Series comparison

-[PULL 00/11] Net patches
+[PULL V3 00/15] Net patches
-The following changes since commit 13356edb87506c148b163b8c7eb0695647d00c2a:
+The following changes since commit 352998df1c53b366413690d95b35f76d0721ebed:
-  Merge tag 'block-pull-request' of https://gitlab.com/stefanha/qemu into staging (2023-01-24 09:45:33 +0000)
+  Merge tag 'i2c-20220314' of https://github.com/philmd/qemu into staging (2022-03-14 14:39:33 +0000)
 are available in the git repository at:
   https://github.com/jasowang/qemu.git tags/net-pull-request
-for you to fetch changes up to 2bd492bca521ee8594f1d5db8dc9aac126fc4f85:
+for you to fetch changes up to 12a195fa343aae2ead1301ce04727bd0ae25eb15:
-  vdpa: fix VHOST_BACKEND_F_IOTLB_ASID flag check (2023-02-02 14:16:48 +0800)
+  vdpa: Expose VHOST_F_LOG_ALL on SVQ (2022-03-15 13:57:44 +0800)
 ----------------------------------------------------------------
+Changes since V2:
+- fix 32bit build errros
 ----------------------------------------------------------------
-Christian Svensson (1):
+Eugenio Pérez (14):
-      net: Increase L2TPv3 buffer to fit jumboframes
+      vhost: Add VhostShadowVirtqueue
       vhost: Add Shadow VirtQueue kick forwarding capabilities
       vhost: Add Shadow VirtQueue call forwarding capabilities
       vhost: Add vhost_svq_valid_features to shadow vq
       virtio: Add vhost_svq_get_vring_addr
       vdpa: adapt vhost_ops callbacks to svq
       vhost: Shadow virtqueue buffers forwarding
       util: Add iova_tree_alloc_map
       util: add iova_tree_find_iova
       vhost: Add VhostIOVATree
       vdpa: Add custom IOTLB translations to SVQ
       vdpa: Adapt vhost_vdpa_get_vring_base to SVQ
       vdpa: Never set log_base addr if SVQ is enabled
       vdpa: Expose VHOST_F_LOG_ALL on SVQ
-Eugenio Pérez (1):
+Jason Wang (1):
-      vdpa: fix VHOST_BACKEND_F_IOTLB_ASID flag check
+      virtio-net: fix map leaking on error during receive
-Fiona Ebner (1):
+ hw/net/virtio-net.c                |   1 +
-      hw/net/vmxnet3: allow VMXNET3_MAX_MTU itself as a value
+ hw/virtio/meson.build              |   2 +-
+ hw/virtio/vhost-iova-tree.c        | 110 +++++++
-Joelle van Dyne (1):
+ hw/virtio/vhost-iova-tree.h        |  27 ++
-      vmnet: stop recieving events when VM is stopped
+ hw/virtio/vhost-shadow-virtqueue.c | 636 +++++++++++++++++++++++++++++++++++++
+ hw/virtio/vhost-shadow-virtqueue.h |  87 +++++
-Laurent Vivier (2):
+ hw/virtio/vhost-vdpa.c             | 522 +++++++++++++++++++++++++++++-
-      tests/qtest: netdev: test stream and dgram backends
+ include/hw/virtio/vhost-vdpa.h     |   8 +
-      net: stream: add a new option to automatically reconnect
+ include/qemu/iova-tree.h           |  38 ++-
+ util/iova-tree.c                   | 170 ++++++++++
-Qiang Liu (2):
+files changed, 1584 insertions(+), 17 deletions(-)
-      hw/net/lan9118: log [read|write]b when mode_16bit is enabled rather than abort
+ create mode 100644 hw/virtio/vhost-iova-tree.c
-      hw/net/can/xlnx-zynqmp-can: fix assertion failures in transfer_fifo()
+ create mode 100644 hw/virtio/vhost-iova-tree.h
+ create mode 100644 hw/virtio/vhost-shadow-virtqueue.c
-Thomas Huth (3):
+ create mode 100644 hw/virtio/vhost-shadow-virtqueue.h
       net: Move the code to collect available NIC models to a separate function
       net: Restore printing of the help text with "-nic help"
       net: Replace "Supported NIC models" with "Available NIC models"
  hw/net/can/xlnx-zynqmp-can.c |   9 +-
  hw/net/lan9118.c             |  17 +-
  hw/net/vmxnet3.c             |   2 +-
  hw/pci/pci.c                 |  29 +--
  include/net/net.h            |  14 ++
  net/l2tpv3.c                 |   2 +-
  net/net.c                    |  50 +++-
  net/stream.c                 |  53 ++++-
  net/vhost-vdpa.c             |   2 +-
  net/vmnet-common.m           |  48 +++-
  net/vmnet_int.h              |   2 +
  qapi/net.json                |   7 +-
  qemu-options.hx              |   6 +-
  tests/qtest/meson.build      |   2 +
  tests/qtest/netdev-socket.c  | 549 +++++++++++++++++++++++++++++++++++++++++++
 files changed, 730 insertions(+), 62 deletions(-)
  create mode 100644 tests/qtest/netdev-socket.c

-[PULL 08/11] hw/net/can/xlnx-zynqmp-can: fix assertion failures in transfer_fifo()
+[PULL V3 01/15] virtio-net: fix map leaking on error during receive
-From: Qiang Liu <cyruscyliu@gmail.com>
+Commit bedd7e93d0196 ("virtio-net: fix use after unmap/free for sg")
 tries to fix the use after free of the sg by caching the virtqueue
 elements in an array and unmap them at once after receiving the
 packets, But it forgot to unmap the cached elements on error which
 will lead to leaking of mapping and other unexpected results.
-Check fifos before poping data from and pushing data into it.
+Fixing this by detaching the cached elements on error. This addresses
 CVE-2022-26353.
-Fixes: 98e5d7a2b726 ("hw/net/can: Introduce Xilinx ZynqMP CAN controller")
+Reported-by: Victor Tom <vv474172261@gmail.com>
-Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1425
+Cc: qemu-stable@nongnu.org
-Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1427
+Fixes: CVE-2022-26353
-Reported-by: Qiang Liu <cyruscyliu@gmail.com>
+Fixes: bedd7e93d0196 ("virtio-net: fix use after unmap/free for sg")
-Signed-off-by: Qiang Liu <cyruscyliu@gmail.com>
+Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
 Signed-off-by: Jason Wang <jasowang@redhat.com>
 ---
- hw/net/can/xlnx-zynqmp-can.c | 9 ++++++++-
+ hw/net/virtio-net.c | 1 +
-file changed, 8 insertions(+), 1 deletion(-)
+file changed, 1 insertion(+)
-diff --git a/hw/net/can/xlnx-zynqmp-can.c b/hw/net/can/xlnx-zynqmp-can.c
+diff --git a/hw/net/virtio-net.c b/hw/net/virtio-net.c
 index XXXXXXX..XXXXXXX 100644
---- a/hw/net/can/xlnx-zynqmp-can.c
+--- a/hw/net/virtio-net.c
-+++ b/hw/net/can/xlnx-zynqmp-can.c
++++ b/hw/net/virtio-net.c
-@@ -XXX,XX +XXX,XX @@ static void transfer_fifo(XlnxZynqMPCANState *s, Fifo32 *fifo)
+@@ -XXX,XX +XXX,XX @@ static ssize_t virtio_net_receive_rcu(NetClientState *nc, const uint8_t *buf,
  err:
      for (j = 0; j < i; j++) {
 +        virtqueue_detach_element(q->rx_vq, elems[j], lens[j]);
          g_free(elems[j]);
      }
-     while (!fifo32_is_empty(fifo)) {
-+        if (fifo32_num_used(fifo) < (4 * CAN_FRAME_SIZE)) {
-+            g_autofree char *path = object_get_canonical_path(OBJECT(s));
-+            qemu_log_mask(LOG_GUEST_ERROR, "%s: data left in the fifo is not"
-+                          " enough for transfer.\n", path);
-+            break;
-+        }
-         for (i = 0; i < CAN_FRAME_SIZE; i++) {
-             data[i] = fifo32_pop(fifo);
-         }
-@@ -XXX,XX +XXX,XX @@ static void transfer_fifo(XlnxZynqMPCANState *s, Fifo32 *fifo)
-              * acknowledged. The XlnxZynqMPCAN core receives any message
-              * that it transmits.
-              */
--            if (fifo32_is_full(&s->rx_fifo)) {
-+            if (fifo32_is_full(&s->rx_fifo) ||
-+                    (fifo32_num_free(&s->rx_fifo) < (4 * CAN_FRAME_SIZE))) {
-                 ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, RXOFLW, 1);
-             } else {
-                 for (i = 0; i < CAN_FRAME_SIZE; i++) {
 --
 .7.4

-[PULL 09/11] tests/qtest: netdev: test stream and dgram backends
+[PULL V3 02/15] vhost: Add VhostShadowVirtqueue
-From: Laurent Vivier <lvivier@redhat.com>
+From: Eugenio Pérez <eperezma@redhat.com>
-Signed-off-by: Laurent Vivier <lvivier@redhat.com>
+Vhost shadow virtqueue (SVQ) is an intermediate jump for virtqueue
 notifications and buffers, allowing qemu to track them. While qemu is
 forwarding the buffers and virtqueue changes, it is able to commit the
 memory it's being dirtied, the same way regular qemu's VirtIO devices
 do.
 This commit only exposes basic SVQ allocation and free. Next patches of
 the series add functionality like notifications and buffers forwarding.
 Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
 Acked-by: Michael S. Tsirkin <mst@redhat.com>
-Acked-by: Thomas Huth <thuth@redhat.com>
 Signed-off-by: Jason Wang <jasowang@redhat.com>
 ---
- tests/qtest/meson.build     |   2 +
+ hw/virtio/meson.build              |  2 +-
- tests/qtest/netdev-socket.c | 448 ++++++++++++++++++++++++++++++++++++++++++++
+ hw/virtio/vhost-shadow-virtqueue.c | 62 ++++++++++++++++++++++++++++++++++++++
-files changed, 450 insertions(+)
+ hw/virtio/vhost-shadow-virtqueue.h | 28 +++++++++++++++++
- create mode 100644 tests/qtest/netdev-socket.c
+files changed, 91 insertions(+), 1 deletion(-)
  create mode 100644 hw/virtio/vhost-shadow-virtqueue.c
  create mode 100644 hw/virtio/vhost-shadow-virtqueue.h
-diff --git a/tests/qtest/meson.build b/tests/qtest/meson.build
+diff --git a/hw/virtio/meson.build b/hw/virtio/meson.build
 index XXXXXXX..XXXXXXX 100644
---- a/tests/qtest/meson.build
+--- a/hw/virtio/meson.build
-+++ b/tests/qtest/meson.build
++++ b/hw/virtio/meson.build
-@@ -XXX,XX +XXX,XX @@ qtests_generic = [
+@@ -XXX,XX +XXX,XX @@ softmmu_ss.add(when: 'CONFIG_ALL', if_true: files('vhost-stub.c'))
-   'test-hmp',
-   'qos-test',
+ virtio_ss = ss.source_set()
-   'readconfig-test',
+ virtio_ss.add(files('virtio.c'))
-+  'netdev-socket',
+-virtio_ss.add(when: 'CONFIG_VHOST', if_true: files('vhost.c', 'vhost-backend.c'))
- ]
++virtio_ss.add(when: 'CONFIG_VHOST', if_true: files('vhost.c', 'vhost-backend.c', 'vhost-shadow-virtqueue.c'))
- if config_host.has_key('CONFIG_MODULES')
+ virtio_ss.add(when: 'CONFIG_VHOST_USER', if_true: files('vhost-user.c'))
-   qtests_generic += [ 'modules-test' ]
+ virtio_ss.add(when: 'CONFIG_VHOST_VDPA', if_true: files('vhost-vdpa.c'))
-@@ -XXX,XX +XXX,XX @@ qtests = {
+ virtio_ss.add(when: 'CONFIG_VIRTIO_BALLOON', if_true: files('virtio-balloon.c'))
-   'tpm-tis-device-swtpm-test': [io, tpmemu_files, 'tpm-tis-util.c'],
+diff --git a/hw/virtio/vhost-shadow-virtqueue.c b/hw/virtio/vhost-shadow-virtqueue.c
    'tpm-tis-device-test': [io, tpmemu_files, 'tpm-tis-util.c'],
    'vmgenid-test': files('boot-sector.c', 'acpi-utils.c'),
 +  'netdev-socket': files('netdev-socket.c', '../unit/socket-helpers.c'),
  }
  gvnc = dependency('gvnc-1.0', required: false)
 diff --git a/tests/qtest/netdev-socket.c b/tests/qtest/netdev-socket.c
 new file mode 100644
 index XXXXXXX..XXXXXXX
 --- /dev/null
-+++ b/tests/qtest/netdev-socket.c
++++ b/hw/virtio/vhost-shadow-virtqueue.c
 @@ -XXX,XX +XXX,XX @@
 +/*
-+ * QTest testcase for netdev stream and dgram
++ * vhost shadow virtqueue
 + *
-+ * Copyright (c) 2022 Red Hat, Inc.
++ * SPDX-FileCopyrightText: Red Hat, Inc. 2021
 + * SPDX-FileContributor: Author: Eugenio Pérez <eperezma@redhat.com>
 + *
 + * SPDX-License-Identifier: GPL-2.0-or-later
 + */
 +
 +#include "qemu/osdep.h"
-+#include "qemu/sockets.h"
++#include "hw/virtio/vhost-shadow-virtqueue.h"
 +#include <glib/gstdio.h>
 +#include "../unit/socket-helpers.h"
 +#include "libqtest.h"
 +
-+#define CONNECTION_TIMEOUT    5
++#include "qemu/error-report.h"
 +
-+#define EXPECT_STATE(q, e, t)                             \
++/**
-+do {                                                      \
++ * Creates vhost shadow virtqueue, and instructs the vhost device to use the
-+    char *resp = NULL;                                    \
++ * shadow methods and file descriptors.
-+    g_test_timer_start();                                 \
++ *
-+    do {                                                  \
++ * Returns the new virtqueue or NULL.
-+        g_free(resp);                                     \
++ *
-+        resp = qtest_hmp(q, "info network");              \
++ * In case of error, reason is reported through error_report.
-+        if (t) {                                          \
++ */
-+            strrchr(resp, t)[0] = 0;                      \
++VhostShadowVirtqueue *vhost_svq_new(void)
-+        }                                                 \
++{
-+        if (g_str_equal(resp, e)) {                       \
++    g_autofree VhostShadowVirtqueue *svq = g_new0(VhostShadowVirtqueue, 1);
-+            break;                                        \
++    int r;
 +        }                                                 \
 +    } while (g_test_timer_elapsed() < CONNECTION_TIMEOUT); \
 +    g_assert_cmpstr(resp, ==, e);                         \
 +    g_free(resp);                                         \
 +} while (0)
 +
-+static gchar *tmpdir;
++    r = event_notifier_init(&svq->hdev_kick, 0);
-+
++    if (r != 0) {
-+static int inet_get_free_port_socket_ipv4(int sock)
++        error_report("Couldn't create kick event notifier: %s (%d)",
-+{
++                     g_strerror(errno), errno);
-+    struct sockaddr_in addr;
++        goto err_init_hdev_kick;
 +    socklen_t len;
 +
 +    memset(&addr, 0, sizeof(addr));
 +    addr.sin_family = AF_INET;
 +    addr.sin_addr.s_addr = INADDR_ANY;
 +    addr.sin_port = 0;
 +    if (bind(sock, (struct sockaddr *)&addr, sizeof(addr)) < 0) {
 +        return -1;
 +    }
 +
-+    len = sizeof(addr);
++    r = event_notifier_init(&svq->hdev_call, 0);
-+    if (getsockname(sock,  (struct sockaddr *)&addr, &len) < 0) {
++    if (r != 0) {
-+        return -1;
++        error_report("Couldn't create call event notifier: %s (%d)",
 +                     g_strerror(errno), errno);
 +        goto err_init_hdev_call;
 +    }
 +
-+    return ntohs(addr.sin_port);
++    return g_steal_pointer(&svq);
 +
 +err_init_hdev_call:
 +    event_notifier_cleanup(&svq->hdev_kick);
 +
 +err_init_hdev_kick:
 +    return NULL;
 +}
 +
-+static int inet_get_free_port_socket_ipv6(int sock)
++/**
 + * Free the resources of the shadow virtqueue.
 + *
 + * @pvq: gpointer to SVQ so it can be used by autofree functions.
 + */
 +void vhost_svq_free(gpointer pvq)
 +{
-+    struct sockaddr_in6 addr;
++    VhostShadowVirtqueue *vq = pvq;
-+    socklen_t len;
++    event_notifier_cleanup(&vq->hdev_kick);
 +    event_notifier_cleanup(&vq->hdev_call);
 +    g_free(vq);
 +}
 diff --git a/hw/virtio/vhost-shadow-virtqueue.h b/hw/virtio/vhost-shadow-virtqueue.h
 new file mode 100644
 index XXXXXXX..XXXXXXX
 --- /dev/null
 +++ b/hw/virtio/vhost-shadow-virtqueue.h
@@ -XXX,XX +XXX,XX @@
 +/*
 + * vhost shadow virtqueue
 + *
 + * SPDX-FileCopyrightText: Red Hat, Inc. 2021
 + * SPDX-FileContributor: Author: Eugenio Pérez <eperezma@redhat.com>
 + *
 + * SPDX-License-Identifier: GPL-2.0-or-later
 + */
 +
-+    memset(&addr, 0, sizeof(addr));
++#ifndef VHOST_SHADOW_VIRTQUEUE_H
-+    addr.sin6_family = AF_INET6;
++#define VHOST_SHADOW_VIRTQUEUE_H
 +    addr.sin6_addr = in6addr_any;
 +    addr.sin6_port = 0;
 +    if (bind(sock, (struct sockaddr *)&addr, sizeof(addr)) < 0) {
 +        return -1;
 +    }
 +
-+    len = sizeof(addr);
++#include "qemu/event_notifier.h"
 +    if (getsockname(sock,  (struct sockaddr *)&addr, &len) < 0) {
 +        return -1;
 +    }
 +
-+    return ntohs(addr.sin6_port);
++/* Shadow virtqueue to relay notifications */
-+}
++typedef struct VhostShadowVirtqueue {
 +    /* Shadow kick notifier, sent to vhost */
 +    EventNotifier hdev_kick;
 +    /* Shadow call notifier, sent to vhost */
 +    EventNotifier hdev_call;
 +} VhostShadowVirtqueue;
 +
-+static int inet_get_free_port_multiple(int nb, int *port, bool ipv6)
++VhostShadowVirtqueue *vhost_svq_new(void);
 +{
 +    int sock[nb];
 +    int i;
 +
-+    for (i = 0; i < nb; i++) {
++void vhost_svq_free(gpointer vq);
-+        sock[i] = socket(ipv6 ? AF_INET6 : AF_INET, SOCK_STREAM, 0);
++G_DEFINE_AUTOPTR_CLEANUP_FUNC(VhostShadowVirtqueue, vhost_svq_free);
 +        if (sock[i] < 0) {
 +            break;
 +        }
 +        port[i] = ipv6 ? inet_get_free_port_socket_ipv6(sock[i]) :
 +                         inet_get_free_port_socket_ipv4(sock[i]);
 +        if (port[i] == -1) {
 +            break;
 +        }
 +    }
 +
-+    nb = i;
-+    for (i = 0; i < nb; i++) {
-+        closesocket(sock[i]);
-+    }
-+
-+    return nb;
-+}
-+
-+static int inet_get_free_port(bool ipv6)
-+{
-+    int nb, port;
-+
-+    nb = inet_get_free_port_multiple(1, &port, ipv6);
-+    g_assert_cmpint(nb, ==, 1);
-+
-+    return port;
-+}
-+
-+static void test_stream_inet_ipv4(void)
-+{
-+    QTestState *qts0, *qts1;
-+    char *expect;
-+    int port;
-+
-+    port = inet_get_free_port(false);
-+    qts0 = qtest_initf("-nodefaults -M none "
-+                       "-netdev stream,id=st0,server=true,addr.type=inet,"
-+                       "addr.ipv4=on,addr.ipv6=off,"
-+                       "addr.host=127.0.0.1,addr.port=%d", port);
-+
-+    EXPECT_STATE(qts0, "st0: index=0,type=stream,\r\n", 0);
-+
-+    qts1 = qtest_initf("-nodefaults -M none "
-+                       "-netdev stream,server=false,id=st0,addr.type=inet,"
-+                       "addr.ipv4=on,addr.ipv6=off,"
-+                       "addr.host=127.0.0.1,addr.port=%d", port);
-+
-+    expect = g_strdup_printf("st0: index=0,type=stream,tcp:127.0.0.1:%d\r\n",
-+                             port);
-+    EXPECT_STATE(qts1, expect, 0);
-+    g_free(expect);
-+
-+    /* the port is unknown, check only the address */
-+    EXPECT_STATE(qts0, "st0: index=0,type=stream,tcp:127.0.0.1", ':');
-+
-+    qtest_quit(qts1);
-+    qtest_quit(qts0);
-+}
-+
-+static void test_stream_inet_ipv6(void)
-+{
-+    QTestState *qts0, *qts1;
-+    char *expect;
-+    int port;
-+
-+    port = inet_get_free_port(true);
-+    qts0 = qtest_initf("-nodefaults -M none "
-+                       "-netdev stream,id=st0,server=true,addr.type=inet,"
-+                       "addr.ipv4=off,addr.ipv6=on,"
-+                       "addr.host=::1,addr.port=%d", port);
-+
-+    EXPECT_STATE(qts0, "st0: index=0,type=stream,\r\n", 0);
-+
-+    qts1 = qtest_initf("-nodefaults -M none "
-+                       "-netdev stream,server=false,id=st0,addr.type=inet,"
-+                       "addr.ipv4=off,addr.ipv6=on,"
-+                       "addr.host=::1,addr.port=%d", port);
-+
-+    expect = g_strdup_printf("st0: index=0,type=stream,tcp:::1:%d\r\n",
-+                             port);
-+    EXPECT_STATE(qts1, expect, 0);
-+    g_free(expect);
-+
-+    /* the port is unknown, check only the address */
-+    EXPECT_STATE(qts0, "st0: index=0,type=stream,tcp:::1", ':');
-+
-+    qtest_quit(qts1);
-+    qtest_quit(qts0);
-+}
-+
-+static void test_stream_unix(void)
-+{
-+    QTestState *qts0, *qts1;
-+    char *expect;
-+    gchar *path;
-+
-+    path = g_strconcat(tmpdir, "/stream_unix", NULL);
-+
-+    qts0 = qtest_initf("-nodefaults -M none "
-+                       "-netdev stream,id=st0,server=true,"
-+                       "addr.type=unix,addr.path=%s,",
-+                       path);
-+
-+    EXPECT_STATE(qts0, "st0: index=0,type=stream,\r\n", 0);
-+
-+    qts1 = qtest_initf("-nodefaults -M none "
-+                       "-netdev stream,id=st0,server=false,"
-+                       "addr.type=unix,addr.path=%s",
-+                       path);
-+
-+    expect = g_strdup_printf("st0: index=0,type=stream,unix:%s\r\n", path);
-+    EXPECT_STATE(qts1, expect, 0);
-+    EXPECT_STATE(qts0, expect, 0);
-+    g_free(expect);
-+    g_free(path);
-+
-+    qtest_quit(qts1);
-+    qtest_quit(qts0);
-+}
-+
-+#ifdef CONFIG_LINUX
-+static void test_stream_unix_abstract(void)
-+{
-+    QTestState *qts0, *qts1;
-+    char *expect;
-+    gchar *path;
-+
-+    path = g_strconcat(tmpdir, "/stream_unix_abstract", NULL);
-+
-+    qts0 = qtest_initf("-nodefaults -M none "
-+                       "-netdev stream,id=st0,server=true,"
-+                       "addr.type=unix,addr.path=%s,"
-+                       "addr.abstract=on",
-+                       path);
-+
-+    EXPECT_STATE(qts0, "st0: index=0,type=stream,\r\n", 0);
-+
-+    qts1 = qtest_initf("-nodefaults -M none "
-+                       "-netdev stream,id=st0,server=false,"
-+                       "addr.type=unix,addr.path=%s,addr.abstract=on",
-+                       path);
-+
-+    expect = g_strdup_printf("st0: index=0,type=stream,unix:%s\r\n", path);
-+    EXPECT_STATE(qts1, expect, 0);
-+    EXPECT_STATE(qts0, expect, 0);
-+    g_free(expect);
-+    g_free(path);
-+
-+    qtest_quit(qts1);
-+    qtest_quit(qts0);
-+}
 +#endif
-+
-+#ifndef _WIN32
-+static void test_stream_fd(void)
-+{
-+    QTestState *qts0, *qts1;
-+    int sock[2];
-+    int ret;
-+
-+    ret = socketpair(AF_LOCAL, SOCK_STREAM, 0, sock);
-+    g_assert_true(ret == 0);
-+
-+    qts0 = qtest_initf("-nodefaults -M none "
-+                       "-netdev stream,id=st0,addr.type=fd,addr.str=%d",
-+                       sock[0]);
-+
-+    EXPECT_STATE(qts0, "st0: index=0,type=stream,unix:\r\n", 0);
-+
-+    qts1 = qtest_initf("-nodefaults -M none "
-+                       "-netdev stream,id=st0,addr.type=fd,addr.str=%d",
-+                       sock[1]);
-+
-+    EXPECT_STATE(qts1, "st0: index=0,type=stream,unix:\r\n", 0);
-+    EXPECT_STATE(qts0, "st0: index=0,type=stream,unix:\r\n", 0);
-+
-+    qtest_quit(qts1);
-+    qtest_quit(qts0);
-+
-+    closesocket(sock[0]);
-+    closesocket(sock[1]);
-+}
-+#endif
-+
-+static void test_dgram_inet(void)
-+{
-+    QTestState *qts0, *qts1;
-+    char *expect;
-+    int port[2];
-+    int nb;
-+
-+    nb = inet_get_free_port_multiple(2, port, false);
-+    g_assert_cmpint(nb, ==, 2);
-+
-+    qts0 = qtest_initf("-nodefaults -M none "
-+                       "-netdev dgram,id=st0,"
-+                       "local.type=inet,local.host=127.0.0.1,local.port=%d,"
-+                       "remote.type=inet,remote.host=127.0.0.1,remote.port=%d",
-+                        port[0], port[1]);
-+
-+    expect = g_strdup_printf("st0: index=0,type=dgram,"
-+                             "udp=127.0.0.1:%d/127.0.0.1:%d\r\n",
-+                             port[0], port[1]);
-+    EXPECT_STATE(qts0, expect, 0);
-+    g_free(expect);
-+
-+    qts1 = qtest_initf("-nodefaults -M none "
-+                       "-netdev dgram,id=st0,"
-+                       "local.type=inet,local.host=127.0.0.1,local.port=%d,"
-+                       "remote.type=inet,remote.host=127.0.0.1,remote.port=%d",
-+                        port[1], port[0]);
-+
-+    expect = g_strdup_printf("st0: index=0,type=dgram,"
-+                             "udp=127.0.0.1:%d/127.0.0.1:%d\r\n",
-+                             port[1], port[0]);
-+    EXPECT_STATE(qts1, expect, 0);
-+    g_free(expect);
-+
-+    qtest_quit(qts1);
-+    qtest_quit(qts0);
-+}
-+
-+#ifndef _WIN32
-+static void test_dgram_mcast(void)
-+{
-+    QTestState *qts;
-+
-+    qts = qtest_initf("-nodefaults -M none "
-+                      "-netdev dgram,id=st0,"
-+                      "remote.type=inet,remote.host=230.0.0.1,remote.port=1234");
-+
-+    EXPECT_STATE(qts, "st0: index=0,type=dgram,mcast=230.0.0.1:1234\r\n", 0);
-+
-+    qtest_quit(qts);
-+}
-+
-+static void test_dgram_unix(void)
-+{
-+    QTestState *qts0, *qts1;
-+    char *expect;
-+    gchar *path0, *path1;
-+
-+    path0 = g_strconcat(tmpdir, "/dgram_unix0", NULL);
-+    path1 = g_strconcat(tmpdir, "/dgram_unix1", NULL);
-+
-+    qts0 = qtest_initf("-nodefaults -M none "
-+                       "-netdev dgram,id=st0,local.type=unix,local.path=%s,"
-+                       "remote.type=unix,remote.path=%s",
-+                       path0, path1);
-+
-+    expect = g_strdup_printf("st0: index=0,type=dgram,udp=%s:%s\r\n",
-+                             path0, path1);
-+    EXPECT_STATE(qts0, expect, 0);
-+    g_free(expect);
-+
-+    qts1 = qtest_initf("-nodefaults -M none "
-+                       "-netdev dgram,id=st0,local.type=unix,local.path=%s,"
-+                       "remote.type=unix,remote.path=%s",
-+                       path1, path0);
-+
-+
-+    expect = g_strdup_printf("st0: index=0,type=dgram,udp=%s:%s\r\n",
-+                             path1, path0);
-+    EXPECT_STATE(qts1, expect, 0);
-+    g_free(expect);
-+
-+    unlink(path0);
-+    g_free(path0);
-+    unlink(path1);
-+    g_free(path1);
-+
-+    qtest_quit(qts1);
-+    qtest_quit(qts0);
-+}
-+
-+static void test_dgram_fd(void)
-+{
-+    QTestState *qts0, *qts1;
-+    char *expect;
-+    int ret;
-+    int sv[2];
-+
-+    ret = socketpair(PF_UNIX, SOCK_DGRAM, 0, sv);
-+    g_assert_cmpint(ret, !=, -1);
-+
-+    qts0 = qtest_initf("-nodefaults -M none "
-+                       "-netdev dgram,id=st0,local.type=fd,local.str=%d",
-+                       sv[0]);
-+
-+    expect = g_strdup_printf("st0: index=0,type=dgram,fd=%d unix\r\n", sv[0]);
-+    EXPECT_STATE(qts0, expect, 0);
-+    g_free(expect);
-+
-+    qts1 = qtest_initf("-nodefaults -M none "
-+                       "-netdev dgram,id=st0,local.type=fd,local.str=%d",
-+                       sv[1]);
-+
-+
-+    expect = g_strdup_printf("st0: index=0,type=dgram,fd=%d unix\r\n", sv[1]);
-+    EXPECT_STATE(qts1, expect, 0);
-+    g_free(expect);
-+
-+    qtest_quit(qts1);
-+    qtest_quit(qts0);
-+
-+    closesocket(sv[0]);
-+    closesocket(sv[1]);
-+}
-+#endif
-+
-+int main(int argc, char **argv)
-+{
-+    int ret;
-+    bool has_ipv4, has_ipv6, has_afunix;
-+    g_autoptr(GError) err = NULL;
-+
-+    socket_init();
-+    g_test_init(&argc, &argv, NULL);
-+
-+    if (socket_check_protocol_support(&has_ipv4, &has_ipv6) < 0) {
-+        g_error("socket_check_protocol_support() failed\n");
-+    }
-+
-+    tmpdir = g_dir_make_tmp("netdev-socket.XXXXXX", &err);
-+    if (tmpdir == NULL) {
-+        g_error("Can't create temporary directory in %s: %s",
-+                g_get_tmp_dir(), err->message);
-+    }
-+
-+    if (has_ipv4) {
-+        qtest_add_func("/netdev/stream/inet/ipv4", test_stream_inet_ipv4);
-+        qtest_add_func("/netdev/dgram/inet", test_dgram_inet);
-+#ifndef _WIN32
-+        qtest_add_func("/netdev/dgram/mcast", test_dgram_mcast);
-+#endif
-+    }
-+    if (has_ipv6) {
-+        qtest_add_func("/netdev/stream/inet/ipv6", test_stream_inet_ipv6);
-+    }
-+
-+    socket_check_afunix_support(&has_afunix);
-+    if (has_afunix) {
-+#ifndef _WIN32
-+        qtest_add_func("/netdev/dgram/unix", test_dgram_unix);
-+#endif
-+        qtest_add_func("/netdev/stream/unix", test_stream_unix);
-+#ifdef CONFIG_LINUX
-+        qtest_add_func("/netdev/stream/unix/abstract",
-+                       test_stream_unix_abstract);
-+#endif
-+#ifndef _WIN32
-+        qtest_add_func("/netdev/stream/fd", test_stream_fd);
-+        qtest_add_func("/netdev/dgram/fd", test_dgram_fd);
-+#endif
-+    }
-+
-+    ret = g_test_run();
-+
-+    g_rmdir(tmpdir);
-+    g_free(tmpdir);
-+
-+    return ret;
-+}
 --
 .7.4

