f15f335

Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1745922

The corruption happens when we do a write that
   * writes to two or more unallocated clusters at once
   * doesn't fully cover the first sector
   * doesn't fully cover the last sector
   * uses luks encryption

In this case, when allocating the new clusters we COW both areas
prior to the write and after the write, and we encrypt them.

The above mentioned commit accidentally made it so we encrypt the
second COW area using the physical cluster offset of the first area.

The problem is that offset_in_cluster in do_perform_cow_encrypt
can be larger that the cluster size, thus cluster_offset
will no longer point to the start of the cluster at which encrypted
area starts.

Next patch in this series will refactor the code to avoid all these
assumptions.

In the bugreport that was triggered by rebasing a luks image to new,
zero filled base, which lot of such writes, and causes some files
with zero areas to contain garbage there instead.
But as described above it can happen elsewhere as well

Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-id: 20190915203655.21638-2-mlevitsk@redhat.com
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block/qcow2-cluster.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

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 bool coroutine_fn do_perform_cow_encrypt(BlockDriverState *bs,
         assert(QEMU_IS_ALIGNED(offset_in_cluster, BDRV_SECTOR_SIZE));
         assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
         assert(s->crypto);
-        if (qcow2_co_encrypt(bs, cluster_offset,
-                             src_cluster_offset + offset_in_cluster,
-                             buffer, bytes) < 0) {
+        if (qcow2_co_encrypt(bs,
+                start_of_cluster(s, cluster_offset + offset_in_cluster),
+                src_cluster_offset + offset_in_cluster,
+                buffer, bytes) < 0) {
             return false;
         }
     }
-- 
2.21.0

From: Maxim Levitsky <mlevitsk@redhat.com>

* Change the qcow2_co_{encrypt|decrypt} to just receive full host and
  guest offsets and use this function directly instead of calling
  do_perform_cow_encrypt (which is removed by that patch).

* Adjust qcow2_co_encdec to take full host and guest offsets as well.

* Document the qcow2_co_{encrypt|decrypt} arguments
  to prevent the bug fixed in former commit from hopefully
  happening again.

Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Message-id: 20190915203655.21638-3-mlevitsk@redhat.com
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
[mreitz: Let perform_cow() return the error value returned by
         qcow2_co_encrypt(), as proposed by Vladimir]
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block/qcow2.h         |  8 +++---
 block/qcow2-cluster.c | 41 +++++++++-------------------
 block/qcow2-threads.c | 63 +++++++++++++++++++++++++++++++++----------
 block/qcow2.c         |  5 ++--
 4 files changed, 69 insertions(+), 48 deletions(-)

diff --git a/block/qcow2.h b/block/qcow2.h
index XXXXXXX..XXXXXXX 100644
--- a/block/qcow2.h
+++ b/block/qcow2.h
@@ -XXX,XX +XXX,XX @@ ssize_t coroutine_fn
 qcow2_co_decompress(BlockDriverState *bs, void *dest, size_t dest_size,
                     const void *src, size_t src_size);
 int coroutine_fn
-qcow2_co_encrypt(BlockDriverState *bs, uint64_t file_cluster_offset,
-                 uint64_t offset, void *buf, size_t len);
+qcow2_co_encrypt(BlockDriverState *bs, uint64_t host_offset,
+                 uint64_t guest_offset, void *buf, size_t len);
 int coroutine_fn
-qcow2_co_decrypt(BlockDriverState *bs, uint64_t file_cluster_offset,
-                 uint64_t offset, void *buf, size_t len);
+qcow2_co_decrypt(BlockDriverState *bs, uint64_t host_offset,
+                 uint64_t guest_offset, void *buf, size_t len);
 
 #endif
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 coroutine_fn do_perform_cow_read(BlockDriverState *bs,
     return 0;
 }
 
