Series comparison

-[PULL 00/33] Block patches
+[PULL 0/8] Block patches
-The following changes since commit 8507c9d5c9a62de2a0e281b640f995e26eac46af:
+The following changes since commit 6c769690ac845fa62642a5f93b4e4bd906adab95:
-  Merge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging (2020-11-03 15:59:44 +0000)
+  Merge remote-tracking branch 'remotes/vsementsov/tags/pull-simplebench-2021-05-04' into staging (2021-05-21 12:02:34 +0100)
 are available in the Git repository at:
   https://gitlab.com/stefanha/qemu.git tags/block-pull-request
-for you to fetch changes up to fc107d86840b3364e922c26cf7631b7fd38ce523:
+for you to fetch changes up to 0a6f0c76a030710780ce10d6347a70f098024d21:
-  util/vfio-helpers: Assert offset is aligned to page size (2020-11-03 19:06:23 +0000)
+  coroutine-sleep: introduce qemu_co_sleep (2021-05-21 18:22:33 +0100)
 ----------------------------------------------------------------
-Pull request for 5.2
+Pull request
-NVMe fixes to solve IOMMU issues on non-x86 and error message/tracing
+(Resent due to an email preparation mistake.)
 improvements. Elena Afanasova's ioeventfd fixes are also included.
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ----------------------------------------------------------------
-Elena Afanasova (2):
+Paolo Bonzini (6):
-  accel/kvm: add PIO ioeventfds only in case kvm_eventfds_allowed is
+  coroutine-sleep: use a stack-allocated timer
-    true
+  coroutine-sleep: disallow NULL QemuCoSleepState** argument
-  softmmu/memory: fix memory_region_ioeventfd_equal()
+  coroutine-sleep: allow qemu_co_sleep_wake that wakes nothing
   coroutine-sleep: move timer out of QemuCoSleepState
   coroutine-sleep: replace QemuCoSleepState pointer with struct in the
     API
   coroutine-sleep: introduce qemu_co_sleep
-Eric Auger (4):
+Philippe Mathieu-Daudé (1):
-  block/nvme: Change size and alignment of IDENTIFY response buffer
+  bitops.h: Improve find_xxx_bit() documentation
   block/nvme: Change size and alignment of queue
   block/nvme: Change size and alignment of prp_list_pages
   block/nvme: Align iov's va and size on host page size
-Philippe Mathieu-Daudé (27):
+Zenghui Yu (1):
-  MAINTAINERS: Cover "block/nvme.h" file
+  multi-process: Initialize variables declared with g_auto*
   block/nvme: Use hex format to display offset in trace events
   block/nvme: Report warning with warn_report()
   block/nvme: Trace controller capabilities
   block/nvme: Trace nvme_poll_queue() per queue
   block/nvme: Improve nvme_free_req_queue_wait() trace information
   block/nvme: Trace queue pair creation/deletion
   block/nvme: Move definitions before structure declarations
   block/nvme: Use unsigned integer for queue counter/size
   block/nvme: Make nvme_identify() return boolean indicating error
   block/nvme: Make nvme_init_queue() return boolean indicating error
   block/nvme: Introduce Completion Queue definitions
   block/nvme: Use definitions instead of magic values in add_io_queue()
   block/nvme: Correctly initialize Admin Queue Attributes
   block/nvme: Simplify ADMIN queue access
   block/nvme: Simplify nvme_cmd_sync()
   block/nvme: Set request_alignment at initialization
   block/nvme: Correct minimum device page size
   block/nvme: Fix use of write-only doorbells page on Aarch64 arch
   block/nvme: Fix nvme_submit_command() on big-endian host
   util/vfio-helpers: Improve reporting unsupported IOMMU type
   util/vfio-helpers: Trace PCI I/O config accesses
   util/vfio-helpers: Trace PCI BAR region info
   util/vfio-helpers: Trace where BARs are mapped
   util/vfio-helpers: Improve DMA trace events
   util/vfio-helpers: Convert vfio_dump_mapping to trace events
   util/vfio-helpers: Assert offset is aligned to page size
- MAINTAINERS          |   2 +
+ include/qemu/bitops.h       | 15 ++++++--
- include/block/nvme.h |  18 ++--
+ include/qemu/coroutine.h    | 27 ++++++++-----
- accel/kvm/kvm-all.c  |   6 +-
+ block/block-copy.c          | 10 ++---
- block/nvme.c         | 209 ++++++++++++++++++++++++-------------------
+ block/nbd.c                 | 14 +++----
- softmmu/memory.c     |  11 ++-
+ hw/remote/memory.c          |  5 +--
- util/vfio-helpers.c  |  43 +++++----
+ hw/remote/proxy.c           |  3 +-
- block/trace-events   |  30 ++++---
+ util/qemu-coroutine-sleep.c | 75 +++++++++++++++++++------------------
- util/trace-events    |  10 ++-
+files changed, 79 insertions(+), 70 deletions(-)
 files changed, 195 insertions(+), 134 deletions(-)
 --
-.28.0
+.31.1

-[PULL 01/33] accel/kvm: add PIO ioeventfds only in case kvm_eventfds_allowed is true
+[PULL 1/8] multi-process: Initialize variables declared with g_auto*
-From: Elena Afanasova <eafanasova@gmail.com>
+From: Zenghui Yu <yuzenghui@huawei.com>
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
+Quote docs/devel/style.rst (section "Automatic memory deallocation"):
-Signed-off-by: Elena Afanasova <eafanasova@gmail.com>
-Message-Id: <20201017210102.26036-1-eafanasova@gmail.com>
+* Variables declared with g_auto* MUST always be initialized,
   otherwise the cleanup function will use uninitialized stack memory
 Initialize @name properly to get rid of the compilation error (using
 gcc-7.3.0 on CentOS):
 ../hw/remote/proxy.c: In function 'pci_proxy_dev_realize':
 /usr/include/glib-2.0/glib/glib-autocleanups.h:28:3: error: 'name' may be used uninitialized in this function [-Werror=maybe-uninitialized]
    g_free (*pp);
    ^~~~~~~~~~~~
 ../hw/remote/proxy.c:350:30: note: 'name' was declared here
              g_autofree char *name;
                               ^~~~
 Signed-off-by: Zenghui Yu <yuzenghui@huawei.com>
 Reviewed-by: Jagannathan Raman <jag.raman@oracle.com>
 Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
 Reviewed-by: Miroslav Rezanina <mrezanin@redhat.com>
 Message-id: 20210312112143.1369-1-yuzenghui@huawei.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
- accel/kvm/kvm-all.c | 6 ++++--
+ hw/remote/memory.c | 5 ++---
-file changed, 4 insertions(+), 2 deletions(-)
+ hw/remote/proxy.c  | 3 +--
 files changed, 3 insertions(+), 5 deletions(-)
-diff --git a/accel/kvm/kvm-all.c b/accel/kvm/kvm-all.c
+diff --git a/hw/remote/memory.c b/hw/remote/memory.c
 index XXXXXXX..XXXXXXX 100644
---- a/accel/kvm/kvm-all.c
+--- a/hw/remote/memory.c
-+++ b/accel/kvm/kvm-all.c
++++ b/hw/remote/memory.c
-@@ -XXX,XX +XXX,XX @@ static int kvm_init(MachineState *ms)
+@@ -XXX,XX +XXX,XX @@ void remote_sysmem_reconfig(MPQemuMsg *msg, Error **errp)
-     kvm_memory_listener_register(s, &s->memory_listener,
+     remote_sysmem_reset();
-                                  &address_space_memory, 0);
--    memory_listener_register(&kvm_io_listener,
+-    for (region = 0; region < msg->num_fds; region++) {
--                             &address_space_io);
+-        g_autofree char *name;
-+    if (kvm_eventfds_allowed) {
++    for (region = 0; region < msg->num_fds; region++, suffix++) {
-+        memory_listener_register(&kvm_io_listener,
++        g_autofree char *name = g_strdup_printf("remote-mem-%u", suffix);
-+                                 &address_space_io);
+         subregion = g_new(MemoryRegion, 1);
-+    }
+-        name = g_strdup_printf("remote-mem-%u", suffix++);
-     memory_listener_register(&kvm_coalesced_pio_listener,
+         memory_region_init_ram_from_fd(subregion, NULL,
-                              &address_space_io);
+                                        name, sysmem_info->sizes[region],
+                                        true, msg->fds[region],
 diff --git a/hw/remote/proxy.c b/hw/remote/proxy.c
 index XXXXXXX..XXXXXXX 100644
 --- a/hw/remote/proxy.c
 +++ b/hw/remote/proxy.c
@@ -XXX,XX +XXX,XX @@ static void probe_pci_info(PCIDevice *dev, Error **errp)
                     PCI_BASE_ADDRESS_SPACE_IO : PCI_BASE_ADDRESS_SPACE_MEMORY;
          if (size) {
 -            g_autofree char *name;
 +            g_autofree char *name = g_strdup_printf("bar-region-%d", i);
              pdev->region[i].dev = pdev;
              pdev->region[i].present = true;
              if (type == PCI_BASE_ADDRESS_SPACE_MEMORY) {
                  pdev->region[i].memory = true;
              }
 -            name = g_strdup_printf("bar-region-%d", i);
              memory_region_init_io(&pdev->region[i].mr, OBJECT(pdev),
                                    &proxy_mr_ops, &pdev->region[i],
                                    name, size);
 --
-.28.0
+.31.1

-[PULL 26/33] block/nvme: Fix nvme_submit_command() on big-endian host
+[PULL 2/8] bitops.h: Improve find_xxx_bit() documentation
 From: Philippe Mathieu-Daudé <philmd@redhat.com>
-The Completion Queue Command Identifier is a 16-bit value,
+Document the following functions return the bitmap size
-so nvme_submit_command() is unlikely to work on big-endian
+if no matching bit is found:
 hosts, as the relevant bits are truncated.
 Fix by using the correct byte-swap function.
-Fixes: bdd6a90a9e5 ("block: Add VFIO based NVMe driver")
+- find_first_bit
-Reported-by: Keith Busch <kbusch@kernel.org>
+- find_next_bit
 - find_last_bit
 - find_first_zero_bit
 - find_next_zero_bit
 Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
 Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
 Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Message-id: 20201029093306.1063879-25-philmd@redhat.com
+Message-id: 20210510200758.2623154-2-philmd@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
 ---
- block/nvme.c | 2 +-
+ include/qemu/bitops.h | 15 ++++++++++++---
-file changed, 1 insertion(+), 1 deletion(-)
+file changed, 12 insertions(+), 3 deletions(-)
-diff --git a/block/nvme.c b/block/nvme.c
+diff --git a/include/qemu/bitops.h b/include/qemu/bitops.h
 index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
+--- a/include/qemu/bitops.h
-+++ b/block/nvme.c
++++ b/include/qemu/bitops.h
-@@ -XXX,XX +XXX,XX @@ static void nvme_submit_command(NVMeQueuePair *q, NVMeRequest *req,
+@@ -XXX,XX +XXX,XX @@ static inline int test_bit(long nr, const unsigned long *addr)
-     assert(!req->cb);
+  * @addr: The address to start the search at
-     req->cb = cb;
+  * @size: The maximum size to search
-     req->opaque = opaque;
+  *
--    cmd->cid = cpu_to_le32(req->cid);
+- * Returns the bit number of the first set bit, or size.
-+    cmd->cid = cpu_to_le16(req->cid);
++ * Returns the bit number of the last set bit,
++ * or @size if there is no set bit in the bitmap.
-     trace_nvme_submit_command(q->s, q->index, req->cid);
+  */
-     nvme_trace_command(cmd);
+ unsigned long find_last_bit(const unsigned long *addr,
                              unsigned long size);
@@ -XXX,XX +XXX,XX @@ unsigned long find_last_bit(const unsigned long *addr,
   * @addr: The address to base the search on
   * @offset: The bitnumber to start searching at
   * @size: The bitmap size in bits
 + *
 + * Returns the bit number of the next set bit,
 + * or @size if there are no further set bits in the bitmap.
   */
  unsigned long find_next_bit(const unsigned long *addr,
                              unsigned long size,
@@ -XXX,XX +XXX,XX @@ unsigned long find_next_bit(const unsigned long *addr,
   * @addr: The address to base the search on
   * @offset: The bitnumber to start searching at
   * @size: The bitmap size in bits
 + *
 + * Returns the bit number of the next cleared bit,
 + * or @size if there are no further clear bits in the bitmap.
   */
  unsigned long find_next_zero_bit(const unsigned long *addr,
@@ -XXX,XX +XXX,XX @@ unsigned long find_next_zero_bit(const unsigned long *addr,
   * @addr: The address to start the search at
   * @size: The maximum size to search
   *
 - * Returns the bit number of the first set bit.
 + * Returns the bit number of the first set bit,
 + * or @size if there is no set bit in the bitmap.
   */
  static inline unsigned long find_first_bit(const unsigned long *addr,
                                             unsigned long size)
@@ -XXX,XX +XXX,XX @@ static inline unsigned long find_first_bit(const unsigned long *addr,
   * @addr: The address to start the search at
   * @size: The maximum size to search
   *
 - * Returns the bit number of the first cleared bit.
 + * Returns the bit number of the first cleared bit,
 + * or @size if there is no clear bit in the bitmap.
   */
  static inline unsigned long find_first_zero_bit(const unsigned long *addr,
                                                  unsigned long size)
 --
-.28.0
+.31.1

-[PULL 02/33] softmmu/memory: fix memory_region_ioeventfd_equal()
+[PULL 3/8] coroutine-sleep: use a stack-allocated timer
-From: Elena Afanasova <eafanasova@gmail.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-Eventfd can be registered with a zero length when fast_mmio is true.
+The lifetime of the timer is well-known (it cannot outlive
-Handle this case properly when dispatching through QEMU.
+qemu_co_sleep_ns_wakeable, because it's deleted by the time the
 coroutine resumes), so it is not necessary to place it on the heap.
-Signed-off-by: Elena Afanasova <eafanasova@gmail.com>
+Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Message-id: cf71a62eb04e61932ff8ffdd02e0b2aab4f495a0.camel@gmail.com
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
 Message-id: 20210517100548.28806-2-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
 ---
- softmmu/memory.c | 11 +++++++++--
+ util/qemu-coroutine-sleep.c | 9 ++++-----
-file changed, 9 insertions(+), 2 deletions(-)
+file changed, 4 insertions(+), 5 deletions(-)
-diff --git a/softmmu/memory.c b/softmmu/memory.c
+diff --git a/util/qemu-coroutine-sleep.c b/util/qemu-coroutine-sleep.c
 index XXXXXXX..XXXXXXX 100644
---- a/softmmu/memory.c
+--- a/util/qemu-coroutine-sleep.c
-+++ b/softmmu/memory.c
++++ b/util/qemu-coroutine-sleep.c
-@@ -XXX,XX +XXX,XX @@ static bool memory_region_ioeventfd_before(MemoryRegionIoeventfd *a,
+@@ -XXX,XX +XXX,XX @@ static const char *qemu_co_sleep_ns__scheduled = "qemu_co_sleep_ns";
- static bool memory_region_ioeventfd_equal(MemoryRegionIoeventfd *a,
-                                           MemoryRegionIoeventfd *b)
+ struct QemuCoSleepState {
- {
+     Coroutine *co;
--    return !memory_region_ioeventfd_before(a, b)
+-    QEMUTimer *ts;
--        && !memory_region_ioeventfd_before(b, a);
++    QEMUTimer ts;
-+    if (int128_eq(a->addr.start, b->addr.start) &&
+     QemuCoSleepState **user_state_pointer;
-+        (!int128_nz(a->addr.size) || !int128_nz(b->addr.size) ||
+ };
-+         (int128_eq(a->addr.size, b->addr.size) &&
-+          (a->match_data == b->match_data) &&
+@@ -XXX,XX +XXX,XX @@ void qemu_co_sleep_wake(QemuCoSleepState *sleep_state)
-+          ((a->match_data && (a->data == b->data)) || !a->match_data) &&
+     if (sleep_state->user_state_pointer) {
-+          (a->e == b->e))))
+         *sleep_state->user_state_pointer = NULL;
-+        return true;
+     }
-+
+-    timer_del(sleep_state->ts);
-+    return false;
++    timer_del(&sleep_state->ts);
      aio_co_wake(sleep_state->co);
  }
- /* Range of memory in the global map.  Addresses are absolute. */
+@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
      AioContext *ctx = qemu_get_current_aio_context();
      QemuCoSleepState state = {
          .co = qemu_coroutine_self(),
 -        .ts = aio_timer_new(ctx, type, SCALE_NS, co_sleep_cb, &state),
          .user_state_pointer = sleep_state,
      };
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
          abort();
      }
 +    aio_timer_init(ctx, &state.ts, type, SCALE_NS, co_sleep_cb, &state);
      if (sleep_state) {
          *sleep_state = &state;
      }
 -    timer_mod(state.ts, qemu_clock_get_ns(type) + ns);
 +    timer_mod(&state.ts, qemu_clock_get_ns(type) + ns);
      qemu_coroutine_yield();
      if (sleep_state) {
          /*
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
           */
          assert(*sleep_state == NULL);
      }
 -    timer_free(state.ts);
  }
 --
-.28.0
+.31.1

-[PULL 03/33] MAINTAINERS: Cover "block/nvme.h" file
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-The "block/nvme.h" header is shared by both the NVMe block
-driver and the NVMe emulated device. Add the 'F:' entry on
-both sections, so all maintainers/reviewers are notified
-when it is changed.
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
-Message-Id: <20200701140634.25994-1-philmd@redhat.com>
----
- MAINTAINERS | 2 ++
-file changed, 2 insertions(+)
-diff --git a/MAINTAINERS b/MAINTAINERS
-index XXXXXXX..XXXXXXX 100644
---- a/MAINTAINERS
-+++ b/MAINTAINERS
-@@ -XXX,XX +XXX,XX @@ M: Klaus Jensen <its@irrelevant.dk>
- L: qemu-block@nongnu.org
- S: Supported
- F: hw/block/nvme*
-+F: include/block/nvme.h
- F: tests/qtest/nvme-test.c
- F: docs/specs/nvme.txt
- T: git git://git.infradead.org/qemu-nvme.git nvme-next
-@@ -XXX,XX +XXX,XX @@ R: Fam Zheng <fam@euphon.net>
- L: qemu-block@nongnu.org
- S: Supported
- F: block/nvme*
-+F: include/block/nvme.h
- T: git https://github.com/stefanha/qemu.git block
- Bootdevice
---
-.28.0

-[PULL 04/33] block/nvme: Use hex format to display offset in trace events
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-Use the same format used for the hw/vfio/ trace events.
-Suggested-by: Eric Auger <eric.auger@redhat.com>
-Reviewed-by: Eric Auger <eric.auger@redhat.com>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201029093306.1063879-3-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- block/trace-events | 12 ++++++------
-file changed, 6 insertions(+), 6 deletions(-)
-diff --git a/block/trace-events b/block/trace-events
-index XXXXXXX..XXXXXXX 100644
---- a/block/trace-events
-+++ b/block/trace-events
-@@ -XXX,XX +XXX,XX @@ nvme_submit_command(void *s, int index, int cid) "s %p queue %d cid %d"
- nvme_submit_command_raw(int c0, int c1, int c2, int c3, int c4, int c5, int c6, int c7) "%02x %02x %02x %02x %02x %02x %02x %02x"
- nvme_handle_event(void *s) "s %p"
- nvme_poll_cb(void *s) "s %p"
--nvme_prw_aligned(void *s, int is_write, uint64_t offset, uint64_t bytes, int flags, int niov) "s %p is_write %d offset %"PRId64" bytes %"PRId64" flags %d niov %d"
--nvme_write_zeroes(void *s, uint64_t offset, uint64_t bytes, int flags) "s %p offset %"PRId64" bytes %"PRId64" flags %d"
-+nvme_prw_aligned(void *s, int is_write, uint64_t offset, uint64_t bytes, int flags, int niov) "s %p is_write %d offset 0x%"PRIx64" bytes %"PRId64" flags %d niov %d"
-+nvme_write_zeroes(void *s, uint64_t offset, uint64_t bytes, int flags) "s %p offset 0x%"PRIx64" bytes %"PRId64" flags %d"
- nvme_qiov_unaligned(const void *qiov, int n, void *base, size_t size, int align) "qiov %p n %d base %p size 0x%zx align 0x%x"
--nvme_prw_buffered(void *s, uint64_t offset, uint64_t bytes, int niov, int is_write) "s %p offset %"PRId64" bytes %"PRId64" niov %d is_write %d"
--nvme_rw_done(void *s, int is_write, uint64_t offset, uint64_t bytes, int ret) "s %p is_write %d offset %"PRId64" bytes %"PRId64" ret %d"
--nvme_dsm(void *s, uint64_t offset, uint64_t bytes) "s %p offset %"PRId64" bytes %"PRId64""
--nvme_dsm_done(void *s, uint64_t offset, uint64_t bytes, int ret) "s %p offset %"PRId64" bytes %"PRId64" ret %d"
-+nvme_prw_buffered(void *s, uint64_t offset, uint64_t bytes, int niov, int is_write) "s %p offset 0x%"PRIx64" bytes %"PRId64" niov %d is_write %d"
-+nvme_rw_done(void *s, int is_write, uint64_t offset, uint64_t bytes, int ret) "s %p is_write %d offset 0x%"PRIx64" bytes %"PRId64" ret %d"
-+nvme_dsm(void *s, uint64_t offset, uint64_t bytes) "s %p offset 0x%"PRIx64" bytes %"PRId64""
-+nvme_dsm_done(void *s, uint64_t offset, uint64_t bytes, int ret) "s %p offset 0x%"PRIx64" bytes %"PRId64" ret %d"
- nvme_dma_map_flush(void *s) "s %p"
- nvme_free_req_queue_wait(void *q) "q %p"
- nvme_cmd_map_qiov(void *s, void *cmd, void *req, void *qiov, int entries) "s %p cmd %p req %p qiov %p entries %d"
---
-.28.0

-[PULL 05/33] block/nvme: Report warning with warn_report()
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-Instead of displaying warning on stderr, use warn_report()
-which also displays it on the monitor.
-Reviewed-by: Eric Auger <eric.auger@redhat.com>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201029093306.1063879-4-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- block/nvme.c | 4 ++--
-file changed, 2 insertions(+), 2 deletions(-)
-diff --git a/block/nvme.c b/block/nvme.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
-+++ b/block/nvme.c
-@@ -XXX,XX +XXX,XX @@ static bool nvme_process_completion(NVMeQueuePair *q)
-         }
-         cid = le16_to_cpu(c->cid);
-         if (cid == 0 || cid > NVME_QUEUE_SIZE) {
--            fprintf(stderr, "Unexpected CID in completion queue: %" PRIu32 "\n",
--                    cid);
-+            warn_report("NVMe: Unexpected CID in completion queue: %"PRIu32", "
-+                        "queue size: %u", cid, NVME_QUEUE_SIZE);
-             continue;
-         }
-         trace_nvme_complete_command(s, q->index, cid);
---
-.28.0

-[PULL 06/33] block/nvme: Trace controller capabilities
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-Controllers have different capabilities and report them in the
-CAP register. We are particularly interested by the page size
-limits.
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Reviewed-by: Eric Auger <eric.auger@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201029093306.1063879-5-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- block/nvme.c       | 13 +++++++++++++
- block/trace-events |  2 ++
-files changed, 15 insertions(+)
-diff --git a/block/nvme.c b/block/nvme.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
-+++ b/block/nvme.c
-@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
-      * Initialization". */
-     cap = le64_to_cpu(regs->cap);
-+    trace_nvme_controller_capability_raw(cap);
-+    trace_nvme_controller_capability("Maximum Queue Entries Supported",
-+                                     1 + NVME_CAP_MQES(cap));
-+    trace_nvme_controller_capability("Contiguous Queues Required",
-+                                     NVME_CAP_CQR(cap));
-+    trace_nvme_controller_capability("Doorbell Stride",
-+                                     2 << (2 + NVME_CAP_DSTRD(cap)));
-+    trace_nvme_controller_capability("Subsystem Reset Supported",
-+                                     NVME_CAP_NSSRS(cap));
-+    trace_nvme_controller_capability("Memory Page Size Minimum",
-+                                     1 << (12 + NVME_CAP_MPSMIN(cap)));
-+    trace_nvme_controller_capability("Memory Page Size Maximum",
-+                                     1 << (12 + NVME_CAP_MPSMAX(cap)));
-     if (!NVME_CAP_CSS(cap)) {
-         error_setg(errp, "Device doesn't support NVMe command set");
-         ret = -EINVAL;
-diff --git a/block/trace-events b/block/trace-events
-index XXXXXXX..XXXXXXX 100644
---- a/block/trace-events
-+++ b/block/trace-events
-@@ -XXX,XX +XXX,XX @@ qed_aio_write_postfill(void *s, void *acb, uint64_t start, size_t len, uint64_t
- qed_aio_write_main(void *s, void *acb, int ret, uint64_t offset, size_t len) "s %p acb %p ret %d offset %"PRIu64" len %zu"
- # nvme.c
-+nvme_controller_capability_raw(uint64_t value) "0x%08"PRIx64
-+nvme_controller_capability(const char *desc, uint64_t value) "%s: %"PRIu64
- nvme_kick(void *s, int queue) "s %p queue %d"
- nvme_dma_flush_queue_wait(void *s) "s %p"
- nvme_error(int cmd_specific, int sq_head, int sqid, int cid, int status) "cmd_specific %d sq_head %d sqid %d cid %d status 0x%x"
---
-.28.0