-New patch
+[PULL V3 03/15] vhost: Add Shadow VirtQueue kick forwarding capabilities
+From: Eugenio Pérez <eperezma@redhat.com>
 At this mode no buffer forwarding will be performed in SVQ mode: Qemu
 will just forward the guest's kicks to the device.
 Host memory notifiers regions are left out for simplicity, and they will
 not be addressed in this series.
 Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
 Acked-by: Michael S. Tsirkin <mst@redhat.com>
 Signed-off-by: Jason Wang <jasowang@redhat.com>
 ---
  hw/virtio/vhost-shadow-virtqueue.c |  55 ++++++++++++++
  hw/virtio/vhost-shadow-virtqueue.h |  14 ++++
  hw/virtio/vhost-vdpa.c             | 144 ++++++++++++++++++++++++++++++++++++-
  include/hw/virtio/vhost-vdpa.h     |   4 ++
 files changed, 215 insertions(+), 2 deletions(-)
 diff --git a/hw/virtio/vhost-shadow-virtqueue.c b/hw/virtio/vhost-shadow-virtqueue.c
 index XXXXXXX..XXXXXXX 100644
 --- a/hw/virtio/vhost-shadow-virtqueue.c
 +++ b/hw/virtio/vhost-shadow-virtqueue.c
@@ -XXX,XX +XXX,XX @@
  #include "hw/virtio/vhost-shadow-virtqueue.h"
  #include "qemu/error-report.h"
 +#include "qemu/main-loop.h"
 +#include "linux-headers/linux/vhost.h"
 +
 +/**
 + * Forward guest notifications.
 + *
 + * @n: guest kick event notifier, the one that guest set to notify svq.
 + */
 +static void vhost_handle_guest_kick(EventNotifier *n)
 +{
 +    VhostShadowVirtqueue *svq = container_of(n, VhostShadowVirtqueue, svq_kick);
 +    event_notifier_test_and_clear(n);
 +    event_notifier_set(&svq->hdev_kick);
 +}
 +
 +/**
 + * Set a new file descriptor for the guest to kick the SVQ and notify for avail
 + *
 + * @svq: The svq
 + * @svq_kick_fd: The svq kick fd
 + *
 + * Note that the SVQ will never close the old file descriptor.
 + */
 +void vhost_svq_set_svq_kick_fd(VhostShadowVirtqueue *svq, int svq_kick_fd)
 +{
 +    EventNotifier *svq_kick = &svq->svq_kick;
 +    bool poll_stop = VHOST_FILE_UNBIND != event_notifier_get_fd(svq_kick);
 +    bool poll_start = svq_kick_fd != VHOST_FILE_UNBIND;
 +
 +    if (poll_stop) {
 +        event_notifier_set_handler(svq_kick, NULL);
 +    }
 +
 +    /*
 +     * event_notifier_set_handler already checks for guest's notifications if
 +     * they arrive at the new file descriptor in the switch, so there is no
 +     * need to explicitly check for them.
 +     */
 +    if (poll_start) {
 +        event_notifier_init_fd(svq_kick, svq_kick_fd);
 +        event_notifier_set(svq_kick);
 +        event_notifier_set_handler(svq_kick, vhost_handle_guest_kick);
 +    }
 +}
 +
 +/**
 + * Stop the shadow virtqueue operation.
 + * @svq: Shadow Virtqueue
 + */
 +void vhost_svq_stop(VhostShadowVirtqueue *svq)
 +{
 +    event_notifier_set_handler(&svq->svq_kick, NULL);
 +}
  /**
   * Creates vhost shadow virtqueue, and instructs the vhost device to use the
@@ -XXX,XX +XXX,XX @@ VhostShadowVirtqueue *vhost_svq_new(void)
          goto err_init_hdev_call;
      }
 +    event_notifier_init_fd(&svq->svq_kick, VHOST_FILE_UNBIND);
      return g_steal_pointer(&svq);
  err_init_hdev_call:
@@ -XXX,XX +XXX,XX @@ err_init_hdev_kick:
  void vhost_svq_free(gpointer pvq)
  {
      VhostShadowVirtqueue *vq = pvq;
 +    vhost_svq_stop(vq);
      event_notifier_cleanup(&vq->hdev_kick);
      event_notifier_cleanup(&vq->hdev_call);
      g_free(vq);
 diff --git a/hw/virtio/vhost-shadow-virtqueue.h b/hw/virtio/vhost-shadow-virtqueue.h
 index XXXXXXX..XXXXXXX 100644
 --- a/hw/virtio/vhost-shadow-virtqueue.h
 +++ b/hw/virtio/vhost-shadow-virtqueue.h
@@ -XXX,XX +XXX,XX @@ typedef struct VhostShadowVirtqueue {
      EventNotifier hdev_kick;
      /* Shadow call notifier, sent to vhost */
      EventNotifier hdev_call;
 +
 +    /*
 +     * Borrowed virtqueue's guest to host notifier. To borrow it in this event
 +     * notifier allows to recover the VhostShadowVirtqueue from the event loop
 +     * easily. If we use the VirtQueue's one, we don't have an easy way to
 +     * retrieve VhostShadowVirtqueue.
 +     *
 +     * So shadow virtqueue must not clean it, or we would lose VirtQueue one.
 +     */
 +    EventNotifier svq_kick;
  } VhostShadowVirtqueue;
 +void vhost_svq_set_svq_kick_fd(VhostShadowVirtqueue *svq, int svq_kick_fd);
 +
 +void vhost_svq_stop(VhostShadowVirtqueue *svq);
 +
  VhostShadowVirtqueue *vhost_svq_new(void);
  void vhost_svq_free(gpointer vq);
 diff --git a/hw/virtio/vhost-vdpa.c b/hw/virtio/vhost-vdpa.c
 index XXXXXXX..XXXXXXX 100644
 --- a/hw/virtio/vhost-vdpa.c
 +++ b/hw/virtio/vhost-vdpa.c
@@ -XXX,XX +XXX,XX @@
  #include "hw/virtio/vhost.h"
  #include "hw/virtio/vhost-backend.h"
  #include "hw/virtio/virtio-net.h"
 +#include "hw/virtio/vhost-shadow-virtqueue.h"
  #include "hw/virtio/vhost-vdpa.h"
  #include "exec/address-spaces.h"
  #include "qemu/main-loop.h"
  #include "cpu.h"
  #include "trace.h"
  #include "qemu-common.h"
 +#include "qapi/error.h"
  /*
   * Return one past the end of the end of section. Be careful with uint64_t
@@ -XXX,XX +XXX,XX @@ static bool vhost_vdpa_one_time_request(struct vhost_dev *dev)
      return v->index != 0;
  }
 +static int vhost_vdpa_init_svq(struct vhost_dev *hdev, struct vhost_vdpa *v,
 +                               Error **errp)
 +{
 +    g_autoptr(GPtrArray) shadow_vqs = NULL;
 +
 +    if (!v->shadow_vqs_enabled) {
 +        return 0;
 +    }
 +
 +    shadow_vqs = g_ptr_array_new_full(hdev->nvqs, vhost_svq_free);
 +    for (unsigned n = 0; n < hdev->nvqs; ++n) {
 +        g_autoptr(VhostShadowVirtqueue) svq = vhost_svq_new();
 +
 +        if (unlikely(!svq)) {
 +            error_setg(errp, "Cannot create svq %u", n);
 +            return -1;
 +        }
 +        g_ptr_array_add(shadow_vqs, g_steal_pointer(&svq));
 +    }
 +
 +    v->shadow_vqs = g_steal_pointer(&shadow_vqs);
 +    return 0;
 +}
 +
  static int vhost_vdpa_init(struct vhost_dev *dev, void *opaque, Error **errp)
  {
      struct vhost_vdpa *v;
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_init(struct vhost_dev *dev, void *opaque, Error **errp)
      dev->opaque =  opaque ;
      v->listener = vhost_vdpa_memory_listener;
      v->msg_type = VHOST_IOTLB_MSG_V2;
 +    ret = vhost_vdpa_init_svq(dev, v, errp);
 +    if (ret) {
 +        goto err;
 +    }
      vhost_vdpa_get_iova_range(v);
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_init(struct vhost_dev *dev, void *opaque, Error **errp)
                                 VIRTIO_CONFIG_S_DRIVER);
      return 0;
 +
 +err:
 +    ram_block_discard_disable(false);
 +    return ret;
  }
  static void vhost_vdpa_host_notifier_uninit(struct vhost_dev *dev,
@@ -XXX,XX +XXX,XX @@ static void vhost_vdpa_host_notifiers_uninit(struct vhost_dev *dev, int n)
  static void vhost_vdpa_host_notifiers_init(struct vhost_dev *dev)
  {
 +    struct vhost_vdpa *v = dev->opaque;
      int i;
 +    if (v->shadow_vqs_enabled) {
 +        /* FIXME SVQ is not compatible with host notifiers mr */
 +        return;
 +    }
 +
      for (i = dev->vq_index; i < dev->vq_index + dev->nvqs; i++) {
          if (vhost_vdpa_host_notifier_init(dev, i)) {
              goto err;
@@ -XXX,XX +XXX,XX @@ err:
      return;
  }
 +static void vhost_vdpa_svq_cleanup(struct vhost_dev *dev)
 +{
 +    struct vhost_vdpa *v = dev->opaque;
 +    size_t idx;
 +
 +    if (!v->shadow_vqs) {
 +        return;
 +    }
 +
 +    for (idx = 0; idx < v->shadow_vqs->len; ++idx) {
 +        vhost_svq_stop(g_ptr_array_index(v->shadow_vqs, idx));
 +    }
 +    g_ptr_array_free(v->shadow_vqs, true);
 +}
 +
  static int vhost_vdpa_cleanup(struct vhost_dev *dev)
  {
      struct vhost_vdpa *v;
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_cleanup(struct vhost_dev *dev)
      trace_vhost_vdpa_cleanup(dev, v);
      vhost_vdpa_host_notifiers_uninit(dev, dev->nvqs);
      memory_listener_unregister(&v->listener);
 +    vhost_vdpa_svq_cleanup(dev);
      dev->opaque = NULL;
      ram_block_discard_disable(false);
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_get_device_id(struct vhost_dev *dev,
      return ret;
  }
 +static void vhost_vdpa_reset_svq(struct vhost_vdpa *v)
 +{
 +    if (!v->shadow_vqs_enabled) {
 +        return;
 +    }
 +
 +    for (unsigned i = 0; i < v->shadow_vqs->len; ++i) {
 +        VhostShadowVirtqueue *svq = g_ptr_array_index(v->shadow_vqs, i);
 +        vhost_svq_stop(svq);
 +    }
 +}
 +
  static int vhost_vdpa_reset_device(struct vhost_dev *dev)
  {
 +    struct vhost_vdpa *v = dev->opaque;
      int ret;
      uint8_t status = 0;
 +    vhost_vdpa_reset_svq(v);
 +
      ret = vhost_vdpa_call(dev, VHOST_VDPA_SET_STATUS, &status);
      trace_vhost_vdpa_reset_device(dev, status);
      return ret;
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_get_config(struct vhost_dev *dev, uint8_t *config,
      return ret;
   }
 +static int vhost_vdpa_set_vring_dev_kick(struct vhost_dev *dev,
 +                                         struct vhost_vring_file *file)
 +{
 +    trace_vhost_vdpa_set_vring_kick(dev, file->index, file->fd);
 +    return vhost_vdpa_call(dev, VHOST_SET_VRING_KICK, file);
 +}
 +
 +/**
 + * Set the shadow virtqueue descriptors to the device
 + *
 + * @dev: The vhost device model
 + * @svq: The shadow virtqueue
 + * @idx: The index of the virtqueue in the vhost device
 + * @errp: Error
 + */
 +static bool vhost_vdpa_svq_setup(struct vhost_dev *dev,
 +                                 VhostShadowVirtqueue *svq, unsigned idx,
 +                                 Error **errp)
 +{
 +    struct vhost_vring_file file = {
 +        .index = dev->vq_index + idx,
 +    };
 +    const EventNotifier *event_notifier = &svq->hdev_kick;
 +    int r;
 +
 +    file.fd = event_notifier_get_fd(event_notifier);
 +    r = vhost_vdpa_set_vring_dev_kick(dev, &file);
 +    if (unlikely(r != 0)) {
 +        error_setg_errno(errp, -r, "Can't set device kick fd");
 +    }
 +
 +    return r == 0;
 +}
 +
 +static bool vhost_vdpa_svqs_start(struct vhost_dev *dev)
 +{
 +    struct vhost_vdpa *v = dev->opaque;
 +    Error *err = NULL;
 +    unsigned i;
 +
 +    if (!v->shadow_vqs) {
 +        return true;
 +    }
 +
 +    for (i = 0; i < v->shadow_vqs->len; ++i) {
 +        VhostShadowVirtqueue *svq = g_ptr_array_index(v->shadow_vqs, i);
 +        bool ok = vhost_vdpa_svq_setup(dev, svq, i, &err);
 +        if (unlikely(!ok)) {
 +            error_reportf_err(err, "Cannot setup SVQ %u: ", i);
 +            return false;
 +        }
 +    }
 +
 +    return true;
 +}
 +
  static int vhost_vdpa_dev_start(struct vhost_dev *dev, bool started)
  {
      struct vhost_vdpa *v = dev->opaque;
 +    bool ok;
      trace_vhost_vdpa_dev_start(dev, started);
      if (started) {
          vhost_vdpa_host_notifiers_init(dev);
 +        ok = vhost_vdpa_svqs_start(dev);
 +        if (unlikely(!ok)) {
 +            return -1;
 +        }
          vhost_vdpa_set_vring_ready(dev);
      } else {
          vhost_vdpa_host_notifiers_uninit(dev, dev->nvqs);
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_get_vring_base(struct vhost_dev *dev,
  static int vhost_vdpa_set_vring_kick(struct vhost_dev *dev,
                                         struct vhost_vring_file *file)
  {
 -    trace_vhost_vdpa_set_vring_kick(dev, file->index, file->fd);
 -    return vhost_vdpa_call(dev, VHOST_SET_VRING_KICK, file);
 +    struct vhost_vdpa *v = dev->opaque;
 +    int vdpa_idx = file->index - dev->vq_index;
 +
 +    if (v->shadow_vqs_enabled) {
 +        VhostShadowVirtqueue *svq = g_ptr_array_index(v->shadow_vqs, vdpa_idx);
 +        vhost_svq_set_svq_kick_fd(svq, file->fd);
 +        return 0;
 +    } else {
 +        return vhost_vdpa_set_vring_dev_kick(dev, file);
 +    }
  }
  static int vhost_vdpa_set_vring_call(struct vhost_dev *dev,
 diff --git a/include/hw/virtio/vhost-vdpa.h b/include/hw/virtio/vhost-vdpa.h
 index XXXXXXX..XXXXXXX 100644
 --- a/include/hw/virtio/vhost-vdpa.h
 +++ b/include/hw/virtio/vhost-vdpa.h
@@ -XXX,XX +XXX,XX @@
  #ifndef HW_VIRTIO_VHOST_VDPA_H
  #define HW_VIRTIO_VHOST_VDPA_H
 +#include <gmodule.h>
 +
  #include "hw/virtio/virtio.h"
  #include "standard-headers/linux/vhost_types.h"
@@ -XXX,XX +XXX,XX @@ typedef struct vhost_vdpa {
      bool iotlb_batch_begin_sent;
      MemoryListener listener;
      struct vhost_vdpa_iova_range iova_range;
 +    bool shadow_vqs_enabled;
 +    GPtrArray *shadow_vqs;
      struct vhost_dev *dev;
      VhostVDPAHostNotifier notifier[VIRTIO_QUEUE_MAX];
  } VhostVDPA;
 --
 .7.4

-[PULL 07/11] vmnet: stop recieving events when VM is stopped
+[PULL V3 04/15] vhost: Add Shadow VirtQueue call forwarding capabilities
-From: Joelle van Dyne <j@getutm.app>
+From: Eugenio Pérez <eperezma@redhat.com>
-When the VM is stopped using the HMP command "stop", soon the handler will
+This will make qemu aware of the device used buffers, allowing it to
-stop reading from the vmnet interface. This causes a flood of
+write the guest memory with its contents if needed.
 `VMNET_INTERFACE_PACKETS_AVAILABLE` events to arrive and puts the host CPU
 at 100%. We fix this by removing the event handler from vmnet when the VM
 is no longer in a running state and restore it when we return to a running
 state.
-Signed-off-by: Joelle van Dyne <j@getutm.app>
+Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
 Acked-by: Michael S. Tsirkin <mst@redhat.com>
 Signed-off-by: Jason Wang <jasowang@redhat.com>
 ---
- net/vmnet-common.m | 48 +++++++++++++++++++++++++++++++++++-------------
+ hw/virtio/vhost-shadow-virtqueue.c | 38 ++++++++++++++++++++++++++++++++++++++
- net/vmnet_int.h    |  2 ++
+ hw/virtio/vhost-shadow-virtqueue.h |  4 ++++
-files changed, 37 insertions(+), 13 deletions(-)
+ hw/virtio/vhost-vdpa.c             | 31 +++++++++++++++++++++++++++++--
 files changed, 71 insertions(+), 2 deletions(-)
-diff --git a/net/vmnet-common.m b/net/vmnet-common.m
+diff --git a/hw/virtio/vhost-shadow-virtqueue.c b/hw/virtio/vhost-shadow-virtqueue.c
 index XXXXXXX..XXXXXXX 100644
---- a/net/vmnet-common.m
+--- a/hw/virtio/vhost-shadow-virtqueue.c
-+++ b/net/vmnet-common.m
++++ b/hw/virtio/vhost-shadow-virtqueue.c
-@@ -XXX,XX +XXX,XX @@
+@@ -XXX,XX +XXX,XX @@ static void vhost_handle_guest_kick(EventNotifier *n)
  #include "clients.h"
  #include "qemu/error-report.h"
  #include "qapi/error.h"
 +#include "sysemu/runstate.h"
  #include <vmnet/vmnet.h>
  #include <dispatch/dispatch.h>
@@ -XXX,XX +XXX,XX @@ static void vmnet_bufs_init(VmnetState *s)
      }
  }
+ /**
++ * Forward vhost notifications
++ *
++ * @n: hdev call event notifier, the one that device set to notify svq.
++ */
++static void vhost_svq_handle_call(EventNotifier *n)
++{
++    VhostShadowVirtqueue *svq = container_of(n, VhostShadowVirtqueue,
++                                             hdev_call);
++    event_notifier_test_and_clear(n);
++    event_notifier_set(&svq->svq_call);
++}
++
 +/**
-+ * Called on state change to un-register/re-register handlers
++ * Set the call notifier for the SVQ to call the guest
 + *
 + * @svq: Shadow virtqueue
 + * @call_fd: call notifier
 + *
 + * Called on BQL context.
 + */
-+static void vmnet_vm_state_change_cb(void *opaque, bool running, RunState state)
++void vhost_svq_set_svq_call_fd(VhostShadowVirtqueue *svq, int call_fd)
 +{
-+    VmnetState *s = opaque;
++    if (call_fd == VHOST_FILE_UNBIND) {
-+
++        /*
-+    if (running) {
++         * Fail event_notifier_set if called handling device call.
-+        vmnet_interface_set_event_callback(
++         *
-+            s->vmnet_if,
++         * SVQ still needs device notifications, since it needs to keep
-+            VMNET_INTERFACE_PACKETS_AVAILABLE,
++         * forwarding used buffers even with the unbind.
-+            s->if_queue,
++         */
-+            ^(interface_event_t event_id, xpc_object_t event) {
++        memset(&svq->svq_call, 0, sizeof(svq->svq_call));
 +                assert(event_id == VMNET_INTERFACE_PACKETS_AVAILABLE);
 +                /*
 +                 * This function is being called from a non qemu thread, so
 +                 * we only schedule a BH, and do the rest of the io completion
 +                 * handling from vmnet_send_bh() which runs in a qemu context.
 +                 */
 +                qemu_bh_schedule(s->send_bh);
 +            });
 +    } else {
-+        vmnet_interface_set_event_callback(
++        event_notifier_init_fd(&svq->svq_call, call_fd);
 +            s->vmnet_if,
 +            VMNET_INTERFACE_PACKETS_AVAILABLE,
 +            NULL,
 +            NULL);
 +    }
 +}
- int vmnet_if_create(NetClientState *nc,
-                     xpc_object_t if_desc,
-@@ -XXX,XX +XXX,XX @@ int vmnet_if_create(NetClientState *nc,
-     s->packets_send_current_pos = 0;
-     s->packets_send_end_pos = 0;
--    vmnet_interface_set_event_callback(
--        s->vmnet_if,
--        VMNET_INTERFACE_PACKETS_AVAILABLE,
--        s->if_queue,
--        ^(interface_event_t event_id, xpc_object_t event) {
--            assert(event_id == VMNET_INTERFACE_PACKETS_AVAILABLE);
--            /*
--             * This function is being called from a non qemu thread, so
--             * we only schedule a BH, and do the rest of the io completion
--             * handling from vmnet_send_bh() which runs in a qemu context.
--             */
--            qemu_bh_schedule(s->send_bh);
--        });
-+    vmnet_vm_state_change_cb(s, 1, RUN_STATE_RUNNING);
 +
-+    s->change = qemu_add_vm_change_state_handler(vmnet_vm_state_change_cb, s);
++/**
+  * Set a new file descriptor for the guest to kick the SVQ and notify for avail
-     return 0;
+  *
   * @svq: The svq
@@ -XXX,XX +XXX,XX @@ VhostShadowVirtqueue *vhost_svq_new(void)
      }
      event_notifier_init_fd(&svq->svq_kick, VHOST_FILE_UNBIND);
 +    event_notifier_set_handler(&svq->hdev_call, vhost_svq_handle_call);
      return g_steal_pointer(&svq);
  err_init_hdev_call:
@@ -XXX,XX +XXX,XX @@ void vhost_svq_free(gpointer pvq)
      VhostShadowVirtqueue *vq = pvq;
      vhost_svq_stop(vq);
      event_notifier_cleanup(&vq->hdev_kick);
 +    event_notifier_set_handler(&vq->hdev_call, NULL);
      event_notifier_cleanup(&vq->hdev_call);
      g_free(vq);
  }
-@@ -XXX,XX +XXX,XX @@ void vmnet_cleanup_common(NetClientState *nc)
+diff --git a/hw/virtio/vhost-shadow-virtqueue.h b/hw/virtio/vhost-shadow-virtqueue.h
-         return;
+index XXXXXXX..XXXXXXX 100644
 --- a/hw/virtio/vhost-shadow-virtqueue.h
 +++ b/hw/virtio/vhost-shadow-virtqueue.h
@@ -XXX,XX +XXX,XX @@ typedef struct VhostShadowVirtqueue {
       * So shadow virtqueue must not clean it, or we would lose VirtQueue one.
       */
      EventNotifier svq_kick;
 +
 +    /* Guest's call notifier, where the SVQ calls guest. */
 +    EventNotifier svq_call;
  } VhostShadowVirtqueue;
  void vhost_svq_set_svq_kick_fd(VhostShadowVirtqueue *svq, int svq_kick_fd);
 +void vhost_svq_set_svq_call_fd(VhostShadowVirtqueue *svq, int call_fd);
  void vhost_svq_stop(VhostShadowVirtqueue *svq);
 diff --git a/hw/virtio/vhost-vdpa.c b/hw/virtio/vhost-vdpa.c
 index XXXXXXX..XXXXXXX 100644
 --- a/hw/virtio/vhost-vdpa.c
 +++ b/hw/virtio/vhost-vdpa.c
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_vring_dev_kick(struct vhost_dev *dev,
      return vhost_vdpa_call(dev, VHOST_SET_VRING_KICK, file);
  }
 +static int vhost_vdpa_set_vring_dev_call(struct vhost_dev *dev,
 +                                         struct vhost_vring_file *file)
 +{
 +    trace_vhost_vdpa_set_vring_call(dev, file->index, file->fd);
 +    return vhost_vdpa_call(dev, VHOST_SET_VRING_CALL, file);
 +}
 +
  /**
   * Set the shadow virtqueue descriptors to the device
   *
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_vring_dev_kick(struct vhost_dev *dev,
   * @svq: The shadow virtqueue
   * @idx: The index of the virtqueue in the vhost device
   * @errp: Error
 + *
 + * Note that this function does not rewind kick file descriptor if cannot set
 + * call one.
   */
  static bool vhost_vdpa_svq_setup(struct vhost_dev *dev,
                                   VhostShadowVirtqueue *svq, unsigned idx,
@@ -XXX,XX +XXX,XX @@ static bool vhost_vdpa_svq_setup(struct vhost_dev *dev,
      r = vhost_vdpa_set_vring_dev_kick(dev, &file);
      if (unlikely(r != 0)) {
          error_setg_errno(errp, -r, "Can't set device kick fd");
 +        return false;
 +    }
 +
 +    event_notifier = &svq->hdev_call;
 +    file.fd = event_notifier_get_fd(event_notifier);
 +    r = vhost_vdpa_set_vring_dev_call(dev, &file);
 +    if (unlikely(r != 0)) {
 +        error_setg_errno(errp, -r, "Can't set device call fd");
      }
-+    vmnet_vm_state_change_cb(s, 0, RUN_STATE_SHUTDOWN);
+     return r == 0;
-+    qemu_del_vm_change_state_handler(s->change);
+@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_vring_kick(struct vhost_dev *dev,
-     if_stopped_sem = dispatch_semaphore_create(0);
+ static int vhost_vdpa_set_vring_call(struct vhost_dev *dev,
-     vmnet_stop_interface(
+                                        struct vhost_vring_file *file)
-         s->vmnet_if,
+ {
-diff --git a/net/vmnet_int.h b/net/vmnet_int.h
+-    trace_vhost_vdpa_set_vring_call(dev, file->index, file->fd);
-index XXXXXXX..XXXXXXX 100644
+-    return vhost_vdpa_call(dev, VHOST_SET_VRING_CALL, file);
---- a/net/vmnet_int.h
++    struct vhost_vdpa *v = dev->opaque;
 +++ b/net/vmnet_int.h
@@ -XXX,XX +XXX,XX @@ typedef struct VmnetState {
      int packets_send_end_pos;
      struct iovec iov_buf[VMNET_PACKETS_LIMIT];
 +
-+    VMChangeStateEntry *change;
++    if (v->shadow_vqs_enabled) {
- } VmnetState;
++        int vdpa_idx = file->index - dev->vq_index;
++        VhostShadowVirtqueue *svq = g_ptr_array_index(v->shadow_vqs, vdpa_idx);
- const char *vmnet_status_map_str(vmnet_return_t status);
++
 +        vhost_svq_set_svq_call_fd(svq, file->fd);
 +        return 0;
 +    } else {
 +        return vhost_vdpa_set_vring_dev_call(dev, file);
 +    }
  }
  static int vhost_vdpa_get_features(struct vhost_dev *dev,
 --
 .7.4

-New patch
+[PULL V3 05/15] vhost: Add vhost_svq_valid_features to shadow vq
+From: Eugenio Pérez <eperezma@redhat.com>
+This allows SVQ to negotiate features with the guest and the device. For
+the device, SVQ is a driver. While this function bypasses all
+non-transport features, it needs to disable the features that SVQ does
+not support when forwarding buffers. This includes packed vq layout,
+indirect descriptors or event idx.
+Future changes can add support to offer more features to the guest,
+since the use of VirtQueue gives this for free. This is left out at the
+moment for simplicity.
+Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
+Acked-by: Michael S. Tsirkin <mst@redhat.com>
+Signed-off-by: Jason Wang <jasowang@redhat.com>
+---
+ hw/virtio/vhost-shadow-virtqueue.c | 44 ++++++++++++++++++++++++++++++++++++++
+ hw/virtio/vhost-shadow-virtqueue.h |  2 ++
+ hw/virtio/vhost-vdpa.c             | 15 +++++++++++++
+files changed, 61 insertions(+)
+diff --git a/hw/virtio/vhost-shadow-virtqueue.c b/hw/virtio/vhost-shadow-virtqueue.c
+index XXXXXXX..XXXXXXX 100644
+--- a/hw/virtio/vhost-shadow-virtqueue.c
++++ b/hw/virtio/vhost-shadow-virtqueue.c
+@@ -XXX,XX +XXX,XX @@
+ #include "hw/virtio/vhost-shadow-virtqueue.h"
+ #include "qemu/error-report.h"
++#include "qapi/error.h"
+ #include "qemu/main-loop.h"
+ #include "linux-headers/linux/vhost.h"
+ /**
++ * Validate the transport device features that both guests can use with the SVQ
++ * and SVQs can use with the device.
++ *
++ * @dev_features: The features
++ * @errp: Error pointer
++ */
++bool vhost_svq_valid_features(uint64_t features, Error **errp)
++{
++    bool ok = true;
++    uint64_t svq_features = features;
++
++    for (uint64_t b = VIRTIO_TRANSPORT_F_START; b <= VIRTIO_TRANSPORT_F_END;
++         ++b) {
++        switch (b) {
++        case VIRTIO_F_ANY_LAYOUT:
++            continue;
++
++        case VIRTIO_F_ACCESS_PLATFORM:
++            /* SVQ trust in the host's IOMMU to translate addresses */
++        case VIRTIO_F_VERSION_1:
++            /* SVQ trust that the guest vring is little endian */
++            if (!(svq_features & BIT_ULL(b))) {
++                svq_features |= BIT_ULL(b);
++                ok = false;
++            }
++            continue;
++
++        default:
++            if (svq_features & BIT_ULL(b)) {
++                svq_features &= ~BIT_ULL(b);
++                ok = false;
++            }
++        }
++    }
++
++    if (!ok) {
++        error_setg(errp, "SVQ Invalid device feature flags, offer: 0x%"PRIx64
++                         ", ok: 0x%"PRIx64, features, svq_features);
++    }
++    return ok;
++}
++
++/**
+  * Forward guest notifications.
+  *
+  * @n: guest kick event notifier, the one that guest set to notify svq.
+diff --git a/hw/virtio/vhost-shadow-virtqueue.h b/hw/virtio/vhost-shadow-virtqueue.h
+index XXXXXXX..XXXXXXX 100644
+--- a/hw/virtio/vhost-shadow-virtqueue.h
++++ b/hw/virtio/vhost-shadow-virtqueue.h
+@@ -XXX,XX +XXX,XX @@ typedef struct VhostShadowVirtqueue {
+     EventNotifier svq_call;
+ } VhostShadowVirtqueue;
++bool vhost_svq_valid_features(uint64_t features, Error **errp);
++
+ void vhost_svq_set_svq_kick_fd(VhostShadowVirtqueue *svq, int svq_kick_fd);
+ void vhost_svq_set_svq_call_fd(VhostShadowVirtqueue *svq, int call_fd);
+diff --git a/hw/virtio/vhost-vdpa.c b/hw/virtio/vhost-vdpa.c
+index XXXXXXX..XXXXXXX 100644
+--- a/hw/virtio/vhost-vdpa.c
++++ b/hw/virtio/vhost-vdpa.c
+@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_init_svq(struct vhost_dev *hdev, struct vhost_vdpa *v,
+                                Error **errp)
+ {
+     g_autoptr(GPtrArray) shadow_vqs = NULL;
++    uint64_t dev_features, svq_features;
++    int r;
++    bool ok;
+     if (!v->shadow_vqs_enabled) {
+         return 0;
+     }
++    r = hdev->vhost_ops->vhost_get_features(hdev, &dev_features);
++    if (r != 0) {
++        error_setg_errno(errp, -r, "Can't get vdpa device features");
++        return r;
++    }
++
++    svq_features = dev_features;
++    ok = vhost_svq_valid_features(svq_features, errp);
++    if (unlikely(!ok)) {
++        return -1;
++    }
++
+     shadow_vqs = g_ptr_array_new_full(hdev->nvqs, vhost_svq_free);
+     for (unsigned n = 0; n < hdev->nvqs; ++n) {
+         g_autoptr(VhostShadowVirtqueue) svq = vhost_svq_new();
+--
+.7.4

-New patch
+[PULL V3 06/15] virtio: Add vhost_svq_get_vring_addr
+From: Eugenio Pérez <eperezma@redhat.com>
+It reports the shadow virtqueue address from qemu virtual address space.
+Since this will be different from the guest's vaddr, but the device can
+access it, SVQ takes special care about its alignment & lack of garbage
+data. It assumes that IOMMU will work in host_page_size ranges for that.
+Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
+Acked-by: Michael S. Tsirkin <mst@redhat.com>
+Signed-off-by: Jason Wang <jasowang@redhat.com>
+---
+ hw/virtio/vhost-shadow-virtqueue.c | 29 +++++++++++++++++++++++++++++
+ hw/virtio/vhost-shadow-virtqueue.h |  9 +++++++++
+files changed, 38 insertions(+)
+diff --git a/hw/virtio/vhost-shadow-virtqueue.c b/hw/virtio/vhost-shadow-virtqueue.c
+index XXXXXXX..XXXXXXX 100644
+--- a/hw/virtio/vhost-shadow-virtqueue.c
++++ b/hw/virtio/vhost-shadow-virtqueue.c
+@@ -XXX,XX +XXX,XX @@ void vhost_svq_set_svq_call_fd(VhostShadowVirtqueue *svq, int call_fd)
+ }
+ /**
++ * Get the shadow vq vring address.
++ * @svq: Shadow virtqueue
++ * @addr: Destination to store address
++ */
++void vhost_svq_get_vring_addr(const VhostShadowVirtqueue *svq,
++                              struct vhost_vring_addr *addr)
++{
++    addr->desc_user_addr = (uint64_t)(intptr_t)svq->vring.desc;
++    addr->avail_user_addr = (uint64_t)(intptr_t)svq->vring.avail;
++    addr->used_user_addr = (uint64_t)(intptr_t)svq->vring.used;
++}
++
++size_t vhost_svq_driver_area_size(const VhostShadowVirtqueue *svq)
++{
++    size_t desc_size = sizeof(vring_desc_t) * svq->vring.num;
++    size_t avail_size = offsetof(vring_avail_t, ring) +
++                                             sizeof(uint16_t) * svq->vring.num;
++
++    return ROUND_UP(desc_size + avail_size, qemu_real_host_page_size);
++}
++
++size_t vhost_svq_device_area_size(const VhostShadowVirtqueue *svq)
++{
++    size_t used_size = offsetof(vring_used_t, ring) +
++                                    sizeof(vring_used_elem_t) * svq->vring.num;
++    return ROUND_UP(used_size, qemu_real_host_page_size);
++}
++
++/**
+  * Set a new file descriptor for the guest to kick the SVQ and notify for avail
+  *
+  * @svq: The svq
+diff --git a/hw/virtio/vhost-shadow-virtqueue.h b/hw/virtio/vhost-shadow-virtqueue.h
+index XXXXXXX..XXXXXXX 100644
+--- a/hw/virtio/vhost-shadow-virtqueue.h
++++ b/hw/virtio/vhost-shadow-virtqueue.h
+@@ -XXX,XX +XXX,XX @@
+ #define VHOST_SHADOW_VIRTQUEUE_H
+ #include "qemu/event_notifier.h"
++#include "hw/virtio/virtio.h"
++#include "standard-headers/linux/vhost_types.h"
+ /* Shadow virtqueue to relay notifications */
+ typedef struct VhostShadowVirtqueue {
++    /* Shadow vring */
++    struct vring vring;
++
+     /* Shadow kick notifier, sent to vhost */
+     EventNotifier hdev_kick;
+     /* Shadow call notifier, sent to vhost */
+@@ -XXX,XX +XXX,XX @@ bool vhost_svq_valid_features(uint64_t features, Error **errp);
+ void vhost_svq_set_svq_kick_fd(VhostShadowVirtqueue *svq, int svq_kick_fd);
+ void vhost_svq_set_svq_call_fd(VhostShadowVirtqueue *svq, int call_fd);
++void vhost_svq_get_vring_addr(const VhostShadowVirtqueue *svq,
++                              struct vhost_vring_addr *addr);
++size_t vhost_svq_driver_area_size(const VhostShadowVirtqueue *svq);
++size_t vhost_svq_device_area_size(const VhostShadowVirtqueue *svq);
+ void vhost_svq_stop(VhostShadowVirtqueue *svq);
+--
+.7.4

-New patch
+[PULL V3 07/15] vdpa: adapt vhost_ops callbacks to svq
+From: Eugenio Pérez <eperezma@redhat.com>
+First half of the buffers forwarding part, preparing vhost-vdpa
+callbacks to SVQ to offer it. QEMU cannot enable it at this moment, so
+this is effectively dead code at the moment, but it helps to reduce
+patch size.
+Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
+Acked-by: Michael S. Tsirkin <mst@redhat.com>
+Signed-off-by: Jason Wang <jasowang@redhat.com>
+---
+ hw/virtio/vhost-vdpa.c | 48 +++++++++++++++++++++++++++++++++++++++++-------
+file changed, 41 insertions(+), 7 deletions(-)
+diff --git a/hw/virtio/vhost-vdpa.c b/hw/virtio/vhost-vdpa.c
+index XXXXXXX..XXXXXXX 100644
+--- a/hw/virtio/vhost-vdpa.c
++++ b/hw/virtio/vhost-vdpa.c
+@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_get_config(struct vhost_dev *dev, uint8_t *config,
+     return ret;
+  }
++static int vhost_vdpa_set_dev_vring_base(struct vhost_dev *dev,
++                                         struct vhost_vring_state *ring)
++{
++    trace_vhost_vdpa_set_vring_base(dev, ring->index, ring->num);
++    return vhost_vdpa_call(dev, VHOST_SET_VRING_BASE, ring);
++}
++
+ static int vhost_vdpa_set_vring_dev_kick(struct vhost_dev *dev,
+                                          struct vhost_vring_file *file)
+ {
+@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_vring_dev_call(struct vhost_dev *dev,
+     return vhost_vdpa_call(dev, VHOST_SET_VRING_CALL, file);
+ }
++static int vhost_vdpa_set_vring_dev_addr(struct vhost_dev *dev,
++                                         struct vhost_vring_addr *addr)
++{
++    trace_vhost_vdpa_set_vring_addr(dev, addr->index, addr->flags,
++                                addr->desc_user_addr, addr->used_user_addr,
++                                addr->avail_user_addr,
++                                addr->log_guest_addr);
++
++    return vhost_vdpa_call(dev, VHOST_SET_VRING_ADDR, addr);
++
++}
++
+ /**
+  * Set the shadow virtqueue descriptors to the device
+  *
+@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_log_base(struct vhost_dev *dev, uint64_t base,
+ static int vhost_vdpa_set_vring_addr(struct vhost_dev *dev,
+                                        struct vhost_vring_addr *addr)
+ {
+-    trace_vhost_vdpa_set_vring_addr(dev, addr->index, addr->flags,
+-                                    addr->desc_user_addr, addr->used_user_addr,
+-                                    addr->avail_user_addr,
+-                                    addr->log_guest_addr);
+-    return vhost_vdpa_call(dev, VHOST_SET_VRING_ADDR, addr);
++    struct vhost_vdpa *v = dev->opaque;
++
++    if (v->shadow_vqs_enabled) {
++        /*
++         * Device vring addr was set at device start. SVQ base is handled by
++         * VirtQueue code.
++         */
++        return 0;
++    }
++
++    return vhost_vdpa_set_vring_dev_addr(dev, addr);
+ }
+ static int vhost_vdpa_set_vring_num(struct vhost_dev *dev,
+@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_vring_num(struct vhost_dev *dev,
+ static int vhost_vdpa_set_vring_base(struct vhost_dev *dev,
+                                        struct vhost_vring_state *ring)
+ {
+-    trace_vhost_vdpa_set_vring_base(dev, ring->index, ring->num);
+-    return vhost_vdpa_call(dev, VHOST_SET_VRING_BASE, ring);
++    struct vhost_vdpa *v = dev->opaque;
++
++    if (v->shadow_vqs_enabled) {
++        /*
++         * Device vring base was set at device start. SVQ base is handled by
++         * VirtQueue code.
++         */
++        return 0;
++    }
++
++    return vhost_vdpa_set_dev_vring_base(dev, ring);
+ }
+ static int vhost_vdpa_get_vring_base(struct vhost_dev *dev,
+--
+.7.4