-static bool coroutine_fn do_perform_cow_encrypt(BlockDriverState *bs,
-                                                uint64_t src_cluster_offset,
-                                                uint64_t cluster_offset,
-                                                unsigned offset_in_cluster,
-                                                uint8_t *buffer,
-                                                unsigned bytes)
-{
-    if (bytes && bs->encrypted) {
-        BDRVQcow2State *s = bs->opaque;
-        assert(QEMU_IS_ALIGNED(offset_in_cluster, BDRV_SECTOR_SIZE));
-        assert(QEMU_IS_ALIGNED(bytes, BDRV_SECTOR_SIZE));
-        assert(s->crypto);
-        if (qcow2_co_encrypt(bs,
-                start_of_cluster(s, cluster_offset + offset_in_cluster),
-                src_cluster_offset + offset_in_cluster,
-                buffer, bytes) < 0) {
-            return false;
-        }
-    }
-    return true;
-}
-
 static int coroutine_fn do_perform_cow_write(BlockDriverState *bs,
                                              uint64_t cluster_offset,
                                              unsigned offset_in_cluster,
@@ -XXX,XX +XXX,XX @@ static int perform_cow(BlockDriverState *bs, QCowL2Meta *m)
 
     /* Encrypt the data if necessary before writing it */
     if (bs->encrypted) {
-        if (!do_perform_cow_encrypt(bs, m->offset, m->alloc_offset,
-                                    start->offset, start_buffer,
-                                    start->nb_bytes) ||
-            !do_perform_cow_encrypt(bs, m->offset, m->alloc_offset,
-                                    end->offset, end_buffer, end->nb_bytes)) {
-            ret = -EIO;
+        ret = qcow2_co_encrypt(bs,
+                               m->alloc_offset + start->offset,
+                               m->offset + start->offset,
+                               start_buffer, start->nb_bytes);
+        if (ret < 0) {
+            goto fail;
+        }
+
+        ret = qcow2_co_encrypt(bs,
+                               m->alloc_offset + end->offset,
+                               m->offset + end->offset,
+                               end_buffer, end->nb_bytes);
+        if (ret < 0) {
             goto fail;
         }
     }
diff --git a/block/qcow2-threads.c b/block/qcow2-threads.c
index XXXXXXX..XXXXXXX 100644
--- a/block/qcow2-threads.c
+++ b/block/qcow2-threads.c
@@ -XXX,XX +XXX,XX @@ static int qcow2_encdec_pool_func(void *opaque)
 }
 
 static int coroutine_fn
-qcow2_co_encdec(BlockDriverState *bs, uint64_t file_cluster_offset,
-                  uint64_t offset, void *buf, size_t len, Qcow2EncDecFunc func)
+qcow2_co_encdec(BlockDriverState *bs, uint64_t host_offset,
+                uint64_t guest_offset, void *buf, size_t len,
+                Qcow2EncDecFunc func)
 {
     BDRVQcow2State *s = bs->opaque;
     Qcow2EncDecData arg = {
         .block = s->crypto,
-        .offset = s->crypt_physical_offset ?
-                      file_cluster_offset + offset_into_cluster(s, offset) :
-                      offset,
+        .offset = s->crypt_physical_offset ? host_offset : guest_offset,
         .buf = buf,
         .len = len,
         .func = func,
     };
 
-    return qcow2_co_process(bs, qcow2_encdec_pool_func, &arg);
+    assert(QEMU_IS_ALIGNED(guest_offset, BDRV_SECTOR_SIZE));
+    assert(QEMU_IS_ALIGNED(host_offset, BDRV_SECTOR_SIZE));
+    assert(QEMU_IS_ALIGNED(len, BDRV_SECTOR_SIZE));
+    assert(s->crypto);
+
+    return len == 0 ? 0 : qcow2_co_process(bs, qcow2_encdec_pool_func, &arg);
 }
 