-[PULL 07/33] block/nvme: Trace nvme_poll_queue() per queue
+[PULL 4/8] coroutine-sleep: disallow NULL QemuCoSleepState** argument
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-As we want to enable multiple queues, report the event
+Simplify the code by removing conditionals.  qemu_co_sleep_ns
-in each nvme_poll_queue() call, rather than once in
+can simply point the argument to an on-stack temporary.
 the callback calling nvme_poll_queues().
-Reviewed-by: Eric Auger <eric.auger@redhat.com>
+Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
+Message-id: 20210517100548.28806-3-pbonzini@redhat.com
 Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
 Message-id: 20201029093306.1063879-6-philmd@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
 ---
- block/nvme.c       | 2 +-
+ include/qemu/coroutine.h    |  5 +++--
- block/trace-events | 2 +-
+ util/qemu-coroutine-sleep.c | 18 +++++-------------
-files changed, 2 insertions(+), 2 deletions(-)
+files changed, 8 insertions(+), 15 deletions(-)
-diff --git a/block/nvme.c b/block/nvme.c
+diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
 index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
+--- a/include/qemu/coroutine.h
-+++ b/block/nvme.c
++++ b/include/qemu/coroutine.h
-@@ -XXX,XX +XXX,XX @@ static bool nvme_poll_queue(NVMeQueuePair *q)
+@@ -XXX,XX +XXX,XX @@ typedef struct QemuCoSleepState QemuCoSleepState;
-     const size_t cqe_offset = q->cq.head * NVME_CQ_ENTRY_BYTES;
-     NvmeCqe *cqe = (NvmeCqe *)&q->cq.queue[cqe_offset];
+ /**
+  * Yield the coroutine for a given duration. During this yield, @sleep_state
-+    trace_nvme_poll_queue(q->s, q->index);
+- * (if not NULL) is set to an opaque pointer, which may be used for
-     /*
++ * is set to an opaque pointer, which may be used for
-      * Do an early check for completions. q->lock isn't needed because
+  * qemu_co_sleep_wake(). Be careful, the pointer is set back to zero when the
-      * nvme_process_completion() only runs in the event loop thread and
+  * timer fires. Don't save the obtained value to other variables and don't call
-@@ -XXX,XX +XXX,XX @@ static bool nvme_poll_cb(void *opaque)
+  * qemu_co_sleep_wake from another aio context.
-     BDRVNVMeState *s = container_of(e, BDRVNVMeState,
+@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
-                                     irq_notifier[MSIX_SHARED_IRQ_IDX]);
+                                             QemuCoSleepState **sleep_state);
+ static inline void coroutine_fn qemu_co_sleep_ns(QEMUClockType type, int64_t ns)
--    trace_nvme_poll_cb(s);
+ {
-     return nvme_poll_queues(s);
+-    qemu_co_sleep_ns_wakeable(type, ns, NULL);
 +    QemuCoSleepState *unused = NULL;
 +    qemu_co_sleep_ns_wakeable(type, ns, &unused);
  }
-diff --git a/block/trace-events b/block/trace-events
+ /**
 diff --git a/util/qemu-coroutine-sleep.c b/util/qemu-coroutine-sleep.c
 index XXXXXXX..XXXXXXX 100644
---- a/block/trace-events
+--- a/util/qemu-coroutine-sleep.c
-+++ b/block/trace-events
++++ b/util/qemu-coroutine-sleep.c
-@@ -XXX,XX +XXX,XX @@ nvme_complete_command(void *s, int index, int cid) "s %p queue %d cid %d"
+@@ -XXX,XX +XXX,XX @@ void qemu_co_sleep_wake(QemuCoSleepState *sleep_state)
- nvme_submit_command(void *s, int index, int cid) "s %p queue %d cid %d"
+                                            qemu_co_sleep_ns__scheduled, NULL);
- nvme_submit_command_raw(int c0, int c1, int c2, int c3, int c4, int c5, int c6, int c7) "%02x %02x %02x %02x %02x %02x %02x %02x"
- nvme_handle_event(void *s) "s %p"
+     assert(scheduled == qemu_co_sleep_ns__scheduled);
--nvme_poll_cb(void *s) "s %p"
+-    if (sleep_state->user_state_pointer) {
-+nvme_poll_queue(void *s, unsigned q_index) "s %p q #%u"
+-        *sleep_state->user_state_pointer = NULL;
- nvme_prw_aligned(void *s, int is_write, uint64_t offset, uint64_t bytes, int flags, int niov) "s %p is_write %d offset 0x%"PRIx64" bytes %"PRId64" flags %d niov %d"
+-    }
- nvme_write_zeroes(void *s, uint64_t offset, uint64_t bytes, int flags) "s %p offset 0x%"PRIx64" bytes %"PRId64" flags %d"
++    *sleep_state->user_state_pointer = NULL;
- nvme_qiov_unaligned(const void *qiov, int n, void *base, size_t size, int align) "qiov %p n %d base %p size 0x%zx align 0x%x"
+     timer_del(&sleep_state->ts);
      aio_co_wake(sleep_state->co);
  }
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
      }
      aio_timer_init(ctx, &state.ts, type, SCALE_NS, co_sleep_cb, &state);
 -    if (sleep_state) {
 -        *sleep_state = &state;
 -    }
 +    *sleep_state = &state;
      timer_mod(&state.ts, qemu_clock_get_ns(type) + ns);
      qemu_coroutine_yield();
 -    if (sleep_state) {
 -        /*
 -         * Note that *sleep_state is cleared during qemu_co_sleep_wake
 -         * before resuming this coroutine.
 -         */
 -        assert(*sleep_state == NULL);
 -    }
 +
 +    /* qemu_co_sleep_wake clears *sleep_state before resuming this coroutine.  */
 +    assert(*sleep_state == NULL);
  }
 --
-.28.0
+.31.1

-[PULL 08/33] block/nvme: Improve nvme_free_req_queue_wait() trace information
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-What we want to trace is the block driver state and the queue index.
-Suggested-by: Stefan Hajnoczi <stefanha@redhat.com>
-Reviewed-by: Eric Auger <eric.auger@redhat.com>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201029093306.1063879-7-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- block/nvme.c       | 2 +-
- block/trace-events | 2 +-
-files changed, 2 insertions(+), 2 deletions(-)
-diff --git a/block/nvme.c b/block/nvme.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
-+++ b/block/nvme.c
-@@ -XXX,XX +XXX,XX @@ static NVMeRequest *nvme_get_free_req(NVMeQueuePair *q)
-     while (q->free_req_head == -1) {
-         if (qemu_in_coroutine()) {
--            trace_nvme_free_req_queue_wait(q);
-+            trace_nvme_free_req_queue_wait(q->s, q->index);
-             qemu_co_queue_wait(&q->free_req_queue, &q->lock);
-         } else {
-             qemu_mutex_unlock(&q->lock);
-diff --git a/block/trace-events b/block/trace-events
-index XXXXXXX..XXXXXXX 100644
---- a/block/trace-events
-+++ b/block/trace-events
-@@ -XXX,XX +XXX,XX @@ nvme_rw_done(void *s, int is_write, uint64_t offset, uint64_t bytes, int ret) "s
- nvme_dsm(void *s, uint64_t offset, uint64_t bytes) "s %p offset 0x%"PRIx64" bytes %"PRId64""
- nvme_dsm_done(void *s, uint64_t offset, uint64_t bytes, int ret) "s %p offset 0x%"PRIx64" bytes %"PRId64" ret %d"
- nvme_dma_map_flush(void *s) "s %p"
--nvme_free_req_queue_wait(void *q) "q %p"
-+nvme_free_req_queue_wait(void *s, unsigned q_index) "s %p q #%u"
- nvme_cmd_map_qiov(void *s, void *cmd, void *req, void *qiov, int entries) "s %p cmd %p req %p qiov %p entries %d"
- nvme_cmd_map_qiov_pages(void *s, int i, uint64_t page) "s %p page[%d] 0x%"PRIx64
- nvme_cmd_map_qiov_iov(void *s, int i, void *page, int pages) "s %p iov[%d] %p pages %d"
---
-.28.0

-[PULL 09/33] block/nvme: Trace queue pair creation/deletion
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-Reviewed-by: Eric Auger <eric.auger@redhat.com>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201029093306.1063879-8-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- block/nvme.c       | 3 +++
- block/trace-events | 2 ++
-files changed, 5 insertions(+)
-diff --git a/block/nvme.c b/block/nvme.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
-+++ b/block/nvme.c
-@@ -XXX,XX +XXX,XX @@ static void nvme_init_queue(BDRVNVMeState *s, NVMeQueue *q,
- static void nvme_free_queue_pair(NVMeQueuePair *q)
- {
-+    trace_nvme_free_queue_pair(q->index, q);
-     if (q->completion_bh) {
-         qemu_bh_delete(q->completion_bh);
-     }
-@@ -XXX,XX +XXX,XX @@ static NVMeQueuePair *nvme_create_queue_pair(BDRVNVMeState *s,
-     if (!q) {
-         return NULL;
-     }
-+    trace_nvme_create_queue_pair(idx, q, size, aio_context,
-+                                 event_notifier_get_fd(s->irq_notifier));
-     q->prp_list_pages = qemu_try_memalign(s->page_size,
-                                           s->page_size * NVME_NUM_REQS);
-     if (!q->prp_list_pages) {
-diff --git a/block/trace-events b/block/trace-events
-index XXXXXXX..XXXXXXX 100644
---- a/block/trace-events
-+++ b/block/trace-events
-@@ -XXX,XX +XXX,XX @@ nvme_dsm(void *s, uint64_t offset, uint64_t bytes) "s %p offset 0x%"PRIx64" byte
- nvme_dsm_done(void *s, uint64_t offset, uint64_t bytes, int ret) "s %p offset 0x%"PRIx64" bytes %"PRId64" ret %d"
- nvme_dma_map_flush(void *s) "s %p"
- nvme_free_req_queue_wait(void *s, unsigned q_index) "s %p q #%u"
-+nvme_create_queue_pair(unsigned q_index, void *q, unsigned size, void *aio_context, int fd) "index %u q %p size %u aioctx %p fd %d"
-+nvme_free_queue_pair(unsigned q_index, void *q) "index %u q %p"
- nvme_cmd_map_qiov(void *s, void *cmd, void *req, void *qiov, int entries) "s %p cmd %p req %p qiov %p entries %d"
- nvme_cmd_map_qiov_pages(void *s, int i, uint64_t page) "s %p page[%d] 0x%"PRIx64
- nvme_cmd_map_qiov_iov(void *s, int i, void *page, int pages) "s %p iov[%d] %p pages %d"
---
-.28.0

-[PULL 10/33] block/nvme: Move definitions before structure declarations
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-To be able to use some definitions in structure declarations,
-move them earlier. No logical change.
-Reviewed-by: Eric Auger <eric.auger@redhat.com>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201029093306.1063879-9-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- block/nvme.c | 19 ++++++++++---------
-file changed, 10 insertions(+), 9 deletions(-)
-diff --git a/block/nvme.c b/block/nvme.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
-+++ b/block/nvme.c
-@@ -XXX,XX +XXX,XX @@
- typedef struct BDRVNVMeState BDRVNVMeState;
-+/* Same index is used for queues and IRQs */
-+#define INDEX_ADMIN     0
-+#define INDEX_IO(n)     (1 + n)
-+
-+/* This driver shares a single MSIX IRQ for the admin and I/O queues */
-+enum {
-+    MSIX_SHARED_IRQ_IDX = 0,
-+    MSIX_IRQ_COUNT = 1
-+};
-+
- typedef struct {
-     int32_t  head, tail;
-     uint8_t  *queue;
-@@ -XXX,XX +XXX,XX @@ typedef struct {
-     QEMUBH      *completion_bh;
- } NVMeQueuePair;
--#define INDEX_ADMIN     0
--#define INDEX_IO(n)     (1 + n)
--
--/* This driver shares a single MSIX IRQ for the admin and I/O queues */
--enum {
--    MSIX_SHARED_IRQ_IDX = 0,
--    MSIX_IRQ_COUNT = 1
--};
--
- struct BDRVNVMeState {
-     AioContext *aio_context;
-     QEMUVFIOState *vfio;
---
-.28.0

-[PULL 18/33] block/nvme: Simplify nvme_cmd_sync()
+[PULL 5/8] coroutine-sleep: allow qemu_co_sleep_wake that wakes nothing
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-As all commands use the ADMIN queue, it is pointless to pass
+All callers of qemu_co_sleep_wake are checking whether they are passing
-it as argument each time. Remove the argument, and rename the
+a NULL argument inside the pointer-to-pointer: do the check in
-function as nvme_admin_cmd_sync() to make this new behavior
+qemu_co_sleep_wake itself.
 clearer.
-Reviewed-by: Eric Auger <eric.auger@redhat.com>
+As a side effect, qemu_co_sleep_wake can be called more than once and
-Tested-by: Eric Auger <eric.auger@redhat.com>
+it will only wake the coroutine once; after the first time, the argument
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
+will be set to NULL via *sleep_state->user_state_pointer.  However, this
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
+would not be safe unless co_sleep_cb keeps using the QemuCoSleepState*
-Message-id: 20201029093306.1063879-17-philmd@redhat.com
+directly, so make it go through the pointer-to-pointer instead.
 Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
 Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
 Message-id: 20210517100548.28806-4-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
 ---
- block/nvme.c | 19 ++++++++++---------
+ block/block-copy.c          |  4 +---
-file changed, 10 insertions(+), 9 deletions(-)
+ block/nbd.c                 |  8 ++------
  util/qemu-coroutine-sleep.c | 21 ++++++++++++---------
 files changed, 15 insertions(+), 18 deletions(-)
-diff --git a/block/nvme.c b/block/nvme.c
+diff --git a/block/block-copy.c b/block/block-copy.c
 index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
+--- a/block/block-copy.c
-+++ b/block/nvme.c
++++ b/block/block-copy.c
-@@ -XXX,XX +XXX,XX @@ static void nvme_submit_command(NVMeQueuePair *q, NVMeRequest *req,
+@@ -XXX,XX +XXX,XX @@ out:
-     qemu_mutex_unlock(&q->lock);
  void block_copy_kick(BlockCopyCallState *call_state)
  {
 -    if (call_state->sleep_state) {
 -        qemu_co_sleep_wake(call_state->sleep_state);
 -    }
 +    qemu_co_sleep_wake(call_state->sleep_state);
  }
--static void nvme_cmd_sync_cb(void *opaque, int ret)
+ /*
-+static void nvme_admin_cmd_sync_cb(void *opaque, int ret)
+diff --git a/block/nbd.c b/block/nbd.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/nbd.c
 +++ b/block/nbd.c
@@ -XXX,XX +XXX,XX @@ static void coroutine_fn nbd_client_co_drain_begin(BlockDriverState *bs)
      BDRVNBDState *s = (BDRVNBDState *)bs->opaque;
      s->drained = true;
 -    if (s->connection_co_sleep_ns_state) {
 -        qemu_co_sleep_wake(s->connection_co_sleep_ns_state);
 -    }
 +    qemu_co_sleep_wake(s->connection_co_sleep_ns_state);
      nbd_co_establish_connection_cancel(bs, false);
@@ -XXX,XX +XXX,XX @@ static void nbd_teardown_connection(BlockDriverState *bs)
      s->state = NBD_CLIENT_QUIT;
      if (s->connection_co) {
 -        if (s->connection_co_sleep_ns_state) {
 -            qemu_co_sleep_wake(s->connection_co_sleep_ns_state);
 -        }
 +        qemu_co_sleep_wake(s->connection_co_sleep_ns_state);
          nbd_co_establish_connection_cancel(bs, true);
      }
      if (qemu_in_coroutine()) {
 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 @@ struct QemuCoSleepState {
  void qemu_co_sleep_wake(QemuCoSleepState *sleep_state)
  {
-     int *pret = opaque;
+-    /* Write of schedule protected by barrier write in aio_co_schedule */
-     *pret = ret;
+-    const char *scheduled = qatomic_cmpxchg(&sleep_state->co->scheduled,
-     aio_wait_kick();
+-                                           qemu_co_sleep_ns__scheduled, NULL);
 +    if (sleep_state) {
 +        /* Write of schedule protected by barrier write in aio_co_schedule */
 +        const char *scheduled = qatomic_cmpxchg(&sleep_state->co->scheduled,
 +                                                qemu_co_sleep_ns__scheduled, NULL);
 -    assert(scheduled == qemu_co_sleep_ns__scheduled);
 -    *sleep_state->user_state_pointer = NULL;
 -    timer_del(&sleep_state->ts);
 -    aio_co_wake(sleep_state->co);
 +        assert(scheduled == qemu_co_sleep_ns__scheduled);
 +        *sleep_state->user_state_pointer = NULL;
 +        timer_del(&sleep_state->ts);
 +        aio_co_wake(sleep_state->co);
 +    }
  }
--static int nvme_cmd_sync(BlockDriverState *bs, NVMeQueuePair *q,
+ static void co_sleep_cb(void *opaque)
 -                         NvmeCmd *cmd)
 +static int nvme_admin_cmd_sync(BlockDriverState *bs, NvmeCmd *cmd)
  {
-+    BDRVNVMeState *s = bs->opaque;
+-    qemu_co_sleep_wake(opaque);
-+    NVMeQueuePair *q = s->queues[INDEX_ADMIN];
++    QemuCoSleepState **sleep_state = opaque;
-     AioContext *aio_context = bdrv_get_aio_context(bs);
++    qemu_co_sleep_wake(*sleep_state);
-     NVMeRequest *req;
+ }
-     int ret = -EINPROGRESS;
-@@ -XXX,XX +XXX,XX @@ static int nvme_cmd_sync(BlockDriverState *bs, NVMeQueuePair *q,
+ void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
-     if (!req) {
+@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
-         return -EBUSY;
+         abort();
      }
--    nvme_submit_command(q, req, cmd, nvme_cmd_sync_cb, &ret);
-+    nvme_submit_command(q, req, cmd, nvme_admin_cmd_sync_cb, &ret);
+-    aio_timer_init(ctx, &state.ts, type, SCALE_NS, co_sleep_cb, &state);
++    aio_timer_init(ctx, &state.ts, type, SCALE_NS, co_sleep_cb, sleep_state);
-     AIO_WAIT_WHILE(aio_context, ret == -EINPROGRESS);
+     *sleep_state = &state;
-     return ret;
+     timer_mod(&state.ts, qemu_clock_get_ns(type) + ns);
-@@ -XXX,XX +XXX,XX @@ static bool nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
+     qemu_coroutine_yield();
      memset(id, 0, sizeof(*id));
      cmd.dptr.prp1 = cpu_to_le64(iova);
 -    if (nvme_cmd_sync(bs, s->queues[INDEX_ADMIN], &cmd)) {
 +    if (nvme_admin_cmd_sync(bs, &cmd)) {
          error_setg(errp, "Failed to identify controller");
          goto out;
      }
@@ -XXX,XX +XXX,XX @@ static bool nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
      memset(id, 0, sizeof(*id));
      cmd.cdw10 = 0;
      cmd.nsid = cpu_to_le32(namespace);
 -    if (nvme_cmd_sync(bs, s->queues[INDEX_ADMIN], &cmd)) {
 +    if (nvme_admin_cmd_sync(bs, &cmd)) {
          error_setg(errp, "Failed to identify namespace");
          goto out;
      }
@@ -XXX,XX +XXX,XX @@ static bool nvme_add_io_queue(BlockDriverState *bs, Error **errp)
          .cdw10 = cpu_to_le32(((queue_size - 1) << 16) | n),
          .cdw11 = cpu_to_le32(NVME_CQ_IEN | NVME_CQ_PC),
      };
 -    if (nvme_cmd_sync(bs, s->queues[INDEX_ADMIN], &cmd)) {
 +    if (nvme_admin_cmd_sync(bs, &cmd)) {
          error_setg(errp, "Failed to create CQ io queue [%u]", n);
          goto out_error;
      }
@@ -XXX,XX +XXX,XX @@ static bool nvme_add_io_queue(BlockDriverState *bs, Error **errp)
          .cdw10 = cpu_to_le32(((queue_size - 1) << 16) | n),
          .cdw11 = cpu_to_le32(NVME_SQ_PC | (n << 16)),
      };
 -    if (nvme_cmd_sync(bs, s->queues[INDEX_ADMIN], &cmd)) {
 +    if (nvme_admin_cmd_sync(bs, &cmd)) {
          error_setg(errp, "Failed to create SQ io queue [%u]", n);
          goto out_error;
      }
@@ -XXX,XX +XXX,XX @@ static int nvme_enable_disable_write_cache(BlockDriverState *bs, bool enable,
          .cdw11 = cpu_to_le32(enable ? 0x01 : 0x00),
      };
 -    ret = nvme_cmd_sync(bs, s->queues[INDEX_ADMIN], &cmd);
 +    ret = nvme_admin_cmd_sync(bs, &cmd);
      if (ret) {
          error_setg(errp, "Failed to configure NVMe write cache");
      }
 --
-.28.0
+.31.1

-[PULL 11/33] block/nvme: Use unsigned integer for queue counter/size
+[PULL 6/8] coroutine-sleep: move timer out of QemuCoSleepState
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-We can not have negative queue count/size/index, use unsigned type.
+This simplification is enabled by the previous patch.  Now aio_co_wake
-Rename 'nr_queues' as 'queue_count' to match the spec naming.
+will only be called once, therefore we do not care about a spurious
 firing of the timer after a qemu_co_sleep_wake.
-Reviewed-by: Eric Auger <eric.auger@redhat.com>
+Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
+Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
+Message-id: 20210517100548.28806-5-pbonzini@redhat.com
 Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
 Message-id: 20201029093306.1063879-10-philmd@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
 ---
- block/nvme.c       | 38 ++++++++++++++++++--------------------
+ util/qemu-coroutine-sleep.c | 8 ++++----
- block/trace-events | 10 +++++-----
+file changed, 4 insertions(+), 4 deletions(-)
 files changed, 23 insertions(+), 25 deletions(-)
-diff --git a/block/nvme.c b/block/nvme.c
+diff --git a/util/qemu-coroutine-sleep.c b/util/qemu-coroutine-sleep.c
 index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
+--- a/util/qemu-coroutine-sleep.c
-+++ b/block/nvme.c
++++ b/util/qemu-coroutine-sleep.c
-@@ -XXX,XX +XXX,XX @@ struct BDRVNVMeState {
+@@ -XXX,XX +XXX,XX @@ static const char *qemu_co_sleep_ns__scheduled = "qemu_co_sleep_ns";
-      * [1..]: io queues.
-      */
+ struct QemuCoSleepState {
-     NVMeQueuePair **queues;
+     Coroutine *co;
--    int nr_queues;
+-    QEMUTimer ts;
-+    unsigned queue_count;
+     QemuCoSleepState **user_state_pointer;
      size_t page_size;
      /* How many uint32_t elements does each doorbell entry take. */
      size_t doorbell_scale;
@@ -XXX,XX +XXX,XX @@ static QemuOptsList runtime_opts = {
  };
- static void nvme_init_queue(BDRVNVMeState *s, NVMeQueue *q,
+@@ -XXX,XX +XXX,XX @@ void qemu_co_sleep_wake(QemuCoSleepState *sleep_state)
--                            int nentries, int entry_bytes, Error **errp)
-+                            unsigned nentries, size_t entry_bytes, Error **errp)
+         assert(scheduled == qemu_co_sleep_ns__scheduled);
          *sleep_state->user_state_pointer = NULL;
 -        timer_del(&sleep_state->ts);
          aio_co_wake(sleep_state->co);
      }
  }
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
                                              QemuCoSleepState **sleep_state)
  {
-     size_t bytes;
+     AioContext *ctx = qemu_get_current_aio_context();
-     int r;
++    QEMUTimer ts;
-@@ -XXX,XX +XXX,XX @@ static void nvme_free_req_queue_cb(void *opaque)
+     QemuCoSleepState state = {
+         .co = qemu_coroutine_self(),
- static NVMeQueuePair *nvme_create_queue_pair(BDRVNVMeState *s,
+         .user_state_pointer = sleep_state,
-                                              AioContext *aio_context,
+@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
--                                             int idx, int size,
+         abort();
 +                                             unsigned idx, size_t size,
                                               Error **errp)
  {
      int i, r;
@@ -XXX,XX +XXX,XX @@ static bool nvme_poll_queues(BDRVNVMeState *s)
      bool progress = false;
      int i;
 -    for (i = 0; i < s->nr_queues; i++) {
 +    for (i = 0; i < s->queue_count; i++) {
          if (nvme_poll_queue(s->queues[i])) {
              progress = true;
          }
@@ -XXX,XX +XXX,XX @@ static void nvme_handle_event(EventNotifier *n)
  static bool nvme_add_io_queue(BlockDriverState *bs, Error **errp)
  {
      BDRVNVMeState *s = bs->opaque;
 -    int n = s->nr_queues;
 +    unsigned n = s->queue_count;
      NVMeQueuePair *q;
      NvmeCmd cmd;
 -    int queue_size = NVME_QUEUE_SIZE;
 +    unsigned queue_size = NVME_QUEUE_SIZE;
      q = nvme_create_queue_pair(s, bdrv_get_aio_context(bs),
                                 n, queue_size, errp);
@@ -XXX,XX +XXX,XX @@ static bool nvme_add_io_queue(BlockDriverState *bs, Error **errp)
          .cdw11 = cpu_to_le32(0x3),
      };
      if (nvme_cmd_sync(bs, s->queues[INDEX_ADMIN], &cmd)) {
 -        error_setg(errp, "Failed to create CQ io queue [%d]", n);
 +        error_setg(errp, "Failed to create CQ io queue [%u]", n);
          goto out_error;
      }
-     cmd = (NvmeCmd) {
-@@ -XXX,XX +XXX,XX @@ static bool nvme_add_io_queue(BlockDriverState *bs, Error **errp)
+-    aio_timer_init(ctx, &state.ts, type, SCALE_NS, co_sleep_cb, sleep_state);
-         .cdw11 = cpu_to_le32(0x1 | (n << 16)),
++    aio_timer_init(ctx, &ts, type, SCALE_NS, co_sleep_cb, sleep_state);
-     };
+     *sleep_state = &state;
-     if (nvme_cmd_sync(bs, s->queues[INDEX_ADMIN], &cmd)) {
+-    timer_mod(&state.ts, qemu_clock_get_ns(type) + ns);
--        error_setg(errp, "Failed to create SQ io queue [%d]", n);
++    timer_mod(&ts, qemu_clock_get_ns(type) + ns);
-+        error_setg(errp, "Failed to create SQ io queue [%u]", n);
+     qemu_coroutine_yield();
-         goto out_error;
++    timer_del(&ts);
-     }
-     s->queues = g_renew(NVMeQueuePair *, s->queues, n + 1);
+     /* qemu_co_sleep_wake clears *sleep_state before resuming this coroutine.  */
-     s->queues[n] = q;
+     assert(*sleep_state == NULL);
 -    s->nr_queues++;
 +    s->queue_count++;
      return true;
  out_error:
      nvme_free_queue_pair(q);
@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
          ret = -EINVAL;
          goto out;
      }
 -    s->nr_queues = 1;
 +    s->queue_count = 1;
      QEMU_BUILD_BUG_ON(NVME_QUEUE_SIZE & 0xF000);
      regs->aqa = cpu_to_le32((NVME_QUEUE_SIZE << AQA_ACQS_SHIFT) |
                              (NVME_QUEUE_SIZE << AQA_ASQS_SHIFT));
@@ -XXX,XX +XXX,XX @@ static int nvme_enable_disable_write_cache(BlockDriverState *bs, bool enable,
  static void nvme_close(BlockDriverState *bs)
  {
 -    int i;
      BDRVNVMeState *s = bs->opaque;
 -    for (i = 0; i < s->nr_queues; ++i) {
 +    for (unsigned i = 0; i < s->queue_count; ++i) {
          nvme_free_queue_pair(s->queues[i]);
      }
      g_free(s->queues);
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int nvme_co_prw_aligned(BlockDriverState *bs,
      };
      trace_nvme_prw_aligned(s, is_write, offset, bytes, flags, qiov->niov);
 -    assert(s->nr_queues > 1);
 +    assert(s->queue_count > 1);
      req = nvme_get_free_req(ioq);
      assert(req);
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int nvme_co_flush(BlockDriverState *bs)
          .ret = -EINPROGRESS,
      };
 -    assert(s->nr_queues > 1);
 +    assert(s->queue_count > 1);
      req = nvme_get_free_req(ioq);
      assert(req);
      nvme_submit_command(ioq, req, &cmd, nvme_rw_cb, &data);
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int nvme_co_pwrite_zeroes(BlockDriverState *bs,
      cmd.cdw12 = cpu_to_le32(cdw12);
      trace_nvme_write_zeroes(s, offset, bytes, flags);
 -    assert(s->nr_queues > 1);
 +    assert(s->queue_count > 1);
      req = nvme_get_free_req(ioq);
      assert(req);
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn nvme_co_pdiscard(BlockDriverState *bs,
          return -ENOTSUP;
      }
 -    assert(s->nr_queues > 1);
 +    assert(s->queue_count > 1);
      buf = qemu_try_memalign(s->page_size, s->page_size);
      if (!buf) {
@@ -XXX,XX +XXX,XX @@ static void nvme_detach_aio_context(BlockDriverState *bs)
  {
      BDRVNVMeState *s = bs->opaque;
 -    for (int i = 0; i < s->nr_queues; i++) {
 +    for (unsigned i = 0; i < s->queue_count; i++) {
          NVMeQueuePair *q = s->queues[i];
          qemu_bh_delete(q->completion_bh);
@@ -XXX,XX +XXX,XX @@ static void nvme_attach_aio_context(BlockDriverState *bs,
      aio_set_event_notifier(new_context, &s->irq_notifier[MSIX_SHARED_IRQ_IDX],
                             false, nvme_handle_event, nvme_poll_cb);
 -    for (int i = 0; i < s->nr_queues; i++) {
 +    for (unsigned i = 0; i < s->queue_count; i++) {
          NVMeQueuePair *q = s->queues[i];
          q->completion_bh =
@@ -XXX,XX +XXX,XX @@ static void nvme_aio_plug(BlockDriverState *bs)
  static void nvme_aio_unplug(BlockDriverState *bs)
  {
 -    int i;
      BDRVNVMeState *s = bs->opaque;
      assert(s->plugged);
      s->plugged = false;
 -    for (i = INDEX_IO(0); i < s->nr_queues; i++) {
 +    for (unsigned i = INDEX_IO(0); i < s->queue_count; i++) {
          NVMeQueuePair *q = s->queues[i];
          qemu_mutex_lock(&q->lock);
          nvme_kick(q);
 diff --git a/block/trace-events b/block/trace-events
 index XXXXXXX..XXXXXXX 100644
 --- a/block/trace-events
 +++ b/block/trace-events
@@ -XXX,XX +XXX,XX @@ qed_aio_write_main(void *s, void *acb, int ret, uint64_t offset, size_t len) "s
  # nvme.c
  nvme_controller_capability_raw(uint64_t value) "0x%08"PRIx64
  nvme_controller_capability(const char *desc, uint64_t value) "%s: %"PRIu64
 -nvme_kick(void *s, int queue) "s %p queue %d"
 +nvme_kick(void *s, unsigned q_index) "s %p q #%u"
  nvme_dma_flush_queue_wait(void *s) "s %p"
  nvme_error(int cmd_specific, int sq_head, int sqid, int cid, int status) "cmd_specific %d sq_head %d sqid %d cid %d status 0x%x"
 -nvme_process_completion(void *s, int index, int inflight) "s %p queue %d inflight %d"
 -nvme_process_completion_queue_plugged(void *s, int index) "s %p queue %d"
 -nvme_complete_command(void *s, int index, int cid) "s %p queue %d cid %d"
 -nvme_submit_command(void *s, int index, int cid) "s %p queue %d cid %d"
 +nvme_process_completion(void *s, unsigned q_index, int inflight) "s %p q #%u inflight %d"
 +nvme_process_completion_queue_plugged(void *s, unsigned q_index) "s %p q #%u"
 +nvme_complete_command(void *s, unsigned q_index, int cid) "s %p q #%u cid %d"
 +nvme_submit_command(void *s, unsigned q_index, int cid) "s %p q #%u cid %d"
  nvme_submit_command_raw(int c0, int c1, int c2, int c3, int c4, int c5, int c6, int c7) "%02x %02x %02x %02x %02x %02x %02x %02x"
  nvme_handle_event(void *s) "s %p"
  nvme_poll_queue(void *s, unsigned q_index) "s %p q #%u"
 --
-.28.0
+.31.1