-[PULL 04/11] hw/net/lan9118: log [read|write]b when mode_16bit is enabled rather than abort
+[PULL V3 08/15] vhost: Shadow virtqueue buffers forwarding
-From: Qiang Liu <cyruscyliu@gmail.com>
+From: Eugenio Pérez <eperezma@redhat.com>
-This patch replaces hw_error to guest error log for [read|write]b
+Initial version of shadow virtqueue that actually forward buffers. There
-accesses when mode_16bit is enabled. This avoids aborting qemu.
+is no iommu support at the moment, and that will be addressed in future
+patches of this series. Since all vhost-vdpa devices use forced IOMMU,
-Fixes: 1248f8d4cbc3 ("hw/lan9118: Add basic 16-bit mode support.")
+this means that SVQ is not usable at this point of the series on any
-Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1433
+device.
-Reported-by: Qiang Liu <cyruscyliu@gmail.com>
-Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
+For simplicity it only supports modern devices, that expects vring
-Signed-off-by: Qiang Liu <cyruscyliu@gmail.com>
+in little endian, with split ring and no event idx or indirect
-Suggested-by: Philippe Mathieu-Daudé <philmd@linaro.org>
+descriptors. Support for them will not be added in this series.
 It reuses the VirtQueue code for the device part. The driver part is
 based on Linux's virtio_ring driver, but with stripped functionality
 and optimizations so it's easier to review.
 However, forwarding buffers have some particular pieces: One of the most
 unexpected ones is that a guest's buffer can expand through more than
 one descriptor in SVQ. While this is handled gracefully by qemu's
 emulated virtio devices, it may cause unexpected SVQ queue full. This
 patch also solves it by checking for this condition at both guest's
 kicks and device's calls. The code may be more elegant in the future if
 SVQ code runs in its own iocontext.
 Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
 Acked-by: Michael S. Tsirkin <mst@redhat.com>
 Signed-off-by: Jason Wang <jasowang@redhat.com>
 ---
- hw/net/lan9118.c | 17 ++++++++---------
+ hw/virtio/vhost-shadow-virtqueue.c | 352 ++++++++++++++++++++++++++++++++++++-
-file changed, 8 insertions(+), 9 deletions(-)
+ hw/virtio/vhost-shadow-virtqueue.h |  26 +++
+ hw/virtio/vhost-vdpa.c             | 155 +++++++++++++++-
-diff --git a/hw/net/lan9118.c b/hw/net/lan9118.c
+files changed, 522 insertions(+), 11 deletions(-)
 diff --git a/hw/virtio/vhost-shadow-virtqueue.c b/hw/virtio/vhost-shadow-virtqueue.c
 index XXXXXXX..XXXXXXX 100644
---- a/hw/net/lan9118.c
+--- a/hw/virtio/vhost-shadow-virtqueue.c
-+++ b/hw/net/lan9118.c
++++ b/hw/virtio/vhost-shadow-virtqueue.c
 @@ -XXX,XX +XXX,XX @@
- #include "migration/vmstate.h"
+ #include "qemu/error-report.h"
- #include "net/net.h"
+ #include "qapi/error.h"
- #include "net/eth.h"
+ #include "qemu/main-loop.h"
--#include "hw/hw.h"
++#include "qemu/log.h"
- #include "hw/irq.h"
++#include "qemu/memalign.h"
- #include "hw/net/lan9118.h"
+ #include "linux-headers/linux/vhost.h"
- #include "hw/ptimer.h"
-@@ -XXX,XX +XXX,XX @@
+ /**
- #ifdef DEBUG_LAN9118
+@@ -XXX,XX +XXX,XX @@ bool vhost_svq_valid_features(uint64_t features, Error **errp)
- #define DPRINTF(fmt, ...) \
+ }
- do { printf("lan9118: " fmt , ## __VA_ARGS__); } while (0)
--#define BADF(fmt, ...) \
+ /**
--do { hw_error("lan9118: error: " fmt , ## __VA_ARGS__);} while (0)
+- * Forward guest notifications.
- #else
++ * Number of descriptors that the SVQ can make available from the guest.
- #define DPRINTF(fmt, ...) do {} while(0)
++ *
--#define BADF(fmt, ...) \
++ * @svq: The svq
--do { fprintf(stderr, "lan9118: error: " fmt , ## __VA_ARGS__);} while (0)
++ */
- #endif
++static uint16_t vhost_svq_available_slots(const VhostShadowVirtqueue *svq)
++{
- /* The tx and rx fifo ports are a range of aliased 32-bit registers */
++    return svq->vring.num - (svq->shadow_avail_idx - svq->shadow_used_idx);
-@@ -XXX,XX +XXX,XX @@ static uint32_t do_phy_read(lan9118_state *s, int reg)
++}
-     case 30: /* Interrupt mask */
++
-         return s->phy_int_mask;
++static void vhost_vring_write_descs(VhostShadowVirtqueue *svq,
-     default:
++                                    const struct iovec *iovec, size_t num,
--        BADF("PHY read reg %d\n", reg);
++                                    bool more_descs, bool write)
 +{
 +    uint16_t i = svq->free_head, last = svq->free_head;
 +    unsigned n;
 +    uint16_t flags = write ? cpu_to_le16(VRING_DESC_F_WRITE) : 0;
 +    vring_desc_t *descs = svq->vring.desc;
 +
 +    if (num == 0) {
 +        return;
 +    }
 +
 +    for (n = 0; n < num; n++) {
 +        if (more_descs || (n + 1 < num)) {
 +            descs[i].flags = flags | cpu_to_le16(VRING_DESC_F_NEXT);
 +        } else {
 +            descs[i].flags = flags;
 +        }
 +        descs[i].addr = cpu_to_le64((hwaddr)(intptr_t)iovec[n].iov_base);
 +        descs[i].len = cpu_to_le32(iovec[n].iov_len);
 +
 +        last = i;
 +        i = cpu_to_le16(descs[i].next);
 +    }
 +
 +    svq->free_head = le16_to_cpu(descs[last].next);
 +}
 +
 +static bool vhost_svq_add_split(VhostShadowVirtqueue *svq,
 +                                VirtQueueElement *elem, unsigned *head)
 +{
 +    unsigned avail_idx;
 +    vring_avail_t *avail = svq->vring.avail;
 +
 +    *head = svq->free_head;
 +
 +    /* We need some descriptors here */
 +    if (unlikely(!elem->out_num && !elem->in_num)) {
 +        qemu_log_mask(LOG_GUEST_ERROR,
-+                      "do_phy_read: PHY read reg %d\n", reg);
++                      "Guest provided element with no descriptors");
-         return 0;
++        return false;
 +    }
 +
 +    vhost_vring_write_descs(svq, elem->out_sg, elem->out_num, elem->in_num > 0,
 +                            false);
 +    vhost_vring_write_descs(svq, elem->in_sg, elem->in_num, false, true);
 +
 +    /*
 +     * Put the entry in the available array (but don't update avail->idx until
 +     * they do sync).
 +     */
 +    avail_idx = svq->shadow_avail_idx & (svq->vring.num - 1);
 +    avail->ring[avail_idx] = cpu_to_le16(*head);
 +    svq->shadow_avail_idx++;
 +
 +    /* Update the avail index after write the descriptor */
 +    smp_wmb();
 +    avail->idx = cpu_to_le16(svq->shadow_avail_idx);
 +
 +    return true;
 +}
 +
 +static bool vhost_svq_add(VhostShadowVirtqueue *svq, VirtQueueElement *elem)
 +{
 +    unsigned qemu_head;
 +    bool ok = vhost_svq_add_split(svq, elem, &qemu_head);
 +    if (unlikely(!ok)) {
 +        return false;
 +    }
 +
 +    svq->ring_id_maps[qemu_head] = elem;
 +    return true;
 +}
 +
 +static void vhost_svq_kick(VhostShadowVirtqueue *svq)
 +{
 +    /*
 +     * We need to expose the available array entries before checking the used
 +     * flags
 +     */
 +    smp_mb();
 +    if (svq->vring.used->flags & VRING_USED_F_NO_NOTIFY) {
 +        return;
 +    }
 +
 +    event_notifier_set(&svq->hdev_kick);
 +}
 +
 +/**
 + * Forward available buffers.
 + *
 + * @svq: Shadow VirtQueue
 + *
 + * Note that this function does not guarantee that all guest's available
 + * buffers are available to the device in SVQ avail ring. The guest may have
 + * exposed a GPA / GIOVA contiguous buffer, but it may not be contiguous in
 + * qemu vaddr.
 + *
 + * If that happens, guest's kick notifications will be disabled until the
 + * device uses some buffers.
 + */
 +static void vhost_handle_guest_kick(VhostShadowVirtqueue *svq)
 +{
 +    /* Clear event notifier */
 +    event_notifier_test_and_clear(&svq->svq_kick);
 +
 +    /* Forward to the device as many available buffers as possible */
 +    do {
 +        virtio_queue_set_notification(svq->vq, false);
 +
 +        while (true) {
 +            VirtQueueElement *elem;
 +            bool ok;
 +
 +            if (svq->next_guest_avail_elem) {
 +                elem = g_steal_pointer(&svq->next_guest_avail_elem);
 +            } else {
 +                elem = virtqueue_pop(svq->vq, sizeof(*elem));
 +            }
 +
 +            if (!elem) {
 +                break;
 +            }
 +
 +            if (elem->out_num + elem->in_num > vhost_svq_available_slots(svq)) {
 +                /*
 +                 * This condition is possible since a contiguous buffer in GPA
 +                 * does not imply a contiguous buffer in qemu's VA
 +                 * scatter-gather segments. If that happens, the buffer exposed
 +                 * to the device needs to be a chain of descriptors at this
 +                 * moment.
 +                 *
 +                 * SVQ cannot hold more available buffers if we are here:
 +                 * queue the current guest descriptor and ignore further kicks
 +                 * until some elements are used.
 +                 */
 +                svq->next_guest_avail_elem = elem;
 +                return;
 +            }
 +
 +            ok = vhost_svq_add(svq, elem);
 +            if (unlikely(!ok)) {
 +                /* VQ is broken, just return and ignore any other kicks */
 +                return;
 +            }
 +            vhost_svq_kick(svq);
 +        }
 +
 +        virtio_queue_set_notification(svq->vq, true);
 +    } while (!virtio_queue_empty(svq->vq));
 +}
 +
 +/**
 + * Handle guest's kick.
   *
   * @n: guest kick event notifier, the one that guest set to notify svq.
   */
 -static void vhost_handle_guest_kick(EventNotifier *n)
 +static void vhost_handle_guest_kick_notifier(EventNotifier *n)
  {
      VhostShadowVirtqueue *svq = container_of(n, VhostShadowVirtqueue, svq_kick);
      event_notifier_test_and_clear(n);
 -    event_notifier_set(&svq->hdev_kick);
 +    vhost_handle_guest_kick(svq);
 +}
 +
 +static bool vhost_svq_more_used(VhostShadowVirtqueue *svq)
 +{
 +    if (svq->last_used_idx != svq->shadow_used_idx) {
 +        return true;
 +    }
 +
 +    svq->shadow_used_idx = cpu_to_le16(svq->vring.used->idx);
 +
 +    return svq->last_used_idx != svq->shadow_used_idx;
  }
  /**
 - * Forward vhost notifications
 + * Enable vhost device calls after disable them.
 + *
 + * @svq: The svq
 + *
 + * It returns false if there are pending used buffers from the vhost device,
 + * avoiding the possible races between SVQ checking for more work and enabling
 + * callbacks. True if SVQ used vring has no more pending buffers.
 + */
 +static bool vhost_svq_enable_notification(VhostShadowVirtqueue *svq)
 +{
 +    svq->vring.avail->flags &= ~cpu_to_le16(VRING_AVAIL_F_NO_INTERRUPT);
 +    /* Make sure the flag is written before the read of used_idx */
 +    smp_mb();
 +    return !vhost_svq_more_used(svq);
 +}
 +
 +static void vhost_svq_disable_notification(VhostShadowVirtqueue *svq)
 +{
 +    svq->vring.avail->flags |= cpu_to_le16(VRING_AVAIL_F_NO_INTERRUPT);
 +}
 +
 +static VirtQueueElement *vhost_svq_get_buf(VhostShadowVirtqueue *svq,
 +                                           uint32_t *len)
 +{
 +    vring_desc_t *descs = svq->vring.desc;
 +    const vring_used_t *used = svq->vring.used;
 +    vring_used_elem_t used_elem;
 +    uint16_t last_used;
 +
 +    if (!vhost_svq_more_used(svq)) {
 +        return NULL;
 +    }
 +
 +    /* Only get used array entries after they have been exposed by dev */
 +    smp_rmb();
 +    last_used = svq->last_used_idx & (svq->vring.num - 1);
 +    used_elem.id = le32_to_cpu(used->ring[last_used].id);
 +    used_elem.len = le32_to_cpu(used->ring[last_used].len);
 +
 +    svq->last_used_idx++;
 +    if (unlikely(used_elem.id >= svq->vring.num)) {
 +        qemu_log_mask(LOG_GUEST_ERROR, "Device %s says index %u is used",
 +                      svq->vdev->name, used_elem.id);
 +        return NULL;
 +    }
 +
 +    if (unlikely(!svq->ring_id_maps[used_elem.id])) {
 +        qemu_log_mask(LOG_GUEST_ERROR,
 +            "Device %s says index %u is used, but it was not available",
 +            svq->vdev->name, used_elem.id);
 +        return NULL;
 +    }
 +
 +    descs[used_elem.id].next = svq->free_head;
 +    svq->free_head = used_elem.id;
 +
 +    *len = used_elem.len;
 +    return g_steal_pointer(&svq->ring_id_maps[used_elem.id]);
 +}
 +
 +static void vhost_svq_flush(VhostShadowVirtqueue *svq,
 +                            bool check_for_avail_queue)
 +{
 +    VirtQueue *vq = svq->vq;
 +
 +    /* Forward as many used buffers as possible. */
 +    do {
 +        unsigned i = 0;
 +
 +        vhost_svq_disable_notification(svq);
 +        while (true) {
 +            uint32_t len;
 +            g_autofree VirtQueueElement *elem = vhost_svq_get_buf(svq, &len);
 +            if (!elem) {
 +                break;
 +            }
 +
 +            if (unlikely(i >= svq->vring.num)) {
 +                qemu_log_mask(LOG_GUEST_ERROR,
 +                         "More than %u used buffers obtained in a %u size SVQ",
 +                         i, svq->vring.num);
 +                virtqueue_fill(vq, elem, len, i);
 +                virtqueue_flush(vq, i);
 +                return;
 +            }
 +            virtqueue_fill(vq, elem, len, i++);
 +        }
 +
 +        virtqueue_flush(vq, i);
 +        event_notifier_set(&svq->svq_call);
 +
 +        if (check_for_avail_queue && svq->next_guest_avail_elem) {
 +            /*
 +             * Avail ring was full when vhost_svq_flush was called, so it's a
 +             * good moment to make more descriptors available if possible.
 +             */
 +            vhost_handle_guest_kick(svq);
 +        }
 +    } while (!vhost_svq_enable_notification(svq));
 +}
 +
 +/**
 + * Forward used buffers.
   *
   * @n: hdev call event notifier, the one that device set to notify svq.
 + *
 + * Note that we are not making any buffers available in the loop, there is no
 + * way that it runs more than virtqueue size times.
   */
  static void vhost_svq_handle_call(EventNotifier *n)
  {
      VhostShadowVirtqueue *svq = container_of(n, VhostShadowVirtqueue,
                                               hdev_call);
      event_notifier_test_and_clear(n);
 -    event_notifier_set(&svq->svq_call);
 +    vhost_svq_flush(svq, true);
  }
  /**
@@ -XXX,XX +XXX,XX @@ void vhost_svq_set_svq_kick_fd(VhostShadowVirtqueue *svq, int svq_kick_fd)
      if (poll_start) {
          event_notifier_init_fd(svq_kick, svq_kick_fd);
          event_notifier_set(svq_kick);
 -        event_notifier_set_handler(svq_kick, vhost_handle_guest_kick);
 +        event_notifier_set_handler(svq_kick, vhost_handle_guest_kick_notifier);
 +    }
 +}
 +
 +/**
 + * Start the shadow virtqueue operation.
 + *
 + * @svq: Shadow Virtqueue
 + * @vdev: VirtIO device
 + * @vq: Virtqueue to shadow
 + */
 +void vhost_svq_start(VhostShadowVirtqueue *svq, VirtIODevice *vdev,
 +                     VirtQueue *vq)
 +{
 +    size_t desc_size, driver_size, device_size;
 +
 +    svq->next_guest_avail_elem = NULL;
 +    svq->shadow_avail_idx = 0;
 +    svq->shadow_used_idx = 0;
 +    svq->last_used_idx = 0;
 +    svq->vdev = vdev;
 +    svq->vq = vq;
 +
 +    svq->vring.num = virtio_queue_get_num(vdev, virtio_get_queue_index(vq));
 +    driver_size = vhost_svq_driver_area_size(svq);
 +    device_size = vhost_svq_device_area_size(svq);
 +    svq->vring.desc = qemu_memalign(qemu_real_host_page_size, driver_size);
 +    desc_size = sizeof(vring_desc_t) * svq->vring.num;
 +    svq->vring.avail = (void *)((char *)svq->vring.desc + desc_size);
 +    memset(svq->vring.desc, 0, driver_size);
 +    svq->vring.used = qemu_memalign(qemu_real_host_page_size, device_size);
 +    memset(svq->vring.used, 0, device_size);
 +    svq->ring_id_maps = g_new0(VirtQueueElement *, svq->vring.num);
 +    for (unsigned i = 0; i < svq->vring.num - 1; i++) {
 +        svq->vring.desc[i].next = cpu_to_le16(i + 1);
      }
  }
-@@ -XXX,XX +XXX,XX @@ static void do_phy_write(lan9118_state *s, int reg, uint32_t val)
-         phy_update_irq(s);
+@@ -XXX,XX +XXX,XX @@ void vhost_svq_set_svq_kick_fd(VhostShadowVirtqueue *svq, int svq_kick_fd)
-         break;
+ void vhost_svq_stop(VhostShadowVirtqueue *svq)
-     default:
+ {
--        BADF("PHY write reg %d = 0x%04x\n", reg, val);
+     event_notifier_set_handler(&svq->svq_kick, NULL);
-+        qemu_log_mask(LOG_GUEST_ERROR,
++    g_autofree VirtQueueElement *next_avail_elem = NULL;
-+                      "do_phy_write: PHY write reg %d = 0x%04x\n", reg, val);
++
 +    if (!svq->vq) {
 +        return;
 +    }
 +
 +    /* Send all pending used descriptors to guest */
 +    vhost_svq_flush(svq, false);
 +
 +    for (unsigned i = 0; i < svq->vring.num; ++i) {
 +        g_autofree VirtQueueElement *elem = NULL;
 +        elem = g_steal_pointer(&svq->ring_id_maps[i]);
 +        if (elem) {
 +            virtqueue_detach_element(svq->vq, elem, 0);
 +        }
 +    }
 +
 +    next_avail_elem = g_steal_pointer(&svq->next_guest_avail_elem);
 +    if (next_avail_elem) {
 +        virtqueue_detach_element(svq->vq, next_avail_elem, 0);
 +    }
 +    svq->vq = NULL;
 +    g_free(svq->ring_id_maps);
 +    qemu_vfree(svq->vring.desc);
 +    qemu_vfree(svq->vring.used);
  }
  /**
 diff --git a/hw/virtio/vhost-shadow-virtqueue.h b/hw/virtio/vhost-shadow-virtqueue.h
 index XXXXXXX..XXXXXXX 100644
 --- a/hw/virtio/vhost-shadow-virtqueue.h
 +++ b/hw/virtio/vhost-shadow-virtqueue.h
@@ -XXX,XX +XXX,XX @@ typedef struct VhostShadowVirtqueue {
      /* Guest's call notifier, where the SVQ calls guest. */
      EventNotifier svq_call;
 +
 +    /* Virtio queue shadowing */
 +    VirtQueue *vq;
 +
 +    /* Virtio device */
 +    VirtIODevice *vdev;
 +
 +    /* Map for use the guest's descriptors */
 +    VirtQueueElement **ring_id_maps;
 +
 +    /* Next VirtQueue element that guest made available */
 +    VirtQueueElement *next_guest_avail_elem;
 +
 +    /* Next head to expose to the device */
 +    uint16_t shadow_avail_idx;
 +
 +    /* Next free descriptor */
 +    uint16_t free_head;
 +
 +    /* Last seen used idx */
 +    uint16_t shadow_used_idx;
 +
 +    /* Next head to consume from the device */
 +    uint16_t last_used_idx;
  } VhostShadowVirtqueue;
  bool vhost_svq_valid_features(uint64_t features, Error **errp);
@@ -XXX,XX +XXX,XX @@ void vhost_svq_get_vring_addr(const VhostShadowVirtqueue *svq,
  size_t vhost_svq_driver_area_size(const VhostShadowVirtqueue *svq);
  size_t vhost_svq_device_area_size(const VhostShadowVirtqueue *svq);
 +void vhost_svq_start(VhostShadowVirtqueue *svq, VirtIODevice *vdev,
 +                     VirtQueue *vq);
  void vhost_svq_stop(VhostShadowVirtqueue *svq);
  VhostShadowVirtqueue *vhost_svq_new(void);
 diff --git a/hw/virtio/vhost-vdpa.c b/hw/virtio/vhost-vdpa.c
 index XXXXXXX..XXXXXXX 100644
 --- a/hw/virtio/vhost-vdpa.c
 +++ b/hw/virtio/vhost-vdpa.c
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_vring_dev_addr(struct vhost_dev *dev,
   * Note that this function does not rewind kick file descriptor if cannot set
   * call one.
   */
 -static bool vhost_vdpa_svq_setup(struct vhost_dev *dev,
 -                                 VhostShadowVirtqueue *svq, unsigned idx,
 -                                 Error **errp)
 +static int vhost_vdpa_svq_set_fds(struct vhost_dev *dev,
 +                                  VhostShadowVirtqueue *svq, unsigned idx,
 +                                  Error **errp)
  {
      struct vhost_vring_file file = {
          .index = dev->vq_index + idx,
@@ -XXX,XX +XXX,XX @@ static bool vhost_vdpa_svq_setup(struct vhost_dev *dev,
      r = vhost_vdpa_set_vring_dev_kick(dev, &file);
      if (unlikely(r != 0)) {
          error_setg_errno(errp, -r, "Can't set device kick fd");
 -        return false;
 +        return r;
      }
+     event_notifier = &svq->hdev_call;
+@@ -XXX,XX +XXX,XX @@ static bool vhost_vdpa_svq_setup(struct vhost_dev *dev,
+         error_setg_errno(errp, -r, "Can't set device call fd");
+     }
++    return r;
++}
++
++/**
++ * Unmap a SVQ area in the device
++ */
++static bool vhost_vdpa_svq_unmap_ring(struct vhost_vdpa *v, hwaddr iova,
++                                      hwaddr size)
++{
++    int r;
++
++    size = ROUND_UP(size, qemu_real_host_page_size);
++    r = vhost_vdpa_dma_unmap(v, iova, size);
++    return r == 0;
++}
++
++static bool vhost_vdpa_svq_unmap_rings(struct vhost_dev *dev,
++                                       const VhostShadowVirtqueue *svq)
++{
++    struct vhost_vdpa *v = dev->opaque;
++    struct vhost_vring_addr svq_addr;
++    size_t device_size = vhost_svq_device_area_size(svq);
++    size_t driver_size = vhost_svq_driver_area_size(svq);
++    bool ok;
++
++    vhost_svq_get_vring_addr(svq, &svq_addr);
++
++    ok = vhost_vdpa_svq_unmap_ring(v, svq_addr.desc_user_addr, driver_size);
++    if (unlikely(!ok)) {
++        return false;
++    }
++
++    return vhost_vdpa_svq_unmap_ring(v, svq_addr.used_user_addr, device_size);
++}
++
++/**
++ * Map the shadow virtqueue rings in the device
++ *
++ * @dev: The vhost device
++ * @svq: The shadow virtqueue
++ * @addr: Assigned IOVA addresses
++ * @errp: Error pointer
++ */
++static bool vhost_vdpa_svq_map_rings(struct vhost_dev *dev,
++                                     const VhostShadowVirtqueue *svq,
++                                     struct vhost_vring_addr *addr,
++                                     Error **errp)
++{
++    struct vhost_vdpa *v = dev->opaque;
++    size_t device_size = vhost_svq_device_area_size(svq);
++    size_t driver_size = vhost_svq_driver_area_size(svq);
++    int r;
++
++    ERRP_GUARD();
++    vhost_svq_get_vring_addr(svq, addr);
++
++    r = vhost_vdpa_dma_map(v, addr->desc_user_addr, driver_size,
++                           (void *)(uintptr_t)addr->desc_user_addr, true);
++    if (unlikely(r != 0)) {
++        error_setg_errno(errp, -r, "Cannot create vq driver region: ");
++        return false;
++    }
++
++    r = vhost_vdpa_dma_map(v, addr->used_user_addr, device_size,
++                           (void *)(intptr_t)addr->used_user_addr, false);
++    if (unlikely(r != 0)) {
++        error_setg_errno(errp, -r, "Cannot create vq device region: ");
++    }
++
++    return r == 0;
++}
++
++static bool vhost_vdpa_svq_setup(struct vhost_dev *dev,
++                                 VhostShadowVirtqueue *svq, unsigned idx,
++                                 Error **errp)
++{
++    uint16_t vq_index = dev->vq_index + idx;
++    struct vhost_vring_state s = {
++        .index = vq_index,
++    };
++    int r;
++
++    r = vhost_vdpa_set_dev_vring_base(dev, &s);
++    if (unlikely(r)) {
++        error_setg_errno(errp, -r, "Cannot set vring base");
++        return false;
++    }
++
++    r = vhost_vdpa_svq_set_fds(dev, svq, idx, errp);
+     return r == 0;
  }
-@@ -XXX,XX +XXX,XX @@ static void lan9118_16bit_mode_write(void *opaque, hwaddr offset,
+@@ -XXX,XX +XXX,XX @@ static bool vhost_vdpa_svqs_start(struct vhost_dev *dev)
          return;
      }
--    hw_error("lan9118_write: Bad size 0x%x\n", size);
+     for (i = 0; i < v->shadow_vqs->len; ++i) {
-+    qemu_log_mask(LOG_GUEST_ERROR,
++        VirtQueue *vq = virtio_get_queue(dev->vdev, dev->vq_index + i);
-+                  "lan9118_16bit_mode_write: Bad size 0x%x\n", size);
+         VhostShadowVirtqueue *svq = g_ptr_array_index(v->shadow_vqs, i);
- }
++        struct vhost_vring_addr addr = {
++            .index = i,
- static uint64_t lan9118_readl(void *opaque, hwaddr offset,
++        };
-@@ -XXX,XX +XXX,XX @@ static uint64_t lan9118_16bit_mode_read(void *opaque, hwaddr offset,
++        int r;
-         return lan9118_readl(opaque, offset, size);
+         bool ok = vhost_vdpa_svq_setup(dev, svq, i, &err);
          if (unlikely(!ok)) {
 -            error_reportf_err(err, "Cannot setup SVQ %u: ", i);
 +            goto err;
 +        }
 +
 +        vhost_svq_start(svq, dev->vdev, vq);
 +        ok = vhost_vdpa_svq_map_rings(dev, svq, &addr, &err);
 +        if (unlikely(!ok)) {
 +            goto err_map;
 +        }
 +
 +        /* Override vring GPA set by vhost subsystem */
 +        r = vhost_vdpa_set_vring_dev_addr(dev, &addr);
 +        if (unlikely(r != 0)) {
 +            error_setg_errno(&err, -r, "Cannot set device address");
 +            goto err_set_addr;
 +        }
 +    }
 +
 +    return true;
 +
 +err_set_addr:
 +    vhost_vdpa_svq_unmap_rings(dev, g_ptr_array_index(v->shadow_vqs, i));
 +
 +err_map:
 +    vhost_svq_stop(g_ptr_array_index(v->shadow_vqs, i));
 +
 +err:
 +    error_reportf_err(err, "Cannot setup SVQ %u: ", i);
 +    for (unsigned j = 0; j < i; ++j) {
 +        VhostShadowVirtqueue *svq = g_ptr_array_index(v->shadow_vqs, j);
 +        vhost_vdpa_svq_unmap_rings(dev, svq);
 +        vhost_svq_stop(svq);
 +    }
 +
 +    return false;
 +}
 +
 +static bool vhost_vdpa_svqs_stop(struct vhost_dev *dev)
 +{
 +    struct vhost_vdpa *v = dev->opaque;
 +
 +    if (!v->shadow_vqs) {
 +        return true;
 +    }
 +
 +    for (unsigned i = 0; i < v->shadow_vqs->len; ++i) {
 +        VhostShadowVirtqueue *svq = g_ptr_array_index(v->shadow_vqs, i);
 +        bool ok = vhost_vdpa_svq_unmap_rings(dev, svq);
 +        if (unlikely(!ok)) {
              return false;
          }
      }
+@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_dev_start(struct vhost_dev *dev, bool started)
--    hw_error("lan9118_read: Bad size 0x%x\n", size);
+         }
-+    qemu_log_mask(LOG_GUEST_ERROR,
+         vhost_vdpa_set_vring_ready(dev);
-+                  "lan9118_16bit_mode_read: Bad size 0x%x\n", size);
+     } else {
-     return 0;
++        ok = vhost_vdpa_svqs_stop(dev);
- }
++        if (unlikely(!ok)) {
 +            return -1;
 +        }
          vhost_vdpa_host_notifiers_uninit(dev, dev->nvqs);
      }
 --
 .7.4

-[PULL 06/11] net: Increase L2TPv3 buffer to fit jumboframes
+[PULL V3 09/15] util: Add iova_tree_alloc_map
-From: Christian Svensson <blue@cmd.nu>
+From: Eugenio Pérez <eperezma@redhat.com>
-Increase the allocated buffer size to fit larger packets.
+This iova tree function allows it to look for a hole in allocated
-Given that jumboframes can commonly be up to 9000 bytes the closest suitable
+regions and return a totally new translation for a given translated
-value seems to be 16 KiB.
+address.
-Tested by running qemu towards a Linux L2TPv3 endpoint and pushing
+It's usage is mainly to allow devices to access qemu address space,
-jumboframe traffic through the interfaces.
+remapping guest's one into a new iova space where qemu can add chunks of
+addresses.
-Signed-off-by: Christian Svensson <blue@cmd.nu>
 Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
 Reviewed-by: Peter Xu <peterx@redhat.com>
 Acked-by: Michael S. Tsirkin <mst@redhat.com>
 Signed-off-by: Jason Wang <jasowang@redhat.com>
 ---
- net/l2tpv3.c | 2 +-
+ include/qemu/iova-tree.h |  18 +++++++
-file changed, 1 insertion(+), 1 deletion(-)
+ util/iova-tree.c         | 136 +++++++++++++++++++++++++++++++++++++++++++++++
+files changed, 154 insertions(+)
-diff --git a/net/l2tpv3.c b/net/l2tpv3.c
 diff --git a/include/qemu/iova-tree.h b/include/qemu/iova-tree.h
 index XXXXXXX..XXXXXXX 100644
---- a/net/l2tpv3.c
+--- a/include/qemu/iova-tree.h
-+++ b/net/l2tpv3.c
++++ b/include/qemu/iova-tree.h
 @@ -XXX,XX +XXX,XX @@
-  */
+ #define  IOVA_OK           (0)
+ #define  IOVA_ERR_INVALID  (-1) /* Invalid parameters */
- #define BUFFER_ALIGN sysconf(_SC_PAGESIZE)
+ #define  IOVA_ERR_OVERLAP  (-2) /* IOVA range overlapped */
--#define BUFFER_SIZE 2048
++#define  IOVA_ERR_NOMEM    (-3) /* Cannot allocate */
-+#define BUFFER_SIZE 16384
- #define IOVSIZE 2
+ typedef struct IOVATree IOVATree;
- #define MAX_L2TPV3_MSGCNT 64
+ typedef struct DMAMap {
- #define MAX_L2TPV3_IOVCNT (MAX_L2TPV3_MSGCNT * IOVSIZE)
+@@ -XXX,XX +XXX,XX @@ const DMAMap *iova_tree_find_address(const IOVATree *tree, hwaddr iova);
  void iova_tree_foreach(IOVATree *tree, iova_tree_iterator iterator);
  /**
 + * iova_tree_alloc_map:
 + *
 + * @tree: the iova tree to allocate from
 + * @map: the new map (as translated addr & size) to allocate in the iova region
 + * @iova_begin: the minimum address of the allocation
 + * @iova_end: the maximum addressable direction of the allocation
 + *
 + * Allocates a new region of a given size, between iova_min and iova_max.
 + *
 + * Return: Same as iova_tree_insert, but cannot overlap and can return error if
 + * iova tree is out of free contiguous range. The caller gets the assigned iova
 + * in map->iova.
 + */
 +int iova_tree_alloc_map(IOVATree *tree, DMAMap *map, hwaddr iova_begin,
 +                        hwaddr iova_end);
 +
 +/**
   * iova_tree_destroy:
   *
   * @tree: the iova tree to destroy
 diff --git a/util/iova-tree.c b/util/iova-tree.c
 index XXXXXXX..XXXXXXX 100644
 --- a/util/iova-tree.c
 +++ b/util/iova-tree.c
@@ -XXX,XX +XXX,XX @@ struct IOVATree {
      GTree *tree;
  };
 +/* Args to pass to iova_tree_alloc foreach function. */
 +struct IOVATreeAllocArgs {
 +    /* Size of the desired allocation */
 +    size_t new_size;
 +
 +    /* The minimum address allowed in the allocation */
 +    hwaddr iova_begin;
 +
 +    /* Map at the left of the hole, can be NULL if "this" is first one */
 +    const DMAMap *prev;
 +
 +    /* Map at the right of the hole, can be NULL if "prev" is the last one */
 +    const DMAMap *this;
 +
 +    /* If found, we fill in the IOVA here */
 +    hwaddr iova_result;
 +
 +    /* Whether have we found a valid IOVA */
 +    bool iova_found;
 +};
 +
 +/**
 + * Iterate args to the next hole
 + *
 + * @args: The alloc arguments
 + * @next: The next mapping in the tree. Can be NULL to signal the last one
 + */
 +static void iova_tree_alloc_args_iterate(struct IOVATreeAllocArgs *args,
 +                                         const DMAMap *next)
 +{
 +    args->prev = args->this;
 +    args->this = next;
 +}
 +
  static int iova_tree_compare(gconstpointer a, gconstpointer b, gpointer data)
  {
      const DMAMap *m1 = a, *m2 = b;
@@ -XXX,XX +XXX,XX @@ int iova_tree_remove(IOVATree *tree, const DMAMap *map)
      return IOVA_OK;
  }
 +/**
 + * Try to find an unallocated IOVA range between prev and this elements.
 + *
 + * @args: Arguments to allocation
 + *
 + * Cases:
 + *
 + * (1) !prev, !this: No entries allocated, always succeed
 + *
 + * (2) !prev, this: We're iterating at the 1st element.
 + *
 + * (3) prev, !this: We're iterating at the last element.
 + *
 + * (4) prev, this: this is the most common case, we'll try to find a hole
 + * between "prev" and "this" mapping.
 + *
 + * Note that this function assumes the last valid iova is HWADDR_MAX, but it
 + * searches linearly so it's easy to discard the result if it's not the case.
 + */
 +static void iova_tree_alloc_map_in_hole(struct IOVATreeAllocArgs *args)
 +{
 +    const DMAMap *prev = args->prev, *this = args->this;
 +    uint64_t hole_start, hole_last;
 +
 +    if (this && this->iova + this->size < args->iova_begin) {
 +        return;
 +    }
 +
 +    hole_start = MAX(prev ? prev->iova + prev->size + 1 : 0, args->iova_begin);
 +    hole_last = this ? this->iova : HWADDR_MAX;
 +
 +    if (hole_last - hole_start > args->new_size) {
 +        args->iova_result = hole_start;
 +        args->iova_found = true;
 +    }
 +}
 +
 +/**
 + * Foreach dma node in the tree, compare if there is a hole with its previous
 + * node (or minimum iova address allowed) and the node.
 + *
 + * @key: Node iterating
 + * @value: Node iterating
 + * @pargs: Struct to communicate with the outside world
 + *
 + * Return: false to keep iterating, true if needs break.
 + */
 +static gboolean iova_tree_alloc_traverse(gpointer key, gpointer value,
 +                                         gpointer pargs)
 +{
 +    struct IOVATreeAllocArgs *args = pargs;
 +    DMAMap *node = value;
 +
 +    assert(key == value);
 +
 +    iova_tree_alloc_args_iterate(args, node);
 +    iova_tree_alloc_map_in_hole(args);
 +    return args->iova_found;
 +}
 +
 +int iova_tree_alloc_map(IOVATree *tree, DMAMap *map, hwaddr iova_begin,
 +                        hwaddr iova_last)
 +{
 +    struct IOVATreeAllocArgs args = {
 +        .new_size = map->size,
 +        .iova_begin = iova_begin,
 +    };
 +
 +    if (unlikely(iova_last < iova_begin)) {
 +        return IOVA_ERR_INVALID;
 +    }
 +
 +    /*
 +     * Find a valid hole for the mapping
 +     *
 +     * Assuming low iova_begin, so no need to do a binary search to
 +     * locate the first node.
 +     *
 +     * TODO: Replace all this with g_tree_node_first/next/last when available
 +     * (from glib since 2.68). To do it with g_tree_foreach complicates the
 +     * code a lot.
 +     *
 +     */
 +    g_tree_foreach(tree->tree, iova_tree_alloc_traverse, &args);
 +    if (!args.iova_found) {
 +        /*
 +         * Either tree is empty or the last hole is still not checked.
 +         * g_tree_foreach does not compare (last, iova_last] range, so we check
 +         * it here.
 +         */
 +        iova_tree_alloc_args_iterate(&args, NULL);
 +        iova_tree_alloc_map_in_hole(&args);
 +    }
 +
 +    if (!args.iova_found || args.iova_result + map->size > iova_last) {
 +        return IOVA_ERR_NOMEM;
 +    }
 +
 +    map->iova = args.iova_result;
 +    return iova_tree_insert(tree, map);
 +}
 +
  void iova_tree_destroy(IOVATree *tree)
  {
      g_tree_destroy(tree->tree);
 --
 .7.4

-[PULL 01/11] net: Move the code to collect available NIC models to a separate function
+[PULL V3 10/15] util: add iova_tree_find_iova
-From: Thomas Huth <thuth@redhat.com>
+From: Eugenio Pérez <eperezma@redhat.com>
-The code that collects the available NIC models is not really specific
+This function does the reverse operation of iova_tree_find: To look for
-to PCI anymore and will be required in the next patch, too, so let's
+a mapping that match a translated address so we can do the reverse.
 move this into a new separate function in net.c instead.
-Signed-off-by: Thomas Huth <thuth@redhat.com>
+This have linear complexity instead of logarithmic, but it supports
 overlapping HVA. Future developments could reduce it.
 Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
 Acked-by: Michael S. Tsirkin <mst@redhat.com>
 Signed-off-by: Jason Wang <jasowang@redhat.com>
 ---
- hw/pci/pci.c      | 29 +----------------------------
+ include/qemu/iova-tree.h | 20 +++++++++++++++++++-
- include/net/net.h | 14 ++++++++++++++
+ util/iova-tree.c         | 34 ++++++++++++++++++++++++++++++++++
- net/net.c         | 34 ++++++++++++++++++++++++++++++++++
+files changed, 53 insertions(+), 1 deletion(-)
 files changed, 49 insertions(+), 28 deletions(-)
-diff --git a/hw/pci/pci.c b/hw/pci/pci.c
+diff --git a/include/qemu/iova-tree.h b/include/qemu/iova-tree.h
 index XXXXXXX..XXXXXXX 100644
---- a/hw/pci/pci.c
+--- a/include/qemu/iova-tree.h
-+++ b/hw/pci/pci.c
++++ b/include/qemu/iova-tree.h
-@@ -XXX,XX +XXX,XX @@ PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus,
+@@ -XXX,XX +XXX,XX @@ int iova_tree_remove(IOVATree *tree, const DMAMap *map);
-                                const char *default_devaddr)
+  * @tree: the iova tree to search from
- {
+  * @map: the mapping to search
-     const char *devaddr = nd->devaddr ? nd->devaddr : default_devaddr;
+  *
--    GSList *list;
+- * Search for a mapping in the iova tree that overlaps with the
-     GPtrArray *pci_nic_models;
++ * Search for a mapping in the iova tree that iova overlaps with the
-     PCIBus *bus;
+  * mapping range specified.  Only the first found mapping will be
-     PCIDevice *pci_dev;
+  * returned.
-@@ -XXX,XX +XXX,XX @@ PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus,
+  *
-         nd->model = g_strdup("virtio-net-pci");
+@@ -XXX,XX +XXX,XX @@ int iova_tree_remove(IOVATree *tree, const DMAMap *map);
-     }
+ const DMAMap *iova_tree_find(const IOVATree *tree, const DMAMap *map);
--    list = object_class_get_list_sorted(TYPE_PCI_DEVICE, false);
+ /**
--    pci_nic_models = g_ptr_array_new();
++ * iova_tree_find_iova:
--    while (list) {
++ *
--        DeviceClass *dc = OBJECT_CLASS_CHECK(DeviceClass, list->data,
++ * @tree: the iova tree to search from
--                                             TYPE_DEVICE);
++ * @map: the mapping to search
--        GSList *next;
++ *
--        if (test_bit(DEVICE_CATEGORY_NETWORK, dc->categories) &&
++ * Search for a mapping in the iova tree that translated_addr overlaps with the
--            dc->user_creatable) {
++ * mapping range specified.  Only the first found mapping will be
--            const char *name = object_class_get_name(list->data);
++ * returned.
--            /*
++ *
--             * A network device might also be something else than a NIC, see
++ * Return: DMAMap pointer if found, or NULL if not found.  Note that
--             * e.g. the "rocker" device. Thus we have to look for the "netdev"
++ * the returned DMAMap pointer is maintained internally.  User should
--             * property, too. Unfortunately, some devices like virtio-net only
++ * only read the content but never modify or free the content.  Also,
--             * create this property during instance_init, so we have to create
++ * user is responsible to make sure the pointer is valid (say, no
--             * a temporary instance here to be able to check it.
++ * concurrent deletion in progress).
--             */
++ */
--            Object *obj = object_new_with_class(OBJECT_CLASS(dc));
++const DMAMap *iova_tree_find_iova(const IOVATree *tree, const DMAMap *map);
--            if (object_property_find(obj, "netdev")) {
++
--                g_ptr_array_add(pci_nic_models, (gpointer)name);
++/**
--            }
+  * iova_tree_find_address:
--            object_unref(obj);
+  *
--        }
+  * @tree: the iova tree to search from
--        next = list->next;
+diff --git a/util/iova-tree.c b/util/iova-tree.c
 -        g_slist_free_1(list);
 -        list = next;
 -    }
 -    g_ptr_array_add(pci_nic_models, NULL);
 +    pci_nic_models = qemu_get_nic_models(TYPE_PCI_DEVICE);
      if (qemu_show_nic_models(nd->model, (const char **)pci_nic_models->pdata)) {
          exit(0);
 diff --git a/include/net/net.h b/include/net/net.h
 index XXXXXXX..XXXXXXX 100644
---- a/include/net/net.h
+--- a/util/iova-tree.c
-+++ b/include/net/net.h
++++ b/util/iova-tree.c
-@@ -XXX,XX +XXX,XX @@ void net_socket_rs_init(SocketReadState *rs,
+@@ -XXX,XX +XXX,XX @@ struct IOVATreeAllocArgs {
-                         bool vnet_hdr);
+     bool iova_found;
- NetClientState *qemu_get_peer(NetClientState *nc, int queue_index);
+ };
-+/**
++typedef struct IOVATreeFindIOVAArgs {
-+ * qemu_get_nic_models:
++    const DMAMap *needle;
-+ * @device_type: Defines which devices should be taken into consideration
++    const DMAMap *result;
-+ *               (e.g. TYPE_DEVICE for all devices, or TYPE_PCI_DEVICE for PCI)
++} IOVATreeFindIOVAArgs;
 + *
 + * Get an array of pointers to names of NIC devices that are available in
 + * the QEMU binary. The array is terminated with a NULL pointer entry.
 + * The caller is responsible for freeing the memory when it is not required
 + * anymore, e.g. with g_ptr_array_free(..., true).
 + *
 + * Returns: Pointer to the array that contains the pointers to the names.
 + */
 +GPtrArray *qemu_get_nic_models(const char *device_type);
 +
- /* NIC info */
+ /**
+  * Iterate args to the next hole
- #define MAX_NICS 8
+  *
-diff --git a/net/net.c b/net/net.c
+@@ -XXX,XX +XXX,XX @@ const DMAMap *iova_tree_find(const IOVATree *tree, const DMAMap *map)
-index XXXXXXX..XXXXXXX 100644
+     return g_tree_lookup(tree->tree, map);
 --- a/net/net.c
 +++ b/net/net.c
@@ -XXX,XX +XXX,XX @@ static int nic_get_free_idx(void)
      return -1;
  }
-+GPtrArray *qemu_get_nic_models(const char *device_type)
++static gboolean iova_tree_find_address_iterator(gpointer key, gpointer value,
 +                                                gpointer data)
 +{
-+    GPtrArray *nic_models = g_ptr_array_new();
++    const DMAMap *map = key;
-+    GSList *list = object_class_get_list_sorted(device_type, false);
++    IOVATreeFindIOVAArgs *args = data;
 +    const DMAMap *needle;
 +
-+    while (list) {
++    g_assert(key == value);
-+        DeviceClass *dc = OBJECT_CLASS_CHECK(DeviceClass, list->data,
++
-+                                             TYPE_DEVICE);
++    needle = args->needle;
-+        GSList *next;
++    if (map->translated_addr + map->size < needle->translated_addr ||
-+        if (test_bit(DEVICE_CATEGORY_NETWORK, dc->categories) &&
++        needle->translated_addr + needle->size < map->translated_addr) {
-+            dc->user_creatable) {
++        return false;
 +            const char *name = object_class_get_name(list->data);
 +            /*
 +             * A network device might also be something else than a NIC, see
 +             * e.g. the "rocker" device. Thus we have to look for the "netdev"
 +             * property, too. Unfortunately, some devices like virtio-net only
 +             * create this property during instance_init, so we have to create
 +             * a temporary instance here to be able to check it.
 +             */
 +            Object *obj = object_new_with_class(OBJECT_CLASS(dc));
 +            if (object_property_find(obj, "netdev")) {
 +                g_ptr_array_add(nic_models, (gpointer)name);
 +            }
 +            object_unref(obj);
 +        }
 +        next = list->next;
 +        g_slist_free_1(list);
 +        list = next;
 +    }
-+    g_ptr_array_add(nic_models, NULL);
 +
-+    return nic_models;
++    args->result = map;
 +    return true;
 +}
 +
- int qemu_show_nic_models(const char *arg, const char *const *models)
++const DMAMap *iova_tree_find_iova(const IOVATree *tree, const DMAMap *map)
 +{
 +    IOVATreeFindIOVAArgs args = {
 +        .needle = map,
 +    };
 +
 +    g_tree_foreach(tree->tree, iova_tree_find_address_iterator, &args);
 +    return args.result;
 +}
 +
  const DMAMap *iova_tree_find_address(const IOVATree *tree, hwaddr iova)
  {
-     int i;
+     const DMAMap map = { .iova = iova, .size = 0 };
 --
 .7.4