+/*
+ * qcow2_co_encrypt()
+ *
+ * Encrypts one or more contiguous aligned sectors
+ *
+ * @host_offset - underlying storage offset of the first sector of the
+ * data to be encrypted
+ *
+ * @guest_offset - guest (virtual) offset of the first sector of the
+ * data to be encrypted
+ *
+ * @buf - buffer with the data to encrypt, that after encryption
+ *        will be written to the underlying storage device at
+ *        @host_offset
+ *
+ * @len - length of the buffer (must be a BDRV_SECTOR_SIZE multiple)
+ *
+ * Depending on the encryption method, @host_offset and/or @guest_offset
+ * may be used for generating the initialization vector for
+ * encryption.
+ *
+ * Note that while the whole range must be aligned on sectors, it
+ * does not have to be aligned on clusters and can also cross cluster
+ * boundaries
+ */
 int coroutine_fn
-qcow2_co_encrypt(BlockDriverState *bs, uint64_t file_cluster_offset,
-                 uint64_t offset, void *buf, size_t len)
+qcow2_co_encrypt(BlockDriverState *bs, uint64_t host_offset,
+                 uint64_t guest_offset, void *buf, size_t len)
 {
-    return qcow2_co_encdec(bs, file_cluster_offset, offset, buf, len,
-                             qcrypto_block_encrypt);
+    return qcow2_co_encdec(bs, host_offset, guest_offset, buf, len,
+                           qcrypto_block_encrypt);
 }
 
+/*
+ * qcow2_co_decrypt()
+ *
+ * Decrypts one or more contiguous aligned sectors
+ * Similar to qcow2_co_encrypt
+ */
 int coroutine_fn
-qcow2_co_decrypt(BlockDriverState *bs, uint64_t file_cluster_offset,
-                 uint64_t offset, void *buf, size_t len)
+qcow2_co_decrypt(BlockDriverState *bs, uint64_t host_offset,
+                 uint64_t guest_offset, void *buf, size_t len)
 {
-    return qcow2_co_encdec(bs, file_cluster_offset, offset, buf, len,
-                             qcrypto_block_decrypt);
+    return qcow2_co_encdec(bs, host_offset, guest_offset, buf, len,
+                           qcrypto_block_decrypt);
 }
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 coroutine_fn int qcow2_co_preadv_part(BlockDriverState *bs,
 
                 assert(QEMU_IS_ALIGNED(offset, BDRV_SECTOR_SIZE));
                 assert(QEMU_IS_ALIGNED(cur_bytes, BDRV_SECTOR_SIZE));
-                if (qcow2_co_decrypt(bs, cluster_offset, offset,
+                if (qcow2_co_decrypt(bs, cluster_offset + offset_in_cluster,
+                                     offset,
                                      cluster_data, cur_bytes) < 0) {
                     ret = -EIO;
                     goto fail;
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_pwritev_part(
             qemu_iovec_to_buf(qiov, qiov_offset + bytes_done,
                               cluster_data, cur_bytes);
 
-            if (qcow2_co_encrypt(bs, cluster_offset, offset,
+            if (qcow2_co_encrypt(bs, cluster_offset + offset_in_cluster, offset,
                                  cluster_data, cur_bytes) < 0) {
                 ret = -EIO;
                 goto out_unlocked;
-- 
2.21.0

From: Maxim Levitsky <mlevitsk@redhat.com>

Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Tested-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-id: 20190915203655.21638-4-mlevitsk@redhat.com
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 tests/qemu-iotests/263     | 91 ++++++++++++++++++++++++++++++++++++++
 tests/qemu-iotests/263.out | 40 +++++++++++++++++
 tests/qemu-iotests/group   |  1 +
 3 files changed, 132 insertions(+)
 create mode 100755 tests/qemu-iotests/263
 create mode 100644 tests/qemu-iotests/263.out