-[PULL 32/33] util/vfio-helpers: Convert vfio_dump_mapping to trace events
+[PULL 7/8] coroutine-sleep: replace QemuCoSleepState pointer with struct in the API
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-The QEMU_VFIO_DEBUG definition is only modifiable at build-time.
+Right now, users of qemu_co_sleep_ns_wakeable are simply passing
-Trace events can be enabled at run-time. As we prefer the latter,
+a pointer to QemuCoSleepState by reference to the function.  But
-convert qemu_vfio_dump_mappings() to use trace events instead
+QemuCoSleepState really is just a Coroutine*; making the
-of fprintf().
+content of the struct public is just as efficient and lets us
+skip the user_state_pointer indirection.
-Reviewed-by: Fam Zheng <fam@euphon.net>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
+Since the usage is changed, take the occasion to rename the
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
+struct to QemuCoSleep.
-Message-id: 20201103020733.2303148-7-philmd@redhat.com
 Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
 Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
 Message-id: 20210517100548.28806-6-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
 ---
- util/vfio-helpers.c | 19 ++++---------------
+ include/qemu/coroutine.h    | 23 +++++++++++----------
- util/trace-events   |  1 +
+ block/block-copy.c          |  8 ++++----
-files changed, 5 insertions(+), 15 deletions(-)
+ block/nbd.c                 | 10 ++++-----
+ util/qemu-coroutine-sleep.c | 41 ++++++++++++++++---------------------
-diff --git a/util/vfio-helpers.c b/util/vfio-helpers.c
+files changed, 39 insertions(+), 43 deletions(-)
-index XXXXXXX..XXXXXXX 100644
---- a/util/vfio-helpers.c
+diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
-+++ b/util/vfio-helpers.c
+index XXXXXXX..XXXXXXX 100644
-@@ -XXX,XX +XXX,XX @@ QEMUVFIOState *qemu_vfio_open_pci(const char *device, Error **errp)
+--- a/include/qemu/coroutine.h
-     return s;
++++ b/include/qemu/coroutine.h
- }
+@@ -XXX,XX +XXX,XX @@ void qemu_co_rwlock_wrlock(CoRwlock *lock);
+  */
--static void qemu_vfio_dump_mapping(IOVAMapping *m)
+ void qemu_co_rwlock_unlock(CoRwlock *lock);
 -typedef struct QemuCoSleepState QemuCoSleepState;
 +typedef struct QemuCoSleep {
 +    Coroutine *to_wake;
 +} QemuCoSleep;
  /**
 - * Yield the coroutine for a given duration. During this yield, @sleep_state
 - * is set to an opaque pointer, which may be used for
 - * qemu_co_sleep_wake(). Be careful, the pointer is set back to zero when the
 - * timer fires. Don't save the obtained value to other variables and don't call
 - * qemu_co_sleep_wake from another aio context.
 + * Yield the coroutine for a given duration. Initializes @w so that,
 + * during this yield, it can be passed to qemu_co_sleep_wake() to
 + * terminate the sleep.
   */
 -void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
 -                                            QemuCoSleepState **sleep_state);
 +void coroutine_fn qemu_co_sleep_ns_wakeable(QemuCoSleep *w,
 +                                            QEMUClockType type, int64_t ns);
 +
  static inline void coroutine_fn qemu_co_sleep_ns(QEMUClockType type, int64_t ns)
  {
 -    QemuCoSleepState *unused = NULL;
 -    qemu_co_sleep_ns_wakeable(type, ns, &unused);
 +    QemuCoSleep w = { 0 };
 +    qemu_co_sleep_ns_wakeable(&w, type, ns);
  }
  /**
@@ -XXX,XX +XXX,XX @@ static inline void coroutine_fn qemu_co_sleep_ns(QEMUClockType type, int64_t ns)
   * qemu_co_sleep_ns() and should be checked to be non-NULL before calling
   * qemu_co_sleep_wake().
   */
 -void qemu_co_sleep_wake(QemuCoSleepState *sleep_state);
 +void qemu_co_sleep_wake(QemuCoSleep *w);
  /**
   * Yield until a file descriptor becomes readable
 diff --git a/block/block-copy.c b/block/block-copy.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/block-copy.c
 +++ b/block/block-copy.c
@@ -XXX,XX +XXX,XX @@ typedef struct BlockCopyCallState {
      /* State */
      int ret;
      bool finished;
 -    QemuCoSleepState *sleep_state;
 +    QemuCoSleep sleep;
      bool cancelled;
      /* OUT parameters */
@@ -XXX,XX +XXX,XX @@ block_copy_dirty_clusters(BlockCopyCallState *call_state)
                  if (ns > 0) {
                      block_copy_task_end(task, -EAGAIN);
                      g_free(task);
 -                    qemu_co_sleep_ns_wakeable(QEMU_CLOCK_REALTIME, ns,
 -                                              &call_state->sleep_state);
 +                    qemu_co_sleep_ns_wakeable(&call_state->sleep,
 +                                              QEMU_CLOCK_REALTIME, ns);
                      continue;
                  }
              }
@@ -XXX,XX +XXX,XX @@ out:
  void block_copy_kick(BlockCopyCallState *call_state)
  {
 -    qemu_co_sleep_wake(call_state->sleep_state);
 +    qemu_co_sleep_wake(&call_state->sleep);
  }
  /*
 diff --git a/block/nbd.c b/block/nbd.c
 index XXXXXXX..XXXXXXX 100644
 --- a/block/nbd.c
 +++ b/block/nbd.c
@@ -XXX,XX +XXX,XX @@ typedef struct BDRVNBDState {
      CoQueue free_sema;
      Coroutine *connection_co;
      Coroutine *teardown_co;
 -    QemuCoSleepState *connection_co_sleep_ns_state;
 +    QemuCoSleep reconnect_sleep;
      bool drained;
      bool wait_drained_end;
      int in_flight;
@@ -XXX,XX +XXX,XX @@ static void coroutine_fn nbd_client_co_drain_begin(BlockDriverState *bs)
      BDRVNBDState *s = (BDRVNBDState *)bs->opaque;
      s->drained = true;
 -    qemu_co_sleep_wake(s->connection_co_sleep_ns_state);
 +    qemu_co_sleep_wake(&s->reconnect_sleep);
      nbd_co_establish_connection_cancel(bs, false);
@@ -XXX,XX +XXX,XX @@ static void nbd_teardown_connection(BlockDriverState *bs)
      s->state = NBD_CLIENT_QUIT;
      if (s->connection_co) {
 -        qemu_co_sleep_wake(s->connection_co_sleep_ns_state);
 +        qemu_co_sleep_wake(&s->reconnect_sleep);
          nbd_co_establish_connection_cancel(bs, true);
      }
      if (qemu_in_coroutine()) {
@@ -XXX,XX +XXX,XX @@ static coroutine_fn void nbd_co_reconnect_loop(BDRVNBDState *s)
              }
              bdrv_inc_in_flight(s->bs);
          } else {
 -            qemu_co_sleep_ns_wakeable(QEMU_CLOCK_REALTIME, timeout,
 -                                      &s->connection_co_sleep_ns_state);
 +            qemu_co_sleep_ns_wakeable(&s->reconnect_sleep,
 +                                      QEMU_CLOCK_REALTIME, timeout);
              if (s->drained) {
                  continue;
              }
 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 const char *qemu_co_sleep_ns__scheduled = "qemu_co_sleep_ns";
 -struct QemuCoSleepState {
 +void qemu_co_sleep_wake(QemuCoSleep *w)
 +{
      Coroutine *co;
 -    QemuCoSleepState **user_state_pointer;
 -};
 -void qemu_co_sleep_wake(QemuCoSleepState *sleep_state)
 -{
--    if (QEMU_VFIO_DEBUG) {
+-    if (sleep_state) {
--        printf("  vfio mapping %p %" PRIx64 " to %" PRIx64 "\n", m->host,
++    co = w->to_wake;
--               (uint64_t)m->size, (uint64_t)m->iova);
++    w->to_wake = NULL;
--    }
++    if (co) {
--}
+         /* Write of schedule protected by barrier write in aio_co_schedule */
--
+-        const char *scheduled = qatomic_cmpxchg(&sleep_state->co->scheduled,
- static void qemu_vfio_dump_mappings(QEMUVFIOState *s)
++        const char *scheduled = qatomic_cmpxchg(&co->scheduled,
- {
+                                                 qemu_co_sleep_ns__scheduled, NULL);
--    int i;
--
+         assert(scheduled == qemu_co_sleep_ns__scheduled);
--    if (QEMU_VFIO_DEBUG) {
+-        *sleep_state->user_state_pointer = NULL;
--        printf("vfio mappings\n");
+-        aio_co_wake(sleep_state->co);
--        for (i = 0; i < s->nr_mappings; ++i) {
++        aio_co_wake(co);
 -            qemu_vfio_dump_mapping(&s->mappings[i]);
 -        }
 +    for (int i = 0; i < s->nr_mappings; ++i) {
 +        trace_qemu_vfio_dump_mapping(s->mappings[i].host,
 +                                     s->mappings[i].iova,
 +                                     s->mappings[i].size);
      }
  }
-diff --git a/util/trace-events b/util/trace-events
+ static void co_sleep_cb(void *opaque)
-index XXXXXXX..XXXXXXX 100644
+ {
---- a/util/trace-events
+-    QemuCoSleepState **sleep_state = opaque;
-+++ b/util/trace-events
+-    qemu_co_sleep_wake(*sleep_state);
-@@ -XXX,XX +XXX,XX @@ qemu_mutex_unlock(void *mutex, const char *file, const int line) "released mutex
++    QemuCoSleep *w = opaque;
- qemu_vfio_dma_reset_temporary(void *s) "s %p"
++    qemu_co_sleep_wake(w);
- qemu_vfio_ram_block_added(void *s, void *p, size_t size) "s %p host %p size 0x%zx"
+ }
- qemu_vfio_ram_block_removed(void *s, void *p, size_t size) "s %p host %p size 0x%zx"
-+qemu_vfio_dump_mapping(void *host, uint64_t iova, size_t size) "vfio mapping %p to iova 0x%08" PRIx64 " size 0x%zx"
+-void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
- qemu_vfio_find_mapping(void *s, void *p) "s %p host %p"
+-                                            QemuCoSleepState **sleep_state)
- qemu_vfio_new_mapping(void *s, void *host, size_t size, int index, uint64_t iova) "s %p host %p size 0x%zx index %d iova 0x%"PRIx64
++void coroutine_fn qemu_co_sleep_ns_wakeable(QemuCoSleep *w,
- qemu_vfio_do_mapping(void *s, void *host, uint64_t iova, size_t size) "s %p host %p <-> iova 0x%"PRIx64 " size 0x%zx"
++                                            QEMUClockType type, int64_t ns)
  {
 +    Coroutine *co = qemu_coroutine_self();
      AioContext *ctx = qemu_get_current_aio_context();
      QEMUTimer ts;
 -    QemuCoSleepState state = {
 -        .co = qemu_coroutine_self(),
 -        .user_state_pointer = sleep_state,
 -    };
 -    const char *scheduled = qatomic_cmpxchg(&state.co->scheduled, NULL,
 -                                           qemu_co_sleep_ns__scheduled);
 +    const char *scheduled = qatomic_cmpxchg(&co->scheduled, NULL,
 +                                            qemu_co_sleep_ns__scheduled);
      if (scheduled) {
          fprintf(stderr,
                  "%s: Co-routine was already scheduled in '%s'\n",
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
          abort();
      }
 -    aio_timer_init(ctx, &ts, type, SCALE_NS, co_sleep_cb, sleep_state);
 -    *sleep_state = &state;
 +    w->to_wake = co;
 +    aio_timer_init(ctx, &ts, type, SCALE_NS, co_sleep_cb, w),
      timer_mod(&ts, qemu_clock_get_ns(type) + ns);
      qemu_coroutine_yield();
      timer_del(&ts);
 -    /* qemu_co_sleep_wake clears *sleep_state before resuming this coroutine.  */
 -    assert(*sleep_state == NULL);
 +    /* w->to_wake is cleared before resuming this coroutine.  */
 +    assert(w->to_wake == NULL);
  }
 --
-.28.0
+.31.1

-[PULL 12/33] block/nvme: Make nvme_identify() return boolean indicating error
+[PULL 8/8] coroutine-sleep: introduce qemu_co_sleep
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
+From: Paolo Bonzini <pbonzini@redhat.com>
-Just for consistency, following the example documented since
+Allow using QemuCoSleep to sleep forever until woken by qemu_co_sleep_wake.
-commit e3fe3988d7 ("error: Document Error API usage rules"),
+This makes the logic of qemu_co_sleep_ns_wakeable easy to understand.
 return a boolean value indicating an error is set or not.
 Directly pass errp as the local_err is not requested in our
 case.
-Tested-by: Eric Auger <eric.auger@redhat.com>
+In the future we will introduce an API that can work even if the
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
+sleep and wake happen from different threads.  For now, initializing
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
+w->to_wake after timer_mod is fine because the timer can only fire in
-Message-id: 20201029093306.1063879-11-philmd@redhat.com
+the same AioContext.
 Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
 Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
 Message-id: 20210517100548.28806-7-pbonzini@redhat.com
 Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
 ---
- block/nvme.c | 12 +++++++-----
+ include/qemu/coroutine.h    |  5 +++++
-file changed, 7 insertions(+), 5 deletions(-)
+ util/qemu-coroutine-sleep.c | 26 +++++++++++++++++++-------
 files changed, 24 insertions(+), 7 deletions(-)
-diff --git a/block/nvme.c b/block/nvme.c
+diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
 index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
+--- a/include/qemu/coroutine.h
-+++ b/block/nvme.c
++++ b/include/qemu/coroutine.h
-@@ -XXX,XX +XXX,XX @@ static int nvme_cmd_sync(BlockDriverState *bs, NVMeQueuePair *q,
+@@ -XXX,XX +XXX,XX @@ typedef struct QemuCoSleep {
-     return ret;
+ void coroutine_fn qemu_co_sleep_ns_wakeable(QemuCoSleep *w,
                                              QEMUClockType type, int64_t ns);
 +/**
 + * Yield the coroutine until the next call to qemu_co_sleep_wake.
 + */
 +void coroutine_fn qemu_co_sleep(QemuCoSleep *w);
 +
  static inline void coroutine_fn qemu_co_sleep_ns(QEMUClockType type, int64_t ns)
  {
      QemuCoSleep w = { 0 };
 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)
      qemu_co_sleep_wake(w);
  }
--static void nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
+-void coroutine_fn qemu_co_sleep_ns_wakeable(QemuCoSleep *w,
-+/* Returns true on success, false on failure. */
+-                                            QEMUClockType type, int64_t ns)
-+static bool nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
++void coroutine_fn qemu_co_sleep(QemuCoSleep *w)
  {
-     BDRVNVMeState *s = bs->opaque;
+     Coroutine *co = qemu_coroutine_self();
-+    bool ret = false;
+-    AioContext *ctx = qemu_get_current_aio_context();
-     union {
+-    QEMUTimer ts;
-         NvmeIdCtrl ctrl;
-         NvmeIdNs ns;
+     const char *scheduled = qatomic_cmpxchg(&co->scheduled, NULL,
-@@ -XXX,XX +XXX,XX @@ static void nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
+                                             qemu_co_sleep_ns__scheduled);
-         goto out;
+@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QemuCoSleep *w,
      }
-+    ret = true;
+     w->to_wake = co;
-     s->blkshift = lbaf->ds;
+-    aio_timer_init(ctx, &ts, type, SCALE_NS, co_sleep_cb, w),
- out:
+-    timer_mod(&ts, qemu_clock_get_ns(type) + ns);
-     qemu_vfio_dma_unmap(s->vfio, id);
+     qemu_coroutine_yield();
-     qemu_vfree(id);
+-    timer_del(&ts);
      /* w->to_wake is cleared before resuming this coroutine.  */
      assert(w->to_wake == NULL);
  }
 +
-+    return ret;
++void coroutine_fn qemu_co_sleep_ns_wakeable(QemuCoSleep *w,
- }
++                                            QEMUClockType type, int64_t ns)
++{
- static bool nvme_poll_queue(NVMeQueuePair *q)
++    AioContext *ctx = qemu_get_current_aio_context();
-@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
++    QEMUTimer ts;
-     uint64_t cap;
++
-     uint64_t timeout_ms;
++    aio_timer_init(ctx, &ts, type, SCALE_NS, co_sleep_cb, w);
-     uint64_t deadline, now;
++    timer_mod(&ts, qemu_clock_get_ns(type) + ns);
--    Error *local_err = NULL;
++
-     volatile NvmeBar *regs = NULL;
++    /*
++     * The timer will fire in the current AiOContext, so the callback
-     qemu_co_mutex_init(&s->dma_map_lock);
++     * must happen after qemu_co_sleep yields and there is no race
-@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
++     * between timer_mod and qemu_co_sleep.
-                            &s->irq_notifier[MSIX_SHARED_IRQ_IDX],
++     */
-                            false, nvme_handle_event, nvme_poll_cb);
++    qemu_co_sleep(w);
++    timer_del(&ts);
--    nvme_identify(bs, namespace, &local_err);
++}
 -    if (local_err) {
 -        error_propagate(errp, local_err);
 +    if (!nvme_identify(bs, namespace, errp)) {
          ret = -EIO;
          goto out;
      }
 --
-.28.0
+.31.1

-[PULL 13/33] block/nvme: Make nvme_init_queue() return boolean indicating error
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-Just for consistency, following the example documented since
-commit e3fe3988d7 ("error: Document Error API usage rules"),
-return a boolean value indicating an error is set or not.
-Directly pass errp as the local_err is not requested in our
-case. This simplifies a bit nvme_create_queue_pair().
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201029093306.1063879-12-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- block/nvme.c | 16 +++++++---------
-file changed, 7 insertions(+), 9 deletions(-)
-diff --git a/block/nvme.c b/block/nvme.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
-+++ b/block/nvme.c
-@@ -XXX,XX +XXX,XX @@ static QemuOptsList runtime_opts = {
-     },
- };
--static void nvme_init_queue(BDRVNVMeState *s, NVMeQueue *q,
-+/* Returns true on success, false on failure. */
-+static bool nvme_init_queue(BDRVNVMeState *s, NVMeQueue *q,
-                             unsigned nentries, size_t entry_bytes, Error **errp)
- {
-     size_t bytes;
-@@ -XXX,XX +XXX,XX @@ static void nvme_init_queue(BDRVNVMeState *s, NVMeQueue *q,
-     q->queue = qemu_try_memalign(s->page_size, bytes);
-     if (!q->queue) {
-         error_setg(errp, "Cannot allocate queue");
--        return;
-+        return false;
-     }
-     memset(q->queue, 0, bytes);
-     r = qemu_vfio_dma_map(s->vfio, q->queue, bytes, false, &q->iova);
-     if (r) {
-         error_setg(errp, "Cannot map queue");
-+        return false;
-     }
-+    return true;
- }
- static void nvme_free_queue_pair(NVMeQueuePair *q)
-@@ -XXX,XX +XXX,XX @@ static NVMeQueuePair *nvme_create_queue_pair(BDRVNVMeState *s,
-                                              Error **errp)
- {
-     int i, r;
--    Error *local_err = NULL;
-     NVMeQueuePair *q;
-     uint64_t prp_list_iova;
-@@ -XXX,XX +XXX,XX @@ static NVMeQueuePair *nvme_create_queue_pair(BDRVNVMeState *s,
-         req->prp_list_iova = prp_list_iova + i * s->page_size;
-     }
--    nvme_init_queue(s, &q->sq, size, NVME_SQ_ENTRY_BYTES, &local_err);
--    if (local_err) {
--        error_propagate(errp, local_err);
-+    if (!nvme_init_queue(s, &q->sq, size, NVME_SQ_ENTRY_BYTES, errp)) {
-         goto fail;
-     }
-     q->sq.doorbell = &s->doorbells[idx * s->doorbell_scale].sq_tail;
--    nvme_init_queue(s, &q->cq, size, NVME_CQ_ENTRY_BYTES, &local_err);
--    if (local_err) {
--        error_propagate(errp, local_err);
-+    if (!nvme_init_queue(s, &q->cq, size, NVME_CQ_ENTRY_BYTES, errp)) {
-         goto fail;
-     }
-     q->cq.doorbell = &s->doorbells[idx * s->doorbell_scale].cq_head;
---
-.28.0

-[PULL 14/33] block/nvme: Introduce Completion Queue definitions
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-Rename Submission Queue flags with 'Sq' to differentiate
-submission queue flags from command queue flags, and introduce
-Completion Queue flag definitions.
-Reviewed-by: Eric Auger <eric.auger@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Message-id: 20201029093306.1063879-13-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- include/block/nvme.h | 18 ++++++++++++------
-file changed, 12 insertions(+), 6 deletions(-)
-diff --git a/include/block/nvme.h b/include/block/nvme.h
-index XXXXXXX..XXXXXXX 100644
---- a/include/block/nvme.h
-+++ b/include/block/nvme.h
-@@ -XXX,XX +XXX,XX @@ typedef struct QEMU_PACKED NvmeCreateCq {
- #define NVME_CQ_FLAGS_PC(cq_flags)  (cq_flags & 0x1)
- #define NVME_CQ_FLAGS_IEN(cq_flags) ((cq_flags >> 1) & 0x1)
-+enum NvmeFlagsCq {
-+    NVME_CQ_PC          = 1,
-+    NVME_CQ_IEN         = 2,
-+};
-+
- typedef struct QEMU_PACKED NvmeCreateSq {
-     uint8_t     opcode;
-     uint8_t     flags;
-@@ -XXX,XX +XXX,XX @@ typedef struct QEMU_PACKED NvmeCreateSq {
- #define NVME_SQ_FLAGS_PC(sq_flags)      (sq_flags & 0x1)
- #define NVME_SQ_FLAGS_QPRIO(sq_flags)   ((sq_flags >> 1) & 0x3)
--enum NvmeQueueFlags {
--    NVME_Q_PC           = 1,
--    NVME_Q_PRIO_URGENT  = 0,
--    NVME_Q_PRIO_HIGH    = 1,
--    NVME_Q_PRIO_NORMAL  = 2,
--    NVME_Q_PRIO_LOW     = 3,
-+enum NvmeFlagsSq {
-+    NVME_SQ_PC          = 1,
-+
-+    NVME_SQ_PRIO_URGENT = 0,
-+    NVME_SQ_PRIO_HIGH   = 1,
-+    NVME_SQ_PRIO_NORMAL = 2,
-+    NVME_SQ_PRIO_LOW    = 3,
- };
- typedef struct QEMU_PACKED NvmeIdentify {
---
-.28.0