-[PULL 10/11] net: stream: add a new option to automatically reconnect
+[PULL V3 11/15] vhost: Add VhostIOVATree
-From: Laurent Vivier <lvivier@redhat.com>
+From: Eugenio Pérez <eperezma@redhat.com>
-In stream mode, if the server shuts down there is currently
+This tree is able to look for a translated address from an IOVA address.
 no way to reconnect the client to a new server without removing
 the NIC device and the netdev backend (or to reboot).
-This patch introduces a reconnect option that specifies a delay
+At first glance it is similar to util/iova-tree. However, SVQ working on
-to try to reconnect with the same parameters.
+devices with limited IOVA space need more capabilities, like allocating
 IOVA chunks or performing reverse translations (qemu addresses to iova).
-Add a new test in qtest to test the reconnect option and the
+The allocation capability, as "assign a free IOVA address to this chunk
-connect/disconnect events.
+of memory in qemu's address space" allows shadow virtqueue to create a
 new address space that is not restricted by guest's addressable one, so
 we can allocate shadow vqs vrings outside of it.
-Signed-off-by: Laurent Vivier <lvivier@redhat.com>
+It duplicates the tree so it can search efficiently in both directions,
 and it will signal overlap if iova or the translated address is present
 in any tree.
 Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
 Acked-by: Michael S. Tsirkin <mst@redhat.com>
 Signed-off-by: Jason Wang <jasowang@redhat.com>
 ---
- net/stream.c                |  53 ++++++++++++++++++++++-
+ hw/virtio/meson.build       |   2 +-
- qapi/net.json               |   7 ++-
+ hw/virtio/vhost-iova-tree.c | 110 ++++++++++++++++++++++++++++++++++++++++++++
- qemu-options.hx             |   6 +--
+ hw/virtio/vhost-iova-tree.h |  27 +++++++++++
- tests/qtest/netdev-socket.c | 101 ++++++++++++++++++++++++++++++++++++++++++++
+files changed, 138 insertions(+), 1 deletion(-)
-files changed, 162 insertions(+), 5 deletions(-)
+ create mode 100644 hw/virtio/vhost-iova-tree.c
  create mode 100644 hw/virtio/vhost-iova-tree.h
-diff --git a/net/stream.c b/net/stream.c
+diff --git a/hw/virtio/meson.build b/hw/virtio/meson.build
 index XXXXXXX..XXXXXXX 100644
---- a/net/stream.c
+--- a/hw/virtio/meson.build
-+++ b/net/stream.c
++++ b/hw/virtio/meson.build
@@ -XXX,XX +XXX,XX @@ softmmu_ss.add(when: 'CONFIG_ALL', if_true: files('vhost-stub.c'))
  virtio_ss = ss.source_set()
  virtio_ss.add(files('virtio.c'))
 -virtio_ss.add(when: 'CONFIG_VHOST', if_true: files('vhost.c', 'vhost-backend.c', 'vhost-shadow-virtqueue.c'))
 +virtio_ss.add(when: 'CONFIG_VHOST', if_true: files('vhost.c', 'vhost-backend.c', 'vhost-shadow-virtqueue.c', 'vhost-iova-tree.c'))
  virtio_ss.add(when: 'CONFIG_VHOST_USER', if_true: files('vhost-user.c'))
  virtio_ss.add(when: 'CONFIG_VHOST_VDPA', if_true: files('vhost-vdpa.c'))
  virtio_ss.add(when: 'CONFIG_VIRTIO_BALLOON', if_true: files('virtio-balloon.c'))
 diff --git a/hw/virtio/vhost-iova-tree.c b/hw/virtio/vhost-iova-tree.c
 new file mode 100644
 index XXXXXXX..XXXXXXX
 --- /dev/null
 +++ b/hw/virtio/vhost-iova-tree.c
 @@ -XXX,XX +XXX,XX @@
- #include "io/channel-socket.h"
++/*
- #include "io/net-listener.h"
++ * vhost software live migration iova tree
- #include "qapi/qapi-events-net.h"
++ *
-+#include "qapi/qapi-visit-sockets.h"
++ * SPDX-FileCopyrightText: Red Hat, Inc. 2021
-+#include "qapi/clone-visitor.h"
++ * SPDX-FileContributor: Author: Eugenio Pérez <eperezma@redhat.com>
++ *
- typedef struct NetStreamState {
++ * SPDX-License-Identifier: GPL-2.0-or-later
-     NetClientState nc;
++ */
-@@ -XXX,XX +XXX,XX @@ typedef struct NetStreamState {
++
-     guint ioc_write_tag;
++#include "qemu/osdep.h"
-     SocketReadState rs;
++#include "qemu/iova-tree.h"
-     unsigned int send_index;      /* number of bytes sent*/
++#include "vhost-iova-tree.h"
-+    uint32_t reconnect;
++
-+    guint timer_tag;
++#define iova_min_addr qemu_real_host_page_size
-+    SocketAddress *addr;
++
- } NetStreamState;
++/**
++ * VhostIOVATree, able to:
- static void net_stream_listen(QIONetListener *listener,
++ * - Translate iova address
-                               QIOChannelSocket *cioc,
++ * - Reverse translate iova address (from translated to iova)
-                               void *opaque);
++ * - Allocate IOVA regions for translated range (linear operation)
-+static void net_stream_arm_reconnect(NetStreamState *s);
++ */
++struct VhostIOVATree {
- static gboolean net_stream_writable(QIOChannel *ioc,
++    /* First addressable iova address in the device */
-                                     GIOCondition condition,
++    uint64_t iova_first;
-@@ -XXX,XX +XXX,XX @@ static gboolean net_stream_send(QIOChannel *ioc,
++
-         qemu_set_info_str(&s->nc, "%s", "");
++    /* Last addressable iova address in the device */
++    uint64_t iova_last;
-         qapi_event_send_netdev_stream_disconnected(s->nc.name);
++
-+        net_stream_arm_reconnect(s);
++    /* IOVA address to qemu memory maps. */
++    IOVATree *iova_taddr_map;
-         return G_SOURCE_REMOVE;
++};
-     }
++
-@@ -XXX,XX +XXX,XX @@ static gboolean net_stream_send(QIOChannel *ioc,
++/**
- static void net_stream_cleanup(NetClientState *nc)
++ * Create a new IOVA tree
- {
++ *
-     NetStreamState *s = DO_UPCAST(NetStreamState, nc, nc);
++ * Returns the new IOVA tree
-+    if (s->timer_tag) {
++ */
-+        g_source_remove(s->timer_tag);
++VhostIOVATree *vhost_iova_tree_new(hwaddr iova_first, hwaddr iova_last)
-+        s->timer_tag = 0;
++{
-+    }
++    VhostIOVATree *tree = g_new(VhostIOVATree, 1);
-+    if (s->addr) {
++
-+        qapi_free_SocketAddress(s->addr);
++    /* Some devices do not like 0 addresses */
-+        s->addr = NULL;
++    tree->iova_first = MAX(iova_first, iova_min_addr);
-+    }
++    tree->iova_last = iova_last;
-     if (s->ioc) {
++
-         if (QIO_CHANNEL_SOCKET(s->ioc)->fd != -1) {
++    tree->iova_taddr_map = iova_tree_new();
-             if (s->ioc_read_tag) {
++    return tree;
@@ -XXX,XX +XXX,XX @@ static void net_stream_client_connected(QIOTask *task, gpointer opaque)
  error:
      object_unref(OBJECT(s->ioc));
      s->ioc = NULL;
 +    net_stream_arm_reconnect(s);
 +}
 +
-+static gboolean net_stream_reconnect(gpointer data)
++/**
 + * Delete an iova tree
 + */
 +void vhost_iova_tree_delete(VhostIOVATree *iova_tree)
 +{
-+    NetStreamState *s = data;
++    iova_tree_destroy(iova_tree->iova_taddr_map);
-+    QIOChannelSocket *sioc;
++    g_free(iova_tree);
 +
 +    s->timer_tag = 0;
 +
 +    sioc = qio_channel_socket_new();
 +    s->ioc = QIO_CHANNEL(sioc);
 +    qio_channel_socket_connect_async(sioc, s->addr,
 +                                     net_stream_client_connected, s,
 +                                     NULL, NULL);
 +    return G_SOURCE_REMOVE;
 +}
 +
-+static void net_stream_arm_reconnect(NetStreamState *s)
++/**
 + * Find the IOVA address stored from a memory address
 + *
 + * @tree: The iova tree
 + * @map: The map with the memory address
 + *
 + * Return the stored mapping, or NULL if not found.
 + */
 +const DMAMap *vhost_iova_tree_find_iova(const VhostIOVATree *tree,
 +                                        const DMAMap *map)
 +{
-+    if (s->reconnect && s->timer_tag == 0) {
++    return iova_tree_find_iova(tree->iova_taddr_map, map);
 +        s->timer_tag = g_timeout_add_seconds(s->reconnect,
 +                                             net_stream_reconnect, s);
 +    }
  }
  static int net_stream_client_init(NetClientState *peer,
                                    const char *model,
                                    const char *name,
                                    SocketAddress *addr,
 +                                  uint32_t reconnect,
                                    Error **errp)
  {
      NetStreamState *s;
@@ -XXX,XX +XXX,XX @@ static int net_stream_client_init(NetClientState *peer,
      s->ioc = QIO_CHANNEL(sioc);
      s->nc.link_down = true;
 +    s->reconnect = reconnect;
 +    if (reconnect) {
 +        s->addr = QAPI_CLONE(SocketAddress, addr);
 +    }
      qio_channel_socket_connect_async(sioc, addr,
                                       net_stream_client_connected, s,
                                       NULL, NULL);
@@ -XXX,XX +XXX,XX @@ int net_init_stream(const Netdev *netdev, const char *name,
      sock = &netdev->u.stream;
      if (!sock->has_server || !sock->server) {
 -        return net_stream_client_init(peer, "stream", name, sock->addr, errp);
 +        return net_stream_client_init(peer, "stream", name, sock->addr,
 +                                      sock->has_reconnect ? sock->reconnect : 0,
 +                                      errp);
 +    }
 +    if (sock->has_reconnect) {
 +        error_setg(errp, "'reconnect' option is incompatible with "
 +                         "socket in server mode");
 +        return -1;
      }
      return net_stream_server_init(peer, "stream", name, sock->addr, errp);
  }
 diff --git a/qapi/net.json b/qapi/net.json
 index XXXXXXX..XXXXXXX 100644
 --- a/qapi/net.json
 +++ b/qapi/net.json
@@ -XXX,XX +XXX,XX @@
  # @addr: socket address to listen on (server=true)
  #        or connect to (server=false)
  # @server: create server socket (default: false)
 +# @reconnect: For a client socket, if a socket is disconnected,
 +#             then attempt a reconnect after the given number of seconds.
 +#             Setting this to zero disables this function. (default: 0)
 +#             (since 8.0)
  #
  # Only SocketAddress types 'unix', 'inet' and 'fd' are supported.
  #
@@ -XXX,XX +XXX,XX @@
  { 'struct': 'NetdevStreamOptions',
    'data': {
      'addr':   'SocketAddress',
 -    '*server': 'bool' } }
 +    '*server': 'bool',
 +    '*reconnect': 'uint32' } }
  ##
  # @NetdevDgramOptions:
 diff --git a/qemu-options.hx b/qemu-options.hx
 index XXXXXXX..XXXXXXX 100644
 --- a/qemu-options.hx
 +++ b/qemu-options.hx
@@ -XXX,XX +XXX,XX @@ DEF("netdev", HAS_ARG, QEMU_OPTION_netdev,
      "-netdev socket,id=str[,fd=h][,udp=host:port][,localaddr=host:port]\n"
      "                configure a network backend to connect to another network\n"
      "                using an UDP tunnel\n"
 -    "-netdev stream,id=str[,server=on|off],addr.type=inet,addr.host=host,addr.port=port[,to=maxport][,numeric=on|off][,keep-alive=on|off][,mptcp=on|off][,addr.ipv4=on|off][,addr.ipv6=on|off]\n"
 -    "-netdev stream,id=str[,server=on|off],addr.type=unix,addr.path=path[,abstract=on|off][,tight=on|off]\n"
 -    "-netdev stream,id=str[,server=on|off],addr.type=fd,addr.str=file-descriptor\n"
 +    "-netdev stream,id=str[,server=on|off],addr.type=inet,addr.host=host,addr.port=port[,to=maxport][,numeric=on|off][,keep-alive=on|off][,mptcp=on|off][,addr.ipv4=on|off][,addr.ipv6=on|off][,reconnect=seconds]\n"
 +    "-netdev stream,id=str[,server=on|off],addr.type=unix,addr.path=path[,abstract=on|off][,tight=on|off][,reconnect=seconds]\n"
 +    "-netdev stream,id=str[,server=on|off],addr.type=fd,addr.str=file-descriptor[,reconnect=seconds]\n"
      "                configure a network backend to connect to another network\n"
      "                using a socket connection in stream mode.\n"
      "-netdev dgram,id=str,remote.type=inet,remote.host=maddr,remote.port=port[,local.type=inet,local.host=addr]\n"
 diff --git a/tests/qtest/netdev-socket.c b/tests/qtest/netdev-socket.c
 index XXXXXXX..XXXXXXX 100644
 --- a/tests/qtest/netdev-socket.c
 +++ b/tests/qtest/netdev-socket.c
@@ -XXX,XX +XXX,XX @@
  #include <glib/gstdio.h>
  #include "../unit/socket-helpers.h"
  #include "libqtest.h"
 +#include "qapi/qmp/qstring.h"
 +#include "qemu/sockets.h"
 +#include "qapi/qobject-input-visitor.h"
 +#include "qapi/qapi-visit-sockets.h"
  #define CONNECTION_TIMEOUT    5
@@ -XXX,XX +XXX,XX @@ static void test_stream_inet_ipv4(void)
      qtest_quit(qts0);
  }
 +static void wait_stream_connected(QTestState *qts, const char *id,
 +                                  SocketAddress **addr)
 +{
 +    QDict *resp, *data;
 +    QString *qstr;
 +    QObject *obj;
 +    Visitor *v = NULL;
 +
 +    resp = qtest_qmp_eventwait_ref(qts, "NETDEV_STREAM_CONNECTED");
 +    g_assert_nonnull(resp);
 +    data = qdict_get_qdict(resp, "data");
 +    g_assert_nonnull(data);
 +
 +    qstr = qobject_to(QString, qdict_get(data, "netdev-id"));
 +    g_assert_nonnull(data);
 +
 +    g_assert(!strcmp(qstring_get_str(qstr), id));
 +
 +    obj = qdict_get(data, "addr");
 +
 +    v = qobject_input_visitor_new(obj);
 +    visit_type_SocketAddress(v, NULL, addr, NULL);
 +    visit_free(v);
 +    qobject_unref(resp);
 +}
 +
-+static void wait_stream_disconnected(QTestState *qts, const char *id)
++/**
 + * Allocate a new mapping
 + *
 + * @tree: The iova tree
 + * @map: The iova map
 + *
 + * Returns:
 + * - IOVA_OK if the map fits in the container
 + * - IOVA_ERR_INVALID if the map does not make sense (like size overflow)
 + * - IOVA_ERR_NOMEM if tree cannot allocate more space.
 + *
 + * It returns assignated iova in map->iova if return value is VHOST_DMA_MAP_OK.
 + */
 +int vhost_iova_tree_map_alloc(VhostIOVATree *tree, DMAMap *map)
 +{
-+    QDict *resp, *data;
++    /* Some vhost devices do not like addr 0. Skip first page */
-+    QString *qstr;
++    hwaddr iova_first = tree->iova_first ?: qemu_real_host_page_size;
 +
-+    resp = qtest_qmp_eventwait_ref(qts, "NETDEV_STREAM_DISCONNECTED");
++    if (map->translated_addr + map->size < map->translated_addr ||
-+    g_assert_nonnull(resp);
++        map->perm == IOMMU_NONE) {
-+    data = qdict_get_qdict(resp, "data");
++        return IOVA_ERR_INVALID;
-+    g_assert_nonnull(data);
++    }
 +
-+    qstr = qobject_to(QString, qdict_get(data, "netdev-id"));
++    /* Allocate a node in IOVA address */
-+    g_assert_nonnull(data);
++    return iova_tree_alloc_map(tree->iova_taddr_map, map, iova_first,
-+
++                               tree->iova_last);
 +    g_assert(!strcmp(qstring_get_str(qstr), id));
 +    qobject_unref(resp);
 +}
 +
-+static void test_stream_inet_reconnect(void)
++/**
 + * Remove existing mappings from iova tree
 + *
 + * @iova_tree: The vhost iova tree
 + * @map: The map to remove
 + */
 +void vhost_iova_tree_remove(VhostIOVATree *iova_tree, const DMAMap *map)
 +{
-+    QTestState *qts0, *qts1;
++    iova_tree_remove(iova_tree->iova_taddr_map, map);
-+    int port;
++}
-+    SocketAddress *addr;
+diff --git a/hw/virtio/vhost-iova-tree.h b/hw/virtio/vhost-iova-tree.h
 new file mode 100644
 index XXXXXXX..XXXXXXX
 --- /dev/null
 +++ b/hw/virtio/vhost-iova-tree.h
@@ -XXX,XX +XXX,XX @@
 +/*
 + * vhost software live migration iova tree
 + *
 + * SPDX-FileCopyrightText: Red Hat, Inc. 2021
 + * SPDX-FileContributor: Author: Eugenio Pérez <eperezma@redhat.com>
 + *
 + * SPDX-License-Identifier: GPL-2.0-or-later
 + */
 +
-+    port = inet_get_free_port(false);
++#ifndef HW_VIRTIO_VHOST_IOVA_TREE_H
-+    qts0 = qtest_initf("-nodefaults -M none "
++#define HW_VIRTIO_VHOST_IOVA_TREE_H
 +                       "-netdev stream,id=st0,server=true,addr.type=inet,"
 +                       "addr.ipv4=on,addr.ipv6=off,"
 +                       "addr.host=127.0.0.1,addr.port=%d", port);
 +
-+    EXPECT_STATE(qts0, "st0: index=0,type=stream,\r\n", 0);
++#include "qemu/iova-tree.h"
 +#include "exec/memory.h"
 +
-+    qts1 = qtest_initf("-nodefaults -M none "
++typedef struct VhostIOVATree VhostIOVATree;
 +                       "-netdev stream,server=false,id=st0,addr.type=inet,"
 +                       "addr.ipv4=on,addr.ipv6=off,reconnect=1,"
 +                       "addr.host=127.0.0.1,addr.port=%d", port);
 +
-+    wait_stream_connected(qts0, "st0", &addr);
++VhostIOVATree *vhost_iova_tree_new(uint64_t iova_first, uint64_t iova_last);
-+    g_assert_cmpint(addr->type, ==, SOCKET_ADDRESS_TYPE_INET);
++void vhost_iova_tree_delete(VhostIOVATree *iova_tree);
-+    g_assert_cmpstr(addr->u.inet.host, ==, "127.0.0.1");
++G_DEFINE_AUTOPTR_CLEANUP_FUNC(VhostIOVATree, vhost_iova_tree_delete);
 +    qapi_free_SocketAddress(addr);
 +
-+    /* kill server */
++const DMAMap *vhost_iova_tree_find_iova(const VhostIOVATree *iova_tree,
-+    qtest_quit(qts0);
++                                        const DMAMap *map);
 +int vhost_iova_tree_map_alloc(VhostIOVATree *iova_tree, DMAMap *map);
 +void vhost_iova_tree_remove(VhostIOVATree *iova_tree, const DMAMap *map);
 +
-+    /* check client has been disconnected */
++#endif
 +    wait_stream_disconnected(qts1, "st0");
 +
 +    /* restart server */
 +    qts0 = qtest_initf("-nodefaults -M none "
 +                       "-netdev stream,id=st0,server=true,addr.type=inet,"
 +                       "addr.ipv4=on,addr.ipv6=off,"
 +                       "addr.host=127.0.0.1,addr.port=%d", port);
 +
 +    /* wait connection events*/
 +    wait_stream_connected(qts0, "st0", &addr);
 +    g_assert_cmpint(addr->type, ==, SOCKET_ADDRESS_TYPE_INET);
 +    g_assert_cmpstr(addr->u.inet.host, ==, "127.0.0.1");
 +    qapi_free_SocketAddress(addr);
 +
 +    wait_stream_connected(qts1, "st0", &addr);
 +    g_assert_cmpint(addr->type, ==, SOCKET_ADDRESS_TYPE_INET);
 +    g_assert_cmpstr(addr->u.inet.host, ==, "127.0.0.1");
 +    g_assert_cmpint(atoi(addr->u.inet.port), ==, port);
 +    qapi_free_SocketAddress(addr);
 +
 +    qtest_quit(qts1);
 +    qtest_quit(qts0);
 +}
 +
  static void test_stream_inet_ipv6(void)
  {
      QTestState *qts0, *qts1;
@@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv)
  #ifndef _WIN32
          qtest_add_func("/netdev/dgram/mcast", test_dgram_mcast);
  #endif
 +        qtest_add_func("/netdev/stream/inet/reconnect",
 +                       test_stream_inet_reconnect);
      }
      if (has_ipv6) {
          qtest_add_func("/netdev/stream/inet/ipv6", test_stream_inet_ipv6);
 --
 .7.4

-[PULL 02/11] net: Restore printing of the help text with "-nic help"
+[PULL V3 12/15] vdpa: Add custom IOTLB translations to SVQ
-From: Thomas Huth <thuth@redhat.com>
+From: Eugenio Pérez <eperezma@redhat.com>
-Running QEMU with "-nic help" used to work in QEMU 5.2 and earlier versions
+Use translations added in VhostIOVATree in SVQ.
-(it showed the available netdev backends), but this feature got broken during
-some refactoring in version 6.0. Let's restore the old behavior, and while
+Only introduce usage here, not allocation and deallocation. As with
-we're at it, let's also print the available NIC models here now since this
+previous patches, we use the dead code paths of shadow_vqs_enabled to
-option can be used to configure both, netdev backend and model in one go.
+avoid commiting too many changes at once. These are impossible to take
+at the moment.
-Fixes: ad6f932fe8 ("net: do not exit on "netdev_add help" monitor command")
-Signed-off-by: Thomas Huth <thuth@redhat.com>
+Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
 Acked-by: Michael S. Tsirkin <mst@redhat.com>
 Signed-off-by: Jason Wang <jasowang@redhat.com>
 ---
- net/net.c | 14 ++++++++++++--
+ hw/virtio/vhost-shadow-virtqueue.c |  86 +++++++++++++++++++++++---
-file changed, 12 insertions(+), 2 deletions(-)
+ hw/virtio/vhost-shadow-virtqueue.h |   6 +-
+ hw/virtio/vhost-vdpa.c             | 122 +++++++++++++++++++++++++++++++------
-diff --git a/net/net.c b/net/net.c
+ include/hw/virtio/vhost-vdpa.h     |   3 +
 files changed, 187 insertions(+), 30 deletions(-)
 diff --git a/hw/virtio/vhost-shadow-virtqueue.c b/hw/virtio/vhost-shadow-virtqueue.c
 index XXXXXXX..XXXXXXX 100644
---- a/net/net.c
+--- a/hw/virtio/vhost-shadow-virtqueue.c
-+++ b/net/net.c
++++ b/hw/virtio/vhost-shadow-virtqueue.c
-@@ -XXX,XX +XXX,XX @@ static int net_param_nic(void *dummy, QemuOpts *opts, Error **errp)
+@@ -XXX,XX +XXX,XX @@ static uint16_t vhost_svq_available_slots(const VhostShadowVirtqueue *svq)
-     const char *type;
+     return svq->vring.num - (svq->shadow_avail_idx - svq->shadow_used_idx);
+ }
-     type = qemu_opt_get(opts, "type");
--    if (type && g_str_equal(type, "none")) {
+-static void vhost_vring_write_descs(VhostShadowVirtqueue *svq,
--        return 0;    /* Nothing to do, default_net is cleared in vl.c */
++/**
-+    if (type) {
++ * Translate addresses between the qemu's virtual address and the SVQ IOVA
-+        if (g_str_equal(type, "none")) {
++ *
-+            return 0;    /* Nothing to do, default_net is cleared in vl.c */
++ * @svq: Shadow VirtQueue
 + * @vaddr: Translated IOVA addresses
 + * @iovec: Source qemu's VA addresses
 + * @num: Length of iovec and minimum length of vaddr
 + */
 +static bool vhost_svq_translate_addr(const VhostShadowVirtqueue *svq,
 +                                     hwaddr *addrs, const struct iovec *iovec,
 +                                     size_t num)
 +{
 +    if (num == 0) {
 +        return true;
 +    }
 +
 +    for (size_t i = 0; i < num; ++i) {
 +        DMAMap needle = {
 +            .translated_addr = (hwaddr)(uintptr_t)iovec[i].iov_base,
 +            .size = iovec[i].iov_len,
 +        };
 +        Int128 needle_last, map_last;
 +        size_t off;
 +
 +        const DMAMap *map = vhost_iova_tree_find_iova(svq->iova_tree, &needle);
 +        /*
 +         * Map cannot be NULL since iova map contains all guest space and
 +         * qemu already has a physical address mapped
 +         */
 +        if (unlikely(!map)) {
 +            qemu_log_mask(LOG_GUEST_ERROR,
 +                          "Invalid address 0x%"HWADDR_PRIx" given by guest",
 +                          needle.translated_addr);
 +            return false;
 +        }
-+        if (is_help_option(type)) {
++
-+            GPtrArray *nic_models = qemu_get_nic_models(TYPE_DEVICE);
++        off = needle.translated_addr - map->translated_addr;
-+            show_netdevs();
++        addrs[i] = map->iova + off;
-+            printf("\n");
++
-+            qemu_show_nic_models(type, (const char **)nic_models->pdata);
++        needle_last = int128_add(int128_make64(needle.translated_addr),
-+            g_ptr_array_free(nic_models, true);
++                                 int128_make64(iovec[i].iov_len));
-+            exit(0);
++        map_last = int128_make64(map->translated_addr + map->size);
 +        if (unlikely(int128_gt(needle_last, map_last))) {
 +            qemu_log_mask(LOG_GUEST_ERROR,
 +                          "Guest buffer expands over iova range");
 +            return false;
 +        }
++    }
++
++    return true;
++}
++
++static void vhost_vring_write_descs(VhostShadowVirtqueue *svq, hwaddr *sg,
+                                     const struct iovec *iovec, size_t num,
+                                     bool more_descs, bool write)
+ {
+@@ -XXX,XX +XXX,XX @@ static void vhost_vring_write_descs(VhostShadowVirtqueue *svq,
+         } else {
+             descs[i].flags = flags;
+         }
+-        descs[i].addr = cpu_to_le64((hwaddr)(intptr_t)iovec[n].iov_base);
++        descs[i].addr = cpu_to_le64(sg[n]);
+         descs[i].len = cpu_to_le32(iovec[n].iov_len);
+         last = i;
+@@ -XXX,XX +XXX,XX @@ static bool vhost_svq_add_split(VhostShadowVirtqueue *svq,
+ {
+     unsigned avail_idx;
+     vring_avail_t *avail = svq->vring.avail;
++    bool ok;
++    g_autofree hwaddr *sgs = g_new(hwaddr, MAX(elem->out_num, elem->in_num));
+     *head = svq->free_head;
+@@ -XXX,XX +XXX,XX @@ static bool vhost_svq_add_split(VhostShadowVirtqueue *svq,
+         return false;
      }
-     idx = nic_get_free_idx();
+-    vhost_vring_write_descs(svq, elem->out_sg, elem->out_num, elem->in_num > 0,
 -                            false);
 -    vhost_vring_write_descs(svq, elem->in_sg, elem->in_num, false, true);
 +    ok = vhost_svq_translate_addr(svq, sgs, elem->out_sg, elem->out_num);
 +    if (unlikely(!ok)) {
 +        return false;
 +    }
 +    vhost_vring_write_descs(svq, sgs, elem->out_sg, elem->out_num,
 +                            elem->in_num > 0, false);
 +
 +
 +    ok = vhost_svq_translate_addr(svq, sgs, elem->in_sg, elem->in_num);
 +    if (unlikely(!ok)) {
 +        return false;
 +    }
 +
 +    vhost_vring_write_descs(svq, sgs, elem->in_sg, elem->in_num, false, true);
      /*
       * Put the entry in the available array (but don't update avail->idx until
@@ -XXX,XX +XXX,XX @@ void vhost_svq_set_svq_call_fd(VhostShadowVirtqueue *svq, int call_fd)
  void vhost_svq_get_vring_addr(const VhostShadowVirtqueue *svq,
                                struct vhost_vring_addr *addr)
  {
 -    addr->desc_user_addr = (uint64_t)(intptr_t)svq->vring.desc;
 -    addr->avail_user_addr = (uint64_t)(intptr_t)svq->vring.avail;
 -    addr->used_user_addr = (uint64_t)(intptr_t)svq->vring.used;
 +    addr->desc_user_addr = (uint64_t)(uintptr_t)svq->vring.desc;
 +    addr->avail_user_addr = (uint64_t)(uintptr_t)svq->vring.avail;
 +    addr->used_user_addr = (uint64_t)(uintptr_t)svq->vring.used;
  }
  size_t vhost_svq_driver_area_size(const VhostShadowVirtqueue *svq)
@@ -XXX,XX +XXX,XX @@ void vhost_svq_stop(VhostShadowVirtqueue *svq)
   * Creates vhost shadow virtqueue, and instructs the vhost device to use the
   * shadow methods and file descriptors.
   *
 + * @iova_tree: Tree to perform descriptors translations
 + *
   * Returns the new virtqueue or NULL.
   *
   * In case of error, reason is reported through error_report.
   */
 -VhostShadowVirtqueue *vhost_svq_new(void)
 +VhostShadowVirtqueue *vhost_svq_new(VhostIOVATree *iova_tree)
  {
      g_autofree VhostShadowVirtqueue *svq = g_new0(VhostShadowVirtqueue, 1);
      int r;
@@ -XXX,XX +XXX,XX @@ VhostShadowVirtqueue *vhost_svq_new(void)
      event_notifier_init_fd(&svq->svq_kick, VHOST_FILE_UNBIND);
      event_notifier_set_handler(&svq->hdev_call, vhost_svq_handle_call);
 +    svq->iova_tree = iova_tree;
      return g_steal_pointer(&svq);
  err_init_hdev_call:
 diff --git a/hw/virtio/vhost-shadow-virtqueue.h b/hw/virtio/vhost-shadow-virtqueue.h
 index XXXXXXX..XXXXXXX 100644
 --- a/hw/virtio/vhost-shadow-virtqueue.h
 +++ b/hw/virtio/vhost-shadow-virtqueue.h
@@ -XXX,XX +XXX,XX @@
  #include "qemu/event_notifier.h"
  #include "hw/virtio/virtio.h"
  #include "standard-headers/linux/vhost_types.h"
 +#include "hw/virtio/vhost-iova-tree.h"
  /* Shadow virtqueue to relay notifications */
  typedef struct VhostShadowVirtqueue {
@@ -XXX,XX +XXX,XX @@ typedef struct VhostShadowVirtqueue {
      /* Virtio device */
      VirtIODevice *vdev;
 +    /* IOVA mapping */
 +    VhostIOVATree *iova_tree;
 +
      /* Map for use the guest's descriptors */
      VirtQueueElement **ring_id_maps;
@@ -XXX,XX +XXX,XX @@ void vhost_svq_start(VhostShadowVirtqueue *svq, VirtIODevice *vdev,
                       VirtQueue *vq);
  void vhost_svq_stop(VhostShadowVirtqueue *svq);
 -VhostShadowVirtqueue *vhost_svq_new(void);
 +VhostShadowVirtqueue *vhost_svq_new(VhostIOVATree *iova_tree);
  void vhost_svq_free(gpointer vq);
  G_DEFINE_AUTOPTR_CLEANUP_FUNC(VhostShadowVirtqueue, vhost_svq_free);
 diff --git a/hw/virtio/vhost-vdpa.c b/hw/virtio/vhost-vdpa.c
 index XXXXXXX..XXXXXXX 100644
 --- a/hw/virtio/vhost-vdpa.c
 +++ b/hw/virtio/vhost-vdpa.c
@@ -XXX,XX +XXX,XX @@ static void vhost_vdpa_listener_region_add(MemoryListener *listener,
                                           vaddr, section->readonly);
      llsize = int128_sub(llend, int128_make64(iova));
 +    if (v->shadow_vqs_enabled) {
 +        DMAMap mem_region = {
 +            .translated_addr = (hwaddr)(uintptr_t)vaddr,
 +            .size = int128_get64(llsize) - 1,
 +            .perm = IOMMU_ACCESS_FLAG(true, section->readonly),
 +        };
 +
 +        int r = vhost_iova_tree_map_alloc(v->iova_tree, &mem_region);
 +        if (unlikely(r != IOVA_OK)) {
 +            error_report("Can't allocate a mapping (%d)", r);
 +            goto fail;
 +        }
 +
 +        iova = mem_region.iova;
 +    }
      vhost_vdpa_iotlb_batch_begin_once(v);
      ret = vhost_vdpa_dma_map(v, iova, int128_get64(llsize),
@@ -XXX,XX +XXX,XX @@ static void vhost_vdpa_listener_region_del(MemoryListener *listener,
      llsize = int128_sub(llend, int128_make64(iova));
 +    if (v->shadow_vqs_enabled) {
 +        const DMAMap *result;
 +        const void *vaddr = memory_region_get_ram_ptr(section->mr) +
 +            section->offset_within_region +
 +            (iova - section->offset_within_address_space);
 +        DMAMap mem_region = {
 +            .translated_addr = (hwaddr)(uintptr_t)vaddr,
 +            .size = int128_get64(llsize) - 1,
 +        };
 +
 +        result = vhost_iova_tree_find_iova(v->iova_tree, &mem_region);
 +        iova = result->iova;
 +        vhost_iova_tree_remove(v->iova_tree, &mem_region);
 +    }
      vhost_vdpa_iotlb_batch_begin_once(v);
      ret = vhost_vdpa_dma_unmap(v, iova, int128_get64(llsize));
      if (ret) {
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_init_svq(struct vhost_dev *hdev, struct vhost_vdpa *v,
      shadow_vqs = g_ptr_array_new_full(hdev->nvqs, vhost_svq_free);
      for (unsigned n = 0; n < hdev->nvqs; ++n) {
 -        g_autoptr(VhostShadowVirtqueue) svq = vhost_svq_new();
 +        g_autoptr(VhostShadowVirtqueue) svq = vhost_svq_new(v->iova_tree);
          if (unlikely(!svq)) {
              error_setg(errp, "Cannot create svq %u", n);
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_svq_set_fds(struct vhost_dev *dev,
  /**
   * Unmap a SVQ area in the device
   */
 -static bool vhost_vdpa_svq_unmap_ring(struct vhost_vdpa *v, hwaddr iova,
 -                                      hwaddr size)
 +static bool vhost_vdpa_svq_unmap_ring(struct vhost_vdpa *v,
 +                                      const DMAMap *needle)
  {
 +    const DMAMap *result = vhost_iova_tree_find_iova(v->iova_tree, needle);
 +    hwaddr size;
      int r;
 -    size = ROUND_UP(size, qemu_real_host_page_size);
 -    r = vhost_vdpa_dma_unmap(v, iova, size);
 +    if (unlikely(!result)) {
 +        error_report("Unable to find SVQ address to unmap");
 +        return false;
 +    }
 +
 +    size = ROUND_UP(result->size, qemu_real_host_page_size);
 +    r = vhost_vdpa_dma_unmap(v, result->iova, size);
      return r == 0;
  }
  static bool vhost_vdpa_svq_unmap_rings(struct vhost_dev *dev,
                                         const VhostShadowVirtqueue *svq)
  {
 +    DMAMap needle = {};
      struct vhost_vdpa *v = dev->opaque;
      struct vhost_vring_addr svq_addr;
 -    size_t device_size = vhost_svq_device_area_size(svq);
 -    size_t driver_size = vhost_svq_driver_area_size(svq);
      bool ok;
      vhost_svq_get_vring_addr(svq, &svq_addr);
 -    ok = vhost_vdpa_svq_unmap_ring(v, svq_addr.desc_user_addr, driver_size);
 +    needle.translated_addr = svq_addr.desc_user_addr;
 +    ok = vhost_vdpa_svq_unmap_ring(v, &needle);
      if (unlikely(!ok)) {
          return false;
      }
 -    return vhost_vdpa_svq_unmap_ring(v, svq_addr.used_user_addr, device_size);
 +    needle.translated_addr = svq_addr.used_user_addr;
 +    return vhost_vdpa_svq_unmap_ring(v, &needle);
 +}
 +
 +/**
 + * Map the SVQ area in the device
 + *
 + * @v: Vhost-vdpa device
 + * @needle: The area to search iova
 + * @errorp: Error pointer
 + */
 +static bool vhost_vdpa_svq_map_ring(struct vhost_vdpa *v, DMAMap *needle,
 +                                    Error **errp)
 +{
 +    int r;
 +
 +    r = vhost_iova_tree_map_alloc(v->iova_tree, needle);
 +    if (unlikely(r != IOVA_OK)) {
 +        error_setg(errp, "Cannot allocate iova (%d)", r);
 +        return false;
 +    }
 +
 +    r = vhost_vdpa_dma_map(v, needle->iova, needle->size + 1,
 +                           (void *)(uintptr_t)needle->translated_addr,
 +                           needle->perm == IOMMU_RO);
 +    if (unlikely(r != 0)) {
 +        error_setg_errno(errp, -r, "Cannot map region to device");
 +        vhost_iova_tree_remove(v->iova_tree, needle);
 +    }
 +
 +    return r == 0;
  }
  /**
@@ -XXX,XX +XXX,XX @@ static bool vhost_vdpa_svq_map_rings(struct vhost_dev *dev,
                                       struct vhost_vring_addr *addr,
                                       Error **errp)
  {
 +    DMAMap device_region, driver_region;
 +    struct vhost_vring_addr svq_addr;
      struct vhost_vdpa *v = dev->opaque;
      size_t device_size = vhost_svq_device_area_size(svq);
      size_t driver_size = vhost_svq_driver_area_size(svq);
 -    int r;
 +    size_t avail_offset;
 +    bool ok;
      ERRP_GUARD();
 -    vhost_svq_get_vring_addr(svq, addr);
 +    vhost_svq_get_vring_addr(svq, &svq_addr);
 -    r = vhost_vdpa_dma_map(v, addr->desc_user_addr, driver_size,
 -                           (void *)(uintptr_t)addr->desc_user_addr, true);
 -    if (unlikely(r != 0)) {
 -        error_setg_errno(errp, -r, "Cannot create vq driver region: ");
 +    driver_region = (DMAMap) {
 +        .translated_addr = svq_addr.desc_user_addr,
 +        .size = driver_size - 1,
 +        .perm = IOMMU_RO,
 +    };
 +    ok = vhost_vdpa_svq_map_ring(v, &driver_region, errp);
 +    if (unlikely(!ok)) {
 +        error_prepend(errp, "Cannot create vq driver region: ");
          return false;
      }
 +    addr->desc_user_addr = driver_region.iova;
 +    avail_offset = svq_addr.avail_user_addr - svq_addr.desc_user_addr;
 +    addr->avail_user_addr = driver_region.iova + avail_offset;
 -    r = vhost_vdpa_dma_map(v, addr->used_user_addr, device_size,
 -                           (void *)(intptr_t)addr->used_user_addr, false);
 -    if (unlikely(r != 0)) {
 -        error_setg_errno(errp, -r, "Cannot create vq device region: ");
 +    device_region = (DMAMap) {
 +        .translated_addr = svq_addr.used_user_addr,
 +        .size = device_size - 1,
 +        .perm = IOMMU_RW,
 +    };
 +    ok = vhost_vdpa_svq_map_ring(v, &device_region, errp);
 +    if (unlikely(!ok)) {
 +        error_prepend(errp, "Cannot create vq device region: ");
 +        vhost_vdpa_svq_unmap_ring(v, &driver_region);
      }
 +    addr->used_user_addr = device_region.iova;
 -    return r == 0;
 +    return ok;
  }
  static bool vhost_vdpa_svq_setup(struct vhost_dev *dev,
 diff --git a/include/hw/virtio/vhost-vdpa.h b/include/hw/virtio/vhost-vdpa.h
 index XXXXXXX..XXXXXXX 100644
 --- a/include/hw/virtio/vhost-vdpa.h
 +++ b/include/hw/virtio/vhost-vdpa.h
@@ -XXX,XX +XXX,XX @@
  #include <gmodule.h>
 +#include "hw/virtio/vhost-iova-tree.h"
  #include "hw/virtio/virtio.h"
  #include "standard-headers/linux/vhost_types.h"
@@ -XXX,XX +XXX,XX @@ typedef struct vhost_vdpa {
      MemoryListener listener;
      struct vhost_vdpa_iova_range iova_range;
      bool shadow_vqs_enabled;
 +    /* IOVA mapping used by the Shadow Virtqueue */
 +    VhostIOVATree *iova_tree;
      GPtrArray *shadow_vqs;
      struct vhost_dev *dev;
      VhostVDPAHostNotifier notifier[VIRTIO_QUEUE_MAX];
 --
 .7.4