diff --git a/tests/qemu-iotests/263 b/tests/qemu-iotests/263
new file mode 100755
index XXXXXXX..XXXXXXX
--- /dev/null
+++ b/tests/qemu-iotests/263
@@ -XXX,XX +XXX,XX @@
+#!/usr/bin/env bash
+#
+# Test encrypted write that crosses cluster boundary of two unallocated clusters
+# Based on 188
+#
+# Copyright (C) 2019 Red Hat, Inc.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+#
+
+# creator
+owner=mlevitsk@redhat.com
+
+seq=`basename $0`
+echo "QA output created by $seq"
+
+status=1	# failure is the default!
+
+_cleanup()
+{
+	_cleanup_test_img
+}
+trap "_cleanup; exit \$status" 0 1 2 3 15
+
+# get standard environment, filters and checks
+. ./common.rc
+. ./common.filter
+
+_supported_fmt qcow2
+_supported_proto generic
+_supported_os Linux
+
+
+size=1M
+
+SECRET="secret,id=sec0,data=astrochicken"
+QEMU_IO_OPTIONS=$QEMU_IO_OPTIONS_NO_FMT
+
+
+_run_test()
+{
+	echo "== reading the whole image =="
+	$QEMU_IO --object $SECRET -c "read -P 0 0 $size" --image-opts "$1" | _filter_qemu_io | _filter_testdir
+
+	echo
+	echo "== write two 512 byte sectors on a cluster boundary =="
+	$QEMU_IO --object $SECRET -c "write -P 0xAA 0xFE00 0x400" --image-opts "$1" | _filter_qemu_io | _filter_testdir
+
+	echo
+	echo "== verify that the rest of the image is not changed =="
+	$QEMU_IO --object $SECRET -c "read -P 0x00 0x00000 0xFE00" --image-opts "$1" | _filter_qemu_io | _filter_testdir
+	$QEMU_IO --object $SECRET -c "read -P 0xAA 0x0FE00 0x400" --image-opts "$1" | _filter_qemu_io | _filter_testdir
+	$QEMU_IO --object $SECRET -c "read -P 0x00 0x10200 0xEFE00" --image-opts "$1" | _filter_qemu_io | _filter_testdir
+
+}
+
+
+echo
+echo "testing LUKS qcow2 encryption"
+echo
+
+_make_test_img --object $SECRET -o "encrypt.format=luks,encrypt.key-secret=sec0,encrypt.iter-time=10,cluster_size=64K" $size
+_run_test "driver=$IMGFMT,encrypt.key-secret=sec0,file.filename=$TEST_IMG"
+_cleanup_test_img
+
+echo
+echo "testing legacy AES qcow2 encryption"
+echo
+
+
+_make_test_img --object $SECRET -o "encrypt.format=aes,encrypt.key-secret=sec0,cluster_size=64K" $size
+_run_test "driver=$IMGFMT,encrypt.key-secret=sec0,file.filename=$TEST_IMG"
+_cleanup_test_img
+
+
+
+# success, all done
+echo "*** done"
+rm -f $seq.full
+status=0
diff --git a/tests/qemu-iotests/263.out b/tests/qemu-iotests/263.out
new file mode 100644
index XXXXXXX..XXXXXXX
--- /dev/null
+++ b/tests/qemu-iotests/263.out
@@ -XXX,XX +XXX,XX @@
+QA output created by 263
+
+testing LUKS qcow2 encryption
+
+Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1048576 encrypt.format=luks encrypt.key-secret=sec0 encrypt.iter-time=10
+== reading the whole image ==
+read 1048576/1048576 bytes at offset 0
+1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+== write two 512 byte sectors on a cluster boundary ==
+wrote 1024/1024 bytes at offset 65024
+1 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+== verify that the rest of the image is not changed ==
+read 65024/65024 bytes at offset 0
+63.500 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+read 1024/1024 bytes at offset 65024
+1 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+read 982528/982528 bytes at offset 66048
+959.500 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+testing legacy AES qcow2 encryption
+
+Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1048576 encrypt.format=aes encrypt.key-secret=sec0
+== reading the whole image ==
+read 1048576/1048576 bytes at offset 0
+1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+== write two 512 byte sectors on a cluster boundary ==
+wrote 1024/1024 bytes at offset 65024
+1 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+== verify that the rest of the image is not changed ==
+read 65024/65024 bytes at offset 0
+63.500 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+read 1024/1024 bytes at offset 65024
+1 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+read 982528/982528 bytes at offset 66048
+959.500 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+*** done
diff --git a/tests/qemu-iotests/group b/tests/qemu-iotests/group
index XXXXXXX..XXXXXXX 100644
--- a/tests/qemu-iotests/group
+++ b/tests/qemu-iotests/group
@@ -XXX,XX +XXX,XX @@
 257 rw
 258 rw quick
 262 rw quick migration
+263 rw quick
 265 rw auto quick
 266 rw quick
-- 
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