-[PULL 15/33] block/nvme: Use definitions instead of magic values in add_io_queue()
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-Replace magic values by definitions, and simplifiy since the
-number of queues will never reach 64K.
-Reviewed-by: Eric Auger <eric.auger@redhat.com>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201029093306.1063879-14-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- block/nvme.c | 9 +++++----
-file changed, 5 insertions(+), 4 deletions(-)
-diff --git a/block/nvme.c b/block/nvme.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
-+++ b/block/nvme.c
-@@ -XXX,XX +XXX,XX @@ static bool nvme_add_io_queue(BlockDriverState *bs, Error **errp)
-     NvmeCmd cmd;
-     unsigned queue_size = NVME_QUEUE_SIZE;
-+    assert(n <= UINT16_MAX);
-     q = nvme_create_queue_pair(s, bdrv_get_aio_context(bs),
-                                n, queue_size, errp);
-     if (!q) {
-@@ -XXX,XX +XXX,XX @@ static bool nvme_add_io_queue(BlockDriverState *bs, Error **errp)
-     cmd = (NvmeCmd) {
-         .opcode = NVME_ADM_CMD_CREATE_CQ,
-         .dptr.prp1 = cpu_to_le64(q->cq.iova),
--        .cdw10 = cpu_to_le32(((queue_size - 1) << 16) | (n & 0xFFFF)),
--        .cdw11 = cpu_to_le32(0x3),
-+        .cdw10 = cpu_to_le32(((queue_size - 1) << 16) | n),
-+        .cdw11 = cpu_to_le32(NVME_CQ_IEN | NVME_CQ_PC),
-     };
-     if (nvme_cmd_sync(bs, s->queues[INDEX_ADMIN], &cmd)) {
-         error_setg(errp, "Failed to create CQ io queue [%u]", n);
-@@ -XXX,XX +XXX,XX @@ static bool nvme_add_io_queue(BlockDriverState *bs, Error **errp)
-     cmd = (NvmeCmd) {
-         .opcode = NVME_ADM_CMD_CREATE_SQ,
-         .dptr.prp1 = cpu_to_le64(q->sq.iova),
--        .cdw10 = cpu_to_le32(((queue_size - 1) << 16) | (n & 0xFFFF)),
--        .cdw11 = cpu_to_le32(0x1 | (n << 16)),
-+        .cdw10 = cpu_to_le32(((queue_size - 1) << 16) | n),
-+        .cdw11 = cpu_to_le32(NVME_SQ_PC | (n << 16)),
-     };
-     if (nvme_cmd_sync(bs, s->queues[INDEX_ADMIN], &cmd)) {
-         error_setg(errp, "Failed to create SQ io queue [%u]", n);
---
-.28.0

-[PULL 16/33] block/nvme: Correctly initialize Admin Queue Attributes
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-From the specification chapter 3.1.8 "AQA - Admin Queue Attributes"
-the Admin Submission Queue Size field is a 0’s based value:
-  Admin Submission Queue Size (ASQS):
-    Defines the size of the Admin Submission Queue in entries.
-    Enabling a controller while this field is cleared to 00h
-    produces undefined results. The minimum size of the Admin
-    Submission Queue is two entries. The maximum size of the
-    Admin Submission Queue is 4096 entries.
-    This is a 0’s based value.
-This bug has never been hit because the device initialization
-uses a single command synchronously :)
-Reviewed-by: Eric Auger <eric.auger@redhat.com>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201029093306.1063879-15-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- block/nvme.c | 6 +++---
-file changed, 3 insertions(+), 3 deletions(-)
-diff --git a/block/nvme.c b/block/nvme.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
-+++ b/block/nvme.c
-@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
-         goto out;
-     }
-     s->queue_count = 1;
--    QEMU_BUILD_BUG_ON(NVME_QUEUE_SIZE & 0xF000);
--    regs->aqa = cpu_to_le32((NVME_QUEUE_SIZE << AQA_ACQS_SHIFT) |
--                            (NVME_QUEUE_SIZE << AQA_ASQS_SHIFT));
-+    QEMU_BUILD_BUG_ON((NVME_QUEUE_SIZE - 1) & 0xF000);
-+    regs->aqa = cpu_to_le32(((NVME_QUEUE_SIZE - 1) << AQA_ACQS_SHIFT) |
-+                            ((NVME_QUEUE_SIZE - 1) << AQA_ASQS_SHIFT));
-     regs->asq = cpu_to_le64(s->queues[INDEX_ADMIN]->sq.iova);
-     regs->acq = cpu_to_le64(s->queues[INDEX_ADMIN]->cq.iova);
---
-.28.0

-[PULL 17/33] block/nvme: Simplify ADMIN queue access
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-We don't need to dereference from BDRVNVMeState each time.
-Use a NVMeQueuePair pointer on the admin queue.
-The nvme_init() becomes easier to review, matching the style
-of nvme_add_io_queue().
-Reviewed-by: Eric Auger <eric.auger@redhat.com>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201029093306.1063879-16-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- block/nvme.c | 12 ++++++------
-file changed, 6 insertions(+), 6 deletions(-)
-diff --git a/block/nvme.c b/block/nvme.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
-+++ b/block/nvme.c
-@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
-                      Error **errp)
- {
-     BDRVNVMeState *s = bs->opaque;
-+    NVMeQueuePair *q;
-     AioContext *aio_context = bdrv_get_aio_context(bs);
-     int ret;
-     uint64_t cap;
-@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
-     /* Set up admin queue. */
-     s->queues = g_new(NVMeQueuePair *, 1);
--    s->queues[INDEX_ADMIN] = nvme_create_queue_pair(s, aio_context, 0,
--                                                          NVME_QUEUE_SIZE,
--                                                          errp);
--    if (!s->queues[INDEX_ADMIN]) {
-+    q = nvme_create_queue_pair(s, aio_context, 0, NVME_QUEUE_SIZE, errp);
-+    if (!q) {
-         ret = -EINVAL;
-         goto out;
-     }
-+    s->queues[INDEX_ADMIN] = q;
-     s->queue_count = 1;
-     QEMU_BUILD_BUG_ON((NVME_QUEUE_SIZE - 1) & 0xF000);
-     regs->aqa = cpu_to_le32(((NVME_QUEUE_SIZE - 1) << AQA_ACQS_SHIFT) |
-                             ((NVME_QUEUE_SIZE - 1) << AQA_ASQS_SHIFT));
--    regs->asq = cpu_to_le64(s->queues[INDEX_ADMIN]->sq.iova);
--    regs->acq = cpu_to_le64(s->queues[INDEX_ADMIN]->cq.iova);
-+    regs->asq = cpu_to_le64(q->sq.iova);
-+    regs->acq = cpu_to_le64(q->cq.iova);
-     /* After setting up all control registers we can enable device now. */
-     regs->cc = cpu_to_le32((ctz32(NVME_CQ_ENTRY_BYTES) << CC_IOCQES_SHIFT) |
---
-.28.0

-[PULL 19/33] block/nvme: Set request_alignment at initialization
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-Commit bdd6a90a9e5 ("block: Add VFIO based NVMe driver")
-sets the request_alignment in nvme_refresh_limits().
-For consistency, also set it during initialization.
-Reported-by: Stefan Hajnoczi <stefanha@redhat.com>
-Reviewed-by: Eric Auger <eric.auger@redhat.com>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201029093306.1063879-18-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- block/nvme.c | 1 +
-file changed, 1 insertion(+)
-diff --git a/block/nvme.c b/block/nvme.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
-+++ b/block/nvme.c
-@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
-     s->page_size = MAX(4096, 1 << NVME_CAP_MPSMIN(cap));
-     s->doorbell_scale = (4 << NVME_CAP_DSTRD(cap)) / sizeof(uint32_t);
-     bs->bl.opt_mem_alignment = s->page_size;
-+    bs->bl.request_alignment = s->page_size;
-     timeout_ms = MIN(500 * NVME_CAP_TO(cap), 30000);
-     /* Reset device to get a clean state. */
---
-.28.0

-[PULL 20/33] block/nvme: Correct minimum device page size
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-While trying to simplify the code using a macro, we forgot
-the 12-bit shift... Correct that.
-Fixes: fad1eb68862 ("block/nvme: Use register definitions from 'block/nvme.h'")
-Reported-by: Eric Auger <eric.auger@redhat.com>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Reviewed-by: Eric Auger <eric.auger@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201029093306.1063879-19-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- block/nvme.c | 2 +-
-file changed, 1 insertion(+), 1 deletion(-)
-diff --git a/block/nvme.c b/block/nvme.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
-+++ b/block/nvme.c
-@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
-         goto out;
-     }
--    s->page_size = MAX(4096, 1 << NVME_CAP_MPSMIN(cap));
-+    s->page_size = 1u << (12 + NVME_CAP_MPSMIN(cap));
-     s->doorbell_scale = (4 << NVME_CAP_DSTRD(cap)) / sizeof(uint32_t);
-     bs->bl.opt_mem_alignment = s->page_size;
-     bs->bl.request_alignment = s->page_size;
---
-.28.0

-[PULL 21/33] block/nvme: Change size and alignment of IDENTIFY response buffer
+Deleted patch
-From: Eric Auger <eric.auger@redhat.com>
-In preparation of 64kB host page support, let's change the size
-and alignment of the IDENTIFY command response buffer so that
-the VFIO DMA MAP succeeds. We align on the host page size.
-Signed-off-by: Eric Auger <eric.auger@redhat.com>
-Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201029093306.1063879-20-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- block/nvme.c | 9 +++++----
-file changed, 5 insertions(+), 4 deletions(-)
-diff --git a/block/nvme.c b/block/nvme.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
-+++ b/block/nvme.c
-@@ -XXX,XX +XXX,XX @@ static bool nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
-         .opcode = NVME_ADM_CMD_IDENTIFY,
-         .cdw10 = cpu_to_le32(0x1),
-     };
-+    size_t id_size = QEMU_ALIGN_UP(sizeof(*id), qemu_real_host_page_size);
--    id = qemu_try_memalign(s->page_size, sizeof(*id));
-+    id = qemu_try_memalign(qemu_real_host_page_size, id_size);
-     if (!id) {
-         error_setg(errp, "Cannot allocate buffer for identify response");
-         goto out;
-     }
--    r = qemu_vfio_dma_map(s->vfio, id, sizeof(*id), true, &iova);
-+    r = qemu_vfio_dma_map(s->vfio, id, id_size, true, &iova);
-     if (r) {
-         error_setg(errp, "Cannot map buffer for DMA");
-         goto out;
-     }
--    memset(id, 0, sizeof(*id));
-+    memset(id, 0, id_size);
-     cmd.dptr.prp1 = cpu_to_le64(iova);
-     if (nvme_admin_cmd_sync(bs, &cmd)) {
-         error_setg(errp, "Failed to identify controller");
-@@ -XXX,XX +XXX,XX @@ static bool nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
-     s->supports_write_zeroes = !!(oncs & NVME_ONCS_WRITE_ZEROES);
-     s->supports_discard = !!(oncs & NVME_ONCS_DSM);
--    memset(id, 0, sizeof(*id));
-+    memset(id, 0, id_size);
-     cmd.cdw10 = 0;
-     cmd.nsid = cpu_to_le32(namespace);
-     if (nvme_admin_cmd_sync(bs, &cmd)) {
---
-.28.0

-[PULL 22/33] block/nvme: Change size and alignment of queue
+Deleted patch
-From: Eric Auger <eric.auger@redhat.com>
-In preparation of 64kB host page support, let's change the size
-and alignment of the queue so that the VFIO DMA MAP succeeds.
-We align on the host page size.
-Signed-off-by: Eric Auger <eric.auger@redhat.com>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201029093306.1063879-21-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- block/nvme.c | 4 ++--
-file changed, 2 insertions(+), 2 deletions(-)
-diff --git a/block/nvme.c b/block/nvme.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
-+++ b/block/nvme.c
-@@ -XXX,XX +XXX,XX @@ static bool nvme_init_queue(BDRVNVMeState *s, NVMeQueue *q,
-     size_t bytes;
-     int r;
--    bytes = ROUND_UP(nentries * entry_bytes, s->page_size);
-+    bytes = ROUND_UP(nentries * entry_bytes, qemu_real_host_page_size);
-     q->head = q->tail = 0;
--    q->queue = qemu_try_memalign(s->page_size, bytes);
-+    q->queue = qemu_try_memalign(qemu_real_host_page_size, bytes);
-     if (!q->queue) {
-         error_setg(errp, "Cannot allocate queue");
-         return false;
---
-.28.0

-[PULL 23/33] block/nvme: Change size and alignment of prp_list_pages
+Deleted patch
-From: Eric Auger <eric.auger@redhat.com>
-In preparation of 64kB host page support, let's change the size
-and alignment of the prp_list_pages so that the VFIO DMA MAP succeeds
-with 64kB host page size. We align on the host page size.
-Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Signed-off-by: Eric Auger <eric.auger@redhat.com>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201029093306.1063879-22-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- block/nvme.c | 11 ++++++-----
-file changed, 6 insertions(+), 5 deletions(-)
-diff --git a/block/nvme.c b/block/nvme.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
-+++ b/block/nvme.c
-@@ -XXX,XX +XXX,XX @@ static NVMeQueuePair *nvme_create_queue_pair(BDRVNVMeState *s,
-     int i, r;
-     NVMeQueuePair *q;
-     uint64_t prp_list_iova;
-+    size_t bytes;
-     q = g_try_new0(NVMeQueuePair, 1);
-     if (!q) {
-@@ -XXX,XX +XXX,XX @@ static NVMeQueuePair *nvme_create_queue_pair(BDRVNVMeState *s,
-     }
-     trace_nvme_create_queue_pair(idx, q, size, aio_context,
-                                  event_notifier_get_fd(s->irq_notifier));
--    q->prp_list_pages = qemu_try_memalign(s->page_size,
--                                          s->page_size * NVME_NUM_REQS);
-+    bytes = QEMU_ALIGN_UP(s->page_size * NVME_NUM_REQS,
-+                          qemu_real_host_page_size);
-+    q->prp_list_pages = qemu_try_memalign(qemu_real_host_page_size, bytes);
-     if (!q->prp_list_pages) {
-         goto fail;
-     }
--    memset(q->prp_list_pages, 0, s->page_size * NVME_NUM_REQS);
-+    memset(q->prp_list_pages, 0, bytes);
-     qemu_mutex_init(&q->lock);
-     q->s = s;
-     q->index = idx;
-     qemu_co_queue_init(&q->free_req_queue);
-     q->completion_bh = aio_bh_new(aio_context, nvme_process_completion_bh, q);
--    r = qemu_vfio_dma_map(s->vfio, q->prp_list_pages,
--                          s->page_size * NVME_NUM_REQS,
-+    r = qemu_vfio_dma_map(s->vfio, q->prp_list_pages, bytes,
-                           false, &prp_list_iova);
-     if (r) {
-         goto fail;
---
-.28.0

-[PULL 24/33] block/nvme: Align iov's va and size on host page size
+Deleted patch
-From: Eric Auger <eric.auger@redhat.com>
-Make sure iov's va and size are properly aligned on the
-host page size.
-Signed-off-by: Eric Auger <eric.auger@redhat.com>
-Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201029093306.1063879-23-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- block/nvme.c | 14 ++++++++------
-file changed, 8 insertions(+), 6 deletions(-)
-diff --git a/block/nvme.c b/block/nvme.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
-+++ b/block/nvme.c
-@@ -XXX,XX +XXX,XX @@ static coroutine_fn int nvme_cmd_map_qiov(BlockDriverState *bs, NvmeCmd *cmd,
-     for (i = 0; i < qiov->niov; ++i) {
-         bool retry = true;
-         uint64_t iova;
-+        size_t len = QEMU_ALIGN_UP(qiov->iov[i].iov_len,
-+                                   qemu_real_host_page_size);
- try_map:
-         r = qemu_vfio_dma_map(s->vfio,
-                               qiov->iov[i].iov_base,
--                              qiov->iov[i].iov_len,
--                              true, &iova);
-+                              len, true, &iova);
-         if (r == -ENOMEM && retry) {
-             retry = false;
-             trace_nvme_dma_flush_queue_wait(s);
-@@ -XXX,XX +XXX,XX @@ static inline bool nvme_qiov_aligned(BlockDriverState *bs,
-     BDRVNVMeState *s = bs->opaque;
-     for (i = 0; i < qiov->niov; ++i) {
--        if (!QEMU_PTR_IS_ALIGNED(qiov->iov[i].iov_base, s->page_size) ||
--            !QEMU_IS_ALIGNED(qiov->iov[i].iov_len, s->page_size)) {
-+        if (!QEMU_PTR_IS_ALIGNED(qiov->iov[i].iov_base,
-+                                 qemu_real_host_page_size) ||
-+            !QEMU_IS_ALIGNED(qiov->iov[i].iov_len, qemu_real_host_page_size)) {
-             trace_nvme_qiov_unaligned(qiov, i, qiov->iov[i].iov_base,
-                                       qiov->iov[i].iov_len, s->page_size);
-             return false;
-@@ -XXX,XX +XXX,XX @@ static int nvme_co_prw(BlockDriverState *bs, uint64_t offset, uint64_t bytes,
-     int r;
-     uint8_t *buf = NULL;
-     QEMUIOVector local_qiov;
--
-+    size_t len = QEMU_ALIGN_UP(bytes, qemu_real_host_page_size);
-     assert(QEMU_IS_ALIGNED(offset, s->page_size));
-     assert(QEMU_IS_ALIGNED(bytes, s->page_size));
-     assert(bytes <= s->max_transfer);
-@@ -XXX,XX +XXX,XX @@ static int nvme_co_prw(BlockDriverState *bs, uint64_t offset, uint64_t bytes,
-     }
-     s->stats.unaligned_accesses++;
-     trace_nvme_prw_buffered(s, offset, bytes, qiov->niov, is_write);
--    buf = qemu_try_memalign(s->page_size, bytes);
-+    buf = qemu_try_memalign(qemu_real_host_page_size, len);
-     if (!buf) {
-         return -ENOMEM;
---
-.28.0

-[PULL 25/33] block/nvme: Fix use of write-only doorbells page on Aarch64 arch
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-qemu_vfio_pci_map_bar() calls mmap(), and mmap(2) states:
-  'offset' must be a multiple of the page size as returned
-   by sysconf(_SC_PAGE_SIZE).
-In commit f68453237b9 we started to use an offset of 4K which
-broke this contract on Aarch64 arch.
-Fix by mapping at offset 0, and and accessing doorbells at offset=4K.
-Fixes: f68453237b9 ("block/nvme: Map doorbells pages write-only")
-Reported-by: Eric Auger <eric.auger@redhat.com>
-Reviewed-by: Eric Auger <eric.auger@redhat.com>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201029093306.1063879-24-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- block/nvme.c | 11 +++++++----
-file changed, 7 insertions(+), 4 deletions(-)
-diff --git a/block/nvme.c b/block/nvme.c
-index XXXXXXX..XXXXXXX 100644
---- a/block/nvme.c
-+++ b/block/nvme.c
-@@ -XXX,XX +XXX,XX @@ typedef struct {
- struct BDRVNVMeState {
-     AioContext *aio_context;
-     QEMUVFIOState *vfio;
-+    void *bar0_wo_map;
-     /* Memory mapped registers */
-     volatile struct {
-         uint32_t sq_tail;
-@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
-         }
-     }
--    s->doorbells = qemu_vfio_pci_map_bar(s->vfio, 0, sizeof(NvmeBar),
--                                         NVME_DOORBELL_SIZE, PROT_WRITE, errp);
-+    s->bar0_wo_map = qemu_vfio_pci_map_bar(s->vfio, 0, 0,
-+                                           sizeof(NvmeBar) + NVME_DOORBELL_SIZE,
-+                                           PROT_WRITE, errp);
-+    s->doorbells = (void *)((uintptr_t)s->bar0_wo_map + sizeof(NvmeBar));
-     if (!s->doorbells) {
-         ret = -EINVAL;
-         goto out;
-@@ -XXX,XX +XXX,XX @@ static void nvme_close(BlockDriverState *bs)
-                            &s->irq_notifier[MSIX_SHARED_IRQ_IDX],
-                            false, NULL, NULL);
-     event_notifier_cleanup(&s->irq_notifier[MSIX_SHARED_IRQ_IDX]);
--    qemu_vfio_pci_unmap_bar(s->vfio, 0, (void *)s->doorbells,
--                            sizeof(NvmeBar), NVME_DOORBELL_SIZE);
-+    qemu_vfio_pci_unmap_bar(s->vfio, 0, s->bar0_wo_map,
-+                            0, sizeof(NvmeBar) + NVME_DOORBELL_SIZE);
-     qemu_vfio_close(s->vfio);
-     g_free(s->device);
---
-.28.0

-[PULL 27/33] util/vfio-helpers: Improve reporting unsupported IOMMU type
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-Change the confuse "VFIO IOMMU check failed" error message by
-the explicit "VFIO IOMMU Type1 is not supported" once.
-Example on POWER:
- $ qemu-system-ppc64 -drive if=none,id=nvme0,file=nvme://0001:01:00.0/1,format=raw
- qemu-system-ppc64: -drive if=none,id=nvme0,file=nvme://0001:01:00.0/1,format=raw: VFIO IOMMU Type1 is not supported
-Suggested-by: Alex Williamson <alex.williamson@redhat.com>
-Reviewed-by: Fam Zheng <fam@euphon.net>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201103020733.2303148-2-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- util/vfio-helpers.c | 2 +-
-file changed, 1 insertion(+), 1 deletion(-)
-diff --git a/util/vfio-helpers.c b/util/vfio-helpers.c
-index XXXXXXX..XXXXXXX 100644
---- a/util/vfio-helpers.c
-+++ b/util/vfio-helpers.c
-@@ -XXX,XX +XXX,XX @@ static int qemu_vfio_init_pci(QEMUVFIOState *s, const char *device,
-     }
-     if (!ioctl(s->container, VFIO_CHECK_EXTENSION, VFIO_TYPE1_IOMMU)) {
--        error_setg_errno(errp, errno, "VFIO IOMMU check failed");
-+        error_setg_errno(errp, errno, "VFIO IOMMU Type1 is not supported");
-         ret = -EINVAL;
-         goto fail_container;
-     }
---
-.28.0

-[PULL 28/33] util/vfio-helpers: Trace PCI I/O config accesses
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-We sometime get kernel panic with some devices on Aarch64
-hosts. Alex Williamson suggests it might be broken PCIe
-root complex. Add trace event to record the latest I/O
-access before crashing. In case, assert our accesses are
-aligned.
-Reviewed-by: Fam Zheng <fam@euphon.net>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201103020733.2303148-3-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- util/vfio-helpers.c | 8 ++++++++
- util/trace-events   | 2 ++
-files changed, 10 insertions(+)
-diff --git a/util/vfio-helpers.c b/util/vfio-helpers.c
-index XXXXXXX..XXXXXXX 100644
---- a/util/vfio-helpers.c
-+++ b/util/vfio-helpers.c
-@@ -XXX,XX +XXX,XX @@ static int qemu_vfio_pci_read_config(QEMUVFIOState *s, void *buf,
- {
-     int ret;
-+    trace_qemu_vfio_pci_read_config(buf, ofs, size,
-+                                    s->config_region_info.offset,
-+                                    s->config_region_info.size);
-+    assert(QEMU_IS_ALIGNED(s->config_region_info.offset + ofs, size));
-     do {
-         ret = pread(s->device, buf, size, s->config_region_info.offset + ofs);
-     } while (ret == -1 && errno == EINTR);
-@@ -XXX,XX +XXX,XX @@ static int qemu_vfio_pci_write_config(QEMUVFIOState *s, void *buf, int size, int
- {
-     int ret;
-+    trace_qemu_vfio_pci_write_config(buf, ofs, size,
-+                                     s->config_region_info.offset,
-+                                     s->config_region_info.size);
-+    assert(QEMU_IS_ALIGNED(s->config_region_info.offset + ofs, size));
-     do {
-         ret = pwrite(s->device, buf, size, s->config_region_info.offset + ofs);
-     } while (ret == -1 && errno == EINTR);
-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_vfio_new_mapping(void *s, void *host, size_t size, int index, uint64_t iova
- qemu_vfio_do_mapping(void *s, void *host, size_t size, uint64_t iova) "s %p host %p size 0x%zx iova 0x%"PRIx64
- qemu_vfio_dma_map(void *s, void *host, size_t size, bool temporary, uint64_t *iova) "s %p host %p size 0x%zx temporary %d iova %p"
- qemu_vfio_dma_unmap(void *s, void *host) "s %p host %p"
-+qemu_vfio_pci_read_config(void *buf, int ofs, int size, uint64_t region_ofs, uint64_t region_size) "read cfg ptr %p ofs 0x%x size 0x%x (region addr 0x%"PRIx64" size 0x%"PRIx64")"
-+qemu_vfio_pci_write_config(void *buf, int ofs, int size, uint64_t region_ofs, uint64_t region_size) "write cfg ptr %p ofs 0x%x size 0x%x (region addr 0x%"PRIx64" size 0x%"PRIx64")"
---
-.28.0