-[PULL 05/11] hw/net/vmxnet3: allow VMXNET3_MAX_MTU itself as a value
+[PULL V3 13/15] vdpa: Adapt vhost_vdpa_get_vring_base to SVQ
-From: Fiona Ebner <f.ebner@proxmox.com>
+From: Eugenio Pérez <eperezma@redhat.com>
-Currently, VMXNET3_MAX_MTU itself (being 9000) is not considered a
+This is needed to achieve migration, so the destination can restore its
-valid value for the MTU, but a guest running ESXi 7.0 might try to
+index.
 set it and fail the assert [0].
-In the Linux kernel, dev->max_mtu itself is a valid value for the MTU
+Setting base as last used idx, so destination will see as available all
-and for the vmxnet3 driver it's 9000, so a guest running Linux will
+the entries that the device did not use, including the in-flight
-also fail the assert when trying to set an MTU of 9000.
+processing ones.
-VMXNET3_MAX_MTU and s->mtu don't seem to be used in relation to buffer
+This is ok for networking, but other kinds of devices might have
-allocations/accesses, so allowing the upper limit itself as a value
+problems with these retransmissions.
 should be fine.
-[0]: https://forum.proxmox.com/threads/114011/
+Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
+Acked-by: Michael S. Tsirkin <mst@redhat.com>
 Fixes: d05dcd94ae ("net: vmxnet3: validate configuration values during activate (CVE-2021-20203)")
 Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
 Signed-off-by: Jason Wang <jasowang@redhat.com>
 ---
- hw/net/vmxnet3.c | 2 +-
+ hw/virtio/vhost-vdpa.c | 17 +++++++++++++++++
-file changed, 1 insertion(+), 1 deletion(-)
+file changed, 17 insertions(+)
-diff --git a/hw/net/vmxnet3.c b/hw/net/vmxnet3.c
+diff --git a/hw/virtio/vhost-vdpa.c b/hw/virtio/vhost-vdpa.c
 index XXXXXXX..XXXXXXX 100644
---- a/hw/net/vmxnet3.c
+--- a/hw/virtio/vhost-vdpa.c
-+++ b/hw/net/vmxnet3.c
++++ b/hw/virtio/vhost-vdpa.c
-@@ -XXX,XX +XXX,XX @@ static void vmxnet3_activate_device(VMXNET3State *s)
+@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_vring_base(struct vhost_dev *dev,
-     vmxnet3_setup_rx_filtering(s);
+ static int vhost_vdpa_get_vring_base(struct vhost_dev *dev,
-     /* Cache fields from shared memory */
+                                        struct vhost_vring_state *ring)
-     s->mtu = VMXNET3_READ_DRV_SHARED32(d, s->drv_shmem, devRead.misc.mtu);
+ {
--    assert(VMXNET3_MIN_MTU <= s->mtu && s->mtu < VMXNET3_MAX_MTU);
++    struct vhost_vdpa *v = dev->opaque;
-+    assert(VMXNET3_MIN_MTU <= s->mtu && s->mtu <= VMXNET3_MAX_MTU);
+     int ret;
-     VMW_CFPRN("MTU is %u", s->mtu);
++    if (v->shadow_vqs_enabled) {
-     s->max_rx_frags =
++        VhostShadowVirtqueue *svq = g_ptr_array_index(v->shadow_vqs,
 +                                                      ring->index);
 +
 +        /*
 +         * Setting base as last used idx, so destination will see as available
 +         * all the entries that the device did not use, including the in-flight
 +         * processing ones.
 +         *
 +         * TODO: This is ok for networking, but other kinds of devices might
 +         * have problems with these retransmissions.
 +         */
 +        ring->num = svq->last_used_idx;
 +        return 0;
 +    }
 +
      ret = vhost_vdpa_call(dev, VHOST_GET_VRING_BASE, ring);
      trace_vhost_vdpa_get_vring_base(dev, ring->index, ring->num);
      return ret;
 --
 .7.4

-[PULL 11/11] vdpa: fix VHOST_BACKEND_F_IOTLB_ASID flag check
+[PULL V3 14/15] vdpa: Never set log_base addr if SVQ is enabled
 From: Eugenio Pérez <eperezma@redhat.com>
-VHOST_BACKEND_F_IOTLB_ASID is the feature bit, not the bitmask. Since
+Setting the log address would make the device start reporting invalid
-the device under test also provided VHOST_BACKEND_F_IOTLB_MSG_V2 and
+dirty memory because the SVQ vrings are located in qemu's memory.
 VHOST_BACKEND_F_IOTLB_BATCH, this went unnoticed.
-Fixes: c1a1008685 ("vdpa: always start CVQ in SVQ mode if possible")
 Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
-Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
+Acked-by: Michael S. Tsirkin <mst@redhat.com>
 Acked-by: Jason Wang <jasowang@redhat.com>
 Signed-off-by: Jason Wang <jasowang@redhat.com>
 ---
- net/vhost-vdpa.c | 2 +-
+ hw/virtio/vhost-vdpa.c | 3 ++-
-file changed, 1 insertion(+), 1 deletion(-)
+file changed, 2 insertions(+), 1 deletion(-)
-diff --git a/net/vhost-vdpa.c b/net/vhost-vdpa.c
+diff --git a/hw/virtio/vhost-vdpa.c b/hw/virtio/vhost-vdpa.c
 index XXXXXXX..XXXXXXX 100644
---- a/net/vhost-vdpa.c
+--- a/hw/virtio/vhost-vdpa.c
-+++ b/net/vhost-vdpa.c
++++ b/hw/virtio/vhost-vdpa.c
-@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_net_cvq_start(NetClientState *nc)
+@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_dev_start(struct vhost_dev *dev, bool started)
-             g_strerror(errno), errno);
+ static int vhost_vdpa_set_log_base(struct vhost_dev *dev, uint64_t base,
-         return -1;
+                                      struct vhost_log *log)
-     }
+ {
--    if (!(backend_features & VHOST_BACKEND_F_IOTLB_ASID) ||
+-    if (vhost_vdpa_one_time_request(dev)) {
-+    if (!(backend_features & BIT_ULL(VHOST_BACKEND_F_IOTLB_ASID)) ||
++    struct vhost_vdpa *v = dev->opaque;
-         !vhost_vdpa_net_valid_svq_features(v->dev->features, NULL)) {
++    if (v->shadow_vqs_enabled || vhost_vdpa_one_time_request(dev)) {
          return 0;
      }
 --
 .7.4

-[PULL 03/11] net: Replace "Supported NIC models" with "Available NIC models"
+[PULL V3 15/15] vdpa: Expose VHOST_F_LOG_ALL on SVQ
-From: Thomas Huth <thuth@redhat.com>
+From: Eugenio Pérez <eperezma@redhat.com>
-Just because a NIC model is compiled into the QEMU binary does not
+SVQ is able to log the dirty bits by itself, so let's use it to not
-necessary mean that it can be used with each and every machine.
+block migration.
 So let's rather talk about "available" models instead of "supported"
 models, just to avoid confusion.
-Reviewed-by: Claudio Fontana <cfontana@suse.de>
+Also, ignore set and clear of VHOST_F_LOG_ALL on set_features if SVQ is
-Signed-off-by: Thomas Huth <thuth@redhat.com>
+enabled. Even if the device supports it, the reports would be nonsense
 because SVQ memory is in the qemu region.
 The log region is still allocated. Future changes might skip that, but
 this series is already long enough.
 Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
 Acked-by: Michael S. Tsirkin <mst@redhat.com>
 Signed-off-by: Jason Wang <jasowang@redhat.com>
 ---
- net/net.c | 2 +-
+ hw/virtio/vhost-vdpa.c         | 39 +++++++++++++++++++++++++++++++++++----
-file changed, 1 insertion(+), 1 deletion(-)
+ include/hw/virtio/vhost-vdpa.h |  1 +
 files changed, 36 insertions(+), 4 deletions(-)
-diff --git a/net/net.c b/net/net.c
+diff --git a/hw/virtio/vhost-vdpa.c b/hw/virtio/vhost-vdpa.c
 index XXXXXXX..XXXXXXX 100644
---- a/net/net.c
+--- a/hw/virtio/vhost-vdpa.c
-+++ b/net/net.c
++++ b/hw/virtio/vhost-vdpa.c
-@@ -XXX,XX +XXX,XX @@ int qemu_show_nic_models(const char *arg, const char *const *models)
+@@ -XXX,XX +XXX,XX @@ static bool vhost_vdpa_one_time_request(struct vhost_dev *dev)
      return v->index != 0;
  }
 +static int vhost_vdpa_get_dev_features(struct vhost_dev *dev,
 +                                       uint64_t *features)
 +{
 +    int ret;
 +
 +    ret = vhost_vdpa_call(dev, VHOST_GET_FEATURES, features);
 +    trace_vhost_vdpa_get_features(dev, *features);
 +    return ret;
 +}
 +
  static int vhost_vdpa_init_svq(struct vhost_dev *hdev, struct vhost_vdpa *v,
                                 Error **errp)
  {
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_init_svq(struct vhost_dev *hdev, struct vhost_vdpa *v,
          return 0;
      }
--    printf("Supported NIC models:\n");
+-    r = hdev->vhost_ops->vhost_get_features(hdev, &dev_features);
-+    printf("Available NIC models:\n");
++    r = vhost_vdpa_get_dev_features(hdev, &dev_features);
-     for (i = 0 ; models[i]; i++) {
+     if (r != 0) {
-         printf("%s\n", models[i]);
+         error_setg_errno(errp, -r, "Can't get vdpa device features");
          return r;
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_mem_table(struct vhost_dev *dev,
  static int vhost_vdpa_set_features(struct vhost_dev *dev,
                                     uint64_t features)
  {
 +    struct vhost_vdpa *v = dev->opaque;
      int ret;
      if (vhost_vdpa_one_time_request(dev)) {
          return 0;
      }
++    if (v->shadow_vqs_enabled) {
++        if ((v->acked_features ^ features) == BIT_ULL(VHOST_F_LOG_ALL)) {
++            /*
++             * QEMU is just trying to enable or disable logging. SVQ handles
++             * this sepparately, so no need to forward this.
++             */
++            v->acked_features = features;
++            return 0;
++        }
++
++        v->acked_features = features;
++
++        /* We must not ack _F_LOG if SVQ is enabled */
++        features &= ~BIT_ULL(VHOST_F_LOG_ALL);
++    }
++
+     trace_vhost_vdpa_set_features(dev, features);
+     ret = vhost_vdpa_call(dev, VHOST_SET_FEATURES, &features);
+     if (ret) {
+@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_vring_call(struct vhost_dev *dev,
+ static int vhost_vdpa_get_features(struct vhost_dev *dev,
+                                      uint64_t *features)
+ {
+-    int ret;
++    struct vhost_vdpa *v = dev->opaque;
++    int ret = vhost_vdpa_get_dev_features(dev, features);
++
++    if (ret == 0 && v->shadow_vqs_enabled) {
++        /* Add SVQ logging capabilities */
++        *features |= BIT_ULL(VHOST_F_LOG_ALL);
++    }
+-    ret = vhost_vdpa_call(dev, VHOST_GET_FEATURES, features);
+-    trace_vhost_vdpa_get_features(dev, *features);
+     return ret;
+ }
+diff --git a/include/hw/virtio/vhost-vdpa.h b/include/hw/virtio/vhost-vdpa.h
+index XXXXXXX..XXXXXXX 100644
+--- a/include/hw/virtio/vhost-vdpa.h
++++ b/include/hw/virtio/vhost-vdpa.h
+@@ -XXX,XX +XXX,XX @@ typedef struct vhost_vdpa {
+     bool iotlb_batch_begin_sent;
+     MemoryListener listener;
+     struct vhost_vdpa_iova_range iova_range;
++    uint64_t acked_features;
+     bool shadow_vqs_enabled;
+     /* IOVA mapping used by the Shadow Virtqueue */
+     VhostIOVATree *iova_tree;
 --
 .7.4

The following changes since commit 13356edb87506c148b163b8c7eb0695647d00c2a:

Merge tag 'block-pull-request' of https://gitlab.com/stefanha/qemu into staging (2023-01-24 09:45:33 +0000)

are available in the git repository at:

https://github.com/jasowang/qemu.git tags/net-pull-request

for you to fetch changes up to 2bd492bca521ee8594f1d5db8dc9aac126fc4f85:

vdpa: fix VHOST_BACKEND_F_IOTLB_ASID flag check (2023-02-02 14:16:48 +0800)

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

----------------------------------------------------------------
Christian Svensson (1):
      net: Increase L2TPv3 buffer to fit jumboframes

Eugenio Pérez (1):
      vdpa: fix VHOST_BACKEND_F_IOTLB_ASID flag check

Fiona Ebner (1):
      hw/net/vmxnet3: allow VMXNET3_MAX_MTU itself as a value

Joelle van Dyne (1):
      vmnet: stop recieving events when VM is stopped

Laurent Vivier (2):
      tests/qtest: netdev: test stream and dgram backends
      net: stream: add a new option to automatically reconnect

Qiang Liu (2):
      hw/net/lan9118: log [read|write]b when mode_16bit is enabled rather than abort
      hw/net/can/xlnx-zynqmp-can: fix assertion failures in transfer_fifo()

Thomas Huth (3):
      net: Move the code to collect available NIC models to a separate function
      net: Restore printing of the help text with "-nic help"
      net: Replace "Supported NIC models" with "Available NIC models"

From: Thomas Huth <thuth@redhat.com>

The code that collects the available NIC models is not really specific
to PCI anymore and will be required in the next patch, too, so let's
move this into a new separate function in net.c instead.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 hw/pci/pci.c      | 29 +----------------------------
 include/net/net.h | 14 ++++++++++++++
 net/net.c         | 34 ++++++++++++++++++++++++++++++++++
 3 files changed, 49 insertions(+), 28 deletions(-)

diff --git a/hw/pci/pci.c b/hw/pci/pci.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/pci/pci.c
+++ b/hw/pci/pci.c
@@ -XXX,XX +XXX,XX @@ PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus,
                                const char *default_devaddr)
 {
     const char *devaddr = nd->devaddr ? nd->devaddr : default_devaddr;
-    GSList *list;
     GPtrArray *pci_nic_models;
     PCIBus *bus;
     PCIDevice *pci_dev;
@@ -XXX,XX +XXX,XX @@ PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus,
         nd->model = g_strdup("virtio-net-pci");
     }
 
-    list = object_class_get_list_sorted(TYPE_PCI_DEVICE, false);
-    pci_nic_models = g_ptr_array_new();
-    while (list) {
-        DeviceClass *dc = OBJECT_CLASS_CHECK(DeviceClass, list->data,
-                                             TYPE_DEVICE);
-        GSList *next;
-        if (test_bit(DEVICE_CATEGORY_NETWORK, dc->categories) &&
-            dc->user_creatable) {
-            const char *name = object_class_get_name(list->data);
-            /*
-             * A network device might also be something else than a NIC, see
-             * e.g. the "rocker" device. Thus we have to look for the "netdev"
-             * property, too. Unfortunately, some devices like virtio-net only
-             * create this property during instance_init, so we have to create
-             * a temporary instance here to be able to check it.
-             */
-            Object *obj = object_new_with_class(OBJECT_CLASS(dc));
-            if (object_property_find(obj, "netdev")) {
-                g_ptr_array_add(pci_nic_models, (gpointer)name);
-            }
-            object_unref(obj);
-        }
-        next = list->next;
-        g_slist_free_1(list);
-        list = next;
-    }
-    g_ptr_array_add(pci_nic_models, NULL);
+    pci_nic_models = qemu_get_nic_models(TYPE_PCI_DEVICE);
 
     if (qemu_show_nic_models(nd->model, (const char **)pci_nic_models->pdata)) {
         exit(0);
diff --git a/include/net/net.h b/include/net/net.h
index XXXXXXX..XXXXXXX 100644
--- a/include/net/net.h
+++ b/include/net/net.h
@@ -XXX,XX +XXX,XX @@ void net_socket_rs_init(SocketReadState *rs,
                         bool vnet_hdr);
 NetClientState *qemu_get_peer(NetClientState *nc, int queue_index);
 
+/**
+ * qemu_get_nic_models:
+ * @device_type: Defines which devices should be taken into consideration
+ *               (e.g. TYPE_DEVICE for all devices, or TYPE_PCI_DEVICE for PCI)
+ *
+ * Get an array of pointers to names of NIC devices that are available in
+ * the QEMU binary. The array is terminated with a NULL pointer entry.
+ * The caller is responsible for freeing the memory when it is not required
+ * anymore, e.g. with g_ptr_array_free(..., true).
+ *
+ * Returns: Pointer to the array that contains the pointers to the names.
+ */
+GPtrArray *qemu_get_nic_models(const char *device_type);
+
 /* NIC info */
 
 #define MAX_NICS 8
diff --git a/net/net.c b/net/net.c
index XXXXXXX..XXXXXXX 100644
--- a/net/net.c
+++ b/net/net.c
@@ -XXX,XX +XXX,XX @@ static int nic_get_free_idx(void)
     return -1;
 }
 
+GPtrArray *qemu_get_nic_models(const char *device_type)
+{
+    GPtrArray *nic_models = g_ptr_array_new();
+    GSList *list = object_class_get_list_sorted(device_type, false);
+
+    while (list) {
+        DeviceClass *dc = OBJECT_CLASS_CHECK(DeviceClass, list->data,
+                                             TYPE_DEVICE);
+        GSList *next;
+        if (test_bit(DEVICE_CATEGORY_NETWORK, dc->categories) &&
+            dc->user_creatable) {
+            const char *name = object_class_get_name(list->data);
+            /*
+             * A network device might also be something else than a NIC, see
+             * e.g. the "rocker" device. Thus we have to look for the "netdev"
+             * property, too. Unfortunately, some devices like virtio-net only
+             * create this property during instance_init, so we have to create
+             * a temporary instance here to be able to check it.
+             */
+            Object *obj = object_new_with_class(OBJECT_CLASS(dc));
+            if (object_property_find(obj, "netdev")) {
+                g_ptr_array_add(nic_models, (gpointer)name);
+            }
+            object_unref(obj);
+        }
+        next = list->next;
+        g_slist_free_1(list);
+        list = next;
+    }
+    g_ptr_array_add(nic_models, NULL);
+
+    return nic_models;
+}
+
 int qemu_show_nic_models(const char *arg, const char *const *models)
 {
     int i;
-- 
2.7.4

From: Thomas Huth <thuth@redhat.com>

Running QEMU with "-nic help" used to work in QEMU 5.2 and earlier versions
(it showed the available netdev backends), but this feature got broken during
some refactoring in version 6.0. Let's restore the old behavior, and while
we're at it, let's also print the available NIC models here now since this
option can be used to configure both, netdev backend and model in one go.

Fixes: ad6f932fe8 ("net: do not exit on "netdev_add help" monitor command")
Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 net/net.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

diff --git a/net/net.c b/net/net.c
index XXXXXXX..XXXXXXX 100644
--- a/net/net.c
+++ b/net/net.c
@@ -XXX,XX +XXX,XX @@ static int net_param_nic(void *dummy, QemuOpts *opts, Error **errp)
     const char *type;
 
     type = qemu_opt_get(opts, "type");
-    if (type && g_str_equal(type, "none")) {
-        return 0;    /* Nothing to do, default_net is cleared in vl.c */
+    if (type) {
+        if (g_str_equal(type, "none")) {
+            return 0;    /* Nothing to do, default_net is cleared in vl.c */
+        }
+        if (is_help_option(type)) {
+            GPtrArray *nic_models = qemu_get_nic_models(TYPE_DEVICE);
+            show_netdevs();
+            printf("\n");
+            qemu_show_nic_models(type, (const char **)nic_models->pdata);
+            g_ptr_array_free(nic_models, true);
+            exit(0);
+        }
     }
 
     idx = nic_get_free_idx();
-- 
2.7.4

From: Qiang Liu <cyruscyliu@gmail.com>

This patch replaces hw_error to guest error log for [read|write]b
accesses when mode_16bit is enabled. This avoids aborting qemu.

Fixes: 1248f8d4cbc3 ("hw/lan9118: Add basic 16-bit mode support.")
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1433
Reported-by: Qiang Liu <cyruscyliu@gmail.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Qiang Liu <cyruscyliu@gmail.com>
Suggested-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 hw/net/lan9118.c | 17 ++++++++---------
 1 file changed, 8 insertions(+), 9 deletions(-)

diff --git a/hw/net/lan9118.c b/hw/net/lan9118.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/net/lan9118.c
+++ b/hw/net/lan9118.c
@@ -XXX,XX +XXX,XX @@
 #include "migration/vmstate.h"
 #include "net/net.h"
 #include "net/eth.h"
-#include "hw/hw.h"
 #include "hw/irq.h"
 #include "hw/net/lan9118.h"
 #include "hw/ptimer.h"
@@ -XXX,XX +XXX,XX @@
 #ifdef DEBUG_LAN9118
 #define DPRINTF(fmt, ...) \
 do { printf("lan9118: " fmt , ## __VA_ARGS__); } while (0)
-#define BADF(fmt, ...) \
-do { hw_error("lan9118: error: " fmt , ## __VA_ARGS__);} while (0)
 #else
 #define DPRINTF(fmt, ...) do {} while(0)
-#define BADF(fmt, ...) \
-do { fprintf(stderr, "lan9118: error: " fmt , ## __VA_ARGS__);} while (0)
 #endif
 
 /* The tx and rx fifo ports are a range of aliased 32-bit registers */
@@ -XXX,XX +XXX,XX @@ static uint32_t do_phy_read(lan9118_state *s, int reg)
     case 30: /* Interrupt mask */
         return s->phy_int_mask;
     default:
-        BADF("PHY read reg %d\n", reg);
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "do_phy_read: PHY read reg %d\n", reg);
         return 0;
     }
 }
@@ -XXX,XX +XXX,XX @@ static void do_phy_write(lan9118_state *s, int reg, uint32_t val)
         phy_update_irq(s);
         break;
     default:
-        BADF("PHY write reg %d = 0x%04x\n", reg, val);
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "do_phy_write: PHY write reg %d = 0x%04x\n", reg, val);
     }
 }
 
@@ -XXX,XX +XXX,XX @@ static void lan9118_16bit_mode_write(void *opaque, hwaddr offset,
         return;
     }
 
-    hw_error("lan9118_write: Bad size 0x%x\n", size);
+    qemu_log_mask(LOG_GUEST_ERROR,
+                  "lan9118_16bit_mode_write: Bad size 0x%x\n", size);
 }
 
 static uint64_t lan9118_readl(void *opaque, hwaddr offset,
@@ -XXX,XX +XXX,XX @@ static uint64_t lan9118_16bit_mode_read(void *opaque, hwaddr offset,
         return lan9118_readl(opaque, offset, size);
     }
 
-    hw_error("lan9118_read: Bad size 0x%x\n", size);
+    qemu_log_mask(LOG_GUEST_ERROR,
+                  "lan9118_16bit_mode_read: Bad size 0x%x\n", size);
     return 0;
 }
 
-- 
2.7.4

From: Fiona Ebner <f.ebner@proxmox.com>

Currently, VMXNET3_MAX_MTU itself (being 9000) is not considered a
valid value for the MTU, but a guest running ESXi 7.0 might try to
set it and fail the assert [0].

In the Linux kernel, dev->max_mtu itself is a valid value for the MTU
and for the vmxnet3 driver it's 9000, so a guest running Linux will
also fail the assert when trying to set an MTU of 9000.

VMXNET3_MAX_MTU and s->mtu don't seem to be used in relation to buffer
allocations/accesses, so allowing the upper limit itself as a value
should be fine.

[0]: https://forum.proxmox.com/threads/114011/

Fixes: d05dcd94ae ("net: vmxnet3: validate configuration values during activate (CVE-2021-20203)")
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 hw/net/vmxnet3.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/hw/net/vmxnet3.c b/hw/net/vmxnet3.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/net/vmxnet3.c
+++ b/hw/net/vmxnet3.c
@@ -XXX,XX +XXX,XX @@ static void vmxnet3_activate_device(VMXNET3State *s)
     vmxnet3_setup_rx_filtering(s);
     /* Cache fields from shared memory */
     s->mtu = VMXNET3_READ_DRV_SHARED32(d, s->drv_shmem, devRead.misc.mtu);
-    assert(VMXNET3_MIN_MTU <= s->mtu && s->mtu < VMXNET3_MAX_MTU);
+    assert(VMXNET3_MIN_MTU <= s->mtu && s->mtu <= VMXNET3_MAX_MTU);
     VMW_CFPRN("MTU is %u", s->mtu);
 
     s->max_rx_frags =
-- 
2.7.4

From: Joelle van Dyne <j@getutm.app>

When the VM is stopped using the HMP command "stop", soon the handler will
stop reading from the vmnet interface. This causes a flood of
`VMNET_INTERFACE_PACKETS_AVAILABLE` events to arrive and puts the host CPU
at 100%. We fix this by removing the event handler from vmnet when the VM
is no longer in a running state and restore it when we return to a running
state.

Signed-off-by: Joelle van Dyne <j@getutm.app>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 net/vmnet-common.m | 48 +++++++++++++++++++++++++++++++++++-------------
 net/vmnet_int.h    |  2 ++
 2 files changed, 37 insertions(+), 13 deletions(-)

diff --git a/net/vmnet-common.m b/net/vmnet-common.m
index XXXXXXX..XXXXXXX 100644
--- a/net/vmnet-common.m
+++ b/net/vmnet-common.m
@@ -XXX,XX +XXX,XX @@
 #include "clients.h"
 #include "qemu/error-report.h"
 #include "qapi/error.h"
+#include "sysemu/runstate.h"
 
 #include <vmnet/vmnet.h>
 #include <dispatch/dispatch.h>
@@ -XXX,XX +XXX,XX @@ static void vmnet_bufs_init(VmnetState *s)
     }
 }
 
+/**
+ * Called on state change to un-register/re-register handlers
+ */
+static void vmnet_vm_state_change_cb(void *opaque, bool running, RunState state)
+{
+    VmnetState *s = opaque;
+
+    if (running) {
+        vmnet_interface_set_event_callback(
+            s->vmnet_if,
+            VMNET_INTERFACE_PACKETS_AVAILABLE,
+            s->if_queue,
+            ^(interface_event_t event_id, xpc_object_t event) {
+                assert(event_id == VMNET_INTERFACE_PACKETS_AVAILABLE);
+                /*
+                 * This function is being called from a non qemu thread, so
+                 * we only schedule a BH, and do the rest of the io completion
+                 * handling from vmnet_send_bh() which runs in a qemu context.
+                 */
+                qemu_bh_schedule(s->send_bh);
+            });
+    } else {
+        vmnet_interface_set_event_callback(
+            s->vmnet_if,
+            VMNET_INTERFACE_PACKETS_AVAILABLE,
+            NULL,
+            NULL);
+    }
+}
 
 int vmnet_if_create(NetClientState *nc,
                     xpc_object_t if_desc,
@@ -XXX,XX +XXX,XX @@ int vmnet_if_create(NetClientState *nc,
     s->packets_send_current_pos = 0;
     s->packets_send_end_pos = 0;
 
-    vmnet_interface_set_event_callback(
-        s->vmnet_if,
-        VMNET_INTERFACE_PACKETS_AVAILABLE,
-        s->if_queue,
-        ^(interface_event_t event_id, xpc_object_t event) {
-            assert(event_id == VMNET_INTERFACE_PACKETS_AVAILABLE);
-            /*
-             * This function is being called from a non qemu thread, so
-             * we only schedule a BH, and do the rest of the io completion
-             * handling from vmnet_send_bh() which runs in a qemu context.
-             */
-            qemu_bh_schedule(s->send_bh);
-        });
+    vmnet_vm_state_change_cb(s, 1, RUN_STATE_RUNNING);
+
+    s->change = qemu_add_vm_change_state_handler(vmnet_vm_state_change_cb, s);
 
     return 0;
 }
@@ -XXX,XX +XXX,XX @@ void vmnet_cleanup_common(NetClientState *nc)
         return;
     }
 
+    vmnet_vm_state_change_cb(s, 0, RUN_STATE_SHUTDOWN);
+    qemu_del_vm_change_state_handler(s->change);
     if_stopped_sem = dispatch_semaphore_create(0);
     vmnet_stop_interface(
         s->vmnet_if,
diff --git a/net/vmnet_int.h b/net/vmnet_int.h
index XXXXXXX..XXXXXXX 100644
--- a/net/vmnet_int.h
+++ b/net/vmnet_int.h
@@ -XXX,XX +XXX,XX @@ typedef struct VmnetState {
     int packets_send_end_pos;
 
     struct iovec iov_buf[VMNET_PACKETS_LIMIT];
+
+    VMChangeStateEntry *change;
 } VmnetState;
 
 const char *vmnet_status_map_str(vmnet_return_t status);
-- 
2.7.4

From: Qiang Liu <cyruscyliu@gmail.com>

Check fifos before poping data from and pushing data into it.

Fixes: 98e5d7a2b726 ("hw/net/can: Introduce Xilinx ZynqMP CAN controller")
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1425
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1427
Reported-by: Qiang Liu <cyruscyliu@gmail.com>
Signed-off-by: Qiang Liu <cyruscyliu@gmail.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 hw/net/can/xlnx-zynqmp-can.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/hw/net/can/xlnx-zynqmp-can.c b/hw/net/can/xlnx-zynqmp-can.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/net/can/xlnx-zynqmp-can.c
+++ b/hw/net/can/xlnx-zynqmp-can.c
@@ -XXX,XX +XXX,XX @@ static void transfer_fifo(XlnxZynqMPCANState *s, Fifo32 *fifo)
     }
 
     while (!fifo32_is_empty(fifo)) {
+        if (fifo32_num_used(fifo) < (4 * CAN_FRAME_SIZE)) {
+            g_autofree char *path = object_get_canonical_path(OBJECT(s));
+            qemu_log_mask(LOG_GUEST_ERROR, "%s: data left in the fifo is not"
+                          " enough for transfer.\n", path);
+            break;
+        }
         for (i = 0; i < CAN_FRAME_SIZE; i++) {
             data[i] = fifo32_pop(fifo);
         }
@@ -XXX,XX +XXX,XX @@ static void transfer_fifo(XlnxZynqMPCANState *s, Fifo32 *fifo)
              * acknowledged. The XlnxZynqMPCAN core receives any message
              * that it transmits.
              */
-            if (fifo32_is_full(&s->rx_fifo)) {
+            if (fifo32_is_full(&s->rx_fifo) ||
+                    (fifo32_num_free(&s->rx_fifo) < (4 * CAN_FRAME_SIZE))) {
                 ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, RXOFLW, 1);
             } else {
                 for (i = 0; i < CAN_FRAME_SIZE; i++) {
-- 
2.7.4

From: Laurent Vivier <lvivier@redhat.com>

Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 tests/qtest/meson.build     |   2 +
 tests/qtest/netdev-socket.c | 448 ++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 450 insertions(+)
 create mode 100644 tests/qtest/netdev-socket.c

diff --git a/tests/qtest/meson.build b/tests/qtest/meson.build
index XXXXXXX..XXXXXXX 100644
--- a/tests/qtest/meson.build
+++ b/tests/qtest/meson.build
@@ -XXX,XX +XXX,XX @@ qtests_generic = [
   'test-hmp',
   'qos-test',
   'readconfig-test',
+  'netdev-socket',
 ]
 if config_host.has_key('CONFIG_MODULES')
   qtests_generic += [ 'modules-test' ]
@@ -XXX,XX +XXX,XX @@ qtests = {
   'tpm-tis-device-swtpm-test': [io, tpmemu_files, 'tpm-tis-util.c'],
   'tpm-tis-device-test': [io, tpmemu_files, 'tpm-tis-util.c'],
   'vmgenid-test': files('boot-sector.c', 'acpi-utils.c'),
+  'netdev-socket': files('netdev-socket.c', '../unit/socket-helpers.c'),
 }
 
 gvnc = dependency('gvnc-1.0', required: false)
diff --git a/tests/qtest/netdev-socket.c b/tests/qtest/netdev-socket.c
new file mode 100644
index XXXXXXX..XXXXXXX
--- /dev/null
+++ b/tests/qtest/netdev-socket.c
@@ -XXX,XX +XXX,XX @@
+/*
+ * QTest testcase for netdev stream and dgram
+ *
+ * Copyright (c) 2022 Red Hat, Inc.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include "qemu/osdep.h"
+#include "qemu/sockets.h"
+#include <glib/gstdio.h>
+#include "../unit/socket-helpers.h"
+#include "libqtest.h"
+
+#define CONNECTION_TIMEOUT    5
+
+#define EXPECT_STATE(q, e, t)                             \
+do {                                                      \
+    char *resp = NULL;                                    \
+    g_test_timer_start();                                 \
+    do {                                                  \
+        g_free(resp);                                     \
+        resp = qtest_hmp(q, "info network");              \
+        if (t) {                                          \
+            strrchr(resp, t)[0] = 0;                      \
+        }                                                 \
+        if (g_str_equal(resp, e)) {                       \
+            break;                                        \
+        }                                                 \
+    } while (g_test_timer_elapsed() < CONNECTION_TIMEOUT); \
+    g_assert_cmpstr(resp, ==, e);                         \
+    g_free(resp);                                         \
+} while (0)
+
+static gchar *tmpdir;
+
+static int inet_get_free_port_socket_ipv4(int sock)
+{
+    struct sockaddr_in addr;
+    socklen_t len;
+
+    memset(&addr, 0, sizeof(addr));
+    addr.sin_family = AF_INET;
+    addr.sin_addr.s_addr = INADDR_ANY;
+    addr.sin_port = 0;
+    if (bind(sock, (struct sockaddr *)&addr, sizeof(addr)) < 0) {
+        return -1;
+    }
+
+    len = sizeof(addr);
+    if (getsockname(sock,  (struct sockaddr *)&addr, &len) < 0) {
+        return -1;
+    }
+
+    return ntohs(addr.sin_port);
+}
+
+static int inet_get_free_port_socket_ipv6(int sock)
+{
+    struct sockaddr_in6 addr;
+    socklen_t len;
+
+    memset(&addr, 0, sizeof(addr));
+    addr.sin6_family = AF_INET6;
+    addr.sin6_addr = in6addr_any;
+    addr.sin6_port = 0;
+    if (bind(sock, (struct sockaddr *)&addr, sizeof(addr)) < 0) {
+        return -1;
+    }
+
+    len = sizeof(addr);
+    if (getsockname(sock,  (struct sockaddr *)&addr, &len) < 0) {
+        return -1;
+    }
+
+    return ntohs(addr.sin6_port);
+}
+
+static int inet_get_free_port_multiple(int nb, int *port, bool ipv6)
+{
+    int sock[nb];
+    int i;
+
+    for (i = 0; i < nb; i++) {
+        sock[i] = socket(ipv6 ? AF_INET6 : AF_INET, SOCK_STREAM, 0);
+        if (sock[i] < 0) {
+            break;
+        }
+        port[i] = ipv6 ? inet_get_free_port_socket_ipv6(sock[i]) :
+                         inet_get_free_port_socket_ipv4(sock[i]);
+        if (port[i] == -1) {
+            break;
+        }
+    }
+
+    nb = i;
+    for (i = 0; i < nb; i++) {
+        closesocket(sock[i]);
+    }
+
+    return nb;
+}
+
+static int inet_get_free_port(bool ipv6)
+{
+    int nb, port;
+
+    nb = inet_get_free_port_multiple(1, &port, ipv6);
+    g_assert_cmpint(nb, ==, 1);
+
+    return port;
+}
+
+static void test_stream_inet_ipv4(void)
+{
+    QTestState *qts0, *qts1;
+    char *expect;
+    int port;
+
+    port = inet_get_free_port(false);
+    qts0 = qtest_initf("-nodefaults -M none "
+                       "-netdev stream,id=st0,server=true,addr.type=inet,"
+                       "addr.ipv4=on,addr.ipv6=off,"
+                       "addr.host=127.0.0.1,addr.port=%d", port);
+
+    EXPECT_STATE(qts0, "st0: index=0,type=stream,\r\n", 0);
+
+    qts1 = qtest_initf("-nodefaults -M none "
+                       "-netdev stream,server=false,id=st0,addr.type=inet,"
+                       "addr.ipv4=on,addr.ipv6=off,"
+                       "addr.host=127.0.0.1,addr.port=%d", port);
+
+    expect = g_strdup_printf("st0: index=0,type=stream,tcp:127.0.0.1:%d\r\n",
+                             port);
+    EXPECT_STATE(qts1, expect, 0);
+    g_free(expect);
+
+    /* the port is unknown, check only the address */
+    EXPECT_STATE(qts0, "st0: index=0,type=stream,tcp:127.0.0.1", ':');
+
+    qtest_quit(qts1);
+    qtest_quit(qts0);
+}
+
+static void test_stream_inet_ipv6(void)
+{
+    QTestState *qts0, *qts1;
+    char *expect;
+    int port;
+
+    port = inet_get_free_port(true);
+    qts0 = qtest_initf("-nodefaults -M none "
+                       "-netdev stream,id=st0,server=true,addr.type=inet,"
+                       "addr.ipv4=off,addr.ipv6=on,"
+                       "addr.host=::1,addr.port=%d", port);
+
+    EXPECT_STATE(qts0, "st0: index=0,type=stream,\r\n", 0);
+
+    qts1 = qtest_initf("-nodefaults -M none "
+                       "-netdev stream,server=false,id=st0,addr.type=inet,"
+                       "addr.ipv4=off,addr.ipv6=on,"
+                       "addr.host=::1,addr.port=%d", port);
+
+    expect = g_strdup_printf("st0: index=0,type=stream,tcp:::1:%d\r\n",
+                             port);
+    EXPECT_STATE(qts1, expect, 0);
+    g_free(expect);
+
+    /* the port is unknown, check only the address */
+    EXPECT_STATE(qts0, "st0: index=0,type=stream,tcp:::1", ':');
+
+    qtest_quit(qts1);
+    qtest_quit(qts0);
+}
+
+static void test_stream_unix(void)
+{
+    QTestState *qts0, *qts1;
+    char *expect;
+    gchar *path;
+
+    path = g_strconcat(tmpdir, "/stream_unix", NULL);
+
+    qts0 = qtest_initf("-nodefaults -M none "
+                       "-netdev stream,id=st0,server=true,"
+                       "addr.type=unix,addr.path=%s,",
+                       path);
+
+    EXPECT_STATE(qts0, "st0: index=0,type=stream,\r\n", 0);
+
+    qts1 = qtest_initf("-nodefaults -M none "
+                       "-netdev stream,id=st0,server=false,"
+                       "addr.type=unix,addr.path=%s",
+                       path);
+
+    expect = g_strdup_printf("st0: index=0,type=stream,unix:%s\r\n", path);
+    EXPECT_STATE(qts1, expect, 0);
+    EXPECT_STATE(qts0, expect, 0);
+    g_free(expect);
+    g_free(path);
+
+    qtest_quit(qts1);
+    qtest_quit(qts0);
+}
+
+#ifdef CONFIG_LINUX
+static void test_stream_unix_abstract(void)
+{
+    QTestState *qts0, *qts1;
+    char *expect;
+    gchar *path;
+
+    path = g_strconcat(tmpdir, "/stream_unix_abstract", NULL);
+
+    qts0 = qtest_initf("-nodefaults -M none "
+                       "-netdev stream,id=st0,server=true,"
+                       "addr.type=unix,addr.path=%s,"
+                       "addr.abstract=on",
+                       path);
+
+    EXPECT_STATE(qts0, "st0: index=0,type=stream,\r\n", 0);
+
+    qts1 = qtest_initf("-nodefaults -M none "
+                       "-netdev stream,id=st0,server=false,"
+                       "addr.type=unix,addr.path=%s,addr.abstract=on",
+                       path);
+
+    expect = g_strdup_printf("st0: index=0,type=stream,unix:%s\r\n", path);
+    EXPECT_STATE(qts1, expect, 0);
+    EXPECT_STATE(qts0, expect, 0);
+    g_free(expect);
+    g_free(path);
+
+    qtest_quit(qts1);
+    qtest_quit(qts0);
+}
+#endif
+
+#ifndef _WIN32
+static void test_stream_fd(void)
+{
+    QTestState *qts0, *qts1;
+    int sock[2];
+    int ret;
+
+    ret = socketpair(AF_LOCAL, SOCK_STREAM, 0, sock);
+    g_assert_true(ret == 0);
+
+    qts0 = qtest_initf("-nodefaults -M none "
+                       "-netdev stream,id=st0,addr.type=fd,addr.str=%d",
+                       sock[0]);
+
+    EXPECT_STATE(qts0, "st0: index=0,type=stream,unix:\r\n", 0);
+
+    qts1 = qtest_initf("-nodefaults -M none "
+                       "-netdev stream,id=st0,addr.type=fd,addr.str=%d",
+                       sock[1]);
+
+    EXPECT_STATE(qts1, "st0: index=0,type=stream,unix:\r\n", 0);
+    EXPECT_STATE(qts0, "st0: index=0,type=stream,unix:\r\n", 0);
+
+    qtest_quit(qts1);
+    qtest_quit(qts0);
+
+    closesocket(sock[0]);
+    closesocket(sock[1]);
+}
+#endif
+
+static void test_dgram_inet(void)
+{
+    QTestState *qts0, *qts1;
+    char *expect;
+    int port[2];
+    int nb;
+
+    nb = inet_get_free_port_multiple(2, port, false);
+    g_assert_cmpint(nb, ==, 2);
+
+    qts0 = qtest_initf("-nodefaults -M none "
+                       "-netdev dgram,id=st0,"
+                       "local.type=inet,local.host=127.0.0.1,local.port=%d,"
+                       "remote.type=inet,remote.host=127.0.0.1,remote.port=%d",
+                        port[0], port[1]);
+
+    expect = g_strdup_printf("st0: index=0,type=dgram,"
+                             "udp=127.0.0.1:%d/127.0.0.1:%d\r\n",
+                             port[0], port[1]);
+    EXPECT_STATE(qts0, expect, 0);
+    g_free(expect);
+
+    qts1 = qtest_initf("-nodefaults -M none "
+                       "-netdev dgram,id=st0,"
+                       "local.type=inet,local.host=127.0.0.1,local.port=%d,"
+                       "remote.type=inet,remote.host=127.0.0.1,remote.port=%d",
+                        port[1], port[0]);
+
+    expect = g_strdup_printf("st0: index=0,type=dgram,"
+                             "udp=127.0.0.1:%d/127.0.0.1:%d\r\n",
+                             port[1], port[0]);
+    EXPECT_STATE(qts1, expect, 0);
+    g_free(expect);
+
+    qtest_quit(qts1);
+    qtest_quit(qts0);
+}
+
+#ifndef _WIN32
+static void test_dgram_mcast(void)
+{
+    QTestState *qts;
+
+    qts = qtest_initf("-nodefaults -M none "
+                      "-netdev dgram,id=st0,"
+                      "remote.type=inet,remote.host=230.0.0.1,remote.port=1234");
+
+    EXPECT_STATE(qts, "st0: index=0,type=dgram,mcast=230.0.0.1:1234\r\n", 0);
+
+    qtest_quit(qts);
+}
+
+static void test_dgram_unix(void)
+{
+    QTestState *qts0, *qts1;
+    char *expect;
+    gchar *path0, *path1;
+
+    path0 = g_strconcat(tmpdir, "/dgram_unix0", NULL);
+    path1 = g_strconcat(tmpdir, "/dgram_unix1", NULL);
+
+    qts0 = qtest_initf("-nodefaults -M none "
+                       "-netdev dgram,id=st0,local.type=unix,local.path=%s,"
+                       "remote.type=unix,remote.path=%s",
+                       path0, path1);
+
+    expect = g_strdup_printf("st0: index=0,type=dgram,udp=%s:%s\r\n",
+                             path0, path1);
+    EXPECT_STATE(qts0, expect, 0);
+    g_free(expect);
+
+    qts1 = qtest_initf("-nodefaults -M none "
+                       "-netdev dgram,id=st0,local.type=unix,local.path=%s,"
+                       "remote.type=unix,remote.path=%s",
+                       path1, path0);
+
+
+    expect = g_strdup_printf("st0: index=0,type=dgram,udp=%s:%s\r\n",
+                             path1, path0);
+    EXPECT_STATE(qts1, expect, 0);
+    g_free(expect);
+
+    unlink(path0);
+    g_free(path0);
+    unlink(path1);
+    g_free(path1);
+
+    qtest_quit(qts1);
+    qtest_quit(qts0);
+}
+
+static void test_dgram_fd(void)
+{
+    QTestState *qts0, *qts1;
+    char *expect;
+    int ret;
+    int sv[2];
+
+    ret = socketpair(PF_UNIX, SOCK_DGRAM, 0, sv);
+    g_assert_cmpint(ret, !=, -1);
+
+    qts0 = qtest_initf("-nodefaults -M none "
+                       "-netdev dgram,id=st0,local.type=fd,local.str=%d",
+                       sv[0]);
+
+    expect = g_strdup_printf("st0: index=0,type=dgram,fd=%d unix\r\n", sv[0]);
+    EXPECT_STATE(qts0, expect, 0);
+    g_free(expect);
+
+    qts1 = qtest_initf("-nodefaults -M none "
+                       "-netdev dgram,id=st0,local.type=fd,local.str=%d",
+                       sv[1]);
+
+
+    expect = g_strdup_printf("st0: index=0,type=dgram,fd=%d unix\r\n", sv[1]);
+    EXPECT_STATE(qts1, expect, 0);
+    g_free(expect);
+
+    qtest_quit(qts1);
+    qtest_quit(qts0);
+
+    closesocket(sv[0]);
+    closesocket(sv[1]);
+}
+#endif
+
+int main(int argc, char **argv)
+{
+    int ret;
+    bool has_ipv4, has_ipv6, has_afunix;
+    g_autoptr(GError) err = NULL;
+
+    socket_init();
+    g_test_init(&argc, &argv, NULL);
+
+    if (socket_check_protocol_support(&has_ipv4, &has_ipv6) < 0) {
+        g_error("socket_check_protocol_support() failed\n");
+    }
+
+    tmpdir = g_dir_make_tmp("netdev-socket.XXXXXX", &err);
+    if (tmpdir == NULL) {
+        g_error("Can't create temporary directory in %s: %s",
+                g_get_tmp_dir(), err->message);
+    }
+
+    if (has_ipv4) {
+        qtest_add_func("/netdev/stream/inet/ipv4", test_stream_inet_ipv4);
+        qtest_add_func("/netdev/dgram/inet", test_dgram_inet);
+#ifndef _WIN32
+        qtest_add_func("/netdev/dgram/mcast", test_dgram_mcast);
+#endif
+    }
+    if (has_ipv6) {
+        qtest_add_func("/netdev/stream/inet/ipv6", test_stream_inet_ipv6);
+    }
+
+    socket_check_afunix_support(&has_afunix);
+    if (has_afunix) {
+#ifndef _WIN32
+        qtest_add_func("/netdev/dgram/unix", test_dgram_unix);
+#endif
+        qtest_add_func("/netdev/stream/unix", test_stream_unix);
+#ifdef CONFIG_LINUX
+        qtest_add_func("/netdev/stream/unix/abstract",
+                       test_stream_unix_abstract);
+#endif
+#ifndef _WIN32
+        qtest_add_func("/netdev/stream/fd", test_stream_fd);
+        qtest_add_func("/netdev/dgram/fd", test_dgram_fd);
+#endif
+    }
+
+    ret = g_test_run();
+
+    g_rmdir(tmpdir);
+    g_free(tmpdir);
+
+    return ret;
+}
-- 
2.7.4

From: Laurent Vivier <lvivier@redhat.com>

In stream mode, if the server shuts down there is currently
no way to reconnect the client to a new server without removing
the NIC device and the netdev backend (or to reboot).

This patch introduces a reconnect option that specifies a delay
to try to reconnect with the same parameters.

Add a new test in qtest to test the reconnect option and the
connect/disconnect events.

Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 net/stream.c                |  53 ++++++++++++++++++++++-
 qapi/net.json               |   7 ++-
 qemu-options.hx             |   6 +--
 tests/qtest/netdev-socket.c | 101 ++++++++++++++++++++++++++++++++++++++++++++
 4 files changed, 162 insertions(+), 5 deletions(-)

diff --git a/net/stream.c b/net/stream.c
index XXXXXXX..XXXXXXX 100644
--- a/net/stream.c
+++ b/net/stream.c
@@ -XXX,XX +XXX,XX @@
 #include "io/channel-socket.h"
 #include "io/net-listener.h"
 #include "qapi/qapi-events-net.h"
+#include "qapi/qapi-visit-sockets.h"
+#include "qapi/clone-visitor.h"
 
 typedef struct NetStreamState {
     NetClientState nc;
@@ -XXX,XX +XXX,XX @@ typedef struct NetStreamState {
     guint ioc_write_tag;
     SocketReadState rs;
     unsigned int send_index;      /* number of bytes sent*/
+    uint32_t reconnect;
+    guint timer_tag;
+    SocketAddress *addr;
 } NetStreamState;
 
 static void net_stream_listen(QIONetListener *listener,
                               QIOChannelSocket *cioc,
                               void *opaque);
+static void net_stream_arm_reconnect(NetStreamState *s);
 
 static gboolean net_stream_writable(QIOChannel *ioc,
                                     GIOCondition condition,
@@ -XXX,XX +XXX,XX @@ static gboolean net_stream_send(QIOChannel *ioc,
         qemu_set_info_str(&s->nc, "%s", "");
 
         qapi_event_send_netdev_stream_disconnected(s->nc.name);
+        net_stream_arm_reconnect(s);
 
         return G_SOURCE_REMOVE;
     }
@@ -XXX,XX +XXX,XX @@ static gboolean net_stream_send(QIOChannel *ioc,
 static void net_stream_cleanup(NetClientState *nc)
 {
     NetStreamState *s = DO_UPCAST(NetStreamState, nc, nc);
+    if (s->timer_tag) {
+        g_source_remove(s->timer_tag);
+        s->timer_tag = 0;
+    }
+    if (s->addr) {
+        qapi_free_SocketAddress(s->addr);
+        s->addr = NULL;
+    }
     if (s->ioc) {
         if (QIO_CHANNEL_SOCKET(s->ioc)->fd != -1) {
             if (s->ioc_read_tag) {
@@ -XXX,XX +XXX,XX @@ static void net_stream_client_connected(QIOTask *task, gpointer opaque)
 error:
     object_unref(OBJECT(s->ioc));
     s->ioc = NULL;
+    net_stream_arm_reconnect(s);
+}
+
+static gboolean net_stream_reconnect(gpointer data)
+{
+    NetStreamState *s = data;
+    QIOChannelSocket *sioc;
+
+    s->timer_tag = 0;
+
+    sioc = qio_channel_socket_new();
+    s->ioc = QIO_CHANNEL(sioc);
+    qio_channel_socket_connect_async(sioc, s->addr,
+                                     net_stream_client_connected, s,
+                                     NULL, NULL);
+    return G_SOURCE_REMOVE;
+}
+
+static void net_stream_arm_reconnect(NetStreamState *s)
+{
+    if (s->reconnect && s->timer_tag == 0) {
+        s->timer_tag = g_timeout_add_seconds(s->reconnect,
+                                             net_stream_reconnect, s);
+    }
 }
 
 static int net_stream_client_init(NetClientState *peer,
                                   const char *model,
                                   const char *name,
                                   SocketAddress *addr,
+                                  uint32_t reconnect,
                                   Error **errp)
 {
     NetStreamState *s;
@@ -XXX,XX +XXX,XX @@ static int net_stream_client_init(NetClientState *peer,
     s->ioc = QIO_CHANNEL(sioc);
     s->nc.link_down = true;
 
+    s->reconnect = reconnect;
+    if (reconnect) {
+        s->addr = QAPI_CLONE(SocketAddress, addr);
+    }
     qio_channel_socket_connect_async(sioc, addr,
                                      net_stream_client_connected, s,
                                      NULL, NULL);
@@ -XXX,XX +XXX,XX @@ int net_init_stream(const Netdev *netdev, const char *name,
     sock = &netdev->u.stream;
 
     if (!sock->has_server || !sock->server) {
-        return net_stream_client_init(peer, "stream", name, sock->addr, errp);
+        return net_stream_client_init(peer, "stream", name, sock->addr,
+                                      sock->has_reconnect ? sock->reconnect : 0,
+                                      errp);
+    }
+    if (sock->has_reconnect) {
+        error_setg(errp, "'reconnect' option is incompatible with "
+                         "socket in server mode");
+        return -1;
     }
     return net_stream_server_init(peer, "stream", name, sock->addr, errp);
 }
diff --git a/qapi/net.json b/qapi/net.json
index XXXXXXX..XXXXXXX 100644
--- a/qapi/net.json
+++ b/qapi/net.json
@@ -XXX,XX +XXX,XX @@
 # @addr: socket address to listen on (server=true)
 #        or connect to (server=false)
 # @server: create server socket (default: false)
+# @reconnect: For a client socket, if a socket is disconnected,
+#             then attempt a reconnect after the given number of seconds.
+#             Setting this to zero disables this function. (default: 0)
+#             (since 8.0)
 #
 # Only SocketAddress types 'unix', 'inet' and 'fd' are supported.
 #
@@ -XXX,XX +XXX,XX @@
 { 'struct': 'NetdevStreamOptions',
   'data': {
     'addr':   'SocketAddress',
-    '*server': 'bool' } }
+    '*server': 'bool',
+    '*reconnect': 'uint32' } }
 
 ##
 # @NetdevDgramOptions:
diff --git a/qemu-options.hx b/qemu-options.hx
index XXXXXXX..XXXXXXX 100644
--- a/qemu-options.hx
+++ b/qemu-options.hx
@@ -XXX,XX +XXX,XX @@ DEF("netdev", HAS_ARG, QEMU_OPTION_netdev,
     "-netdev socket,id=str[,fd=h][,udp=host:port][,localaddr=host:port]\n"
     "                configure a network backend to connect to another network\n"
     "                using an UDP tunnel\n"
-    "-netdev stream,id=str[,server=on|off],addr.type=inet,addr.host=host,addr.port=port[,to=maxport][,numeric=on|off][,keep-alive=on|off][,mptcp=on|off][,addr.ipv4=on|off][,addr.ipv6=on|off]\n"
-    "-netdev stream,id=str[,server=on|off],addr.type=unix,addr.path=path[,abstract=on|off][,tight=on|off]\n"
-    "-netdev stream,id=str[,server=on|off],addr.type=fd,addr.str=file-descriptor\n"
+    "-netdev stream,id=str[,server=on|off],addr.type=inet,addr.host=host,addr.port=port[,to=maxport][,numeric=on|off][,keep-alive=on|off][,mptcp=on|off][,addr.ipv4=on|off][,addr.ipv6=on|off][,reconnect=seconds]\n"
+    "-netdev stream,id=str[,server=on|off],addr.type=unix,addr.path=path[,abstract=on|off][,tight=on|off][,reconnect=seconds]\n"
+    "-netdev stream,id=str[,server=on|off],addr.type=fd,addr.str=file-descriptor[,reconnect=seconds]\n"
     "                configure a network backend to connect to another network\n"
     "                using a socket connection in stream mode.\n"
     "-netdev dgram,id=str,remote.type=inet,remote.host=maddr,remote.port=port[,local.type=inet,local.host=addr]\n"
diff --git a/tests/qtest/netdev-socket.c b/tests/qtest/netdev-socket.c
index XXXXXXX..XXXXXXX 100644
--- a/tests/qtest/netdev-socket.c
+++ b/tests/qtest/netdev-socket.c
@@ -XXX,XX +XXX,XX @@
 #include <glib/gstdio.h>
 #include "../unit/socket-helpers.h"
 #include "libqtest.h"
+#include "qapi/qmp/qstring.h"
+#include "qemu/sockets.h"
+#include "qapi/qobject-input-visitor.h"
+#include "qapi/qapi-visit-sockets.h"
 
 #define CONNECTION_TIMEOUT    5
 
@@ -XXX,XX +XXX,XX @@ static void test_stream_inet_ipv4(void)
     qtest_quit(qts0);
 }
 
+static void wait_stream_connected(QTestState *qts, const char *id,
+                                  SocketAddress **addr)
+{
+    QDict *resp, *data;
+    QString *qstr;
+    QObject *obj;
+    Visitor *v = NULL;
+
+    resp = qtest_qmp_eventwait_ref(qts, "NETDEV_STREAM_CONNECTED");
+    g_assert_nonnull(resp);
+    data = qdict_get_qdict(resp, "data");
+    g_assert_nonnull(data);
+
+    qstr = qobject_to(QString, qdict_get(data, "netdev-id"));
+    g_assert_nonnull(data);
+
+    g_assert(!strcmp(qstring_get_str(qstr), id));
+
+    obj = qdict_get(data, "addr");
+
+    v = qobject_input_visitor_new(obj);
+    visit_type_SocketAddress(v, NULL, addr, NULL);
+    visit_free(v);
+    qobject_unref(resp);
+}
+
+static void wait_stream_disconnected(QTestState *qts, const char *id)
+{
+    QDict *resp, *data;
+    QString *qstr;
+
+    resp = qtest_qmp_eventwait_ref(qts, "NETDEV_STREAM_DISCONNECTED");
+    g_assert_nonnull(resp);
+    data = qdict_get_qdict(resp, "data");
+    g_assert_nonnull(data);
+
+    qstr = qobject_to(QString, qdict_get(data, "netdev-id"));
+    g_assert_nonnull(data);
+
+    g_assert(!strcmp(qstring_get_str(qstr), id));
+    qobject_unref(resp);
+}
+
+static void test_stream_inet_reconnect(void)
+{
+    QTestState *qts0, *qts1;
+    int port;
+    SocketAddress *addr;
+
+    port = inet_get_free_port(false);
+    qts0 = qtest_initf("-nodefaults -M none "
+                       "-netdev stream,id=st0,server=true,addr.type=inet,"
+                       "addr.ipv4=on,addr.ipv6=off,"
+                       "addr.host=127.0.0.1,addr.port=%d", port);
+
+    EXPECT_STATE(qts0, "st0: index=0,type=stream,\r\n", 0);
+
+    qts1 = qtest_initf("-nodefaults -M none "
+                       "-netdev stream,server=false,id=st0,addr.type=inet,"
+                       "addr.ipv4=on,addr.ipv6=off,reconnect=1,"
+                       "addr.host=127.0.0.1,addr.port=%d", port);
+
+    wait_stream_connected(qts0, "st0", &addr);
+    g_assert_cmpint(addr->type, ==, SOCKET_ADDRESS_TYPE_INET);
+    g_assert_cmpstr(addr->u.inet.host, ==, "127.0.0.1");
+    qapi_free_SocketAddress(addr);
+
+    /* kill server */
+    qtest_quit(qts0);
+
+    /* check client has been disconnected */
+    wait_stream_disconnected(qts1, "st0");
+
+    /* restart server */
+    qts0 = qtest_initf("-nodefaults -M none "
+                       "-netdev stream,id=st0,server=true,addr.type=inet,"
+                       "addr.ipv4=on,addr.ipv6=off,"
+                       "addr.host=127.0.0.1,addr.port=%d", port);
+
+    /* wait connection events*/
+    wait_stream_connected(qts0, "st0", &addr);
+    g_assert_cmpint(addr->type, ==, SOCKET_ADDRESS_TYPE_INET);
+    g_assert_cmpstr(addr->u.inet.host, ==, "127.0.0.1");
+    qapi_free_SocketAddress(addr);
+
+    wait_stream_connected(qts1, "st0", &addr);
+    g_assert_cmpint(addr->type, ==, SOCKET_ADDRESS_TYPE_INET);
+    g_assert_cmpstr(addr->u.inet.host, ==, "127.0.0.1");
+    g_assert_cmpint(atoi(addr->u.inet.port), ==, port);
+    qapi_free_SocketAddress(addr);
+
+    qtest_quit(qts1);
+    qtest_quit(qts0);
+}
+
 static void test_stream_inet_ipv6(void)
 {
     QTestState *qts0, *qts1;
@@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv)
 #ifndef _WIN32
         qtest_add_func("/netdev/dgram/mcast", test_dgram_mcast);
 #endif
+        qtest_add_func("/netdev/stream/inet/reconnect",
+                       test_stream_inet_reconnect);
     }
     if (has_ipv6) {
         qtest_add_func("/netdev/stream/inet/ipv6", test_stream_inet_ipv6);
-- 
2.7.4

From: Eugenio Pérez <eperezma@redhat.com>

VHOST_BACKEND_F_IOTLB_ASID is the feature bit, not the bitmask. Since
the device under test also provided VHOST_BACKEND_F_IOTLB_MSG_V2 and
VHOST_BACKEND_F_IOTLB_BATCH, this went unnoticed.

Fixes: c1a1008685 ("vdpa: always start CVQ in SVQ mode if possible")
Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 net/vhost-vdpa.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/net/vhost-vdpa.c b/net/vhost-vdpa.c
index XXXXXXX..XXXXXXX 100644
--- a/net/vhost-vdpa.c
+++ b/net/vhost-vdpa.c
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_net_cvq_start(NetClientState *nc)
             g_strerror(errno), errno);
         return -1;
     }
-    if (!(backend_features & VHOST_BACKEND_F_IOTLB_ASID) ||
+    if (!(backend_features & BIT_ULL(VHOST_BACKEND_F_IOTLB_ASID)) ||
         !vhost_vdpa_net_valid_svq_features(v->dev->features, NULL)) {
         return 0;
     }
-- 
2.7.4

The following changes since commit 352998df1c53b366413690d95b35f76d0721ebed:

Merge tag 'i2c-20220314' of https://github.com/philmd/qemu into staging (2022-03-14 14:39:33 +0000)

are available in the git repository at:

https://github.com/jasowang/qemu.git tags/net-pull-request

for you to fetch changes up to 12a195fa343aae2ead1301ce04727bd0ae25eb15:

vdpa: Expose VHOST_F_LOG_ALL on SVQ (2022-03-15 13:57:44 +0800)

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

Changes since V2:
- fix 32bit build errros

----------------------------------------------------------------
Eugenio Pérez (14):
      vhost: Add VhostShadowVirtqueue
      vhost: Add Shadow VirtQueue kick forwarding capabilities
      vhost: Add Shadow VirtQueue call forwarding capabilities
      vhost: Add vhost_svq_valid_features to shadow vq
      virtio: Add vhost_svq_get_vring_addr
      vdpa: adapt vhost_ops callbacks to svq
      vhost: Shadow virtqueue buffers forwarding
      util: Add iova_tree_alloc_map
      util: add iova_tree_find_iova
      vhost: Add VhostIOVATree
      vdpa: Add custom IOTLB translations to SVQ
      vdpa: Adapt vhost_vdpa_get_vring_base to SVQ
      vdpa: Never set log_base addr if SVQ is enabled
      vdpa: Expose VHOST_F_LOG_ALL on SVQ

Jason Wang (1):
      virtio-net: fix map leaking on error during receive

hw/net/virtio-net.c                |   1 +
 hw/virtio/meson.build              |   2 +-
 hw/virtio/vhost-iova-tree.c        | 110 +++++++
 hw/virtio/vhost-iova-tree.h        |  27 ++
 hw/virtio/vhost-shadow-virtqueue.c | 636 +++++++++++++++++++++++++++++++++++++
 hw/virtio/vhost-shadow-virtqueue.h |  87 +++++
 hw/virtio/vhost-vdpa.c             | 522 +++++++++++++++++++++++++++++-
 include/hw/virtio/vhost-vdpa.h     |   8 +
 include/qemu/iova-tree.h           |  38 ++-
 util/iova-tree.c                   | 170 ++++++++++
 10 files changed, 1584 insertions(+), 17 deletions(-)
 create mode 100644 hw/virtio/vhost-iova-tree.c
 create mode 100644 hw/virtio/vhost-iova-tree.h
 create mode 100644 hw/virtio/vhost-shadow-virtqueue.c
 create mode 100644 hw/virtio/vhost-shadow-virtqueue.h

Commit bedd7e93d0196 ("virtio-net: fix use after unmap/free for sg")
tries to fix the use after free of the sg by caching the virtqueue
elements in an array and unmap them at once after receiving the
packets, But it forgot to unmap the cached elements on error which
will lead to leaking of mapping and other unexpected results.

Fixing this by detaching the cached elements on error. This addresses
CVE-2022-26353.

Reported-by: Victor Tom <vv474172261@gmail.com>
Cc: qemu-stable@nongnu.org
Fixes: CVE-2022-26353
Fixes: bedd7e93d0196 ("virtio-net: fix use after unmap/free for sg")
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 hw/net/virtio-net.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/hw/net/virtio-net.c b/hw/net/virtio-net.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/net/virtio-net.c
+++ b/hw/net/virtio-net.c
@@ -XXX,XX +XXX,XX @@ static ssize_t virtio_net_receive_rcu(NetClientState *nc, const uint8_t *buf,
 
 err:
     for (j = 0; j < i; j++) {
+        virtqueue_detach_element(q->rx_vq, elems[j], lens[j]);
         g_free(elems[j]);
     }
 
-- 
2.7.4

From: Eugenio Pérez <eperezma@redhat.com>

Vhost shadow virtqueue (SVQ) is an intermediate jump for virtqueue
notifications and buffers, allowing qemu to track them. While qemu is
forwarding the buffers and virtqueue changes, it is able to commit the
memory it's being dirtied, the same way regular qemu's VirtIO devices
do.

This commit only exposes basic SVQ allocation and free. Next patches of
the series add functionality like notifications and buffers forwarding.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 hw/virtio/meson.build              |  2 +-
 hw/virtio/vhost-shadow-virtqueue.c | 62 ++++++++++++++++++++++++++++++++++++++
 hw/virtio/vhost-shadow-virtqueue.h | 28 +++++++++++++++++
 3 files changed, 91 insertions(+), 1 deletion(-)
 create mode 100644 hw/virtio/vhost-shadow-virtqueue.c
 create mode 100644 hw/virtio/vhost-shadow-virtqueue.h

diff --git a/hw/virtio/meson.build b/hw/virtio/meson.build
index XXXXXXX..XXXXXXX 100644
--- a/hw/virtio/meson.build
+++ b/hw/virtio/meson.build
@@ -XXX,XX +XXX,XX @@ softmmu_ss.add(when: 'CONFIG_ALL', if_true: files('vhost-stub.c'))
 
 virtio_ss = ss.source_set()
 virtio_ss.add(files('virtio.c'))
-virtio_ss.add(when: 'CONFIG_VHOST', if_true: files('vhost.c', 'vhost-backend.c'))
+virtio_ss.add(when: 'CONFIG_VHOST', if_true: files('vhost.c', 'vhost-backend.c', 'vhost-shadow-virtqueue.c'))
 virtio_ss.add(when: 'CONFIG_VHOST_USER', if_true: files('vhost-user.c'))
 virtio_ss.add(when: 'CONFIG_VHOST_VDPA', if_true: files('vhost-vdpa.c'))
 virtio_ss.add(when: 'CONFIG_VIRTIO_BALLOON', if_true: files('virtio-balloon.c'))
diff --git a/hw/virtio/vhost-shadow-virtqueue.c b/hw/virtio/vhost-shadow-virtqueue.c
new file mode 100644
index XXXXXXX..XXXXXXX
--- /dev/null
+++ b/hw/virtio/vhost-shadow-virtqueue.c
@@ -XXX,XX +XXX,XX @@
+/*
+ * vhost shadow virtqueue
+ *
+ * SPDX-FileCopyrightText: Red Hat, Inc. 2021
+ * SPDX-FileContributor: Author: Eugenio Pérez <eperezma@redhat.com>
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include "qemu/osdep.h"
+#include "hw/virtio/vhost-shadow-virtqueue.h"
+
+#include "qemu/error-report.h"
+
+/**
+ * Creates vhost shadow virtqueue, and instructs the vhost device to use the
+ * shadow methods and file descriptors.
+ *
+ * Returns the new virtqueue or NULL.
+ *
+ * In case of error, reason is reported through error_report.
+ */
+VhostShadowVirtqueue *vhost_svq_new(void)
+{
+    g_autofree VhostShadowVirtqueue *svq = g_new0(VhostShadowVirtqueue, 1);
+    int r;
+
+    r = event_notifier_init(&svq->hdev_kick, 0);
+    if (r != 0) {
+        error_report("Couldn't create kick event notifier: %s (%d)",
+                     g_strerror(errno), errno);
+        goto err_init_hdev_kick;
+    }
+
+    r = event_notifier_init(&svq->hdev_call, 0);
+    if (r != 0) {
+        error_report("Couldn't create call event notifier: %s (%d)",
+                     g_strerror(errno), errno);
+        goto err_init_hdev_call;
+    }
+
+    return g_steal_pointer(&svq);
+
+err_init_hdev_call:
+    event_notifier_cleanup(&svq->hdev_kick);
+
+err_init_hdev_kick:
+    return NULL;
+}
+
+/**
+ * Free the resources of the shadow virtqueue.
+ *
+ * @pvq: gpointer to SVQ so it can be used by autofree functions.
+ */
+void vhost_svq_free(gpointer pvq)
+{
+    VhostShadowVirtqueue *vq = pvq;
+    event_notifier_cleanup(&vq->hdev_kick);
+    event_notifier_cleanup(&vq->hdev_call);
+    g_free(vq);
+}
diff --git a/hw/virtio/vhost-shadow-virtqueue.h b/hw/virtio/vhost-shadow-virtqueue.h
new file mode 100644
index XXXXXXX..XXXXXXX
--- /dev/null
+++ b/hw/virtio/vhost-shadow-virtqueue.h
@@ -XXX,XX +XXX,XX @@
+/*
+ * vhost shadow virtqueue
+ *
+ * SPDX-FileCopyrightText: Red Hat, Inc. 2021
+ * SPDX-FileContributor: Author: Eugenio Pérez <eperezma@redhat.com>
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#ifndef VHOST_SHADOW_VIRTQUEUE_H
+#define VHOST_SHADOW_VIRTQUEUE_H
+
+#include "qemu/event_notifier.h"
+
+/* Shadow virtqueue to relay notifications */
+typedef struct VhostShadowVirtqueue {
+    /* Shadow kick notifier, sent to vhost */
+    EventNotifier hdev_kick;
+    /* Shadow call notifier, sent to vhost */
+    EventNotifier hdev_call;
+} VhostShadowVirtqueue;
+
+VhostShadowVirtqueue *vhost_svq_new(void);
+
+void vhost_svq_free(gpointer vq);
+G_DEFINE_AUTOPTR_CLEANUP_FUNC(VhostShadowVirtqueue, vhost_svq_free);
+
+#endif
-- 
2.7.4

From: Eugenio Pérez <eperezma@redhat.com>

At this mode no buffer forwarding will be performed in SVQ mode: Qemu
will just forward the guest's kicks to the device.