-[PULL 29/33] util/vfio-helpers: Trace PCI BAR region info
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-For debug purpose, trace BAR regions info.
-Reviewed-by: Fam Zheng <fam@euphon.net>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201103020733.2303148-4-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- util/vfio-helpers.c | 8 ++++++++
- util/trace-events   | 1 +
-files changed, 9 insertions(+)
-diff --git a/util/vfio-helpers.c b/util/vfio-helpers.c
-index XXXXXXX..XXXXXXX 100644
---- a/util/vfio-helpers.c
-+++ b/util/vfio-helpers.c
-@@ -XXX,XX +XXX,XX @@ static inline void assert_bar_index_valid(QEMUVFIOState *s, int index)
- static int qemu_vfio_pci_init_bar(QEMUVFIOState *s, int index, Error **errp)
- {
-+    g_autofree char *barname = NULL;
-     assert_bar_index_valid(s, index);
-     s->bar_region_info[index] = (struct vfio_region_info) {
-         .index = VFIO_PCI_BAR0_REGION_INDEX + index,
-@@ -XXX,XX +XXX,XX @@ static int qemu_vfio_pci_init_bar(QEMUVFIOState *s, int index, Error **errp)
-         error_setg_errno(errp, errno, "Failed to get BAR region info");
-         return -errno;
-     }
-+    barname = g_strdup_printf("bar[%d]", index);
-+    trace_qemu_vfio_region_info(barname, s->bar_region_info[index].offset,
-+                                s->bar_region_info[index].size,
-+                                s->bar_region_info[index].cap_offset);
-     return 0;
- }
-@@ -XXX,XX +XXX,XX @@ static int qemu_vfio_init_pci(QEMUVFIOState *s, const char *device,
-         ret = -errno;
-         goto fail;
-     }
-+    trace_qemu_vfio_region_info("config", s->config_region_info.offset,
-+                                s->config_region_info.size,
-+                                s->config_region_info.cap_offset);
-     for (i = 0; i < ARRAY_SIZE(s->bar_region_info); i++) {
-         ret = qemu_vfio_pci_init_bar(s, i, errp);
-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_vfio_dma_map(void *s, void *host, size_t size, bool temporary, uint64_t *io
- qemu_vfio_dma_unmap(void *s, void *host) "s %p host %p"
- qemu_vfio_pci_read_config(void *buf, int ofs, int size, uint64_t region_ofs, uint64_t region_size) "read cfg ptr %p ofs 0x%x size 0x%x (region addr 0x%"PRIx64" size 0x%"PRIx64")"
- qemu_vfio_pci_write_config(void *buf, int ofs, int size, uint64_t region_ofs, uint64_t region_size) "write cfg ptr %p ofs 0x%x size 0x%x (region addr 0x%"PRIx64" size 0x%"PRIx64")"
-+qemu_vfio_region_info(const char *desc, uint64_t region_ofs, uint64_t region_size, uint32_t cap_offset) "region '%s' addr 0x%"PRIx64" size 0x%"PRIx64" cap_ofs 0x%"PRIx32
---
-.28.0

-[PULL 30/33] util/vfio-helpers: Trace where BARs are mapped
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-For debugging purpose, trace where a BAR is mapped.
-Reviewed-by: Fam Zheng <fam@euphon.net>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201103020733.2303148-5-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- util/vfio-helpers.c | 2 ++
- util/trace-events   | 1 +
-files changed, 3 insertions(+)
-diff --git a/util/vfio-helpers.c b/util/vfio-helpers.c
-index XXXXXXX..XXXXXXX 100644
---- a/util/vfio-helpers.c
-+++ b/util/vfio-helpers.c
-@@ -XXX,XX +XXX,XX @@ void *qemu_vfio_pci_map_bar(QEMUVFIOState *s, int index,
-     p = mmap(NULL, MIN(size, s->bar_region_info[index].size - offset),
-              prot, MAP_SHARED,
-              s->device, s->bar_region_info[index].offset + offset);
-+    trace_qemu_vfio_pci_map_bar(index, s->bar_region_info[index].offset ,
-+                                size, offset, p);
-     if (p == MAP_FAILED) {
-         error_setg_errno(errp, errno, "Failed to map BAR region");
-         p = NULL;
-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_vfio_dma_unmap(void *s, void *host) "s %p host %p"
- qemu_vfio_pci_read_config(void *buf, int ofs, int size, uint64_t region_ofs, uint64_t region_size) "read cfg ptr %p ofs 0x%x size 0x%x (region addr 0x%"PRIx64" size 0x%"PRIx64")"
- qemu_vfio_pci_write_config(void *buf, int ofs, int size, uint64_t region_ofs, uint64_t region_size) "write cfg ptr %p ofs 0x%x size 0x%x (region addr 0x%"PRIx64" size 0x%"PRIx64")"
- qemu_vfio_region_info(const char *desc, uint64_t region_ofs, uint64_t region_size, uint32_t cap_offset) "region '%s' addr 0x%"PRIx64" size 0x%"PRIx64" cap_ofs 0x%"PRIx32
-+qemu_vfio_pci_map_bar(int index, uint64_t region_ofs, uint64_t region_size, int ofs, void *host) "map region bar#%d addr 0x%"PRIx64" size 0x%"PRIx64" ofs 0x%x host %p"
---
-.28.0

-[PULL 31/33] util/vfio-helpers: Improve DMA trace events
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-For debugging purpose, trace where DMA regions are mapped.
-Reviewed-by: Fam Zheng <fam@euphon.net>
-Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201103020733.2303148-6-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- util/vfio-helpers.c | 3 ++-
- util/trace-events   | 5 +++--
-files changed, 5 insertions(+), 3 deletions(-)
-diff --git a/util/vfio-helpers.c b/util/vfio-helpers.c
-index XXXXXXX..XXXXXXX 100644
---- a/util/vfio-helpers.c
-+++ b/util/vfio-helpers.c
-@@ -XXX,XX +XXX,XX @@ static int qemu_vfio_do_mapping(QEMUVFIOState *s, void *host, size_t size,
-         .vaddr = (uintptr_t)host,
-         .size = size,
-     };
--    trace_qemu_vfio_do_mapping(s, host, size, iova);
-+    trace_qemu_vfio_do_mapping(s, host, iova, size);
-     if (ioctl(s->container, VFIO_IOMMU_MAP_DMA, &dma_map)) {
-         error_report("VFIO_MAP_DMA failed: %s", strerror(errno));
-@@ -XXX,XX +XXX,XX @@ int qemu_vfio_dma_map(QEMUVFIOState *s, void *host, size_t size,
-             }
-         }
-     }
-+    trace_qemu_vfio_dma_mapped(s, host, iova0, size);
-     if (iova) {
-         *iova = iova0;
-     }
-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_vfio_ram_block_added(void *s, void *p, size_t size) "s %p host %p size 0x%z
- qemu_vfio_ram_block_removed(void *s, void *p, size_t size) "s %p host %p size 0x%zx"
- qemu_vfio_find_mapping(void *s, void *p) "s %p host %p"
- qemu_vfio_new_mapping(void *s, void *host, size_t size, int index, uint64_t iova) "s %p host %p size 0x%zx index %d iova 0x%"PRIx64
--qemu_vfio_do_mapping(void *s, void *host, size_t size, uint64_t iova) "s %p host %p size 0x%zx iova 0x%"PRIx64
--qemu_vfio_dma_map(void *s, void *host, size_t size, bool temporary, uint64_t *iova) "s %p host %p size 0x%zx temporary %d iova %p"
-+qemu_vfio_do_mapping(void *s, void *host, uint64_t iova, size_t size) "s %p host %p <-> iova 0x%"PRIx64 " size 0x%zx"
-+qemu_vfio_dma_map(void *s, void *host, size_t size, bool temporary, uint64_t *iova) "s %p host %p size 0x%zx temporary %d &iova %p"
-+qemu_vfio_dma_mapped(void *s, void *host, uint64_t iova, size_t size) "s %p host %p <-> iova 0x%"PRIx64" size 0x%zx"
- qemu_vfio_dma_unmap(void *s, void *host) "s %p host %p"
- qemu_vfio_pci_read_config(void *buf, int ofs, int size, uint64_t region_ofs, uint64_t region_size) "read cfg ptr %p ofs 0x%x size 0x%x (region addr 0x%"PRIx64" size 0x%"PRIx64")"
- qemu_vfio_pci_write_config(void *buf, int ofs, int size, uint64_t region_ofs, uint64_t region_size) "write cfg ptr %p ofs 0x%x size 0x%x (region addr 0x%"PRIx64" size 0x%"PRIx64")"
---
-.28.0

-[PULL 33/33] util/vfio-helpers: Assert offset is aligned to page size
+Deleted patch
-From: Philippe Mathieu-Daudé <philmd@redhat.com>
-mmap(2) states:
-  'offset' must be a multiple of the page size as returned
-   by sysconf(_SC_PAGE_SIZE).
-Add an assertion to be sure we don't break this contract.
-Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
-Message-id: 20201103020733.2303148-8-philmd@redhat.com
-Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
-Tested-by: Eric Auger <eric.auger@redhat.com>
----
- util/vfio-helpers.c | 1 +
-file changed, 1 insertion(+)
-diff --git a/util/vfio-helpers.c b/util/vfio-helpers.c
-index XXXXXXX..XXXXXXX 100644
---- a/util/vfio-helpers.c
-+++ b/util/vfio-helpers.c
-@@ -XXX,XX +XXX,XX @@ void *qemu_vfio_pci_map_bar(QEMUVFIOState *s, int index,
-                             Error **errp)
- {
-     void *p;
-+    assert(QEMU_IS_ALIGNED(offset, qemu_real_host_page_size));
-     assert_bar_index_valid(s, index);
-     p = mmap(NULL, MIN(size, s->bar_region_info[index].size - offset),
-              prot, MAP_SHARED,
---
-.28.0

The following changes since commit 8507c9d5c9a62de2a0e281b640f995e26eac46af:

Merge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging (2020-11-03 15:59:44 +0000)

are available in the Git repository at:

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

for you to fetch changes up to fc107d86840b3364e922c26cf7631b7fd38ce523:

util/vfio-helpers: Assert offset is aligned to page size (2020-11-03 19:06:23 +0000)

----------------------------------------------------------------
Pull request for 5.2

NVMe fixes to solve IOMMU issues on non-x86 and error message/tracing
improvements. Elena Afanasova's ioeventfd fixes are also included.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

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

Elena Afanasova (2):
  accel/kvm: add PIO ioeventfds only in case kvm_eventfds_allowed is
    true
  softmmu/memory: fix memory_region_ioeventfd_equal()

Eric Auger (4):
  block/nvme: Change size and alignment of IDENTIFY response buffer
  block/nvme: Change size and alignment of queue
  block/nvme: Change size and alignment of prp_list_pages
  block/nvme: Align iov's va and size on host page size

Philippe Mathieu-Daudé (27):
  MAINTAINERS: Cover "block/nvme.h" file
  block/nvme: Use hex format to display offset in trace events
  block/nvme: Report warning with warn_report()
  block/nvme: Trace controller capabilities
  block/nvme: Trace nvme_poll_queue() per queue
  block/nvme: Improve nvme_free_req_queue_wait() trace information
  block/nvme: Trace queue pair creation/deletion
  block/nvme: Move definitions before structure declarations
  block/nvme: Use unsigned integer for queue counter/size
  block/nvme: Make nvme_identify() return boolean indicating error
  block/nvme: Make nvme_init_queue() return boolean indicating error
  block/nvme: Introduce Completion Queue definitions
  block/nvme: Use definitions instead of magic values in add_io_queue()
  block/nvme: Correctly initialize Admin Queue Attributes
  block/nvme: Simplify ADMIN queue access
  block/nvme: Simplify nvme_cmd_sync()
  block/nvme: Set request_alignment at initialization
  block/nvme: Correct minimum device page size
  block/nvme: Fix use of write-only doorbells page on Aarch64 arch
  block/nvme: Fix nvme_submit_command() on big-endian host
  util/vfio-helpers: Improve reporting unsupported IOMMU type
  util/vfio-helpers: Trace PCI I/O config accesses
  util/vfio-helpers: Trace PCI BAR region info
  util/vfio-helpers: Trace where BARs are mapped
  util/vfio-helpers: Improve DMA trace events
  util/vfio-helpers: Convert vfio_dump_mapping to trace events
  util/vfio-helpers: Assert offset is aligned to page size

-- 
2.28.0

From: Elena Afanasova <eafanasova@gmail.com>

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Elena Afanasova <eafanasova@gmail.com>
Message-Id: <20201017210102.26036-1-eafanasova@gmail.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 accel/kvm/kvm-all.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/accel/kvm/kvm-all.c b/accel/kvm/kvm-all.c
index XXXXXXX..XXXXXXX 100644
--- a/accel/kvm/kvm-all.c
+++ b/accel/kvm/kvm-all.c
@@ -XXX,XX +XXX,XX @@ static int kvm_init(MachineState *ms)
 
     kvm_memory_listener_register(s, &s->memory_listener,
                                  &address_space_memory, 0);
-    memory_listener_register(&kvm_io_listener,
-                             &address_space_io);
+    if (kvm_eventfds_allowed) {
+        memory_listener_register(&kvm_io_listener,
+                                 &address_space_io);
+    }
     memory_listener_register(&kvm_coalesced_pio_listener,
                              &address_space_io);
 
-- 
2.28.0

From: Elena Afanasova <eafanasova@gmail.com>

Eventfd can be registered with a zero length when fast_mmio is true.
Handle this case properly when dispatching through QEMU.

Signed-off-by: Elena Afanasova <eafanasova@gmail.com>
Message-id: cf71a62eb04e61932ff8ffdd02e0b2aab4f495a0.camel@gmail.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 softmmu/memory.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/softmmu/memory.c b/softmmu/memory.c
index XXXXXXX..XXXXXXX 100644
--- a/softmmu/memory.c
+++ b/softmmu/memory.c
@@ -XXX,XX +XXX,XX @@ static bool memory_region_ioeventfd_before(MemoryRegionIoeventfd *a,
 static bool memory_region_ioeventfd_equal(MemoryRegionIoeventfd *a,
                                           MemoryRegionIoeventfd *b)
 {
-    return !memory_region_ioeventfd_before(a, b)
-        && !memory_region_ioeventfd_before(b, a);
+    if (int128_eq(a->addr.start, b->addr.start) &&
+        (!int128_nz(a->addr.size) || !int128_nz(b->addr.size) ||
+         (int128_eq(a->addr.size, b->addr.size) &&
+          (a->match_data == b->match_data) &&
+          ((a->match_data && (a->data == b->data)) || !a->match_data) &&
+          (a->e == b->e))))
+        return true;
+
+    return false;
 }
 
 /* Range of memory in the global map.  Addresses are absolute. */
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

The "block/nvme.h" header is shared by both the NVMe block
driver and the NVMe emulated device. Add the 'F:' entry on
both sections, so all maintainers/reviewers are notified
when it is changed.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Message-Id: <20200701140634.25994-1-philmd@redhat.com>
---
 MAINTAINERS | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/MAINTAINERS b/MAINTAINERS
index XXXXXXX..XXXXXXX 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -XXX,XX +XXX,XX @@ M: Klaus Jensen <its@irrelevant.dk>
 L: qemu-block@nongnu.org
 S: Supported
 F: hw/block/nvme*
+F: include/block/nvme.h
 F: tests/qtest/nvme-test.c
 F: docs/specs/nvme.txt
 T: git git://git.infradead.org/qemu-nvme.git nvme-next
@@ -XXX,XX +XXX,XX @@ R: Fam Zheng <fam@euphon.net>
 L: qemu-block@nongnu.org
 S: Supported
 F: block/nvme*
+F: include/block/nvme.h
 T: git https://github.com/stefanha/qemu.git block
 
 Bootdevice
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

Use the same format used for the hw/vfio/ trace events.

Suggested-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-3-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/trace-events | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/block/trace-events b/block/trace-events
index XXXXXXX..XXXXXXX 100644
--- a/block/trace-events
+++ b/block/trace-events
@@ -XXX,XX +XXX,XX @@ nvme_submit_command(void *s, int index, int cid) "s %p queue %d cid %d"
 nvme_submit_command_raw(int c0, int c1, int c2, int c3, int c4, int c5, int c6, int c7) "%02x %02x %02x %02x %02x %02x %02x %02x"
 nvme_handle_event(void *s) "s %p"
 nvme_poll_cb(void *s) "s %p"
-nvme_prw_aligned(void *s, int is_write, uint64_t offset, uint64_t bytes, int flags, int niov) "s %p is_write %d offset %"PRId64" bytes %"PRId64" flags %d niov %d"
-nvme_write_zeroes(void *s, uint64_t offset, uint64_t bytes, int flags) "s %p offset %"PRId64" bytes %"PRId64" flags %d"
+nvme_prw_aligned(void *s, int is_write, uint64_t offset, uint64_t bytes, int flags, int niov) "s %p is_write %d offset 0x%"PRIx64" bytes %"PRId64" flags %d niov %d"
+nvme_write_zeroes(void *s, uint64_t offset, uint64_t bytes, int flags) "s %p offset 0x%"PRIx64" bytes %"PRId64" flags %d"
 nvme_qiov_unaligned(const void *qiov, int n, void *base, size_t size, int align) "qiov %p n %d base %p size 0x%zx align 0x%x"
-nvme_prw_buffered(void *s, uint64_t offset, uint64_t bytes, int niov, int is_write) "s %p offset %"PRId64" bytes %"PRId64" niov %d is_write %d"
-nvme_rw_done(void *s, int is_write, uint64_t offset, uint64_t bytes, int ret) "s %p is_write %d offset %"PRId64" bytes %"PRId64" ret %d"
-nvme_dsm(void *s, uint64_t offset, uint64_t bytes) "s %p offset %"PRId64" bytes %"PRId64""
-nvme_dsm_done(void *s, uint64_t offset, uint64_t bytes, int ret) "s %p offset %"PRId64" bytes %"PRId64" ret %d"
+nvme_prw_buffered(void *s, uint64_t offset, uint64_t bytes, int niov, int is_write) "s %p offset 0x%"PRIx64" bytes %"PRId64" niov %d is_write %d"
+nvme_rw_done(void *s, int is_write, uint64_t offset, uint64_t bytes, int ret) "s %p is_write %d offset 0x%"PRIx64" bytes %"PRId64" ret %d"
+nvme_dsm(void *s, uint64_t offset, uint64_t bytes) "s %p offset 0x%"PRIx64" bytes %"PRId64""
+nvme_dsm_done(void *s, uint64_t offset, uint64_t bytes, int ret) "s %p offset 0x%"PRIx64" bytes %"PRId64" ret %d"
 nvme_dma_map_flush(void *s) "s %p"
 nvme_free_req_queue_wait(void *q) "q %p"
 nvme_cmd_map_qiov(void *s, void *cmd, void *req, void *qiov, int entries) "s %p cmd %p req %p qiov %p entries %d"
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

Instead of displaying warning on stderr, use warn_report()
which also displays it on the monitor.

Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-4-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ static bool nvme_process_completion(NVMeQueuePair *q)
         }
         cid = le16_to_cpu(c->cid);
         if (cid == 0 || cid > NVME_QUEUE_SIZE) {
-            fprintf(stderr, "Unexpected CID in completion queue: %" PRIu32 "\n",
-                    cid);
+            warn_report("NVMe: Unexpected CID in completion queue: %"PRIu32", "
+                        "queue size: %u", cid, NVME_QUEUE_SIZE);
             continue;
         }
         trace_nvme_complete_command(s, q->index, cid);
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

Controllers have different capabilities and report them in the
CAP register. We are particularly interested by the page size
limits.

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-5-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c       | 13 +++++++++++++
 block/trace-events |  2 ++
 2 files changed, 15 insertions(+)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
      * Initialization". */
 
     cap = le64_to_cpu(regs->cap);
+    trace_nvme_controller_capability_raw(cap);
+    trace_nvme_controller_capability("Maximum Queue Entries Supported",
+                                     1 + NVME_CAP_MQES(cap));
+    trace_nvme_controller_capability("Contiguous Queues Required",
+                                     NVME_CAP_CQR(cap));
+    trace_nvme_controller_capability("Doorbell Stride",
+                                     2 << (2 + NVME_CAP_DSTRD(cap)));
+    trace_nvme_controller_capability("Subsystem Reset Supported",
+                                     NVME_CAP_NSSRS(cap));
+    trace_nvme_controller_capability("Memory Page Size Minimum",
+                                     1 << (12 + NVME_CAP_MPSMIN(cap)));
+    trace_nvme_controller_capability("Memory Page Size Maximum",
+                                     1 << (12 + NVME_CAP_MPSMAX(cap)));
     if (!NVME_CAP_CSS(cap)) {
         error_setg(errp, "Device doesn't support NVMe command set");
         ret = -EINVAL;
diff --git a/block/trace-events b/block/trace-events
index XXXXXXX..XXXXXXX 100644
--- a/block/trace-events
+++ b/block/trace-events
@@ -XXX,XX +XXX,XX @@ qed_aio_write_postfill(void *s, void *acb, uint64_t start, size_t len, uint64_t
 qed_aio_write_main(void *s, void *acb, int ret, uint64_t offset, size_t len) "s %p acb %p ret %d offset %"PRIu64" len %zu"
 
 # nvme.c
+nvme_controller_capability_raw(uint64_t value) "0x%08"PRIx64
+nvme_controller_capability(const char *desc, uint64_t value) "%s: %"PRIu64
 nvme_kick(void *s, int queue) "s %p queue %d"
 nvme_dma_flush_queue_wait(void *s) "s %p"
 nvme_error(int cmd_specific, int sq_head, int sqid, int cid, int status) "cmd_specific %d sq_head %d sqid %d cid %d status 0x%x"
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

As we want to enable multiple queues, report the event
in each nvme_poll_queue() call, rather than once in
the callback calling nvme_poll_queues().

Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-6-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c       | 2 +-
 block/trace-events | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ static bool nvme_poll_queue(NVMeQueuePair *q)
     const size_t cqe_offset = q->cq.head * NVME_CQ_ENTRY_BYTES;
     NvmeCqe *cqe = (NvmeCqe *)&q->cq.queue[cqe_offset];
 
+    trace_nvme_poll_queue(q->s, q->index);
     /*
      * Do an early check for completions. q->lock isn't needed because
      * nvme_process_completion() only runs in the event loop thread and
@@ -XXX,XX +XXX,XX @@ static bool nvme_poll_cb(void *opaque)
     BDRVNVMeState *s = container_of(e, BDRVNVMeState,
                                     irq_notifier[MSIX_SHARED_IRQ_IDX]);
 
-    trace_nvme_poll_cb(s);
     return nvme_poll_queues(s);
 }
 
diff --git a/block/trace-events b/block/trace-events
index XXXXXXX..XXXXXXX 100644
--- a/block/trace-events
+++ b/block/trace-events
@@ -XXX,XX +XXX,XX @@ nvme_complete_command(void *s, int index, int cid) "s %p queue %d cid %d"
 nvme_submit_command(void *s, int index, int cid) "s %p queue %d cid %d"
 nvme_submit_command_raw(int c0, int c1, int c2, int c3, int c4, int c5, int c6, int c7) "%02x %02x %02x %02x %02x %02x %02x %02x"
 nvme_handle_event(void *s) "s %p"
-nvme_poll_cb(void *s) "s %p"
+nvme_poll_queue(void *s, unsigned q_index) "s %p q #%u"
 nvme_prw_aligned(void *s, int is_write, uint64_t offset, uint64_t bytes, int flags, int niov) "s %p is_write %d offset 0x%"PRIx64" bytes %"PRId64" flags %d niov %d"
 nvme_write_zeroes(void *s, uint64_t offset, uint64_t bytes, int flags) "s %p offset 0x%"PRIx64" bytes %"PRId64" flags %d"
 nvme_qiov_unaligned(const void *qiov, int n, void *base, size_t size, int align) "qiov %p n %d base %p size 0x%zx align 0x%x"
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

What we want to trace is the block driver state and the queue index.

Suggested-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-7-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c       | 2 +-
 block/trace-events | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ static NVMeRequest *nvme_get_free_req(NVMeQueuePair *q)
 
     while (q->free_req_head == -1) {
         if (qemu_in_coroutine()) {
-            trace_nvme_free_req_queue_wait(q);
+            trace_nvme_free_req_queue_wait(q->s, q->index);
             qemu_co_queue_wait(&q->free_req_queue, &q->lock);
         } else {
             qemu_mutex_unlock(&q->lock);
diff --git a/block/trace-events b/block/trace-events
index XXXXXXX..XXXXXXX 100644
--- a/block/trace-events
+++ b/block/trace-events
@@ -XXX,XX +XXX,XX @@ nvme_rw_done(void *s, int is_write, uint64_t offset, uint64_t bytes, int ret) "s
 nvme_dsm(void *s, uint64_t offset, uint64_t bytes) "s %p offset 0x%"PRIx64" bytes %"PRId64""
 nvme_dsm_done(void *s, uint64_t offset, uint64_t bytes, int ret) "s %p offset 0x%"PRIx64" bytes %"PRId64" ret %d"
 nvme_dma_map_flush(void *s) "s %p"
-nvme_free_req_queue_wait(void *q) "q %p"
+nvme_free_req_queue_wait(void *s, unsigned q_index) "s %p q #%u"
 nvme_cmd_map_qiov(void *s, void *cmd, void *req, void *qiov, int entries) "s %p cmd %p req %p qiov %p entries %d"
 nvme_cmd_map_qiov_pages(void *s, int i, uint64_t page) "s %p page[%d] 0x%"PRIx64
 nvme_cmd_map_qiov_iov(void *s, int i, void *page, int pages) "s %p iov[%d] %p pages %d"
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-8-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c       | 3 +++
 block/trace-events | 2 ++
 2 files changed, 5 insertions(+)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ static void nvme_init_queue(BDRVNVMeState *s, NVMeQueue *q,
 
 static void nvme_free_queue_pair(NVMeQueuePair *q)
 {
+    trace_nvme_free_queue_pair(q->index, q);
     if (q->completion_bh) {
         qemu_bh_delete(q->completion_bh);
     }
@@ -XXX,XX +XXX,XX @@ static NVMeQueuePair *nvme_create_queue_pair(BDRVNVMeState *s,
     if (!q) {
         return NULL;
     }
+    trace_nvme_create_queue_pair(idx, q, size, aio_context,
+                                 event_notifier_get_fd(s->irq_notifier));
     q->prp_list_pages = qemu_try_memalign(s->page_size,
                                           s->page_size * NVME_NUM_REQS);
     if (!q->prp_list_pages) {
diff --git a/block/trace-events b/block/trace-events
index XXXXXXX..XXXXXXX 100644
--- a/block/trace-events
+++ b/block/trace-events
@@ -XXX,XX +XXX,XX @@ nvme_dsm(void *s, uint64_t offset, uint64_t bytes) "s %p offset 0x%"PRIx64" byte
 nvme_dsm_done(void *s, uint64_t offset, uint64_t bytes, int ret) "s %p offset 0x%"PRIx64" bytes %"PRId64" ret %d"
 nvme_dma_map_flush(void *s) "s %p"
 nvme_free_req_queue_wait(void *s, unsigned q_index) "s %p q #%u"
+nvme_create_queue_pair(unsigned q_index, void *q, unsigned size, void *aio_context, int fd) "index %u q %p size %u aioctx %p fd %d"
+nvme_free_queue_pair(unsigned q_index, void *q) "index %u q %p"
 nvme_cmd_map_qiov(void *s, void *cmd, void *req, void *qiov, int entries) "s %p cmd %p req %p qiov %p entries %d"
 nvme_cmd_map_qiov_pages(void *s, int i, uint64_t page) "s %p page[%d] 0x%"PRIx64
 nvme_cmd_map_qiov_iov(void *s, int i, void *page, int pages) "s %p iov[%d] %p pages %d"
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

To be able to use some definitions in structure declarations,
move them earlier. No logical change.

Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-9-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c | 19 ++++++++++---------
 1 file changed, 10 insertions(+), 9 deletions(-)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@
 
 typedef struct BDRVNVMeState BDRVNVMeState;
 
+/* Same index is used for queues and IRQs */
+#define INDEX_ADMIN     0
+#define INDEX_IO(n)     (1 + n)
+
+/* This driver shares a single MSIX IRQ for the admin and I/O queues */
+enum {
+    MSIX_SHARED_IRQ_IDX = 0,
+    MSIX_IRQ_COUNT = 1
+};
+
 typedef struct {
     int32_t  head, tail;
     uint8_t  *queue;
@@ -XXX,XX +XXX,XX @@ typedef struct {
     QEMUBH      *completion_bh;
 } NVMeQueuePair;
 
-#define INDEX_ADMIN     0
-#define INDEX_IO(n)     (1 + n)
-
-/* This driver shares a single MSIX IRQ for the admin and I/O queues */
-enum {
-    MSIX_SHARED_IRQ_IDX = 0,
-    MSIX_IRQ_COUNT = 1
-};
-
 struct BDRVNVMeState {
     AioContext *aio_context;
     QEMUVFIOState *vfio;
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

We can not have negative queue count/size/index, use unsigned type.
Rename 'nr_queues' as 'queue_count' to match the spec naming.

Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-10-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c       | 38 ++++++++++++++++++--------------------
 block/trace-events | 10 +++++-----
 2 files changed, 23 insertions(+), 25 deletions(-)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ struct BDRVNVMeState {
      * [1..]: io queues.
      */
     NVMeQueuePair **queues;
-    int nr_queues;
+    unsigned queue_count;
     size_t page_size;
     /* How many uint32_t elements does each doorbell entry take. */
     size_t doorbell_scale;
@@ -XXX,XX +XXX,XX @@ static QemuOptsList runtime_opts = {
 };
 
 static void nvme_init_queue(BDRVNVMeState *s, NVMeQueue *q,
-                            int nentries, int entry_bytes, Error **errp)
+                            unsigned nentries, size_t entry_bytes, Error **errp)
 {
     size_t bytes;
     int r;
@@ -XXX,XX +XXX,XX @@ static void nvme_free_req_queue_cb(void *opaque)
 
 static NVMeQueuePair *nvme_create_queue_pair(BDRVNVMeState *s,
                                              AioContext *aio_context,
-                                             int idx, int size,
+                                             unsigned idx, size_t size,
                                              Error **errp)
 {
     int i, r;
@@ -XXX,XX +XXX,XX @@ static bool nvme_poll_queues(BDRVNVMeState *s)
     bool progress = false;
     int i;
 
-    for (i = 0; i < s->nr_queues; i++) {
+    for (i = 0; i < s->queue_count; i++) {
         if (nvme_poll_queue(s->queues[i])) {
             progress = true;
         }
@@ -XXX,XX +XXX,XX @@ static void nvme_handle_event(EventNotifier *n)
 static bool nvme_add_io_queue(BlockDriverState *bs, Error **errp)
 {
     BDRVNVMeState *s = bs->opaque;
-    int n = s->nr_queues;
+    unsigned n = s->queue_count;
     NVMeQueuePair *q;
     NvmeCmd cmd;
-    int queue_size = NVME_QUEUE_SIZE;
+    unsigned queue_size = NVME_QUEUE_SIZE;
 
     q = nvme_create_queue_pair(s, bdrv_get_aio_context(bs),
                                n, queue_size, errp);
@@ -XXX,XX +XXX,XX @@ static bool nvme_add_io_queue(BlockDriverState *bs, Error **errp)
         .cdw11 = cpu_to_le32(0x3),
     };
     if (nvme_cmd_sync(bs, s->queues[INDEX_ADMIN], &cmd)) {
-        error_setg(errp, "Failed to create CQ io queue [%d]", n);
+        error_setg(errp, "Failed to create CQ io queue [%u]", n);
         goto out_error;
     }
     cmd = (NvmeCmd) {
@@ -XXX,XX +XXX,XX @@ static bool nvme_add_io_queue(BlockDriverState *bs, Error **errp)
         .cdw11 = cpu_to_le32(0x1 | (n << 16)),
     };
     if (nvme_cmd_sync(bs, s->queues[INDEX_ADMIN], &cmd)) {
-        error_setg(errp, "Failed to create SQ io queue [%d]", n);
+        error_setg(errp, "Failed to create SQ io queue [%u]", n);
         goto out_error;
     }
     s->queues = g_renew(NVMeQueuePair *, s->queues, n + 1);
     s->queues[n] = q;
-    s->nr_queues++;
+    s->queue_count++;
     return true;
 out_error:
     nvme_free_queue_pair(q);
@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
         ret = -EINVAL;
         goto out;
     }
-    s->nr_queues = 1;
+    s->queue_count = 1;
     QEMU_BUILD_BUG_ON(NVME_QUEUE_SIZE & 0xF000);
     regs->aqa = cpu_to_le32((NVME_QUEUE_SIZE << AQA_ACQS_SHIFT) |
                             (NVME_QUEUE_SIZE << AQA_ASQS_SHIFT));
@@ -XXX,XX +XXX,XX @@ static int nvme_enable_disable_write_cache(BlockDriverState *bs, bool enable,
 
 static void nvme_close(BlockDriverState *bs)
 {
-    int i;
     BDRVNVMeState *s = bs->opaque;
 
-    for (i = 0; i < s->nr_queues; ++i) {
+    for (unsigned i = 0; i < s->queue_count; ++i) {
         nvme_free_queue_pair(s->queues[i]);
     }
     g_free(s->queues);
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int nvme_co_prw_aligned(BlockDriverState *bs,
     };
 
     trace_nvme_prw_aligned(s, is_write, offset, bytes, flags, qiov->niov);
-    assert(s->nr_queues > 1);
+    assert(s->queue_count > 1);
     req = nvme_get_free_req(ioq);
     assert(req);
 
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int nvme_co_flush(BlockDriverState *bs)
         .ret = -EINPROGRESS,
     };
 
-    assert(s->nr_queues > 1);
+    assert(s->queue_count > 1);
     req = nvme_get_free_req(ioq);
     assert(req);
     nvme_submit_command(ioq, req, &cmd, nvme_rw_cb, &data);
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int nvme_co_pwrite_zeroes(BlockDriverState *bs,
     cmd.cdw12 = cpu_to_le32(cdw12);
 
     trace_nvme_write_zeroes(s, offset, bytes, flags);
-    assert(s->nr_queues > 1);
+    assert(s->queue_count > 1);
     req = nvme_get_free_req(ioq);
     assert(req);
 
@@ -XXX,XX +XXX,XX @@ static int coroutine_fn nvme_co_pdiscard(BlockDriverState *bs,
         return -ENOTSUP;
     }
 
-    assert(s->nr_queues > 1);
+    assert(s->queue_count > 1);
 
     buf = qemu_try_memalign(s->page_size, s->page_size);
     if (!buf) {
@@ -XXX,XX +XXX,XX @@ static void nvme_detach_aio_context(BlockDriverState *bs)
 {
     BDRVNVMeState *s = bs->opaque;
 
-    for (int i = 0; i < s->nr_queues; i++) {
+    for (unsigned i = 0; i < s->queue_count; i++) {
         NVMeQueuePair *q = s->queues[i];
 
         qemu_bh_delete(q->completion_bh);
@@ -XXX,XX +XXX,XX @@ static void nvme_attach_aio_context(BlockDriverState *bs,
     aio_set_event_notifier(new_context, &s->irq_notifier[MSIX_SHARED_IRQ_IDX],
                            false, nvme_handle_event, nvme_poll_cb);
 
-    for (int i = 0; i < s->nr_queues; i++) {
+    for (unsigned i = 0; i < s->queue_count; i++) {
         NVMeQueuePair *q = s->queues[i];
 
         q->completion_bh =
@@ -XXX,XX +XXX,XX @@ static void nvme_aio_plug(BlockDriverState *bs)
 
 static void nvme_aio_unplug(BlockDriverState *bs)
 {
-    int i;
     BDRVNVMeState *s = bs->opaque;
     assert(s->plugged);
     s->plugged = false;
-    for (i = INDEX_IO(0); i < s->nr_queues; i++) {
+    for (unsigned i = INDEX_IO(0); i < s->queue_count; i++) {
         NVMeQueuePair *q = s->queues[i];
         qemu_mutex_lock(&q->lock);
         nvme_kick(q);
diff --git a/block/trace-events b/block/trace-events
index XXXXXXX..XXXXXXX 100644
--- a/block/trace-events
+++ b/block/trace-events
@@ -XXX,XX +XXX,XX @@ qed_aio_write_main(void *s, void *acb, int ret, uint64_t offset, size_t len) "s
 # nvme.c
 nvme_controller_capability_raw(uint64_t value) "0x%08"PRIx64
 nvme_controller_capability(const char *desc, uint64_t value) "%s: %"PRIu64
-nvme_kick(void *s, int queue) "s %p queue %d"
+nvme_kick(void *s, unsigned q_index) "s %p q #%u"
 nvme_dma_flush_queue_wait(void *s) "s %p"
 nvme_error(int cmd_specific, int sq_head, int sqid, int cid, int status) "cmd_specific %d sq_head %d sqid %d cid %d status 0x%x"
-nvme_process_completion(void *s, int index, int inflight) "s %p queue %d inflight %d"
-nvme_process_completion_queue_plugged(void *s, int index) "s %p queue %d"
-nvme_complete_command(void *s, int index, int cid) "s %p queue %d cid %d"
-nvme_submit_command(void *s, int index, int cid) "s %p queue %d cid %d"
+nvme_process_completion(void *s, unsigned q_index, int inflight) "s %p q #%u inflight %d"
+nvme_process_completion_queue_plugged(void *s, unsigned q_index) "s %p q #%u"
+nvme_complete_command(void *s, unsigned q_index, int cid) "s %p q #%u cid %d"
+nvme_submit_command(void *s, unsigned q_index, int cid) "s %p q #%u cid %d"
 nvme_submit_command_raw(int c0, int c1, int c2, int c3, int c4, int c5, int c6, int c7) "%02x %02x %02x %02x %02x %02x %02x %02x"
 nvme_handle_event(void *s) "s %p"
 nvme_poll_queue(void *s, unsigned q_index) "s %p q #%u"
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

Just for consistency, following the example documented since
commit e3fe3988d7 ("error: Document Error API usage rules"),
return a boolean value indicating an error is set or not.
Directly pass errp as the local_err is not requested in our
case.

Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20201029093306.1063879-11-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ static int nvme_cmd_sync(BlockDriverState *bs, NVMeQueuePair *q,
     return ret;
 }
 
-static void nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
+/* Returns true on success, false on failure. */
+static bool nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
 {
     BDRVNVMeState *s = bs->opaque;
+    bool ret = false;
     union {
         NvmeIdCtrl ctrl;
         NvmeIdNs ns;
@@ -XXX,XX +XXX,XX @@ static void nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
         goto out;
     }
 
+    ret = true;
     s->blkshift = lbaf->ds;
 out:
     qemu_vfio_dma_unmap(s->vfio, id);
     qemu_vfree(id);
+
+    return ret;
 }
 
 static bool nvme_poll_queue(NVMeQueuePair *q)
@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
     uint64_t cap;
     uint64_t timeout_ms;
     uint64_t deadline, now;
-    Error *local_err = NULL;
     volatile NvmeBar *regs = NULL;
 
     qemu_co_mutex_init(&s->dma_map_lock);
@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
                            &s->irq_notifier[MSIX_SHARED_IRQ_IDX],
                            false, nvme_handle_event, nvme_poll_cb);
 
-    nvme_identify(bs, namespace, &local_err);
-    if (local_err) {
-        error_propagate(errp, local_err);
+    if (!nvme_identify(bs, namespace, errp)) {
         ret = -EIO;
         goto out;
     }
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-12-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c | 16 +++++++---------
 1 file changed, 7 insertions(+), 9 deletions(-)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ static QemuOptsList runtime_opts = {
     },
 };
 
-static void nvme_init_queue(BDRVNVMeState *s, NVMeQueue *q,
+/* Returns true on success, false on failure. */
+static bool nvme_init_queue(BDRVNVMeState *s, NVMeQueue *q,
                             unsigned nentries, size_t entry_bytes, Error **errp)
 {
     size_t bytes;
@@ -XXX,XX +XXX,XX @@ static void nvme_init_queue(BDRVNVMeState *s, NVMeQueue *q,
     q->queue = qemu_try_memalign(s->page_size, bytes);
     if (!q->queue) {
         error_setg(errp, "Cannot allocate queue");
-        return;
+        return false;
     }
     memset(q->queue, 0, bytes);
     r = qemu_vfio_dma_map(s->vfio, q->queue, bytes, false, &q->iova);
     if (r) {
         error_setg(errp, "Cannot map queue");
+        return false;
     }
+    return true;
 }
 
 static void nvme_free_queue_pair(NVMeQueuePair *q)
@@ -XXX,XX +XXX,XX @@ static NVMeQueuePair *nvme_create_queue_pair(BDRVNVMeState *s,
                                              Error **errp)
 {
     int i, r;
-    Error *local_err = NULL;
     NVMeQueuePair *q;
     uint64_t prp_list_iova;
 
@@ -XXX,XX +XXX,XX @@ static NVMeQueuePair *nvme_create_queue_pair(BDRVNVMeState *s,
         req->prp_list_iova = prp_list_iova + i * s->page_size;
     }
 
-    nvme_init_queue(s, &q->sq, size, NVME_SQ_ENTRY_BYTES, &local_err);
-    if (local_err) {
-        error_propagate(errp, local_err);
+    if (!nvme_init_queue(s, &q->sq, size, NVME_SQ_ENTRY_BYTES, errp)) {
         goto fail;
     }
     q->sq.doorbell = &s->doorbells[idx * s->doorbell_scale].sq_tail;
 
-    nvme_init_queue(s, &q->cq, size, NVME_CQ_ENTRY_BYTES, &local_err);
-    if (local_err) {
-        error_propagate(errp, local_err);
+    if (!nvme_init_queue(s, &q->cq, size, NVME_CQ_ENTRY_BYTES, errp)) {
         goto fail;
     }
     q->cq.doorbell = &s->doorbells[idx * s->doorbell_scale].cq_head;
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

Rename Submission Queue flags with 'Sq' to differentiate
submission queue flags from command queue flags, and introduce
Completion Queue flag definitions.

Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20201029093306.1063879-13-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 include/block/nvme.h | 18 ++++++++++++------
 1 file changed, 12 insertions(+), 6 deletions(-)

diff --git a/include/block/nvme.h b/include/block/nvme.h
index XXXXXXX..XXXXXXX 100644
--- a/include/block/nvme.h
+++ b/include/block/nvme.h
@@ -XXX,XX +XXX,XX @@ typedef struct QEMU_PACKED NvmeCreateCq {
 #define NVME_CQ_FLAGS_PC(cq_flags)  (cq_flags & 0x1)
 #define NVME_CQ_FLAGS_IEN(cq_flags) ((cq_flags >> 1) & 0x1)
 
+enum NvmeFlagsCq {
+    NVME_CQ_PC          = 1,
+    NVME_CQ_IEN         = 2,
+};
+
 typedef struct QEMU_PACKED NvmeCreateSq {
     uint8_t     opcode;
     uint8_t     flags;
@@ -XXX,XX +XXX,XX @@ typedef struct QEMU_PACKED NvmeCreateSq {
 #define NVME_SQ_FLAGS_PC(sq_flags)      (sq_flags & 0x1)
 #define NVME_SQ_FLAGS_QPRIO(sq_flags)   ((sq_flags >> 1) & 0x3)
 
-enum NvmeQueueFlags {
-    NVME_Q_PC           = 1,
-    NVME_Q_PRIO_URGENT  = 0,
-    NVME_Q_PRIO_HIGH    = 1,
-    NVME_Q_PRIO_NORMAL  = 2,
-    NVME_Q_PRIO_LOW     = 3,
+enum NvmeFlagsSq {
+    NVME_SQ_PC          = 1,
+
+    NVME_SQ_PRIO_URGENT = 0,
+    NVME_SQ_PRIO_HIGH   = 1,
+    NVME_SQ_PRIO_NORMAL = 2,
+    NVME_SQ_PRIO_LOW    = 3,
 };
 
 typedef struct QEMU_PACKED NvmeIdentify {
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

Replace magic values by definitions, and simplifiy since the
number of queues will never reach 64K.

Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-14-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ static bool nvme_add_io_queue(BlockDriverState *bs, Error **errp)
     NvmeCmd cmd;
     unsigned queue_size = NVME_QUEUE_SIZE;
 
+    assert(n <= UINT16_MAX);
     q = nvme_create_queue_pair(s, bdrv_get_aio_context(bs),
                                n, queue_size, errp);
     if (!q) {
@@ -XXX,XX +XXX,XX @@ static bool nvme_add_io_queue(BlockDriverState *bs, Error **errp)
     cmd = (NvmeCmd) {
         .opcode = NVME_ADM_CMD_CREATE_CQ,
         .dptr.prp1 = cpu_to_le64(q->cq.iova),
-        .cdw10 = cpu_to_le32(((queue_size - 1) << 16) | (n & 0xFFFF)),
-        .cdw11 = cpu_to_le32(0x3),
+        .cdw10 = cpu_to_le32(((queue_size - 1) << 16) | n),
+        .cdw11 = cpu_to_le32(NVME_CQ_IEN | NVME_CQ_PC),
     };
     if (nvme_cmd_sync(bs, s->queues[INDEX_ADMIN], &cmd)) {
         error_setg(errp, "Failed to create CQ io queue [%u]", n);
@@ -XXX,XX +XXX,XX @@ static bool nvme_add_io_queue(BlockDriverState *bs, Error **errp)
     cmd = (NvmeCmd) {
         .opcode = NVME_ADM_CMD_CREATE_SQ,
         .dptr.prp1 = cpu_to_le64(q->sq.iova),
-        .cdw10 = cpu_to_le32(((queue_size - 1) << 16) | (n & 0xFFFF)),
-        .cdw11 = cpu_to_le32(0x1 | (n << 16)),
+        .cdw10 = cpu_to_le32(((queue_size - 1) << 16) | n),
+        .cdw11 = cpu_to_le32(NVME_SQ_PC | (n << 16)),
     };
     if (nvme_cmd_sync(bs, s->queues[INDEX_ADMIN], &cmd)) {
         error_setg(errp, "Failed to create SQ io queue [%u]", n);
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

From the specification chapter 3.1.8 "AQA - Admin Queue Attributes"
the Admin Submission Queue Size field is a 0’s based value:

Admin Submission Queue Size (ASQS):

Defines the size of the Admin Submission Queue in entries.
    Enabling a controller while this field is cleared to 00h
    produces undefined results. The minimum size of the Admin
    Submission Queue is two entries. The maximum size of the
    Admin Submission Queue is 4096 entries.
    This is a 0’s based value.

This bug has never been hit because the device initialization
uses a single command synchronously :)

Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-15-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
         goto out;
     }
     s->queue_count = 1;
-    QEMU_BUILD_BUG_ON(NVME_QUEUE_SIZE & 0xF000);
-    regs->aqa = cpu_to_le32((NVME_QUEUE_SIZE << AQA_ACQS_SHIFT) |
-                            (NVME_QUEUE_SIZE << AQA_ASQS_SHIFT));
+    QEMU_BUILD_BUG_ON((NVME_QUEUE_SIZE - 1) & 0xF000);
+    regs->aqa = cpu_to_le32(((NVME_QUEUE_SIZE - 1) << AQA_ACQS_SHIFT) |
+                            ((NVME_QUEUE_SIZE - 1) << AQA_ASQS_SHIFT));
     regs->asq = cpu_to_le64(s->queues[INDEX_ADMIN]->sq.iova);
     regs->acq = cpu_to_le64(s->queues[INDEX_ADMIN]->cq.iova);
 
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

We don't need to dereference from BDRVNVMeState each time.
Use a NVMeQueuePair pointer on the admin queue.
The nvme_init() becomes easier to review, matching the style
of nvme_add_io_queue().

Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-16-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
                      Error **errp)
 {
     BDRVNVMeState *s = bs->opaque;
+    NVMeQueuePair *q;
     AioContext *aio_context = bdrv_get_aio_context(bs);
     int ret;
     uint64_t cap;
@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
 
     /* Set up admin queue. */
     s->queues = g_new(NVMeQueuePair *, 1);
-    s->queues[INDEX_ADMIN] = nvme_create_queue_pair(s, aio_context, 0,
-                                                          NVME_QUEUE_SIZE,
-                                                          errp);
-    if (!s->queues[INDEX_ADMIN]) {
+    q = nvme_create_queue_pair(s, aio_context, 0, NVME_QUEUE_SIZE, errp);
+    if (!q) {
         ret = -EINVAL;
         goto out;
     }
+    s->queues[INDEX_ADMIN] = q;
     s->queue_count = 1;
     QEMU_BUILD_BUG_ON((NVME_QUEUE_SIZE - 1) & 0xF000);
     regs->aqa = cpu_to_le32(((NVME_QUEUE_SIZE - 1) << AQA_ACQS_SHIFT) |
                             ((NVME_QUEUE_SIZE - 1) << AQA_ASQS_SHIFT));
-    regs->asq = cpu_to_le64(s->queues[INDEX_ADMIN]->sq.iova);
-    regs->acq = cpu_to_le64(s->queues[INDEX_ADMIN]->cq.iova);
+    regs->asq = cpu_to_le64(q->sq.iova);
+    regs->acq = cpu_to_le64(q->cq.iova);
 
     /* After setting up all control registers we can enable device now. */
     regs->cc = cpu_to_le32((ctz32(NVME_CQ_ENTRY_BYTES) << CC_IOCQES_SHIFT) |
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

As all commands use the ADMIN queue, it is pointless to pass
it as argument each time. Remove the argument, and rename the
function as nvme_admin_cmd_sync() to make this new behavior
clearer.

Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20201029093306.1063879-17-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c | 19 ++++++++++---------
 1 file changed, 10 insertions(+), 9 deletions(-)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ static void nvme_submit_command(NVMeQueuePair *q, NVMeRequest *req,
     qemu_mutex_unlock(&q->lock);
 }
 
-static void nvme_cmd_sync_cb(void *opaque, int ret)
+static void nvme_admin_cmd_sync_cb(void *opaque, int ret)
 {
     int *pret = opaque;
     *pret = ret;
     aio_wait_kick();
 }
 
-static int nvme_cmd_sync(BlockDriverState *bs, NVMeQueuePair *q,
-                         NvmeCmd *cmd)
+static int nvme_admin_cmd_sync(BlockDriverState *bs, NvmeCmd *cmd)
 {
+    BDRVNVMeState *s = bs->opaque;
+    NVMeQueuePair *q = s->queues[INDEX_ADMIN];
     AioContext *aio_context = bdrv_get_aio_context(bs);
     NVMeRequest *req;
     int ret = -EINPROGRESS;
@@ -XXX,XX +XXX,XX @@ static int nvme_cmd_sync(BlockDriverState *bs, NVMeQueuePair *q,
     if (!req) {
         return -EBUSY;
     }
-    nvme_submit_command(q, req, cmd, nvme_cmd_sync_cb, &ret);
+    nvme_submit_command(q, req, cmd, nvme_admin_cmd_sync_cb, &ret);
 
     AIO_WAIT_WHILE(aio_context, ret == -EINPROGRESS);
     return ret;
@@ -XXX,XX +XXX,XX @@ static bool nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
 
     memset(id, 0, sizeof(*id));
     cmd.dptr.prp1 = cpu_to_le64(iova);
-    if (nvme_cmd_sync(bs, s->queues[INDEX_ADMIN], &cmd)) {
+    if (nvme_admin_cmd_sync(bs, &cmd)) {
         error_setg(errp, "Failed to identify controller");
         goto out;
     }
@@ -XXX,XX +XXX,XX @@ static bool nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
     memset(id, 0, sizeof(*id));
     cmd.cdw10 = 0;
     cmd.nsid = cpu_to_le32(namespace);
-    if (nvme_cmd_sync(bs, s->queues[INDEX_ADMIN], &cmd)) {
+    if (nvme_admin_cmd_sync(bs, &cmd)) {
         error_setg(errp, "Failed to identify namespace");
         goto out;
     }
@@ -XXX,XX +XXX,XX @@ static bool nvme_add_io_queue(BlockDriverState *bs, Error **errp)
         .cdw10 = cpu_to_le32(((queue_size - 1) << 16) | n),
         .cdw11 = cpu_to_le32(NVME_CQ_IEN | NVME_CQ_PC),
     };
-    if (nvme_cmd_sync(bs, s->queues[INDEX_ADMIN], &cmd)) {
+    if (nvme_admin_cmd_sync(bs, &cmd)) {
         error_setg(errp, "Failed to create CQ io queue [%u]", n);
         goto out_error;
     }
@@ -XXX,XX +XXX,XX @@ static bool nvme_add_io_queue(BlockDriverState *bs, Error **errp)
         .cdw10 = cpu_to_le32(((queue_size - 1) << 16) | n),
         .cdw11 = cpu_to_le32(NVME_SQ_PC | (n << 16)),
     };
-    if (nvme_cmd_sync(bs, s->queues[INDEX_ADMIN], &cmd)) {
+    if (nvme_admin_cmd_sync(bs, &cmd)) {
         error_setg(errp, "Failed to create SQ io queue [%u]", n);
         goto out_error;
     }
@@ -XXX,XX +XXX,XX @@ static int nvme_enable_disable_write_cache(BlockDriverState *bs, bool enable,
         .cdw11 = cpu_to_le32(enable ? 0x01 : 0x00),
     };
 
-    ret = nvme_cmd_sync(bs, s->queues[INDEX_ADMIN], &cmd);
+    ret = nvme_admin_cmd_sync(bs, &cmd);
     if (ret) {
         error_setg(errp, "Failed to configure NVMe write cache");
     }
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

Commit bdd6a90a9e5 ("block: Add VFIO based NVMe driver")
sets the request_alignment in nvme_refresh_limits().
For consistency, also set it during initialization.

Reported-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-18-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
     s->page_size = MAX(4096, 1 << NVME_CAP_MPSMIN(cap));
     s->doorbell_scale = (4 << NVME_CAP_DSTRD(cap)) / sizeof(uint32_t);
     bs->bl.opt_mem_alignment = s->page_size;
+    bs->bl.request_alignment = s->page_size;
     timeout_ms = MIN(500 * NVME_CAP_TO(cap), 30000);
 
     /* Reset device to get a clean state. */
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

While trying to simplify the code using a macro, we forgot
the 12-bit shift... Correct that.

Fixes: fad1eb68862 ("block/nvme: Use register definitions from 'block/nvme.h'")
Reported-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-19-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
         goto out;
     }
 
-    s->page_size = MAX(4096, 1 << NVME_CAP_MPSMIN(cap));
+    s->page_size = 1u << (12 + NVME_CAP_MPSMIN(cap));
     s->doorbell_scale = (4 << NVME_CAP_DSTRD(cap)) / sizeof(uint32_t);
     bs->bl.opt_mem_alignment = s->page_size;
     bs->bl.request_alignment = s->page_size;
-- 
2.28.0

From: Eric Auger <eric.auger@redhat.com>

In preparation of 64kB host page support, let's change the size
and alignment of the IDENTIFY command response buffer so that
the VFIO DMA MAP succeeds. We align on the host page size.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-20-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ static bool nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
         .opcode = NVME_ADM_CMD_IDENTIFY,
         .cdw10 = cpu_to_le32(0x1),
     };
+    size_t id_size = QEMU_ALIGN_UP(sizeof(*id), qemu_real_host_page_size);
 
-    id = qemu_try_memalign(s->page_size, sizeof(*id));
+    id = qemu_try_memalign(qemu_real_host_page_size, id_size);
     if (!id) {
         error_setg(errp, "Cannot allocate buffer for identify response");
         goto out;
     }
-    r = qemu_vfio_dma_map(s->vfio, id, sizeof(*id), true, &iova);
+    r = qemu_vfio_dma_map(s->vfio, id, id_size, true, &iova);
     if (r) {
         error_setg(errp, "Cannot map buffer for DMA");
         goto out;
     }
 
-    memset(id, 0, sizeof(*id));
+    memset(id, 0, id_size);
     cmd.dptr.prp1 = cpu_to_le64(iova);
     if (nvme_admin_cmd_sync(bs, &cmd)) {
         error_setg(errp, "Failed to identify controller");
@@ -XXX,XX +XXX,XX @@ static bool nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
     s->supports_write_zeroes = !!(oncs & NVME_ONCS_WRITE_ZEROES);
     s->supports_discard = !!(oncs & NVME_ONCS_DSM);
 
-    memset(id, 0, sizeof(*id));
+    memset(id, 0, id_size);
     cmd.cdw10 = 0;
     cmd.nsid = cpu_to_le32(namespace);
     if (nvme_admin_cmd_sync(bs, &cmd)) {
-- 
2.28.0

From: Eric Auger <eric.auger@redhat.com>

In preparation of 64kB host page support, let's change the size
and alignment of the queue so that the VFIO DMA MAP succeeds.
We align on the host page size.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-21-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ static bool nvme_init_queue(BDRVNVMeState *s, NVMeQueue *q,
     size_t bytes;
     int r;
 
-    bytes = ROUND_UP(nentries * entry_bytes, s->page_size);
+    bytes = ROUND_UP(nentries * entry_bytes, qemu_real_host_page_size);
     q->head = q->tail = 0;
-    q->queue = qemu_try_memalign(s->page_size, bytes);
+    q->queue = qemu_try_memalign(qemu_real_host_page_size, bytes);
     if (!q->queue) {
         error_setg(errp, "Cannot allocate queue");
         return false;
-- 
2.28.0

From: Eric Auger <eric.auger@redhat.com>

In preparation of 64kB host page support, let's change the size
and alignment of the prp_list_pages so that the VFIO DMA MAP succeeds
with 64kB host page size. We align on the host page size.

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-22-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ static NVMeQueuePair *nvme_create_queue_pair(BDRVNVMeState *s,
     int i, r;
     NVMeQueuePair *q;
     uint64_t prp_list_iova;
+    size_t bytes;
 
     q = g_try_new0(NVMeQueuePair, 1);
     if (!q) {
@@ -XXX,XX +XXX,XX @@ static NVMeQueuePair *nvme_create_queue_pair(BDRVNVMeState *s,
     }
     trace_nvme_create_queue_pair(idx, q, size, aio_context,
                                  event_notifier_get_fd(s->irq_notifier));
-    q->prp_list_pages = qemu_try_memalign(s->page_size,
-                                          s->page_size * NVME_NUM_REQS);
+    bytes = QEMU_ALIGN_UP(s->page_size * NVME_NUM_REQS,
+                          qemu_real_host_page_size);
+    q->prp_list_pages = qemu_try_memalign(qemu_real_host_page_size, bytes);
     if (!q->prp_list_pages) {
         goto fail;
     }
-    memset(q->prp_list_pages, 0, s->page_size * NVME_NUM_REQS);
+    memset(q->prp_list_pages, 0, bytes);
     qemu_mutex_init(&q->lock);
     q->s = s;
     q->index = idx;
     qemu_co_queue_init(&q->free_req_queue);
     q->completion_bh = aio_bh_new(aio_context, nvme_process_completion_bh, q);
-    r = qemu_vfio_dma_map(s->vfio, q->prp_list_pages,
-                          s->page_size * NVME_NUM_REQS,
+    r = qemu_vfio_dma_map(s->vfio, q->prp_list_pages, bytes,
                           false, &prp_list_iova);
     if (r) {
         goto fail;
-- 
2.28.0

From: Eric Auger <eric.auger@redhat.com>

Make sure iov's va and size are properly aligned on the
host page size.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-23-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ static coroutine_fn int nvme_cmd_map_qiov(BlockDriverState *bs, NvmeCmd *cmd,
     for (i = 0; i < qiov->niov; ++i) {
         bool retry = true;
         uint64_t iova;
+        size_t len = QEMU_ALIGN_UP(qiov->iov[i].iov_len,
+                                   qemu_real_host_page_size);
 try_map:
         r = qemu_vfio_dma_map(s->vfio,
                               qiov->iov[i].iov_base,
-                              qiov->iov[i].iov_len,
-                              true, &iova);
+                              len, true, &iova);
         if (r == -ENOMEM && retry) {
             retry = false;
             trace_nvme_dma_flush_queue_wait(s);
@@ -XXX,XX +XXX,XX @@ static inline bool nvme_qiov_aligned(BlockDriverState *bs,
     BDRVNVMeState *s = bs->opaque;
 
     for (i = 0; i < qiov->niov; ++i) {
-        if (!QEMU_PTR_IS_ALIGNED(qiov->iov[i].iov_base, s->page_size) ||
-            !QEMU_IS_ALIGNED(qiov->iov[i].iov_len, s->page_size)) {
+        if (!QEMU_PTR_IS_ALIGNED(qiov->iov[i].iov_base,
+                                 qemu_real_host_page_size) ||
+            !QEMU_IS_ALIGNED(qiov->iov[i].iov_len, qemu_real_host_page_size)) {
             trace_nvme_qiov_unaligned(qiov, i, qiov->iov[i].iov_base,
                                       qiov->iov[i].iov_len, s->page_size);
             return false;
@@ -XXX,XX +XXX,XX @@ static int nvme_co_prw(BlockDriverState *bs, uint64_t offset, uint64_t bytes,
     int r;
     uint8_t *buf = NULL;
     QEMUIOVector local_qiov;
-
+    size_t len = QEMU_ALIGN_UP(bytes, qemu_real_host_page_size);
     assert(QEMU_IS_ALIGNED(offset, s->page_size));
     assert(QEMU_IS_ALIGNED(bytes, s->page_size));
     assert(bytes <= s->max_transfer);
@@ -XXX,XX +XXX,XX @@ static int nvme_co_prw(BlockDriverState *bs, uint64_t offset, uint64_t bytes,
     }
     s->stats.unaligned_accesses++;
     trace_nvme_prw_buffered(s, offset, bytes, qiov->niov, is_write);
-    buf = qemu_try_memalign(s->page_size, bytes);
+    buf = qemu_try_memalign(qemu_real_host_page_size, len);
 
     if (!buf) {
         return -ENOMEM;
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

qemu_vfio_pci_map_bar() calls mmap(), and mmap(2) states:

'offset' must be a multiple of the page size as returned
   by sysconf(_SC_PAGE_SIZE).

In commit f68453237b9 we started to use an offset of 4K which
broke this contract on Aarch64 arch.

Fix by mapping at offset 0, and and accessing doorbells at offset=4K.

Fixes: f68453237b9 ("block/nvme: Map doorbells pages write-only")
Reported-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-24-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ typedef struct {
 struct BDRVNVMeState {
     AioContext *aio_context;
     QEMUVFIOState *vfio;
+    void *bar0_wo_map;
     /* Memory mapped registers */
     volatile struct {
         uint32_t sq_tail;
@@ -XXX,XX +XXX,XX @@ static int nvme_init(BlockDriverState *bs, const char *device, int namespace,
         }
     }
 
-    s->doorbells = qemu_vfio_pci_map_bar(s->vfio, 0, sizeof(NvmeBar),
-                                         NVME_DOORBELL_SIZE, PROT_WRITE, errp);
+    s->bar0_wo_map = qemu_vfio_pci_map_bar(s->vfio, 0, 0,
+                                           sizeof(NvmeBar) + NVME_DOORBELL_SIZE,
+                                           PROT_WRITE, errp);
+    s->doorbells = (void *)((uintptr_t)s->bar0_wo_map + sizeof(NvmeBar));
     if (!s->doorbells) {
         ret = -EINVAL;
         goto out;
@@ -XXX,XX +XXX,XX @@ static void nvme_close(BlockDriverState *bs)
                            &s->irq_notifier[MSIX_SHARED_IRQ_IDX],
                            false, NULL, NULL);
     event_notifier_cleanup(&s->irq_notifier[MSIX_SHARED_IRQ_IDX]);
-    qemu_vfio_pci_unmap_bar(s->vfio, 0, (void *)s->doorbells,
-                            sizeof(NvmeBar), NVME_DOORBELL_SIZE);
+    qemu_vfio_pci_unmap_bar(s->vfio, 0, s->bar0_wo_map,
+                            0, sizeof(NvmeBar) + NVME_DOORBELL_SIZE);
     qemu_vfio_close(s->vfio);
 
     g_free(s->device);
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

The Completion Queue Command Identifier is a 16-bit value,
so nvme_submit_command() is unlikely to work on big-endian
hosts, as the relevant bits are truncated.
Fix by using the correct byte-swap function.

Fixes: bdd6a90a9e5 ("block: Add VFIO based NVMe driver")
Reported-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20201029093306.1063879-25-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 block/nvme.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/block/nvme.c b/block/nvme.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -XXX,XX +XXX,XX @@ static void nvme_submit_command(NVMeQueuePair *q, NVMeRequest *req,
     assert(!req->cb);
     req->cb = cb;
     req->opaque = opaque;
-    cmd->cid = cpu_to_le32(req->cid);
+    cmd->cid = cpu_to_le16(req->cid);
 
     trace_nvme_submit_command(q->s, q->index, req->cid);
     nvme_trace_command(cmd);
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

Change the confuse "VFIO IOMMU check failed" error message by
the explicit "VFIO IOMMU Type1 is not supported" once.

Example on POWER:

$ qemu-system-ppc64 -drive if=none,id=nvme0,file=nvme://0001:01:00.0/1,format=raw
 qemu-system-ppc64: -drive if=none,id=nvme0,file=nvme://0001:01:00.0/1,format=raw: VFIO IOMMU Type1 is not supported

Suggested-by: Alex Williamson <alex.williamson@redhat.com>
Reviewed-by: Fam Zheng <fam@euphon.net>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201103020733.2303148-2-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 util/vfio-helpers.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/util/vfio-helpers.c b/util/vfio-helpers.c
index XXXXXXX..XXXXXXX 100644
--- a/util/vfio-helpers.c
+++ b/util/vfio-helpers.c
@@ -XXX,XX +XXX,XX @@ static int qemu_vfio_init_pci(QEMUVFIOState *s, const char *device,
     }
 
     if (!ioctl(s->container, VFIO_CHECK_EXTENSION, VFIO_TYPE1_IOMMU)) {
-        error_setg_errno(errp, errno, "VFIO IOMMU check failed");
+        error_setg_errno(errp, errno, "VFIO IOMMU Type1 is not supported");
         ret = -EINVAL;
         goto fail_container;
     }
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

We sometime get kernel panic with some devices on Aarch64
hosts. Alex Williamson suggests it might be broken PCIe
root complex. Add trace event to record the latest I/O
access before crashing. In case, assert our accesses are
aligned.

Reviewed-by: Fam Zheng <fam@euphon.net>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201103020733.2303148-3-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 util/vfio-helpers.c | 8 ++++++++
 util/trace-events   | 2 ++
 2 files changed, 10 insertions(+)

diff --git a/util/vfio-helpers.c b/util/vfio-helpers.c
index XXXXXXX..XXXXXXX 100644
--- a/util/vfio-helpers.c
+++ b/util/vfio-helpers.c
@@ -XXX,XX +XXX,XX @@ static int qemu_vfio_pci_read_config(QEMUVFIOState *s, void *buf,
 {
     int ret;
 
+    trace_qemu_vfio_pci_read_config(buf, ofs, size,
+                                    s->config_region_info.offset,
+                                    s->config_region_info.size);
+    assert(QEMU_IS_ALIGNED(s->config_region_info.offset + ofs, size));
     do {
         ret = pread(s->device, buf, size, s->config_region_info.offset + ofs);
     } while (ret == -1 && errno == EINTR);
@@ -XXX,XX +XXX,XX @@ static int qemu_vfio_pci_write_config(QEMUVFIOState *s, void *buf, int size, int
 {
     int ret;
 
+    trace_qemu_vfio_pci_write_config(buf, ofs, size,
+                                     s->config_region_info.offset,
+                                     s->config_region_info.size);
+    assert(QEMU_IS_ALIGNED(s->config_region_info.offset + ofs, size));
     do {
         ret = pwrite(s->device, buf, size, s->config_region_info.offset + ofs);
     } while (ret == -1 && errno == EINTR);
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_vfio_new_mapping(void *s, void *host, size_t size, int index, uint64_t iova
 qemu_vfio_do_mapping(void *s, void *host, size_t size, uint64_t iova) "s %p host %p size 0x%zx iova 0x%"PRIx64
 qemu_vfio_dma_map(void *s, void *host, size_t size, bool temporary, uint64_t *iova) "s %p host %p size 0x%zx temporary %d iova %p"
 qemu_vfio_dma_unmap(void *s, void *host) "s %p host %p"
+qemu_vfio_pci_read_config(void *buf, int ofs, int size, uint64_t region_ofs, uint64_t region_size) "read cfg ptr %p ofs 0x%x size 0x%x (region addr 0x%"PRIx64" size 0x%"PRIx64")"
+qemu_vfio_pci_write_config(void *buf, int ofs, int size, uint64_t region_ofs, uint64_t region_size) "write cfg ptr %p ofs 0x%x size 0x%x (region addr 0x%"PRIx64" size 0x%"PRIx64")"
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

For debug purpose, trace BAR regions info.

Reviewed-by: Fam Zheng <fam@euphon.net>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201103020733.2303148-4-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 util/vfio-helpers.c | 8 ++++++++
 util/trace-events   | 1 +
 2 files changed, 9 insertions(+)

diff --git a/util/vfio-helpers.c b/util/vfio-helpers.c
index XXXXXXX..XXXXXXX 100644
--- a/util/vfio-helpers.c
+++ b/util/vfio-helpers.c
@@ -XXX,XX +XXX,XX @@ static inline void assert_bar_index_valid(QEMUVFIOState *s, int index)
 
 static int qemu_vfio_pci_init_bar(QEMUVFIOState *s, int index, Error **errp)
 {
+    g_autofree char *barname = NULL;
     assert_bar_index_valid(s, index);
     s->bar_region_info[index] = (struct vfio_region_info) {
         .index = VFIO_PCI_BAR0_REGION_INDEX + index,
@@ -XXX,XX +XXX,XX @@ static int qemu_vfio_pci_init_bar(QEMUVFIOState *s, int index, Error **errp)
         error_setg_errno(errp, errno, "Failed to get BAR region info");
         return -errno;
     }
+    barname = g_strdup_printf("bar[%d]", index);
+    trace_qemu_vfio_region_info(barname, s->bar_region_info[index].offset,
+                                s->bar_region_info[index].size,
+                                s->bar_region_info[index].cap_offset);
 
     return 0;
 }
@@ -XXX,XX +XXX,XX @@ static int qemu_vfio_init_pci(QEMUVFIOState *s, const char *device,
         ret = -errno;
         goto fail;
     }
+    trace_qemu_vfio_region_info("config", s->config_region_info.offset,
+                                s->config_region_info.size,
+                                s->config_region_info.cap_offset);
 
     for (i = 0; i < ARRAY_SIZE(s->bar_region_info); i++) {
         ret = qemu_vfio_pci_init_bar(s, i, errp);
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_vfio_dma_map(void *s, void *host, size_t size, bool temporary, uint64_t *io
 qemu_vfio_dma_unmap(void *s, void *host) "s %p host %p"
 qemu_vfio_pci_read_config(void *buf, int ofs, int size, uint64_t region_ofs, uint64_t region_size) "read cfg ptr %p ofs 0x%x size 0x%x (region addr 0x%"PRIx64" size 0x%"PRIx64")"
 qemu_vfio_pci_write_config(void *buf, int ofs, int size, uint64_t region_ofs, uint64_t region_size) "write cfg ptr %p ofs 0x%x size 0x%x (region addr 0x%"PRIx64" size 0x%"PRIx64")"
+qemu_vfio_region_info(const char *desc, uint64_t region_ofs, uint64_t region_size, uint32_t cap_offset) "region '%s' addr 0x%"PRIx64" size 0x%"PRIx64" cap_ofs 0x%"PRIx32
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

For debugging purpose, trace where a BAR is mapped.

Reviewed-by: Fam Zheng <fam@euphon.net>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201103020733.2303148-5-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 util/vfio-helpers.c | 2 ++
 util/trace-events   | 1 +
 2 files changed, 3 insertions(+)

diff --git a/util/vfio-helpers.c b/util/vfio-helpers.c
index XXXXXXX..XXXXXXX 100644
--- a/util/vfio-helpers.c
+++ b/util/vfio-helpers.c
@@ -XXX,XX +XXX,XX @@ void *qemu_vfio_pci_map_bar(QEMUVFIOState *s, int index,
     p = mmap(NULL, MIN(size, s->bar_region_info[index].size - offset),
              prot, MAP_SHARED,
              s->device, s->bar_region_info[index].offset + offset);
+    trace_qemu_vfio_pci_map_bar(index, s->bar_region_info[index].offset ,
+                                size, offset, p);
     if (p == MAP_FAILED) {
         error_setg_errno(errp, errno, "Failed to map BAR region");
         p = NULL;
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_vfio_dma_unmap(void *s, void *host) "s %p host %p"
 qemu_vfio_pci_read_config(void *buf, int ofs, int size, uint64_t region_ofs, uint64_t region_size) "read cfg ptr %p ofs 0x%x size 0x%x (region addr 0x%"PRIx64" size 0x%"PRIx64")"
 qemu_vfio_pci_write_config(void *buf, int ofs, int size, uint64_t region_ofs, uint64_t region_size) "write cfg ptr %p ofs 0x%x size 0x%x (region addr 0x%"PRIx64" size 0x%"PRIx64")"
 qemu_vfio_region_info(const char *desc, uint64_t region_ofs, uint64_t region_size, uint32_t cap_offset) "region '%s' addr 0x%"PRIx64" size 0x%"PRIx64" cap_ofs 0x%"PRIx32
+qemu_vfio_pci_map_bar(int index, uint64_t region_ofs, uint64_t region_size, int ofs, void *host) "map region bar#%d addr 0x%"PRIx64" size 0x%"PRIx64" ofs 0x%x host %p"
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

For debugging purpose, trace where DMA regions are mapped.

Reviewed-by: Fam Zheng <fam@euphon.net>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201103020733.2303148-6-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 util/vfio-helpers.c | 3 ++-
 util/trace-events   | 5 +++--
 2 files changed, 5 insertions(+), 3 deletions(-)

diff --git a/util/vfio-helpers.c b/util/vfio-helpers.c
index XXXXXXX..XXXXXXX 100644
--- a/util/vfio-helpers.c
+++ b/util/vfio-helpers.c
@@ -XXX,XX +XXX,XX @@ static int qemu_vfio_do_mapping(QEMUVFIOState *s, void *host, size_t size,
         .vaddr = (uintptr_t)host,
         .size = size,
     };
-    trace_qemu_vfio_do_mapping(s, host, size, iova);
+    trace_qemu_vfio_do_mapping(s, host, iova, size);
 
     if (ioctl(s->container, VFIO_IOMMU_MAP_DMA, &dma_map)) {
         error_report("VFIO_MAP_DMA failed: %s", strerror(errno));
@@ -XXX,XX +XXX,XX @@ int qemu_vfio_dma_map(QEMUVFIOState *s, void *host, size_t size,
             }
         }
     }
+    trace_qemu_vfio_dma_mapped(s, host, iova0, size);
     if (iova) {
         *iova = iova0;
     }
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_vfio_ram_block_added(void *s, void *p, size_t size) "s %p host %p size 0x%z
 qemu_vfio_ram_block_removed(void *s, void *p, size_t size) "s %p host %p size 0x%zx"
 qemu_vfio_find_mapping(void *s, void *p) "s %p host %p"
 qemu_vfio_new_mapping(void *s, void *host, size_t size, int index, uint64_t iova) "s %p host %p size 0x%zx index %d iova 0x%"PRIx64
-qemu_vfio_do_mapping(void *s, void *host, size_t size, uint64_t iova) "s %p host %p size 0x%zx iova 0x%"PRIx64
-qemu_vfio_dma_map(void *s, void *host, size_t size, bool temporary, uint64_t *iova) "s %p host %p size 0x%zx temporary %d iova %p"
+qemu_vfio_do_mapping(void *s, void *host, uint64_t iova, size_t size) "s %p host %p <-> iova 0x%"PRIx64 " size 0x%zx"
+qemu_vfio_dma_map(void *s, void *host, size_t size, bool temporary, uint64_t *iova) "s %p host %p size 0x%zx temporary %d &iova %p"
+qemu_vfio_dma_mapped(void *s, void *host, uint64_t iova, size_t size) "s %p host %p <-> iova 0x%"PRIx64" size 0x%zx"
 qemu_vfio_dma_unmap(void *s, void *host) "s %p host %p"
 qemu_vfio_pci_read_config(void *buf, int ofs, int size, uint64_t region_ofs, uint64_t region_size) "read cfg ptr %p ofs 0x%x size 0x%x (region addr 0x%"PRIx64" size 0x%"PRIx64")"
 qemu_vfio_pci_write_config(void *buf, int ofs, int size, uint64_t region_ofs, uint64_t region_size) "write cfg ptr %p ofs 0x%x size 0x%x (region addr 0x%"PRIx64" size 0x%"PRIx64")"
-- 
2.28.0

From: Philippe Mathieu-Daudé <philmd@redhat.com>

The QEMU_VFIO_DEBUG definition is only modifiable at build-time.
Trace events can be enabled at run-time. As we prefer the latter,
convert qemu_vfio_dump_mappings() to use trace events instead
of fprintf().

Reviewed-by: Fam Zheng <fam@euphon.net>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201103020733.2303148-7-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
---
 util/vfio-helpers.c | 19 ++++---------------
 util/trace-events   |  1 +
 2 files changed, 5 insertions(+), 15 deletions(-)

diff --git a/util/vfio-helpers.c b/util/vfio-helpers.c
index XXXXXXX..XXXXXXX 100644
--- a/util/vfio-helpers.c
+++ b/util/vfio-helpers.c
@@ -XXX,XX +XXX,XX @@ QEMUVFIOState *qemu_vfio_open_pci(const char *device, Error **errp)
     return s;
 }
 
-static void qemu_vfio_dump_mapping(IOVAMapping *m)
-{
-    if (QEMU_VFIO_DEBUG) {
-        printf("  vfio mapping %p %" PRIx64 " to %" PRIx64 "\n", m->host,
-               (uint64_t)m->size, (uint64_t)m->iova);
-    }
-}
-
 static void qemu_vfio_dump_mappings(QEMUVFIOState *s)
 {
-    int i;
-
-    if (QEMU_VFIO_DEBUG) {
-        printf("vfio mappings\n");
-        for (i = 0; i < s->nr_mappings; ++i) {
-            qemu_vfio_dump_mapping(&s->mappings[i]);
-        }
+    for (int i = 0; i < s->nr_mappings; ++i) {
+        trace_qemu_vfio_dump_mapping(s->mappings[i].host,
+                                     s->mappings[i].iova,
+                                     s->mappings[i].size);
     }
 }
 
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_mutex_unlock(void *mutex, const char *file, const int line) "released mutex
 qemu_vfio_dma_reset_temporary(void *s) "s %p"
 qemu_vfio_ram_block_added(void *s, void *p, size_t size) "s %p host %p size 0x%zx"
 qemu_vfio_ram_block_removed(void *s, void *p, size_t size) "s %p host %p size 0x%zx"
+qemu_vfio_dump_mapping(void *host, uint64_t iova, size_t size) "vfio mapping %p to iova 0x%08" PRIx64 " size 0x%zx"
 qemu_vfio_find_mapping(void *s, void *p) "s %p host %p"
 qemu_vfio_new_mapping(void *s, void *host, size_t size, int index, uint64_t iova) "s %p host %p size 0x%zx index %d iova 0x%"PRIx64
 qemu_vfio_do_mapping(void *s, void *host, uint64_t iova, size_t size) "s %p host %p <-> iova 0x%"PRIx64 " size 0x%zx"
-- 
2.28.0

The following changes since commit 6c769690ac845fa62642a5f93b4e4bd906adab95:

Merge remote-tracking branch 'remotes/vsementsov/tags/pull-simplebench-2021-05-04' into staging (2021-05-21 12:02:34 +0100)

are available in the Git repository at:

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

for you to fetch changes up to 0a6f0c76a030710780ce10d6347a70f098024d21:

coroutine-sleep: introduce qemu_co_sleep (2021-05-21 18:22:33 +0100)

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

(Resent due to an email preparation mistake.)