Host memory notifiers regions are left out for simplicity, and they will
not be addressed in this series.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 hw/virtio/vhost-shadow-virtqueue.c |  55 ++++++++++++++
 hw/virtio/vhost-shadow-virtqueue.h |  14 ++++
 hw/virtio/vhost-vdpa.c             | 144 ++++++++++++++++++++++++++++++++++++-
 include/hw/virtio/vhost-vdpa.h     |   4 ++
 4 files changed, 215 insertions(+), 2 deletions(-)

From: Eugenio Pérez <eperezma@redhat.com>

This will make qemu aware of the device used buffers, allowing it to
write the guest memory with its contents if needed.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 hw/virtio/vhost-shadow-virtqueue.c | 38 ++++++++++++++++++++++++++++++++++++++
 hw/virtio/vhost-shadow-virtqueue.h |  4 ++++
 hw/virtio/vhost-vdpa.c             | 31 +++++++++++++++++++++++++++++--
 3 files changed, 71 insertions(+), 2 deletions(-)

diff --git a/hw/virtio/vhost-shadow-virtqueue.c b/hw/virtio/vhost-shadow-virtqueue.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/virtio/vhost-shadow-virtqueue.c
+++ b/hw/virtio/vhost-shadow-virtqueue.c
@@ -XXX,XX +XXX,XX @@ static void vhost_handle_guest_kick(EventNotifier *n)
 }
 
 /**
+ * Forward vhost notifications
+ *
+ * @n: hdev call event notifier, the one that device set to notify svq.
+ */
+static void vhost_svq_handle_call(EventNotifier *n)
+{
+    VhostShadowVirtqueue *svq = container_of(n, VhostShadowVirtqueue,
+                                             hdev_call);
+    event_notifier_test_and_clear(n);
+    event_notifier_set(&svq->svq_call);
+}
+
+/**
+ * Set the call notifier for the SVQ to call the guest
+ *
+ * @svq: Shadow virtqueue
+ * @call_fd: call notifier
+ *
+ * Called on BQL context.
+ */
+void vhost_svq_set_svq_call_fd(VhostShadowVirtqueue *svq, int call_fd)
+{
+    if (call_fd == VHOST_FILE_UNBIND) {
+        /*
+         * Fail event_notifier_set if called handling device call.
+         *
+         * SVQ still needs device notifications, since it needs to keep
+         * forwarding used buffers even with the unbind.
+         */
+        memset(&svq->svq_call, 0, sizeof(svq->svq_call));
+    } else {
+        event_notifier_init_fd(&svq->svq_call, call_fd);
+    }
+}
+
+/**
  * Set a new file descriptor for the guest to kick the SVQ and notify for avail
  *
  * @svq: The svq
@@ -XXX,XX +XXX,XX @@ VhostShadowVirtqueue *vhost_svq_new(void)
     }
 
     event_notifier_init_fd(&svq->svq_kick, VHOST_FILE_UNBIND);
+    event_notifier_set_handler(&svq->hdev_call, vhost_svq_handle_call);
     return g_steal_pointer(&svq);
 
 err_init_hdev_call:
@@ -XXX,XX +XXX,XX @@ void vhost_svq_free(gpointer pvq)
     VhostShadowVirtqueue *vq = pvq;
     vhost_svq_stop(vq);
     event_notifier_cleanup(&vq->hdev_kick);
+    event_notifier_set_handler(&vq->hdev_call, NULL);
     event_notifier_cleanup(&vq->hdev_call);
     g_free(vq);
 }
diff --git a/hw/virtio/vhost-shadow-virtqueue.h b/hw/virtio/vhost-shadow-virtqueue.h
index XXXXXXX..XXXXXXX 100644
--- a/hw/virtio/vhost-shadow-virtqueue.h
+++ b/hw/virtio/vhost-shadow-virtqueue.h
@@ -XXX,XX +XXX,XX @@ typedef struct VhostShadowVirtqueue {
      * So shadow virtqueue must not clean it, or we would lose VirtQueue one.
      */
     EventNotifier svq_kick;
+
+    /* Guest's call notifier, where the SVQ calls guest. */
+    EventNotifier svq_call;
 } VhostShadowVirtqueue;
 
 void vhost_svq_set_svq_kick_fd(VhostShadowVirtqueue *svq, int svq_kick_fd);
+void vhost_svq_set_svq_call_fd(VhostShadowVirtqueue *svq, int call_fd);
 
 void vhost_svq_stop(VhostShadowVirtqueue *svq);
 
diff --git a/hw/virtio/vhost-vdpa.c b/hw/virtio/vhost-vdpa.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/virtio/vhost-vdpa.c
+++ b/hw/virtio/vhost-vdpa.c
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_vring_dev_kick(struct vhost_dev *dev,
     return vhost_vdpa_call(dev, VHOST_SET_VRING_KICK, file);
 }
 
+static int vhost_vdpa_set_vring_dev_call(struct vhost_dev *dev,
+                                         struct vhost_vring_file *file)
+{
+    trace_vhost_vdpa_set_vring_call(dev, file->index, file->fd);
+    return vhost_vdpa_call(dev, VHOST_SET_VRING_CALL, file);
+}
+
 /**
  * Set the shadow virtqueue descriptors to the device
  *
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_vring_dev_kick(struct vhost_dev *dev,
  * @svq: The shadow virtqueue
  * @idx: The index of the virtqueue in the vhost device
  * @errp: Error
+ *
+ * Note that this function does not rewind kick file descriptor if cannot set
+ * call one.
  */
 static bool vhost_vdpa_svq_setup(struct vhost_dev *dev,
                                  VhostShadowVirtqueue *svq, unsigned idx,
@@ -XXX,XX +XXX,XX @@ static bool vhost_vdpa_svq_setup(struct vhost_dev *dev,
     r = vhost_vdpa_set_vring_dev_kick(dev, &file);
     if (unlikely(r != 0)) {
         error_setg_errno(errp, -r, "Can't set device kick fd");
+        return false;
+    }
+
+    event_notifier = &svq->hdev_call;
+    file.fd = event_notifier_get_fd(event_notifier);
+    r = vhost_vdpa_set_vring_dev_call(dev, &file);
+    if (unlikely(r != 0)) {
+        error_setg_errno(errp, -r, "Can't set device call fd");
     }
 
     return r == 0;
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_vring_kick(struct vhost_dev *dev,
 static int vhost_vdpa_set_vring_call(struct vhost_dev *dev,
                                        struct vhost_vring_file *file)
 {
-    trace_vhost_vdpa_set_vring_call(dev, file->index, file->fd);
-    return vhost_vdpa_call(dev, VHOST_SET_VRING_CALL, file);
+    struct vhost_vdpa *v = dev->opaque;
+
+    if (v->shadow_vqs_enabled) {
+        int vdpa_idx = file->index - dev->vq_index;
+        VhostShadowVirtqueue *svq = g_ptr_array_index(v->shadow_vqs, vdpa_idx);
+
+        vhost_svq_set_svq_call_fd(svq, file->fd);
+        return 0;
+    } else {
+        return vhost_vdpa_set_vring_dev_call(dev, file);
+    }
 }
 
 static int vhost_vdpa_get_features(struct vhost_dev *dev,
-- 
2.7.4

From: Eugenio Pérez <eperezma@redhat.com>

This allows SVQ to negotiate features with the guest and the device. For
the device, SVQ is a driver. While this function bypasses all
non-transport features, it needs to disable the features that SVQ does
not support when forwarding buffers. This includes packed vq layout,
indirect descriptors or event idx.

Future changes can add support to offer more features to the guest,
since the use of VirtQueue gives this for free. This is left out at the
moment for simplicity.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 hw/virtio/vhost-shadow-virtqueue.c | 44 ++++++++++++++++++++++++++++++++++++++
 hw/virtio/vhost-shadow-virtqueue.h |  2 ++
 hw/virtio/vhost-vdpa.c             | 15 +++++++++++++
 3 files changed, 61 insertions(+)

From: Eugenio Pérez <eperezma@redhat.com>

It reports the shadow virtqueue address from qemu virtual address space.

Since this will be different from the guest's vaddr, but the device can
access it, SVQ takes special care about its alignment & lack of garbage
data. It assumes that IOMMU will work in host_page_size ranges for that.

diff --git a/hw/virtio/vhost-shadow-virtqueue.c b/hw/virtio/vhost-shadow-virtqueue.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/virtio/vhost-shadow-virtqueue.c
+++ b/hw/virtio/vhost-shadow-virtqueue.c
@@ -XXX,XX +XXX,XX @@ void vhost_svq_set_svq_call_fd(VhostShadowVirtqueue *svq, int call_fd)
 }
 
 /**
+ * Get the shadow vq vring address.
+ * @svq: Shadow virtqueue
+ * @addr: Destination to store address
+ */
+void vhost_svq_get_vring_addr(const VhostShadowVirtqueue *svq,
+                              struct vhost_vring_addr *addr)
+{
+    addr->desc_user_addr = (uint64_t)(intptr_t)svq->vring.desc;
+    addr->avail_user_addr = (uint64_t)(intptr_t)svq->vring.avail;
+    addr->used_user_addr = (uint64_t)(intptr_t)svq->vring.used;
+}
+
+size_t vhost_svq_driver_area_size(const VhostShadowVirtqueue *svq)
+{
+    size_t desc_size = sizeof(vring_desc_t) * svq->vring.num;
+    size_t avail_size = offsetof(vring_avail_t, ring) +
+                                             sizeof(uint16_t) * svq->vring.num;
+
+    return ROUND_UP(desc_size + avail_size, qemu_real_host_page_size);
+}
+
+size_t vhost_svq_device_area_size(const VhostShadowVirtqueue *svq)
+{
+    size_t used_size = offsetof(vring_used_t, ring) +
+                                    sizeof(vring_used_elem_t) * svq->vring.num;
+    return ROUND_UP(used_size, qemu_real_host_page_size);
+}
+
+/**
  * Set a new file descriptor for the guest to kick the SVQ and notify for avail
  *
  * @svq: The svq
diff --git a/hw/virtio/vhost-shadow-virtqueue.h b/hw/virtio/vhost-shadow-virtqueue.h
index XXXXXXX..XXXXXXX 100644
--- a/hw/virtio/vhost-shadow-virtqueue.h
+++ b/hw/virtio/vhost-shadow-virtqueue.h
@@ -XXX,XX +XXX,XX @@
 #define VHOST_SHADOW_VIRTQUEUE_H
 
 #include "qemu/event_notifier.h"
+#include "hw/virtio/virtio.h"
+#include "standard-headers/linux/vhost_types.h"
 
 /* Shadow virtqueue to relay notifications */
 typedef struct VhostShadowVirtqueue {
+    /* Shadow vring */
+    struct vring vring;
+
     /* Shadow kick notifier, sent to vhost */
     EventNotifier hdev_kick;
     /* Shadow call notifier, sent to vhost */
@@ -XXX,XX +XXX,XX @@ bool vhost_svq_valid_features(uint64_t features, Error **errp);
 
 void vhost_svq_set_svq_kick_fd(VhostShadowVirtqueue *svq, int svq_kick_fd);
 void vhost_svq_set_svq_call_fd(VhostShadowVirtqueue *svq, int call_fd);
+void vhost_svq_get_vring_addr(const VhostShadowVirtqueue *svq,
+                              struct vhost_vring_addr *addr);
+size_t vhost_svq_driver_area_size(const VhostShadowVirtqueue *svq);
+size_t vhost_svq_device_area_size(const VhostShadowVirtqueue *svq);
 
 void vhost_svq_stop(VhostShadowVirtqueue *svq);
 
-- 
2.7.4

From: Eugenio Pérez <eperezma@redhat.com>

First half of the buffers forwarding part, preparing vhost-vdpa
callbacks to SVQ to offer it. QEMU cannot enable it at this moment, so
this is effectively dead code at the moment, but it helps to reduce
patch size.

diff --git a/hw/virtio/vhost-vdpa.c b/hw/virtio/vhost-vdpa.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/virtio/vhost-vdpa.c
+++ b/hw/virtio/vhost-vdpa.c
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_get_config(struct vhost_dev *dev, uint8_t *config,
     return ret;
  }
 
+static int vhost_vdpa_set_dev_vring_base(struct vhost_dev *dev,
+                                         struct vhost_vring_state *ring)
+{
+    trace_vhost_vdpa_set_vring_base(dev, ring->index, ring->num);
+    return vhost_vdpa_call(dev, VHOST_SET_VRING_BASE, ring);
+}
+
 static int vhost_vdpa_set_vring_dev_kick(struct vhost_dev *dev,
                                          struct vhost_vring_file *file)
 {
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_vring_dev_call(struct vhost_dev *dev,
     return vhost_vdpa_call(dev, VHOST_SET_VRING_CALL, file);
 }
 
+static int vhost_vdpa_set_vring_dev_addr(struct vhost_dev *dev,
+                                         struct vhost_vring_addr *addr)
+{
+    trace_vhost_vdpa_set_vring_addr(dev, addr->index, addr->flags,
+                                addr->desc_user_addr, addr->used_user_addr,
+                                addr->avail_user_addr,
+                                addr->log_guest_addr);
+
+    return vhost_vdpa_call(dev, VHOST_SET_VRING_ADDR, addr);
+
+}
+
 /**
  * Set the shadow virtqueue descriptors to the device
  *
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_log_base(struct vhost_dev *dev, uint64_t base,
 static int vhost_vdpa_set_vring_addr(struct vhost_dev *dev,
                                        struct vhost_vring_addr *addr)
 {
-    trace_vhost_vdpa_set_vring_addr(dev, addr->index, addr->flags,
-                                    addr->desc_user_addr, addr->used_user_addr,
-                                    addr->avail_user_addr,
-                                    addr->log_guest_addr);
-    return vhost_vdpa_call(dev, VHOST_SET_VRING_ADDR, addr);
+    struct vhost_vdpa *v = dev->opaque;
+
+    if (v->shadow_vqs_enabled) {
+        /*
+         * Device vring addr was set at device start. SVQ base is handled by
+         * VirtQueue code.
+         */
+        return 0;
+    }
+
+    return vhost_vdpa_set_vring_dev_addr(dev, addr);
 }
 
 static int vhost_vdpa_set_vring_num(struct vhost_dev *dev,
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_vring_num(struct vhost_dev *dev,
 static int vhost_vdpa_set_vring_base(struct vhost_dev *dev,
                                        struct vhost_vring_state *ring)
 {
-    trace_vhost_vdpa_set_vring_base(dev, ring->index, ring->num);
-    return vhost_vdpa_call(dev, VHOST_SET_VRING_BASE, ring);
+    struct vhost_vdpa *v = dev->opaque;
+
+    if (v->shadow_vqs_enabled) {
+        /*
+         * Device vring base was set at device start. SVQ base is handled by
+         * VirtQueue code.
+         */
+        return 0;
+    }
+
+    return vhost_vdpa_set_dev_vring_base(dev, ring);
 }
 
 static int vhost_vdpa_get_vring_base(struct vhost_dev *dev,
-- 
2.7.4

From: Eugenio Pérez <eperezma@redhat.com>

Initial version of shadow virtqueue that actually forward buffers. There
is no iommu support at the moment, and that will be addressed in future
patches of this series. Since all vhost-vdpa devices use forced IOMMU,
this means that SVQ is not usable at this point of the series on any
device.

For simplicity it only supports modern devices, that expects vring
in little endian, with split ring and no event idx or indirect
descriptors. Support for them will not be added in this series.

It reuses the VirtQueue code for the device part. The driver part is
based on Linux's virtio_ring driver, but with stripped functionality
and optimizations so it's easier to review.

However, forwarding buffers have some particular pieces: One of the most
unexpected ones is that a guest's buffer can expand through more than
one descriptor in SVQ. While this is handled gracefully by qemu's
emulated virtio devices, it may cause unexpected SVQ queue full. This
patch also solves it by checking for this condition at both guest's
kicks and device's calls. The code may be more elegant in the future if
SVQ code runs in its own iocontext.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 hw/virtio/vhost-shadow-virtqueue.c | 352 ++++++++++++++++++++++++++++++++++++-
 hw/virtio/vhost-shadow-virtqueue.h |  26 +++
 hw/virtio/vhost-vdpa.c             | 155 +++++++++++++++-
 3 files changed, 522 insertions(+), 11 deletions(-)

diff --git a/hw/virtio/vhost-shadow-virtqueue.c b/hw/virtio/vhost-shadow-virtqueue.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/virtio/vhost-shadow-virtqueue.c
+++ b/hw/virtio/vhost-shadow-virtqueue.c
@@ -XXX,XX +XXX,XX @@
 #include "qemu/error-report.h"
 #include "qapi/error.h"
 #include "qemu/main-loop.h"
+#include "qemu/log.h"
+#include "qemu/memalign.h"
 #include "linux-headers/linux/vhost.h"
 
 /**
@@ -XXX,XX +XXX,XX @@ bool vhost_svq_valid_features(uint64_t features, Error **errp)
 }
 
 /**
- * Forward guest notifications.
+ * Number of descriptors that the SVQ can make available from the guest.
+ *
+ * @svq: The svq
+ */
+static uint16_t vhost_svq_available_slots(const VhostShadowVirtqueue *svq)
+{
+    return svq->vring.num - (svq->shadow_avail_idx - svq->shadow_used_idx);
+}
+
+static void vhost_vring_write_descs(VhostShadowVirtqueue *svq,
+                                    const struct iovec *iovec, size_t num,
+                                    bool more_descs, bool write)
+{
+    uint16_t i = svq->free_head, last = svq->free_head;
+    unsigned n;
+    uint16_t flags = write ? cpu_to_le16(VRING_DESC_F_WRITE) : 0;
+    vring_desc_t *descs = svq->vring.desc;
+
+    if (num == 0) {
+        return;
+    }
+
+    for (n = 0; n < num; n++) {
+        if (more_descs || (n + 1 < num)) {
+            descs[i].flags = flags | cpu_to_le16(VRING_DESC_F_NEXT);
+        } else {
+            descs[i].flags = flags;
+        }
+        descs[i].addr = cpu_to_le64((hwaddr)(intptr_t)iovec[n].iov_base);
+        descs[i].len = cpu_to_le32(iovec[n].iov_len);
+
+        last = i;
+        i = cpu_to_le16(descs[i].next);
+    }
+
+    svq->free_head = le16_to_cpu(descs[last].next);
+}
+
+static bool vhost_svq_add_split(VhostShadowVirtqueue *svq,
+                                VirtQueueElement *elem, unsigned *head)
+{
+    unsigned avail_idx;
+    vring_avail_t *avail = svq->vring.avail;
+
+    *head = svq->free_head;
+
+    /* We need some descriptors here */
+    if (unlikely(!elem->out_num && !elem->in_num)) {
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "Guest provided element with no descriptors");
+        return false;
+    }
+
+    vhost_vring_write_descs(svq, elem->out_sg, elem->out_num, elem->in_num > 0,
+                            false);
+    vhost_vring_write_descs(svq, elem->in_sg, elem->in_num, false, true);
+
+    /*
+     * Put the entry in the available array (but don't update avail->idx until
+     * they do sync).
+     */
+    avail_idx = svq->shadow_avail_idx & (svq->vring.num - 1);
+    avail->ring[avail_idx] = cpu_to_le16(*head);
+    svq->shadow_avail_idx++;
+
+    /* Update the avail index after write the descriptor */
+    smp_wmb();
+    avail->idx = cpu_to_le16(svq->shadow_avail_idx);
+
+    return true;
+}
+
+static bool vhost_svq_add(VhostShadowVirtqueue *svq, VirtQueueElement *elem)
+{
+    unsigned qemu_head;
+    bool ok = vhost_svq_add_split(svq, elem, &qemu_head);
+    if (unlikely(!ok)) {
+        return false;
+    }
+
+    svq->ring_id_maps[qemu_head] = elem;
+    return true;
+}
+
+static void vhost_svq_kick(VhostShadowVirtqueue *svq)
+{
+    /*
+     * We need to expose the available array entries before checking the used
+     * flags
+     */
+    smp_mb();
+    if (svq->vring.used->flags & VRING_USED_F_NO_NOTIFY) {
+        return;
+    }
+
+    event_notifier_set(&svq->hdev_kick);
+}
+
+/**
+ * Forward available buffers.
+ *
+ * @svq: Shadow VirtQueue
+ *
+ * Note that this function does not guarantee that all guest's available
+ * buffers are available to the device in SVQ avail ring. The guest may have
+ * exposed a GPA / GIOVA contiguous buffer, but it may not be contiguous in
+ * qemu vaddr.
+ *
+ * If that happens, guest's kick notifications will be disabled until the
+ * device uses some buffers.
+ */
+static void vhost_handle_guest_kick(VhostShadowVirtqueue *svq)
+{
+    /* Clear event notifier */
+    event_notifier_test_and_clear(&svq->svq_kick);
+
+    /* Forward to the device as many available buffers as possible */
+    do {
+        virtio_queue_set_notification(svq->vq, false);
+
+        while (true) {
+            VirtQueueElement *elem;
+            bool ok;
+
+            if (svq->next_guest_avail_elem) {
+                elem = g_steal_pointer(&svq->next_guest_avail_elem);
+            } else {
+                elem = virtqueue_pop(svq->vq, sizeof(*elem));
+            }
+
+            if (!elem) {
+                break;
+            }
+
+            if (elem->out_num + elem->in_num > vhost_svq_available_slots(svq)) {
+                /*
+                 * This condition is possible since a contiguous buffer in GPA
+                 * does not imply a contiguous buffer in qemu's VA
+                 * scatter-gather segments. If that happens, the buffer exposed
+                 * to the device needs to be a chain of descriptors at this
+                 * moment.
+                 *
+                 * SVQ cannot hold more available buffers if we are here:
+                 * queue the current guest descriptor and ignore further kicks
+                 * until some elements are used.
+                 */
+                svq->next_guest_avail_elem = elem;
+                return;
+            }
+
+            ok = vhost_svq_add(svq, elem);
+            if (unlikely(!ok)) {
+                /* VQ is broken, just return and ignore any other kicks */
+                return;
+            }
+            vhost_svq_kick(svq);
+        }
+
+        virtio_queue_set_notification(svq->vq, true);
+    } while (!virtio_queue_empty(svq->vq));
+}
+
+/**
+ * Handle guest's kick.
  *
  * @n: guest kick event notifier, the one that guest set to notify svq.
  */
-static void vhost_handle_guest_kick(EventNotifier *n)
+static void vhost_handle_guest_kick_notifier(EventNotifier *n)
 {
     VhostShadowVirtqueue *svq = container_of(n, VhostShadowVirtqueue, svq_kick);
     event_notifier_test_and_clear(n);
-    event_notifier_set(&svq->hdev_kick);
+    vhost_handle_guest_kick(svq);
+}
+
+static bool vhost_svq_more_used(VhostShadowVirtqueue *svq)
+{
+    if (svq->last_used_idx != svq->shadow_used_idx) {
+        return true;
+    }
+
+    svq->shadow_used_idx = cpu_to_le16(svq->vring.used->idx);
+
+    return svq->last_used_idx != svq->shadow_used_idx;
 }
 
 /**
- * Forward vhost notifications
+ * Enable vhost device calls after disable them.
+ *
+ * @svq: The svq
+ *
+ * It returns false if there are pending used buffers from the vhost device,
+ * avoiding the possible races between SVQ checking for more work and enabling
+ * callbacks. True if SVQ used vring has no more pending buffers.
+ */
+static bool vhost_svq_enable_notification(VhostShadowVirtqueue *svq)
+{
+    svq->vring.avail->flags &= ~cpu_to_le16(VRING_AVAIL_F_NO_INTERRUPT);
+    /* Make sure the flag is written before the read of used_idx */
+    smp_mb();
+    return !vhost_svq_more_used(svq);
+}
+
+static void vhost_svq_disable_notification(VhostShadowVirtqueue *svq)
+{
+    svq->vring.avail->flags |= cpu_to_le16(VRING_AVAIL_F_NO_INTERRUPT);
+}
+
+static VirtQueueElement *vhost_svq_get_buf(VhostShadowVirtqueue *svq,
+                                           uint32_t *len)
+{
+    vring_desc_t *descs = svq->vring.desc;
+    const vring_used_t *used = svq->vring.used;
+    vring_used_elem_t used_elem;
+    uint16_t last_used;
+
+    if (!vhost_svq_more_used(svq)) {
+        return NULL;
+    }
+
+    /* Only get used array entries after they have been exposed by dev */
+    smp_rmb();
+    last_used = svq->last_used_idx & (svq->vring.num - 1);
+    used_elem.id = le32_to_cpu(used->ring[last_used].id);
+    used_elem.len = le32_to_cpu(used->ring[last_used].len);
+
+    svq->last_used_idx++;
+    if (unlikely(used_elem.id >= svq->vring.num)) {
+        qemu_log_mask(LOG_GUEST_ERROR, "Device %s says index %u is used",
+                      svq->vdev->name, used_elem.id);
+        return NULL;
+    }
+
+    if (unlikely(!svq->ring_id_maps[used_elem.id])) {
+        qemu_log_mask(LOG_GUEST_ERROR,
+            "Device %s says index %u is used, but it was not available",
+            svq->vdev->name, used_elem.id);
+        return NULL;
+    }
+
+    descs[used_elem.id].next = svq->free_head;
+    svq->free_head = used_elem.id;
+
+    *len = used_elem.len;
+    return g_steal_pointer(&svq->ring_id_maps[used_elem.id]);
+}
+
+static void vhost_svq_flush(VhostShadowVirtqueue *svq,
+                            bool check_for_avail_queue)
+{
+    VirtQueue *vq = svq->vq;
+
+    /* Forward as many used buffers as possible. */
+    do {
+        unsigned i = 0;
+
+        vhost_svq_disable_notification(svq);
+        while (true) {
+            uint32_t len;
+            g_autofree VirtQueueElement *elem = vhost_svq_get_buf(svq, &len);
+            if (!elem) {
+                break;
+            }
+
+            if (unlikely(i >= svq->vring.num)) {
+                qemu_log_mask(LOG_GUEST_ERROR,
+                         "More than %u used buffers obtained in a %u size SVQ",
+                         i, svq->vring.num);
+                virtqueue_fill(vq, elem, len, i);
+                virtqueue_flush(vq, i);
+                return;
+            }
+            virtqueue_fill(vq, elem, len, i++);
+        }
+
+        virtqueue_flush(vq, i);
+        event_notifier_set(&svq->svq_call);
+
+        if (check_for_avail_queue && svq->next_guest_avail_elem) {
+            /*
+             * Avail ring was full when vhost_svq_flush was called, so it's a
+             * good moment to make more descriptors available if possible.
+             */
+            vhost_handle_guest_kick(svq);
+        }
+    } while (!vhost_svq_enable_notification(svq));
+}
+
+/**
+ * Forward used buffers.
  *
  * @n: hdev call event notifier, the one that device set to notify svq.
+ *
+ * Note that we are not making any buffers available in the loop, there is no
+ * way that it runs more than virtqueue size times.
  */
 static void vhost_svq_handle_call(EventNotifier *n)
 {
     VhostShadowVirtqueue *svq = container_of(n, VhostShadowVirtqueue,
                                              hdev_call);
     event_notifier_test_and_clear(n);
-    event_notifier_set(&svq->svq_call);
+    vhost_svq_flush(svq, true);
 }
 
 /**
@@ -XXX,XX +XXX,XX @@ void vhost_svq_set_svq_kick_fd(VhostShadowVirtqueue *svq, int svq_kick_fd)
     if (poll_start) {
         event_notifier_init_fd(svq_kick, svq_kick_fd);
         event_notifier_set(svq_kick);
-        event_notifier_set_handler(svq_kick, vhost_handle_guest_kick);
+        event_notifier_set_handler(svq_kick, vhost_handle_guest_kick_notifier);
+    }
+}
+
+/**
+ * Start the shadow virtqueue operation.
+ *
+ * @svq: Shadow Virtqueue
+ * @vdev: VirtIO device
+ * @vq: Virtqueue to shadow
+ */
+void vhost_svq_start(VhostShadowVirtqueue *svq, VirtIODevice *vdev,
+                     VirtQueue *vq)
+{
+    size_t desc_size, driver_size, device_size;
+
+    svq->next_guest_avail_elem = NULL;
+    svq->shadow_avail_idx = 0;
+    svq->shadow_used_idx = 0;
+    svq->last_used_idx = 0;
+    svq->vdev = vdev;
+    svq->vq = vq;
+
+    svq->vring.num = virtio_queue_get_num(vdev, virtio_get_queue_index(vq));
+    driver_size = vhost_svq_driver_area_size(svq);
+    device_size = vhost_svq_device_area_size(svq);
+    svq->vring.desc = qemu_memalign(qemu_real_host_page_size, driver_size);
+    desc_size = sizeof(vring_desc_t) * svq->vring.num;
+    svq->vring.avail = (void *)((char *)svq->vring.desc + desc_size);
+    memset(svq->vring.desc, 0, driver_size);
+    svq->vring.used = qemu_memalign(qemu_real_host_page_size, device_size);
+    memset(svq->vring.used, 0, device_size);
+    svq->ring_id_maps = g_new0(VirtQueueElement *, svq->vring.num);
+    for (unsigned i = 0; i < svq->vring.num - 1; i++) {
+        svq->vring.desc[i].next = cpu_to_le16(i + 1);
     }
 }
 
@@ -XXX,XX +XXX,XX @@ void vhost_svq_set_svq_kick_fd(VhostShadowVirtqueue *svq, int svq_kick_fd)
 void vhost_svq_stop(VhostShadowVirtqueue *svq)
 {
     event_notifier_set_handler(&svq->svq_kick, NULL);
+    g_autofree VirtQueueElement *next_avail_elem = NULL;
+
+    if (!svq->vq) {
+        return;
+    }
+
+    /* Send all pending used descriptors to guest */
+    vhost_svq_flush(svq, false);
+
+    for (unsigned i = 0; i < svq->vring.num; ++i) {
+        g_autofree VirtQueueElement *elem = NULL;
+        elem = g_steal_pointer(&svq->ring_id_maps[i]);
+        if (elem) {
+            virtqueue_detach_element(svq->vq, elem, 0);
+        }
+    }
+
+    next_avail_elem = g_steal_pointer(&svq->next_guest_avail_elem);
+    if (next_avail_elem) {
+        virtqueue_detach_element(svq->vq, next_avail_elem, 0);
+    }
+    svq->vq = NULL;
+    g_free(svq->ring_id_maps);
+    qemu_vfree(svq->vring.desc);
+    qemu_vfree(svq->vring.used);
 }
 
 /**
diff --git a/hw/virtio/vhost-shadow-virtqueue.h b/hw/virtio/vhost-shadow-virtqueue.h
index XXXXXXX..XXXXXXX 100644
--- a/hw/virtio/vhost-shadow-virtqueue.h
+++ b/hw/virtio/vhost-shadow-virtqueue.h
@@ -XXX,XX +XXX,XX @@ typedef struct VhostShadowVirtqueue {
 
     /* Guest's call notifier, where the SVQ calls guest. */
     EventNotifier svq_call;
+
+    /* Virtio queue shadowing */
+    VirtQueue *vq;
+
+    /* Virtio device */
+    VirtIODevice *vdev;
+
+    /* Map for use the guest's descriptors */
+    VirtQueueElement **ring_id_maps;
+
+    /* Next VirtQueue element that guest made available */
+    VirtQueueElement *next_guest_avail_elem;
+
+    /* Next head to expose to the device */
+    uint16_t shadow_avail_idx;
+
+    /* Next free descriptor */
+    uint16_t free_head;
+
+    /* Last seen used idx */
+    uint16_t shadow_used_idx;
+
+    /* Next head to consume from the device */
+    uint16_t last_used_idx;
 } VhostShadowVirtqueue;
 
 bool vhost_svq_valid_features(uint64_t features, Error **errp);
@@ -XXX,XX +XXX,XX @@ void vhost_svq_get_vring_addr(const VhostShadowVirtqueue *svq,
 size_t vhost_svq_driver_area_size(const VhostShadowVirtqueue *svq);
 size_t vhost_svq_device_area_size(const VhostShadowVirtqueue *svq);
 
+void vhost_svq_start(VhostShadowVirtqueue *svq, VirtIODevice *vdev,
+                     VirtQueue *vq);
 void vhost_svq_stop(VhostShadowVirtqueue *svq);
 
 VhostShadowVirtqueue *vhost_svq_new(void);
diff --git a/hw/virtio/vhost-vdpa.c b/hw/virtio/vhost-vdpa.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/virtio/vhost-vdpa.c
+++ b/hw/virtio/vhost-vdpa.c
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_vring_dev_addr(struct vhost_dev *dev,
  * Note that this function does not rewind kick file descriptor if cannot set
  * call one.
  */
-static bool vhost_vdpa_svq_setup(struct vhost_dev *dev,
-                                 VhostShadowVirtqueue *svq, unsigned idx,
-                                 Error **errp)
+static int vhost_vdpa_svq_set_fds(struct vhost_dev *dev,
+                                  VhostShadowVirtqueue *svq, unsigned idx,
+                                  Error **errp)
 {
     struct vhost_vring_file file = {
         .index = dev->vq_index + idx,
@@ -XXX,XX +XXX,XX @@ static bool vhost_vdpa_svq_setup(struct vhost_dev *dev,
     r = vhost_vdpa_set_vring_dev_kick(dev, &file);
     if (unlikely(r != 0)) {
         error_setg_errno(errp, -r, "Can't set device kick fd");
-        return false;
+        return r;
     }
 
     event_notifier = &svq->hdev_call;
@@ -XXX,XX +XXX,XX @@ static bool vhost_vdpa_svq_setup(struct vhost_dev *dev,
         error_setg_errno(errp, -r, "Can't set device call fd");
     }
 
+    return r;
+}
+
+/**
+ * Unmap a SVQ area in the device
+ */
+static bool vhost_vdpa_svq_unmap_ring(struct vhost_vdpa *v, hwaddr iova,
+                                      hwaddr size)
+{
+    int r;
+
+    size = ROUND_UP(size, qemu_real_host_page_size);
+    r = vhost_vdpa_dma_unmap(v, iova, size);
+    return r == 0;
+}
+
+static bool vhost_vdpa_svq_unmap_rings(struct vhost_dev *dev,
+                                       const VhostShadowVirtqueue *svq)
+{
+    struct vhost_vdpa *v = dev->opaque;
+    struct vhost_vring_addr svq_addr;
+    size_t device_size = vhost_svq_device_area_size(svq);
+    size_t driver_size = vhost_svq_driver_area_size(svq);
+    bool ok;
+
+    vhost_svq_get_vring_addr(svq, &svq_addr);
+
+    ok = vhost_vdpa_svq_unmap_ring(v, svq_addr.desc_user_addr, driver_size);
+    if (unlikely(!ok)) {
+        return false;
+    }
+
+    return vhost_vdpa_svq_unmap_ring(v, svq_addr.used_user_addr, device_size);
+}
+
+/**
+ * Map the shadow virtqueue rings in the device
+ *
+ * @dev: The vhost device
+ * @svq: The shadow virtqueue
+ * @addr: Assigned IOVA addresses
+ * @errp: Error pointer
+ */
+static bool vhost_vdpa_svq_map_rings(struct vhost_dev *dev,
+                                     const VhostShadowVirtqueue *svq,
+                                     struct vhost_vring_addr *addr,
+                                     Error **errp)
+{
+    struct vhost_vdpa *v = dev->opaque;
+    size_t device_size = vhost_svq_device_area_size(svq);
+    size_t driver_size = vhost_svq_driver_area_size(svq);
+    int r;
+
+    ERRP_GUARD();
+    vhost_svq_get_vring_addr(svq, addr);
+
+    r = vhost_vdpa_dma_map(v, addr->desc_user_addr, driver_size,
+                           (void *)(uintptr_t)addr->desc_user_addr, true);
+    if (unlikely(r != 0)) {
+        error_setg_errno(errp, -r, "Cannot create vq driver region: ");
+        return false;
+    }
+
+    r = vhost_vdpa_dma_map(v, addr->used_user_addr, device_size,
+                           (void *)(intptr_t)addr->used_user_addr, false);
+    if (unlikely(r != 0)) {
+        error_setg_errno(errp, -r, "Cannot create vq device region: ");
+    }
+
+    return r == 0;
+}
+
+static bool vhost_vdpa_svq_setup(struct vhost_dev *dev,
+                                 VhostShadowVirtqueue *svq, unsigned idx,
+                                 Error **errp)
+{
+    uint16_t vq_index = dev->vq_index + idx;
+    struct vhost_vring_state s = {
+        .index = vq_index,
+    };
+    int r;
+
+    r = vhost_vdpa_set_dev_vring_base(dev, &s);
+    if (unlikely(r)) {
+        error_setg_errno(errp, -r, "Cannot set vring base");
+        return false;
+    }
+
+    r = vhost_vdpa_svq_set_fds(dev, svq, idx, errp);
     return r == 0;
 }
 
@@ -XXX,XX +XXX,XX @@ static bool vhost_vdpa_svqs_start(struct vhost_dev *dev)
     }
 
     for (i = 0; i < v->shadow_vqs->len; ++i) {
+        VirtQueue *vq = virtio_get_queue(dev->vdev, dev->vq_index + i);
         VhostShadowVirtqueue *svq = g_ptr_array_index(v->shadow_vqs, i);
+        struct vhost_vring_addr addr = {
+            .index = i,
+        };
+        int r;
         bool ok = vhost_vdpa_svq_setup(dev, svq, i, &err);
         if (unlikely(!ok)) {
-            error_reportf_err(err, "Cannot setup SVQ %u: ", i);
+            goto err;
+        }
+
+        vhost_svq_start(svq, dev->vdev, vq);
+        ok = vhost_vdpa_svq_map_rings(dev, svq, &addr, &err);
+        if (unlikely(!ok)) {
+            goto err_map;
+        }
+
+        /* Override vring GPA set by vhost subsystem */
+        r = vhost_vdpa_set_vring_dev_addr(dev, &addr);
+        if (unlikely(r != 0)) {
+            error_setg_errno(&err, -r, "Cannot set device address");
+            goto err_set_addr;
+        }
+    }
+
+    return true;
+
+err_set_addr:
+    vhost_vdpa_svq_unmap_rings(dev, g_ptr_array_index(v->shadow_vqs, i));
+
+err_map:
+    vhost_svq_stop(g_ptr_array_index(v->shadow_vqs, i));
+
+err:
+    error_reportf_err(err, "Cannot setup SVQ %u: ", i);
+    for (unsigned j = 0; j < i; ++j) {
+        VhostShadowVirtqueue *svq = g_ptr_array_index(v->shadow_vqs, j);
+        vhost_vdpa_svq_unmap_rings(dev, svq);
+        vhost_svq_stop(svq);
+    }
+
+    return false;
+}
+
+static bool vhost_vdpa_svqs_stop(struct vhost_dev *dev)
+{
+    struct vhost_vdpa *v = dev->opaque;
+
+    if (!v->shadow_vqs) {
+        return true;
+    }
+
+    for (unsigned i = 0; i < v->shadow_vqs->len; ++i) {
+        VhostShadowVirtqueue *svq = g_ptr_array_index(v->shadow_vqs, i);
+        bool ok = vhost_vdpa_svq_unmap_rings(dev, svq);
+        if (unlikely(!ok)) {
             return false;
         }
     }
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_dev_start(struct vhost_dev *dev, bool started)
         }
         vhost_vdpa_set_vring_ready(dev);
     } else {
+        ok = vhost_vdpa_svqs_stop(dev);
+        if (unlikely(!ok)) {
+            return -1;
+        }
         vhost_vdpa_host_notifiers_uninit(dev, dev->nvqs);
     }
 
-- 
2.7.4

From: Eugenio Pérez <eperezma@redhat.com>

This iova tree function allows it to look for a hole in allocated
regions and return a totally new translation for a given translated
address.