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

Paolo Bonzini (6):
  coroutine-sleep: use a stack-allocated timer
  coroutine-sleep: disallow NULL QemuCoSleepState** argument
  coroutine-sleep: allow qemu_co_sleep_wake that wakes nothing
  coroutine-sleep: move timer out of QemuCoSleepState
  coroutine-sleep: replace QemuCoSleepState pointer with struct in the
    API
  coroutine-sleep: introduce qemu_co_sleep

Philippe Mathieu-Daudé (1):
  bitops.h: Improve find_xxx_bit() documentation

Zenghui Yu (1):
  multi-process: Initialize variables declared with g_auto*

-- 
2.31.1

From: Zenghui Yu <yuzenghui@huawei.com>

Quote docs/devel/style.rst (section "Automatic memory deallocation"):

* Variables declared with g_auto* MUST always be initialized,
  otherwise the cleanup function will use uninitialized stack memory

Initialize @name properly to get rid of the compilation error (using
gcc-7.3.0 on CentOS):

../hw/remote/proxy.c: In function 'pci_proxy_dev_realize':
/usr/include/glib-2.0/glib/glib-autocleanups.h:28:3: error: 'name' may be used uninitialized in this function [-Werror=maybe-uninitialized]
   g_free (*pp);
   ^~~~~~~~~~~~
../hw/remote/proxy.c:350:30: note: 'name' was declared here
             g_autofree char *name;
                              ^~~~

Signed-off-by: Zenghui Yu <yuzenghui@huawei.com>
Reviewed-by: Jagannathan Raman <jag.raman@oracle.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Miroslav Rezanina <mrezanin@redhat.com>
Message-id: 20210312112143.1369-1-yuzenghui@huawei.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 hw/remote/memory.c | 5 ++---
 hw/remote/proxy.c  | 3 +--
 2 files changed, 3 insertions(+), 5 deletions(-)

diff --git a/hw/remote/memory.c b/hw/remote/memory.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/remote/memory.c
+++ b/hw/remote/memory.c
@@ -XXX,XX +XXX,XX @@ void remote_sysmem_reconfig(MPQemuMsg *msg, Error **errp)
 
     remote_sysmem_reset();
 
-    for (region = 0; region < msg->num_fds; region++) {
-        g_autofree char *name;
+    for (region = 0; region < msg->num_fds; region++, suffix++) {
+        g_autofree char *name = g_strdup_printf("remote-mem-%u", suffix);
         subregion = g_new(MemoryRegion, 1);
-        name = g_strdup_printf("remote-mem-%u", suffix++);
         memory_region_init_ram_from_fd(subregion, NULL,
                                        name, sysmem_info->sizes[region],
                                        true, msg->fds[region],
diff --git a/hw/remote/proxy.c b/hw/remote/proxy.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/remote/proxy.c
+++ b/hw/remote/proxy.c
@@ -XXX,XX +XXX,XX @@ static void probe_pci_info(PCIDevice *dev, Error **errp)
                    PCI_BASE_ADDRESS_SPACE_IO : PCI_BASE_ADDRESS_SPACE_MEMORY;
 
         if (size) {
-            g_autofree char *name;
+            g_autofree char *name = g_strdup_printf("bar-region-%d", i);
             pdev->region[i].dev = pdev;
             pdev->region[i].present = true;
             if (type == PCI_BASE_ADDRESS_SPACE_MEMORY) {
                 pdev->region[i].memory = true;
             }
-            name = g_strdup_printf("bar-region-%d", i);
             memory_region_init_io(&pdev->region[i].mr, OBJECT(pdev),
                                   &proxy_mr_ops, &pdev->region[i],
                                   name, size);
-- 
2.31.1

From: Philippe Mathieu-Daudé <philmd@redhat.com>

Document the following functions return the bitmap size
if no matching bit is found:

- find_first_bit
- find_next_bit
- find_last_bit
- find_first_zero_bit
- find_next_zero_bit

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20210510200758.2623154-2-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 include/qemu/bitops.h | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

diff --git a/include/qemu/bitops.h b/include/qemu/bitops.h
index XXXXXXX..XXXXXXX 100644
--- a/include/qemu/bitops.h
+++ b/include/qemu/bitops.h
@@ -XXX,XX +XXX,XX @@ static inline int test_bit(long nr, const unsigned long *addr)
  * @addr: The address to start the search at
  * @size: The maximum size to search
  *
- * Returns the bit number of the first set bit, or size.
+ * Returns the bit number of the last set bit,
+ * or @size if there is no set bit in the bitmap.
  */
 unsigned long find_last_bit(const unsigned long *addr,
                             unsigned long size);
@@ -XXX,XX +XXX,XX @@ unsigned long find_last_bit(const unsigned long *addr,
  * @addr: The address to base the search on
  * @offset: The bitnumber to start searching at
  * @size: The bitmap size in bits
+ *
+ * Returns the bit number of the next set bit,
+ * or @size if there are no further set bits in the bitmap.
  */
 unsigned long find_next_bit(const unsigned long *addr,
                             unsigned long size,
@@ -XXX,XX +XXX,XX @@ unsigned long find_next_bit(const unsigned long *addr,
  * @addr: The address to base the search on
  * @offset: The bitnumber to start searching at
  * @size: The bitmap size in bits
+ *
+ * Returns the bit number of the next cleared bit,
+ * or @size if there are no further clear bits in the bitmap.
  */
 
 unsigned long find_next_zero_bit(const unsigned long *addr,
@@ -XXX,XX +XXX,XX @@ unsigned long find_next_zero_bit(const unsigned long *addr,
  * @addr: The address to start the search at
  * @size: The maximum size to search
  *
- * Returns the bit number of the first set bit.
+ * Returns the bit number of the first set bit,
+ * or @size if there is no set bit in the bitmap.
  */
 static inline unsigned long find_first_bit(const unsigned long *addr,
                                            unsigned long size)
@@ -XXX,XX +XXX,XX @@ static inline unsigned long find_first_bit(const unsigned long *addr,
  * @addr: The address to start the search at
  * @size: The maximum size to search
  *
- * Returns the bit number of the first cleared bit.
+ * Returns the bit number of the first cleared bit,
+ * or @size if there is no clear bit in the bitmap.
  */
 static inline unsigned long find_first_zero_bit(const unsigned long *addr,
                                                 unsigned long size)
-- 
2.31.1

From: Paolo Bonzini <pbonzini@redhat.com>

The lifetime of the timer is well-known (it cannot outlive
qemu_co_sleep_ns_wakeable, because it's deleted by the time the
coroutine resumes), so it is not necessary to place it on the heap.

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20210517100548.28806-2-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 util/qemu-coroutine-sleep.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

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 const char *qemu_co_sleep_ns__scheduled = "qemu_co_sleep_ns";
 
 struct QemuCoSleepState {
     Coroutine *co;
-    QEMUTimer *ts;
+    QEMUTimer ts;
     QemuCoSleepState **user_state_pointer;
 };
 
@@ -XXX,XX +XXX,XX @@ void qemu_co_sleep_wake(QemuCoSleepState *sleep_state)
     if (sleep_state->user_state_pointer) {
         *sleep_state->user_state_pointer = NULL;
     }
-    timer_del(sleep_state->ts);
+    timer_del(&sleep_state->ts);
     aio_co_wake(sleep_state->co);
 }
 
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
     AioContext *ctx = qemu_get_current_aio_context();
     QemuCoSleepState state = {
         .co = qemu_coroutine_self(),
-        .ts = aio_timer_new(ctx, type, SCALE_NS, co_sleep_cb, &state),
         .user_state_pointer = sleep_state,
     };
 
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
         abort();
     }
 
+    aio_timer_init(ctx, &state.ts, type, SCALE_NS, co_sleep_cb, &state);
     if (sleep_state) {
         *sleep_state = &state;
     }
-    timer_mod(state.ts, qemu_clock_get_ns(type) + ns);
+    timer_mod(&state.ts, qemu_clock_get_ns(type) + ns);
     qemu_coroutine_yield();
     if (sleep_state) {
         /*
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
          */
         assert(*sleep_state == NULL);
     }
-    timer_free(state.ts);
 }
-- 
2.31.1

From: Paolo Bonzini <pbonzini@redhat.com>

Simplify the code by removing conditionals.  qemu_co_sleep_ns
can simply point the argument to an on-stack temporary.

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20210517100548.28806-3-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 include/qemu/coroutine.h    |  5 +++--
 util/qemu-coroutine-sleep.c | 18 +++++-------------
 2 files changed, 8 insertions(+), 15 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 QemuCoSleepState QemuCoSleepState;
 
 /**
  * Yield the coroutine for a given duration. During this yield, @sleep_state
- * (if not NULL) is set to an opaque pointer, which may be used for
+ * is set to an opaque pointer, which may be used for
  * qemu_co_sleep_wake(). Be careful, the pointer is set back to zero when the
  * timer fires. Don't save the obtained value to other variables and don't call
  * qemu_co_sleep_wake from another aio context.
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
                                             QemuCoSleepState **sleep_state);
 static inline void coroutine_fn qemu_co_sleep_ns(QEMUClockType type, int64_t ns)
 {
-    qemu_co_sleep_ns_wakeable(type, ns, NULL);
+    QemuCoSleepState *unused = NULL;
+    qemu_co_sleep_ns_wakeable(type, ns, &unused);
 }
 
 /**
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 @@ void qemu_co_sleep_wake(QemuCoSleepState *sleep_state)
                                            qemu_co_sleep_ns__scheduled, NULL);
 
     assert(scheduled == qemu_co_sleep_ns__scheduled);
-    if (sleep_state->user_state_pointer) {
-        *sleep_state->user_state_pointer = NULL;
-    }
+    *sleep_state->user_state_pointer = NULL;
     timer_del(&sleep_state->ts);
     aio_co_wake(sleep_state->co);
 }
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
     }
 
     aio_timer_init(ctx, &state.ts, type, SCALE_NS, co_sleep_cb, &state);
-    if (sleep_state) {
-        *sleep_state = &state;
-    }
+    *sleep_state = &state;
     timer_mod(&state.ts, qemu_clock_get_ns(type) + ns);
     qemu_coroutine_yield();
-    if (sleep_state) {
-        /*
-         * Note that *sleep_state is cleared during qemu_co_sleep_wake
-         * before resuming this coroutine.
-         */
-        assert(*sleep_state == NULL);
-    }
+
+    /* qemu_co_sleep_wake clears *sleep_state before resuming this coroutine.  */
+    assert(*sleep_state == NULL);
 }
-- 
2.31.1

From: Paolo Bonzini <pbonzini@redhat.com>

All callers of qemu_co_sleep_wake are checking whether they are passing
a NULL argument inside the pointer-to-pointer: do the check in
qemu_co_sleep_wake itself.

As a side effect, qemu_co_sleep_wake can be called more than once and
it will only wake the coroutine once; after the first time, the argument
will be set to NULL via *sleep_state->user_state_pointer.  However, this
would not be safe unless co_sleep_cb keeps using the QemuCoSleepState*
directly, so make it go through the pointer-to-pointer instead.

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20210517100548.28806-4-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 block/block-copy.c          |  4 +---
 block/nbd.c                 |  8 ++------
 util/qemu-coroutine-sleep.c | 21 ++++++++++++---------
 3 files changed, 15 insertions(+), 18 deletions(-)

diff --git a/block/block-copy.c b/block/block-copy.c
index XXXXXXX..XXXXXXX 100644
--- a/block/block-copy.c
+++ b/block/block-copy.c
@@ -XXX,XX +XXX,XX @@ out:
 
 void block_copy_kick(BlockCopyCallState *call_state)
 {
-    if (call_state->sleep_state) {
-        qemu_co_sleep_wake(call_state->sleep_state);
-    }
+    qemu_co_sleep_wake(call_state->sleep_state);
 }
 
 /*
diff --git a/block/nbd.c b/block/nbd.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nbd.c
+++ b/block/nbd.c
@@ -XXX,XX +XXX,XX @@ static void coroutine_fn nbd_client_co_drain_begin(BlockDriverState *bs)
     BDRVNBDState *s = (BDRVNBDState *)bs->opaque;
 
     s->drained = true;
-    if (s->connection_co_sleep_ns_state) {
-        qemu_co_sleep_wake(s->connection_co_sleep_ns_state);
-    }
+    qemu_co_sleep_wake(s->connection_co_sleep_ns_state);
 
     nbd_co_establish_connection_cancel(bs, false);
 
@@ -XXX,XX +XXX,XX @@ static void nbd_teardown_connection(BlockDriverState *bs)
 
     s->state = NBD_CLIENT_QUIT;
     if (s->connection_co) {
-        if (s->connection_co_sleep_ns_state) {
-            qemu_co_sleep_wake(s->connection_co_sleep_ns_state);
-        }
+        qemu_co_sleep_wake(s->connection_co_sleep_ns_state);
         nbd_co_establish_connection_cancel(bs, true);
     }
     if (qemu_in_coroutine()) {
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 @@ struct QemuCoSleepState {
 
 void qemu_co_sleep_wake(QemuCoSleepState *sleep_state)
 {
-    /* Write of schedule protected by barrier write in aio_co_schedule */
-    const char *scheduled = qatomic_cmpxchg(&sleep_state->co->scheduled,
-                                           qemu_co_sleep_ns__scheduled, NULL);
+    if (sleep_state) {
+        /* Write of schedule protected by barrier write in aio_co_schedule */
+        const char *scheduled = qatomic_cmpxchg(&sleep_state->co->scheduled,
+                                                qemu_co_sleep_ns__scheduled, NULL);
 
-    assert(scheduled == qemu_co_sleep_ns__scheduled);
-    *sleep_state->user_state_pointer = NULL;
-    timer_del(&sleep_state->ts);
-    aio_co_wake(sleep_state->co);
+        assert(scheduled == qemu_co_sleep_ns__scheduled);
+        *sleep_state->user_state_pointer = NULL;
+        timer_del(&sleep_state->ts);
+        aio_co_wake(sleep_state->co);
+    }
 }
 
 static void co_sleep_cb(void *opaque)
 {
-    qemu_co_sleep_wake(opaque);
+    QemuCoSleepState **sleep_state = opaque;
+    qemu_co_sleep_wake(*sleep_state);
 }
 
 void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
         abort();
     }
 
-    aio_timer_init(ctx, &state.ts, type, SCALE_NS, co_sleep_cb, &state);
+    aio_timer_init(ctx, &state.ts, type, SCALE_NS, co_sleep_cb, sleep_state);
     *sleep_state = &state;
     timer_mod(&state.ts, qemu_clock_get_ns(type) + ns);
     qemu_coroutine_yield();
-- 
2.31.1

From: Paolo Bonzini <pbonzini@redhat.com>

This simplification is enabled by the previous patch.  Now aio_co_wake
will only be called once, therefore we do not care about a spurious
firing of the timer after a qemu_co_sleep_wake.

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20210517100548.28806-5-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 util/qemu-coroutine-sleep.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

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 const char *qemu_co_sleep_ns__scheduled = "qemu_co_sleep_ns";
 
 struct QemuCoSleepState {
     Coroutine *co;
-    QEMUTimer ts;
     QemuCoSleepState **user_state_pointer;
 };
 
@@ -XXX,XX +XXX,XX @@ void qemu_co_sleep_wake(QemuCoSleepState *sleep_state)
 
         assert(scheduled == qemu_co_sleep_ns__scheduled);
         *sleep_state->user_state_pointer = NULL;
-        timer_del(&sleep_state->ts);
         aio_co_wake(sleep_state->co);
     }
 }
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
                                             QemuCoSleepState **sleep_state)
 {
     AioContext *ctx = qemu_get_current_aio_context();
+    QEMUTimer ts;
     QemuCoSleepState state = {
         .co = qemu_coroutine_self(),
         .user_state_pointer = sleep_state,
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
         abort();
     }
 
-    aio_timer_init(ctx, &state.ts, type, SCALE_NS, co_sleep_cb, sleep_state);
+    aio_timer_init(ctx, &ts, type, SCALE_NS, co_sleep_cb, sleep_state);
     *sleep_state = &state;
-    timer_mod(&state.ts, qemu_clock_get_ns(type) + ns);
+    timer_mod(&ts, qemu_clock_get_ns(type) + ns);
     qemu_coroutine_yield();
+    timer_del(&ts);
 
     /* qemu_co_sleep_wake clears *sleep_state before resuming this coroutine.  */
     assert(*sleep_state == NULL);
-- 
2.31.1

From: Paolo Bonzini <pbonzini@redhat.com>

Right now, users of qemu_co_sleep_ns_wakeable are simply passing
a pointer to QemuCoSleepState by reference to the function.  But
QemuCoSleepState really is just a Coroutine*; making the
content of the struct public is just as efficient and lets us
skip the user_state_pointer indirection.

Since the usage is changed, take the occasion to rename the
struct to QemuCoSleep.

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20210517100548.28806-6-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 include/qemu/coroutine.h    | 23 +++++++++++----------
 block/block-copy.c          |  8 ++++----
 block/nbd.c                 | 10 ++++-----
 util/qemu-coroutine-sleep.c | 41 ++++++++++++++++---------------------
 4 files changed, 39 insertions(+), 43 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 qemu_co_rwlock_wrlock(CoRwlock *lock);
  */
 void qemu_co_rwlock_unlock(CoRwlock *lock);
 
-typedef struct QemuCoSleepState QemuCoSleepState;
+typedef struct QemuCoSleep {
+    Coroutine *to_wake;
+} QemuCoSleep;
 
 /**
- * Yield the coroutine for a given duration. During this yield, @sleep_state
- * is set to an opaque pointer, which may be used for
- * qemu_co_sleep_wake(). Be careful, the pointer is set back to zero when the
- * timer fires. Don't save the obtained value to other variables and don't call
- * qemu_co_sleep_wake from another aio context.
+ * Yield the coroutine for a given duration. Initializes @w so that,
+ * during this yield, it can be passed to qemu_co_sleep_wake() to
+ * terminate the sleep.
  */
-void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
-                                            QemuCoSleepState **sleep_state);
+void coroutine_fn qemu_co_sleep_ns_wakeable(QemuCoSleep *w,
+                                            QEMUClockType type, int64_t ns);
+
 static inline void coroutine_fn qemu_co_sleep_ns(QEMUClockType type, int64_t ns)
 {
-    QemuCoSleepState *unused = NULL;
-    qemu_co_sleep_ns_wakeable(type, ns, &unused);
+    QemuCoSleep w = { 0 };
+    qemu_co_sleep_ns_wakeable(&w, type, ns);
 }
 
 /**
@@ -XXX,XX +XXX,XX @@ static inline void coroutine_fn qemu_co_sleep_ns(QEMUClockType type, int64_t ns)
  * qemu_co_sleep_ns() and should be checked to be non-NULL before calling
  * qemu_co_sleep_wake().
  */
-void qemu_co_sleep_wake(QemuCoSleepState *sleep_state);
+void qemu_co_sleep_wake(QemuCoSleep *w);
 
 /**
  * Yield until a file descriptor becomes readable
diff --git a/block/block-copy.c b/block/block-copy.c
index XXXXXXX..XXXXXXX 100644
--- a/block/block-copy.c
+++ b/block/block-copy.c
@@ -XXX,XX +XXX,XX @@ typedef struct BlockCopyCallState {
     /* State */
     int ret;
     bool finished;
-    QemuCoSleepState *sleep_state;
+    QemuCoSleep sleep;
     bool cancelled;
 
     /* OUT parameters */
@@ -XXX,XX +XXX,XX @@ block_copy_dirty_clusters(BlockCopyCallState *call_state)
                 if (ns > 0) {
                     block_copy_task_end(task, -EAGAIN);
                     g_free(task);
-                    qemu_co_sleep_ns_wakeable(QEMU_CLOCK_REALTIME, ns,
-                                              &call_state->sleep_state);
+                    qemu_co_sleep_ns_wakeable(&call_state->sleep,
+                                              QEMU_CLOCK_REALTIME, ns);
                     continue;
                 }
             }
@@ -XXX,XX +XXX,XX @@ out:
 
 void block_copy_kick(BlockCopyCallState *call_state)
 {
-    qemu_co_sleep_wake(call_state->sleep_state);
+    qemu_co_sleep_wake(&call_state->sleep);
 }
 
 /*
diff --git a/block/nbd.c b/block/nbd.c
index XXXXXXX..XXXXXXX 100644
--- a/block/nbd.c
+++ b/block/nbd.c
@@ -XXX,XX +XXX,XX @@ typedef struct BDRVNBDState {
     CoQueue free_sema;
     Coroutine *connection_co;
     Coroutine *teardown_co;
-    QemuCoSleepState *connection_co_sleep_ns_state;
+    QemuCoSleep reconnect_sleep;
     bool drained;
     bool wait_drained_end;
     int in_flight;
@@ -XXX,XX +XXX,XX @@ static void coroutine_fn nbd_client_co_drain_begin(BlockDriverState *bs)
     BDRVNBDState *s = (BDRVNBDState *)bs->opaque;
 
     s->drained = true;
-    qemu_co_sleep_wake(s->connection_co_sleep_ns_state);
+    qemu_co_sleep_wake(&s->reconnect_sleep);
 
     nbd_co_establish_connection_cancel(bs, false);
 
@@ -XXX,XX +XXX,XX @@ static void nbd_teardown_connection(BlockDriverState *bs)
 
     s->state = NBD_CLIENT_QUIT;
     if (s->connection_co) {
-        qemu_co_sleep_wake(s->connection_co_sleep_ns_state);
+        qemu_co_sleep_wake(&s->reconnect_sleep);
         nbd_co_establish_connection_cancel(bs, true);
     }
     if (qemu_in_coroutine()) {
@@ -XXX,XX +XXX,XX @@ static coroutine_fn void nbd_co_reconnect_loop(BDRVNBDState *s)
             }
             bdrv_inc_in_flight(s->bs);
         } else {
-            qemu_co_sleep_ns_wakeable(QEMU_CLOCK_REALTIME, timeout,
-                                      &s->connection_co_sleep_ns_state);
+            qemu_co_sleep_ns_wakeable(&s->reconnect_sleep,
+                                      QEMU_CLOCK_REALTIME, timeout);
             if (s->drained) {
                 continue;
             }
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 const char *qemu_co_sleep_ns__scheduled = "qemu_co_sleep_ns";
 
-struct QemuCoSleepState {
+void qemu_co_sleep_wake(QemuCoSleep *w)
+{
     Coroutine *co;
-    QemuCoSleepState **user_state_pointer;
-};
 
-void qemu_co_sleep_wake(QemuCoSleepState *sleep_state)
-{
-    if (sleep_state) {
+    co = w->to_wake;
+    w->to_wake = NULL;
+    if (co) {
         /* Write of schedule protected by barrier write in aio_co_schedule */
-        const char *scheduled = qatomic_cmpxchg(&sleep_state->co->scheduled,
+        const char *scheduled = qatomic_cmpxchg(&co->scheduled,
                                                 qemu_co_sleep_ns__scheduled, NULL);
 
         assert(scheduled == qemu_co_sleep_ns__scheduled);
-        *sleep_state->user_state_pointer = NULL;
-        aio_co_wake(sleep_state->co);
+        aio_co_wake(co);
     }
 }
 
 static void co_sleep_cb(void *opaque)
 {
-    QemuCoSleepState **sleep_state = opaque;
-    qemu_co_sleep_wake(*sleep_state);
+    QemuCoSleep *w = opaque;
+    qemu_co_sleep_wake(w);
 }
 
-void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
-                                            QemuCoSleepState **sleep_state)
+void coroutine_fn qemu_co_sleep_ns_wakeable(QemuCoSleep *w,
+                                            QEMUClockType type, int64_t ns)
 {
+    Coroutine *co = qemu_coroutine_self();
     AioContext *ctx = qemu_get_current_aio_context();
     QEMUTimer ts;
-    QemuCoSleepState state = {
-        .co = qemu_coroutine_self(),
-        .user_state_pointer = sleep_state,
-    };
 
-    const char *scheduled = qatomic_cmpxchg(&state.co->scheduled, NULL,
-                                           qemu_co_sleep_ns__scheduled);
+    const char *scheduled = qatomic_cmpxchg(&co->scheduled, NULL,
+                                            qemu_co_sleep_ns__scheduled);
     if (scheduled) {
         fprintf(stderr,
                 "%s: Co-routine was already scheduled in '%s'\n",
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QEMUClockType type, int64_t ns,
         abort();
     }
 
-    aio_timer_init(ctx, &ts, type, SCALE_NS, co_sleep_cb, sleep_state);
-    *sleep_state = &state;
+    w->to_wake = co;
+    aio_timer_init(ctx, &ts, type, SCALE_NS, co_sleep_cb, w),
     timer_mod(&ts, qemu_clock_get_ns(type) + ns);
     qemu_coroutine_yield();
     timer_del(&ts);
 
-    /* qemu_co_sleep_wake clears *sleep_state before resuming this coroutine.  */
-    assert(*sleep_state == NULL);
+    /* w->to_wake is cleared before resuming this coroutine.  */
+    assert(w->to_wake == NULL);
 }
-- 
2.31.1

From: Paolo Bonzini <pbonzini@redhat.com>

Allow using QemuCoSleep to sleep forever until woken by qemu_co_sleep_wake.
This makes the logic of qemu_co_sleep_ns_wakeable easy to understand.

In the future we will introduce an API that can work even if the
sleep and wake happen from different threads.  For now, initializing
w->to_wake after timer_mod is fine because the timer can only fire in
the same AioContext.

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20210517100548.28806-7-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 include/qemu/coroutine.h    |  5 +++++
 util/qemu-coroutine-sleep.c | 26 +++++++++++++++++++-------
 2 files changed, 24 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 QemuCoSleep {
 void coroutine_fn qemu_co_sleep_ns_wakeable(QemuCoSleep *w,
                                             QEMUClockType type, int64_t ns);
 
+/**
+ * Yield the coroutine until the next call to qemu_co_sleep_wake.
+ */
+void coroutine_fn qemu_co_sleep(QemuCoSleep *w);
+
 static inline void coroutine_fn qemu_co_sleep_ns(QEMUClockType type, int64_t ns)
 {
     QemuCoSleep w = { 0 };
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)
     qemu_co_sleep_wake(w);
 }
 
-void coroutine_fn qemu_co_sleep_ns_wakeable(QemuCoSleep *w,
-                                            QEMUClockType type, int64_t ns)
+void coroutine_fn qemu_co_sleep(QemuCoSleep *w)
 {
     Coroutine *co = qemu_coroutine_self();
-    AioContext *ctx = qemu_get_current_aio_context();
-    QEMUTimer ts;
 
     const char *scheduled = qatomic_cmpxchg(&co->scheduled, NULL,
                                             qemu_co_sleep_ns__scheduled);
@@ -XXX,XX +XXX,XX @@ void coroutine_fn qemu_co_sleep_ns_wakeable(QemuCoSleep *w,
     }
 
     w->to_wake = co;
-    aio_timer_init(ctx, &ts, type, SCALE_NS, co_sleep_cb, w),
-    timer_mod(&ts, qemu_clock_get_ns(type) + ns);
     qemu_coroutine_yield();
-    timer_del(&ts);
 
     /* w->to_wake is cleared before resuming this coroutine.  */
     assert(w->to_wake == NULL);
 }
+
+void coroutine_fn qemu_co_sleep_ns_wakeable(QemuCoSleep *w,
+                                            QEMUClockType type, int64_t ns)
+{
+    AioContext *ctx = qemu_get_current_aio_context();
+    QEMUTimer ts;
+
+    aio_timer_init(ctx, &ts, type, SCALE_NS, co_sleep_cb, w);
+    timer_mod(&ts, qemu_clock_get_ns(type) + ns);
+
+    /*
+     * The timer will fire in the current AiOContext, so the callback
+     * must happen after qemu_co_sleep yields and there is no race
+     * between timer_mod and qemu_co_sleep.
+     */
+    qemu_co_sleep(w);
+    timer_del(&ts);
+}
-- 
2.31.1