It's usage is mainly to allow devices to access qemu address space,
remapping guest's one into a new iova space where qemu can add chunks of
addresses.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 include/qemu/iova-tree.h |  18 +++++++
 util/iova-tree.c         | 136 +++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 154 insertions(+)

diff --git a/include/qemu/iova-tree.h b/include/qemu/iova-tree.h
index XXXXXXX..XXXXXXX 100644
--- a/include/qemu/iova-tree.h
+++ b/include/qemu/iova-tree.h
@@ -XXX,XX +XXX,XX @@
 #define  IOVA_OK           (0)
 #define  IOVA_ERR_INVALID  (-1) /* Invalid parameters */
 #define  IOVA_ERR_OVERLAP  (-2) /* IOVA range overlapped */
+#define  IOVA_ERR_NOMEM    (-3) /* Cannot allocate */
 
 typedef struct IOVATree IOVATree;
 typedef struct DMAMap {
@@ -XXX,XX +XXX,XX @@ const DMAMap *iova_tree_find_address(const IOVATree *tree, hwaddr iova);
 void iova_tree_foreach(IOVATree *tree, iova_tree_iterator iterator);
 
 /**
+ * iova_tree_alloc_map:
+ *
+ * @tree: the iova tree to allocate from
+ * @map: the new map (as translated addr & size) to allocate in the iova region
+ * @iova_begin: the minimum address of the allocation
+ * @iova_end: the maximum addressable direction of the allocation
+ *
+ * Allocates a new region of a given size, between iova_min and iova_max.
+ *
+ * Return: Same as iova_tree_insert, but cannot overlap and can return error if
+ * iova tree is out of free contiguous range. The caller gets the assigned iova
+ * in map->iova.
+ */
+int iova_tree_alloc_map(IOVATree *tree, DMAMap *map, hwaddr iova_begin,
+                        hwaddr iova_end);
+
+/**
  * iova_tree_destroy:
  *
  * @tree: the iova tree to destroy
diff --git a/util/iova-tree.c b/util/iova-tree.c
index XXXXXXX..XXXXXXX 100644
--- a/util/iova-tree.c
+++ b/util/iova-tree.c
@@ -XXX,XX +XXX,XX @@ struct IOVATree {
     GTree *tree;
 };
 
+/* Args to pass to iova_tree_alloc foreach function. */
+struct IOVATreeAllocArgs {
+    /* Size of the desired allocation */
+    size_t new_size;
+
+    /* The minimum address allowed in the allocation */
+    hwaddr iova_begin;
+
+    /* Map at the left of the hole, can be NULL if "this" is first one */
+    const DMAMap *prev;
+
+    /* Map at the right of the hole, can be NULL if "prev" is the last one */
+    const DMAMap *this;
+
+    /* If found, we fill in the IOVA here */
+    hwaddr iova_result;
+
+    /* Whether have we found a valid IOVA */
+    bool iova_found;
+};
+
+/**
+ * Iterate args to the next hole
+ *
+ * @args: The alloc arguments
+ * @next: The next mapping in the tree. Can be NULL to signal the last one
+ */
+static void iova_tree_alloc_args_iterate(struct IOVATreeAllocArgs *args,
+                                         const DMAMap *next)
+{
+    args->prev = args->this;
+    args->this = next;
+}
+
 static int iova_tree_compare(gconstpointer a, gconstpointer b, gpointer data)
 {
     const DMAMap *m1 = a, *m2 = b;
@@ -XXX,XX +XXX,XX @@ int iova_tree_remove(IOVATree *tree, const DMAMap *map)
     return IOVA_OK;
 }
 
+/**
+ * Try to find an unallocated IOVA range between prev and this elements.
+ *
+ * @args: Arguments to allocation
+ *
+ * Cases:
+ *
+ * (1) !prev, !this: No entries allocated, always succeed
+ *
+ * (2) !prev, this: We're iterating at the 1st element.
+ *
+ * (3) prev, !this: We're iterating at the last element.
+ *
+ * (4) prev, this: this is the most common case, we'll try to find a hole
+ * between "prev" and "this" mapping.
+ *
+ * Note that this function assumes the last valid iova is HWADDR_MAX, but it
+ * searches linearly so it's easy to discard the result if it's not the case.
+ */
+static void iova_tree_alloc_map_in_hole(struct IOVATreeAllocArgs *args)
+{
+    const DMAMap *prev = args->prev, *this = args->this;
+    uint64_t hole_start, hole_last;
+
+    if (this && this->iova + this->size < args->iova_begin) {
+        return;
+    }
+
+    hole_start = MAX(prev ? prev->iova + prev->size + 1 : 0, args->iova_begin);
+    hole_last = this ? this->iova : HWADDR_MAX;
+
+    if (hole_last - hole_start > args->new_size) {
+        args->iova_result = hole_start;
+        args->iova_found = true;
+    }
+}
+
+/**
+ * Foreach dma node in the tree, compare if there is a hole with its previous
+ * node (or minimum iova address allowed) and the node.
+ *
+ * @key: Node iterating
+ * @value: Node iterating
+ * @pargs: Struct to communicate with the outside world
+ *
+ * Return: false to keep iterating, true if needs break.
+ */
+static gboolean iova_tree_alloc_traverse(gpointer key, gpointer value,
+                                         gpointer pargs)
+{
+    struct IOVATreeAllocArgs *args = pargs;
+    DMAMap *node = value;
+
+    assert(key == value);
+
+    iova_tree_alloc_args_iterate(args, node);
+    iova_tree_alloc_map_in_hole(args);
+    return args->iova_found;
+}
+
+int iova_tree_alloc_map(IOVATree *tree, DMAMap *map, hwaddr iova_begin,
+                        hwaddr iova_last)
+{
+    struct IOVATreeAllocArgs args = {
+        .new_size = map->size,
+        .iova_begin = iova_begin,
+    };
+
+    if (unlikely(iova_last < iova_begin)) {
+        return IOVA_ERR_INVALID;
+    }
+
+    /*
+     * Find a valid hole for the mapping
+     *
+     * Assuming low iova_begin, so no need to do a binary search to
+     * locate the first node.
+     *
+     * TODO: Replace all this with g_tree_node_first/next/last when available
+     * (from glib since 2.68). To do it with g_tree_foreach complicates the
+     * code a lot.
+     *
+     */
+    g_tree_foreach(tree->tree, iova_tree_alloc_traverse, &args);
+    if (!args.iova_found) {
+        /*
+         * Either tree is empty or the last hole is still not checked.
+         * g_tree_foreach does not compare (last, iova_last] range, so we check
+         * it here.
+         */
+        iova_tree_alloc_args_iterate(&args, NULL);
+        iova_tree_alloc_map_in_hole(&args);
+    }
+
+    if (!args.iova_found || args.iova_result + map->size > iova_last) {
+        return IOVA_ERR_NOMEM;
+    }
+
+    map->iova = args.iova_result;
+    return iova_tree_insert(tree, map);
+}
+
 void iova_tree_destroy(IOVATree *tree)
 {
     g_tree_destroy(tree->tree);
-- 
2.7.4

From: Eugenio Pérez <eperezma@redhat.com>

This function does the reverse operation of iova_tree_find: To look for
a mapping that match a translated address so we can do the reverse.

This have linear complexity instead of logarithmic, but it supports
overlapping HVA. Future developments could reduce it.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 include/qemu/iova-tree.h | 20 +++++++++++++++++++-
 util/iova-tree.c         | 34 ++++++++++++++++++++++++++++++++++
 2 files changed, 53 insertions(+), 1 deletion(-)

diff --git a/include/qemu/iova-tree.h b/include/qemu/iova-tree.h
index XXXXXXX..XXXXXXX 100644
--- a/include/qemu/iova-tree.h
+++ b/include/qemu/iova-tree.h
@@ -XXX,XX +XXX,XX @@ int iova_tree_remove(IOVATree *tree, const DMAMap *map);
  * @tree: the iova tree to search from
  * @map: the mapping to search
  *
- * Search for a mapping in the iova tree that overlaps with the
+ * Search for a mapping in the iova tree that iova overlaps with the
  * mapping range specified.  Only the first found mapping will be
  * returned.
  *
@@ -XXX,XX +XXX,XX @@ int iova_tree_remove(IOVATree *tree, const DMAMap *map);
 const DMAMap *iova_tree_find(const IOVATree *tree, const DMAMap *map);
 
 /**
+ * iova_tree_find_iova:
+ *
+ * @tree: the iova tree to search from
+ * @map: the mapping to search
+ *
+ * Search for a mapping in the iova tree that translated_addr overlaps with the
+ * mapping range specified.  Only the first found mapping will be
+ * returned.
+ *
+ * Return: DMAMap pointer if found, or NULL if not found.  Note that
+ * the returned DMAMap pointer is maintained internally.  User should
+ * only read the content but never modify or free the content.  Also,
+ * user is responsible to make sure the pointer is valid (say, no
+ * concurrent deletion in progress).
+ */
+const DMAMap *iova_tree_find_iova(const IOVATree *tree, const DMAMap *map);
+
+/**
  * iova_tree_find_address:
  *
  * @tree: the iova tree to search from
diff --git a/util/iova-tree.c b/util/iova-tree.c
index XXXXXXX..XXXXXXX 100644
--- a/util/iova-tree.c
+++ b/util/iova-tree.c
@@ -XXX,XX +XXX,XX @@ struct IOVATreeAllocArgs {
     bool iova_found;
 };
 
+typedef struct IOVATreeFindIOVAArgs {
+    const DMAMap *needle;
+    const DMAMap *result;
+} IOVATreeFindIOVAArgs;
+
 /**
  * Iterate args to the next hole
  *
@@ -XXX,XX +XXX,XX @@ const DMAMap *iova_tree_find(const IOVATree *tree, const DMAMap *map)
     return g_tree_lookup(tree->tree, map);
 }
 
+static gboolean iova_tree_find_address_iterator(gpointer key, gpointer value,
+                                                gpointer data)
+{
+    const DMAMap *map = key;
+    IOVATreeFindIOVAArgs *args = data;
+    const DMAMap *needle;
+
+    g_assert(key == value);
+
+    needle = args->needle;
+    if (map->translated_addr + map->size < needle->translated_addr ||
+        needle->translated_addr + needle->size < map->translated_addr) {
+        return false;
+    }
+
+    args->result = map;
+    return true;
+}
+
+const DMAMap *iova_tree_find_iova(const IOVATree *tree, const DMAMap *map)
+{
+    IOVATreeFindIOVAArgs args = {
+        .needle = map,
+    };
+
+    g_tree_foreach(tree->tree, iova_tree_find_address_iterator, &args);
+    return args.result;
+}
+
 const DMAMap *iova_tree_find_address(const IOVATree *tree, hwaddr iova)
 {
     const DMAMap map = { .iova = iova, .size = 0 };
-- 
2.7.4

From: Eugenio Pérez <eperezma@redhat.com>

This tree is able to look for a translated address from an IOVA address.

At first glance it is similar to util/iova-tree. However, SVQ working on
devices with limited IOVA space need more capabilities, like allocating
IOVA chunks or performing reverse translations (qemu addresses to iova).

The allocation capability, as "assign a free IOVA address to this chunk
of memory in qemu's address space" allows shadow virtqueue to create a
new address space that is not restricted by guest's addressable one, so
we can allocate shadow vqs vrings outside of it.

It duplicates the tree so it can search efficiently in both directions,
and it will signal overlap if iova or the translated address is present
in any tree.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 hw/virtio/meson.build       |   2 +-
 hw/virtio/vhost-iova-tree.c | 110 ++++++++++++++++++++++++++++++++++++++++++++
 hw/virtio/vhost-iova-tree.h |  27 +++++++++++
 3 files changed, 138 insertions(+), 1 deletion(-)
 create mode 100644 hw/virtio/vhost-iova-tree.c
 create mode 100644 hw/virtio/vhost-iova-tree.h

diff --git a/hw/virtio/meson.build b/hw/virtio/meson.build
index XXXXXXX..XXXXXXX 100644
--- a/hw/virtio/meson.build
+++ b/hw/virtio/meson.build
@@ -XXX,XX +XXX,XX @@ softmmu_ss.add(when: 'CONFIG_ALL', if_true: files('vhost-stub.c'))
 
 virtio_ss = ss.source_set()
 virtio_ss.add(files('virtio.c'))
-virtio_ss.add(when: 'CONFIG_VHOST', if_true: files('vhost.c', 'vhost-backend.c', 'vhost-shadow-virtqueue.c'))
+virtio_ss.add(when: 'CONFIG_VHOST', if_true: files('vhost.c', 'vhost-backend.c', 'vhost-shadow-virtqueue.c', 'vhost-iova-tree.c'))
 virtio_ss.add(when: 'CONFIG_VHOST_USER', if_true: files('vhost-user.c'))
 virtio_ss.add(when: 'CONFIG_VHOST_VDPA', if_true: files('vhost-vdpa.c'))
 virtio_ss.add(when: 'CONFIG_VIRTIO_BALLOON', if_true: files('virtio-balloon.c'))
diff --git a/hw/virtio/vhost-iova-tree.c b/hw/virtio/vhost-iova-tree.c
new file mode 100644
index XXXXXXX..XXXXXXX
--- /dev/null
+++ b/hw/virtio/vhost-iova-tree.c
@@ -XXX,XX +XXX,XX @@
+/*
+ * vhost software live migration iova tree
+ *
+ * SPDX-FileCopyrightText: Red Hat, Inc. 2021
+ * SPDX-FileContributor: Author: Eugenio Pérez <eperezma@redhat.com>
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include "qemu/osdep.h"
+#include "qemu/iova-tree.h"
+#include "vhost-iova-tree.h"
+
+#define iova_min_addr qemu_real_host_page_size
+
+/**
+ * VhostIOVATree, able to:
+ * - Translate iova address
+ * - Reverse translate iova address (from translated to iova)
+ * - Allocate IOVA regions for translated range (linear operation)
+ */
+struct VhostIOVATree {
+    /* First addressable iova address in the device */
+    uint64_t iova_first;
+
+    /* Last addressable iova address in the device */
+    uint64_t iova_last;
+
+    /* IOVA address to qemu memory maps. */
+    IOVATree *iova_taddr_map;
+};
+
+/**
+ * Create a new IOVA tree
+ *
+ * Returns the new IOVA tree
+ */
+VhostIOVATree *vhost_iova_tree_new(hwaddr iova_first, hwaddr iova_last)
+{
+    VhostIOVATree *tree = g_new(VhostIOVATree, 1);
+
+    /* Some devices do not like 0 addresses */
+    tree->iova_first = MAX(iova_first, iova_min_addr);
+    tree->iova_last = iova_last;
+
+    tree->iova_taddr_map = iova_tree_new();
+    return tree;
+}
+
+/**
+ * Delete an iova tree
+ */
+void vhost_iova_tree_delete(VhostIOVATree *iova_tree)
+{
+    iova_tree_destroy(iova_tree->iova_taddr_map);
+    g_free(iova_tree);
+}
+
+/**
+ * Find the IOVA address stored from a memory address
+ *
+ * @tree: The iova tree
+ * @map: The map with the memory address
+ *
+ * Return the stored mapping, or NULL if not found.
+ */
+const DMAMap *vhost_iova_tree_find_iova(const VhostIOVATree *tree,
+                                        const DMAMap *map)
+{
+    return iova_tree_find_iova(tree->iova_taddr_map, map);
+}
+
+/**
+ * Allocate a new mapping
+ *
+ * @tree: The iova tree
+ * @map: The iova map
+ *
+ * Returns:
+ * - IOVA_OK if the map fits in the container
+ * - IOVA_ERR_INVALID if the map does not make sense (like size overflow)
+ * - IOVA_ERR_NOMEM if tree cannot allocate more space.
+ *
+ * It returns assignated iova in map->iova if return value is VHOST_DMA_MAP_OK.
+ */
+int vhost_iova_tree_map_alloc(VhostIOVATree *tree, DMAMap *map)
+{
+    /* Some vhost devices do not like addr 0. Skip first page */
+    hwaddr iova_first = tree->iova_first ?: qemu_real_host_page_size;
+
+    if (map->translated_addr + map->size < map->translated_addr ||
+        map->perm == IOMMU_NONE) {
+        return IOVA_ERR_INVALID;
+    }
+
+    /* Allocate a node in IOVA address */
+    return iova_tree_alloc_map(tree->iova_taddr_map, map, iova_first,
+                               tree->iova_last);
+}
+
+/**
+ * Remove existing mappings from iova tree
+ *
+ * @iova_tree: The vhost iova tree
+ * @map: The map to remove
+ */
+void vhost_iova_tree_remove(VhostIOVATree *iova_tree, const DMAMap *map)
+{
+    iova_tree_remove(iova_tree->iova_taddr_map, map);
+}
diff --git a/hw/virtio/vhost-iova-tree.h b/hw/virtio/vhost-iova-tree.h
new file mode 100644
index XXXXXXX..XXXXXXX
--- /dev/null
+++ b/hw/virtio/vhost-iova-tree.h
@@ -XXX,XX +XXX,XX @@
+/*
+ * vhost software live migration iova tree
+ *
+ * SPDX-FileCopyrightText: Red Hat, Inc. 2021
+ * SPDX-FileContributor: Author: Eugenio Pérez <eperezma@redhat.com>
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#ifndef HW_VIRTIO_VHOST_IOVA_TREE_H
+#define HW_VIRTIO_VHOST_IOVA_TREE_H
+
+#include "qemu/iova-tree.h"
+#include "exec/memory.h"
+
+typedef struct VhostIOVATree VhostIOVATree;
+
+VhostIOVATree *vhost_iova_tree_new(uint64_t iova_first, uint64_t iova_last);
+void vhost_iova_tree_delete(VhostIOVATree *iova_tree);
+G_DEFINE_AUTOPTR_CLEANUP_FUNC(VhostIOVATree, vhost_iova_tree_delete);
+
+const DMAMap *vhost_iova_tree_find_iova(const VhostIOVATree *iova_tree,
+                                        const DMAMap *map);
+int vhost_iova_tree_map_alloc(VhostIOVATree *iova_tree, DMAMap *map);
+void vhost_iova_tree_remove(VhostIOVATree *iova_tree, const DMAMap *map);
+
+#endif
-- 
2.7.4

From: Eugenio Pérez <eperezma@redhat.com>

Use translations added in VhostIOVATree in SVQ.

Only introduce usage here, not allocation and deallocation. As with
previous patches, we use the dead code paths of shadow_vqs_enabled to
avoid commiting too many changes at once. These are impossible to take
at the moment.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 hw/virtio/vhost-shadow-virtqueue.c |  86 +++++++++++++++++++++++---
 hw/virtio/vhost-shadow-virtqueue.h |   6 +-
 hw/virtio/vhost-vdpa.c             | 122 +++++++++++++++++++++++++++++++------
 include/hw/virtio/vhost-vdpa.h     |   3 +
 4 files changed, 187 insertions(+), 30 deletions(-)

diff --git a/hw/virtio/vhost-shadow-virtqueue.c b/hw/virtio/vhost-shadow-virtqueue.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/virtio/vhost-shadow-virtqueue.c
+++ b/hw/virtio/vhost-shadow-virtqueue.c
@@ -XXX,XX +XXX,XX @@ static uint16_t vhost_svq_available_slots(const VhostShadowVirtqueue *svq)
     return svq->vring.num - (svq->shadow_avail_idx - svq->shadow_used_idx);
 }
 
-static void vhost_vring_write_descs(VhostShadowVirtqueue *svq,
+/**
+ * Translate addresses between the qemu's virtual address and the SVQ IOVA
+ *
+ * @svq: Shadow VirtQueue
+ * @vaddr: Translated IOVA addresses
+ * @iovec: Source qemu's VA addresses
+ * @num: Length of iovec and minimum length of vaddr
+ */
+static bool vhost_svq_translate_addr(const VhostShadowVirtqueue *svq,
+                                     hwaddr *addrs, const struct iovec *iovec,
+                                     size_t num)
+{
+    if (num == 0) {
+        return true;
+    }
+
+    for (size_t i = 0; i < num; ++i) {
+        DMAMap needle = {
+            .translated_addr = (hwaddr)(uintptr_t)iovec[i].iov_base,
+            .size = iovec[i].iov_len,
+        };
+        Int128 needle_last, map_last;
+        size_t off;
+
+        const DMAMap *map = vhost_iova_tree_find_iova(svq->iova_tree, &needle);
+        /*
+         * Map cannot be NULL since iova map contains all guest space and
+         * qemu already has a physical address mapped
+         */
+        if (unlikely(!map)) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "Invalid address 0x%"HWADDR_PRIx" given by guest",
+                          needle.translated_addr);
+            return false;
+        }
+
+        off = needle.translated_addr - map->translated_addr;
+        addrs[i] = map->iova + off;
+
+        needle_last = int128_add(int128_make64(needle.translated_addr),
+                                 int128_make64(iovec[i].iov_len));
+        map_last = int128_make64(map->translated_addr + map->size);
+        if (unlikely(int128_gt(needle_last, map_last))) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "Guest buffer expands over iova range");
+            return false;
+        }
+    }
+
+    return true;
+}
+
+static void vhost_vring_write_descs(VhostShadowVirtqueue *svq, hwaddr *sg,
                                     const struct iovec *iovec, size_t num,
                                     bool more_descs, bool write)
 {
@@ -XXX,XX +XXX,XX @@ static void vhost_vring_write_descs(VhostShadowVirtqueue *svq,
         } else {
             descs[i].flags = flags;
         }
-        descs[i].addr = cpu_to_le64((hwaddr)(intptr_t)iovec[n].iov_base);
+        descs[i].addr = cpu_to_le64(sg[n]);
         descs[i].len = cpu_to_le32(iovec[n].iov_len);
 
         last = i;
@@ -XXX,XX +XXX,XX @@ static bool vhost_svq_add_split(VhostShadowVirtqueue *svq,
 {
     unsigned avail_idx;
     vring_avail_t *avail = svq->vring.avail;
+    bool ok;
+    g_autofree hwaddr *sgs = g_new(hwaddr, MAX(elem->out_num, elem->in_num));
 
     *head = svq->free_head;
 
@@ -XXX,XX +XXX,XX @@ static bool vhost_svq_add_split(VhostShadowVirtqueue *svq,
         return false;
     }
 
-    vhost_vring_write_descs(svq, elem->out_sg, elem->out_num, elem->in_num > 0,
-                            false);
-    vhost_vring_write_descs(svq, elem->in_sg, elem->in_num, false, true);
+    ok = vhost_svq_translate_addr(svq, sgs, elem->out_sg, elem->out_num);
+    if (unlikely(!ok)) {
+        return false;
+    }
+    vhost_vring_write_descs(svq, sgs, elem->out_sg, elem->out_num,
+                            elem->in_num > 0, false);
+
+
+    ok = vhost_svq_translate_addr(svq, sgs, elem->in_sg, elem->in_num);
+    if (unlikely(!ok)) {
+        return false;
+    }
+
+    vhost_vring_write_descs(svq, sgs, elem->in_sg, elem->in_num, false, true);
 
     /*
      * Put the entry in the available array (but don't update avail->idx until
@@ -XXX,XX +XXX,XX @@ void vhost_svq_set_svq_call_fd(VhostShadowVirtqueue *svq, int call_fd)
 void vhost_svq_get_vring_addr(const VhostShadowVirtqueue *svq,
                               struct vhost_vring_addr *addr)
 {
-    addr->desc_user_addr = (uint64_t)(intptr_t)svq->vring.desc;
-    addr->avail_user_addr = (uint64_t)(intptr_t)svq->vring.avail;
-    addr->used_user_addr = (uint64_t)(intptr_t)svq->vring.used;
+    addr->desc_user_addr = (uint64_t)(uintptr_t)svq->vring.desc;
+    addr->avail_user_addr = (uint64_t)(uintptr_t)svq->vring.avail;
+    addr->used_user_addr = (uint64_t)(uintptr_t)svq->vring.used;
 }
 
 size_t vhost_svq_driver_area_size(const VhostShadowVirtqueue *svq)
@@ -XXX,XX +XXX,XX @@ void vhost_svq_stop(VhostShadowVirtqueue *svq)
  * Creates vhost shadow virtqueue, and instructs the vhost device to use the
  * shadow methods and file descriptors.
  *
+ * @iova_tree: Tree to perform descriptors translations
+ *
  * Returns the new virtqueue or NULL.
  *
  * In case of error, reason is reported through error_report.
  */
-VhostShadowVirtqueue *vhost_svq_new(void)
+VhostShadowVirtqueue *vhost_svq_new(VhostIOVATree *iova_tree)
 {
     g_autofree VhostShadowVirtqueue *svq = g_new0(VhostShadowVirtqueue, 1);
     int r;
@@ -XXX,XX +XXX,XX @@ VhostShadowVirtqueue *vhost_svq_new(void)
 
     event_notifier_init_fd(&svq->svq_kick, VHOST_FILE_UNBIND);
     event_notifier_set_handler(&svq->hdev_call, vhost_svq_handle_call);
+    svq->iova_tree = iova_tree;
     return g_steal_pointer(&svq);
 
 err_init_hdev_call:
diff --git a/hw/virtio/vhost-shadow-virtqueue.h b/hw/virtio/vhost-shadow-virtqueue.h
index XXXXXXX..XXXXXXX 100644
--- a/hw/virtio/vhost-shadow-virtqueue.h
+++ b/hw/virtio/vhost-shadow-virtqueue.h
@@ -XXX,XX +XXX,XX @@
 #include "qemu/event_notifier.h"
 #include "hw/virtio/virtio.h"
 #include "standard-headers/linux/vhost_types.h"
+#include "hw/virtio/vhost-iova-tree.h"
 
 /* Shadow virtqueue to relay notifications */
 typedef struct VhostShadowVirtqueue {
@@ -XXX,XX +XXX,XX @@ typedef struct VhostShadowVirtqueue {
     /* Virtio device */
     VirtIODevice *vdev;
 
+    /* IOVA mapping */
+    VhostIOVATree *iova_tree;
+
     /* Map for use the guest's descriptors */
     VirtQueueElement **ring_id_maps;
 
@@ -XXX,XX +XXX,XX @@ void vhost_svq_start(VhostShadowVirtqueue *svq, VirtIODevice *vdev,
                      VirtQueue *vq);
 void vhost_svq_stop(VhostShadowVirtqueue *svq);
 
-VhostShadowVirtqueue *vhost_svq_new(void);
+VhostShadowVirtqueue *vhost_svq_new(VhostIOVATree *iova_tree);
 
 void vhost_svq_free(gpointer vq);
 G_DEFINE_AUTOPTR_CLEANUP_FUNC(VhostShadowVirtqueue, vhost_svq_free);
diff --git a/hw/virtio/vhost-vdpa.c b/hw/virtio/vhost-vdpa.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/virtio/vhost-vdpa.c
+++ b/hw/virtio/vhost-vdpa.c
@@ -XXX,XX +XXX,XX @@ static void vhost_vdpa_listener_region_add(MemoryListener *listener,
                                          vaddr, section->readonly);
 
     llsize = int128_sub(llend, int128_make64(iova));
+    if (v->shadow_vqs_enabled) {
+        DMAMap mem_region = {
+            .translated_addr = (hwaddr)(uintptr_t)vaddr,
+            .size = int128_get64(llsize) - 1,
+            .perm = IOMMU_ACCESS_FLAG(true, section->readonly),
+        };
+
+        int r = vhost_iova_tree_map_alloc(v->iova_tree, &mem_region);
+        if (unlikely(r != IOVA_OK)) {
+            error_report("Can't allocate a mapping (%d)", r);
+            goto fail;
+        }
+
+        iova = mem_region.iova;
+    }
 
     vhost_vdpa_iotlb_batch_begin_once(v);
     ret = vhost_vdpa_dma_map(v, iova, int128_get64(llsize),
@@ -XXX,XX +XXX,XX @@ static void vhost_vdpa_listener_region_del(MemoryListener *listener,
 
     llsize = int128_sub(llend, int128_make64(iova));
 
+    if (v->shadow_vqs_enabled) {
+        const DMAMap *result;
+        const void *vaddr = memory_region_get_ram_ptr(section->mr) +
+            section->offset_within_region +
+            (iova - section->offset_within_address_space);
+        DMAMap mem_region = {
+            .translated_addr = (hwaddr)(uintptr_t)vaddr,
+            .size = int128_get64(llsize) - 1,
+        };
+
+        result = vhost_iova_tree_find_iova(v->iova_tree, &mem_region);
+        iova = result->iova;
+        vhost_iova_tree_remove(v->iova_tree, &mem_region);
+    }
     vhost_vdpa_iotlb_batch_begin_once(v);
     ret = vhost_vdpa_dma_unmap(v, iova, int128_get64(llsize));
     if (ret) {
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_init_svq(struct vhost_dev *hdev, struct vhost_vdpa *v,
 
     shadow_vqs = g_ptr_array_new_full(hdev->nvqs, vhost_svq_free);
     for (unsigned n = 0; n < hdev->nvqs; ++n) {
-        g_autoptr(VhostShadowVirtqueue) svq = vhost_svq_new();
+        g_autoptr(VhostShadowVirtqueue) svq = vhost_svq_new(v->iova_tree);
 
         if (unlikely(!svq)) {
             error_setg(errp, "Cannot create svq %u", n);
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_svq_set_fds(struct vhost_dev *dev,
 /**
  * Unmap a SVQ area in the device
  */
-static bool vhost_vdpa_svq_unmap_ring(struct vhost_vdpa *v, hwaddr iova,
-                                      hwaddr size)
+static bool vhost_vdpa_svq_unmap_ring(struct vhost_vdpa *v,
+                                      const DMAMap *needle)
 {
+    const DMAMap *result = vhost_iova_tree_find_iova(v->iova_tree, needle);
+    hwaddr size;
     int r;
 
-    size = ROUND_UP(size, qemu_real_host_page_size);
-    r = vhost_vdpa_dma_unmap(v, iova, size);
+    if (unlikely(!result)) {
+        error_report("Unable to find SVQ address to unmap");
+        return false;
+    }
+
+    size = ROUND_UP(result->size, qemu_real_host_page_size);
+    r = vhost_vdpa_dma_unmap(v, result->iova, size);
     return r == 0;
 }
 
 static bool vhost_vdpa_svq_unmap_rings(struct vhost_dev *dev,
                                        const VhostShadowVirtqueue *svq)
 {
+    DMAMap needle = {};
     struct vhost_vdpa *v = dev->opaque;
     struct vhost_vring_addr svq_addr;
-    size_t device_size = vhost_svq_device_area_size(svq);
-    size_t driver_size = vhost_svq_driver_area_size(svq);
     bool ok;
 
     vhost_svq_get_vring_addr(svq, &svq_addr);
 
-    ok = vhost_vdpa_svq_unmap_ring(v, svq_addr.desc_user_addr, driver_size);
+    needle.translated_addr = svq_addr.desc_user_addr;
+    ok = vhost_vdpa_svq_unmap_ring(v, &needle);
     if (unlikely(!ok)) {
         return false;
     }
 
-    return vhost_vdpa_svq_unmap_ring(v, svq_addr.used_user_addr, device_size);
+    needle.translated_addr = svq_addr.used_user_addr;
+    return vhost_vdpa_svq_unmap_ring(v, &needle);
+}
+
+/**
+ * Map the SVQ area in the device
+ *
+ * @v: Vhost-vdpa device
+ * @needle: The area to search iova
+ * @errorp: Error pointer
+ */
+static bool vhost_vdpa_svq_map_ring(struct vhost_vdpa *v, DMAMap *needle,
+                                    Error **errp)
+{
+    int r;
+
+    r = vhost_iova_tree_map_alloc(v->iova_tree, needle);
+    if (unlikely(r != IOVA_OK)) {
+        error_setg(errp, "Cannot allocate iova (%d)", r);
+        return false;
+    }
+
+    r = vhost_vdpa_dma_map(v, needle->iova, needle->size + 1,
+                           (void *)(uintptr_t)needle->translated_addr,
+                           needle->perm == IOMMU_RO);
+    if (unlikely(r != 0)) {
+        error_setg_errno(errp, -r, "Cannot map region to device");
+        vhost_iova_tree_remove(v->iova_tree, needle);
+    }
+
+    return r == 0;
 }
 
 /**
@@ -XXX,XX +XXX,XX @@ static bool vhost_vdpa_svq_map_rings(struct vhost_dev *dev,
                                      struct vhost_vring_addr *addr,
                                      Error **errp)
 {
+    DMAMap device_region, driver_region;
+    struct vhost_vring_addr svq_addr;
     struct vhost_vdpa *v = dev->opaque;
     size_t device_size = vhost_svq_device_area_size(svq);
     size_t driver_size = vhost_svq_driver_area_size(svq);
-    int r;
+    size_t avail_offset;
+    bool ok;
 
     ERRP_GUARD();
-    vhost_svq_get_vring_addr(svq, addr);
+    vhost_svq_get_vring_addr(svq, &svq_addr);
 
-    r = vhost_vdpa_dma_map(v, addr->desc_user_addr, driver_size,
-                           (void *)(uintptr_t)addr->desc_user_addr, true);
-    if (unlikely(r != 0)) {
-        error_setg_errno(errp, -r, "Cannot create vq driver region: ");
+    driver_region = (DMAMap) {
+        .translated_addr = svq_addr.desc_user_addr,
+        .size = driver_size - 1,
+        .perm = IOMMU_RO,
+    };
+    ok = vhost_vdpa_svq_map_ring(v, &driver_region, errp);
+    if (unlikely(!ok)) {
+        error_prepend(errp, "Cannot create vq driver region: ");
         return false;
     }
+    addr->desc_user_addr = driver_region.iova;
+    avail_offset = svq_addr.avail_user_addr - svq_addr.desc_user_addr;
+    addr->avail_user_addr = driver_region.iova + avail_offset;
 
-    r = vhost_vdpa_dma_map(v, addr->used_user_addr, device_size,
-                           (void *)(intptr_t)addr->used_user_addr, false);
-    if (unlikely(r != 0)) {
-        error_setg_errno(errp, -r, "Cannot create vq device region: ");
+    device_region = (DMAMap) {
+        .translated_addr = svq_addr.used_user_addr,
+        .size = device_size - 1,
+        .perm = IOMMU_RW,
+    };
+    ok = vhost_vdpa_svq_map_ring(v, &device_region, errp);
+    if (unlikely(!ok)) {
+        error_prepend(errp, "Cannot create vq device region: ");
+        vhost_vdpa_svq_unmap_ring(v, &driver_region);
     }
+    addr->used_user_addr = device_region.iova;
 
-    return r == 0;
+    return ok;
 }
 
 static bool vhost_vdpa_svq_setup(struct vhost_dev *dev,
diff --git a/include/hw/virtio/vhost-vdpa.h b/include/hw/virtio/vhost-vdpa.h
index XXXXXXX..XXXXXXX 100644
--- a/include/hw/virtio/vhost-vdpa.h
+++ b/include/hw/virtio/vhost-vdpa.h
@@ -XXX,XX +XXX,XX @@
 
 #include <gmodule.h>
 
+#include "hw/virtio/vhost-iova-tree.h"
 #include "hw/virtio/virtio.h"
 #include "standard-headers/linux/vhost_types.h"
 
@@ -XXX,XX +XXX,XX @@ typedef struct vhost_vdpa {
     MemoryListener listener;
     struct vhost_vdpa_iova_range iova_range;
     bool shadow_vqs_enabled;
+    /* IOVA mapping used by the Shadow Virtqueue */
+    VhostIOVATree *iova_tree;
     GPtrArray *shadow_vqs;
     struct vhost_dev *dev;
     VhostVDPAHostNotifier notifier[VIRTIO_QUEUE_MAX];
-- 
2.7.4

From: Eugenio Pérez <eperezma@redhat.com>

This is needed to achieve migration, so the destination can restore its
index.

Setting base as last used idx, so destination will see as available all
the entries that the device did not use, including the in-flight
processing ones.

This is ok for networking, but other kinds of devices might have
problems with these retransmissions.

diff --git a/hw/virtio/vhost-vdpa.c b/hw/virtio/vhost-vdpa.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/virtio/vhost-vdpa.c
+++ b/hw/virtio/vhost-vdpa.c
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_vring_base(struct vhost_dev *dev,
 static int vhost_vdpa_get_vring_base(struct vhost_dev *dev,
                                        struct vhost_vring_state *ring)
 {
+    struct vhost_vdpa *v = dev->opaque;
     int ret;
 
+    if (v->shadow_vqs_enabled) {
+        VhostShadowVirtqueue *svq = g_ptr_array_index(v->shadow_vqs,
+                                                      ring->index);
+
+        /*
+         * Setting base as last used idx, so destination will see as available
+         * all the entries that the device did not use, including the in-flight
+         * processing ones.
+         *
+         * TODO: This is ok for networking, but other kinds of devices might
+         * have problems with these retransmissions.
+         */
+        ring->num = svq->last_used_idx;
+        return 0;
+    }
+
     ret = vhost_vdpa_call(dev, VHOST_GET_VRING_BASE, ring);
     trace_vhost_vdpa_get_vring_base(dev, ring->index, ring->num);
     return ret;
-- 
2.7.4

From: Eugenio Pérez <eperezma@redhat.com>

SVQ is able to log the dirty bits by itself, so let's use it to not
block migration.

Also, ignore set and clear of VHOST_F_LOG_ALL on set_features if SVQ is
enabled. Even if the device supports it, the reports would be nonsense
because SVQ memory is in the qemu region.

The log region is still allocated. Future changes might skip that, but
this series is already long enough.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 hw/virtio/vhost-vdpa.c         | 39 +++++++++++++++++++++++++++++++++++----
 include/hw/virtio/vhost-vdpa.h |  1 +
 2 files changed, 36 insertions(+), 4 deletions(-)

diff --git a/hw/virtio/vhost-vdpa.c b/hw/virtio/vhost-vdpa.c
index XXXXXXX..XXXXXXX 100644
--- a/hw/virtio/vhost-vdpa.c
+++ b/hw/virtio/vhost-vdpa.c
@@ -XXX,XX +XXX,XX @@ static bool vhost_vdpa_one_time_request(struct vhost_dev *dev)
     return v->index != 0;
 }
 
+static int vhost_vdpa_get_dev_features(struct vhost_dev *dev,
+                                       uint64_t *features)
+{
+    int ret;
+
+    ret = vhost_vdpa_call(dev, VHOST_GET_FEATURES, features);
+    trace_vhost_vdpa_get_features(dev, *features);
+    return ret;
+}
+
 static int vhost_vdpa_init_svq(struct vhost_dev *hdev, struct vhost_vdpa *v,
                                Error **errp)
 {
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_init_svq(struct vhost_dev *hdev, struct vhost_vdpa *v,
         return 0;
     }
 
-    r = hdev->vhost_ops->vhost_get_features(hdev, &dev_features);
+    r = vhost_vdpa_get_dev_features(hdev, &dev_features);
     if (r != 0) {
         error_setg_errno(errp, -r, "Can't get vdpa device features");
         return r;
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_mem_table(struct vhost_dev *dev,
 static int vhost_vdpa_set_features(struct vhost_dev *dev,
                                    uint64_t features)
 {
+    struct vhost_vdpa *v = dev->opaque;
     int ret;
 
     if (vhost_vdpa_one_time_request(dev)) {
         return 0;
     }
 
+    if (v->shadow_vqs_enabled) {
+        if ((v->acked_features ^ features) == BIT_ULL(VHOST_F_LOG_ALL)) {
+            /*
+             * QEMU is just trying to enable or disable logging. SVQ handles
+             * this sepparately, so no need to forward this.
+             */
+            v->acked_features = features;
+            return 0;
+        }
+
+        v->acked_features = features;
+
+        /* We must not ack _F_LOG if SVQ is enabled */
+        features &= ~BIT_ULL(VHOST_F_LOG_ALL);
+    }
+
     trace_vhost_vdpa_set_features(dev, features);
     ret = vhost_vdpa_call(dev, VHOST_SET_FEATURES, &features);
     if (ret) {
@@ -XXX,XX +XXX,XX @@ static int vhost_vdpa_set_vring_call(struct vhost_dev *dev,
 static int vhost_vdpa_get_features(struct vhost_dev *dev,
                                      uint64_t *features)
 {
-    int ret;
+    struct vhost_vdpa *v = dev->opaque;
+    int ret = vhost_vdpa_get_dev_features(dev, features);
+
+    if (ret == 0 && v->shadow_vqs_enabled) {
+        /* Add SVQ logging capabilities */
+        *features |= BIT_ULL(VHOST_F_LOG_ALL);
+    }
 
-    ret = vhost_vdpa_call(dev, VHOST_GET_FEATURES, features);
-    trace_vhost_vdpa_get_features(dev, *features);
     return ret;
 }
 
diff --git a/include/hw/virtio/vhost-vdpa.h b/include/hw/virtio/vhost-vdpa.h
index XXXXXXX..XXXXXXX 100644
--- a/include/hw/virtio/vhost-vdpa.h
+++ b/include/hw/virtio/vhost-vdpa.h
@@ -XXX,XX +XXX,XX @@ typedef struct vhost_vdpa {
     bool iotlb_batch_begin_sent;
     MemoryListener listener;
     struct vhost_vdpa_iova_range iova_range;
+    uint64_t acked_features;
     bool shadow_vqs_enabled;
     /* IOVA mapping used by the Shadow Virtqueue */
     VhostIOVATree *iova_tree;
-- 
2.7.4